Biomedical Informatics In Translational Research

1st Edition Hai Hu download

https://ebookbell.com/product/biomedical-informatics-in-
translational-research-1st-edition-hai-hu-2200766

Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

Biomedical Informatics Computer Applications In Health Care And

Biomedicine 5th Ed 2021 Edward H Shortliffe Editor

https://ebookbell.com/product/biomedical-informatics-computer-
applications-in-health-care-and-biomedicine-5th-ed-2021-edward-h-
shortliffe-editor-31291084

Biomedical Informatics Discovering Knowledge In Big Data 2nd Edition

Andreas Holzinger Auth

https://ebookbell.com/product/biomedical-informatics-discovering-
knowledge-in-big-data-2nd-edition-andreas-holzinger-auth-4696742

Biomedical Informatics Computer Applications In Health Care And

Biomedicine 4th Edition Edward H Shortliffe

https://ebookbell.com/product/biomedical-informatics-computer-
applications-in-health-care-and-biomedicine-4th-edition-edward-h-
shortliffe-220539054

Evaluation Methods In Biomedical Informatics 2ed Friedman C

https://ebookbell.com/product/evaluation-methods-in-biomedical-
informatics-2ed-friedman-c-2043100
Pediatric Biomedical Informatics Computer Applications In Pediatric
Research 1st Edition S Andrew Spooner Md

https://ebookbell.com/product/pediatric-biomedical-informatics-
computer-applications-in-pediatric-research-1st-edition-s-andrew-
spooner-md-4406750

Pediatric Biomedical Informatics Computer Applications In Pediatric

Research 2nd Edition John J Hutton Eds

https://ebookbell.com/product/pediatric-biomedical-informatics-
computer-applications-in-pediatric-research-2nd-edition-john-j-hutton-
eds-5609536

Methods In Biomedical Informatics A Pragmatic Approach 1st Edition

Neil Sarkar Eds

https://ebookbell.com/product/methods-in-biomedical-informatics-a-
pragmatic-approach-1st-edition-neil-sarkar-eds-4556560

Advances In Biomedical Informatics 1st Edition Dawn E Holmes

https://ebookbell.com/product/advances-in-biomedical-informatics-1st-
edition-dawn-e-holmes-6790554

Optimization And Data Analysis In Biomedical Informatics 1st Edition

Peter A Dimaggio Jr

https://ebookbell.com/product/optimization-and-data-analysis-in-
biomedical-informatics-1st-edition-peter-a-dimaggio-jr-2628002
 Biomedical Informatics
in Translational Research
Artech House Series
Bioinformatics & Biomedical Imaging
Series Editors
Stephen T. C. Wong, The Methodist Hospital and Weill Cornell Medical College
Guang-Zhong Yang, Imperial College

Advances in Diagnostic and Therapeutic Ultrasound Imaging, Jasjit S. Suri,

Chirinjeev Kathuria, Ruey-Feng Chang, Filippo Molinari, and Aaron Fenster,
editors
Biological Database Modeling, Jake Chen and Amandeep S. Sidhu, editors
Biomedical Informatics in Translational Research, Hai Hu, Michael Liebman,
and Richard Mural
Genome Sequencing Technology and Algorithms, Sun Kim, Haixu Tang, and
Elaine R. Mardis, editors
Life Science Automation Fundamentals and Applications, Mingjun Zhang,
Bradley Nelson, and Robin Felder, editors
Microscopic Image Analysis for Life Science Applications, Jens Rittscher,
Stephen T. C. Wong, and Raghu Machiraju, editors
Next Generation Artificial Vision Systems: Reverse Engineering the Human
Visual System, Maria Petrou and Anil Bharath, editors
Systems Bioinformatics: An Engineering Case-Based Approach, Gil Alterovitz
and Marco F. Ramoni, editors
 Biomedical Informatics
in Translational Research

Hai Hu
Richard J. Mural
Michael N. Liebman
Editors

artechhouse.com
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the U. S. Library of Congress.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library.

ISBN-13: 978-1-59693-038-4

Cover design by Igor Valdman

© 2008 ARTECH HOUSE, INC.

685 Canton Street
Norwood, MA 02062

All rights reserved. Printed and bound in the United States of America. No part of this book
may be reproduced or utilized in any form or by any means, electronic or mechanical, includ-
ing photocopying, recording, or by any information storage and retrieval system, without
permission in writing from the publisher.
All terms mentioned in this book that are known to be trademarks or service marks have
been appropriately capitalized. Artech House cannot attest to the accuracy of this informa-
tion. Use of a term in this book should not be regarded as affecting the validity of any trade-
mark or service mark.

10 9 8 7 6 5 4 3 2 1
 To the patients,
whose quality of life we strive to improve!
Contents
Preface xiii

CHAPTER 1
Biomedical Informatics in Translational Research 1
1.1 Evolution of Terminology 3
1.1.1 Translational Research 3
1.1.2 Systems Biology 4
1.1.3 Personalized Medicine 4
References 9

CHAPTER 2
The Clinical Perspective 11
2.1 Introduction 11
2.2 Ethics in Clinical Research 12
2.3 Regulatory Policies for Protecting a Research Subject’s Privacy 13
2.4 Informed Consent 15
2.5 Collecting Clinical Data: Developing and Administering Survey
2.5 Instruments 17
2.6 Issues Important to Biomedical Informatics 18
2.6.1 Data Tracking and Centralization 18
2.6.2 Deidentifying Data 19
2.6.3 Quality Assurance 20
2.6.4 Data Transfer from the Health Care Clinic to the Research
2.6.4 Setting 21
2.7 Standard Operating Procedures 23
2.8 Developing and Implementing a Research Protocol 23
2.8.1 Developing a Research Protocol 24
2.8.2 Implementing the Research Protocol 28
2.9 Summary 29
References 29

CHAPTER 3
Tissue Banking: Collection, Processing, and Pathologic Characterization of
Biospecimens for Research 31
3.1 Introduction 31
3.1.1 A Biorepository’s Mandate 31
3.1.2 Overview of Current Tissue Banking Practices 32

vii
viii Contents

3.2 Consenting and Clinical Data Acquisition 33

3.3 Blood Collection, Processing, and Storage 33
3.4 Tissue Collection, Processing, Archiving, and Annotation 35
3.4.1 Tissue Collection 35
3.4.2 Tissue Processing 36
3.4.3 Tissue Archiving and Storage 37
3.4.4 Pathologic Characterization of Tissue Samples 39
3.5 Conclusion 41
References 41

CHAPTER 4
Biological Perspective 43
4.1 Background for “Omics” Technologies 43
4.2 Basic Biology and Definitions 44
4.2.1 A Historical Perspective 44
4.2.2 Biological Processes 44
4.2.3 Some Definitions 45
4.3 Very Basic Biochemistry 46
4.3.1 DNA 46
4.3.2 RNA 47
4.3.3 Proteins 50
4.4 Summary 52
References 52

CHAPTER 5
Genomics Studies 55
5.1 Introduction 55
5.2 Genomic Technologies Used for DNA Analysis 56
5.2.1 DNA Sequencing 56
5.2.2 Genotyping 58
5.2.3 Array-Based Comparative Genomic Hybridization 64
5.3 Genomic Technology Used for RNA Analysis 69
5.3.1 Real-Time PCR 69
5.3.2 Microarrays 70
5.3.3 Chips for Alternative Splicing Analysis (GeneChip Exon) 76
5.4 Translational Research Case Studies 78
5.4.1 Case 1 79
5.4.2 Case 2 79
5.5 Summary 80
References 80

CHAPTER 6
Proteomics 85
6.1 Introduction 85
6.2 Clinical Specimens 87
6.2.1 Body Fluids 87
6.2.2 Tissue 89
Contents ix

6.3 Proteomics Technologies 90

6.3.1 Two-Dimensional Gel Electrophoresis 91
6.3.2 MALDI-TOF 93
6.3.3 Liquid Chromatography Mass Spectrometry 95
6.3.4 Protein Arrays 101
6.4 Analysis of Proteomics Data 103
6.4.1 2D DIGE Data Analysis 103
6.4.2 SELDI-TOF/MALDI-TOF Data Analysis 103
6.4.3 Shotgun Proteomics Data Analysis 104
6.5 Summary 105
References 105

CHAPTER 7
Data Tracking Systems 111
7.1 Introduction 111
7.1.1 Definition of a Data Tracking System 111
7.1.2 Why Use a Data Tracking System? 112
7.2 Overview of Data Tracking Systems 113
7.2.1 Historical Review 113
7.2.2 Available Resources 114
7.2.3 Data Tracking Systems in the Life Sciences 114
7.3 Major Requirements of a Data Tracking System for Biomedical
7.3 Informatics Research 119
7.3.1 General Requirements 120
7.3.2 Front-End Requirements 120
7.3.3 Back-End Requirements 121
7.3.4 Field-Specific Requirements 121
7.3.5 Additional Points 126
7.4 Ways to Establish a Data Tracking System 127
7.4.1 Buy a System Off the Shelf 127
7.4.2 Develop a System 129
7.4.3 Pursue a Hybrid Approach 132
7.5 Deployment Challenges and Other Notes 133
7.5.1 Resistance from End Users 133
7.5.2 Training 134
7.5.3 Mismatches Between System Features and Real Needs 135
7.5.4 Protocol Changes and Other Evolutions 135
7.5.5 Data Tracking System as a Data Source 136
7.6 Summary 136
References 136

CHAPTER 8
Data Centralization 141
8.1 An Overview of Data Centralization 142
8.2 Types of Data in Question 145
8.2.1 In-house Patient-Centric Clinical, Genomic, and Proteomic Data 147
8.2.2 Publicly Available Annotation and Experimental Data 149
x Contents

8.2.3 Data Format Standards 155

8.3 DW Development for Integrative Biomedical Informatics Research 157
8.3.1 Selection of the Developing Partner—Experiences in the Field 157
8.3.2 DW Requirements 158
8.3.3 Data Source Selection 159
8.3.4 Hardware and the Database Management Systems Selection 160
8.3.5 DW Structural Models—Integrated, Federated, or Hybrid 161
8.3.6 Data Models: Dimensional Models, Data Marts, and
8.3.6 Normalization Levels 162
8.3.7 Data Models: EAV, Entity-Relationship, and Object-Oriented
8.3.6 Modules 162
8.3.8 Data Models: Handling of the Temporal Information 164
8.3.9 Data Extraction, Cleansing, Transformation, and Loading 166
8.3.10 Tuning and QA 167
8.3.11 Changes—The Dynamic Nature 167
8.4 Use of the DW 168
8.5 Example Case 169
8.6 Summary 171
References 171

CHAPTER 9
Data Analysis 175
9.1 The Nature and Diversity of Research Data in Translational Medicine 176
9.1.1 Where Data Reside 176
9.1.2 Operational Versus Analytical Data Systems 177
9.1.3 Data Warehouses 177
9.1.4 Data Preprocessing 178
9.2 Data Analysis Methods and Techniques 179
9.2.1 Generalized Forms of Analysis 179
9.2.2 Significance Testing 180
9.2.3 Predictive Modeling 182
9.2.4 Clustering 184
9.2.5 Evaluation and Validation Methodologies 185
9.3 Analysis of High-Throughput Genomic and Proteomic Data 187
9.3.1 Genomic Data 188
9.3.2 Proteomic Data 190
9.3.3 Functional Determination 192
9.4 Analysis of Clinical Data 194
9.5 Analysis of Textual Data 195
9.5.1 Data Sources 195
9.5.2 Biological Entity 196
9.5.3 Mining Relations Between Named Entities 197
9.6 Integrative Analysis and Application Examples 199
9.7 Data Analysis Tools and Resources 200
9.8 Summary 202
References 202
Contents xi

CHAPTER 10
Research and Application: Examples 207
10.1 Introduction 207
10.2 deCODE Genetics 208
10.2.1 Data Repository Development and Data Centralization 208
10.2.2 Genomic Studies 210
10.2.3 Application 212
10.3 Windber Research Institute 214
10.3.1 Clinical Data Collection and Storage 215
10.3.2 Data Tracking 216
10.3.3 Data Centralization 217
10.3.4 Genomic and Proteomic Studies 217
10.3.5 Data Analysis, Data Mining, and Data Visualization 219
10.3.6 Outcomes Summary 222
10.4 Conclusions 224
References 224

CHAPTER 11
Clinical Examples: A Biomedical Informatics Approach 227
11.1 Understanding the Role of Biomarkers and Diagnostics 227
11.2 Understanding the Difference Between Pathways and Networks 228
11.3 How Biomarkers/Diagnostics and Pathways/Networks Are Linked 228
11.4 Breast Cancer 229
11.5 Menopause 233
11.6 Coagulation/DIC 240
11.7 Conclusions 247
References 247

About the Editors 249

About the Contributors 250

Index 255
Preface
There are multiple definitions of “Biomedical Informatics.” We have taken one that
broadly defines this multidisciplinary subject as the management and usage of bio-
medical information encompassing clinical informatics, public health informatics,
and bioinformatics. This definition is increasingly important as new concepts and
technologies enter into medical practice and related basic research, and require new
types of information management and data analysis that relies on sophisticated sta-
tistical and computational technologies.
In particular, this book focuses on the application of biomedical informatics for
translational research. Translational research is often seen as the rapid inclusion of
the results of basic biological research into clinical practice, (i.e., “bench to bed-
side”). We have found that it is equally important to have clinical needs feeding into
the framing of basic research questions, more of a “bedside – bench – bedside” cycle
which further requires a strong biomedical informatics base. The need for merging
genomic, proteomic, and other “omic” data into the study of human diseases
requires using computational and statistical technologies due to the sheer volume of
the involved data. From the clinical perspective, it requires the identification of clin-
ically relevant questions, and supports the application of the results from research
back into clinical practice. From the molecular study perspective, it involves the
application of advanced analytical technologies in the study of human bio-speci-
mens. From the informatics perspective, it involves management of the large data
sets generated in the study. From the data analysis perspective, it involves deploy-
ment of existing computational and statistical methods and algorithms, and the
development of new methods to extract knowledge from the underlying data.
Our vision of biomedical informatics was formed and reduced to practice
beginning in 2003 at the Windber Research Institute. In June 2004, Drs. Liebman
and Hu presented some of our results at the Cambridge Healthtech Institute’s Con-
ference on Bioinformatics. The presentations drew the attention of Mr. Wayne
Yuhasz, the Executive Acquisition Editor of the Artech Publishing House, who sub-
sequently contacted us and initiated this book project. In late 2005, the editorial
team was strengthened by the addition of Dr. Mural. In preparing the manuscript
for this book, the editors divided their responsibilities as follows; Liebman was
responsible for Chapters 1 and 11, Mural was responsible for Chapters 4 and 6, Hu
was responsible for Chapters 7, 8, 9, and 10, Mural and Hu were jointly responsible
for Chapters 2, 3, and 5, and Hu was responsible for all the administrative work
associated with this book project.
As the editors, we sincerely thank all the contributors of this book who all work
in the frontiers of biomedical informatics and translational research amongst differ-
ent component fields. We highly appreciate the organizational support of the

xiii
xiv Preface

Windber Research Institute. We are also very grateful to the funding from the U.S.
Department of Defense to the Clinical Breast Care Project and the Gynecological
Disease Program for which the Windber Research Institute is one of the major par-
ticipants, which provided us with the opportunity to conceive, develop, and imple-
ment a biomedical informatics infrastructure, which in turn served as a rich resource
especially when examples were needed to illustrate our points on biomedical infor-
matics in this book. The funds for these two projects are managed through Henry
Jackson Foundation for the Advancement of Military Medicine, Rockville, MD,
USA.
Finally, we thank Mr. Wayne Yuhasz of Artech House, Inc., for his initiation of
this project, and for his advice and patience throughout the development and com-
pletion of the manuscript.
Hai Hu,
Richard J. Mural,
Michael N. Liebman
July, 2008
Windber, PA
 CHAPTER 1

Biomedical Informatics in Translational

Research
Michael N. Liebman

This book’s goal is to present the elements of and relationships among

“omics”-based approaches and clinical data perspectives while establishing the crit-
ical need to understand, define, refine, and address actual clinical needs. A critical
need exists to address the real issues that arise when a physician is faced with a
patient and the need to make clinical choices that will impact the patient, the
patient’s quality life, and the patient’s family. Bridging this gap between the clinical
need and the available technologies, clinical data, and clinician input is the role that
biomedical informatics can play in driving the evolution of patient care in the
postgenome era.
Translational research and personalized medicine have become buzzwords that
follow on the aspirations of systems biology and the postgenome era. They empha-
size the need to apply high-resolution and high-throughput technologies to medical
applications as they were originally conceived, intended, and funded. These terms,
however, have evolved with diverse definitions as molecular and clinical research-
ers, clinicians and patients, venture capitalists and investment bankers, and the pop-
ular press have all attempted to identify the value proposition that should evolve
from the overall investment in genomics, proteomics, metabolomics, and “omics”
projects in general. In general, the development of “omics”-based technologies and
their “integration” into systems biology have approached the complex problems of
health care from a technology-focused, bottoms-up approach. The transition from
the fundamental application of these approaches into true clinical utility remains
elusive as outlined next.
The generation of data far exceeds the ability to convert it into useful clinical
value, and, unfortunately, this trend continues in spite of weekly, even daily, reports
of the discovery of new disease-related genes and so forth. In Figure 1.1, data is con-
verted into information when redundancies are removed and it is “cleaned” to
remove spurious results; information becomes knowledge when its interpretation
leads to new discoveries (e.g., biomarkers, pathways, gene correlations); and
knowledge evolves to clinical utility when it is finally incorporated into the actual
practice of medicine (e.g., biomarkers become diagnostics and, eventually, causal
diagnostics). Within this context, unfortunately, the gap between data/information
and clinical utility continues to grow with time.

1
2 Biomedical Informatics in Translational Research

Data
Information

Amount
Gap

Knowledge
Gap
Gap
Clinical Utility
Time
Figure 1.1 Relationship of data to information, knowledge, clinical utility, and the increasing gap
between technology and science.

A significant opportunity has presented itself in which data generation can be

accomplished within a top-down approach, but this requires using clinician insight
and clinical data to identify and prioritize the relevant questions and use them to
drive experimental design. The integration of this clinical perspective with
“omics”-based data represents the field of biomedical informatics as described in
this book. In this manner, it becomes essential to stratify the complexity of the dis-
ease as well as to stratify the molecular characteristics of the patient to provide new
insight into the concepts and characteristics associated with the disease process as
separate from those associated with the patient who is exhibiting the disease. The
ultimate goal of these actions is to improve patient care and quality-of-life issues for
the patient and his or her family. This goal cannot be achieved without a close link
between the clinician and clinical perspective, which must drive the genomic,
genetic, epidemiologic, and proteomic research.
This book is intended to present the components and their relationship from
both “omics”-based approaches and clinical data perspectives, while establishing
the critical need to understand, define, refine, and address the clinical needs that
appear when a physician is faced with a patient and the need to make clinical
choices.
In constructing this book, perhaps the first questions to address are “What is
biomedical informatics?” and “How does it differ from bioinformatics and medical
informatics? To establish the framework for this book and for this book to succeed,
other terms also require examination and definition: systems biology, translational
medicine or research, and personalized medicine. This book is not intended to be
simply a series of definitions, but rather a presentation of the attempt to integrate a
wide range of both clinical and molecular views and perspectives of the patient into
a single, coherent patient-centric view. The goal of this reorientation is to enable the
patient, her clinical history and state, the physician, and the molecular and/or clini-
cal researcher to be better equipped to handle the data, information, and potential
knowledge that results from the application of the advanced technologies of both
omics and diagnostics, to identify and tackle real clinical problems along the path
toward improving the patient’s quality of life.
To define biomedical informatics, we probably have to examine how we view
disease. The evolution from risk (genetics/genotype) to disease (expressed pheno-
1.1 Evolution of Terminology 3

type) should be viewed as a continuum, not as distinctly separable states, and dis-
ease itself should be viewed as an ongoing process, not a state fixed in time. This is
an important distinction from the current application of diagnostics and therapeutic
intervention and will impact the drug development process. Biomedical informatics
requires access to longitudinal patient medical histories, not simply clinical trial
data.
If we add clinical data to current bioinformatics practices, we establish the fol-
lowing relationships:

Clinical observations + Molecular/genetic information à Clinical correlations

Clinical observations + Biological process/pathway knowledge à Clinical mechanism

Clinical correlation points us in the right direction, but the clinical mechanism
directs us to the best target for diagnostic or therapeutic development. Biomedical
informatics is the catalyst for the conversion from correlation to mechanism.
Although bioinformatics provides the fundamental knowledge about general bio-
logical processes, it is biomedical informatics, with the inclusion of clinical observa-
tions, that enables this knowledge to be brought to bear on drug and diagnostic
development and, ultimately, clinical practice. Its value cannot be underestimated.

1.1 Evolution of Terminology

A simple Google search indicates incredible activity in the areas discussed thus far,
with more than 1,990,000 hits for “personalized medicine” (compared to more
than 293,000,000 for “medicine”), more than 1,910,000 for “translational medi-
cine or research,” and more than 146,000,000 for “systems biology.” This proba-
bly reflects the diversity in interpretation and application of these terms rather than
deep, focused efforts along specific research tracks. More importantly, these con-
cepts have generally evolved from a focus on technology rather than a focus on clin-
ical need, although their stated goals are directed to both understanding the
fundamental science and improving patient care. An ongoing problem that exists
within the scientific community is the perception that more data means more
knowledge. This reflects an incomplete appreciation of the significant chasm
between data and knowledge, and the even greater gap that exists when we evaluate
knowledge in terms of clinical utility.
Several operational definitions are used in this book and it is important to
understand their context as we examine molecular-based technologies, clinical
observations and limitations, and clinical decision making. Whereas it is critical to
understand these concepts individually, it is their synergies that will provide the
basis for improving patient care and quality of life.

1.1.1 Translational Research

Translational research is focused on the conversion of laboratory results into clini-
cal utility, but to be successful, translational research must actually start in the clinic
and not at the lab bench. This critical change is necessary because good
translational research must begin with the identification, elucidation, and commu-
4 Biomedical Informatics in Translational Research

nication of clinically significant problems into the laboratory for research and reso-
lution. The only true measure of success of this targeted research is in terms of what
translates into the clinic. Although this may seem logical, examples of such true suc-
cesses are somewhat limited because the driver of much of academic research still
focuses on research that may be significant for enhancing our understanding of biol-
ogy, but does not necessarily transcend into addressing more direct clinical needs
[1–5].
Chapter 2 addresses the clinical perspective that is necessary to support this view
of translational research, and Chapter 3 discusses the critical aspects of sample and
data collection and the quality control issues needed to support clinically based
research.

1.1.2 Systems Biology

Systems biology is most commonly interpreted as the aggregation and integration of
multiple approaches to analyze and define a system, that is, the “omics” perspective,
and then analysis of the behavior of the system based on these perspectives. This bot-
tom-up approach can only bring together those views that are available through the
application of existing (and evolving) technologies. It is easy to see that this
approach can be limited by the expectation that these technologies will provide a
complete picture of the entity being studied rather than multiple views, each with its
own contextual limitation. Of course, these technologies are neither comprehensive
enough to provide a complete picture of a patient, nor can they necessarily produce
data of equivalent quality or content. A more suitable approach to systems biology
may involve a top-down approach, first examining the behavior of the intact system
(e.g., a patient with or without disease symptoms), to more fully identify the critical
question and then determine the technology or technologies most appropriate to
addressing these questions. It is clear, however, that all results must be integrated
into a comprehensive data model [6, 7]. In the case of biomedical informatics, this
model should be patient-centric to enable the exploration of relationships among
the multiple views presented by the different technologies. It should be clear that this
top-down approach aligns directly with the translational medicine definition given
earlier.
Chapter 4 provides a general overview of the range of data in the “omics” fields,
as well as issues related to experimental design and implementation. Chapter 5
focuses on issues specifically related to genomic studies, and Chapter 6 presents the
proteomic perspective.

1.1.3 Personalized Medicine

Personalized medicine has focused on optimizing treatment to maximize efficacy
and minimize risk (i.e., therapeutic medicine), using the genetic makeup of the
patient. However, per the Wikipedia website, “Medicine is … concerned with main-
taining or restoring human health through the study, diagnosis, treatment and possi-
ble prevention of disease and injury.” So, ideally, personalized medicine should
incorporate and promote a significant component of preventive medicine, thus
aligning more closely with the non-U.S. clinical perspective in which prevention is a
1.1 Evolution of Terminology 5

major focus of health care that is frequently based on a single-payer system. More
importantly, restricting the approach to only include genetic information may sig-
nificantly limit its application to clinical practice and its ability to support broader
research goals. Within the frameworks of translational research and systems biol-
ogy as stated earlier, however, personalized medicine can evolve to achieve its
broadest goals, the improvement of patient care [8].
The key to developing personalized medicine as previously described involves
data tracking, data integration, and data analysis (e.g., visualization and data min-
ing). These specific topics are covered in Chapters 7, 8, and 9. A key element to this
integration and analysis involves the development of a patient-centric data model
and its potential implementation across both the research and clinical domains.
Some of the challenges facing biomedical informatics include the following:

1. Patient records are not universally available in electronic form.

2. The data is soft data; that is, clinical observations may be qualitative in
nature.
3. Quantitative results may require significant detail about the underlying test
and reagents used.
4. Medical terminology may be ambiguous across different specialties.
5. Patient confidentiality must be maintained.
6. Patient consent must be obtained to use data in a particular study.
7. Diseases as we know them today are typically composites of multiple
subtypes that reflect an individual’s genetic makeup and response.
8. Diseases are frequently observed well beyond their initiation, which results
in comorbidities and the reduced ability to provide effective treatments.
9. Disease etiologies require synchronization of patient records, which is not
currently available for most diseases.
10.Methodologies evolve as do standards of care, practice guidelines,
diagnostic procedures, and so forth. Many of these have analogies in the
bioinformatics domain.

Chapters 10 and Chapter 11 focus on examples that tie together the compo-
nents laid out in the earlier chapters. Chapter 10 focuses on the Clinical Breast Care
Project (CBCP) as an example of the integration of the clinical and molecular infra-
structures to support a broad-based research program ranging from clinical data
and tissue collection to molecular characterization and analysis in terms of the
translational research model addressed earlier, but in a bottom-up approach.
Chapter 11 focuses on an example of the top-down approach, in which the
problems associated with clinical decision making for patient treatment in breast
cancer are viewed from the perspective of successes and gaps in our appreciation of
the full scope of patient–physician issues. Additional examples of the clinical per-
spective are presented about stratifying individuals in their transition toward meno-
pause and about examination of disorders in blood coagulation.
Confronted with an emphasis on treatment rather than prevention, these activi-
ties have focused on the development of diagnostics and/or therapeutics to directly
impact treatment rather than understanding the fundamental aspects of the under-
6 Biomedical Informatics in Translational Research

lying diseases or the physiological (and psychological) state of the patient. It is

readily observed that breast cancer appears/behaves differently in premenopausal
versus postmenopausal women. Is this the same disease in a developmentally differ-
ent host or does it reflect different diseases beyond estrogen receptor (ER) and
progestrone receptor (PR) status? Also, statistical evidence relates risk for breast
cancer to smoking, alcohol use, and body weight factors.
Are these risks uniform throughout a patient’s lifetime? Not likely! For exam-
ple, the breast undergoes developmental changes continuously during the in utero
to postmenopause transition (Figure 1.2), and the mechanistic basis for risk is
probably related to differences in gene expression, protein expression, protein modi-
fication, and so forth that will accompany these developmental changes. Biomedi-
cal informatics approaches this problem by analyzing the clinical data
epidemiologically to determine what level of exposure at what age may be most crit-
ical for establishing risk from one or more of these factors and then combining this
with molecular characterization of the underlying physiological changes (Figure
1.3). In this manner, more than a simple correlation between these risk factors and
breast disease can be realized because the molecular processes, including
upregulation and downregulation of the gene/protein pathway, can be identified;
these are potentially mechanistically related to observed risk. This enhances the like-
lihood for identifying new diagnostics as well as therapeutics and perhaps, more
importantly, establishing a more highly refined basis for making lifestyle and envi-
ronmental choices for the patient and the physician.
The complexity of the underlying biological relationship between patient and
disease has not been adequately addressed by translational research, systems biol-
ogy, or personalized medicine to date and requires refocusing their potential to
describe these underlying processes. A fundamental aspect of this complexity is the
fact that disease, although frequently described in terms of “disease state,” actually
represents a process that evolves over time, through a integrative relationship
involving the patient’s genetics and his interaction with lifestyle and environmental
factors, always starting significantly before any symptoms may appear.

Genetic risk + Lifestyle/environment/exposure [F(t)] à Disease

Breast Development

Cumulative development

Lactation

Menarch Menopause
Peri-
menopause
Child-bearing
Fetal stages
Figure 1.2 Major stages of breast development in a woman’s lifetime.
1.1 Evolution of Terminology 7

Risk = (Age, Development)

Figure 1.3 Assessment of critical lifestyle risk factors over a patient’s lifetime.

The key to tackling this complex relationship comes from the integration of the
approaches defined earlier, but the essential requirement is to define separately the
characteristics of the patient from those of the disease.
Patients possess intrinsic characteristics, that is, those derived from their genetic
makeup and its expression in their underlying physiological structures, and extrin-
sic factors such as their lifestyle and environmental exposures (e.g., physical activ-
ity, smoking behaviors, alcohol consumption). The disease represents a pattern of
actions that interact with and are modified by these characteristics. Thus the conun-
drum is how to identify and separate the extensible definitions of disease from those
that are dependent on the patient in terms of intrinsic and extrinsic characteristics.
Thus, the presentation of disease exceeds the simple sum of its parts, namely, the
patient and the disease process. For biomarkers/diagnostics to be effective in speci-
fying targets or levels for appropriate intervention, it becomes critical to interpret
and resolve the complexity of the disease–patient interaction.
To better define the patient, we have been developing a personalized health
record (PHR) [9] that spans the life history of the patient and treats the data in a
chronological record including, but not limited to, family history (at a genetic anal-
ysis level), history of diagnoses/illnesses, treatments and response, and lifetime
exposure to environmental and lifestyle risk factors (e.g., smoking, body mass
index, alcohol consumption).
A key difference in establishing a personalized health record versus an elec-
tronic medical record is that the PHR is temporally focused; that is, it creates a
timeline for the patient around all clinical, lifestyle, and environmental parameters,
while most electronic medical records (EMR) focus on specific medical incidents.
The PHR provides the base to both represent and model temporal data and relate it
accurately to the patient. This is critical to understanding the underlying physiologi-
cal state of a patient, which includes co-occurring disease or treatment, as well as
lifestyle and environmental factors, all of which may impact diagnosis, prognosis,
and response to treatment.
8 Biomedical Informatics in Translational Research

An essential complement to the PHR in establishing its utility for clinical deci-
sion making involves the need to be able to model a patient with respect to both
qualitative and quantitative time relationships and for the purpose of identifying
critical co-occurrence information about diseases and risks. This enables a physician
to both represent the detailed history of a specific patient and to also com-
pare/search the database for other patients with similar patterns of disease for use in
evaluating treatment options. The examination of co-occurrence of disease [10] or
systemic problems is thus examined from both an epidemiologic perspective as well
as a mechanistic perspective, using pathway databases and pathway simulation/rea-
soning methodologies to support the evaluation of genotypic variation or drug
interactions in the patient.
Enhancing the accuracy in defining disease is extremely difficult because of the
tendency to group potential disease subtypes under simple classifications, for exam-
ple, breast cancer. Among the elements critical to more accurately define a disease
subtype is the need to identify quantifiable characteristics, such as disease progres-
sion or subpathologies, which can readily complement the current trend toward
stratification using gene expression technology alone. Utilization of the concepts of
disease progression and disease process as noted should be invaluable for measuring
and analyzing clinical parameters longitudinally—not just at time of diagnosis.
Note that most diagnostics/biomarkers have been developed because of their corre-
lative relationship with a disease diagnosis or state, not because of a quantifiable
recognition of their mechanistic relationship to disease symptoms and so forth.
Optimally, we need to define disease stratification as the longitudinal path or vector
through all clinical parameters that can be observed for a patient and which may
occur in multiple dimensions, each one reflecting a clinical observation. The CBCP
(see Chapter 10) measures more than 600 clinical parameters and additional molec-
ular descriptors longitudinally; thus, a patient can be algorithmically represented as
moving through a 600-dimensional space!
Disease stratification involves identifying the significant clusters of this longitu-
dinal progression, and the concomitant reduction of dimensionality necessary to
describe the disease pathway. Disease staging, then, is observation of how far along
a disease path a patient has progressed [11], although clinical ambiguities may be
present when two patients exhibit similar diagnostic values but are on different dis-
ease paths. This can result from the limitations of using a correlative relationship
with the biomarker and the fact that these patients may be at different time points
along their separate disease paths that overlap in one or more dimensions (clinical
parameters). Conversely, two patients may appear “diagnostically” different
although they are on the same disease path but have been observed at different dis-
ease stages. This dilemma is faced by physicians daily and its resolution relies on the
experience and knowledge of the practitioner. To truly develop diagnostics,
biomarkers, and the area of personalized medicine, it will be critical to analyze and
interpret the longitudinal nature of disease in a more quantifiable manner to reflect
its true complexity.
This process of stratification of the patient versus stratification of the disease is
not one that is readily solvable with current patterns of patient information/record
keeping, sole dependency on genomic information, and so forth, but will require
1.1 Evolution of Terminology 9

extensive, recursive modeling of the complexity of the relationship that truly defines
the patient–disease relationship.
There also must be an evaluation, based on issues of quality of life for the
patient and his or her family, access to technology, and cost–benefit analysis of the
application, to determine for which diseases and which patients this analysis will
become critical. These issues will quickly move beyond the question of access to
technology to touch on cultural and ethical boundaries and sensitivities that cur-
rently exist. But the reality of personalized medicine and the development of effec-
tive diagnostics that truly support improvement of quality of life for the patient
must anticipate and deliver on the integration of all of these factors to become truly
effective. Biomedical informatics, as outlined in this book, will be an enabler of
these changes.

References

[1] Mathew, J. P., et al., “From Bytes to Bedside: Data Integration and Computational Biology
for Translational Cancer Research,” PLoS Comput. Biol., Vol. 23, No. 2, February 2007,
p. e12.
[2] Littman, B. H., et al., “What’s Next in Translational Medicine?” Clin. Sci. (Lond.), Vol.
112, No. 4, February 2007, pp. 217–227.
[3] The Translational Research Working Group (NCI/NIH) defines translational research as
follows: Translational research transforms scientific discoveries arising from laboratory,
clinical, or population studies into clinical applications to reduce cancer incidence, morbid-
ity, and mortality.” Available at http://www.cancer.gov/trwg/TRWG-definition-and-TR-
continuum.
[4] Wikipedia. “Translational Medicine.”
[5] Liebman, M. N., “An Engineering Approach to Translation Medicine,” Am. Sci., Vol. 93,
No. 4, July/August 2005, pp. 296–300.
[6] Huang, S., and J. Wikswo, “Dimensions of Systems Biology,” Rev. Physiol. Biochem.
Pharmacol., Vol. 157, 2006, pp. 81–104.
[7] Liebman, M. N., “Systems Biology: Top-Down or Bottom-Up?” Bio. IT World, March
2004.
[8] Sikora, K.,“Personalized Medicine for Cancer: From Molecular Signature to Therapeutic
Choice,” Adv. Cancer Res., Vol. 96, 2007, pp. 345–369.
[9] Hu, H., et al., “Biomedical Informatics: Development of a Comprehensive Data Warehouse
for Clinical and Genomic Breast Cancer Research,” Pharmacogenomics, Vol. 5, No. 7,
2004, pp. 933–941.
[10] Maskery, S. M., et al., “Co-Occurrence Analysis for Discovery of Novel Patterns of Breast
Cancer Pathology,” IEEE Trans. on Information in Biomedicine, Vol. 10, No. 3, July
2006, pp. 1–7.
[11] Liebman, M. N., “Information Processing Method for Disease Stratification and Assess-
ment of Disease Progression,” European Patent EP1399868.
 CHAPTER 2

The Clinical Perspective

Lee Bronfman, Craig D. Shriver, and Emily Gutchell

2.1 Introduction

Biomedical informatics research begins with clinical questions: How can the growth
of a tumor be stopped? How do certain diseases metastasize? Is there a less invasive
way than surgical biopsy to determine the presence of breast cancer? The questions
begin broadly and become very specific: Which women at high risk of developing
breast cancer will benefit from the use of tamoxifen and which will not? What is the
optimal combination of chemotherapy for a breast cancer that has metastasized to
the bones? Why does Herceptin work for some breast cancer patients but not
others?
It is within the clinical setting that the need for biomedical research is revealed,
and to this setting that the advancements of biomedical research return. No one
truly needs to be convinced of the need for clinical research. There is intrinsic agree-
ment that as human beings we have extraordinary needs for quality health care.
Although debates ensue regarding how research is conducted, it is rarely debated
that research is an essential enterprise and that it should be conducted with the high-
est level of integrity. A research project that meets the highest standards will begin
with a well-devised plan for launching the project from the clinical setting. In the
competitive biomedical informatics research environment of today, a successful
research project integrates the normal flow of clinical processes with the unique
demands of research. Health care clinics conducting biomedical research are likely
to be busy, with thousands of patient visits a year. Clinical research staff must
accomplish their work within the normal clinic business of the day, which poses
many challenges. A biomedical research project that is well integrated at the health
care clinic setting exhibits the following qualities:

• Clinicians and clinic staff are well oriented to the goals of the research.
• Clinic staff are well trained in the processes related to the research.
• The actions of leadership support the research mission.
• An organized, well-documented process is followed to obtain a patient’s
informed consent to participate in research.

11
12 The Clinical Perspective

• Biospecimens are collected, temporarily stored, transported to, and perma-

nently stored at the biorepository in which they will reside in such a way that
the specimen is preserved for optimal immediate and future research.
• Well-designed data management systems serve and support the biomedical
informatics platform of the project and are continually well maintained and
managed. Clinical research staff managing biomedical information have
appropriate, ongoing training in the informatics systems they use. Access to
these systems is secure and limited to essential staff.
• Well-defined quality assurance processes are in place to monitor informed
consent and the collection, storage, and transmission of all biomedical data
collected, including biospecimens.

This chapter discusses important clinical perspectives involved in implementing

and maintaining a biomedical research project. Important individual issues within
the clinical health care setting are discussed first: ethics, privacy protection, subject
consent, and quality assurance. Research protocol development and implementa-
tion are discussed later in the chapter. Although this chapter is intended to be useful
to all biomedical researchers, it is focused operationally on the U.S. environment of
regulation in health care research.

2.2 Ethics in Clinical Research

A basic discussion of ethics in biomedical research begins with two key documents,
the Nuremberg Code and the Declaration of Helsinki [1, 2]. The Nuremberg Code
was developed in 1947 amid the findings of a Nuremberg military tribunal that
addressed the war crimes of a group of Nazi physicians. The Nuremberg Code is a
directive of 10 ethical principles of research involving human subjects and is the first
official offering of formal guidelines on the subject. Nearly 20 years later, the Decla-
ration of Helsinki was introduced by the World Medical Association and adopted at
the 18th World Medical Assembly in Helsinki, in June 1964. Adding to the core
principles of the Nuremberg Code, it also outlines basic principles for ethical
research. Amendments to the Declaration of Helsinki have been adopted five times
between 1975 and 2000. Since then, notes of clarification have additionally been
adopted. In themselves, the many revisions to the Declaration of Helsinki reflect the
complex challenges of developing a comprehensive code of ethics for human use
research. Regulations and guidelines are continually reevaluated for unforeseen
gaps as well as the unintended consequences of overregulation exposed by the needs
of researchers and human subjects today.
In the United States, it was not until 1974 that federal legislation was adopted to
protect human subjects in medical research. The act was called the National
Research Act [3]. The legislation charged a commission, the National Commission
for Protection of Human Subjects of Biomedical and Behavioral Research, with
addressing the need for federal guidelines and mandates to protect human subjects.
In 1979 that commission produced the Belmont Report [4]. The Belmont Report is a
summary of ethical principles intended to provide national guidelines to all those
involved in human subject research. The Belmont Report focuses on three basic ethi-
2.3 Regulatory Policies for Protecting a Research Subject’s Privacy 13

cal concepts in research: respect for persons, beneficence, and justice. Also legislated
by the National Research Act was the establishment of institutional review boards
(IRBs) for reviewing research involving human subjects. The Department of Health
and Human Services (DHHS) Office for Human Research Protections (OHRP) is
the oversight body for IRBs in the United States. Federal regulations governing the
protection of human subjects, under the statutory authority of the OHRP, can be
found in Title 45, Part 46, Subparts A, B, C, and D, of the Code of Federal Regula-
tions (CFR) [5]. Subpart A defines basic human subject protections that 15 other
departments and agencies within the federal government also adopted. Because the
language of Subpart A is identically reproduced in each of their different chapters of
the CFR, it is referred to as the Common Rule [6].
Returning to the international scene, another document addressing human sub-
ject protections in biomedical research was published in the early 1990s. In an effort
to address unique concerns of human subject protections for research conducted in
developing countries, the World Health Organization collaborated with the Coun-
cil for International Organizations of Medical Sciences (CIOMS), to publish the
International Ethical Guidelines for Biomedical Research Involving Human Sub-
jects in 1993, and again in 2002 [7].
IRBs and institutional ethics committees (IECs) exist to protect the rights and
the welfare of human subjects participating in research. These oversight bodies
require assurances that staff involved in research are trained in the ethics of human
subject protection. Many IRBs require research staff to complete established
trainings, usually offered online. References to available trainings in human subject
protection can be found later in Section 2.8.2.

2.3 Regulatory Policies for Protecting a Research Subject’s Privacy

Regulatory policies for safeguarding the privacy of personal information vary

among countries and have an impact on the use of data in research. Privacy policies
governing personal data originate from different perspectives, such as banking,
e-commerce, business, and health care, in different countries. Regardless, safe-
guarding the privacy of research subjects is an aspect of ethical practice that is now
highly legislated. Researchers in Canada are impacted by the Personal Information
Protection and Electronic Document Act [8]; researchers in the European Union, by
the European Union Directive on Data Protection [9]. Researchers in the United
States must comply with the Health Insurance Portability and Accountability Act
(HIPAA) [10].
HIPAA was enacted to address many disparate needs of health care consumers
and government agencies such as protecting employee health insurance when work-
ers changed or lost their jobs, reducing waste and fraud, and simplifying adminis-
trative processes. One of the many mandates of HIPAA produced the HIPAA
Privacy Rule, which requires agencies to protect an individual’s health information
[11]. The Privacy Rule is administered through DHHS’s Office for Civil Rights.
Although HIPAA was enacted in 1996, the HIPAA Privacy Rule was implemented
in April 2003.
14 The Clinical Perspective

Among the many mandates of the HIPAA Privacy Rule, it sets standards that
limit the use and disclosure of protected health information (PHI) for research pur-
poses. The Privacy Rule does not replace or modify the Common Rule or any of the
other human subject protections in the CFR. The Privacy Rule works with these
other regulations to increase human subject protections. The intent of the HIPAA
Privacy Rule is to protect the privacy of individually identifiable health information
by establishing conditions for its use and disclosure. All “covered entities” are sub-
ject to the HIPAA Privacy Rule. A covered entity is a health care provider, health
plan, or health care clearinghouse. Researchers working within or collaboratively
with covered entities become bound by provisions of the Privacy Rule as well. Thus,
many researchers must obtain written permission from study enrollees, in the form
of a HIPAA authorization, before the enrollee’s protected health information can be
used in research.
An enrollee’s PHI is any information that can identify the individual. This
includes the following 18 identifiers: names, locations smaller than a state, dates
that relate directly to an individual, telephone and fax numbers, e-mail addresses,
Social Security numbers, medical record numbers, prescription numbers, health
plan beneficiary numbers, other account numbers, certificate/license numbers, vehi-
cle identification numbers, serial numbers, license plate numbers, device identifi-
ers/serial numbers, web URLs and IP addresses, biometric identifiers such as
fingerprints and voiceprints, photographic images, and any other unique identifying
number or code. Given this exhaustive list and the nature of clinical and demo-
graphic data collected for biomedical research projects (such as dates of birth, zip
codes, dates of medical diagnostic tests and procedures, and family medical histo-
ries), it is likely that the biomedical researcher will need to obtain written HIPAA
authorization from research subjects.
For protocols collecting very limited PHI, a researcher may elect to use a limited
data set as defined by the Privacy Rule. A limited data set is one in which all of the
following identifiers pertaining to the research participant, as well as the partici-
pant’s relatives, employers, and household members, have been removed from the
recorded data: names, addresses (town, city, state, and zip codes are aspects of
addresses that may remain in a limited data set), telephone and fax numbers, e-mail
addresses, Social Security numbers, medical record numbers, health plan beneficiary
numbers, account numbers, certificate/license numbers, vehicle identification num-
bers, serial numbers, license plate numbers, device identifiers/serial numbers, web
URLs and IP addresses, biometric identifiers such as fingerprints and voiceprints,
and full-face photographic images and any comparable images.
If using a limited data set, written authorization from the research subject is not
required. However, a data use agreement between the covered entity and the
researcher is required, ensuring that specific safeguards will be taken with the PHI
that remains within the data set. The IRB may require evidence of the data use agree-
ment in the protocol submission or may require a reference to the data use agree-
ment within the informed consent document. Informed consent is covered in the
next section.
A final consideration for researchers is to use only de-identified health informa-
tion, as defined by the Privacy Rule, for which there are no restrictions on use and
discloser. There are two ways to de-identify health information:
2.4 Informed Consent 15

1. Remove all of 18 unique identifiers (listed earlier) specified by the Privacy

Rule, thereby assuring no information can identify the individual.
2. Statistically deidentify information whereby a statistician certifies that there
is a very small risk that the information could be used to identify the
individual.

In most cases, biomedical informatics researchers will have a need for robust
data sets that do include some elements of personally identifiable data, and thus
must obtain a HIPAA authorization from research subjects. HIPAA authorization is
different from, but may be combined with, the informed consent document. It must
be for a specific research study. A single, general HIPAA authorization for all
research within an organization is not permitted. It must contain core elements and
required statements as mandated by the Privacy Rule. The National Institutes of
Health provides user friendly guidance on the core elements and required state-
ments at http://privacyruleandresearch.nih.gov/authorization.asp. Additionally,
the requirements are published in the Code of Federal Regulations, Title 45, Part
164.508(c) and (d).
The HIPAA authorization does not need to be reviewed by the IRB or privacy
board, unless it is combined with the informed consent document, in which case the
IRB will review it as part of informed consent. The HIPAA authorization form need
not expire, and the fact that it does not must be stated. All research participants
must receive a signed copy of their HIPAA authorization.

2.4 Informed Consent

Informed consent must be obtained from every research subject. This is a federal
requirement mandated by the CFR, Title 21, Part 50, Section 50.25, and Title 45,
Part 46, Section 116. The intent of the regulation is to provide research subjects
with all of the necessary information needed to make an informed decision about
whether or not to participate in the research. There are many required elements in a
consent form. Each is listed within the code noted above. Among them, of key
importance to the biomedical researcher are a description of the study’s research
purposes and procedures, identification of any procedures in the protocol that are
considered experimental, a disclosure of alternative treatments that may be of bene-
fit to the patient, the expected duration of a subject’s participation, a description of
any reasonably foreseeable risks, a statement describing the extent to which confi-
dentiality of records will be maintained, and, where appropriate, a statement of sig-
nificant new findings that developed during the course of the research, which could
impact the subject’s willingness to continue to participate. Additionally, patients
must be informed about the types of data that will be collected for the research
project, how it will be privately and securely stored, and how it will be analyzed.
Many IRBs recommend that consent forms be written at an eighth-grade reading
level. This becomes challenging when genomic and proteomic investigations are the
goal of the researcher. Mark Hochhauser, Ph.D., is a readability consultant and
advocates the use of plain English for consent forms. He recommends including a
table of contents in a consent form and following a question and answer format [12].
16 The Clinical Perspective

Microsoft Word offers functions for calculating readability scores for a docu-
ment. These can be optionally displayed when using the Spelling and Grammar
tools. A student’s or children’s dictionary can also be a useful tool. It is important
that the consent form be written in the language that the research subjects use; there-
fore, more than one language version of the consent form may be necessary.
Be sure to include all of the required elements as outlined in the CFR in your
informed consent document. These are likely to be reviewed carefully by the IRB.
An excerpt from the informed consent document used by the Clinical Breast
Care Project (CBCP) is shown below. CBCP is a biomedical research collaboration
between Walter Reed Army Medical Center in Washington, D.C., and Windber
Research Institute in Windber, Pennsylvania. The CBCP seeks to decrease morbidity
and mortality from breast cancer and breast disease by providing patients with
state-of-the-art clinical care, including advances in risk reduction. The CBCP is also
committed to a translational research goal of rapidly translating research findings
into clinical practice. The CBCP has a state-of-the-art biospecimen repository to
which research subjects contribute blood and tissue samples. Multiple question-
naires containing clinical data are collected on each patient and biospecimens are,
thus, extensively annotated. Researchers use the repository to conduct proteomic
and genomic research. Clinical data are analyzed in response to proteomic and
genomic analysis and findings. Biomedical informatics scientists develop informa-
tics structures and expand the informatics environment on which the research relies
for rapid translation into the clinical environment. Given the challenges of ade-
quately describing research with many goals in easy-to-understand language, the
section of the informed consent document used by the CBCP that describes the study
is two pages long. The following is an excerpt from that section:

The tissue and blood samples will undergo the primary research studies as described
in the next paragraphs.

This breast tissue can be used for many types of laboratory research looking at
cell changes during breast cancer development, identification of risk factors that
lead to breast cancer, learning why and how breast cancer spreads to lymph nodes,
and breast tissue biology. Some of this type of research on tissue is genetic research.
These research studies may include examination of the gene markers of the breast
cells or lymph nodes, as well as other parts of the breast cells.
The primary research uses of the blood and tissue samples are to study the
genetic makeup, protein changes, and to look for other research markers that may
be associated with breast disease. This information will be analyzed and stored in a
computer database using your assigned code number.
Any remaining blood and tissue samples left over after these studies are done
will remain frozen indefinitely until needed for approved research projects. There
may be future uses of the remaining frozen samples of your blood that are not
known to us at this time. If other research is to be conducted using these samples,
this research also will only be conducted after the samples have been stripped of all
identifiers. It is not possible at this time to predict all the potential future research
uses of the samples. You will be given the opportunity at the end of this consent form
2.5 Collecting Clinical Data: Developing and Administering Survey Instruments 17

to indicate if you will allow your blood and tissue samples to be used for future
unknown breast disease research.
Although not required by the CFR, an important item on a consent form for
proteomic and genomic research, and biomedical research in general, is an option to
consent for future research. If approved by the IRB, this will allow researchers to
conduct proteomic and genomic research in the future without having to obtain a
second consent from the subject. Many ethicists and, in fact, most patients prefer
this approach to consenting in advance for related, but presently unknown, future
research on specimens and data collected now. It avoids a cold call from the
research team in what might be years after the patient has successfully completed a
course of treatment for the disease and readdressing a difficult time in order to
obtain an updated consent; such a call could cause needless emotional turmoil. If
the patient should die of his or her disease, and the research team then inadvertently
contacts the family, this too could have significant emotional effects on all con-
cerned. Enlightened IRBs, especially those with good patient or layperson represen-
tation, understand this important point about allowing the patients to consent now
for future unknown research uses of their donated biospecimens and data. For stud-
ies that employ future use consenting, a reliable process must be in place that identi-
fies to the appropriate study personnel in a timely manner which participants’ data
may not be used for future research. In the experience of the CBCP, only about 2%
of the participants refuse to consent to future use research.
Obtaining consent from patients for future research uses of their clinical data,
including imaging data, as well as their biospecimens could prove pivotal to the suc-
cess of a biomedical research project in the genomic age.

2.5 Collecting Clinical Data: Developing and Administering Survey

Instruments

The array of clinical data collected by a biomedical research project will depend on
the goals of the project. Representatives from the clinical setting and the research
setting will be involved in determining what information needs to be collected.
Some consulting firms specialize in the development of research survey instruments.
These firms offer expertise in instrument development, optimal interviewing for-
mats given the characteristics of the instrument, training of staff for standardized
administration of the instrument, and data collection and data management ser-
vices. American Institutes for Research (http://www.air.org/), RTI International
(http://www.rti.org/), and RAND Survey Research Group (http://www.rand.org/
srg/) are organizations that offer survey development consultation. If a project
requires a lower cost solution, companies such as ObjectPlanet (http://www.
objectplanet.com/), Apian (http://www.apian.com), and PREZZA Technologies
(http://www.prezzatech.com/biz/company/) offer software for designing and
implementing surveys.
Many research projects develop their own survey instrument(s). It is helpful to
convene a multidisciplinary task force to successfully compile all data elements nec-
essary for the project. Staff with expertise along the continuum of collecting, pro-
cessing, storing, and analyzing data to be collected should be consulted. These are
18 The Clinical Perspective

clinical experts and scientists, business administrators at the organizations involved

in the research, research specialists adept at consenting and administering question-
naires to research subjects, information technology (IT) specialists in programming,
web services, and database management, and data quality specialists. Each of these
knows something helpful about obstacles and solutions for successfully obtaining
and securing data at their organization and/or their level of enterprise.
Data acquisition can easily become a lofty goal. The ideal to pursue the most
complete and accurate information available should be tempered with an under-
standing of the realities of data collection. Research participants range in their abili-
ties to give reliable, thorough information. Not all participants can answer an
extensive survey in one interview. It is helpful to prioritize essential information
needed by the project from nonessential information in keeping a survey from
becoming too exhaustive for the participants. Consider other venues such as elec-
tronic medical records within the health care informatics system from which data
can be drawn, keeping in mind that patients must be informed about and agree to
the data being collected about them. The following are example data types collected
in the CBCP: family cancer histories, dates of menarche and menopause, hormone
replacement therapy use and duration, hormonal birth control use and duration,
pregnancy and birth histories, intake of tobacco, alcohol, fat, caffeine, and under-
arm cosmetic use and duration. Additionally, diagnosis and treatment data are col-
lected, such as mammogram dates and findings, ultrasound dates and findings, CT
scans, x-rays, bone scans, PET scans, MRI, chemotherapy, radiation therapy, hor-
monal and endocrine adjuvant therapies, pathologic diagnoses from biopsies,
lumpectomies, mastectomies, and so forth.
The method by which the survey instrument will be administered must be cho-
sen. Although there are many ways to administer a research questionnaire (mail,
Internet, telephone, computer, chart abstraction, focus groups, and so on), the per-
sonal, one-on-one interview conducted by qualified and trained staff may yield the
highest quality data. Once the content of the instrument has been established, the
reliability of the responses it generates depends on the instrument being adminis-
tered the same way to every respondent. This reliability is achieved by standardizing
administration across interviewers [American Institutes for Research, personal
communication, January 2004].
To standardize administration across interviewers, training is required. Provid-
ing staff with professional training on administering the survey will clarify difficult
or confusing questions, establish trust between interviewer and interviewee,
empower research staff to confidently administer the survey, and result in better
data collection. It may allow for a longer, more extensive survey instrument to be
reliably completed by minimizing the questionnaire fatigue that patients experience
when filling out lengthy research tools.

2.6 Issues Important to Biomedical Informatics

2.6.1 Data Tracking and Centralization

A biomedical informatics study collects many data for multiple purposes. Demo-
graphic data on participants are essential for managing the demands of daily
2.6 Issues Important to Biomedical Informatics 19

research in the health care clinic. A data tracking system is needed to keep track of
all such data. Different types of data need to be tracked in different ways. For exam-
ple, data of a PHI nature, such as names, addresses, phone numbers, medical record
numbers, and dates of birth, are often elements that should be readily available to
clinic staff but blinded to researchers (with some exceptions such as date of birth
that are allowed in the “limited data set” defined in Section 2.3.). Non-PHI data,
such as dates and results of diagnostic tests, procedural interventions and outcomes,
medical therapies and outcomes, personal medical histories, social histories, family
histories of certain illnesses, stress levels, caffeine intake, x-ray images, ultrasound
images, tissue and blood biospecimens, should all be properly tracked (refer to
Chapter 7), and ideally are all subsequently centralized into a database of a data
warehouse nature (refer to Chapter 8), to enable researchers to analyze them syner-
gistically as much as possible. The data and findings generated by research scientists
in bench practice comprise another data domain that should be integrated with the
whole to allow comprehensive data analysis.
A clinical data tracking system needs to cover many aspects of clinical data and
biospecimen collection issues. The system not only needs to track normal clinical
data collection, specimen annotations, and lab values and test results, but also needs
to properly handle possible erroneous events such as an illegible value, an incor-
rectly transcribed value, a misunderstood survey question, a preliminary report
finding differing from the final report finding, a clinical lab system temporarily not
available, and a missing specimen—these are all examples of the need for checks
and balances in the processes of data collection, data entry, data quality control,
and data transmission, storage, and reporting. Well-executed data management
processes that provide effective stewardship at each step along the way contribute to
an essential and solid foundation for the research project. A more comprehensive
discussion of clinical data tracking is provided in Chapter 7.
The clinical team at the health care facility is responsible for the initial collec-
tion of most biomedical data gathered on the patient. Staff members, often research
nurses with clinical expertise, administer the survey(s) or questionnaire(s) associ-
ated with the research, and collect further medical information such as lab values
and test results. Once collected, either the same staff, or different staff manage the
tasks of quality review and data entry of what can be multiple data instruments.
After data entry is complete, another round of quality review/quality control tasks
should be implemented to ensure the data have been entered accurately and
thoroughly.

2.6.2 Deidentifying Data

Questionnaires, medical images such as mammograms and ultrasound images, and
biospecimens will need to be deidentified to the extent stated on the informed con-
sent and HIPAA authorizations. Yet, these data must also be tracked in an orga-
nized fashion for proper data management. A research-specific ID number—not the
subject’s Social Security number or hospital number or other personal identi-
fier—needs to identify all biospecimens and data collected for each enrolling indi-
vidual. The project must plan for how a research subject’s ID number will be
assigned, where and how the connection between the assigned ID number and the
20 The Clinical Perspective

subject’s name will be stored, and who, besides the principal investigator, will have
access to this information.
To this end, a database containing the patient’s demographic information that is
needed by clinic staff, as well as the patient’s unique, research-specific ID number
must be maintained in a secure way. Access to it should be password protected and
available to essential staff only. Documents containing both the patient’s name and
unique research ID number should be used and viewed only by essential staff in the
clinic. Paper files requiring storage that link the participant’s name with his or her
research-specific ID number should be locked and viewed only by essential staff.
Such files should be shredded when disposal becomes necessary. All data that will be
entered into the biomedical informatics platforms, and analyzed for research pur-
poses must contain the research-specific ID number of the patient, and must not con-
tain other patient identifiers such as names or medical record numbers, assuming
this level of deidentification has been promised to the patient within the informed
consent. Thus, the researchers conducting research analysis will see only a research
ID for the patient, but those clinical research staff who need the information will
have access to both the patient’s name and research ID number.
Biospecimens and questionnaires require a tracking barcode or unique identifier
of their own. That identifier links to the research subject’s specific ID number and
also links to the date these data are collected. This becomes essential when multiple
surveys, images, and biospecimens from one research subject are collected through-
out the lifetime of the project. One participant may have five yearly mammograms
on file in the biomedical informatics repository. Each mammogram needs to be iden-
tified as belonging to that unique patient (research-specific ID number), and each
needs to be identified as a unique data element of its own, associated with the date it
was obtained (e.g., a unique barcode ID number that maps back to the research-spe-
cific ID number and the date the mammogram was obtained).

2.6.3 Quality Assurance

The clinical team will need to communicate closely with the data management staff
that receive and review the questionnaires, biospecimens, and other clinical data
such as mammograms and ultrasound images. It is vital to have open communica-
tion channels for resolving discrepancies, as well as electronic data tracking tools for
making changes to data collected. Neither medical charting, nor the processes of
retrieving medical and social histories from patients are 100% reliable.
The process of obtaining as accurate and complete a data record as possible
from a patient can require much back and forth between data collectors and data
reviewers, depending on the volume of patients participating in the research study,
the volume of data collected, and the precision with which data is collected. For
example, a research nurse may record that a mammogram was completed on a
patient the same day she completed a research questionnaire, but forgets to include
the actual mammogram result on the questionnaire. The research staff conducting
the quality review of the questionnaire identifies the missing value and must commu-
nicate with the research nurse to obtain it.
At the CBCP, the Windber Research Institute developed a user-friendly online
data tracking tool that streamlines the process of identifying, communicating, and
2.6 Issues Important to Biomedical Informatics 21

resolving discrepancies in clinical data collected. The tool is called the Quality
Assurance Issue Tracking System (QAIT) [13]. There are thousands of enrollees in
the CBCP so far, and the project collects many hundreds of data elements from each
enrollee. Without precise methods for correcting errors, the project would lose vital
data strength and compromise its ability to conduct excellent research. The QAIT
provides secure online communication between the data entry technicians and the
quality specialists who review all the questionnaires for completeness and accuracy
before data are entered. The process is straightforward. When an issue requires clar-
ification or correction based on standard operating procedures, it is entered into the
QAIT by a data entry technician. A quality reviewer then opens that issue in the
QAIT, via secure online access, and assesses what needs to be done to resolve the
issue. This may require verifying information in a medical record or verifying infor-
mation through the research nurse who initially collected it. When the quality
reviewer successfully resolves the issue, the resolution is entered into the QAIT.
That issue is then closed by a data entry technician. The steps, dates, and times
along the way are recorded electronically by the QAIT. At any given moment, the
current stage of the issue’s progress is easily identified by the users. When the issue is
resolved, the data entry technician enters the correct value into the clinical data
tracking system CLWS, which is described in more detail in Chapter 7.
These processes are the standard operating procedures for quality checks and
balances using the QAIT, before ultimate data entry into the CLWS. Once data are
entered, further quality checks are applied in the form of a computer program called
QA Metrics, which implements hundreds of established quality assurance (QA)
rules to flag discordant data, given their relationship to each other. In addition to
data problem resolution, the QAIT provides reports that enable management to
better supervise, train, and track proficiency of staff involved in the data collection
and quality assurance processes. Figure 2.1 depicts a list of issues that require clari-
fication and/or correction by users of the QAIT.

2.6.4 Data Transfer from the Health Care Clinic to the Research Setting
The biomedical information collected within the clinical setting must become avail-
able to the research specialists and scientists who will use and analyze the informa-
tion. This may involve manual and/or electronic transfer of information. It may be a
matter of entry into the data tracking systems and repositories that are on site or a
matter of shipping to sites where the data and specimen management systems are
located. Regardless, quality control practices that ensure the integrity of clinical
data and biological specimens during transfer are of paramount importance.
Quality control methods for transferring biological specimens to the repository
are covered in Chapter 3. As previously discussed, all clinical data should undergo
quality review before being cleared for data entry. The data entry processes them-
selves should also have a quality oversight component. If the data tracking system is
one whereby data entry occurs directly, such as a survey administered electronically
on the Internet, with no opportunity for quality review before data entry, it is neces-
sary for quality control procedures to begin at the point after which the data is
entered, but before it is analyzed by the researchers.
 22
The Clinical Perspective
Figure 2.1 The list of current issues being generated in the QAIT online tool used by the Clinical Breast Care Project. Notice the status column. Issues that are ready
for a quality specialist to respond to are marked “Ready for QA.” Issues that are still being formulated by a data entry technician are marked “Data Entry.” Issues that
have been resolved by a quality specialist are marked “QA: (no pending issue)” along with the quality specialist’s name.
2.7 Standard Operating Procedures 23

The challenge of the quality review process is to ensure that the clinical data col-
lected are as accurate and complete as possible. An organized data management sys-
tem such as the QAIT tool described earlier, which tracks corrections and changes
to data, is invaluable for maintaining integrity when processing clinical data. Coor-
dination of the quality review takes continual involvement and oversight. A stan-
dard operating procedure that details the quality control process that clinical data
undergo may be essential for projects that collect many clinical data elements.

2.7 Standard Operating Procedures

Written policies and procedures can greatly assist in ensuring standardized adher-
ence to regulations, guidelines, local policies, and institutional directives. Standard
operating procedures (SOPs) provide staff with easy-to-locate guidance. They
should be living documents that are developed, shared with appropriate staff,
reviewed, and revised periodically. The format of SOPs should be standardized.
Staff can then rely on the existence of SOPs within a familiar format that is easy to
find within the clinic and/or research setting.
Each SOP should be titled with the subject or standard it addresses. It should
then state why the organization has this standard and who must comply with it.
Finally, the steps of the procedure should be clearly delineated. A biomedical
research project will require an SOP for the collection, storage, and shipping of
biospecimens. Also important is an SOP for gaining informed consent from
patients. It is also useful to have an SOP for receiving, reviewing, and entering ques-
tionnaire data. Figure 2.2 shows the table of contents for a SOP that delineates how
the core questionnaire from the CBCP is processed after it has been administered to
a research participant.
SOPs define the scope of practice for an individual or a program. They provide
guidelines for new members of the team. They document roles and responsibilities
for the research team, and they set the standards of good practice for the operation
of the research project. It is worthwhile to keep these benefits in mind when faced
with how time consuming it is to write useful SOPs and the fact that they must be
periodically reviewed and updated to reflect changes. Make every effort to write
SOPs that can be, and are, enforced in practice. Deviations from them may not be
viewed kindly by an auditor.

2.8 Developing and Implementing a Research Protocol

The research protocol must be developed and submitted to the IRB affiliated with
the organization where the research will be conducted, and must be approved prior
to the beginning of any research. As previously discussed, IRBs exist to ensure that
research conducted within the organization is well designed and without undue risk
to participants, and that human subjects are protected from unethical practices.
Augmented by implications of the genomic era, as well as regulatory mandates to
protect personal health information, protecting privacy has also become an impor-
tant concern to IRBs. The specific requirements for writing and submitting a
24 The Clinical Perspective

Figure 2.2 Table of contents for a SOP used in the Clinical Breast Care Project to standardize the
way research questionnaires are processed.

research protocol to an IRB are unique to any given organization, yet because the
goals of an IRB are the same everywhere, the required elements of a protocol are
fairly consistent. Organizations such as the U.S. Food and Drug Administration
(http://www.fda.gov/oc/gcp/guidance.html#guidance), the Office for Human
Research Protections (http://www.hhs.gov/ohrp/education/#activities), the
National Institutes of Health (http://clinicalcenter.nih.gov/researchers/train-
ing/ippcr.shtml), and the Society of Clinical Research Associates
(http://www.socra.org) offer courses and educational material on developing clini-
cal research protocols. Reed and Jones-Wells [14] suggest that a protocol can be
thought of as a blueprint for what is to be done, how it is to be done, and who is to
assume ultimate responsibility for it. In an effective protocol, these three basic
questions of what, how, and who are fully addressed.

2.8.1 Developing a Research Protocol

The cover page of a protocol includes the title of the study; the name, title, and con-
tact information of the principal investigator, associate investigators, or collabora-
tors; and, if required, medical monitor information. Figure 2.3 shows a sample cover
page for a protocol.
The principal investigator (PI)is the individual primarily responsible for the
actual execution of the clinical investigation. He or she is responsible for the conduct
of the study, obtaining the consent of research subjects, providing necessary reports,
2.8 Developing and Implementing a Research Protocol 25

Figure 2.3 Sample cover page of a protocol to study breast disease by collecting biological speci-
mens from research subjects at two different clinical sites, and storing and analyzing the specimens
at a research facility.

and maintaining study documents. Some or all of these responsibilities can be,
should be, and often are delegated to other trained professionals with the requisite
expertise, under the supervision of the PI. The PI is usually affiliated with the institu-
tion where the research is to be conducted; if that is not the case, however, an
administrative point of contact will be identified.
Associate investigators are other key personnel involved in the governance of
the study who are affiliated with the sponsoring institution. For the example of a
26 The Clinical Perspective

protocol to study breast disease, these may be a breast surgeon, a pathologist, a med-
ical oncologist, a radiation oncologist, a radiologist, and so forth.
Collaborating personnel are all other key personnel involved in the governance
of the study that come from outside the sponsoring institution. In the example, the
collaborating personnel identified could be the vice president of the research facility
where the biorepository is located, and the breast surgeon consenting patients to the
protocol at the second site.
A medical monitor is a physician who monitors the study from a medical safety
perspective and is not involved in any way with the research study. When a study is
assessed to pose greater than minimal risk to its subjects, a medical monitor will
likely be necessary. The responsibilities of a medical monitor include:

1. Periodic review of research files, medical records, and so on;

2. Review of annual progress reports, adverse events, protocol deviations;
3. The ability to terminate the study or participation in the study if there is
concern about the welfare of enrolled subjects.

Many research studies only pose minimal risk to participants and may not
require a medical monitor. Minimal risk to human subject participants is defined
within the CFR: “Minimal risk means that the probability and magnitude of harm
or discomfort anticipated in the research are not greater in and of themselves than
those ordinarily encountered in daily life or during the performance of routine phys-
ical or psychological examinations or tests” [15].
The body of a research protocol should be composed of the following sections,
or variations thereof:

Objectives
The objectives of the protocol are stated, including the research hypothesis if any.
Research objectives should be consistent with the Plan and Data Analysis sections.
The medical application or medical importance and usefulness of the results of the
study are often cited in the objectives.

Background and Significance

The Background and Significance section of a protocol is necessary to justify the
need for the project and explain the rationale. It includes a summary of relevant lit-
erature. Making this section easy to understand, especially to those outside of the
specialty of the study, is important.

Plan and Data Analysis

The protocol’s plan is comprised of many subsections that together delineate the
broad scope of processes and methodologies that will be implemented by the proto-
col. Each subsection must be addressed briefly. This assures the IRB that the plan for
implementing the protocol has been thoroughly conceived. The subsections include,
but may not be limited to:
2.8 Developing and Implementing a Research Protocol 27

• Study design;
• Recruitment methods, with inclusion and exclusion criteria for subject partic-
ipation;
• An estimate of the sample size planned for the study;
• Anticipated start date and expected completion date of the study;
• Data collection practices, including types of data collected as well as practices
for ensuring patient confidentiality, including processes for handling PHI as
defined by HIPAA;
• Data analysis practices, that is, a general description of how the data collected
will be analyzed and by whom.

When a study includes the collection and storage of human biological speci-
mens, additional information will be required. Such information will include but
may not be limited to:

• Obtaining informed consent for future use of specimens and clinical data;
• Specimen collection procedures;
• Specimen shipping procedures;
• Specimen storage procedures;
• Specimen confidentiality, and how it will be maintained;
• Who will have access to specimens;
• Length of time specimens will be stored;
• Withdrawal and destruction of specimens.

The research office associated with the IRB that will approve your protocol may
have valuable guidelines and/or standard formats for submitting the plan, as well as
other sections of the protocol.

References
Include a section of references that have been cited in the protocol.

Budget
The budget for the research project is submitted with the protocol. Individual IRBs
usually provide their own specific, required budget formats.

Attachments Relevant to the Research

The IRB will require copies of documents to be used in conducting the research such
as consent forms, questionnaires/survey instruments, brochures describing the
research, and marketing material. If additional, collaborative protocols or grant
proposals are affiliated with the project, these will be required by the IRB.
Developing a protocol is a time-consuming, labor-intensive undertaking. There
will be many drafts, revisions, and consultations with associate investigators before
a final document is submitted to an IRB. Once the review process begins with the
IRB, correspondence and communication will require additional clarification and
revision before a final version is approved. Although all IRBs must adhere to the
Common Rule, individual IRBs may function quite differently from one another. It
28 The Clinical Perspective

behooves the principal investigator and study coordinator to become familiar with
the requirements of their IRB. Attention to detail in this phase of research will prove
beneficial as the study is conducted.
After the IRB has granted initial approval of a research protocol, a yearly, con-
tinuing review will be required. The IRB will provide the format for what is called a
Continuing Review application. This is likely to include a progress report of the
number of enrollees during the past year, as well as the duration of the study,
changes in such things as survey instruments, consent forms, or any changes to the
protocol’s activities that require approval from an IRB. If any significant events
requiring regulatory reporting have occurred, the original report to the IRB of the
event is included in the progress report.

2.8.2 Implementing the Research Protocol

Once the protocol, consent form, questionnaire/survey instrument(s), and HIPAA
authorization have been approved by the IRB, and the biomedical informatics sys-
tems and repositories are functioning to receive data and specimens, a planned
implementation of the research protocol can take place. Clinical staff responsible for
specimen acquisition need training in exactly how specimens will be obtained, trans-
ported within the clinic, stored temporarily, and shipped to the repository. Team
planning and “dry runs” for every aspect of the acquisition process, but with no
actual specimen, should be considered. This will help ensure that the first batch of
biological specimens sent to the repository location will not be compromised due to
unforeseen problems. Similar “dry runs” should be done with subject enrollment
and clinical data management.
All staff responsible for obtaining informed consent, administering survey
instruments to subjects, and conducting quality oversight on the biomedical infor-
mation collected will need to be thoroughly knowledgeable about the protocol and
proficient in their respective specialties. Completion of a live or online course in the
protection of human subjects is recommended for all staff directly involved in the
research and is required by many IRBs. Courses may be offered through professional
organizations such as the Society of Clinical Research Associates [16]. The Office
for Human Research Protections and the National Cancer Institute offer free
web-based resources for human subjects training [17, 18]. Information about inter-
national codes and ethics for human subject protections is also covered by these
sources. The Collaborative IRB Training Initiative (CITI), out of the University of
Miami, provides a comprehensive selection of educational modules for biomedical
and social and behavioral researchers. Continuing education modules are also
offered as a refresher to ensure that the research staff remains competent in the area
of Human Subject Protection [19]. Staff responsible for consenting subjects to the
protocol, as well as completing the questionnaire/survey instrument, should be
encouraged to keep track of any issues or problems experienced with the consent
form or the survey instrument. Staff meetings are a good forum to discuss “best
practices” in administration of the instrument and consenting of subjects. Also dis-
cuss any questions frequently asked by subjects to help determine which questions
on the survey may need special attention and clarification. This will augment reli-
ability and standardization of both the administration of the survey instrument and
2.9 Summary 29

the obtaining of informed consent. A successful biomedical informatics research

initiative will include a solid orientation and continuing education programs for
staff.

2.9 Summary

In summary, preparing the documents and the processes for ensuring a well-run
protocol is a time-consuming endeavor that requires collaboration from many
experts. The administration of the protocol must be planned, the formal protocol
written, and the informed consent and HIPAA authorization documents written. A
survey instrument must be developed. Approval from the IRB of the organization(s)
conducting research must be obtained. All staff participating in research activities
must be trained for their unique roles in the study as well as in research ethics,
including the protection of human subjects. Data management systems for ensuring
integrity and confidentiality of biospecimens and clinical data must be developed,
implemented, and continually managed. Quality oversight and ongoing staff educa-
tion must be conducted with regularity. Biospecimens and clinical data must be suc-
cessfully transferred from the clinic setting to the setting in which the scientific
research will be carried forward.

References

[1] “The Nuremberg Code,” available at http://www.hhs.gov/ohrp/references/nurcode.htm.

[2] World Medical Association, http://www.wma.net/e/policy/b3.htm.
[3] U.S. Congress, House of Representatives, Bill H.R. 7724, Public Law 93-348, July 12,
1974.
[4] “The Belmont Report,” available at http://www.hhs.gov/ohrp/humansubjects/guidance/
belmont.htm.
[5] U.S. Department of Health and Human Services, http://www.hhs.gov/ohrp/humansubjects/
guidance/45cfr46.htm.
[6] U.S. Department of Health and Human Services, http://www.hhs.gov/ohrp/
45CFRpt46faq.html.
[7] Council for International Organizations of Medical Sciences (CIOMS),
http://www.cioms.ch/frame_guidelines_nov_2002.htm.
[8] “The Personal Information Protection and Electronic Document Act,” available at
http://www.privcom.gc.ca/legislation/02_06_01_e.asp.
[9] “Eur-Lex, Directive 95/46/EC of the European Parliament and of the Council of 24 Octo-
ber 1995 on the Protection of Individuals with Regard to the Processing of Personal Data
and on the Free Movement of Such Data,” available at http://eur-lex.europa.eu/LexUriServ/
LexUriServ.do?uri=CELEX:31995L0046:EN:HTML.
[10] “Public Law 104-191,” available at http://aspe.hhs.gov/admnsimp/pl104191.htm.
[11] “Summary of the HIPAA Privacy Rule,” available at http://www.hhs.gov/ocr/
privacysummary.pdf.
[12] Hochhauser, M., “Informed Consent Reading, Understanding and Plain English,” SoCRA
Source, Issue 42, November 2004, pp. 24–26.
[13] Zhang, Y., et al., “QAIT: A Quality Assurance Issue Tracking Tool to Facilitate the
Enhancement of Clinical Data Quality” (submitted for publication).
30 The Clinical Perspective

[14] Reed, E., and A. Jones-Wells, “Writing a Clinical Protocol: The Mechanics,” in Principles
and Practice of Clinical Research, J. I. Gallin (Ed.), San Diego, CA: Academic Press, 2002,
p. 439.
[15] Code of Federal Regulations, Title 45 Public Welfare; Subtitle A; Department of Health and
Human Services, Part 46, Subsection 46.102.
[16] Society of Clinical Research Associates, http://www.socra.org.
[17] U.S. Department of Health and Human Services, Office for Human Research Protections,
http://www.hhs.gov/ohrp/education/#materials.
[18] “National Cancer Institute, Human Participant Protections Education for Research
Teams,” available at http://cme.cancer.gov/clinicaltrials/learning/humanparticipant-
protections.asp.
[19] “CITI Collaborative Institutional Training Initiative,” available at http://www.citiprogram.
org.
 CHAPTER 3

Tissue Banking: Collection, Processing,

and Pathologic Characterization of
Biospecimens for Research
Jeffrey Hooke, Leigh Fantacone, and Craig Shriver

This chapter provides an overview of tissue banking practices and is intended to

serve as a guide that can be adapted, as appropriate, to the mission and scientific
needs of individual biospecimen resources. Details about laboratory methods and
procedures can be found elsewhere [1]. The principles set forth in this discussion are
based on a model developed by the Clinical Breast Care Project (CBCP), a joint
breast cancer research project between Walter Reed Army Medical Center and
Windber Research Institute that was established in 2001. Although the focus of the
CBCP is breast disease, the principles described herein can be applied, with only
slight modifications, to other organ systems as well.

3.1 Introduction

3.1.1 A Biorepository’s Mandate

In today’s era of high-throughput genomic and proteomic research, there is a
demand for high-quality and well-annotated tissue. Unfortunately, human speci-
mens when available are often unsuitable for genomic and proteomics research
because of the type of preservation method (i.e., formalin-fixed paraffin-embedded
tissue instead of frozen tissue) and the absence of supporting clinical and demo-
graphic data. Although emerging technologies for retrieval of ribonucleic acid
(RNA) and protein from formalin-fixed paraffin-embedded (FFPE) tissue may per-
mit the use of these samples for genomics and proteomics research, the lack of uni-
form tissue acquisition procedures and standardized protocols for specimen
handling may limit the usefulness of many biologic specimens and preclude compar-
isons of research results obtained at different institutions. Finally, the type of
informed consent obtained from many biorepositories is not sufficiently robust to
allow the unrestricted use of these specimens for long-term follow-up.
The goal of a biorepository is to increase the quantity and variety of high-qual-
ity samples. All types of biospecimens (diseased and nondiseased tissue, blood and
its various components, bone marrow, and so on) are potentially useful for

31
32 Tissue Banking

research, and great care must be exercised to ensure that they are handled in a man-
ner that will not compromise their value. The discussion that follows outlines the
components required for establishing and managing a successful biorepository in
today’s high-throughput world of translational genomics and proteomics research.

3.1.2 Overview of Current Tissue Banking Practices

Human tissue repositories have existed in the United States in one form or another
for more than 100 years. These repositories are located at a variety of sites includ-
ing the Armed Forces Institute of Pathology (AFIP), the National Institutes of Health
(NIH) and its sponsored facilities, other federal agencies, diagnostic pathology
laboratories, academic institutions, hospital-based research laboratories, commer-
cial institutions, and nonprofit organizations. These tissue collections are diverse,
ranging from large multisite operations to informal collections in a researcher’s
freezer [2].
The National Pathology Repository, located at the AFIP, is perhaps the largest
collection of pathology material in the world, housing more than 50 million tis-
sue-mounted slides and 30 million paraffin-embedded tissue blocks [3]. The tissue
repositories supported by the NIH are smaller than those of the AFIP, but the NIH is
the largest funding source for tissue repositories. For example, the National Cancer
Institute (NCI), a component of the NIH and the federal government’s principal
agency for cancer research, supports a variety of research initiatives including the
Cooperative Human Tissue Network (CHTN) [4], the Clinical Trial Cooperative
Group Human Tissue Resources [5], and the Specialized Programs of Research
Excellence (SPOREs) [6]. Collectively, the tissue repositories of university- and hos-
pital-based laboratories represent the oldest and largest collection of tissue speci-
mens in the United States, with some paraffin-embedded tissue blocks dating back
more than a century [2].
In many countries, government-sponsored biorepository initiatives have been
established, particularly in recent years, to help understand and manage the health
care needs of their people. The Swedish National Biobanking Program is a nation-
wide collaboration that was created to “increase the quality, usefulness, efficiency,
and accessibility of Swedish biobanks for health-related research as well as for clini-
cal care and treatment” [7]. The Swedish program has at least 10 unique biobanks,
with a total collection estimated between 50 and 100 million samples [8]. Similar
projects exist in Canada, Norway, and the United Kingdom. The UK Biobank,
started in 2007, is a large, long-term study in the United Kingdom, which is investi-
gating the contributions of genetic and environmental exposures to the development
of disease. The biobank plans to collect blood and urine samples, along with exten-
sive clinical and demographic data, from 500,000 volunteers ages 40 to 69. These
patients will be followed for 25 years [9].
One of the best known private industry endeavors is led by deCode Genetics in
Iceland. Created in 1996, deCode has created a bank of genetic samples from
100,000 volunteers. Its mission is to identify human genes associated with common
diseases using population studies and to apply the knowledge gained to guide drug
discovery and development. So far, the company has isolated genes believed to be
involved in cardiovascular disease and cancer [10].
3.2 Consenting and Clinical Data Acquisition 33

3.2 Consenting and Clinical Data Acquisition

Whereas biospecimens used by basic researchers may require only limited clinical
information, biospecimens collected for translational research (e.g., target identifi-
cation or validation) require more in-depth associated clinical data, such as medical
and family histories, treatment, and clinical outcomes data. All data collected must
be entered into the repository’s biomedical informatics system, where it can be inte-
grated with data from the genomics and proteomics research platforms. Uniform,
approved protocols must be used to enroll donors and obtain biospecimens. Donor
consent and authorization to utilize blood and tissue samples is typically done by
dedicated personnel such as nurse case managers and research nurses, who have
received appropriate training in the consenting process and in the rules and regula-
tions governing the use of human subjects in research, and can act on behalf of the
principal investigator in the consenting process. The study coordinators must
describe the study to the patient, obtain consent, and, whenever possible, complete
the study-specific questionnaire with the donor.
Potential participants should be given information related to the kinds of
biospecimens that may be taken and the kinds of studies the samples will be used
for. Note that in a clinical setting potential participants for tissue and blood dona-
tion typically present with an imaged or palpable lesion that needs further evalua-
tion. As a result, disease-negative patients cannot be prospectively enrolled as
control subjects. Should tissue samples be collected from a patient with no patho-
logic findings, the samples can be used as controls. The consent for the collection
and utilization of specimens should be explicit and separate from the routine surgi-
cal consent. Donors should be given ample opportunity to ask questions and to
decline participation if they choose. At every stage, personnel should make the
potential donors feel comfortable and emphasize that their confidentiality and
health care needs are not compromised. Once all clinical data have been collected,
they must be entered into the repository’s biomedical informatics system. Proce-
dures for ensuring accurate data entry are crucial. Accurate data entry is accom-
plished by using standardized terminology and forms. Automated techniques to flag
discrepancies and track errors and reconciliation also help to ensure the integrity of
the information being collected. The consenting process and issues related to it were
discussed in greater detail in Chapter 2.

3.3 Blood Collection, Processing, and Storage

Peripheral blood is an attractive sample source for biomedical research because a

minimally invasive procedure is all that is required to collect a sample (i.e., a simple
blood stick versus a tissue biopsy). The spectrum of genetic and physiological events
associated with tumors and other diseased states can be potential targets for molec-
ular detection in the blood of patients. The analysis of blood RNA, for example,
holds promise for noninvasive gene expression profiling, and the discovery of
tumor-derived deoxyribonucleic acid (DNA) in the plasma of cancer patients has
provided an alternative method for cancer detection, monitoring, and prognosis
[11, 12].
34 Tissue Banking

Blood samples should be collected as early in the diagnosis as possible. Samples

collected in vivo (while cancerous tissue is still present within the body) are best for
these molecular studies. In particular, blood samples taken prior to biopsy are ideal
since tissue has not yet been manipulated and the circulating blood has not been sub-
jected to medical agents and physiological responses associated with surgical proce-
dures. The amount of blood that is drawn from a patient and the types of collection
tubes that are required depend on the specific objectives of a researcher. In each case,
it is imperative that the entire process from sample collection to analysis be stan-
dardized so that reliable and reproducible results are achieved. Nurses,
phlebotomists, and laboratory personnel play critical roles in this process.
In contrast to DNA and protein—both of which are relatively stable in blood
preparations—RNA is unstable in vitro and poses unique technical challenges.
Enabling technologies such as quantitative reverse transcriptase polymerase reac-
tion (RT-PCR) and microarrays require high-quality, highly purified RNA samples.
Preparing such samples is difficult due to the presence of endogenous RNAs that eas-
ily degrade the RNA and to the induced expression of certain genes after the blood is
drawn. Both RNA degradation and in vitro gene induction can lead to an underesti-
mation or overestimation of gene transcript number. Fortunately, techniques such
as the PAXgene Blood RNA System (QIAGEN Inc.) now exist that allow for the iso-
lation, stabilization, and purification of cell-associated blood RNA. This technology
has allowed even isolated laboratories that lack RNA-handling expertise to venture
into the arena of gene expression profiling.
For biorepositories involved in cancer research, two or three tubes of blood may
be required; however, the actual amount of blood that is drawn should be limited to
meet the objectives of the particular study. Three blood tubes in common usage are
the serum tube (BD Vacutainer with red closure), the plasma tube (BD Vacutainer
with green closure), and the PAXgene RNA tube. For genomic and proteomic-based
applications, the tubes are utilized as follows:

• Serum tube: Fractionated into serum and clot. Serum used for analysis of
blood proteins. Clot used for a variety of downstream DNA analyses.
• Plasma tube: Fractionated into plasma and cell fraction. Plasma used for anal-
ysis of blood proteins and identification of plasma-associated DNA. Cell frac-
tion used for cell-associated DNA studies.
• PAXgene RNA tube: Used exclusively for the isolation and stabilization of
RNA.

Both the serum and plasma tubes are fractionated into their components using a
centrifuge, the components aliquotted into cryovials, and then placed into a −80°C
mechanical freezer. The PAXgene tube is incubated at room temperature for at least
2 hr (overnight is best) to allow complete lysis of blood cells, before processing or
freezing. At all times standard operating procedures must be adhered to. For blood
collection, the standard procedure is to fill collection tubes in a set order (serum,
plasma, RNA), followed by inversion of the tubes to allow proper mixing with the
reagent in each tube (serum tube: clot activator; plasma tube: sodium heparin;
PAXgene RNA tube: RNA stabilizing reagent). In the case of the PAXgene RNA
tube, failure to completely fill the PAXgene tube or failure to invert the tube follow-
3.4 Tissue Collection, Processing, Archiving, and Annotation 35

ing collection may result in inadequate RNA yield [13]. Standard procedures for
processing blood samples should detail the appropriate supplies to be used,
centrifugation speed and times, aliquot volumes, storage temperatures, and ship-
ping specifications (Figure 3.1). A bar-coding system to track the samples from the
time of collection through the time of storage and utilization is advantageous. The
frozen blood samples can be maintained on site in a −80°C freezer or transported
via overnight service (e.g., FedEx or private courier) to another site in a carefully
packaged container containing dry ice. Standardized and carefully monitored ship-
ping procedures should be integrated with a biorepository’s informatics system
whenever possible.

3.4 Tissue Collection, Processing, Archiving, and Annotation

3.4.1 Tissue Collection

Most existing biorepositories collect tissue from patients undergoing surgery or
other diagnostic procedure. A typical scenario for tissue collection is as follows:
Prospective patients are counseled and informed consent is obtained for tissue sam-

Blood Collection

Red top Green top PAXgene

Vacutainer tube Vacutainer tube RNA tube

Centrifuge @ 3000 rpm

for 10 min

1. Room temp for 4hrs

Serum Clot Plasma Cells 2. Store @ -80°C
3. Ship on dry ice

1. Store @ -80°C
2. Ship on dry ice

Protein DNA Protein DNA RNA

Figure 3.1 Blood processing workflow.

Exploring the Variety of Random
Documents with Different Content
(Cerambyx L.), amongst the Coleoptera, are without them, and probably
several others; and amongst the Diptera, all those with a membranous or
variable head. Those of the remaining orders, with the exception, perhaps,
of some Hymenoptera and Lepidoptera, are furnished with these organs;
and in the Coleoptera all the predaceous tribes, as well as most of those
that are herbivorous or granivorous, and the Gnats and other Tipulidans
(Tipulariæ Latr.) in the Diptera, are also distinguished by them. In the
larvæ of the dragon-flies (Libellula L.), and other Neuroptera, they are
composed of many facets as in those of the perfect insect, from which
they differ chiefly in being smaller. But in the other insects of this
description they are simple, and resemble those of the Arachnida, and
many aptera. These simple eyes vary in their number, in different genera
and tribes, from one to six on each side of the head. Thus the larva of
Telephorus, and the saw-flies, has only one[237]; that of Cicindela three,
the two posterior ones being large with a red pupil surrounded by a paler
iris, which adds to the fierce aspect of this animal; and the anterior one
very minute. Those of the tortoise-beetles also (Cassida) have three[238];
of Staphylinus, four; of Timarcha (the bloody-nosed beetle) five; of
Carabus, and the Lepidoptera in general, six. In the last they are of
different sizes, and generally arranged in a circle: in that of Hemerobius
there are five in a circle, with one central one[239]. The appearance of
these globules, which are often not visible but under a powerful lens, is so
different from that of the eyes of a butterfly or moth, or other perfect
insect, that it has been doubted whether they actually perform the office
of eyes, but without reason. They occupy the usual station of those
organs, being situated in many instances upon a protuberance which
appears to incase them; and seem of a construction closely analogous to
that of the eyes of spiders, and the stemmata or ocelli of Hymenoptera,
which have been satisfactorily proved to be organs of vision. In the larva
of a moth not yet ascertained to exist in this country, Attacus Tau, and
probably other species, the eyes, after the skin has been changed a few
times, are no longer to be seen[240].
Antennæ. Most larvæ are provided with organs near the base of the
mandibles, which from their situation and figure may be regarded as
antennæ. Fabricius has asserted that the larvæ of the saw-flies (Tenthredo
L.) have no antennæ; but in this he was mistaken, for though very short,
they are discoverable in them, as he might have learned by consulting De
Geer[241]. In the majority of Neuropterous larvæ, they almost precisely
resemble those of the perfect insect. In all the rest they are very different.
The antennæ of Coleopterous larvæ are usually either filiform or
setaceous, consisting of four or five joints, nearly equal in length. Those of
Lepidopterous larvæ are commonly conical, as are those likewise of
Chrysomela and Coccinella &c. amongst the Coleoptera, and very short,
composed of two or three joints, of which the last is much thinner than
the first, and ends in one or two hairs or bristles. These antennæ the larva
has the power of protruding or retracting at pleasure. Lyonnet informs us,
that the caterpillar of the great goat-moth (Cossus ligniperda) can draw
the joints of its antennæ one within the other, so as nearly to conceal the
whole[242]. The larva of the common gnat has two long incurved
setaceous antennæ, fringed with hairs at some distance from their apex,
which consist only of a single joint[243]. The greater number of Dipterous
larvæ, however, all indeed except the Tipulidans (Tipulariæ Latr.), and
many belonging to the Coleoptera and Hymenoptera orders (as those of
Curculio, Apion, Apis, &c.), are wholly deprived of antennæ. It is a general
rule, that the antennæ of larvæ are shorter than the same organs in the
perfect insect, the tribe Ephemerina perhaps affording the only example in
which the reverse of this takes place[244].
Mouth. All larvæ have a mouth situated in the head, by which they receive
their food, and furnished with one or more instruments for the purpose of
mastication and deglutition. These instruments, in all the orders except
Lepidoptera, some Neuroptera and Diptera, bear a general resemblance to
the same parts in the perfect insect. In larvæ of the Coleopterous,
Lepidopterous, and Hymenopterous orders, we can distinguish for the
most part an upper and under lip; two pairs of jaws answering to the
mandibulæ and maxillæ; and two, four, or six palpi[245]: and some of
these instruments may be found in most Diptera. Each of these parts
require separate notice.
Upper-lip (Labrum). The mouth of almost all larvæ, except some of the
order Diptera, are provided with a distinct upper-lip, for retaining their
food during mastication. As the construction of this part does not widely
differ from that of the perfect insect, which will hereafter be more fully
described, it is only necessary to observe, that it is usually a transverse
moveable plate, attached posteriorly to the nasus (clypeus F.), and
situated just above the mandibles[246].
Upper-jaws (Mandibulæ). The most usual figure of these, which are of a
hard horny consistence[247], is that of two slightly concave, oblong, or
triangular plates, often at their lower extremity of considerable thickness,
and of very irregular form, the base of which is filled with powerful
muscles, and planted in the side of the mouth so as to move transversely.
The other extremity can be made to meet or diverge like the claws of
pincers, and are divided into one or more tooth-like indentations, by
means of which the food of the larva is cut[248]. This is their construction
in the larvæ of all Lepidoptera, and in many of those of the other orders.
They frequently, however, assume a different form, though their situation
is always the same. Thus in the larvæ of the capricorn beetles (Cerambyx
L.) and of other wood-boring species, they are shaped like the half of a
cone, the inner sides of which, applying close to each other, form a pair of
powerful grindstones, capable of comminuting the hardest timber[249]. M.
Cuvier has observed, with regard to the mandibulæ of those of stag-
beetles (Lucanus), that besides their teeth at the extremity, they have
towards their base a flat striated molary surface; so that they both cut and
grind their ligneous food[250]. It seems to have escaped him, that a similar
structure takes place in many perfect insects of the lamellicorn tribe, as I
shall hereafter show you. In the larvæ of the water-beetles (Dytiscus L.),
ant-lions (Myrmeleon L.), and lace-winged flies (Hemerobius L.), they
resemble somewhat the forceps at the tail of an ear-wig, being long and
incurved; and, what is more remarkable, hollow and perforated at the end,
so as to serve as a channel for conveying into the larva's mouth the juices
of the prey which by their aid it has seized. Reaumur even asserts, that
the larva of Myrmeleon has no other entrance into its throat than through
these tubular mandibles[251]. That of the rove-beetles (Staphylinus L.),
and of many other Coleopterous genera, have these organs of this
forcipate construction, without being perforated[252]. In the larva of the
carnivorous flies, and many other Diptera, are two black incurved subulate
parts, connected at the base, and capable of being protruded out of, and
retracted into, the head, through the skin of which they are usually visible.
As I informed you in a former letter[253], these mandibles are used for
walking as well as feeding: they are parallel to each other, and are neither
formed for cutting nor grinding like the mandibles of other insects, but
merely detach particles of food by digging into it and tearing the fibres
asunder. In this operation they are probably assisted by an acutely
triangular dart-like instrument of a horny substance, which in some species
(Musca vomitoria) is placed between the two. In others this part is
wanting. Some Dipterous larvæ have two similar mandibles, but instead of
being parallel, they are placed one above the other; others (Musca
domestica and meridiana) have but one such mandible, and some have no
perceptible mandible of any kind. The mandibles of the larva of the crane-
flies (Tipula), which are transverse and unguiform, do not act against each
other, but against two other fixed, internally concave and externally
convex, and dentated pieces[254].
Under-jaws (Maxillæ). These are a pair of organs, usually of a softer
consistence, placed immediately under the upper-jaws; but as they are
usually so formed and situated as not to have any action upon each other,
it is probable that in general they rather assist in submitting the food to
the action of the mandibulæ, than in the comminution of it. In
Lepidopterous larvæ they appear to be conical or cylindrical (at least in
that of the cossus so admirably figured by Lyonnet[255]), and to consist of
two joints; which may, I imagine, be analogous to the upper and lower
portions of which the maxillæ of perfect insects usually consist. The last of
these joints is surmounted by two smaller jointed palpiform organs. If any
part of the maxillæ can act upon each other, it is these organs or palpi;
but it is evident they are not calculated for mastication, although they may
assist in the retention of the substance to be masticated. In a figure given
by Reaumur of the under side of the head of another lepidopterous larva
(Erminea Pomonella), the maxillæ consist of a single joint, and appear to
be crowned by chelate palpi[256]: a circumstance which is also observable
in that of a common species of stag-beetle (Lucanus parallelipipedus), the
weevil of the water-hemlock (Lixus paraplecticus[257]), and other insects.
In general the maxillæ of larvæ are without the lobe or lobes discoverable
in those of most perfect insects, this part being usually represented by a
kind of nipple, or palpiform jointed process, strictly analogous to the
interior maxillary palpi of the predaceous coleoptera; but in most of the
lamellicorn beetles the lobe exists in its proper form[258], as it does
likewise in that of the capricorn-beetle before noticed (Callidium
violaceum[259]). In the former instance, it is armed with spines or claws;
but in the latter it is unarmed, and rounded at the end. In the larva of
Cicindela campestris, the base of the maxilla runs in a transverse direction
from the mentum, to which, as is usually the case, it is attached. From this
at right angles proceeds the lobe, from the outer side of which the feeler
emerges; and the inner part terminates in an unguiform joint, ending in
two or three bristles. The structure in the larvæ of water-beetles (Dytiscus
L.) is different, for they appear to be without maxillæ[260]; but the case
really seems to be, that these organs are represented by the first joint of
what M. Cuvier calls their palpi[261]; from which proceed the real palpi, the
interior one being very short, and consisting only of a single joint. These
maxillæ of larvæ were regarded by Reaumur and other writers as parts of
the under-lip, on each side of which they are situated; and indeed, as well
as those in the perfect insect, they form a part of the same machine,
being connected by their base with the mentum, which is part of the
labium, but they are clearly analogous to the maxillæ of the imago. They
are not to be found in the larvæ of many Dipterous insects, and perhaps in
some species belonging to other orders. In some Neuropterous larvæ, as
those of the Libellulina MacLeay, the maxillæ are of a substance quite as
solid and horny as the mandibles, which in every respect they
resemble[262].
Under-lip (Labium). Between the two maxillæ in the larvæ of most of the
insects under consideration is a part termed by Reaumur the middle
division of the under-lip, but which is in fact analogous to the whole of
that organ in the imago. This organ varies in shape, being sometimes
quadrangular, often conical, &c. Interiorly it is frequently connected with a
more fleshy protuberance, called the tongue by Reaumur[263], and
supplying the place of the ligula in the perfect insect. On each side of the
apex of the under-lip is a minute feeler, and in the middle between these
in the Lepidoptera and many others, is a filiform organ, which I shall call
the spinneret (Fusulus), through which the larva draws the silken thread
employed in fabricating its cocoon, preparatory to assuming the pupa
state, and for other purposes[264]. This organ is found only in those larvæ
which have the power of spinning silk; that is, in all Lepidoptera, most
Hymenoptera, Trichoptera, some Neuroptera, and even a Dipterous
insect[265]. This tube, Lyonnet had reason to believe, is composed of
longitudinal slips, alternately corneous and membranous, so as to give the
insect the power of contracting its diameter, and thus making the thread
thicker or smaller. There is only a single orifice at the end, which is cut
obliquely, somewhat like a pen, only with less obliquity, and without a
point, the opening being below, so as to be conveniently applicable to the
bodies on which the larva is placed. Reaumur conceived that this spinneret
had two orifices; but Lyonnet ascertained this to be a mistake, the two silk
tubes uniting into one before they reach the orifice. From the contractile
nature of the sides and the form of the orifice, combined with the power
the insect has of moving it in every direction, results the great difference
which we see in the breadth and form of the threads, some being seven or
eight times as thick as others, some cylindrical, others flat, others
channelled, and others of different thickness in different parts[266]. In the
larvæ of many Diptera the under-lip is merely a small tubercle, which can
be protruded from the insect's mouth by pressure[267].
One of the most remarkable prepensile instruments, in which the art and
skill of a Divine Mechanician are singularly conspicuous, and which appears
to be without a parallel in the insect world, may be seen in the under-lip of
the various species of dragon-fly (Libellula L.). In other larvæ this part is
usually small and inconspicuous, and serves merely for retaining the food
and assisting in its deglutition; but in these it is by far the largest organ of
the mouth, which when closed it entirely conceals; and it not only retains
but actually seizes the animal's prey, by means of a very singular pair of
jaws with which it is furnished. Conceive your under-lip (to have recourse,
as Reaumur on another occasion[268], to such comparison,) to be horny
instead of fleshy, and to be elongated perpendicularly downwards[269], so
as to wrap over your chin and extend to its bottom,—that this elongation
is there expanded into a triangular convex plate[270], attached to it by a
joint[271], so as to bend upwards again and fold over the face as high as
the nose, concealing not only the chin and the first-mentioned elongation,
but the mouth and part of the cheeks[272]: conceive, moreover, that to the
end of this last-mentioned plate are fixed two other convex ones, so broad
as to cover the whole nose and temples[273],—that these can open at
pleasure, transversely like a pair of jaws, so as to expose the nose and
mouth, and that their inner edges where they meet are cut into numerous
sharp teeth or spines, or armed with one or more long and sharp
claws[274]:—you will then have as accurate an idea as my powers of
description can give, of the strange conformation of the under-lip in the
larvæ of the tribes of Libellulina; which conceals the mouth and face
precisely as I have supposed a similar construction of your lip would do
yours. You will probably admit that your own visage would present an
appearance not very engaging while concealed by such a mask; but it
would strike still more awe into the spectators, were they to see you first
open the two upper jaw-like plates, which would project from each temple
like the blinders of a horse; and next, having by means of the joint at your
chin let down the whole apparatus and uncovered your face, employ them
in seizing any food that presented itself, and conveying it to your mouth.
Yet this procedure is that adopted by the larvæ provided with this strange
organ. While it is at rest, it applies close to and covers the face. When the
insects would make use of it, they unfold it like an arm, catch the prey at
which they aim by means of the mandibuliform plates, and then partly
refold it so as to hold the prey to the mouth in a convenient position for
the operation of the two pairs of jaws with which they are provided.
Reaumur once found one of them thus holding and devouring a large
tadpole;—a sufficient proof that Swammerdam was greatly deceived in
imagining earth to be the food of animals so tremendously armed and
fitted for carnivorous purposes. Such an under-lip as I have described is
found in the tribe of dragon-flies (Libellulina); varied, however,
considerably in its figure in the different genera. In the larva of Libellula
Fab., such as Libellula depressa, &c. it is of the shape above described; so
exactly resembling a mask, that if Entomologists ever went to
masquerades, they could not more effectually relieve the insipidity of such
amusements and attract the attention of the demoiselles, than by
appearing at the supper table with a mask of this construction, and
serving themselves by its assistance. It would be difficult, to be sure, by
mechanism to supply the place of the muscles with which in the insect it is
amply provided: but Merlin, or his successor, has surmounted greater
obstacles. In the larva of the Fabrician Æshnæ (Libellula grandis, &c. L.),
this apparatus is not convex but flat: so that, though it equally conceals
the face, it does not so accurately resemble a mask; and the jaws at its
apex are not convex plates, but rather two single conical teeth[275]. It is,
as to its general shape, similarly constructed in Agrion Fab. (L. Virgo, &c.
L.); but the first joint is more remarkably elongated, the jaws more
precisely resemble jaws than in any of the rest, and are armed with three
long, very sharp teeth: between them also there is a lozenge-shaped
opening, through which, when the apparatus is closed, is protruded a
circular sort of nipple, apparently analogous to the ligula[276]. Libellula
ænea, L., which is the type of another tribe (Cordulia Leach), has a mask
somewhat different from all the above, the jaws being armed with a
moveable claw and an internal tooth[277]. You will admire the wisdom of
this admirable contrivance, when you reflect that these larvæ are not fitted
to pursue their prey with rapidity, like most predaceous animals; but that
they steal upon them, as De Geer observes[278], as a cat does upon a
bird, very slowly, and as if they counted their steps; and then, by a sudden
evolution of this machine, take them as it were by surprise, when they
think themselves safe. De Geer says, it is very difficult for other insects to
elude their attacks, and that he has even seen them devour very small
fishes[279]. As these animals are found in almost every ditch, you will
doubtless lose no time in examining for yourself an instance of so singular
a construction.
Feelers (Palpi). In the orders Diptera and Hymenoptera are many larvæ in
which these organs have not been certainly discovered; yet Reaumur in
that of a common fly (M. meridiana L.) found four retractile nipples[280]
which seem analogous to them; and Latreille has observed, that below the
mandibles of those of ants are four minute points, two on each side[281]:
but in all other larvæ their existence is more clearly ascertained. The
maxillary palpi vary in number, many having two on each maxilla and
others only one. In the perfect insect the former is one of the
distinguishing characters of the predaceous beetles (Entomophagi Latr.),
but in the larvæ it is more widely extended; since even in the caterpillars
of Lepidoptera the inner lobe of the maxilla which represents this feeler is
jointed, which is precisely the case with the beetles just named. Cuvier has
observed this circumstance in the larva of the stag-beetle[282]; and it
belongs to many other Coleoptera that have only a pair of maxillary palpi
in the perfect state. The labial palpi are always two, emerging usually one
on each side from the apex of the under-lip. With regard to the form of
the palpi, those of the Lepidoptera are mostly conical; in other orders they
are sometimes setaceous and sometimes filiform. Their termination is
generally simple, but sometimes the last joint is divided. They are for the
most part very short, and the labial shorter than the maxillary. The latter
never exceed four joints[283], which seems the most natural number; and
the former are limited to three. Both vary between these numbers, and
one joint. The joints, though commonly simple, are sometimes branched.
This is the case with one I met with in considerable numbers upon the
Turnip, in October 1808, the second joint of the palpi of which sends forth
near the apex an internal branch. In the larva of the Cossus, as Lyonnet
informs us[284], the joints of the palpi are retractile, so that the whole of
the organ may be nearly withdrawn.

After thus describing the head of larvæ, and its principal organs, we must
next say something upon the remainder of the body, or what constitutes
the
2. Trunk and Abdomen: which I shall consider under one article. These are
composed of several segments or rings, to which the feet and other
appendages of the body are fixed. The form of these segments, or that of
their vertical section, varies considerably: in many Lepidoptera, the wire-
worm, &c., it would be nearly circular; in others a greater or less segment
of a circle would represent it; and in some, perhaps, it would consist of
two such segments applied together. Their lower surface is generally
nearly plane. Their most natural number, without the head and including
the anal segment, is twelve: this they seldom exceed, and perhaps never
fourteen. The three first segments are those which represent the trunk of
the perfect insect, and to which the six anterior legs when present are
affixed. In general, they differ from the remaining segments only in being
shorter, and in many cases less distinctly characterized; but in
Neuropterous larvæ, those of Dytisci, and some other Coleoptera, they are
longer than the succeeding ones, and pretty nearly resemble the trunk of
the animal in its last state. The surface of the trunk and abdomen will be
considered under a subsequent head; I shall not, therefore, describe it
here. The conformation of the different segments varies but little, except
of the terminal one, or tail, which in different larvæ takes various figures.
In most, this part is obtuse and rounded; in others acute or acuminate; in
others truncate; and in others emarginate, or with a wider sinus, and with
intermediate modifications of shape which it would be endless to
particularize. In some, also, it is simple and unarmed; in others beset with
horns, spines, radii, and tubercles of different forms, some of which will
come under future consideration. The parts connected with the trunk and
abdomen which will require separate consideration, are the legs, the
spiracles, and various appendages.
Legs. It may be stated generally that the larvæ of the orders Coleoptera,
Lepidoptera, and Neuroptera, have legs; and that those of the orders
Hymenoptera and Diptera have none. This must be understood, however,
with some exceptions. Thus the larvæ of some Coleoptera, as the weevil
tribes (Curculio L.) have no legs, unless we may call by that name certain
fleshy tubercles besmeared with gluten, which assist them in their
motions[285]; while those of Tenthredo and Sirex in the order
Hymenoptera are furnished with these organs. At present I know no
Dipterous larva that may be said to have real legs, unless we are to regard
as such certain tentacula formed upon a different model from the legs of
other larvæ[286]. Rösel has, I think, figured a Lepidopterous apode. No
Neuropterous one has yet been discovered.
The legs of larvæ are of two kinds; either horny and composed of joints,
or fleshy and without joints[287]. The first of these, as I observed in a
former letter[288], are the principal instruments of locomotion, and the last
are to be regarded chiefly as props and stays by which the animal keeps
its long body from trailing, or by which it takes hold of surfaces; while the
other legs, or where there are none, the annuli of its body, regulate its
motions. The former have been commonly called true legs (pedes veri),
because they are persistent, being found in the perfect insect as well as in
the larva; and the latter spurious legs (pedes spurii), because they are
caducous, being found in the larva only. Instead of these not very
appropriate names, I shall employ for the former the simple term legs, and
for the latter prolegs (propedes)[289].
The legs, when present, are always in number six, and attached by pairs
to the underside of the three first segments of the trunk. They are of a
horny substance, and consist usually of the same parts as those of the
perfect insect; namely, coxa, trochanter, femur, tibia, and tarsus,
suspended to each other by membranous ligaments: these parts are less
distinctly marked in some than in others. Thus in the legs of a caterpillar,
or the grub of a capricorn-beetle, at first you would think there were only
three or four joints besides the claw; but upon a nearer inspection, you
would discover at the base of the leg the rudiments of two others[290], in
the latter represented indeed by the fleshy protuberance from which the
legs emerge. In the larvæ of the predaceous Coleoptera, the hip and
trochanter are as conspicuous nearly as in the perfect insect; and the
tarsus, which still consists of only a single joint, is armed with two
claws[291]. In those of the Neuroptera order, in which all the joints are
very conspicuous, the tarsi are jointed, as well as two-clawed[292]. The
legs of larvæ are usually shorter than those of the perfect insect, and
scarcely differ from each other in shape, for they all gradually decrease in
diameter from the base to the apex. This is the most usual conformation
of them in Lepidopterous, Hymenopterous, and some Coleopterous larvæ,
(those of the capricorn-beetles are very short and minute, so as to be
scarcely visible,) in which they are so small as to be concealed by the body
of the insect[293]. In Neuropterous larvæ, however, and several
Coleoptera, as those of Dytiscus, Staphylinus, Coccinella, &c., they more
resemble the legs of the perfect insect, the joints being more elongated,
and the femoral one projecting beyond the body[294].
You will find no other than true legs in most Coleopterous, Neuropterous,
and Hymenopterous larvæ. But those of the saw-flies (Tenthredo L.), and
all caterpillars, have besides a number of prolegs: a few Dipterous larvæ
also, are provided with some organs nearly analogous to them. These
prolegs are fleshy, commonly conical or cylindrical, and sometimes
retractile protuberances, usually attached by pairs to the underside of that
part of the body that represents the abdomen of the future fly[295]. They
vary in conformation and in number; some having but one, others as many
as eighteen.
With regard to their conformation, they may be divided into two principal
sections: first, those furnished with terminal claws; and secondly, those
deprived of them. Each of which may be divided into smaller sections,
founded on the general figure of the prolegs, and arrangement of the
claws or hooks.
i. The prolegs of almost all Lepidopterous larvæ are furnished with a set of
minute slender horny hooks, crotchets, or claws, of different lengths,
somewhat resembling fish-hooks; which either partially or wholly surround
the apex like a pallisade. By means of these claws, of which there are from
forty to sixty in each proleg, a short and a long one arranged alternately,
the insect is enabled to cling to smooth surfaces, to grasp the smallest
twigs to which the legs could not possibly adhere: a circumstance which
the flexible nature of the prolegs greatly facilitates[296]. Claws nearly
similar are found on the prolegs of some Dipterous larvæ[297], but not in
any of those of the other orders. These last, however, are seldom either so
numerous, or arranged in the same manner, as in caterpillars. When the
sole of the foot is open, the claws with which it is more or less surrounded
are turned outwards, and are in a situation to lay hold of any surface; but
when the animal wishes to let go its hold, it begins to draw in the skin of
the sole, and in proportion as this is retracted, the claws turn their points
inwards, so as not to impede its motion[298].
The prolegs with claws may be further divided into four different kinds.
1. In the larvæ of the great majority of butterflies and moths they assume
the form of a truncated cone, the lower and smaller end of which is
expanded into a semicircular or subtriangular plate, having the inner half
of its circumference beset with the claws above mentioned; and, from its
great power of dilating and contracting, admirably adapted for performing
the offices of a foot. Jungius calls these legs pedes elephantini[299]; and
the term is not altogether inapplicable, since they exhibit considerable
resemblance to the clumsy but accommodating leg and foot of the gigantic
animal he alludes to.
2. The larvæ of many minute moths, particularly of the Fabrician genera
Tortrix and Tinea—those which live in convoluted leaves, the interior of
fruits, &c., as well as the Cossus, and some other large moths,—have their
prolegs of a form not very unlike those of the preceding class, but shorter,
and without any terminal expansion; the apex, moreover, is wholly, instead
of half, surrounded with claws[300]; the additional provision of which,
together with a centrical kind of nipple capable of being protruded or
retracted, in some measure, though imperfectly, supplies the place of the
more flexible plate-like expansion present in the first class.
3. The third class is composed of a very few Lepidopterous larvæ which
have their prolegs very thick and conical at the base, but afterwards
remarkably slender, long, and cylindrical, so as exactly to assume the
shape of a wooden leg[301]. These, as in the first class, are expanded at
the end into a flat plate: but this is wholly circular, is surrounded with
claws, and has also in the middle a retractile nipple, as in the preceding
class. In Cossus, at least in an American species (Cossus Robiniæ),
described by Professor Peck[302], the anal prolegs have the claws only on
their exterior half.
4. The remaining description of unguiferous prolegs, if they may not rather
be deemed a kind of tentacula, are those of certain Diptera, provided with
no true legs; which differ from the three preceding classes, either in their
shape, or the arrangement of their claws. In one kind of those remarkable
larvæ, which from their long respiratory anal tubes Reaumur denominates
"rat-tailed," that of Elophilus pendulus, there are fourteen of these prolegs,
affixed by pairs to the ventral segments, the twelve posterior ones of
which are subconical, and truncate at the apex, which is surrounded with
two circles of very minute claws, those of the inner being much more
numerous and shorter than those of the exterior circle; while the anterior
pair terminate in a flat expansion, and in shape almost exactly resemble
those of a mole[303]. The prolegs of the larvæ of a kind of gnat called by
De Geer Tipula amphibia, and of Syrphus mystaceus F., (Musca plumata
De Geer,) are nearly of a similar construction, but in the last are armed
with three claws only[304]. Long moveable claws also distinguish the
singular prolegs before described[305] of another gnat (Tanypus maculatus
Meig., Tipula De Geer). The case-worms (Trichoptera K.) and some others,
have two prolegs at the anus, each furnished with a single claw[306].
ii. The prolegs deprived of claws are found in the larva of the
Hymenopterous tribe of saw-flies (Tenthredo L.), in those of some
Lepidoptera (Hepialus F. &c.), and in some few Coleopterous and Dipterous
genera. Those of the former are of the shape of a truncated cone, and
resemble the second class of unguiculate prolegs, except in the defect of
claws. In the latter they are a mere retractile nipple-like protuberance, in
some species so small as scarcely to be perceptible. In all they aid in
progressive motion; but it is by laying hold of surfaces, and so enabling
the body more readily to push itself forward by annular contraction and
dilatation, and not by taking steps, of which all prolegs are incapable: to
assist in this purpose the protuberance sometimes secretes a gluten[307],
which supplies the place of claws. Some larvæ have the power of
voluntarily dilating certain portions of the underside of their body, so as to
assume nearly the shape and to perform the functions of prolegs. In a
Coleopterous (?) subcortical one from Brazil, before alluded to, there are
four round and nearly flat areas in each ventral segment of the abdomen,
but the last very little raised above the surface, and rough, somewhat like
a file; and besides these, the base of the anal segment has ten of these
little rough spaces, but of a different shape, being nearly linear, placed in a
double series, five on each side. Doubtless these may be regarded as a
kind of prolegs, which enable the animal to push itself along between the
bark and the wood[308].
In considering, in the next place, the number and situation of the prolegs,
it will contribute to distinctness to advert to these circumstances as they
occur in the different orders furnished with these organs.
To begin with the Lepidoptera.—Lepidopterous larvæ have either ten,
eight, six, or two prolegs, seldom more[309], and never fewer. Of these,
with a very few exceptions, two are attached to the last or anal, and the
rest, when present, to one or more of the sixth, seventh, eighth, and ninth
segments of the body: none are ever found on the fourth, fifth, tenth, or
eleventh segments.
1. Where ten prolegs are present, as is the case in by far the greatest
proportion of Lepidopterous larvæ, there is constantly an anal pair, and a
pair on each of the four intermediate segments just mentioned.
2. In caterpillars, which like those of a few species of the genera Sphinx,
Pyralis, and of the Bombycidæ, &c. have eight legs, they are placed in
three different ways. In those which have an anal pair, the remaining six
are in some fixed to the sixth, seventh, and eighth; in others, to the
seventh, eighth, and ninth segments. In those which, like Cerura Vinula,
and several other species of the same family, have no anal prolegs; the
whole eight emerge from the sixth, seventh, eighth, and ninth segments.
3. The Hemigeometers, as Noctua Gamma, &c. have only six legs: namely,
an anal pair, and two ventral ones, situated on the eighth and ninth
segments.
4. The larvæ of the Geometers (Geometræ F.) have but four prolegs; of
which two are anal, and two spring from the ninth segment. It should be
observed, however, that the larvæ of Hemigeometers, and even of some
of those that have ten prolegs, where the four anterior ones are much
shorter than the rest, move in the same way as the Geometers. This even
prevails in a few where these organs are all of equal length.
5. Many of the larvæ of Tinea L. which live in the interior of fruits, seeds,
&c., have but one pair of prolegs, which are attached to the anal segment.
6. The larvæ of Haworth's genus Apoda (Hepialus Testudo and Asellus F.),
remarkable for their slug-like shape and appearance, move by the aid of
two lateral longitudinal pustule-like protuberances, which leave a trace of a
gummy slime in their course.
Hymenoptera.—The larvæ of the different tribes of Tenthredo L., almost
the only Hymenopterous insects in which prolegs are present, have a
variable number of these organs; some sixteen, as the saw-fly of the
willow (T. lutea L.), and this is the most numerous tribe of them, including
the modern genera, Cimbex F., Pterophorus, &c. Others have fourteen, as
that of the cherry (T. cerasi L.); and many others with only nine joints to
their antennæ. A third class have only twelve, as that of the rose (T. Rosæ
L.), but this contains but few species. The last class contains those that
have no prolegs at all, but only the six horny ones appended to the trunk.
Of this tribe, the caterpillars of which have a very different aspect from the
preceding, are those of the genus Lyda F. (T. crythrocephala L.)[310]. Two
of the prolegs are anal, and the rest intermediate, and none are furnished
with claws. This circumstance, in conjunction with the greater number of
prolegs, except in the case of Lyda, will always serve as a mark to
distinguish these fausses chenilles, as the French call the larvæ of saw-
flies, from true caterpillars. The dorsal prolegs of a species of Cynips
described by Reaumur have been before noticed.
Coleoptera.—The larvæ of insects of this order are so little known or
attended to, that no very accurate generalization of them in this respect is
practicable. Many of them, in addition to their six horny legs, have a proleg
at the anus; which in many cases appears to be the last segment of the
abdomen, forming an obtuse angle with the remainder of it, so as to
support that part of the body, and prevent it from trailing; and in some
instances, as in Chrysomela Populi, a common beetle, secreting a slimy
matter to fix itself[311]. In the larvæ of Staphylinidæ this proleg is very
long and cylindrical; in that of Cicindela it is shorter, and in shape a
truncated cone rather compressed; it is very short, also, in those of the
Silphæ that I have seen. In the wire-worm (Elater Segetum) it is a minute
retractile tubercle, placed in a nearly semicircular space, shut in by the last
dorsal segment, which becomes also ventral at the anus. This space is in
fact the last ventral segment. This seems characteristic of the genus[312].
From the underside of the body of the common meal-worm (Tenebrio
Molitor), at the junction of the two last segments, when the animal walks,
there issues a fleshy part, furnished below with two rather hard, long, and
moveable pediform pieces, which the animal uses in walking[313]. In the
larva of another beetle, whose ravages have been before noticed, under
the name of the cadelle[314] (Trogosita mauritanica), a pair of prolegs are
said to be found under the anal segment; and in that of the bloody-nose
beetle (Timarcha tenebricosa), that segment is bifid. That of the weevil of
the common water-hemlock (Lixus paraplecticus F.) exhibits a singular
anomaly: prolegs occupy the usual station of the true legs, being attached
to the three segments representing the trunk[315]. This insect, however,
does not appear to use them in moving. A pair in each of the twelve
segments of the body are found in the grub of another weevil (Hypera
Rumicis Germ.), the nine last pair being the shortest, which all assist the
insect in walking[316]. But the greatest number of prolegs is to be found in
the Brazil subcortical larva lately mentioned. Besides the six horny legs of
the trunk, this remarkable animal has four prolegs on each of the seven
intermediate abdominal segments, and five on each side of the base of the
last, making the whole number of prolegs, if so they may be called,
amount to forty-four: a far greater number than is to be found in any larva
at present known. When I wrote to you upon the motions of insects, I
informed you that some larvæ moved by means of legs upon their
back[317], but I was not then aware that any were furnished with them
both on the back and the belly at the same time. By the kindness of Mr.
Joseph Sparshall of Norwich, a very ardent and indefatigable entomologist,
I am in possession of the larva of Rhagium fasciatum, a timber-feeding
beetle. This animal on the ten intermediate segments of the underside of
the body, which in the centre form a fleshy protuberance, has on it a
double series of rasps, as it were, consisting each of two rows of oblique
oblong prominences; and on the seven intermediate dorsal segments there
are also in the centre seven rasps of three or four rows each, of similar
prominences: so that this animal at the same time can push itself along
both by dorsal and ventral prolegs. It is worthy of observation, that a pair
of these rasps is between the second and third pair of true legs.
Diptera.—The larva of a little gnat, Tipula stercoraria De Geer[318]
(Chironomus Meig.?), drags itself along by the assistance of a single
tubercle, placed on the underside of the first segment of the body, which
the animal has the power of lengthening or contracting[319]. That of
another beautiful Chironomus (C. plumosus), remarkable for the feathered
antennæ of the male[320], has two short prolegs, or pediform but not
retractile tentacula in the same situation[321]. Others, as that of Tanypus
maculatus, &c. have two pairs, one attached to the anal and the other to
the first segment[322]. Tipula amphibia De Geer in this state has ten
prolegs, placed by pairs on the fourth, fifth, eighth, ninth, and tenth dorsal
segments[323]; and Scæva Pyrastri F., one of the aphidivorous flies, has
not fewer than forty-two, arranged in a sextuple series, seven in each
row[324].
It may not be useless to close this long description of the legs of larvæ
with a tabular view of them, founded chiefly upon these organs; which
afford very obvious marks of distinction.

I. Larvæ without legs.

i. With a corneous head of determinate shape (coleopterous and
hymenopterous apods—Culicidæ, some Tipulidæ, &c. amongst
the Diptera).
ii. With a membranaceous head of indeterminate shape (Muscidæ,
Syrphidæ, and other Diptera).
II. Larvæ with legs.
i. With legs only, and with or without an anal proleg (Neuroptera,
and many Coleoptera).
1. Joints short and conical (Elater, Cerambycidæ, &c.).
2. Joints long and subfiliform (Staphylinus, Coccinella, Cicindela,
&c.).
ii. Prolegs only (many Tipulidæ, and some subcutaneous
Lepidopterous larvæ, &c.).
iii. Both legs and prolegs (Lepidoptera, Tenthredinidæ, and some
Coleoptera).
1. Without claws (Tenthredinidæ, &c.).
 2. With claws (Lepidoptera, &c.).

I should next say something upon the spiracles, or breathing-pores, or any

other external apparatus for the purpose of respiration, in larvæ; but I
think it will be best to reserve the consideration of these for a subsequent
Letter. We will therefore conclude this detailed description of their parts in
their first state, with some account of their other.
iii. Appendages. The generality of larvæ have no other external organs
than those already described; but in several of them we observe various
kinds of retractile ones and others—protuberances—horn-like processes—
rays, &c.; which, though not properly coming either under any of the
above parts, or under the clothing of these animals, yet require to be
noticed. Upon these I shall now enlarge a little.
You must have observed upon the back of the last segment but one of the
caterpillar of the silk-worm a horn-like process, rising at first nearly
perpendicularly, and then bending forward. A similar horn, though
confined in the genus Bombyx to the silk-worm and a few others, if we
may believe Madame Merian, who, however, often makes great mistakes,
is found in the beautiful caterpillar of one of the largest and finest moths
that we know (Erebus Strix[325]), the glory of the Noctuidæ, and in most
of those of the hawk-moths (Sphinx F.) [S. Porcellus, Vitis, and a few
others excepted; in some of which, as S. Labruscæ, &c., this anal horn is
replaced by a gibbosity, and in others, as S. Œnotheræ, by a callous eye-
like plate[326]] in the same situation, but much longer[327], and commonly
curving backwards over the tail[328]. Sometimes, however, as in S. ocellata
and S. Stellatarum, it is perfectly straight. These organs towards the apex
are horny, and often end in a sharp point; nearer the base they are fleshy.
They are without any true joint[329], yet the insect can elevate or depress
them at pleasure. Under a lens, they usually appear covered with spinous
eminences, arranged like scales. The use of these horns is quite unknown:
Goedart fancies that they secrete a potent poison, and are intended as
instruments of defence; but both suppositions are altogether unfounded. It
has been remarked, that the body of those caterpillars which have these
horns, is firmer, and yields less to the touch than that of those which have
no such appendages[330]. The larva of a small timber-devouring beetle
(Lymexylon dermestoides F.) has, like the above caterpillars, a long horn,
and in the same situation: it has also a singular protuberance on the first
segment[331]. Upon some other caterpillars, as in Bombyx Stigma F., a
singular pair of horn-like appendages arises from the back of the second
segment of the body, excluding the head. In a tawny-coloured one from
Georgia, with a transverse row of short black spines on each segment,
these horns are half an inch long, black, covered with spinous eminences,
rather thickest at the base, and terminate in a little knob. They appear to
articulate with the body at the lower extremity. I have another species,
black, with narrow longitudinal yellow stripes, in which these horns are of
equal thickness at base and apex, but with the same terminal knob.
Danais Archippus has a pair of tentacula at the head, and another pair, but
shorter, at the tail; and D. Gylippus has, besides these, two in the middle
of the body[332].
We are equally ignorant of the use of the upright horn found upon the
back of the fourth segment in the larva of some moths (Noctua Psi, and
tridens F.) which is of a construction quite different from that of those last
described. It is cylindrical, slightly thinner at the apex, which is obtuse,
fleshy, incapable of motion, of a black colour, and about two lines long. On
the same segment, also, in the case-worms (Trichoptera K.) are three
fleshy conical eminences, which the animal can inflate or depress, so that
they sometimes totally disappear, and then in an instant swell out again.
When retracted, they form a tunnel-shaped cavity, varying in depth[333].
Reaumur conjectured that these eminences were connected with
respiration, and one circumstance seems in favour of this conjecture, that
this segment has not the respiratory threads observable in the subsequent
ones. Latreille mentions certain fleshy naked eminences placed upon the
ninth and tenth segments of some hairy caterpillars, which, like those just
mentioned, the animal can elevate more or less. They are often little
cones; but when it would shorten them, the summit is drawn in, and a
tunnel appears where before there was a pyramid[334].
In a former Letter I gave you a short account of the remarkable Y-shaped,
as it should seem, scent-organs (Osmateria) of the beautiful caterpillar of
the swallow-tailed butterfly (Papilio Machaon L.), and others of the
Equites[335]; I will now speak of them more fully. That found in the former
is situated at the anterior margin at the back of the first segment, close to
the head, from which at first view it seems to proceed. At the bottom it is
simple, but divides towards the middle, like the letter Y, into two forks, of
a fleshy substance[336], which it can lengthen, as a snail does its horns, to
five times their ordinary extent, or retract them within the stalk, so as
wholly to conceal them. Sometimes it protrudes one fork, keeping the
other retracted; and often withdraws the whole apparatus for hours
together under the skin, and its place is only marked by two tawny-
coloured dots, so that an ordinary spectator would not suspect the
existence of such an instrument[337]. Unfortunately this larva is rare in this
country, so that I can scarcely flatter you with the hope of seeing this
curious organ in a living specimen[338], unless you choose to import a
parcel of its eggs from the south of Europe, where it is common. This you
will think rather a wild proposition; but why should not Entomologists
import the eggs of rare insects, as well as botanists the seeds of rare
plants? But if you will be satisfied with the dissection of a dead specimen,
I have several, done by the ingenious Mr. Abbott of Georgia, in which this
part is well exhibited[339].
Another small caterpillar, as it should seem, of a geometer, prepared by
the same gentleman, exhibits a pair of similar horns on the fifth and sixth
segments: in these the common base from which the fork proceeds is very
short and wide, and each branch grows gradually more slender from the
base to the apex, where it is involute. Whether these are retractile, or
whether they correspond with those of P. Machaon in their nature and use,
cannot be ascertained from a dead specimen: as they belong to a larva of
a quite different tribe of Lepidoptera, the probability is, that they
essentially differ. Two globose retractile vesicles issue from the ninth and
tenth segments of those of Arctia chrysorrhea, &c.[340]
A great number of Lepidopterous larvæ, particularly those which are
smooth and of a moderate size, have between the under-lip and fore-legs
a slender transverse opening, containing a teat-like protuberance of the
same construction as the furcate horn of the caterpillar of the beautiful
mountain-butterfly, Parnassius Apollo; and, like that, can either be wholly
retracted and concealed, or by pressure be extended to the length of one
of the legs. In some larvæ this part is of a subhemispherical figure,
generally single, but sometimes double. It is commonly, however, more
slender and conical; and when of this shape, it is sometimes
quadruple[341]. The use of this part is not very clearly known: some have
supposed it to be a second spinneret, and to be of use in fabricating the
cocoon; but it is more probable that it secretes some other kind of fluid,
and is connected with defence.
The singular organ in a similar situation, evidently for that purpose, with
which the puss-moth endeavours to annoy its assailants, has been
described in a former Letter, to which I refer you[342]. Bonnet, who was
the first that discovered this organ, ascertained that it might be cut off
without injury to either larva or imago. He also remarked in a caterpillar
found in the wild succory (Cichorium Intybus) another short, black, needle-
shaped organ between the conical part just described and the under-
lip[343]. De Geer mentions a remarkable fleshy horn-like style, which issues
from the lower side of the first segment, between the head and the legs of
the case-worms (Trichoptera): he does not describe it as retractile, or it
might be regarded as analogous to those of Lepidoptera similarly situated,
that I have just noticed[344]. In that of the emperor-moth (Saturnia
Pavonia), there are perforated tubercles, which when the animal is
molested spirt forth a transparent fluid[345].
The horn-like appendage of the puss-moth (Cerura Vinula) is situated at
the tail of the insect, and is composed of two distinct cylindrical diverging
branches, each about four lines long, not united at the base. Each of these
is hollow, and includes a smaller cylindrical piece, which can be protruded
at pleasure, and withdrawn again, as a pencil within its case; or, rather, as
the horns of a snail. The two outer horns are tolerably firm, moveable at
their base, and beset with black spines; the interior tentacula are fleshy,
moveable in every direction, and in full-grown larvæ of a rose colour. The
animal seldom protrudes them, unless in some way disturbed; and
frequently it approximates the two outer cases so closely that they
resemble a single horn. It appears to use these inner horns, when
protruded, as a kind of whip to drive away the flies, especially the
Ichneumons, that alight upon its body. When touched in any place, it will
unsheath one of them, and sometimes both, and with them strike the
place where it is incommoded[346]. A similar organ is found in some other
Bombycidæ, as B. Tau and Furcula F. Reaumur mentions a caterpillar that
to this kind of tail added the resemblance of two ears, or two cylindrical
bodies, terminating in a point, which emerged from the first segment
behind the head[347]. In another observed by the same author, the legs
were replaced by a single horn, but which did not appear to send forth an
internal one: from the back of its fourth segment also emerged a single
conical or pyramidal fleshy eminence or cleft, terminating in two
points[348]. Some of the tropical butterflies also, as may be seen in the
figures of Madame Merian, have two diverging anal horns instead of anal
prolegs; but it does not appear that they incase tentacula[349]. Wherever
these caudal horns are found, the above prolegs are wanting[350]. Two
conical anal horns also distinguish the caterpillar of one of the moths
called Prominents, Notodonta camelina; but these are not terminal, but on
the back of the last segment but one[351]. In that of another British moth,
N. ziczac F., there are three dorsal prominences, one near the anus, and
two more in the middle[352]. Some Geometers (G. fuliginosa, &c.) have
two erect horns on the eleventh segment, and others (G. syringaria, &c.)
two recurved ones on the eighth[353]. I must not here omit to mention the
curious hooks emerging from two tubercles on the back of the eighth
segment of the ferocious larva of that beautiful tiger-beetle, the Cicindela
campestris L., not uncommon on warm sunny banks. This animal with
incessant labour, as we are informed by M. Desmarets, digs a cylindrical
burrow, to the enormous depth, the size of the animal considered, of
eighteen inches. To effect this, it carries out small masses of earth upon its
large concave head; and having often occasion to rest in ascending this
height, by means of these hooks[354] it fixes itself to the sides of its
burrow, and, having finally arrived at its mouth, casts off its burthen.
When these insects lie in wait for their prey, their head, probably in
conjunction with the first segment of the body, accurately stops the mouth
of the burrow, so as to form an exact level with the surrounding soil; and
thus careless insects, walking over it without perceiving the snare, are
seized in a moment and devoured[355].
Another kind of appendage, which is found in some larvæ, is the organ
employed by them to carry the excrement; with which, instead of letting it
fall to the ground, they form a kind of umbrella to shelter and probably
conceal them. All the tortoise-beetles (Cassida L.) have instruments for
this purpose, as well as an Indian genus (Imatidium Latr.) very nearly
related to them. This instrument is a kind of fork, half as long as the body,
consisting of two branches, growing gradually smaller from the base to the
summit, where they terminate in a very fine point, of a substance rather
horny, and attached to the body near the anal orifice. They are armed on
the outside with short spines, from the base for about a third of their
length. When this fork, as it usually is, is laid parallel to the back, with its
points towards the head, the anal aperture points the same way. When the
animal walks, the fork points the other way, and is in the same line with
the body, and the anus assumes a prone position[356].
The larvæ of a genus of flies (Volucella Geoffr.) remarkable for inhabiting
the nests of humble bees, are distinguished on their upper side by six
long, diverging, pointed, membranous radii; placed in a semicircle round
the anus[357]: what the particular use of these organs may be, has not
been conjectured. Another in my collection has only four upper radii, but
below the anus are two fleshy filiform tentacula. One of a Tipulidan
described by Reaumur, has also four upper teeth; but instead of two
subanal tentacula, has six[358]. The singular larva of another of this tribe
(Chironomus plumosus) has on the two last segments four long, fleshy,
filiform, flexible tentacula, often interlaced with each other; which,
according to the same illustrious author, are used by the animal to fix its
caudal extremity, like the geometers, that the other end may be at liberty.
Besides these organs round the anus, it has also four other oval ones, of
uncertain use: not to mention the two prolegs, which M. Latreille thinks
are air-tubes[359]. Jointed anal organs are observable in other larvæ: those
in that of a saw-fly described by De Geer (Lyda F.) consist of three
joints[360]; in that of Hister cadaverinus, a carnivorous beetle, of two[361].
The larva as well as the pupa and imago of Ephemera is furnished with
three long diverging multiarticulate tails, which are probably useful as a
kind of rudder to assist and direct their motions. That of the smaller
dragon flies (Agrion F.) is furnished with three long vertical laminæ, by
moving which, as fish do their tails, from side to side, the animal makes its
way in the water[362]. That singular one, also, with a hooked head, figured
by Reaumur, has a single swimming lamina, or fin, shaped like a fan, and
placed in a vertical position under the tail[363].
The whole circumference of the body in some coleopterous larvæ,—for
instance, in that of the tortoise-beetle lately mentioned,—is surrounded
with appendages like rays. These are sometimes simple, rough with very
short spinous points[364]; but I have a dipterous larva, in which these radii
themselves are beautifully pinnated by a fringe of longish spines on each
side. Reaumur has described the grub of a beetle, the genus of which is
uncertain, and which feeds upon the larva of Aleyrodes Proletella, whose
body is margined on each side by eight triangular fleshy mammular
processes, terminating each in a bristle, which give it a remarkable
aspect[365]. The curious scent-organs with which the larva of Chrysomela
Populi is fringed have been before fully described; and therefore I shall
only mention them here[366].
In the larvæ of the lace-winged flies (Hemerobius), and ant-lions
(Myrmeleon), the anus is furnished with a small fleshy retractile cylinder,
from which proceeds the silken thread that forms the cocoon inclosing the
pupa[367]. Providence has many different ways of performing the same
operation. From the structure of the oral organs of these animals, the silk
could not conveniently be furnished by the mouth; the Allwise Creator has
therefore instructed and fitted them to render it by a spinneret at the
other extremity of the body.
The respiratory anal appendages of many Dipterous larvæ will be fully
described in a subsequent Letter: I shall therefore now only further
observe upon this subject, that although there is seldom any alteration in
the form of these appendages &c. in the same species, the caterpillars of
two moths (Cerura Vinula and Attacus Tau), however, are exceptions. The
former, when young, has two hairy projecting ear-like protuberances,
which it entirely loses, as I have myself observed, before it assumes the
pupa; and the latter, in like manner, after its third change of skin, is
deprived of its bent thorn-like points which attend it when young[368]. It is
remarkable that these last larvæ, when just excluded from the egg, are
also entirely destitute of these appendages; they soon, however, appear,
from slight elevations which mark their situation, and rapidly acquire their
usual form[369]. Changes of a similar kind, hitherto unobserved, may
probably take place in other species.
iii. Figure. I am next to consider the general figure or shape of larvæ. All
of them, with but few exceptions[370], agree in having a body more or less
constricted at intervals into a series of rings or segments; usually in
number, twelve; often nearly equal in length, but sometimes in this respect
very dissimilar[371]. The general outline or shape of the body is extremely
various: most frequently it approaches to cylindrical, as in most of the
caterpillars of Lepidoptera, and of the Hymenopterous tribe of saw-flies
(Tenthredo L.). The next most common figure is that more or less oblong
or oval one, sometimes approaching to conical, found in many of the larvæ
usually called grubs; such as those of the weevil (Curculio L.) and of the
capricorn (Cerambyx L.), and other coleopterous tribes; of bees, and all
Hymenopterous insects but the saw-flies; and also of a large number of
flies (Diptera). In some the figure approaches to fusiform, as in most of
the moths of the Fabrician genus Lithosia. In others, as in those of the
water-beetles (Dytiscus, &c.), it approaches to an obovate shape, being
widest towards the head, and terminating in a point at the anus. In others,
again, it is linear; an example of which is that of Staphylinus. Some are
convex, and others gibbous, above, and flat underneath; as those of
Silpha, Chrysomela and many other beetles. Others are flat, both above
and below, and depressed like a leaf; a remarkable instance of which has
been before noticed[372]. Some are very long, as those of most
Lepidoptera; others very short, as that of the ant-lion (Myrmeleon). Many
other peculiarities of form in individuals might be instanced; but a dry
enumeration of these would be of no great use to you. They can only be
advantageously learned by the study of good figures, and by watching the
actual metamorphosis of the singularly-formed larvæ that you meet with.
Instead, therefore, of any further specification of individual forms, I shall
now endeavour to give you, as far as my own knowledge of them and the
information I can collect from other sources will enable me, a larger and
more general view of the kinds of larvæ; for analytical inquiries lose half
their value and importance unless we proceed to apply them synthetically,
by forming, if possible, into groups the objects with which we are
individually acquainted.
Partial attempts at a synthetical arrangement with regard to the larvæ of
Lepidoptera and the saw-flies (Tenthredo L.) have been made both by
Reaumur and De Geer. M. Latreille also has recently given a Tableau
méthodique et général of articulated animals furnished with jointed legs,
considered in their first state[373]. The former of these is chiefly founded
upon the number of the prolegs, and the latter upon the metamorphosis,
prolegs, habits, head, and parts of the mouth, without any other notice of
the configuration. Mr. Wm. MacLeay, who, though young in years, is old in
science and critical acumen, has started a perfectly new hypothesis upon
this subject. In the progress of his inquiries into the natural arrangement
of animals, particularly of insects in the Linnean sense, he has been the
first to observe, that the relation which organized objects bear to each
other is of two kinds; one of real affinity, and the other only of analogy, or
resemblance. This important distinction, upon which I shall enlarge in a
future Letter, when I come to treat of Systems of Entomology, he has
applied, in a way quite original, to larvæ in general, but more particularly
to those of the Coleoptera order. For the basis of his system he assumes a
relation of analogy between the larvæ of Insects that in the progress of
their metamorphosis assume wings, and those that do not, which form his
class Ametabola, so that the prototypes of the former shall be found
amongst the latter[374]. But though Mr. MacLeay appears to consider the
analogy between these two as primary, he extends it in a secondary sense
to the Crustacea, at least in several instances[375]. Upon this occasion he
very judiciously remarks, that "in terming larvæ Chilognathiformes or
Chilopodiformes, it is not meant that they are Scolopendræ or Iuli, or even
near to them in affinity; but only that they are so constructed that certain
analogical circumstances attending them strongly remind us of these
Ametabola[376]." This remark you will bear in mind while I am treating of
this subject. It should seem from another part of the same paragraph, that
the comparison which our learned Physiologist recommends, is between
the young of the Ametabola and the larvæ of the corresponding groups of
Coleoptera. This must be understood to refer chiefly to the young of the
Chilopoda and Chilognatha, since they approach nearer to them in that
state, having then only six legs; but the rest of the Ametabola should
certainly be brought to this comparison in their adult state: and even the
two former orders in that state more strongly resemble numerous
coleopterous larvæ, than when they are young and much shorter. I before
called your attention to the remarkable circumstance that contrasts very
many larvæ of Hexapod insects that become winged in their perfect state
with adult Myriapoda: namely, that in one the progress to this state is by
losing their prolegs and shortening their body; while in the other, the
reverse of this takes place, numerous prolegs and additional segments
being gained before they arrive at maturity[377]. As the multiplication of
organs is a sign of imperfection, it may be affirmed of the former of these
tribes, that their progress is towards greater perfection; while that of the
other may be called a degradation. As larvæ may be regarded as a
stepping-stone by which approach is made from the apterous to the
winged tribes of Insects, it seems most consistent with general analogy
that each should connect with the other in that state in which the
resemblance is greatest. Now the Myriapoda resemble larvæ, as we have
just seen, most when in their adult state; therefore the comparison should
be between larvæ and adult Myriapoda.
Mr. MacLeay divides coleopterous larvæ into five tribes thus characterized:
—
1. A carnivorous hexapod larva, with an elongate linear flattened body,
having a large head armed with two sharp falciform mandibles, and
furnished with six granular eyes on each side. This kind he denominates
Chilopodiform, as having for its type in the Ametabola, Scolopendra L. The
examples he gives are Carabus and Dytiscus.
2. A herbivorous hexapod larva, with a long and almost cylindrical body, so
fashioned that the posterior extremity being curved under the breast, the
animal when at rest necessarily lies like an Iulus on its side. This tribe he
denominates Chilognathiform, from Iulus L. His examples are, the larvæ of
Petalocerous insects, as Scarabæus L., Lucanus L. &c.
3. An apod larva, having scarcely the rudiments of antennæ, but which is
furnished instead of feet with fat fleshy tubercles; which, when continued
along the back and belly, give the animal a facility of moving in whatever
way it may be placed. These he denominates Vermiform, from certain of
the Vermes intestina and Mollusca of Linné which he has associated with
his Annulosa[378]. His examples are, Curculio L. and Cerambyx L.
4. A hexapod and distinctly antenniferous larva, with a subovate rather
conical body, of which the second segment is longer and of a different
form from the others, so as to give the appearance of a thorax. His
denomination for these is Anopluriform, from Pediculus L., forming Dr.
Leach's Anoplura. His examples are, Coccinella and Chrysomela L.
5. A hexapod antenniferous larva of an oblong form, having like the former
vestiges of a thorax, besides two or more articulated or inarticulated
setaceous or corneous appendages to the last segment of the abdomen.
This tribe he calls Thysanuriform, from Lepisma and Podura L., forming M.
Latreille's order Thysanura. His example is Meloe with a note of
interrogation[379].
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookbell.com