IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO.

4, APRIL 2006

385

Computer Analysis of Computed Tomography Scans

of the Lung: A Survey
Ingrid Sluimer, Arnold Schilham, Mathias Prokop, and Bram van Ginneken*, Member, IEEE

AbstractCurrent computed tomography (CT) technology allows for near isotropic, submillimeter resolution acquisition of the
complete chest in a single breath hold. These thin-slice chest scans
have become indispensable in thoracic radiology, but have also substantially increased the data load for radiologists. Automating the
analysis of such data is, therefore, a necessity and this has created a rapidly developing research area in medical imaging. This
paper presents a review of the literature on computer analysis of
the lungs in CT scans and addresses segmentation of various pulmonary structures, registration of chest scans, and applications
aimed at detection, classification and quantification of chest abnormalities. In addition, research trends and challenges are identified
and directions for future research are discussed.
Index TermsAirway disease, chest, computer-aided diagnosis,
CT, emphysema quantification, interstitial lung disease, literature
review, literature survey, lung cancer, nodule characterization,
nodule detection, nodule size measurements, pulmonary embolism, registration, segmentation.

I. INTRODUCTION
ANN
BI
BO
CAD
CBIR
CLE
COPD
CT
CTA
DPLD
EC
FN
FDA
FP
GG
HIST(x)
HRCT
HU
ILD

Artificial neural network.

Area under the ROC curve.
Bullae index.
Bronchiolitis obliterans.
Computer-aided detection/diagnosis.
Content-based image retrieval.
Centrilobular emphysema.
Chronic obstructive pulmonary disease.
Computed tomography.
CT angiography.
Diffuse parenchymal lung disease.
Explosion controlled region growing.
False negative.
Food and Drug Administration.
False positive.
Ground glass.
Lowest x-th percentile of the histogram.
High-resolution computed tomography.
Hounsfield unit.
Interstitial lung disease.

Manuscript received May 27, 2005; revised November 17, 2005. Asterisk indicates corresponding author.
I. Sluimer and A. Schilham are with the Image Sciences Institute, University Medical Center Utrecht, 3584 CX Utrecht, The Netherlands (e-mail: [email protected]; arnold@[email protected]; http://www.isi.uu.nl/).
M. Prokop is with the Department of Radiology, University Medical Center
Utrecht, 3584 CX Utrecht, The Netherlands (e-mail: [email protected]).
*B. van Ginneken is with the Image Sciences Institute, University Medical
Center Utrecht, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands (e-mail:
[email protected]; URL: http://www.isi.uu.nl/).
Digital Object Identifier 10.1109/TMI.2005.862753

IPF
PF
LAA
LDC
LVRS
MC
MLD
pCa
PCP
PE
PET
PF
PI
PLE
PSE
PVE
SPECT
ROC
ROI
TN
TP
UIP
VOI

Idiopathic pulmonary fibrosis.

Pulmonary function tests.
Low attenuation area(s).
Linear discriminant classifier.
Lung volume reduction surgery.
Metastatic calcification.
Mean lung density.
Probability of cancer.
Pneumocystis carinii pneumonia.
Pulmonary embolism.
Positron emission tomography.
Pulmonary function.
Pixel index.
Panlobular emphysema.
Paraseptal emphysema.
Partial volume effect.
Single photon emission computed tomography.
Receiver operating characteristic.
Region of interest.
True negative.
True positive.
Usual interstitial pneumonia.
Volume of interest.
II. INTRODUCTION

OMPUTED tomography (CT) for the body has been available since 1975. Originally, CT was not considered a technique particularly well suited for the thorax. Low resolution resulted in large partial volume effects (PVEs) and the large difference in attenuation values between tissue and air (currently
a main reason for the effectiveness of CT in thoracic imaging)
made it difficult to correctly interprete small lesions. In 1977,
Kollins [1] concluded that CT had revolutionized neuroradiology and its impact in abdominal and pelvic imaging had been
similarly great but the ultimate role of computed tomography
in the study of diseases of the chest is not as certain.
Technical strides forward have since completely transformed
CT and thoracic imaging with it. Early on, the essential role of
image processing was recognized, for example in the work on
the Mayo Clinics dynamic spatial reconstructor [2]. Two advances in particular have had the most repercussions for CT of
the chest. About twenty years ago, improved axial resolution
made HRCT scans possible. Slices of 1 mm thick could provide anatomical detail of the lungs similar to that available from
gross pathological specimens [3]. However, scanner speed limitations at the time meant that a 1-cm gap between slices was
necessary to cover the entire thorax while avoiding breathing artifacts. This limitation has been effectively removed in the last

0278-0062/$20.00 2006 IEEE

386

IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006

decade with the advent of multi-detector-row scanners that can

acquire up to 64 1-mm slices simultaneously per rotation and
perform each rotation in less than a second. Present day scanners allow for isotropic acquisition of the complete chest with
submillimeter resolution well within a single breath hold. Compared to other modalities, CT excels in the imaging of the lungs.
A major challenge accompanying these spectacular improvements is dealing with the enormous increase in images that are
generated and have to be reported on. This challenge has been
coined the data explosion by Rubin [4]. It is becoming clear that
computer vision techniques are essential to facilitate CT interpretation. As a consequence, the rapid developments in chest CT
acquisition techniques have been followed by a sharp increase
in research on computer analysis of thoracic CT scans.
This paper aims to provide an overview of the literature on
computer analysis of CT images of the human lungs. For our
purposes here, we define this area as including segmentation of
various pulmonary structures (Section II), registration of chest
scans (Section III), and applications aimed at detection, classification and quantification of chest abnormalities (Section IV).
The work on segmentation can be further subdivided into
segmentations of the complete lung fields (Section II-A), the
bronchial tree (Section II-B), the vascular tree (Section II-C)
and the lung lobes (Section II-D). Application areas involve emphysema (Section IV-A), lung cancer (Section IV-B1, IV-B2,
IV-B3), Pulmonary embolisms (PEs) (Section IVC), (bronchial)
signs of airway diseases (Section IV-D), and differential diagnosis of lung disease (Section IV-E).
The survey covers the period 19992004, but relevant work
before 1999 has been included as well. Details of the literature
collection procedure can be found in the Appendix. Studies on
reconstruction algorithms, scans parameters and image quality
improvement, and visualization techniques have been excluded,
except for some studies that pertain particularly to chest CT.
There is a large body of work on virtual bronchoscopy. Work on
image analysis in this field, mainly about airway segmentation,
has been included. Studies that focus on the application of such
techniques and on effective visualization are considered outside
the scope of this survey.
In terms of methodology, most reviewed works employ
standard image processing techniques, such as filtering, region
growing and connected component analysis, mathematical
morphology, etc. which can be found in standard textbooks,
e.g., [5] and [6]. A large number of studies investigates the
detection and classification of abnormalities. This field is often
referred to as computer-aided detection or diagnosis (CAD).
The typical set-up of a CAD system is preprocessing, segmentation, candidate extraction, feature extraction, and classification.
The first four steps are usually considered to be part of image
processing. The final classification step deals with patterns that
are represented as points in a feature space. Finding decision
boundaries in these vector spaces is a central problem in pattern
recognition theory. Popular techniques such as ANNs or LDCs
can be found in many textbooks, e.g., [7] and [8].
III. SEGMENTATION
Segmentation is often a necessary first step to computer
analysis. In CT of the lungs, the various anatomical entities that

can require segmentation are the lungs themselves, the airways,

the vessels and the lung lobes. Each of these is discussed in
more detail in the following subsections.
A. Lung Segmentation
Any computer system that analyzes the lungs and does not
work on manually delineated regions of interest must incorporate an automatic lung segmentation. Armato and Sensakovic
[9] illustrated the importance of accurate segmentation as a preprocessing step in a CAD scheme. In a nodule detection setting, they showed that 5% 17% of the lung nodules in their
test data was missed due to the preprocessing segmentation,
depending on whether or not the segmentation algorithm was
adapted specifically to the nodule detection task.
As the lung is essentially a bag of air in the body, it shows up
as a dark region in CT scans. This contrast between lung and surrounding tissues forms the basis for the majority of the segmentation schemes. Most of the methods are rule-based [9][14].
The main lung volume is found in one of two ways. Gray-level
thresholding and component analysis can be used, after which
the objects that are the lungs are identified by imposing restrictions on size and location. Alternatively, the volume is found by
region growing from the trachea. The trachea itself is found by
searching for two-dimensional (2-D) circular air-filled regions
in the first slices of the scan, or by searching for a three-dimensional (3-D) tubular air-filled object located centrally in the top
half of the scan. After identification of the combined lung and
airway volume, separation of left and right lung and removal
of the trachea and mainstem bronchi are performed. This is
followed by morphological processing to obtain lung volumes
without holes and with smooth borders. Li and Reinhardt [15]
have taken a statistical rather than rule-based approach. A 3-D
active shape model provided an approximate segmentation of a
single lung and was combined with multiple 2-D refinements by
snakes to capture fine details and shape variance not present in
the statistics of the training data.
Often it remains unclear at what points the segmentations are
supposed to cut through the major bronchi and vessels in the
hilar area where they enter the lung. Manual delineations also
show much variance around the hilum, depending on definitions (or a lack thereof) and personal preferences with respect to
smoothness. Ukil and Reinhardt[16] used a segmentation of the
bronchial tree to improve their lung segmentation in this region.
Other parts of the segmentation requiring special attention are
the posterior and anterior junctions between left and right lung.
These junctions can be very narrow and consequently of low
contrast due to the PVE. The common solution to this problem
is to heuristically define a search region in each affected axial
slice and identify a separating junction line inside it. The line is
either defined as a shortest distance in anterior-posterior direction [12][14], or the soft tissue interface is traced by searching
for a minimum cost path through the gray values [9][11].
A challenge that has not yet been met is the segmentation of
lungs affected by high density pathologies that are connected
to the lung border. Due to a lack of contrast between lung and
surrounding tissues, rule-based thresholding methods will fail
to segment these pathological parts of the lung. The statistical
method in [15] employed a shape model and should, therefore, be

SLUIMER et al.: COMPUTER ANALYSIS OF COMPUTED TOMOGRAPHY SCANS OF THE LUNG: A SURVEY

more robust against pathology, but it was not tested on such scans.
Sluimer et al. [17] proposed registration with a presegmented reference scan. Another approach could be to identify surrounding
structures, such as the rib cage and the diaphragm, and combine
those algorithms into a comprehensive segmentation scheme.
Apart from the segmentation of scans containing pathology,
another direction of research that should be focused on are the
incorporation of user-feedback in automatic systems or alternatively the development of user-interactive segmentation tools,
as well as the implementation of automatic means of failure detection. In clinical practice, it is often most important to know
when the automatic algorithm failed without having to review
every segmentation result, and to be able to quickly improve the
result without having to resort to complete manual delineation.

387

From Table I it is apparent that proper validation of the

models and a good description of the data used is not always
covered in the articles. It appears that humans outperform
the discussed segmentation algorithms. Furthermore, reported
calculation times range from several minutes up to one hour per
scan, depending on the comprehensiveness of the segmentation.
Clearly, improved performance of automatic segmentation is
still wanted.
More challenges lie ahead in the segmentation of pathological
and low-dose scans. The problems to overcome are twofold:
leakage and obstructions. The source of these problems might be
artificial (noise, heart movement, high-density implants, PVE)
or real (mucus, pathology). As Table I shows, only the work
of Tschirren et al. [32] was validated for low-dose and severe
pathology, for which they reported encouraging results.

B. Segmentation of Airways
The airways exhibit a tree structure (the tracheobronchial
tree) of roughly cylindrical branches of decreasing radius.
The trachea bifurcates into the left and right main bronchus.
These bronchi repeatedly bifurcate (or trifurcate) into smaller
bronchi, up to the 23th generation [18]. The bronchial lumen
is (normally) filled with air, surrounded by the bronchial wall
which has a relatively high CT value.
On a normal dose CT scan, an expert might trace bronchi up to
generation 7. After that, the PVE is too severe, smearing lumen
and bronchus wall into an indistinguishable mass. In the last
decade, a number of methods have been proposed to (semi) automatically segment the tracheobronchial tree [19][33]. Based
upon such segmentations, computerized schemes have been developed to label the different bronchi such that branches with
problems can be pinpointed anatomically [22], [34], [35]. There
are also a number of schemes proposed to measure the geometrical properties of the bronchi at user given locations [36][38],
which can be used to diagnose a number of respiratory diseases
(see also the small overview of semi-automated measurements
of airway dimensions given by Mller and Coxson [39]).
The proposed methods for airways segmentation can be split
up into four main strategies: (i) knowledge-based segmentation
[19], [20]; (ii) region growing/wave propagation [19], [20], [22],
[25], [27][33], [40]; (iii) centreline extraction [23], [24], [26];
(iv) mathematical morphology [25], [41], [42]. Indeed, many of
the proposed schemes combine two or more of these strategies.
An example of the region growing/wave propagation strategy
is explosion controlled region growing (EC), first introduced by
Mori et al. [43], and used in [22], [25], [28]. EC is an iterative
region growing process with increasing threshold value until the
total number of voxels grown increases too much in one iteration
step.
Mathematical morphology methods focus more on identifying regions that might be part of the airways. For example
Aykac et al. [42] used 2-D gray-scale reconstruction (gray-scale
closings with increasing kernel size) to find gray-scale valleys.
In Table I, the different studies on tracheobronchial tree segmentation are listed. The items in the table include a description of the data (number of scans, slice thickness, and radiation
dose), a short description of the algorithm, whether the method
is automatic or needs manual seed points, if it is a fully 3-D
model, and the reported performance.

C. Segmentation of Vessels
Each lung contains an arterial and a venous vessel tree. Where
the pulmonary arteries and veins enter the lungs, their diameter
can be up to 30 mm. As they branch, vessel diameters decrease.
On a normal CT scan vessels can be seen up to 510 mm from
the pleura. The arteries follow the course of the bronchial tree
(when the bronchial wall is thickened, bronchus and artery have
the appearance of a signet ring).
A segmentation of the vessel trees can be of interest for
matching follow-up scans and to remove FPs of CAD schemes,
for example in the case of nodule detection (see Section IV-B1).
Conversely, vessel segmentation can provide a VOI for abnormalities that occur inside the vessels, e.g., PEs (see Section IV-C). For the latter task, contrast material is administered,
which can make the vessel segmentation task easier.
The number of studies on pulmonary vessel segmentation is
limited. Shikata et al. [45] enhanced the vessels with a filter based
on the Hessian tensor [21] that responds to bright elongated structures. Different filter scales were used depending on the distance
to the chest wall. The filtered data was thresholded to provide a
first segmentation. Seed points on vessel centerlines were used
to initialize a tracking algorithm also based on the local Hessian
tensor that could detect bifurcations that may be missed by the
vessel filter. In an experiment on five normal scans, the method
showed high sensitivity compared to manual labelings, but specificity was not evaluated. Wu et al. [46] did not use elongatedness to obtain a first segmentation, but instead employed a locally
adaptive threshold. Regulated morphological processing was applied to obtain a representation of the segmented structures in
terms of fuzzy spheres which were connected by a tracking algorithm. Only the robustness to noise was evaluated. Kiraly et
al. [47] presented a vessel tree segmentation algorithm with the
aim of segmenting the arterial subtree distal to a site of PE. A
fixed threshold and removal of small structures was used to obtain an initial segmentation. The plane perpendicular to the embolism was determined. In a VOI located distal from this plane,
a tree was extracted by skeletonization. Rules for branch sizes
and branching angles were used to remove false branches and
separate connected subtrees. The output of the algorithm is the
volume of the lung affected by PE. The same group of authors
used a similar vessel segmentation to visualize the densities inside the vessel which can be useful for radiological evaluation of

388

IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006

TABLE I
STUDIES ON TRACHEOBRONCHIAL TREE SEGMENTATION. FOR EACH STUDY, THE NUMBER (#) OF SCANS USED AND THEIR SLICE THICKNESS
(mm) IS GIVEN. THE METHOD IS BRIEFLY DESCRIBED AND IT IS STATED WHETHER IT IS 2-D OR 3-D, AND IF A MANUAL SEED IS NEEDED
(AUTO). PERFORMANCE REPORTS THE EVALUATION METHOD AND THE RESULTS

scans that may contain PE [48]. In the design of CAD systems

to detect PE, several vessel segmentation procedures are briefly
described [49], [50] but these are not specifically evaluated.
One of the main future challenges is the automatic separation
of arterial and venous trees. Furthermore, algorithms need to be
validated on larger data sets and their robustness in the presence
of pathology and noise is yet unclear.
D. Lobar Segmentation
Lungs consists of distinct anatomical compartments called
lobes. The left lung contains two lobes, and the right lung three.
The lobes are separated by fissures, which are thin sheets of
tissue. The major and minor fissures separate the lobes and are
visible on thin-slice CT. Lobar segmentation and fissure detection are, however, not equivalent: The major and minor fissures
can be incomplete, while at the same time other, accessory fissures can be visible [51], [52]. Vascular and bronchial trees do

not cross the lobar boundaries and, therefore, in the absence of

fissures the positions of the lobes can be inferred from the course
of vessels and bronchi.
It is often clinically important to determine whether a disease
affects one or more lobes, for example when lobar resection is
considered. In addition, the extraction of quantitative parameters per lobe can provide valuable information. Lobe segmentation may also prove useful in intrapatient registration.
There are several strategies that can be employed to segment
the lobes. The most obvious is to detect the fissures directly,
by locating sheet-like bright structures in 3-D or, as is also
common, line-like structures in 2-D slices. Knowledge about
the typical shape of lobes and their positions within the lungs
can be exploited. Finally, the regions containing the lobe
borders are almost devoid of larger blood vessels, so a vessel
segmentation can be used to infer lobar boundaries. Similarly,
a segmentation of the airway tree can be used.

SLUIMER et al.: COMPUTER ANALYSIS OF COMPUTED TOMOGRAPHY SCANS OF THE LUNG: A SURVEY

Zhang et al. [53], [54] performed fissure detection. A rough estimation of fissure positions was obtained through atlas registration. Subsequently fissures were delineated with ridge detection
in 2-D slices. The method allowed for interaction through manually provided anchor points. Performance was evaluated by comparing computer results with manual tracings in 22 scans of 12
subjects. Kuhnigk et al. [55] enhanced the fissures by extracting
the vasculature with region growing, computing a distance transform on this segmentation and adding that to the original data.
Next, the lobes were segmented by an interactive watershed computed from this fissure-enhanced data. The method was evaluated by measuring interobserver differences on a set of five scans.
Saita and co-workers have described various related systems for
fissure detection. In a recent version [56], a four step approach
was presented. First the vessels were segmented using a line filter
[21]. From this segmentation search volumes were determined.
A set of filters was applied to enhance sheet-like structures in
these volumes. Subsequently the fissures were found by morphological processing. A qualitative evaluation on 20 low-dose
scans was given. Wang et al. [57] presented a 2-D algorithm that
detected line-like structures through energy minimization. Once
initialized in one slice, the algorithm propagated through the
scan and used shape information from the previous slices. The
method was tested on scans from four patients.
We conclude that previous work shows encouraging results
but automatic lobe segmentation is still largely unsolved, especially in the presence of incomplete fissures and pathology, an
issue not specifically addressed in any paper.
The lobes are further subdivided in segments: ten for the right
lung and eight for the left. The segmental boundaries can only
be estimated from the course of bronchi and veins. Automatic
identification of segments is a completely open research area.
IV. REGISTRATION
Bringing images into spatial alignment, referred to as registration or matching, is one of the most common procedures
and also one of the most active research areas in medical image
analysis [58][60]. A plethora of algorithms has been proposed,
many of these general in the sense that they could be applied
as-is to chest CT. Publications on chest CT most often employ
elastic registration and typically include some dedicated modifications to standard approaches.
There are four reasons for matching CT lung scans.
Matching a CT scan to another scan of the same patient
from a different modality, typically a PET scan.
Matching to a follow-up CT scan of the same patient
for effective visual or automatic comparison to detect
or quantify interval change and/or monitor response to
therapy.
Intrapatient matching to scans acquired at a different inspiration level to study ventilation or to extract functional
information.
Interpatient matching, possibly to an atlas, to guide segmentations or detect deviations from normal appearance.
A recent example of intermodality matching is the work of
Mattes et al. [61]. PET and CT chest scans were matched with
a rigid deformation followed by an elastic deformation based

389

on cubic B-splines. Mutual information was used as similarity

criterion in a hierarchical multiresolution framework with a
quasi- Newton optimization algorithm. This approach, similar
to the system described in [62], is typical of contemporary
volume-based elastic matching algorithms. Other works on
PET-CT matching used thresholding to find the lung contours
and chamfer-matching [63], preprocessing to make both images similar in appearance followed by multi-resolution elastic
matching based on minimization of the squared difference
image [64], and translations of several VOIs using mutual
information [65].
Matching of follow-up scans serves many purposes, such
as pairing and comparing nodules [66], [67] or, more generally, displaying similar slices to a radiologist [68] for which
several commercial workstations already offer automatic solutions. Some studies use rigid deformations. Betke et al. [66]
determined a rigid alignment based on a small number of automatically identified anatomical landmarks and the lung surfaces
and succeeded in finding corresponding nodules in 56 out of 58
cases. Yamamoto et al. [68] detected similar slices in two studies
based on anatomical landmarks and simple features computed
over complete slices. Dougherty et al. [69] used an optical
flow method. Stewart et al. [70] recently described a hybrid
registration scheme that combined feature- and intensity-based
approaches which was applied to chest CT scans. The latter three
studies presented no quantitative results and showed experiments
on small numbers of scans only. Blaffert and Wiemker [67]
compared rigid, affine and elastic registration of follow-up scans
with and without the use of segmentation masks and concluded
that affine registration on the lung mask is a good compromise
between speed, accuracy and robustness. The possibility of using
the aligned scans to detect changes automatically in a CAD
system is hinted at in some studies [70], [71] but actual application of CAD systems on follow-up scans is currently limited to
commercial workstations for lung nodule work-up, e.g., [72].
Several studies focus specifically on the registration of
scans acquired at different moments of the respiratory cycle.
Matching inspiration and expiration scans is challenging because of the substantial, locally varying deformations that take
place during breathing. Fan et al. [73] described a system to
assess several functional parameters from registered scans, such
as regional lung volume changes, local changes in air content
and the distribution of ventilation. Their registration algorithm
combined feature point matching with lung surface matching
and constraints on the optical flow of mass and the smoothness
of the deformation. Other studies focused on radiotherapy
applications: Correcting for tumor motion is a major challenge
in radiotherapy treatment planning. Boldea et al. [74] used the
popular demons algorithm [75] to validate the use of active
breath control during radiotherapy treatment. In [76], an elastic
registration method was described that was tested on 2-D slices.
Both methods require long computation times. Kaus et al. [77]
compared volumetric B-spline registration with surface based
registration using manual segmentations of heart and lungs
and found that the surface based method is much faster with
comparable accuracy.
Registration across individuals is another area. Similar
methods can be employed, but many anatomical landmarks

390

IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006

that can be used for intrapatient registration are not always

available in two scans of different patients. Li et al. [78] have
constructed a human lung atlas as a general tool that can be used
for many purposes. Examples are guiding of segmentations and
establishing a normal range for local functional measurements
that could be used to detect early indications of disease. An
earlier work [79] showed that atlas registration could be used to
pinpoint the approximate location of lobar fissures which was
used later for fissure segmentation [54] (see also Section II-D).
Sluimer et al. [17] used atlas registration to segment the lungs
in scans with dense pathologies.
In any application, the proper registration method is a compromise between computation time and demands for accuracy
and robustness. Providing fast and fully automatic registration
with an accuracy comparable to manual indication of corresponding points is still a major challenge, especially in the presence of large (ventilation) deformations, pathology and noise.
No studies which address these issues specifically have been
published. On the other hand, several authors conclude that the
performance of their method already suffices for their purposes.
V. COMPUTERIZED DETECTION, QUANTIFICATION,
AND CLASSIFICATION
Segmentation and registration are often precursors to a host
of specific application-dependent image analysis systems.
In this section, compound systems on automated detection,
quantification and classification of pulmonary disorders are
discussed. These are grouped by clinical application area: emphysema (Section IV-A), lung cancer (Section IV-B1, IV-B2,
and IV-B3), PE (Section IVC), (bronchial) signs of airway
diseases (Section IV-D), and differential diagnosis of lung
disease (Section IV-E).
A. Emphysema
Emphysema is a pathology of the lung, characterized by the
destruction of lung tissue. This deficiency can be measured by
aberrations of PF tests, expressing the performance of the lungs
(e.g., residual lung capacity and diffusion of gases) as compared
to the expected performance. However, PF tests can only distinguish the progression of emphysema in rough stages: normal,
mild, or severe. Studies have shown that a volume of tissue
amounting to about a third of the total lung volume has to be
destructed before PF tests indicate a significant deviation from
healthy lungs [80]. There is a clear need for an in vivo diagnostic
tool for emphysema which can diagnose emphysema at an earlier stage and is more sensitive to small changes in the progression of the disease. The best candidate for this tool is CT. Two
earlier overviews of emphysema quantifications are by Madani
et al. [81] and Mller and Coxson [39].
Because emphysema shows up on CT as areas with abnormally low attenuation coefficients (close to that of air),
visual CT scoring of emphysema is feasible. A lot of studies
on emphysema quantification with CT focus on a reliable
automatic method as a replacement for the insensitive lung
function tests [37], [82][88] and subjective visual scoring
[83], [84], [87][94]. Indeed most of these show that computer
scoring of CT is better suited for emphysema detection than PF
tests, and more objective than visual scoring. In fact, for some

studies on emphysema, CT has already become the accepted

gold standard for quantification [83], [90], [95].
The emphysema detection studies mostly focus on extracting
from CT a single value expressing the emphysematous fraction
of the lungs. As emphysema is identified by air voxels in the
lungs, thresholding seems the best way to obtain such a value.
Usually this results in calculating one of the following.
Fraction of lung voxels with intensity values below a
given threshold
(density mask method [96], or pixel
[97]).
index (PI)
Lung area occupied by low attenuation areas larger than
some minimum size, with intensities below threshold :
LAA%
.
Mean lung density (MLD).
Fraction of histogram covered by a given lowest per.
centile:
Visually it is possible to distinguish a variety of emphysema
patterns describing the sizes and the spatial distribution of the lesions (which can be termed bullae or LAAs in literature). However, few studies [84], [88], [89], [92], [98][101] tried to incorporate this additional knowledge in the emphysema quantification. Texture features to determine if a voxel represents emphysema or not were only used by [80] and [102].
In Table II, an overview is given of emphysema quantification
studies. The table describes the data sets, the quantification
methods, distinguishes 2-D and 3-D algorithms, and reports
performance. The table covers researches concerning the
(long-term) reproducibility and trends with time of automated
CT scores [86], [95], [103][105], asthma scoring [88], and
more informative descriptive emphysema staging [80], [84],
[88][90], [98][102]. Studies that are primarily concerned with
the validation of single value replacements for visual and PF
test based grading of emphysema are not displayed in Table II.
A number of emphysema studies [82], [87], [90] focused
on the selection of patients for lung volume reduction surgery
(LVRS) instead of emphysema quantification. When emphysema is severe and concentrated in one lobe, a patient might be
eligible for LVRS, in which the emphysematous lobe is resected.
The challenges for future research on emphysema quantification lie in quantification of emphysema patterns and prognosis
of the disease. These topics are even more challenging for noisy,
low-dose screening data. It is well known from older and relatively recent studies that the typically used emphysema scores
are sensitive to noise, slice thickness, reconstruction filters and
level of inspiration [97], [104], [106][108]. How to correct for
this is an open area. Another issue is whether LAAs are in fact
responsible for decreased PF; as there are many examples of
scores over 10%, it might be that
asymptomatic patients with
the true disease is in fact not characterized by the dead space.
Possibly there is a role for CAD in solving this medical issue.
The CAD techniques developed for emphysema quantification might also be used for quantification of airtrapping (airtrapping refers to the retention of air in the lungs during expiration).
So far we havent found any studies of automatic quantification
of airtrapping, but the visual techniques used in medical literature (see e.g., [93]) are the same as those used for the assessment of emphysema, which suggest that CAD could be useful
for quantification of airtrapping.

SLUIMER et al.: COMPUTER ANALYSIS OF COMPUTED TOMOGRAPHY SCANS OF THE LUNG: A SURVEY

391

TABLE II
STUDIES ON EMPHYSEMA QUANTIFICATION. FOR EACH STUDY THE NUMBER OF NORMAL (#N) AND ABNORMAL (#A) SCANS USED IS GIVEN (WITH POSSIBLY A
NOTE TO IT). THE METHOD IS BRIEFLY DESCRIBED AND IT IS STATED WHETHER IT IS 2-D OR FULLY 3-D. PERFORMANCE REPORTS THE
EVALUATION METHOD AND THE RESULTS

B. Lung Cancer
Much of published CAD research is focused on detecting
lung cancer, which is the main cause of cancer deaths. Especially since the start of a number of lung cancer CT screening
programs, this CAD field has taken a prominent place in medical
related literature. The main focus over the past years has been to
aid the radiologists in the detection of lung nodules. Two related
areas of research cover nodule size measurements and the characterization of nodule appearance. Both are used to attempt to
estimate of the probability of malignancy. Published work for all
three areas is described in more detail in the subsections IV-B-1
through IV-B-3 below.
In general, it can be concluded that for the development of
systems that can be used in clinical practice it is necessary that
the algorithms are trained and tested on larger numbers of cases.
The collection of well-characterized cases requires a huge effort
and cannot be done on a single site. The availability of common

databases [110] could spur further development of the CAD systems substantially. It will also be increasingly important to measure system performance in a clinical setting and evaluate the
usefulness of CAD as a second reader for both experienced and
inexperienced radiologists.
1) Lung Cancer: Detection of Pulmonary Nodules: Over the
years covered in this survey (19982004), the number of articles about automated nodule detection has roughly doubled each
year. However, many of these articles deal with a small extension to a previously published method, or a different database
is used for testing. If the results in the follow-up paper are not
significantly different from the earlier results, only the latest version is discussed in this section.
As a rule, nodule detection systems consist of several steps:
a) preprocessing; b) candidate detection; c) false positive reduction; d) classification. Most often the preprocessing stage is
used to restrict the search space to the lungs and to reduce noise
and image artifacts. As will be discussed below, there are many

392

IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006

TABLE III
THE REPORTED BEST PERFORMANCES OF VARIOUS NODULE DETECTION SYSTEMS AND THE FP REDUCTION TECHNIQUES EMPLOYED. IF THE
PERFORMANCE IS OBTAINED BY MODIFYING A PREVIOUSLY REPORTED SYSTEM, THE ORIGINAL ALGORITHM IS GIVEN IN THE THIRD COLUMN. THE
DATA PARAMETERS (NUMBER OF SCANS, NUMBER OF PATIENTS, SLICE THICKNESS, AND RADIATION DOSE) MENTIONED IN THE DATA
COLUMN REFER TO THE DATA USED TO OBTAIN THE LISTED BEST PERFORMANCE

ways to generate nodule candidates, but amongst those candidate there are always many (obvious) false positives. Therefore,
one tries to cheaply and drastically reduce the number of these
FPs [step c)] before going to the more computationally expensive classification step [step d)]. Still, after the classification
stage, many false positives exist, and much of current research
on nodule detection is in fact not focused on the detection part,
but on FP reduction instead. Stages b)d) of nodule detection
systems will be covered in the following subsections.
Table III gives an overview of the different CAD models that
are covered in this survey. The performances given are the best
performances if previous models have been extended, or if new
databases have been used in later articles. If there is explicit

mention of a false positive reduction scheme in the study description, this is mentioned in the table.
a) Candidate detection schemes: For finding nodule candidates, the following techniques have been reported: multiple
gray-level thresholding [111][113], mathematical morphology
[114][117], genetic algorithm template matching of Gaussian
spheres and discs [118], [119], clustering [120][123], connected component analysis of thresholded images [124], [125],
thresholding [126][128], detection of (half) circles in thresholded images [129], gray-level distance transform [123], and
filters enhancing (spherical) structures [130][132].
The multiple gray-level thresholding technique tried to find
connected components of similar intensity, and to remove

SLUIMER et al.: COMPUTER ANALYSIS OF COMPUTED TOMOGRAPHY SCANS OF THE LUNG: A SURVEY

attached vessels. For the schemes described, mathematical

morphology covered a number of convolution filters: variable
N-Quoit filter [114], [116], [133], [134], selective marking and
depth constrained cost map [115], top-hat and sieve filter [117].
The used clustering methods differed in clustering technique
and in the features used for clustering: Kanazawa et al. [120],
Kubo et al. [122], Yamada et al. [123] applied fuzzy clustering
to intensity values and Gurcan et al. [121] used -means clustering on intensities in the original and median-filtered images.
b) Reduction of false positives: A good example of the
shift in focus from nodule detection to false-positive reduction
is the set of papers published by Armato and co-workers. In
[111], their nodule detection scheme was described, containing
the steps: preprocessing, candidate detection, and classification.
In the first step, the lungs were segmented. Candidates were
found by multiple gray-level thresholding. Using a number
of geometric and gray-level features and linear discriminant
analysis (LDA), the results of leave-one-out classification were
the detection of 70% of the nodules marked by experts, with
on average 3 FPs per image slice (i.e., about 8090 FPs per
scan). In later papers, Armato and co-workers concentrated on
schemes to reduce the number of FPs: rule-based [135], [136],
LDA [136], [137], and massive training ANN [136], [138]. The
most successful of these techniques reported a nodule detection
rate of 80.3% with on average 4.8 FPs/scan as opposed to
27.4 FPs/scan without FP reduction [138].
Saita et al. [125] added an FP reduction step to the nodule detection algorithm by Oda et al. [124]. For FP reduction prominent anatomical structures in or near the lungs were (roughly)
extracted: bones, mediastinum, and vessels. Usage of the positions of nodule candidates relative to these structures resulted in
a 100% detection rate with 2.6 FPs/scan. The original paper of
Oda et al. [124] reported 59% detected at 19.2 FPs/scan.
The model by Gurcan et al. [121] was extended with FP reduction by Ge et al. [139]: They added a 3-D gradient field as
an extra feature for the LDC. As a result the area under the ROC
increased from 0.91 to 0.93.
curve
Lee and co-workers published their detection method in
[118], and introduced an FP reduction step in [140]. In the
latter work, they added five new gray-level features and tuned
the threshold parameters of the original model. The sensitivity
of the model remained 72.4%, but the FP rate dropped from
30.8 to 5.5 per scan.
c) Classification: There is a number of classification
techniques used in the final stage of the nodule detection systems: rule-based or linear classifier [112], [118], [120], [126],
[127], [130], [131]; LDA [111], [121]; template matching
[141]; nearest cluster [114], [116]; Markov random field [133];
neural network [117], [142]; Bayesian classifier [119], [143].
The CAD schemes described in [115], [122][124], [129],
[134] did not state explicitly how a label was determined.
The most common features for classification were gray-level
features, shape descriptions, and spatial and size information.
A growing area of interest related to nodule detection is that
of nodule matching. This specific problem pertains to the localization of previously detected nodules in a follow-up scan.
Brown et al. [141] described the construction of an anatomical a priori fuzzy model which was used in combination with

393

image primitives matching to find nodules. From the results obtained from a scan, a patient-specific model was tuned which
could be applied to follow-up scans.
Ko and Betke [112] used multiple gray-level thresholding to
find nodule candidates in both original and follow-up scans.
Using position, shape, and volume information a rule-based
classification found nodule matches between scans.
With the FDA approval1 of several commercial CAD systems
for nodule detection, it seems that CAD for this field has come
to an acceptable performance. This performance is not perfect
yet, but the increased chance of finding a nodule with the help of
CAD and the achievable workload reduction for the radiologist
demand for usage of these systems in CT screenings as well
as daily hospital practice. For future work, research leading to
improved detection of ground glass opacities should have top
priority.
2) Lung Cancer: Characterization of Pulmonary Nodules: The pulmonary nodule is a dilemma for the radiologist.
Most large nodules (diameter 1 cm) in subjects at high risk
for cancer are malignant, but current CT scanners allow for
the detection of small nodules with diameters well below a
centimeter. Such nodules are extremely common and the vast
majority of them is benign [144]. Follow-up procedures to determine malignancy are often invasive, and induce risks for the
patient [145]. It is, however, of crucial importance for patient
management to determine as soon as possible whether nodules
are malignant, because symptoms of lung cancer often dont
appear until the malignancy is advanced and unresectable. As
a result, the 5 year survival for a patient diagnosed with lung
cancer is only 10%15%, but for patients in which early stage
lung cancer has been completely resected, this increases to
65%80% [146].
Several attempts have been made to design computer systems
that can help to estimate the pCa. For the design of such systems
it is obviously of crucial importance to know which characteristics point toward malignancy. This is also important for radiologists, and the subject of clinical research. It is becoming clear
that rules of thumb that apply to larger nodules do not always
hold for smaller nodules. For recent overview articles, see [144],
[146][151].
Clinical information such as old age, male sex, a history of
smoking, a history of cancer, and exposure to certain chemical
compounds increases the pCa, while other factors decrease this
probability. Bayesian analysis to include this information in the
diagnostic process was proposed by [152], [153], but applied to
nodule characterization in chest radiographs only.
The most important characteristics appear to be nodule size
and growth rate. To accurately estimate this, precise segmentation of nodules is essential. This topic is further discussed in
Section IV-B3. Another possibility is to perform a contrast-enhanced CT scan. A malignant tumor with a diameter over 2
mm must exhibit angiogenesis, which leads to contrast enhancement in the nodule [154]. To accurately determine enhancement in small nodules, again, precise segmentation is essential,
as was shown by Wormanns et al. [154]. Another noninvasive
follow-up procedure is PET with 18-fluorodeoxyglucose [155].
1The FDA is the government agency responsible for regulating medical devices in the USA.

394

IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006

TABLE IV
STUDIES ON NODULE CHARACTERIZATION. FOR EACH STUDY THE NUMBER OF BENIGN AND MALIGNANT NODULES IS GIVEN AND THE TYPE OF DATA IS LISTED
(THIN-SLICE REFERS TO AROUND 1-mm THICKNESS, THICK SLICES ARE 3 mm OR MORE). THE FEATURES AND CLASSIFIER ARE BRIEFLY DESCRIBED, THE TYPE
OF ANALYSIS (2-D/3-D) IS INDICATED, AND REPORTED RESULTS ARE SUMMARIZED

The possibilities of computer analysis to estimate pCa from PET

and CT scans simultaneously, or from both contrast and noncontrast CT have to our knowledge not yet been investigated. In this
section, we restrict our attention to the estimation of pCa from
noncontrast CT scans.
Table IV provides an overview of published studies on nodule
characterization. All studies used a combination of features to
characterize the size, shape and internal structure of the nodules.
In that way, they indirectly encoded radiologists knowledge
about indicators of malignancy. For example, certain calcification patterns are reliable indicators of benignity, while other types
can be present in malignancies [144], [149], [151]. The presence
of fat in a nodule points toward benign hamartoma [146]. The
border of nodules can be smooth, lobulated (consisting of multiple lobes) or spiculated. The former points toward benignity,
and the latter to malignancy. According to Takashima et al. [168],
[169] any nodule deviating significantly from a spherical shape
is probably benign. It also appears that concave margins and a
polygonal shape are typical for benign lesions [168], [169].
It is important to realize that thin sections are needed for good
characterization; an ill-defined small lesion on thick slices may
appear well-defined at high resolution. Moreover, 3-D computer
analysis is hardly possible using thick slices. Nevertheless, several studies used thick slices only (Table IV). The use of low-

dose may also negatively affect the possibility to make a reliable diagnosis, especially for segmentation (Section IV-B3). It
has also been reported that the use of different reconstruction
filters affected the likelihood that radiologists rate a nodule as
calcified [170]. When human ratings of nodule characteristics
are used to train computer systems, it is important to realize that
such ratings are not always reliable and reproducible [171].
A clear trend is to switch from 2-D analysis on thick slices
to 3-D analysis on thin-slice data. Recent studies tended to use
more data, but the size of training and testing databases remains a limitation; all studies resorted to leave-one-out evaluation strategies. Often good results were reported but comparisons between systems cannot be made as no standard database is employed. Performance depends heavily on the data;
when performance of radiologists was measured it ranged from
[165] to
[163]. It is interesting to note
that in some studies [163], [167] stand-alone CAD systems outperformed radiologists.
Several systems for content-based image retrieval (CBIR)
that retrieve similar cases from a database for a given nodule
at hand have also been proposed. Display of similar cases with
known classification may help radiologists to make a diagnosis.
Modest improvements in observer performance when such
similar cases are provided were reported in [165].

SLUIMER et al.: COMPUTER ANALYSIS OF COMPUTED TOMOGRAPHY SCANS OF THE LUNG: A SURVEY

Signs that are not in the direct neighborhood of the nodule

under consideration can influence the probability of malignancy.
An example is the presence of metastases elsewhere [172]. It is a
very difficult taskwell beyond the capabilities of current computer analysis systemsto find such signs automatically in general. However, often findings in the direct vicinity of a nodule
are important, such as the presence of peripheral subpleural lesions and pleural tags [168], [169] and could be integrated in
more advanced CAD systems.
When malignancy is suspected, staging the tumor is required.
CT scans are important for staging [173] but including computer
analysis in tumor staging has not been attempted yet.
Further research on nodule characterization should be focused on the integration of multiple features, extracted from
both patient history and several examinationsnot just a single
CT scanin order to compute a reliable pCa.
3) Lung Cancer: Nodule Size Measurements: The size of a
nodule highly correlates with the chance of malignancy. Growth
rate is another important indicator. Benign nodules typically
have either a very small ( 1 month, e.g., for inflammation or
pneumonia) or a very large doubling time ( 16 months). The
volume doubling time for cancers is typically between 40 and
360 days [146]. To measure nodule size and growth rate, accurate segmentation is needed. Segmentation is also required for
shape characterization (Section IV-B2) and as part of nodule detection schemes (Section IV-B1). For these reasons, nodule segmentation has received considerable attention. Older studies use
2-D outlines, but recent algorithms segment in 3-D. The major
industrial vendors currently all provide automatic nodule segmentation in their chest workstations, although they typically
require a manually indicated seed point.
It is difficult or impossible to obtain a ground truth for nodule
segmentations in real clinical data. Manual outlines, provided
by experts, have been used often, but it has been shown that
there can be interobserver and intraobserver differences in
such outlines [174]. Size measurements can also be affected
by slice thickness [175] and acquisition parameters [176].
Several studies have, therefore, used phantoms for algorithm
validation and some of these phantoms even contain structures
mimicking part-solid nodules [177]. However, it is very difficult
to realistically model the wide variety of pulmonary structures
encountered in patients using phantoms.
In absence of a ground truth, algorithms can also be evaluated
in terms of their reproducibility. Most published algorithms
require a manually indicated seed point, and a slightly different
seed point should not lead to substantially different segmentation.
This has been investigated in several studies [178][180]. More
importantly, it has been investigated if consecutive scanning of
the same patient leads to reproducible nodule size measurements
[179][181]. Reasons for measurement deviations in repeated
scans are image noise (especially evident in low dose scans),
PVE and variations in inspiration level. Wormanns and Diederich
[144] reported 95% limits of agreement for volume were around
20% using commercial software with manual interaction by a
radiologist. Note also that algorithms that make errors in volume
measurements for example when compared to manual segmentation, can still be reliable to determine growth rates, if the errors
are systematic. This was observed by Mullally et al. [182].

395

The excellent contrast between tissue and air on CT makes

segmentation of an isolated solid nodule of reasonable size a
simple task. But difficulties arise when a) the nodule is small,
so that PVE play an important role; b) the nodule is connected
to vasculature or other structures such as the pleura, fissures or
abnormalities; c) the nodule is part-solid or nonsolid, in which
case it can be difficult to define the boundary; d) the data is noisy
(typical for low-dose scans). The algorithms described in this
section have been designed to cope with difficulties a) to c).
The PVE can be dealt with by supersampling the VOI around
the nodule. A discussion of different ways to do this is given
in [183]; a comparison between binary methods and methods
that take the PVE into account is given in [177], [184]. Typically, the segmentation of solid nodules is performed by a dedicated algorithm that performs thresholding or region growing
while constraining the shape of the grown nodule, or by template matching orin the case of 2-D processing, dynamic programming (typically used in older studies, or e.g., [185]). In
[182], fixed and variable thresholds and shape based segmentation methods were compared. Attached structures were removed
by post-processing, usually with a sequence of mathematical
morphological operators or basic image processing operations
involving connected component labeling, distance transforms,
etc. Vessel and pleura each had their own sequences.
Kostis et al. [183] described 3-D algorithms for four different types of nodules, depending on their attachments. The
effect of varying parameters was described in detail and evaluation was on 16 scans for which a follow-up is available. The
method appeared to improve upon a 2-D segmentation method
from the same research group [186]. Fan et al. [178] described
a method based on template matching. The union of the template and a thresholded VOI was taken and the segmentation
was refined using certain rules. The method was extended and
used in a recent study [72] which described a complete nodule
detection, matching and size measurement system. Shen et al.
[187] described an algorithm for nodules attached to the chest
wall. Using 2-D projections, the chest wall was removed from
the segmentation. Fetita et al. [188] used dedicated sequences
of mathematical morphological operators to segment isolated,
juxtavascular and peripheral nodules but evaluation was only
performed qualitatively. Kuhnigk et al. [180] focused specifically on large, not necessarily spherical tumors with possibly
complex attachments to vessels and pleura. Their algorithm had
reasonably low (4.7%) interscan variability.
In the case of part-solid or nonsolid nodules, these approaches
typically fail. Several alternative approaches have been proposed
to deal particularly with such nodules. Okada et al. [189] used
a two-step method based on scale-space analysis that first described the nodule by an ellipsoid which was subsequently deformed by attracting the boundary to the gradient. Zhang et al.
[190] employed Markov random field modeling for ground glass
nodule segmentation. In both cases, further evaluation is necessary to better assess the value of these techniques.
Finally, there are alternative approaches that do not directly
segment the nodule but infer growth from the Jacobian matrix
of an elastic registration of two consecutive nodules [191].
Given the large number of algorithms proposed, a comparative study on a common database with a reliable reference stan-

396

dard would be very worthwhile. The data collected by the Lung

Image Database Consortium [110] could be used for such a
study. An advanced system for nodule segmentation will likely
consist of multiple algorithms each tailored for a particular type
of nodule and attachments and a recipe for choosing the best
algorithm for a nodule at hand. Systems that allow more user
interaction to correct segmentation errors made by automatic algorithms with a simple and intuitive user interface could be very
valuable in clinical practice and deserve more research. Other
future challenges are automatic correction for inspiration level
and suppression of inaccuracies caused by acquisition noise.
C. Pulmonary Embolism
Suspected acute PE is the main indication for pulmonary CT
angiography (CTA, chest CT with arterial contrast). PE is a lifethreatening complication which should be diagnosed promptly
and treated with anti-coagulants. Thin-slice CT is the preferred
modality when PE is suspected because it has high sensitivity,
relatively high specificity and it can establish an alternative diagnosis in up to one third of patients [18].
The direct signs of PE are intraluminal filling defects and lack
of enhancement of pulmonary arteries. Acute emboli can get
trapped at bifurcations or in peripheral arteries. Complete occlusion of vessels by clots is possible but residual perfusion in
the periphery is more common. This suggests two ways to detect PE: the direct detection of the clots from the HU values
in the vessels or indirect detection by localization of perfusion
defects. Herzog et al. [192] presented a visualization method
for the latter. The lungs were segmented by a simple algorithm,
vessel and bronchi were excluded by thresholding, and the average local attenuation in the remaining voxels was superimposed on the original data by color coding. Instead of presenting
these results to a radiologist through visualization, these measurements could also prove valuable for a CAD system.
The few published CAD systems in this area have focused
on direct detection. Masutani et al. [49] developed a scheme
tested on 19 scans with 3.0 mm slices. A sensitivity of 85%
was obtained at 2.6 false positives per case but scans with poor
image quality in which vessel segmentation failed were excluded. Vessels were segmented with standard techniques such
as hysteresis thresholding, region growing and mathematical
morphology. For each voxel in the vessel segmentation, the
HU value and a local contrast measure were determined and
shape features based on second order derivatives were computed. Rules were used to select suspicious voxels. These were
grouped into PE candidates and size measurements were added
to the feature set. Rules with variable thresholds provided the
final classification. Zhou et al. [50] developed a segmentation
scheme for pulmonary vessels based on histogram analysis and
an expectation-maximization algorithm followed by a tracking
algorithm. The vessel tree was subdivided in distinct regions
and intensity, shape and edge features were computed and
entered in a rule-based classifier to detect PE. When tested on 6
cases of thin-slice CT data, 58% of manually indicated emboli
were detected at the expense of 10.5 false positives per case.
Pichon et al. [48] presented a method to segment the vessel tree
and the intraluminal values were used to colorcode a rendering
of the arterial tree to make locations of clot more evident. The

IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006

main purpose of this work was to generate these renderings for

use in clinical practice, but with a simple threshold on the result
of their algorithm, component labeling and constraints on size
they were also able to construct a CAD system with 86% PE
detection sensitivity at the expense of 2 false positives per TP.
PE appears to be an area in which CAD has high potential,
because of the encouraging results reported so far and the enormous clinical importance of prompt diagnosis. PE is, after cardiac disease, the second cause of sudden, unexpected death with
over 50 000 deaths in the USA alone every year. Several major
industrial vendors are, therefore, active in this area and have recently demonstrated prototypes of PE CAD systems. Considering the relatively small number of studies published so far,
the small number of cases on which they have been tested, the
modest performance obtained, and the simple sets of features
and classifiers employed, there should be ample room for further improvement.
Apart from detection, quantification of PE is also important.
Qanadli et al. [193] proposed a quantitative index, but automatic
quantification of PE has not been reported yet.
D. Airways Diseases (Bronchial Signs)
Signs of airways diseases that the bronchi may exhibit on
thin-slice CT examinations are wall thickening and dilation or
narrowing [194], [195]. These three clear signs lend themselves
well to automated analysis. Thus far, the work on this has centered on measurements of bronchial cross sections on 2-D transverse slices, in analogy to current methods of visual assessment
by a radiologist.
The bronchi are usually of equal size as their accompanying
artery. Normal bronchovascular cross-sectional pairs, therefore,
appear on CT as a figure eight with one hole open (the bronchus)
and one hole closed (the artery). It is the size relation between
vessel and bronchus that indicates bronchial narrowing or dilatation. A few studies focus on the detection of normal [196] as
well as abnormal [197] bronchovascular pairs. In other studies
these pairs were indicated manually and automated size measurements were performed [36], [37], [198], [199]. Such size
measurements always incorporated a correction for nonperpendicular cross sections and were always validated on phantoms.
Nakano et al. [37] correlated bronchial size measurements to
clinical parameters of COPD in smokers.
Two-dimensional analysis reflects the way in which a radiologist currently assesses the airways. This restricts the measurements to specific favorable cross sections in a transverse slice.
Three-dimensional analysis will soon become the standard in
this area, however. The combination of measurement and 3-D
bronchial segmentation techniques (Section II-B) will provide
more data for the analysis which can be of added clinical value.
Techniques developed for comprehensive virtual bronchoscopy
systems, e.g., [47] provide the basis for 3-D airway analysis.
E. Diffuse Lung Disease
There is a large group of disorders that primarily affect the
lung parenchyma. This group is referred to by the generic term
diffuse parenchymal lung disease (DPLD), but terms such as
Interstitial lung disease (ILD) or infiltrative lung disease are also

SLUIMER et al.: COMPUTER ANALYSIS OF COMPUTED TOMOGRAPHY SCANS OF THE LUNG: A SURVEY

encountered. The DPLDs account for about 15% of respiratory

practice [200].
Thin slice CT plays an important role in the detection, diagnosis and follow-up of these disorders. They are characterized
by specific abnormal findings mostly texture-like in appearance
and it is the cooccurrence of several such findings that can point
toward a specific diagnosis [194].
It is for this reason that computer analysis of DPLD is commonly viewed as a texture analysis problem. Systems designed
to quantify certain lung diseases and/or differentiate between
them, are without fail based on the vector space paradigm [7].
This means that ROIs are represented by features that are input
to a classifier which is trained to (re)produce category labels.
Most often the application of these systems revolves around the
detection of abnormal tissue and its simultaneous classification
into several textural categories [201][207]. One study focusses
exclusively on the detection of abnormalities without further
classification [208]. A different field of application (employing
the same pattern recognition tools) is that of content-based image
retrieval, in which one group has been active [209][212].
With respect to implementation, various choices are made
about (i) what ROIs to use, (ii) what features, and (iii) what classifier. ROI sizes typically lie in the range of 31 31 to 96 96
pixels. The smallest ROIs are 9 9 pixels [204], the largest encompass an entire lung field [202]. For the features representing
the ROIs, a discriminatory set can be automatically chosen by
supervised feature selection from families of textural features
known from pattern recognition theory, such as histogram statistics, features from filterbanks or cooccurrence matrices, runlength parameters or fractal features [201][203], [208], [213].
Alternatively, features may be designed for the task at hand
[205], or a combination of both can be used [209]. Kauczor et
al.[204] use a combination of two artificial neural networks into
which the raw pixel values in the ROIs are input as features.
The ANN (also employed by Uchiyama et al. [205]) is only
one example of classifier. We find also LDCs [201], Bayesian
classifiers [202], [203], [213] and -nearest-neighbor classifiers
[208], [209].
Table V provides a summary of the described studies with respect to their objective, the data that was used and the evaluation.
Descriptions of purpose and the lists of used classification categories show considerable variation, illustrating that in this field
there is no consensus yet on a common goal and exact definitions. Although follow-up measurements for the quantification
of disease progression are mentioned as a possible area of interest, in none of the studies this task is performed explicitly.
A noticeable common problem to all studies is the establishment of a reliable reference standard. This is due to the
fact that texture classification in thin-slice CT scans is a complex task even for experts: Low intraobserver and interobserver
agreements of about 50% [203] reflect this. We see that in most
studies the opinion of a single expert is used as a reference standard. In some cases, multiple radiologists are involved and a
consensus is taken. For samples on which there is no agreement
between the radiologists, the computer may be regarded as another observer. It can then be ascertained whether its behavior
is statistically equivalent to that of the radiologists, avoiding the
use of an explicit reference standard.

397

For completely automated differential diagnosis of DPLD,

much more information will need to be incorporated into the
computer systems than is the case at present. The scans need to
be analyzed on a global rather than a local scale. This means that
the full 3-D character of modern scan data should be exploited
and information from the entire scan (e.g., from all regions of
interest) should be combined to come to a single diagnosis. In
this process, also, the use of anatomical knowledge is indispensable. Shyu et al. [209] are the only group incorporating forms
of global and anatomical knowledge, combining features from
several pathological regions and anatomical indicators per slice.
The regions are, however, manually delineated rather than automatically detected. Recently also the first work appeared analyzing 3-D objects in complete scans [207], although these 3-D
objects were combined from multiple candidates detected beforehand on 2-D slices. Another source of information that is of
paramount diagnostic value but is not being taken into account
so far is clinical data. Fukushima et al. [214] showed that an expert system can come to reasonable diagnoses of DPLD when
image analysis results (in their case rated by radiologists, not
automatically) are augmented by clinical parameters.
We conclude that with the systems described in this section
the first steps toward more advanced processing schemes have
been taken, but that in computer analysis of DPLD, the question
on what exactly to aim for and how to achieve it is still open.
VI. DISCUSSION
This final section consists of four parts. First we summarize
what has been achieved so far and what is the state of the art.
Then we identify a number of challenges for academic research.
Several remarks to this end have already been raised specifically
in each of the topics in Sections IIV; here we unify those issues.
Third, we specifically focus on the discrepancy between theoretical availability of an algorithm and its availability in clinical
routine, and try to explain its causes. Finally we provide an outlook to new developments that can be expected in the coming
years.
A. State of the Art
Initially, the clinical use of computer analysis of CT of the
lungs did not reach much further than the quantification of emphysema, a direct consequence of its relatively simple detection on CT. Simple tools for emphysema quantification are now
available on commercial workstations.
Ongoing improvements in scanner speed and quality have
broadened the field of applications. Currently, the emphasis
lies on detection and analysis of pulmonary nodules. We see
CAD for lung cancer follow in the footsteps of CAD for breast
cancer: Several commercial systems for automated nodule
detection have acquired FDA approval and research effort is
shifting from detection to characterization and follow-up. In
this field, industrial R&D efforts probably outweigh those of
academic research. Commercial workstations are now capable
of automatic nodule volumetry and nodule detection. These
workstations perform segmentation, but beyond complete lung
segmentation, these segmentation techniques are not directly
available yet for clinical use.

398

IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006

TABLE V
STUDIES ON THE (TEXTURE) ANALYSIS OF DPLD. FOR EACH STUDY A DESCRIPTION OF THE PURPOSE IS GIVEN AND OF THE CATEGORIES INVOLVED (DISEASE OR
TEXTURAL). THE DATA USED IS DESCRIBED ON THE LEVEL OF SUBJECTS/SCANS, AS WELL AS THE LEVEL OF SLICES/ROIS. IT IS STATED HOW
A REFERENCE STANDARD WAS SET, AND A CHARACTERISTIC RESULT IS GIVEN

Apart from the pulmonary nodule, the lungs may contain

much more pathology. Segmentation, registration and further image analysis of each anatomical structure has been
researched, as is evident from this survey, but has not reached
the status of routine clinical use.
B. Research Challenges
For all methods of analysis the present trend is to move from
(axial) 2-D to 3-D processing, born out of possibility as much
as necessity.
This inevitably leads to a need to create suitable reference
standards in 3-D. This represents a major hurdle for evaluation
of tasks such as segmentation, for which the reference standard
is classically set using manually traced outlines. It is a prohibitively large amount of work to create manual delineations of
structures of interest in thorax exams that typically consist of

over 300 slices. More effort should be spent on the development

of interactive tools (e.g., [215], [216]). Apart from their use in
research for easy establishment of a ground truth, these tools can
be used in clinical practice to manually correct the (as of yet) unsatisfactory automated segmentation outcomes. Moreover, these
tools can be developed in such a way as to enforce consistency
of manual segmentations in 3-D, as manual delineations performed on 2-D slices often lack such consistency over larger
volumes.
Maybe the largest challenge is to design algorithms that are
robust against pathological and anatomical variety, image noise
and differences in acquisition parameters. Many of the algorithms reviewed here have only been tested on very select patient
populations that do not necessarily represent clinical practice. It
is well possible that fundamentally different methodological approaches are necessary to achieve this increased robustness.

SLUIMER et al.: COMPUTER ANALYSIS OF COMPUTED TOMOGRAPHY SCANS OF THE LUNG: A SURVEY

Proper validation is indispensable in order to assess the

quality and application range of an algorithm. Algorithms
should be tested in context of their intended practical clinical
use, with due consequences for the choice of test data and
measures of performance. Test data should reflect the variety of
pathology, noise levels and acquisition variety encountered in
clinical practice. However, a lack of robustness does not have
to hinder clinical use of a system as long as it is compensated
by adequate means of failure detection. Failure detection is an
important part of an automated system, but this fact is typically
ignored in academic research. A clinical user should not be
required to examine closely either the input data, or the (intermediate) results to estimate the validity of the final outcome
of an automated analysis. Construction of adequate failure
detection mechanisms is far from trivial and in our opinion this
topic deserves more research.
Ideally, common, carefully annotated and representative
databases should be available for algorithm benchmarking.
This does, however, require consensus on the definition and
relevance of the clinical questions to be answered, which has
not been reached for all discussed areas of application. Detection of lung nodules is a task with a clear definition of purpose
and a vast range of algorithms in need of competitive testing,
presenting a prime example for which a common database can
and should be assembled. This necessity has been recognized
and steps in this direction are being undertaken [110].
In many of the studies describing application oriented work
(Section IV), research efforts are necessarily focused on the final
stages of the comprehensive image analysis system, and preprocessing modules (such as lung segmentation) are implemented
in a suboptimal way. Many of these systems would benefit from
incorporation of existing methods from other research groups.
This is a common problem to science as well as industry and can
not be solved easily (making research software publicly available might be a solution). It is important to be aware of the limitations this might set on system performance. Situations should
be avoided where effort goes into refinement of the final stages
of a compound system when the true bottleneck lies elsewhere.
C. Clinical Application
The advent of multislice CT has made isotropic imaging of
the lung a standard technique. Since nonisotropic data was one
of the main problems in computer vision of CT data, it should
now be easy to use various computer assisted techniques in clinical routine practice. Unfortunately, of the many potential applications for the lungs described in this survey, only very few are
available on the market and even fewer are provided by the large
CT manufacturers.
What are the reasons for this discrepancy? For an algorithm
to succeed in clinical practice, there needs to be a market for
it (a real clinical need by a sufficient amount of radiologists),
it needs to be fast and easy to use (patient throughput is ever
increasing with multislice CT), it needs to be reliable (correct
results in the vast amount of cases) and it should be correctable
(easy adjustments in case the algorithm fails). Because radiologists are used to working without such algorithms, it is not easy
to market new tools. Only when radiologists fail miserably, such

399

as in the case of nodule volumetry and detection of small nodules, and when there is a growing need to perform such tasks
(lung cancer screening depends on nodule detection and volumetry) do such algorithms find their way into clinical practice.
In addition, regulatory bodies such as the FDA require positive
proof of the benefit of such techniques and their use may be
hampered by medico-legal considerations.
Many if not all techniques described in this survey have the
potential to be important in clinical practice. How much so,
however, will mainly depend on their ease of use and their reliability. As mentioned in the previous section, improving reliability and robustness is a major research challenge. In addition, most algorithms have not been optimized for speed, and
even if they had, they probably would take longer than most radiologists would accept. New concepts are, therefore, needed
to implement such algorithms in clinical practice. One potential solution could be automated preprocessing of the data as
soon as it is sent from the scanner to the CT workstation: such
preprocessing should include the time-consuming steps of any
algorithm that is made available. The radiologist then should
have the results at the fingertips and can easily useand ideally also modifythe results in any patient in whom he deems
a particular computer-assisted analysis valuable. By making results of CAD easily accessible in every patient, radiologists will
be much more prone to using these results. Speed and ease of
use, therefore, will be the determining factors as soon as performance of algorithms has risen beyond a certain basic level.
The fact that many good algorithms are not available in clinical practice demonstrates a general dilemma in the image processing community: Research money is mainly available for
basic work and algorithm development but not for implementation and optimization of technique. Workflow issues will ultimately determine whether a new technique is practicable or
not. Close collaboration between academic image analysis researchers, radiologists and industry is mandatory for success.
D. Outlook
The past few years have seen the introduction of nodule volumetry and detection into clinical workstations. Techniques for
classification of nodules are ready to follow relatively soon but
probably will be hampered by medico-legal considerations. The
available techniques will have to be upgraded to suit the increased exposure to clinical cases, which will invariably lead to
new problems that have previously not been considered. However, within a rather short period of time, nodule detection, measurement and classification can be expected to be standard in
clinical practice.
A new hot application is the automated detection of PE, a
task that has been previously considered trivial but now becomes
more cumbersome because many more vessels are visible and
have to be evaluated on multislice CT data sets. Techniques that
indicate potential emboli, and later may even help in differentiating emboli from artifacts would be very welcome in clinical
practice.
Registration of various types of scans is already available for
PET/CT and SPECT/CT data. Registration of lung nodules is
already implemented in commercial workstations but should
become even more reliable in the future. The registration of

400

IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006

follow-up scans should have a bright future, mainly because it is

a daily problem for radiologists. This problem is becoming increasingly important with the growing workload with new multislice CT systems but it is not trivial to solve because of varying
inspirational levels.
Emphysema quantification has been around for many years
but is still limited to research settings because the clinical benefit of a precise quantification is not yet evident for the vast majority of patients. Problems such as the correction for varying
inspirational levels or the correction for image noise have not
yet been fully solved. While simple threshold-based techniques
have been made available on clinical workstations, there remains a substantial amount of work to be done before such techniques will be fully established in everyday practice.
Evaluation of airway morphology, especially airway thickness, may become an important topic in the future because airways are much more readily responsive to treatment, and computer analysis could be used as a faster way to study the effect
of a new drug than analysis of lung function or patient outcome.
Classification and quantification of interstitial lung disease
is difficult, and even experienced chest radiologists frequently
struggle with differential diagnoses. Automated schemes that
indicate a percentage of affected lung or the probability of a
certain disease would certainly be welcome, but require much
more research.
The combination of these issuesairway evaluation, regional
texture analysis of parenchymal structure, quantification of
ventilation, perfusion and airtrapping, etc.turns MSCT into a
modality that is suited for functional analysis of the lung, in addition to assessment of anatomy. An outlook on the possibilities
for functional chest imaging with CT is given in [217]. But only
robust automated image analysis and availability of reference
data on both normal and abnormal variability will make such
a comprehensive analysis valuable in clinical practice. This
defines a challenging research agenda for the next decade.
A quick analysis of the roughly 300 publications considered
for this survey reveals that the amount of publications in this
field has grown on average by a factor 1.5 per year over the past
five years: a practical indicator to the increase in research effort
accompanying the modern (r)evolution of CT. Given the clinical
importance of chest CT and the existing challenges that we have
identified, it is to be expected that the computer analysis of chest
CT scans will remain a very active research field in the years to
come.
APPENDIX
LITERATURE COLLECTION
Of the following journals, all publications on chest CT
analysis in the years 1999 through 2004 were included: Medical Physics, Radiology, IEEE TRANSACTIONS ON MEDICAL
IMAGING (TMI), Med. Image Anal. (MedIA), Acad. Radiol.,
American Journal of Roentgenology (AJR), American Journal
of Respiratory and Critical Care Medicine and the European
Journal of Radiology. For some journals differing time spans
were covered for reasons of availability. These were TMI
(19882004), MedIA (19972004), Acad. Radiol. (20012004)
and AJR (20002002). The following conference proceedings
of the years 1999 through 2004 were searched: SPIE Medical

Imaging, Medical Image Computing and Computer Assisted

Intervention (MICCAI), Information Processing in Medical
Imaging (IPMI), and the IEEE Symposium on Biomedical
Imaging (ISBI). Contributions in alternate sources were identified from citations in the aforementioned publications and from
searches using the online database search engines PubMed,
IEEEXplore, and the Science Citation Index. The key that
was sought on was (computed tomography OR CT) AND
(chest OR thorax OR lung OR lungs OR pulmonary) AND
(computer OR computerized OR automatic OR computerised
OR automated) and this resulted in 1659, 197, and 883 hits,
respectively. When similar contributions have appeared only
the most recent or comprehensive paper has been included.
REFERENCES
[1] S. A. Kollins, Computed tomography of the pulmonary parenchyma
and chest wall, Radiolog. Clin. N. Am., vol. 15, no. 3, pp. 297308,
1977.
[2] E. L. Ritman, J. H. Kinsey, R. A. Robb, B. K. Gilbert, L. D. Harris, and E.
H. Wood, Three-dimensional imaging of heart, lungs, and circulation,
Science, vol. 210, pp. 273280, 1980.
[3] M. A. Meziane, R. H. Hruban, E. A. Zerhouni, <Au: Are the initials W.
P. S. appropriate here?> N. F. Khouri, E. K. Fishman, G. M. Hutchins,
and S. S. Siegelman, High resolution CT of the lung parenchyma with
pathologic correlation, Radiographics, vol. 8, no. 1, pp. 2754, 1988.
[4] G. D. Rubin, Data explosion: the challenge of multidetector-row CT,
Eur. J. Radiol., vol. 36, pp. 7480, 2000.
[5] M. Sonka, V. Hlavac, and R. Boyle, Image Processing, Analysis, and
Machine Vision, 2nd ed. Pacific Grove, CA: PWS, 1999.
[6] R. C. Gonzalez and R. E. Woods, Digital Image Processing, 2nd
ed. New York: Addison-Wesley, 2002.
[7] R. O. Duda, P. E. Hart, and D. G. Stork, Pattern Classification, 2nd
ed. New York: Wiley, 2001.
[8] C. M. Bishop, Neural Networks for Pattern Recognition. Oxford,
U.K.: Oxford Univ. Press, 1995.
[9] S. G. Armato and W. F. Sensakovic, Automated lung segmentation for
thoracic CT, Acad. Radiol., vol. 11, no. 9, pp. 10111021, 2004.
[10] M. S. Brown, M. F. McNitt-Gray, N. J. Mankovich, J. G. Goldin, J.
Hiller, L. S. Wilson, and D. R. Aberle, Method for segmenting chest
CT image data using an anatomical model: preliminary results, IEEE
Trans. Med. Imag., vol. 16, no. 6, pp. 828839, Dec. 1997.
[11] S. Hu, E. A. Hoffman, and J. M. Reinhardt, Automatic lung segmentation for accurate quantitation of volumetric X-ray CT images, IEEE
Trans. Med. Imag., vol. 20, no. 6, pp. 490498, Jun. 2001.
[12] A. Silva, J. S. Silva, B. S. Santos, and C. Ferreira, Fast pulmonary contour extraction in X-ray CT images: a methodology and quality assessment, Proc. SPIE, vol. 4321, pp. 216224, 2001.
[13] B. Zheng, J. K. Leader, G. S. Maitz, B. E. Chapman, C. R. Fuhrman,
R. M. Rogers, F. C. Sciurba, A. Perez, P. Thompson, W. F. Good, and
D. Gur, A simple method for automated lung segmentation in X-ray
CT images, Proc. SPIE (Medical Imaging), vol. 5032, pp. 14551463,
2003.
[14] J. K. Leader, B. Zheng, R. M. Rogers, F. C. Sciurba, A. Perez, B. E.
Chapman, S. Patel, C. R. Fuhrman, and D. Gur, Automated lung segmentation in X-ray computed tomography, Acad. Radiol., vol. 10, no.
11, pp. 12241236, 2003.
[15] B. Li and J. M. Reinhardt, Automatic generation of 3-D shape models
and their application to tomographic image segmentation, Proc. SPIE
(Medical Imaging), vol. 4322, pp. 311322, 2001.
[16] S. Ukil and J. M. Reinhardt, Smoothing lung segmentation surfaces in
3-D X-ray CT images using anatomic guidance, Proc. SPIE (Medical
Imaging), vol. 5370, pp. 10661075, 2004.
[17] I. C. Sluimer, M. Niemeijer, and B. van Ginneken, Lung field segmentation from thin-slice CT scans in presence of severe pathology, Proc.
SPIE (Medical Imaging), vol. 5370, pp. 14471455, 2004.
[18] Spiral and Multislice Computed Tomography of the Body, M. Prokop
and M. Galanski, Eds., Thieme, Stuttgart, Germany, 2003.
[19] M. Sonka, W. Park, and E. A. Hoffman, Rule-based detection of intrathoracic airway trees, IEEE Trans. Med. Imag., vol. 15, no. 3, pp.
314326, Jun. 1996.

SLUIMER et al.: COMPUTER ANALYSIS OF COMPUTED TOMOGRAPHY SCANS OF THE LUNG: A SURVEY

[20] W. Park, E. A. Hoffman, and M. Sonka, Segmentation of intrathoracic

airway trees: a fuzzy logic approach, IEEE Trans. Med. Imag., vol. 17,
no. 4, pp. 489497, Aug. 1998.
[21] Y. Sato, S. Nakajima, N. Shiraga, H. Atsumi, S. Yoshida, T. Koller, G.
Gerig, and R. Kikinis, Three-dimensional multi-scale line filter for segmentation and visualization of curvilinear structures in medical images,
Med. Image Anal., vol. 2, no. 2, pp. 143168, 1998.
[22] K. Mori, J. Hasegawa, Y. Suenaga, and J. Toriwaki, Automated
anatomical labeling of the bronchial branch and its application to the
virtual bronchoscopy system, IEEE Trans. Med. Imag., vol. 19, no. 2,
pp. 103114, Feb. 2000.
[23] T. Deschamps and L. D. Cohen, Fast extraction of minimal paths in 3-D
images and applications to virtual endoscopy, Med. Image Anal., vol.
1, pp. 119, 2001.
[24] S. R. Aylward and E. Bullitt, Initialization, noise, singularities, and
scale in height ridge traversal for tubular object centerline extraction,
IEEE Trans. Med. Imag., vol. 21, no. 2, pp. 6175, Feb. 2002.
[25] A. P. Kiraly, W. E. Higgins, G. McLennan, E. A. Hoffman, and J. M.
Reinhardt, Three-dimensional human airway segmentation methods
for clinical virtual bronchoscopy, Acad. Radiol., vol. 9, no. 10, pp.
11531168, 2002.
[26] R. D. Swift, A. P. Kiraly, A. J. Sherbondy, A. L. Austin, E. A. Hoffman,
G. McLennan, and W. E. Higgins, Automatic axis generation for virtual
bronchoscopic assessment of major airway obstructions, Comput. Med.
Imag. Graphics, vol. 26, pp. 103118, 2002.
[27] J. Tschirren, K. Palagyi, J. M. Reinhardt, E. A. Hoffman, and M. Sonka,
Segmentation, skeletonization, and branchpoint matchinga fully automated quantitative evaluation of human intrathoracic airway trees,
in Lecture Notes in Computer Science. Berlin, Germany: SpringerVerlag, 2002, vol. 2489, Medical Image Computing and Computer-Assisted Intervention, pp. 1219.
[28] T. Kitasaka, K. Mori, Y. Suenaga, J. Hasegawa, and J. Toriwaki, A
method for segmenting bronchial trees from 3-D chest X-ray CT
images, in Lecture Notes in Computer Science. Berlin, Germany:
Springer-Verlag, 2003, vol. 2879, , pp. 603610.
[29] K. C. Yu, E. L. Ritman, A. P. Kiraly, S. Y. Wan, M. Zamir, and W. E. Higgins, Toward reliable multi-generational analysis of anatomical trees
in 3-D high-resolution CT images, Proc. SPIE (Medical Imaging), vol.
5031, pp. 178186, 2003.
[30] D. Mayer, D. Bartz, J. Fischer, S. Ley, A. del Rio, S. Thust, H.
Kauczor, and C. P. Heussel, Hybrid segmentation and virtual bronchoscopy based on CT images, Acad. Radiol., vol. 11, no. 5, pp.
551565, 2004.
[31] H. Singh, M. Crawford, J. Curtin, and R. Zwiggelaar, Automated 3-D
segmentation of the lung airway tree using gain-based region growing
approach, in Lecture Notes in Computer Science. Berlin, Germany:
Springer-Verlag, 2004, vol. 3217, Medical Image Computing and Computer-Assisted Intervention, pp. 975982.
[32] J. Tschirren, E. A. Hoffman, G. McLennan, and M. Sonka, Airway tree
segmentation using adaptive regions of interest, Proc. SPIE (Medical
Imaging), vol. 5369, pp. 117124, 2004.
[33] S. Zidowitz, A. H. Schmidt, A. Kriete, S. Krass, and H. O. Peitgen,
Steps toward a patient individual geometric model of the bronchial-tree
used for functional simulations, Proc. SPIE (Medical Imaging), vol.
5369, pp. 125131, 2004.
[34] H. Kitaoka, Y. Park, J. Tschirren, J. Reinhardt, M. Sonka, G. McLennan,
and E. A. Hoffman, Automated nomenclature labeling of the bronchial
tree in 3-D-CT lung images, in Lecture Notes in Computer Science. Berlin, Germany: Springer-Verlag, 2002, vol. 2489, Medical
Image Computing and Computer-Assisted Intervention, pp. 111.
[35] K. Palagyi, J. Tschirren, and M. Sonka, Quantitative analysis of
intrathoracic airway trees: methods and validation, in Lecture Notes
in Computer Science. Berlin, Germany: Springer-Verlag, 2003, vol.
2732, Information Processing in Medical Imaging, pp. 222233.
[36] J. M. Reinhardt, N. D. DSouza, and E. A. Hoffman, Accurate measurement of intrathoracic airways, IEEE Trans. Med. Imag., vol. 16, no. 6,
pp. 820827, Dec. 1997.
[37] Y. Nakano, S. Muro, H. Sakai, T. Hirai, K. Chin, M. Tsukino,
K. Nishimura, H. Itoh, P. D. Par, J. C. Hogg, and M. Mishima,
Computed tomographic measurements of airway dimensions and
emphysema in smokers, Am. J. Respir. Crit. Care Med., vol. 162, no.
3, pp. 11021108, 2000.
[38] J. R. Ederle, C. P. Heussel, J. Hast, B. Fischer, E. J. R. van Beek, S. Ley,
M. Thelen, and H. U. Kauczor, Evaluation of changes in central airway
dimensions, lung area and mean lung density at paired inspiratory/expiratory high-resolution computed tomography, Eur. Radiol., vol. 13, pp.
24542461, 2003.

401

[39] N. L. Mller and H. Coxson, Chronic obstructive pulmonary disease:

imaging the lungs in patients with chronic obstructive pulmonary disease, Thorax, vol. 57, no. 11, pp. 982985, 2002.
[40] T. Schlathlter, C. Lorenz, I. C. Carlsen, S. Renisch, and T. Deschamps,
Simultaneous segmentation and tree reconstruction of the airways for
virtual bronchoscopy, Proc. SPIE, vol. 4684, pp. 103113, 2002.
[41] C. I. Fetita and F. Prteux, Quantitative 3-D CT bronchography, in
Proc. IEEE Int. Symp. Biomedical Imaging, 2002, pp. 221224.
[42] D. Aykac, E. A. Hoffman, G. McLennan, and J. M. Reinhardt, Segmentation and analysis of the human airway tree from three-dimensional
X-ray CT images, IEEE Trans. Med. Imag., vol. 22, no. 8, pp. 940950,
Aug. 2003.
[43] K. Mori, J. I. Hasegawa, J. I. Toriwaki, H. Anno, and K. Katada, Recognition of bronchus in three-dimensional X-ray CT images with application to virtualized bronchoscopy system, in Proc. Int. Conf. Pattern
Recognition, vol. 3, 1996, pp. 528532.
[44] S. Y. Wan, E. L. Ritman, and W. E. Higgins, Multi-generational analysis
and visualization of the vascular tree in 3-D micro-CT images, Comput.
Biol. Med., vol. 32, no. 2, pp. 5571, 2002.
[45] H. Shikata, E. A. Hoffman, and M. Sonka, Automated segmentation of
pulmonary vascular tree from 3-D CT images, Proc. SPIE, vol. 5369,
pp. 107116, 2004.
[46] C. Wu, G. Agam, A. S. Roy, and S. G. Armato, Regulated morphology
approach to fuzzy shape analysis with application to blood vessel extraction in thoracic CT scans, Proc. SPIE, vol. 5370, pp. 12621270, 2004.
[47] A. P. Kiraly, E. Pichon, D. P. Naidich, and C. L. Novak, Analysis of
arterial subtrees affected by pulmonary emboli, Proc. SPIE, vol. 5370,
pp. 17201729, 2004.
[48] E. Pichon, C. L. Novak, A. P. Kiraly, and D. P. Naidich, A novel method
for pulmonary emboli visualization from high-resolution CT images,
Proc. SPIE, vol. 5367, pp. 161170, 2004.
[49] Y. Masutani, H. MacMahon, and K. Doi, Computerized detection
of pulmonary embolism in spiral CT angiography based on volumetric image analysis, IEEE Trans. Med. Imag., vol. 21, no. 12, pp.
15171523, Dec. 2002.
[50] C. Zhou, L. M. Hadjiiski, B. Sahiner, H. P. Chan, S. Patel, P. N. Cascade,
E. A. Kazerooni, and J. Wei, Computerized detection of pulmonary
embolism in 3-D computed tomographic (CT) images: vessel tracking
and segmentation techniques, Proc. SPIE, vol. 5032, pp. 16131620,
2003.
[51] K. Hayashi, A. Aziz, K. Ashizawa, H. Hayashi, K. Nagaoki, and H. Otsuji, Radiographic and CT appearances of the major fissures, Radiographics, vol. 21, no. 4, pp. 861874, 2001.
[52] A. Yildiz, F. Glpinar, C. Mukadder, M. N. Ducea, C. zer, and F. D.
Apaydin, HRCT evaluation of the accessory fissures of the lung, Eur.
J. Radiol., vol. 49, pp. 245249, 2004.
[53] L. Zhang, E. A. Hoffman, and J. M. Reinhardt, Lung lobe segmentation
by graph search with 3-D shape constraints, Proc. SPIE, vol. 4321, pp.
204215, 2001.
[54]
, Atlas-driven lung lobe segmentation in volumetric X-ray CT images, Proc. SPIE, vol. 5031, pp. 306315, 2003.
[55] J. M. Kuhnigk, H. K. Hahn, M. Hindennach, V. Dicken, S. Krass, and H.
O. Peitgen, Lung lobe segmentation by anatomy-guided 3-D watershed
transform, Proc. SPIE, vol. 5032, pp. 14821490, 2003.
[56] S. Saita, M. Yasutomo, M. Kubo, Y. Kawata, N. Niki, K. Eguchi, H.
Ohmatsu, R. Kakinuma, M. Kaneko, M. Kusumoto, M. Moriyama, and
M. Sasagawa, An extraction algorithm of pulmonary fissures from
multi-slice CT image, Proc. SPIE, vol. 5370, pp. 15901597, 2004.
[57] J. Wang, M. Betke, and J. P. Ko, Shape-based curve growing model
and adaptive regularization for pulmonary fissure segmentation in CT,
in Lecture Notes in Computer Science. Berlin, Germany: SpringerVerlag, 2003, vol. 3216, Medical Image Computing and Computer-Assisted Intervention, pp. 541548.
[58] J. B. A. Maintz and M. A. Viergever, A survey of medical image registration, Med. Image Anal., vol. 2, pp. 136, 1998.
[59] H. Lester and S. R. Arridge, A survey of hierarchical nonlinear medical
image registration, Pattern Recogn., vol. 32, pp. 129149, 1999.
[60] J. P. W. Pluim, J. B. A. Maintz, and M. A. Viergever, Mutual-information-based registration of medical images: a survey, IEEE Trans. Med.
Imag., vol. 22, no. 8, pp. 9861004, Aug. 2003.
[61] D. Mattes, D. R. Haynor, H. Vesselle, T. K. Lewellen, and W. Eubank,
PET-CT image registration in the chest using free-form deformations,
IEEE Trans. Med. Imag., vol. 22, no. 1, pp. 120128, Jan. 2003.
[62] D. Rueckert, L. I. Sonoda, C. Hayes, D. L. G. Hill, M. O. Leach, and
D. J. Hawkes, Nonrigid registration using free-form deformations: application to breast MR images, IEEE Trans. Med. Imag., vol. 18, no. 8,
pp. 712721, Aug. 1999.

402

[63] J. Cai, J. C. H. Chu, D. Recine, M. Sharma, C. Nguyen, R. Rodebaugh,

V. A. Saxena, and A. Ali, CT and PET lung image registration and
fusion in radiotherapy treatment planning using the chamfer-matching
method, Int. J. Radiat. Oncol. Biology Physics, vol. 43, pp.
883891, 1999.
[64] Y. C. Tai, K. P. Lin, C. K. Hoh, S. C. H. Huang, and E. J. Hoffman, Utilization of 3-D elastic transformation in the registration of chest X-ray
CT and whole body PET, IEEE Trans. Nucl. Sci., vol. 44, no. 4, pp.
16061621, Aug. 1997.
[65] Y. E. Erdi, K. Rosenzweig, A. K. Erdi, H. A. Macapinlac, Y. C. Hu, L. E.
Braban, J. L. Humm, O. D. Squire, C. S. Chui, S. M. Larson, and E. D.
Yorke, Radiotherapy treatment planning for patients with nonsmall cell
lung cancer using positron emission tomography (PET), Radiotherapy
Oncol., vol. 62, pp. 5160, 2002.
[66] M. Betke, H. Hong, and J. P. Ko, Landmark detection in the chest and
registration of lung surfaces with an application to nodule registration,
Med. Image Anal., vol. 7, pp. 265281, 2003.
[67] T. Blaffert and R. Wiemker, Comparison of different follow-up lung
registration methods with and without segmentation, Proc. SPIE, vol.
5370, pp. 17011708, 2004.
[68] T. Yamamoto, M. Kubo, Y. Kawata, N. Niki, K. Eguchi, H. Ohmatsu, R.
Kakinuma, M. Kaneko, M. Kusumoto, N. Moriyama, K. Mori, and H.
Nishiyama, Automatic slice image matching algorithm for comparative
reading using data of the pulmonary structure, Proc. SPIE, vol. 5370,
pp. 10911100, 2004.
[69] L. Dougherty, J. C. Asmuth, and W. B. Gefter, Alignment of CT lung
volumes with an optical flow method, Acad. Radiol., vol. 10, no. 3, pp.
249254, 2003.
[70] C. V. Stewart, Y. L. Lee, and C. L. Tsai, An uncertainty-driven hybrid of
intensity-based and feature-based registration with application to retinal
and lung CT images, in Lecture Notes in Computer Science. Berlin,
Germany: Springer-Verlag, 2003, vol. 3216, Medical Image Computing
and Computer-Assisted Intervention, pp. 870877.
[71] K. Woods, L. Fan, C. W. Chen, and Y. Wan, Model supported image
registration and warping for change detection in computer-aided diagnosis, in Proc. 29th Applied Imagery Pattern Recognition Workshop,
2000, pp. 180186.
[72] C. L. Novak, H. Shen, B. L. Odry, J. P. Ko, and D. P. Naidich, A system
for automatic detection of lung nodules exhibiting growth, Proc. SPIE,
vol. 5370, pp. 1021, 2004.
[73] L. Fan, C. W. Chen, J. M. Reinhardt, and E. A. Hoffman, Evaluation
and application of 3-D lung warping and registration model using HRCT
images, Proc. SPIE, vol. 4321, pp. 234243, 2001.
[74] V. Boldea, D. Sarrut, and S. Clippe, Lung deformation estimation
with nonrigid registration for radiotherapy treatment, in Lecture Notes
in Computer Science. Berlin, Germany: Springer-Verlag, 2003, vol.
2878, Medical Image Computing and Computer-Assisted Intervention,
pp. 770777.
[75] J. P. Thirion, Image matching as a diffusion process: an analogy
with Maxwells demons, in Med. Image Anal., 1998, vol. 2, pp.
243260.
[76] L. Weruaga, J. Morales, L. Nunez, and R. Verdu, Estimating volumetric
motion in human thorax with parametric matching constraints, IEEE
Trans. Med. Imag., vol. 22, no. 6, pp. 766772, Jun. 2003.
[77] M. R. Kaus, T. Netsch, S. Kabus, V. Pekar, T. McNutt, and B. Fischer,
Estimation of organ motion from 4D CT for 4D radiation therapy planning of lung cancer, in Lecture Notes in Computer Science. Berlin,
Germany: Springer-Verlag, 2004, vol. 3217, Medical Image Computing
and Computer-Assisted Intervention, pp. 10171024.
[78] B. Li, G. E. Christensen, E. A. Hoffman, G. McLennan, and J. M. Reinhardt, Establishing a normative atlas of the human lung: Intersubject
warping and registration of volumetric CT images, Acad. Radiol., vol.
10, no. 3, pp. 255265, 2003.
[79] L. Zhang and J. M. Reinhardt, 3-D pulmonary CT image registration
with a standard lung atlas, Proc. SPIE, vol. 3978, pp. 6777, 2000.
[80] R. Uppaluri, T. Mitsa, M. Sonka, E. A. Hoffman, and G. McLennan,
Quantification of pulmonary emphysema from lung computed tomography images, Am. J. Respir. Crit. Care Med., vol. 156, pp. 248254,
1997.
[81] A. Madani, C. Keyzer, and P. Gevenois, Quantitative computed tomography assessment of lung structure and function in pulmonary emphysema, Eur. Respir. J., vol. 18, no. 4, pp. 720730, 2001.
[82] H. O. Coxson, R. M. Rogers, K. P. Whittall, Y. DYachkova, P. D. Par,
F. C. Sciurba, and J. C. Hogg, A quantification of the lung surface area
in emphysema using computed tomography, Am. J. Respir. Crit. Care
Med., vol. 159, no. 3, pp. 8511073, 1999.

IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006

[83] M. Kohlhaufl, P. Brand, C. Rock, T. Radons, G. Scheuch, T. Meyer, H.

Schulz, K. J. Pfeifer, K. Haussinger, and J. Heyder, Noninvasive diagnosis of emphysema. Aerosol morphometry and aerosol bolus dispersion
in comparison to HRCT, Am. J. Respir. Crit. Care Med., vol. 160, no.
3, pp. 913918, 1999.
[84] Y. Nakano, H. Sakai, S. Muro, T. Hirai, Y. Oku, K. Nishimura, and
M. Mishima, Comparison of low attenuation areas on computed tomographic scans between inner and outer segments of the lung in patients
with chronic obstructive pulmonary disease: incidence and contribution
to lung function, Thorax, vol. 54, no. 5, pp. 384389, 1999.
[85] K. J. Park, C. J. Bergin, and J. L. Clausen, Quantitation of emphysema
with three-dimensional CT densitometry: comparison with two-dimensional analysis, visual emphysema scores, and pulmonary function test
results, Radiology, vol. 211, no. 2, pp. 541547, 1999.
[86] K. Soejima, K. Yamaguchi, E. Kohda, K. Takeshita, Y. Ito, H. Mastubara,
T. Oguma, T. Inoue, Y. Okubo, K. Amakawa, H. Tateno, and T. Shiomi, Longitudinal follow-up study of smoking-induced lung density
changes by high-resolution computed tomography, Am. J. Respir. Crit.
Care Med., vol. 161, no. 4, pp. 12641273, 2000.
[87] A. R. Hunsaker, E. P. Ingenito, J. J. Reilly, and P. Costello, Lung volume
reduction surgery for emphysema: correlation of CT and V/Q imaging
with physiologic mechanisms of improvement in lung function, Radiology, vol. 222, no. 2, pp. 491498, 2002.
[88] F. Mitsunobu, K. Ashida, Y. Hosaki, H. Tsugeno, M. Okamoto, K.
Nishida, S. Takata, T. Yokoi, M. Mishima, and Y. Tanizaki, Complexity
of terminal airspace geometry assessed by computed tomography in
asthma, Am. J. Respir. Crit. Care Med., vol. 167, no. 3, pp. 411417,
2003.
[89] R. A. Blechschmidt, R. Werthschtzky, and U. Lrcher, Automated
CT image evaluation of the lung: a morphology-based concept, IEEE
Trans. Med. Imag., vol. 20, no. 5, pp. 434442, May 2001.
[90] K. Cederlund, U. Tyln, L. Jorfeldt, and P. Aspelin, Classification of
emphysema in candidates for lung volume reduction surgery, Chest,
vol. 122, pp. 590596, 2002.
[91] W. J. Kostis, S. C. Fluture, D. F. Yankelevitz, and C. I. Henschke,
Method for analysis and display of distribution of emphysema in CT
scans, Proc. SPIE, vol. 5032, pp. 199206, 2003.
[92] W. J. Kostis, S. C. Fluture, A. O. Farooqi, D. F. Yankelevitz, and
C. I. Henschke, Volumetric assessment of emphysema on low-dose
screening CT scans, Proc. SPIE, vol. 5032, pp. 827834, 2003.
[93] N. Tanaka, T. Matsumoto, G. Miura, T. Emoto, N. Matsunaga, K. Ueda,
and D. A. Lynch, Air trapping at CT: high prevalence in asymptomatic
subjects with normal pulmonary function, Radiology, vol. 227, pp.
776785, 2003.
[94] T. Hara, A. Yamamoto, X. Zhou, S. Iwano, and S. Itoh, Automated
detection system for pulmonary emphysema on 3-D chest images, Proc.
SPIE, vol. 5370, pp. 915919, 2004.
[95] D. S. Gierada, R. D. Yusen, T. K. Pilgram, L. Crouch, R. M. Slone, K.
T. Bae, S. S. Lefrak, and J. D. Cooper, Repeatability of quantitative CT
indexes of emphysema in patients evaluated for lung volume reduction
surgery, Radiology, vol. 220, pp. 448454, 2001.
[96] N. L. Muller, C. A. Staples, R. R. Miller, and R. T. Abboud, Density
mask. An objective method to quantitate emphysema using computed
tomography, Chest, vol. 94, no. 4, pp. 782787, 1988.
[97] W. Kalender, R. Rienmuller, W. Seissler, J. Behr, M. Welke, and H.
Fichte, Measurement of pulmonary parenchymal attenuation: use of
spirometric gating with quantitative CT, Radiology, vol. 175, no. 1, pp.
265268, 1990.
[98] M. Mishima, T. Hirai, H. Itoh, Y. Nakano, H. Sakai, S. Muro, K.
Nishimura, Y. Oku, K. Chin, M. Ohi, T. Nakamura, J. H. T. Bates, A.
M. Alencar, and B. Suki, Complexity of terminal airspace geometry
assessed by lung computed tomography in normal subjects and patients
with chronic obstructive pulmonary disease, Proc. Nat. Acad. Sci.
(USA), vol. 96, no. 16, pp. 88298834, 1999.
[99] J. Guo, J. M. Reinhardt, H. Kitaoka, L. Zhang, M. Sonka, G. McLennan,
and E. A. Hoffman, Integrated system for CT-based assessment of
parenchymal lung disease, in Proc. IEEE Int. Symp. Biomedical
Imaging, 2002, pp. 871874.
[100] J. Nagao, T. Aiguchi, K. Mori, Y. Suenaga, J. I. Toriwaki, M. Mori,
and H. Natori, A CAD system for quantifying COPD based on 3-D
CT images, in Lecture Notes in Computer Science. Berlin, Germany:
Springer-Verlag, 2003, vol. 2878, Medical Image Computing and Computer-Assisted Intervention, pp. 730737.
[101] J. S. Silva, A. Silva, B. S. Santos, and J. Madeira, Detection and 3-D
representation of pulmonary air bubbles in HRCT volumes, Proc. SPIE,
vol. 5031, pp. 430439, 2003.

SLUIMER et al.: COMPUTER ANALYSIS OF COMPUTED TOMOGRAPHY SCANS OF THE LUNG: A SURVEY

[102] M. N. Prasad and A. Sowmya, Multi-level emphysema diagnosis of

HRCT lung images in an incremental framework, Proc. SPIE, vol.
5370, pp. 4250, 2004.
[103] B. C. Stoel, H. A. Vrooman, J. Stolk, and J. H. C. Reiber, Sources
of error in lung densitometry with CT, Invest. Radiol., vol. 34, pp.
303309, 1999.
[104] K. L. Boedeker, M. F. McNitt-Gray, S. R. Rogers, D. A. Truong, M.
S. Brown, D. W. Gjertson, and J. G. Goldin, Emphysema: effect of
reconstruction algorithm on CT imaging measures, Radiology, vol. 232,
no. 1, pp. 295301, 2004.
[105] D. G. Parr, B. C. Stoel, J. Stolk, P. G. Nightingale, and R. A. Stockley,
Influence of calibration on densitometric studies of emphysema progression using computed tomography, Am. J. Respir. Crit. Care Med.,
vol. 170, no. 8, pp. 883890, 2004.
[106] G. J. Kemerink, R. J. S. Lamers, B. J. Pellis, H. H. Kruize, and J. M.
A. van Engelshoven, On segmentation of lung parenchyma in quantitative computed tomography of the lung, Med. Phys., vol. 25, no. 12, pp.
24322439, 1998.
[107] G. J. Kemerink, H. H. Kruize, R. J. S. Lamers, and J. M. A. van
Engelshoven, CT lung densitometry: dependence of CT number
histograms on sample volume and consequences for scan protocol
comparability, J. Comput. Assist. Tomogr., vol. 21, no. 6, pp. 948954,
1997.
[108] S. B. Shaker, A. Dirksen, L. C. Laursen, L. T. Skovgaard, and N. H.
Holstein-Rathlou, Volume adjustment of lung density by computed tomography scans in patients with emphysema, Acta Radiologica, vol.
45, pp. 417423, 2004.
[109] W. Weder, R. Thurnheer, U. Stammberger, M. Burge, E. W. Russi, and
K. E. Bloch, Radiologic emphysema morphology is associated with
outcome after surgical lung volume reduction, Ann. Thoracic Surg., vol.
64, no. 2, pp. 313320, 1997.
[110] S. G. Armato, G. McLennan, M. F. McNitt-Gray, C. R. Meyer, D.
Yankelevitz, D. R. Aberle, C. I. Henschke, E. A. Hoffman, E. A. Kazerooni, H. MacMahon, A. P. Reeves, B. Y. Croft, and L. P. Clarke, Lung
image database consortium: developing a resource for the medical
imaging research community, Radiology, vol. 232, no. 3, pp. 739748,
2004.
[111] S. G. Armato, M. L. Giger, C. J. Moran, J. T. Blackburn, K. Doi, and
H. MacMahon, Computerized detection of pulmonary nodules on CT
scans, Radiographics, vol. 19, no. 5, pp. 13031311, 1999.
[112] J. P. Ko and M. Betke, Chest CT: automated nodule detection and assessment of change over timepreliminary experience, Radiology, vol.
218, no. 1, pp. 267273, 2001.
[113] B. Zhao, M. S. Ginsberg, R. A. Lefkowitz, L. Jiang, C. Cooper, and L.
H. Schwartz, Application of the LDM algorithm to identify small lung
nodules on low-dose MSCT scans, Proc. SPIE, vol. 5370, pp. 818823,
2004.
[114] T. Ezoe, H. Takizawa, S. Yamamoto, A. Shimizu, T. Matsumoto, Y.
Tateno, T. Iinuma, and M. Matsumoto, An automatic detection method
of lung cancers including ground glass opacities from chest X-ray CT
images, Proc. SPIE, vol. 4684, pp. 16721680, 2002.
[115] C. I. Fetita, F. Prteux, C. Beigelman-Aubry, and P. Grenier, 3-D
automated lung nodule segmentation in HRCT, in Lecture Notes in
Computer Science. Berlin, Germany: Springer-Verlag, 2003, vol.
2878, Medical Image Computing and Computer-Assisted Intervention,
pp. 626634.
[116] M. Tanino, H. Takizawa, S. Yamamoto, T. Matsumoto, Y. Tateno, and T.
Iinuma, A detection method of ground glass opacities in chest x-ray CT
images using automatic clustering techniques, Proc. SPIE, vol. 5032,
pp. 17281737, 2003.
[117] K. Awai, K. Murao, A. Ozawa, M. Komi, H. Hayakawa, S. Hori, and Y.
Nishimura, Pulmonary nodules at chest CT: effect of computer-aided
diagnosis on radiologists detection performance, Radiology, vol. 230,
no. 2, pp. 347352, 2004.
[118] Y. Lee, T. Hara, H. Fujita, S. Itoh, and T. Ishigaki, Automated detection of pulmonary nodules in helical CT images based on an improved
template-matching technique, IEEE Trans. Med. Imag., vol. 20, no. 7,
pp. 595604, Jul. 2001.
[119] A. Farag, A. El-Baz, G. G. Gimelfarb, R. Falk, and S. G. Hushek,
Automatic detection and recognition of lung abnormalities in helical
CT images using deformable templates, in Lecture Notes in Computer
Science. Berlin, Germany: Springer-Verlag, 2004, vol. 3217, Medical
Image Computing and Computer-Assisted Intervention, pp. 856864.
[120] K. Kanazawa, Y. Kawata, N. Niki, H. Satoh, H. Ohmatsu, R. Kakinuma,
M. Kaneko, N. Moriyama, and K. Eguchi, Computer-aided diagnosis
for pulmonary nodules based on helical CT images, Comput. Med.
Imag. Graphics, vol. 22, no. 2, pp. 157167, 1998.

403

[121] M. N. Gurcan, B. Sahiner, N. Petrick, H. P. Chan, E. A. Kazerooni, P.

N. Cascade, and L. Hadjiiski, Lung nodule detection on thoracic computed tomography images: preliminary evaluation of a computer-aided
diagnosis system, Med. Phys., vol. 29, no. 11, pp. 25522558, 2002.
[122] M. Kubo, K. Kubota, N. Yamada, Y. Kawata, N. Niki, K. Eguchi,
H. Ohmatsu, R. Kakinuma, M. Kaneko, M. Kusumoto, K. Mori, H.
Nishiyama, and N. Moriyama, A CAD system for lung cancer based on
low dose single-slice CT image, Proc. SPIE, vol. 4684, pp. 16721680,
2002.
[123] N. Yamada, M. Kubo, Y. Kawata, N. Niki, K. Eguchi, H. Omatsu, R.
Kakinuma, M. Kaneko, M. Kusumoto, H. Nishiyama, and N. Moriyama,
ROI extraction of chest CT images using adaptive opening filter, Proc.
SPIE, vol. 5032, pp. 869876, 2003.
[124] T. Oda, M. Kubo, Y. Kawata, N. Niki, K. Eguchi, H. Ohmatsu, R.
Kakinuma, M. Kaneko, M. Kusumoto, N. Moriyama, K. Mori, and H.
Nishiyama, A detection algorithm of lung cancer candidate nodules on
multi-slice CT images, Proc. SPIE, vol. 4684, pp. 13541361, 2002.
[125] S. Saita, T. Oda, M. Kubo, Y. Kawata, N. Niki, M. Sasagawa, H.
Ohmatsu, R. Kakinuma, M. Kaneko, M. Kusumoto, K. Eguchi, H.
Nishiyama, K. Mori, and N. Moriyama, Nodule detection algorithm
based on multi-slice CT images for lung cancer screening, Proc. SPIE,
vol. 5370, pp. 10831090, 2004.
[126] G. Q. Wei, L. Fan, and J. Qian, Automatic detection of nodules attached
to vessels in lung CT by volume projection analysis, in Lecture Notes
in Computer Science. Berlin, Germany: Springer-Verlag, 2002, vol.
2488, Medical Image Computing and Computer-Assisted Intervention,
pp. 746752.
[127] Y. Mekadaa, T. Kusanagi, Y. Hayase, K. Mori, J. I. Hasegawa, J. I. Toriwaki, M. Mori, and H. Natori, Detection of small nodules from 3-D
chest X-ray CT images based on shape features, in Proc. CARS, vol.
1256, 2003, pp. 971976.
[128] A. A. Enquobahrie, A. P. Reeves, D. F. Yankelevitz, and C. I. Henschke,
Automated detection of pulmonary nodules from whole lung helical
CT scans: performance comparison for isolated and attached nodules,
Proc. SPIE, vol. 5370, pp. 791800, 2004.
[129] R. Wiemker, P. Rogalla, A. Zwartkruis, and T. Blaffert, Computer aided
lung nodule detection on high resolution CT data, Proc. SPIE, vol.
4684, pp. 677688, 2002.
[130] S. Chang, H. Emoto, D. N. Metaxas, and L. Axel, Pulmonary
micronodule detection from 3-D chest CT, in Lecture Notes in
Computer Science. Berlin, Germany: Springer-Verlag, 2004, vol.
3217, Medical Image Computing and Computer-Assisted Intervention,
pp. 821828.
[131] Q. Li and K. Doi, New selective enhancement filter and its application
for significant improvement of nodule detection on computed tomography, Proc. SPIE, vol. 5370, pp. 19, 2004.
[132] D. S. Paik, C. F. Beaulieu, G. D. Rubin, B. Acar, R. B. J. Jeffrey, J. Yee, J.
Dey, and S. Napel, Surface normal overlap: a computer-aided detection
algorithm with application to colonic polyps and lung nodules in helical
CT, IEEE Trans. Med. Imag., vol. 23, no. 6, pp. 661675, Jun. 2004.
[133] H. Takizawa, S. Yamamoto, T. Matsumoto, Y. Tateno, T. Iinuma, and M.
Matsumoto, Recognition of lung nodules from X-ray CT images using
3-D Markov random field models, Proc. SPIE, vol. 4684, pp. 716725,
2002.
[134] G. Fukano, H. Takizawa, K. Shigemoto, S. Yamamoto, T. Matsumoto,
Y. Tateno, and T. Iinuma, Recognition method of lung nodules using
blood vessel extraction techniques and 3-D object models, Proc. SPIE,
vol. 5032, pp. 190196, 2003.
[135] S. G. Armato, M. L. Giger, and H. MacMahon, Automated detection
of lung nodules in CT scans: preliminary results, Med. Phys., vol. 28,
no. 8, pp. 15521561, 2001.
[136] H. Arimura, S. Katsuragawa, K. Suzuki, F. Li, J. Shiraishi, S. Sone, and
K. Doi, Computerized scheme for automated detection of lung nodules
in low-dose computed tomography images for lung cancer screening,
Acad. Radiol., vol. 11, no. 6, pp. 617629, 2004.
[137] S. G. Armato, F. Li, M. L. Giger, H. MacMahon, S. Sone, and K. Doi,
Lung cancer: performance of automated lung nodule detection applied
to cancers missed in a CT screening program, Radiology, vol. 225, no.
3, pp. 685692, 2002.
[138] K. Suzuki, S. G. Armato, F. Li, S. Sone, and K. Doi, Massive training
artificial neural network (MTANN) for reduction of false positives in
computerized detection of lung nodules in low-dose computed tomography, Med. Phys., vol. 30, no. 7, pp. 16021617, 2003.
[139] Z. Ge, B. Sahiner, H. P. Chan, L. M. Hadjiiski, J. Wei, N. Bogot, P. N.
Cascade, E. A. Kazerooni, and C. Zhou, Computer aided detection of
lung nodules: false positive reduction using a 3-D gradient field method,
Proc. SPIE, vol. 5370, pp. 10761082, 2004.

404

[140] Y. Lee, D. Y. Tsai, T. Hara, H. Fujita, S. Itoh, and T. Ishigaki, Improvement in automated detection of pulmonary nodules on helical X-ray CT
images, Proc. SPIE, vol. 5370, pp. 824832, 2004.
[141] M. S. Brown, M. F. McNitt-Gray, J. G. Goldin, R. D. Suh, J. W. Sayre,
and D. R. Aberle, Patient-specific models for lung nodule detection and
surveillance in CT images, IEEE Trans. Med. Imag., vol. 20, no. 12, pp.
12421250, Dec. 2001.
[142] X. Zhang, G. McLennan, E. A. Hoffman, and M. Sonka, Computerized
detection of pulmonary nodules using cellular neural networks in CT
images, Proc. SPIE, vol. 5370, pp. 3041, 2004.
[143] C. C. McCulloch, R. A. Kaucic, P. R. Mendona, D. J. Walter, and R. S.
Avila, Model-based detection of lung nodules in computed tomography
exams, Acad. Radiol., vol. 11, no. 3, pp. 258266, 2004.
[144] D. Wormanns and S. Diederich, Characterization of small pulmonary
nodules by CT, Eur. Radiol., vol. 14, pp. 13801391, 2004.
[145] S. Diederich, D. Wormanns, M. Semik, M. Thomas, H. Lenzen, N. Roos,
and W. Heindel, Screening for early lung cancer with low-dose spiral
CT: prevalence in 817 asymptomatic smokers, Radiology, vol. 222, no.
3, pp. 773781, 2002.
[146] B. B. Tan, K. R. Flaherty, E. A. Kazerooni, and M. D. Iannettoni, The
solitary pulmonary nodule, Chest, vol. 123, no. 90 010, pp. 89S96S,
2003.
[147] D. Ost and A. Fein, Evaluation and management of the solitary pulmonary nodule, Am. J. Respir. Crit. Care Med., vol. 162, no. 3, pp.
782787, 2000.
[148] D. F. Yankelevitz and C. I. Henschke, Lung cancer: small pulmonary
nodules, Radiologic Clinics of North America, vol. 38, pp. 19, 2000.
[149] J. J. Erasmus, J. E. Connolly, H. P. McAdams, and V. L. Roggli, Solitary
pulmonary nodules: part I. Morphologic evaluation for differentiation of
benign and malignant lesions, Radiographics, vol. 20, no. 1, pp. 4358,
2000.
[150] J. J. Erasmus, H. P. McAdams, and J. E. Connolly, Solitary pulmonary nodules: part II. Evaluation of the indeterminate nodule,
Radiographics, vol. 20, no. 1, pp. 5966, 2000.
[151] J. P. Ko and D. P. Naidich, Lung nodule detection and characterization with multi-slice CT, Radiologic Clin. N. Am., vol. 41, no. 3, pp.
575597, 2003.
[152] J. Gurney, Determining the likelihood of malignancy in solitary pulmonary nodules with Bayesian analysis. part I. Theory, Radiology, vol.
186, no. 2, pp. 405413, 1993.
[153] J. Gurney, D. Lyddon, and J. McKay, Determining the likelihood of
malignancy in solitary pulmonary nodules with Bayesian analysis. Part
II. Application, Radiology, vol. 186, no. 2, pp. 415422, 1993.
[154] D. Wormanns, E. Klotz, U. Dregger, F. Beyer, and W. Heindel, Diagnostic performance of different measurement methods for lung nodule
enhancement at quantitative contrast-enhanced computed tomography,
Proc. SPIE, vol. 5372, pp. 455462, 2004.
[155] M. K. Gould, G. D. Sanders, P. G. Barnett, C. E. Rydzak, C. C. Maclean,
M. B. McClellan, and D. K. Owens, Cost-effectiveness of alternative
management strategies for patients with solitary pulmonary nodules,
Ann. Intern. Med., vol. 138, no. 9, pp. 724735, 2003.
[156] C. I. Henschke, D. F. Yankelevitz, I. Mateescu, D. W. Brettle, T. G.
Rainey, and F. S. Weingard, Neural networks for the analysis of small
pulmonary nodules, Clin. Imag., vol. 21, no. 6, pp. 390399, 1997.
[157] M. F. McNitt-Gray, E. M. Hart, N. Wyckoff, J. W. Sayre, J. G. Goldin,
and D. R. Aberle, A pattern classification approach to characterizing
solitary pulmonary nodules imaged on high resolution CT: preliminary
results, Med. Phys., vol. 26, no. 6, pp. 880888, 1999.
[158] M. F. McNitt-Gray, N. Wyckoff, J. W. Sayre, J. G. Goldin, and D. R.
Aberle, The effects of co-occurrence matrix based texture parameters
on the classification of solitary pulmonary nodules imaged on computed
tomography, Comput. Med. Imag. Graphics, vol. 23, no. 6, pp. 339348,
1999.
[159] Y. Kawata, N. Niki, H. Ohmatsu, M. Kusumoto, R. Kakinuma, K.
Mori, H. Nishiyama, K. Eguchi, M. Kaneko, and N. Moriyama, Computer-aided diagnosis of pulmonary nodules using three-dimensional
thoracic CT images, Lecture Notes in Computer Science, vol. 2208,
pp. 13931394, 2001.
[160] Y. Kawata, N. Niki, H. Ohamatsu, M. Kusumoto, R. Kakinuma, K. Mori,
K. Yamada, H. Nishiyama, K. Eguchi, M. Kaneko, and N. Moriyama,
Pulmonary nodule classification based on CT density distribution using
3-D thoracic CT images, Proc. SPIE, vol. 5369, pp. 413422, 2004.
[161] Y. Kawata, N. Niki, H. Ohmatsu, M. Kusumoto, R. Kakinuma, K. Mori,
H. Nishiyama, K. Eguchi, M. Kaneko, and N. Moriyama, Searching
similar images for classification of pulmonary nodules in three-dimensional CT images, in Proc. IEEE Int. Symp. Biomedical Imaging, 2002,
pp. 189193.

IEEE TRANSACTIONS ON MEDICAL IMAGING, VOL. 25, NO. 4, APRIL 2006

[162] Y. Kawata, N. Niki, H. Ohmatsu, M. Kusumoto, R. Kakinuma, K. Yamada, K. Mori, H. Nishiyama, K. Eguchi, M. Kaneko, and N. Moriyama,
Pulmonary nodule classification based on nodule retrieval from 3-D
thoracic CT image database, Lecture Notes in Computer Science, vol.
3217, pp. 838846, 2004.
[163] Y. Matsuki, K. Nakamura, H. Watanabe, T. Aoki, H. Nakata, S. Katsuragawa, and K. Doi, Usefulness of an artificial neural network for
differentiating benign from malignant pulmonary nodules on high-resolution CT: evaluation with receiver operating characteristic analysis,
Am. J. Roentgenol., vol. 178, no. 3, pp. 657663, 2002.
[164] S. C. B. Lo, M. T. F. Li-Yueh Hsu, Y. M. F. Lure, and H. Zhao, Classification of lung nodules in diagnostic CT: an approach based on 3-D vascular features, nodule density distributions, and shape features, Proc.
SPIE, vol. 5032, pp. 183189, 2003.
[165] Q. Li, F. Li, J. Shiraishi, S. Katsuragawa, S. Sone, and K. Doi, Investigation of new psychophysical measures for evaluation of similar images
on thoracic computed tomography for distinction between benign and
malignant nodules, Med. Phys., vol. 30, no. 10, pp. 25842593, 2003.
[166] S. Armato and H. MacMahon, Automated lung segmentation and computer-aided diagnosis for thoracic CT scans, in Proc. CARS, vol. 1256,
2003, pp. 977982.
[167] M. Aoyama, Q. Li, S. Katsuragawa, F. Li, S. Sone, and K. Doi, Computerized scheme for determination of the likelihood measure of malignancy for pulmonary nodules on low-dose CT images, Med. Phys., vol.
30, no. 3, pp. 387394, 2003.
[168] S. Takashima, S. Sone, F. Li, Y. Maruyama, M. Hasegawa, T. Matsushita,
F. Takayama, and M. Kadoya, Small solitary pulmonary nodules ( 1
cm) detected at population-based CT screening for lung cancer: reliable
high-resolution CT features of benign lesions, Am. J. Roentgenol., vol.
180, no. 4, pp. 955964, 2003.
[169] S. Takashima, S. Sone, F. Li, Y. Maruyama, M. Hasegawa, and M.
Kadoya, Indeterminate solitary pulmonary nodules revealed at population-based CT screening of the lung: using first follow-up diagnostic
CT to differentiate benign and malignant lesions, Am. J. Roentgenol.,
vol. 180, no. 5, pp. 12551263, 2003.
[170] P. F. Judy, R. Riva, Y. Kadota, and F. L. Jacobson, Calcification classifications of small nodules identified during CT lung-cancer screening,
Proc. SPIE, vol. 5370, pp. 330333, 2004.
[171] J. Burns, L. B. Haramati, K. Whitney, and M. N. Zelefsky, Consistency of reporting basic characteristics of lung nodules and masses on
computed tomography, Acad. Radiol., vol. 11, pp. 233237, 2004.
[172] Y. Yuan, T. Matsumoto, A. Hiyama, G. Miura, N. Tanaka, T. Emoto, T.
Kawamura, and N. Matsunaga, The probability of malignancy in small
pulmonary nodules coexisting with potentially operable lung cancer detected by CT, Eur. Radiol., vol. 13, no. 11, pp. 24472453, 2003.
[173] J. P. Ko, E. A. Drucker, J. A. O. Shepard, C. F. Mountain, C. Dresler, B.
Sabloff, and T. C. McLoud, CT depiction of regional nodal stations for
lung cancer staging, Am. J. Roentgenol., vol. 174, no. 3, pp. 775782,
2000.
[174] D. Wormanns, S. Diederich, M. G. Lentschig, F. Winter, and W. Heindel,
Spiral CT of pulmonary nodules: interobserver variation in assessment
of lesion size, Eur. Radiol., vol. 10, pp. 710713, 2000.
[175] H. T. Winer-Muram, S. G. Jennings, C. A. Meyer, Y. Liang, A. M. Aisen,
R. D. Tarver, and R. C. McGarry, Effect of varying CT section width
on volumetric measurement of lung tumors and application of compensatory equations, Radiology, vol. 229, no. 1, pp. 184194, 2003.
[176] J. Hsieh and K. Karau, Theoretical prediction of lung nodule measurement accuracy under different acquisition and reconstruction conditions, Proc. SPIE, vol. 5369, pp. 406412, 2004.
[177] J. P. Ko, H. Rusinek, E. L. Jacobs, J. S. Babb, M. Betke, G. McGuinness,
and D. P. Naidich, Small pulmonary nodules: volume measurement at
chest CTphantom study, Radiology, vol. 228, no. 3, pp. 864870,
2003.
[178] L. Fan, J. Qian, B. Odry, H. Shen, D. Naidich, G. Kohl, and E. Klotz,
Automatic segmentation of pulmonary nodules by using dynamic 3-D
cross-correlation for interactive CAD systems, Proc. SPIE (Medical
Imaging), vol. 4684, pp. 13621369, 2002.
[179] D. Wormanns, G. Kohl, E. Klotz, A. Marheine, F. Beyer, W. Heindel,
and S. Diederich, Volumetric measurements of pulmonary nodules at
multi-row detector CT: in vivo reproducibility, Eur. Radiol., vol. 14, no.
1, pp. 8692, 2004.
[180] J. M. Kuhnigk, V. Dicken, L. Bornemann, D. Wormanns, S. Krass,
and H. O. Peitgen, Fast automated segmentation and reproducible
volumetry of pulmonary metastases in CT-scans for therapy monitoring, in Lecture Notes in Computer Science. Berlin, Germany:
Springer-Verlag, 2004, vol. 3217, Medical Image Computing and
Computer-Assisted Intervention, pp. 933941.

SLUIMER et al.: COMPUTER ANALYSIS OF COMPUTED TOMOGRAPHY SCANS OF THE LUNG: A SURVEY

[181] D. Wormanns, G. Kohl, E. Klotz, W. Heindel, and S. Diederich, Clinical

evaluation of the reproducibility of volume measurements of pulmonary
nodules, Proc. SPIE, vol. 4684, pp. 316322, 2002.
[182] W. Mullally, M. Betke, and J. Wang, Segmentation of nodules on chest
computed tomography for growth assessment, Med. Phys., vol. 31, pp.
839848, 2004.
[183] W. Kostis, A. Reeves, D. Yankelevitz, and C. Henschke, Three-dimensional segmentation and growth rate estimation of small pulmonary nodules in helical CT images, IEEE Trans. Med. Imag., vol. 22, no. 10, pp.
12591274, 2003.
[184] E. Bomsztyk and H. Rusinek, Volumetric analysis of lung nodules
using a hybrid algorithm, Proc. SPIE, vol. 5370, pp. 2229, 2004.
[185] N. Xu, N. Ahuja, and R. Bansal, Automated lung nodule segmentation
using dynamic programming and EM based classification, Proc. SPIE,
vol. 4684, pp. 666676, 2002.
[186] B. Zhao, D. Yankelevitz, A. Reeves, and C. Henschke, Two-dimensional multi-criterion segmentation of pulmonary nodules on helical CT
images, Med. Phys., vol. 26, no. 6, pp. 889895, 1999.
[187] H. Shen, B. Goebel, and B. Odry, A new algorithm for local surface
smoothing with application to chest wall nodule segmentation in lung
CT data, Proc. SPIE, vol. 5370, pp. 15191526, 2004.
[188] C. I. Fetita, F. Prteux, C. Beigelman-Aubry, and P. Grenier, 3-D automated lung nodule segmentation in HRCT, Lecture Notes in Computer
Science, vol. 2878, pp. 626634, 2004.
[189] K. Okada, D. Comaniciu, and A. Krishnan, Robust 3-D segmentation
of pulmonary nodules in multislice CT images, in Lecture Notes in
Computer Science. Berlin, Germany: Springer-Verlag, 2004, vol.
3217, MICCAI, pp. 881889.
[190] L. Zhang, M. Fang, D. P. Naidich, and C. L. Novak, Consistent interactive segmentation of pulmonary ground glass nodules identified in CT
studies, Proc. SPIE, vol. 5370, pp. 17091719, 2004.
[191] Y. Kawata, N. Niki, H. Ohmatsu, M. Kusumoto, R. Kakinuma, K. Mori,
H. Nishiyama, K. Eguchi, M. Kaneko, and N. Moriyama, Analysis of
pulmonary nodule evolutions using a sequence of three-dimensional thoracic CT images, in Lecture Notes in Computer Science, ser. . Berlin,
Germany: Springer-Verlag, 2001, vol. 2208, Medical Image Computing
and Computer-Assisted Intervention, pp. 103110.
[192] P. Herzog, J. E. Wildberger, M. Niethammer, S. Schaller, and J. Schoepf,
CT perfusion imaging of the lung in pulmonary embolism, Acad. Radiol., vol. 10, pp. 11321146, 2003.
[193] S. D. Qanadli, M. El Hajjam, A. Vieillard-Baron, T. Joseph, B.
Mesurolle, V. L. Oliva, O. Barre, F. Bruckert, O. Dubourg, and P.
Lacombe, New CT index to quantify arterial obstruction in pulmonary
embolism: comparison with angiographic index and echocardiography,
Am. J. Roentgenol., vol. 176, no. 6, pp. 14151420, 2001.
[194] W. Webb, N. Mller, and D. Naidich, High Resolution CT of the Lung,
3rd ed. Philadelphia, PA: Lippincott Williams & Wilkins, 2001.
[195] G. G. King, N. L. Mller, and P. D. Par, Evaluation of airways in
obstructive pulmonary disease using high-resolution computed tomography, Am. J. Respir. Crit. Care Med., vol. 159, no. 3, pp. 9921004,
1999.
[196] M. N. Prasad and A. Sowmya, Detection of bronchovascular pairs
on HRCT lung images through relational learning, in Proc. IEEE Int.
Symp. Biomedical Imaging, 2004, pp. 11351138.
[197] F. Chabat, X. P. Hu, D. M. Hansell, and G. Z. Yang, ERS transform for
the automated detection of bronchial abnormalities on CT of the lungs,
IEEE Trans. Med. Imag., vol. 20, no. 9, pp. 942952, Sep. 2001.
[198] G. G. King, N. L. Mller, K. P. Whittall, Q. S. Xiang, and P. D. Par,
An analysis algorithm for measuring airway lumen and wall areas from
high-resolution computed tomographic data, Am. J. Respir. Crit. Care
Med., vol. 161, pp. 574580, 2000.
[199] O. I. Saba, E. A. Hoffman, and J. M. Reinhardt, Maximizing quantitative
accuracy of lung airway lumen and wall measures obtained from x-ray CT
imaging, J. Appl. Physiol., vol. 95, no. 3, pp. 10631075, 2003.
[200] BTS guidelines on the diagnosis, assessment and treatment of diffuse parenchymal lung disease in adults, Thorax, vol. 54, pp. S24S30,
1999.

405

[201] S. Delorme, M. A. Keller-Reichenbecher, I. Zuna, W. Schlegel, and G.

Van Kaick, Usual interstitial pneumonia: quantitative assessment of
high-resolution computed tomography findings by computer-assisted
texture-based image analysis, Invest. Radiol., vol. 32, no. 9, pp.
566574, 1997.
[202] R. Uppaluri, E. Hoffman, M. Sonka, G. Hunninghake, and G.
McLennan, Interstitial lung disease. a quantitative study using the
adaptive multiple feature method, Am. J. Respir. Crit. Care Med., vol.
159, no. 2, pp. 5191073, 1999.
[203] R. Uppaluri, E. A. Hoffman, M. Sonka, P. G. Hartley, G. W. Hunninghake, and G. McLennan, Computer recognition of regional lung disease
patterns, Am. J. Respir. Crit. Care Med., vol. 160, no. 2, pp. 648654,
1999.
[204] H. U. Kauczor, K. Heitmann, C. P. Heussel, D. Marwede, T. Uthmann,
and M. Thelen, Automatic detection and quantification of ground-glass
opacities on high-resolution CT using multiple neural networks: comparison with a density mask, Am. J. Roentgenol., vol. 175, no. 5, pp.
13291334, 2000.
[205] Y. Uchiyama, S. Katsuragawa, H. Abe, J. Shiraishi, F. Li, Q. Li, C.
T. Zhang, K. Suzuki, and K. Doi, Quantitative computerized analysis
of diffuse lung disease in high-resolution computed tomography, Med.
Phys., vol. 30, no. 9, pp. 24402454, 2003.
[206] A. Shamsheyeva and A. Sowmya, Tuning kernel function parameters
of support vector machines for segmentation of lung disease patterns in
high-resolution computed tomography images, Proc. SPIE, vol. 5370,
pp. 15481557, 2004.
[207] B. Zheng, J. K. Leader, C. R. Fuhrman, F. C. Sciurba, and D. Gur, Automated detection and classification of interstitial lung diseases from low
dose {CT} images, Proc. SPIE, vol. 5370, pp. 849856, 2004.
[208] I. Sluimer, P. F. van Waes, M. A. Viergever, and B. van Ginneken, Computer-aided diagnosis in high resolution CT of the lungs, Med. Phys.,
vol. 30, no. 12, pp. 30813090, 2003.
[209] C. R. Shyu, C. Brodley, A. Kak, and A. Kosaka, ASSERT: a physician-in-the-loop content-based retrieval system for HRCT image
databases, Comput. Vis. Image Understanding, vol. 75, no. 1/2, pp.
111132, 1999.
[210] C. R. Shyu, C. Pavlopoulou, A. C. Kak, and C. E. Brodley, Using
human perceptual categories for content-based retrieval from a medical image database, Comput. Vis. Image Understanding, vol. 88, pp.
119151, 2002.
[211] S. D. MacArthur, C. E. Brodley, A. C. Kak, and L. S. Broderick, Interactive content-based image retrieval using relevance feedback, Comput.
Vis. Image Understanding, vol. 88, pp. 5575, 2002.
[212] J. G. Dy, C. E. Brodley, A. Kak, L. S. Broderick, and A. M. Aisen, Unsupervised feature selection applied to content-based retrieval of lung
images, IEEE Trans. Pattern Anal. Mach. Intell., vol. 25, no. 3, pp.
373378, Mar. 2003.
[213] F. Chabat, G. Z. Yang, and D. M. Hansell, Obstructive lung diseases:
texture classification for differentiation at CT, Radiology, vol. 228, pp.
871877, 2003.
[214] A. Fukushima, K. Ashizawa, T. Yamaguchi, N. Matsuyama, H. Hayashi,
I. Kida, Y. Imafuku, A. Egawa, S. Kimura, K. Nagaoki, S. Honda, S. Katsuragawa, K. Doi, and K. Hayashi, Application of an artificial neural
network to high-resolution CT: usefulness in differential diagnosis of
diffuse lung disease, Am. J. Roentgenol., vol. 183, no. 2, pp. 297305,
2004.
[215] I. Wolf, A. Eid, M. Vetter, P. Hassenpflug, and H. P. Meinzer, Extension of 2-D segmentation methods into 3-D by means of Coons-patch
interpolation, Proc. SPIE, vol. 5032, pp. 14131420, 2003.
[216] Y. Kang, K. Engelke, and W. A. Kalender, Interactive 3-D editing tools
for image segmentation, Med. Image Anal., vol. 8, no. 1, pp. 3546,
2004.
[217] E. A. Hoffman, A. V. Clough, G. E. Christensen, C. long Lin, G.
McLennan, J. M. Reinhardt, B. A. Simon, M. Sonka, M. H. Tawhai,
E. J. van Beek, and G. Wang, The comprehensive imaging-based
analysis of the lung: a forum for team science, Acad. Radiol., vol. 11,
pp. 13701380, 2004.