Nephropathy Prediction Using Deep Learning Models

ABSTRACT
Kidney cancer is a significant global health issue, requiring advanced diagnostic
methods for early and accurate detection. Our project introduces an innovative
approach to detecting kidney cancer, using cutting-edge deep learning models,
Python programming, and a user-friendly web interface developed with Flask.
We start by analyzing CT scan images, employing the state-of-the-art
MobileNet architecture to create a model that excels in identifying kidney tumor
tissues. With a remarkable training and validation accuracy of 99.00%, our
model instills confidence in pinpointing cancerous growths. Our CT scan
dataset includes over 7,000 images, ensuring thorough training and testing of
our model. Additionally, we incorporate a complementary method using blood
test samples. Leveraging an Artificial Neural Network (ANN) model, we
achieve high accuracy in detecting kidney cancer, highlighting the effectiveness
of integrating imaging and biochemical data. The blood test dataset features 400
individual records with various clinical markers, enabling us to leverage diverse
patient information for more accurate diagnoses. Our user-friendly web
interface streamlines the process, allowing healthcare professionals and patients
to input data easily and receive rapid results. In summary, our project offers a
groundbreaking approach to kidney cancer detection, utilizing Python-based
deep learning models and an accessible web interface. By combining CT scan
analysis and blood test data, we provide a comprehensive diagnostic solution
with the potential to transform kidney cancer diagnosis and treatment.

TABLE OF CONTENTS

Department Of CSE(Artificial Intelligence And Machine Learning) 1

 Nephropathy Prediction Using Deep Learning Models

CHAPTER CONTENT PAGE NO.

NO.
1 INTRODUCTION 3
1.1 MOTIVATION 4
2 LITERATURE SURVEY 5
2.1 PROBLEMS WITH EXISTING 10
SYSTEM

3 PROBLEM STATEMENT 11
3.1 OBJECTIVES 12
4 SYSTEM REQUIREMENTS 13
4.1 SOFTWARE REQUIREMENTS
4.2 HARDWARE REQUIREMENTS
5 METHODOLOGY 14
6 RESULT AND SNAPSHOTS 16
7 APPLICATIONS 27
8 REFERENCES 28
CONCLUSION 30

CHAPTER 1
INTRODUCTION
Introduction to Nephropathy prediction using ANN:

Department Of CSE(Artificial Intelligence And Machine Learning) 2

 Nephropathy Prediction Using Deep Learning Models

The evolving landscape of kidney cancer detection and management is

witnessing a paradigm shift driven by advancements in medical imaging, data
analytics, and personalized medicine. Conventional diagnostic methods, such as
CT scans and MRIs, though valuable, have limitations in capturing the intricate
nuances of the disease. The integration of imaging data with structured clinical
information from blood samples and patient records represents a ground
breaking approach. This convergence aims to construct predictive models for
early-stage detection and prognosis, offering a more precise understanding of
kidney cancer pathophysiology. The incorporation of genomic and molecular
profiling further refines this approach, allowing for tailored treatments based on
specific genetic mutations and molecular markers.

Machine learning algorithms play a pivotal role in this transformative

landscape, enabling the analysis of vast datasets encompassing imaging,
clinical, and molecular information. These algorithms contribute to the
development of sophisticated predictive models, enhancing diagnostic accuracy
and paving the way for personalized treatment strategies. Telemedicine and
remote monitoring technologies emerge as key facilitators, ensuring broader
access to specialized care, especially in remote or underserved areas.
Additionally, the integration of patient-reported outcomes, advancements in
immunotherapy, and collaborative research initiatives underscore the holistic
and patient-centered nature of this approach.

1.1 MOTIVATION

Motivated by the alarming statistics surrounding kidney cancer and the critical
need for early detection methods, this project seeks to harness the power of deep
learning models. With kidney cancer becoming increasingly prevalent,
especially among the aging population, there is a pressing need for innovative
approaches to improve diagnosis accuracy and timeliness. The chapter delves

Department Of CSE(Artificial Intelligence And Machine Learning) 3

 Nephropathy Prediction Using Deep Learning Models

into the significance of Convolutional Neural Networks (CNNs) in medical

image analysis and Artificial Neural Networks (ANNs) in handling blood
sample data. The effectiveness of deep learning models, such as CNNs, for
tumour segmentation and classification is discussed. By emphasizing the
importance of suitable loss functions in model performance, the chapter sets the
stage for the proposed research's focus on kidney cancer classification using a
combination of CT scan images and blood sample test values. The chapter
concludes with an overview of the project's dual-model approach and the
development of a web application for user-friendly kidney cancer detection,
ultimately contributing to improved patient outcomes and healthcare efficacy.

CHAPTER 2

LITERATURE SURVEY
1) Deep learning for end-to-end kidney cancer diagnosis on multi-phase
abdominal computed tomography

AUTHORS: Kwang-Hyun Uhm, Seung-Won Jung, Moon Hyung Choi, Hong-

Kyu Shin, Jae-Ik Yoo, Se Won Oh, Jee Young Kim, Hyun Gi Kim, Young Joon
Lee, Seo Yeon Youn, Sung-Hoo Hong, Sung-Jea Ko

Department Of CSE(Artificial Intelligence And Machine Learning) 4

 Nephropathy Prediction Using Deep Learning Models

In 2020, it is estimated that 73,750 kidney cancer cases were diagnosed, and 14,830 people
died from cancer in the United States. Preoperative multi-phase abdominal computed
tomography (CT) is often used for detecting lesions and classifying histologic subtypes of
renal tumor to avoid unnecessary biopsy or surgery. However, there exists inter-observer
variability due to subtle differences in the imaging features of tumor subtypes, which makes
decisions on treatment challenging. While deep learning has been recently applied to the
automated diagnosis of renal tumor, classification of a wide range of subtype classes has not
been sufficiently studied yet. In this paper, we propose an end-to-end deep learning model for
the differential diagnosis of five major histologic subtypes of renal tumors including both
benign and malignant tumors on multi-phase CT. Our model is a unified framework to
simultaneously identify lesions and classify subtypes for the diagnosis without manual
intervention. We trained and tested the model using CT data from 308 patients who
underwent nephrectomy for renal tumors. The model achieved an area under the curve (AUC)
of 0.889, and outperformed radiologists for most subtypes. We further validated the model on
an independent dataset of 184 patients from The Cancer Imaging Archive (TCIA). The AUC
for this dataset was 0.855, and the model performed comparably to the radiologists. These
results indicate that our model can achieve similar or better diagnostic performance than
radiologists in differentiating a wide range of renal tumors on multi-phase CT.

2) Radiology Imaging Scans for Early Diagnosis of Kidney Tumors: A Review

of Data Analytics-Based Machine Learning and Deep Learning Approaches

AUTHORS: Maha gharaibeh, Dalia Alz’bi, Malak Abdullah, Ismail Hmeidi,

Mohammad Rustom AI Nasar

Plenty of disease types exist in world communities that can be explained by humans’
lifestyles or the economic, social, genetic, and other factors of the country of residence.
Recently, most research has focused on studying common diseases in the population to
reduce death risks, take the best procedure for treatment, and enhance the healthcare level of
the communities. Kidney Disease is one of the common diseases that have affected our
societies. Sectionicularly Kidney Tumors (KT) are the 10th most prevalent tumor for men

Department Of CSE(Artificial Intelligence And Machine Learning) 5

 Nephropathy Prediction Using Deep Learning Models

and women worldwide. Overall, the lifetime likelihood of developing a kidney tumor for
males is about 1 in 466 (2.02 percent) and it is around 1 in 80 (1.03 percent) for females.
Still, more research is needed on new diagnostic, early, and innovative methods regarding
finding an appropriate treatment method for KT. Compared to the tedious and time-
consuming traditional diagnosis, automatic detection algorithms of machine learning can save
diagnosis time, improve test accuracy, and reduce costs. Previous studies have shown that
deep learning can play a role in dealing with complex tasks, diagnosis and segmentation, and
classification of Kidney Tumors, one of the most malignant tumors. The goals of this review
article on deep learning in radiology imaging are to summarize what has already been
accomplished, determine the techniques used by the researchers in previous years in
diagnosing Kidney Tumors through medical imaging, and identify some promising future
avenues, whether in terms of applications or technological developments, as well as
identifying common problems, describing ways to expand the data set, summarizing the
knowledge and best practices, and determining remaining challenges and future directions.

3) Kidney segmentation from computed tomography images using deep neural

network

AUTHORS: Luana Bat ista da Cruz, José Denes Lima Araújo, Jonnison Lima
Ferreira, Joao Otavio Bandeira Diniz, Aristofanes Correa Silva, João Dallyson
Sousa de Almeida, Anselmo Cardoso de Paiva, Marcelo GATT ass

The precise segmentation of kidneys and kidney tumors can help medical specialists to
diagnose diseases and improve treatment planning, which is highly required in clinical
practice. Manual segmentation of the kidneys is extremely time-consuming and prone to
variability between different specialists due to their heterogeneity. Because of this hard work,
computational techniques, such as deep convolutional neural networks, have become popular
in kidney segmentation tasks to assist in the early diagnosis of kidney tumors. In this study,
we propose an automatic method to delimit the kidneys in computed tomography (CT)

Department Of CSE(Artificial Intelligence And Machine Learning) 6

 Nephropathy Prediction Using Deep Learning Models

images using image processing techniques and deep convolutional neural networks (CNNs)
to minimize false positives.

Methods:
The proposed method has four main steps: (1) acquisition of the KiTS19 dataset, (2) scope
reduction using Alex Net, (3) initial segmentation using U-Net 2D, and (4) false positive
reduction using image processing to maintain the largest elements (kidneys).
Results:
The proposed method was evaluated in 210 CTs from the KiTS19 database and obtained the
best result with an average Dice coefficient of 96.33%, an average Jacquard index of 93.02%,
an average sensitivity of 97.42%, an average specificity of 99.94% and an average accuracy
of 99.92%. In the KiTS19 challenge, it presented an average Dice coefficient of 93.03%.
Conclusion:
In our method, we demonstrated that the kidney segmentation problem in CT can be solved
efficiently using deep neural networks to define the scope of the problem and segment the
kidneys with high precision and with the use of image processing techniques to reduce false
positives.

4) The classification of renal cancer in 3- phase CT images using a deep

learning method

AUTHORS: Seokmin Han, Sung Il Hwang, Hak Jong Lee

In this research, we exploit an image-based deep learning framework to distinguish three

major subtypes of renal cell carcinoma (clear cell, papillary, and chromophore) using images
acquired with computed tomography (CT). A biopsy-proven benchmarking dataset was built
from 169 renal cancer cases. In each case, images were acquired at three phases (phase 1,
before injection of the contrast agent; phase 2, 1 min after the injection; phase 3, 5 min after
the injection). After image acquisition, rectangular ROI (region of interest) in each phase
image was marked by radiologists. After cropping the ROIs, a combination weight was
multiplied to the three-phase ROI images and the linearly combined images were fed into a

Department Of CSE(Artificial Intelligence And Machine Learning) 7

 Nephropathy Prediction Using Deep Learning Models

deep learning neural network after concatenation. A deep learning neural network was trained
to classify the subtypes of renal cell carcinoma, using the drawn ROIs as inputs and the
biopsy results as labels. The network showed about 0.85 accuracy, 0.64-0.98 sensitivity,
0.83-0.93 specificity, and 0.9 AUC. The proposed framework which is based on deep
learning method and ROIs provided by radiologists showed promising results in renal cell
subtype classification. We hope it will help future research on this subject and it can
cooperate with radiologists in classifying the subtype of lesion in real clinical situation.

5) An optimized SVM-based possibilistic fuzzy c-means clustering algorithm

for tumor segmentation

AUTHORS: Duggirala SR, Kollem S, Rama Linga Reddy K

To design an efficient partial differential equation-based total variation method for denoising
and possibilistic fuzzy c-means clustering algorithm for segmentation and these methods
presented the more detailed information of the MRI medical images compared to traditional
methods. In this article, the pipeline of the proposed method described by two modules like
pre-processing and segmentation. In pre-processing, noisy image is decomposed using no
subsampled contourlet transform and it contains highpass contourlet coefficient (i.e., noisy
coefficient) is removed by the threshold method as well. After reconstruction, the primary
denoised image is enhanced by an improved partial differential equation-based total variation
method in terms of image details like edges, boundaries, etc. In segmentation, the enhanced
primary denoised image is segmented by an improved possibilistic fuzzy c-means clustering
algorithm that avoids limitations in possibilistic c-means, fuzzy c-means, and K-means
clustering. Next, a support vector machine classifier is utilized to identify brain tissues into
gray matter, white matter, cerebrospinal fluid, and tumor part. The parameters were optimally
selected by a grey wolf optimization algorithm for the classification of brain tissues. The

Department Of CSE(Artificial Intelligence And Machine Learning) 8

 Nephropathy Prediction Using Deep Learning Models

performance of the proposed method is computed with reference to peak signal-to-noise ratio,
mean square error, structural similarity index, sensitivity, specificity, and accuracy. The
experimental results claimed that the proposed method is better than the traditional methods.

2.1 PROBLEMS WITH EXISTING SYSTEM

 Limited Accuracy: CNN-based systems struggle with intricate patterns in

medical data, leading to potential false diagnoses and adverse patient
outcomes.
 Data Size and Diversity: Insufficient and non-diverse datasets hinder
CNNs' generalization, impacting their effectiveness in handling variations
in patient demographics and imaging conditions.
 Model Complexity: CNNs require intricate architectures and tuning,
posing challenges in training, deployment, and maintenance, making
them less straightforward solutions.
 Resource Intensive: High computational demands make CNNs costly,
limiting accessibility in smaller healthcare facilities or regions with
limited resources.
 Interpretability: CNNs' "black-box" nature in decision-making processes
can be a disadvantage in medical contexts, where understanding
diagnoses is crucial for trust and informed decisions.
 Dependency on Imaging Quality: CNN performance is sensitive to data
quality; lower-quality medical images may degrade diagnostic accuracy.

Department Of CSE(Artificial Intelligence And Machine Learning) 9

 Nephropathy Prediction Using Deep Learning Models

 Scalability Challenges: Adapting CNN systems for broader disease

detection or different modalities may require significant redevelopment
efforts.
 Regulatory Compliance: Meeting stringent medical regulatory standards
adds complexity to the system's development and deployment.

CHAPTER 3
PROBLEM STATEMENT
Kidney cancer poses a significant global health challenge, necessitating early
detection for improved patient outcomes. Existing diagnostic methods may lack
efficiency or sensitivity, leading to delayed diagnoses and treatment initiation.
This project addresses the need for a more effective and timely detection system
by harnessing the power of deep learning models. The challenge is to develop a
comprehensive solution that integrates CT scan image analysis using MobileNet
and blood test sample analysis employing an Artificial Neural Network (ANN).
The aim is to achieve high accuracy in both modalities, leveraging a diverse
dataset of 5,077 CT scan images and 400 blood test records. The project also
aims to provide a user-friendly interface through Flask, ensuring accessibility
for healthcare professionals and users alike. This research lays the groundwork
for an innovative kidney cancer detection system capable of early diagnosis,
ultimately enhancing patient outcomes and addressing a critical aspect of global
healthcare.

Department Of CSE(Artificial Intelligence And Machine Learning) 10

 Nephropathy Prediction Using Deep Learning Models

3.1 OBJECTIVES

1. Input Design involves the transformation of user-oriented input descriptions

into a computer-based system, playing a crucial role in preventing data input
errors and guiding management towards obtaining accurate information from
the computerized system.

2. Achieving this objective involves the creation of user-friendly screens

tailored for efficient data entry, especially when dealing with large volumes of
information. The primary goal is to design inputs that facilitate error-free data
entry, ensuring that data manipulation tasks are easily performed. Additionally,
these screens offer convenient facilities for viewing records.

3. The design of data entry screens ensures that entered data undergoes
validation checks to ensure its accuracy. By providing user-friendly screens, the
system allows for the easy and error-free input of data, with appropriate
messages provided to guide users and prevent confusion. Ultimately, the
objective of input design is to establish a layout that is not only easy to follow
but also enhances the overall data input process.

Department Of CSE(Artificial Intelligence And Machine Learning) 11

 Nephropathy Prediction Using Deep Learning Models

CHAPTER 4
SYSTEM REQUIREMENTS

4.1 SOFTWARE REQUIREMENTS:

 Operating system : Windows 10 Pro.

 Coding Language : Python.
 Web Framework : Flask.
 Frontend : HTML, CSS, JavaScript.

4.2 HARDWARE REQUIREMENTS:

 System : Pentium i3 Processor.

 Hard Disk : 500 GB.
 Monitor : 15’’ LED
 Input Devices : Keyboard, Mouse
 Ram : 8 GB

Department Of CSE(Artificial Intelligence And Machine Learning) 12

 Nephropathy Prediction Using Deep Learning Models

CHAPTER 5

METHODOLOGY

5.1 FLOWCHART

Department Of CSE(Artificial Intelligence And Machine Learning) 13

 Nephropathy Prediction Using Deep Learning Models

CHAPTER 6

RESULT AND SNAPSHOTS

Department Of CSE(Artificial Intelligence And Machine Learning) 14

 Nephropathy Prediction Using Deep Learning Models

6.1 Homepage

6.2 Admin Login

Department Of CSE(Artificial Intelligence And Machine Learning) 15

 Nephropathy Prediction Using Deep Learning Models

6.3 Upload Dataset

6.4 Dataset Uploaded

Department Of CSE(Artificial Intelligence And Machine Learning) 16

 Nephropathy Prediction Using Deep Learning Models

6.5 Train/Test Dataset

Department Of CSE(Artificial Intelligence And Machine Learning) 17

 Nephropathy Prediction Using Deep Learning Models

6.6 Attributes

Department Of CSE(Artificial Intelligence And Machine Learning) 18

 Nephropathy Prediction Using Deep Learning Models

6.7 Prediction is Normal

6.8 Prediction is Abnormal

Department Of CSE(Artificial Intelligence And Machine Learning) 19

 Nephropathy Prediction Using Deep Learning Models

6.9 Confusion Matrix

Department Of CSE(Artificial Intelligence And Machine Learning) 20

 Nephropathy Prediction Using Deep Learning Models

6.10 Detection Through Image

Department Of CSE(Artificial Intelligence And Machine Learning) 21

 Nephropathy Prediction Using Deep Learning Models

6.11 Normal

Department Of CSE(Artificial Intelligence And Machine Learning) 22

 Nephropathy Prediction Using Deep Learning Models

6.12 Tumor

Department Of CSE(Artificial Intelligence And Machine Learning) 23

 Nephropathy Prediction Using Deep Learning Models

6.13 Confusion Matrix and Accuracy

Department Of CSE(Artificial Intelligence And Machine Learning) 24

 Nephropathy Prediction Using Deep Learning Models

6.14 Model Loss and Accuracy

APPLICATIONS

Department Of CSE(Artificial Intelligence And Machine Learning) 25

 Nephropathy Prediction Using Deep Learning Models

1. Early Detection: Deep learning models analyze CT scans and blood tests for
timely and accurate identification of kidney cancer, aiding in early detection.
2. Treatment Planning: Early cancer detection enhances treatment planning,
enabling healthcare professionals to develop personalized and effective
strategies for patients.
3. Reduction of False Results: The project aims to minimize false positives and
negatives in kidney cancer detection, preventing unnecessary treatments or
overlooking actual cases.
4. Comprehensive Assessment: Combining medical imaging and blood tests
provides a holistic patient assessment, allowing healthcare professionals to
consider both anatomical and biochemical factors.
5. Web-Based Tools: A user-friendly web interface facilitates easy interaction
for healthcare professionals and patients, offering rapid results for diagnostic
purposes using CT scan images or blood test data.
6. Research Contribution: The project contributes to medical research, utilizing
deep learning for kidney cancer detection, with potential advancements in
cancer detection methodologies.
7. Public Health Impact: Accurate and early detection improves overall survival
rates and reduces invasive treatments, contributing to a broader positive impact
on public health.

REFERENCES
[1] Matta Preethi Reddy, MD Farhan Mohiuddin, Shruthi Budde, Gajula
Jayanth, Ch Rajendra Prasad, Srikanth Yalabaka. “A Deep Learning Model for

Department Of CSE(Artificial Intelligence And Machine Learning) 26

 Nephropathy Prediction Using Deep Learning Models

Traffic Sign Detect ion and Recognition using Convolution Neural Network”,
Published: 24 June 2022.
[2] Kollem, S., Prasad, C. R., Ajayan, J., Malathy, V., & Subbarao, A. (2022).
“Brain tumor MRI image segmentation using an optimized multi-kernel FCM
method with a pre-processing stage. Multimedia Tools and Applications”, 1-30.
[3] Akram Z*, Kareem MS, Mughal B, Ahmed Z, and Aziz S. “Cancerous
Tumor Segmentation of Kidney Images and Prediction of Tumor Using Medical
Image Segmentation and Deep Learning Techniques”, Published: 09 Mar 2021.
[4] Kwang-Hyun Uhm, Seung-Won Jung, Moon Hyung Choi, Hong-Kyu Shin,
Jae-Ik Yoo, Se Won Oh, Jee Young Kim, Hyun Gi Kim, Young Joon Lee, Seo
Yeon Youn, Sung-Hoo Hong, Sung-Jea Ko. “Deep learning for end-to-end
kidney cancer diagnosis on multi-phase abdominal computed tomography”,
Published: 18 June 2021.
[5] Maha gharaibeh, Dalia Alz’bi, Malak Abdullah, Ismail Hmeidi, Mohammad
Rustom AI Nasar, "Radiology Imaging Scans for Early Diagnosis of Kidney
Tumors: A Review of Data Analytics-Based Machine Learning and Deep
Learning Approaches”, Doi: 10.3390/bdcc6010029, Published: 8 March 2022.
[6] Kollem, Sreedhar, et al. "Image denoising for magnetic resonance imaging
medical images using improved generalized cross ‐validation based on the
diffusivity function." International Journal of Imaging Systems and Technology
32.4 (2022): 1263-1285.
[7] Ehwa Yang, Chan Kyo Kim, Yi Guan, Bang-Bon Koo, Jae-Hun Kim, “3D
multi-scale residual fully convolutional neural network for segmentation of
extremely .
[8] Luana Bat ista da Cruz, José Denes Lima Araújo, Jonnison Lima Ferreira,
Joao Otavio Bandeira Diniz, Aristofanes Correa Silva, João Dallyson Sousa de
Almeida, Anselmo Cardoso de Paiva, Marcelo Gatt ass, “Kidney segmentation

Department Of CSE(Artificial Intelligence And Machine Learning) 27

 Nephropathy Prediction Using Deep Learning Models

from computed tomography images using deep neural network”,

Doi:10.1016/j.compbiomed.2020.10390621, published: 6 July2020
[9] Seokmin Han, Sung IL Hwang, Hak Jong Lee, “The classification of renal
cancer in 3- phase CT images using a deep learning method”, published: 16
May 2019.
[10] Duggirala SR, Kollem S, Rama Linga Reddy K (2021) an optimized SVM-
based possibilistic fuzzy c-means clustering algorithm for tumor segmentation.
Multimed Tools Appl 80(1):409–437

CONCLUSION
In this project, we have introduced a groundbreaking approach to kidney cancer
detection, leveraging advanced deep learning models and a user-friendly web
interface. The culmination of our efforts has resulted in a comprehensive system
that excels in accuracy, accessibility, and diagnostic depth.
Our dual-model architecture, featuring the MobileNet-based CT scan image
analysis and an Artificial Neural Network (ANN) for blood test data analysis,

Department Of CSE(Artificial Intelligence And Machine Learning) 28

 Nephropathy Prediction Using Deep Learning Models

has demonstrated remarkable accuracy levels. Achieving training and validation

accuracies of 99.00% for CT scan images and 90% and 97%, respectively, for
blood test data, our models outperform existing systems, providing healthcare
professionals with a robust tool for precise diagnoses.
The integration of diverse and extensive datasets, comprising a wide range of
CT scan images and comprehensive blood test data, ensures that our models can
generalize effectively and adapt to various clinical scenarios. This dataset
richness is crucial for achieving the highest standards of diagnostic reliability.
The user-friendly web interface, developed using Flask, simplifies the
diagnostic process, making it accessible to healthcare professionals and patients
alike. This interface reduces barriers to entry, facilitating the rapid and accurate
assessment of kidney cancer risk.
In conclusion, our project represents a significant advancement in kidney cancer
detection, offering a holistic, accurate, and accessible solution to enhance
patient care. By combining the power of deep learning models, diverse datasets,
and a user-friendly interface, we have contributed to the field of medical
diagnostics, paving the way for improved early detection and management of
kidney cancer, ultimately leading to better patient outcomes.

Department Of CSE(Artificial Intelligence And Machine Learning) 29