Automatic Screening of Pediatric Renal Ultrasound Abnormalities: Deep Learning and Transfer Learning Approach
BackgroundIn recent years, the progress and generalization surrounding portable ultrasonic probes has made ultrasound (US) a useful tool for physicians when making a diagnosis. With the advent of machine learning and deep learning, the development of a computer-aided diagnost...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
JMIR Publications
2022-11-01
|
Series: | JMIR Medical Informatics |
Online Access: | https://medinform.jmir.org/2022/11/e40878 |
_version_ | 1797734722150858752 |
---|---|
author | Ming-Chin Tsai Henry Horng-Shing Lu Yueh-Chuan Chang Yung-Chieh Huang Lin-Shien Fu |
author_facet | Ming-Chin Tsai Henry Horng-Shing Lu Yueh-Chuan Chang Yung-Chieh Huang Lin-Shien Fu |
author_sort | Ming-Chin Tsai |
collection | DOAJ |
description |
BackgroundIn recent years, the progress and generalization surrounding portable ultrasonic probes has made ultrasound (US) a useful tool for physicians when making a diagnosis. With the advent of machine learning and deep learning, the development of a computer-aided diagnostic system for screening renal US abnormalities can assist general practitioners in the early detection of pediatric kidney diseases.
ObjectiveIn this paper, we sought to evaluate the diagnostic performance of deep learning techniques to classify kidney images as normal and abnormal.
MethodsWe chose 330 normal and 1269 abnormal pediatric renal US images for establishing a model for artificial intelligence. The abnormal images involved stones, cysts, hyperechogenicity, space-occupying lesions, and hydronephrosis. We performed preprocessing of the original images for subsequent deep learning. We redefined the final connecting layers for classification of the extracted features as abnormal or normal from the ResNet-50 pretrained model. The performances of the model were tested by a validation data set using area under the receiver operating characteristic curve, accuracy, specificity, and sensitivity.
ResultsThe deep learning model, 94 MB parameters in size, based on ResNet-50, was built for classifying normal and abnormal images. The accuracy, (%)/area under curve, of the validated images of stone, cyst, hyperechogenicity, space-occupying lesions, and hydronephrosis were 93.2/0.973, 91.6/0.940, 89.9/0.940, 91.3/0.934, and 94.1/0.996, respectively. The accuracy of normal image classification in the validation data set was 90.1%. Overall accuracy of (%)/area under curve was 92.9/0.959..
ConclusionsWe established a useful, computer-aided model for automatic classification of pediatric renal US images in terms of normal and abnormal categories. |
first_indexed | 2024-03-12T12:47:27Z |
format | Article |
id | doaj.art-7b2006feb9804eaeb2b0a3b14246c212 |
institution | Directory Open Access Journal |
issn | 2291-9694 |
language | English |
last_indexed | 2024-03-12T12:47:27Z |
publishDate | 2022-11-01 |
publisher | JMIR Publications |
record_format | Article |
series | JMIR Medical Informatics |
spelling | doaj.art-7b2006feb9804eaeb2b0a3b14246c2122023-08-28T23:13:08ZengJMIR PublicationsJMIR Medical Informatics2291-96942022-11-011011e4087810.2196/40878Automatic Screening of Pediatric Renal Ultrasound Abnormalities: Deep Learning and Transfer Learning ApproachMing-Chin Tsaihttps://orcid.org/0000-0002-7206-4228Henry Horng-Shing Luhttps://orcid.org/0000-0002-4392-3361Yueh-Chuan Changhttps://orcid.org/0000-0001-6529-1724Yung-Chieh Huanghttps://orcid.org/0000-0003-2945-4404Lin-Shien Fuhttps://orcid.org/0000-0003-4770-3208 BackgroundIn recent years, the progress and generalization surrounding portable ultrasonic probes has made ultrasound (US) a useful tool for physicians when making a diagnosis. With the advent of machine learning and deep learning, the development of a computer-aided diagnostic system for screening renal US abnormalities can assist general practitioners in the early detection of pediatric kidney diseases. ObjectiveIn this paper, we sought to evaluate the diagnostic performance of deep learning techniques to classify kidney images as normal and abnormal. MethodsWe chose 330 normal and 1269 abnormal pediatric renal US images for establishing a model for artificial intelligence. The abnormal images involved stones, cysts, hyperechogenicity, space-occupying lesions, and hydronephrosis. We performed preprocessing of the original images for subsequent deep learning. We redefined the final connecting layers for classification of the extracted features as abnormal or normal from the ResNet-50 pretrained model. The performances of the model were tested by a validation data set using area under the receiver operating characteristic curve, accuracy, specificity, and sensitivity. ResultsThe deep learning model, 94 MB parameters in size, based on ResNet-50, was built for classifying normal and abnormal images. The accuracy, (%)/area under curve, of the validated images of stone, cyst, hyperechogenicity, space-occupying lesions, and hydronephrosis were 93.2/0.973, 91.6/0.940, 89.9/0.940, 91.3/0.934, and 94.1/0.996, respectively. The accuracy of normal image classification in the validation data set was 90.1%. Overall accuracy of (%)/area under curve was 92.9/0.959.. ConclusionsWe established a useful, computer-aided model for automatic classification of pediatric renal US images in terms of normal and abnormal categories.https://medinform.jmir.org/2022/11/e40878 |
spellingShingle | Ming-Chin Tsai Henry Horng-Shing Lu Yueh-Chuan Chang Yung-Chieh Huang Lin-Shien Fu Automatic Screening of Pediatric Renal Ultrasound Abnormalities: Deep Learning and Transfer Learning Approach JMIR Medical Informatics |
title | Automatic Screening of Pediatric Renal Ultrasound Abnormalities: Deep Learning and Transfer Learning Approach |
title_full | Automatic Screening of Pediatric Renal Ultrasound Abnormalities: Deep Learning and Transfer Learning Approach |
title_fullStr | Automatic Screening of Pediatric Renal Ultrasound Abnormalities: Deep Learning and Transfer Learning Approach |
title_full_unstemmed | Automatic Screening of Pediatric Renal Ultrasound Abnormalities: Deep Learning and Transfer Learning Approach |
title_short | Automatic Screening of Pediatric Renal Ultrasound Abnormalities: Deep Learning and Transfer Learning Approach |
title_sort | automatic screening of pediatric renal ultrasound abnormalities deep learning and transfer learning approach |
url | https://medinform.jmir.org/2022/11/e40878 |
work_keys_str_mv | AT mingchintsai automaticscreeningofpediatricrenalultrasoundabnormalitiesdeeplearningandtransferlearningapproach AT henryhorngshinglu automaticscreeningofpediatricrenalultrasoundabnormalitiesdeeplearningandtransferlearningapproach AT yuehchuanchang automaticscreeningofpediatricrenalultrasoundabnormalitiesdeeplearningandtransferlearningapproach AT yungchiehhuang automaticscreeningofpediatricrenalultrasoundabnormalitiesdeeplearningandtransferlearningapproach AT linshienfu automaticscreeningofpediatricrenalultrasoundabnormalitiesdeeplearningandtransferlearningapproach |