Deep learning for histopathological segmentation of smooth muscle in the urinary bladder
Abstract Background Histological assessment of smooth muscle is a critical step particularly in staging malignant tumors in various internal organs including the urinary bladder. Nonetheless, manual segmentation and classification of muscular tissues by pathologists is often challenging. Therefore,...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
BMC
2023-07-01
|
Series: | BMC Medical Informatics and Decision Making |
Subjects: | |
Online Access: | https://doi.org/10.1186/s12911-023-02222-3 |
_version_ | 1797778845125836800 |
---|---|
author | Sridevi K. Subramanya Rui Li Ying Wang Hiroshi Miyamoto Feng Cui |
author_facet | Sridevi K. Subramanya Rui Li Ying Wang Hiroshi Miyamoto Feng Cui |
author_sort | Sridevi K. Subramanya |
collection | DOAJ |
description | Abstract Background Histological assessment of smooth muscle is a critical step particularly in staging malignant tumors in various internal organs including the urinary bladder. Nonetheless, manual segmentation and classification of muscular tissues by pathologists is often challenging. Therefore, a fully automated and reliable smooth muscle image segmentation system is in high demand. Methods To characterize muscle fibers in the urinary bladder, including muscularis mucosa (MM) and muscularis propria (MP), we assessed 277 histological images from surgical specimens, using two well-known deep learning (DL) model groups, one including VGG16, ResNet18, SqueezeNet, and MobileNetV2, considered as a patch-based approach, and the other including U-Net, MA-Net, DeepLabv3 + , and FPN, considered as a pixel-based approach. All the trained models in both the groups were evaluated at pixel-level for their performance. Results For segmenting MP and non-MP (including MM) regions, MobileNetV2, in the patch-based approach and U-Net, in the pixel-based approach outperformed their peers in the groups with mean Jaccard Index equal to 0.74 and 0.79, and mean Dice co-efficient equal to 0.82 and 0.88, respectively. We also demonstrated the strengths and weaknesses of the models in terms of speed and prediction accuracy. Conclusions This work not only creates a benchmark for future development of tools for the histological segmentation of smooth muscle but also provides an effective DL-based diagnostic system for accurate pathological staging of bladder cancer. |
first_indexed | 2024-03-12T23:23:25Z |
format | Article |
id | doaj.art-91ddb5549fc146a2b097dc2fc4c9a39a |
institution | Directory Open Access Journal |
issn | 1472-6947 |
language | English |
last_indexed | 2024-03-12T23:23:25Z |
publishDate | 2023-07-01 |
publisher | BMC |
record_format | Article |
series | BMC Medical Informatics and Decision Making |
spelling | doaj.art-91ddb5549fc146a2b097dc2fc4c9a39a2023-07-16T11:18:54ZengBMCBMC Medical Informatics and Decision Making1472-69472023-07-0123111610.1186/s12911-023-02222-3Deep learning for histopathological segmentation of smooth muscle in the urinary bladderSridevi K. Subramanya0Rui Li1Ying Wang2Hiroshi Miyamoto3Feng Cui4Thomas H. Gosnell School of Life Sciences, Rochester Institute of TechnologyGolisano College of Computing and Information Sciences, Rochester Institute of TechnologyDepartment of Pathology and Laboratory Medicine, University of Rochester Medical CenterDepartment of Pathology and Laboratory Medicine, University of Rochester Medical CenterThomas H. Gosnell School of Life Sciences, Rochester Institute of TechnologyAbstract Background Histological assessment of smooth muscle is a critical step particularly in staging malignant tumors in various internal organs including the urinary bladder. Nonetheless, manual segmentation and classification of muscular tissues by pathologists is often challenging. Therefore, a fully automated and reliable smooth muscle image segmentation system is in high demand. Methods To characterize muscle fibers in the urinary bladder, including muscularis mucosa (MM) and muscularis propria (MP), we assessed 277 histological images from surgical specimens, using two well-known deep learning (DL) model groups, one including VGG16, ResNet18, SqueezeNet, and MobileNetV2, considered as a patch-based approach, and the other including U-Net, MA-Net, DeepLabv3 + , and FPN, considered as a pixel-based approach. All the trained models in both the groups were evaluated at pixel-level for their performance. Results For segmenting MP and non-MP (including MM) regions, MobileNetV2, in the patch-based approach and U-Net, in the pixel-based approach outperformed their peers in the groups with mean Jaccard Index equal to 0.74 and 0.79, and mean Dice co-efficient equal to 0.82 and 0.88, respectively. We also demonstrated the strengths and weaknesses of the models in terms of speed and prediction accuracy. Conclusions This work not only creates a benchmark for future development of tools for the histological segmentation of smooth muscle but also provides an effective DL-based diagnostic system for accurate pathological staging of bladder cancer.https://doi.org/10.1186/s12911-023-02222-3Deep learningSmooth muscleBladder cancerPathological slidesSegmentation |
spellingShingle | Sridevi K. Subramanya Rui Li Ying Wang Hiroshi Miyamoto Feng Cui Deep learning for histopathological segmentation of smooth muscle in the urinary bladder BMC Medical Informatics and Decision Making Deep learning Smooth muscle Bladder cancer Pathological slides Segmentation |
title | Deep learning for histopathological segmentation of smooth muscle in the urinary bladder |
title_full | Deep learning for histopathological segmentation of smooth muscle in the urinary bladder |
title_fullStr | Deep learning for histopathological segmentation of smooth muscle in the urinary bladder |
title_full_unstemmed | Deep learning for histopathological segmentation of smooth muscle in the urinary bladder |
title_short | Deep learning for histopathological segmentation of smooth muscle in the urinary bladder |
title_sort | deep learning for histopathological segmentation of smooth muscle in the urinary bladder |
topic | Deep learning Smooth muscle Bladder cancer Pathological slides Segmentation |
url | https://doi.org/10.1186/s12911-023-02222-3 |
work_keys_str_mv | AT srideviksubramanya deeplearningforhistopathologicalsegmentationofsmoothmuscleintheurinarybladder AT ruili deeplearningforhistopathologicalsegmentationofsmoothmuscleintheurinarybladder AT yingwang deeplearningforhistopathologicalsegmentationofsmoothmuscleintheurinarybladder AT hiroshimiyamoto deeplearningforhistopathologicalsegmentationofsmoothmuscleintheurinarybladder AT fengcui deeplearningforhistopathologicalsegmentationofsmoothmuscleintheurinarybladder |