A generalized AI method for pathology cancer diagnosis and prognosis prediction based on transfer learning and hierarchical split

被引:1
作者
Yin, Pengzhi [1 ]
Zhou, Zehao [2 ]
Liu, Jingze [2 ]
Jiang, Nan [3 ]
Zhang, Junchao [1 ]
Liu, Shiyu [3 ]
Wang, Feiyang [3 ]
Wang, Li [4 ]
机构
[1] Cent South Univ, Sch Automat, Changsha 410083, Peoples R China
[2] Xinjiang Univ, Sch Software, Urumqi 830001, Peoples R China
[3] Cent South Univ, XiangYa Sch Med, Changsha 410083, Peoples R China
[4] Tsinghua Univ, Coll Comp Sci & Technol, Beijing 100084, Peoples R China
基金
中国国家自然科学基金;
关键词
artificial intelligence; regular network; hierarchical split; dual transfer learning; cancer diagnosis and prognosis prediction; NETWORK;
D O I
10.1088/1361-6560/aced34
中图分类号
R318 [生物医学工程];
学科分类号
0831 ;
摘要
Objective. This study aims to propose a generalized AI method for pathology cancer diagnosis and prognosis prediction based on transfer learning and hierarchical split. Approach. We present a neural network framework for cancer diagnosis and prognosis prediction in pathological images. To enhance the network's depth and width, we employ a hierarchical split block (HS-Block) to create an AI-aided diagnosis system suitable for semi-supervised clinical settings with limited labeled samples and cross-domain tasks. By incorporating a lightweight convolution unit based on the HS-Block, we improve the feature information extraction capabilities of a regular network (RegNet). Additionally, we integrate a Convolutional Block Attention Module into the first and last convolutions to optimize the extraction of global features and local details. To address limited sample labels, we employ a dual-transfer learning (DTL) mechanism named DTL-HS-Regnet, enabling semi-supervised learning in clinical settings. Main results. Our proposed DTL-HS-Regnet model outperforms other advanced deep-learning models in three different types of cancer diagnosis tasks. It demonstrates superior feature extraction ability, achieving an average sensitivity, specificity, accuracy, and F1 score of 0.9987, 1.0000, 1.0000 and 0.9992, respectively. Furthermore, we evaluate the model's capability to directly extract prognosis prediction information from pathological images by constructing patient cohorts. The results show that the correlation between DTL-HS-Regnet predictions and the presence of cancer-associated fibroblasts is comparable to that of pathologists. Significance. Our proposed AI method offers a generalized approach for cancer diagnosis and prognosis prediction in pathology. The outstanding performance of the DTL-HS-Regnet model demonstrates its potential for improving current practices in image digital pathology, expanding the boundaries of cancer treatment in two critical areas.
引用
收藏
页数:18
相关论文
共 50 条
[1]  
Bello I, 2021, ADV NEUR IN, V34
[2]   Machine Learning for Risk Prediction of Oesophago-Gastric Cancer in Primary Care: Comparison with Existing Risk-Assessment Tools [J].
Briggs, Emma ;
de Kamps, Marc ;
Hamilton, Willie ;
Johnson, Owen ;
McInerney, Ciaran D. ;
Neal, Richard D. .
CANCERS, 2022, 14 (20)
[3]   The immune contexture and Immunoscore in cancer prognosis and therapeutic efficacy [J].
Bruni, Daniela ;
Angell, Helen K. ;
Galon, Jerome .
NATURE REVIEWS CANCER, 2020, 20 (11) :662-680
[4]   Clinical-grade computational pathology using weakly supervised deep learning on whole slide images [J].
Campanella, Gabriele ;
Hanna, Matthew G. ;
Geneslaw, Luke ;
Miraflor, Allen ;
Silva, Vitor Werneck Krauss ;
Busam, Klaus J. ;
Brogi, Edi ;
Reuter, Victor E. ;
Klimstra, David S. ;
Fuchs, Thomas J. .
NATURE MEDICINE, 2019, 25 (08) :1301-+
[5]   Absolute quantification of somatic DNA alterations in human cancer [J].
Carter, Scott L. ;
Cibulskis, Kristian ;
Helman, Elena ;
McKenna, Aaron ;
Shen, Hui ;
Zack, Travis ;
Laird, Peter W. ;
Onofrio, Robert C. ;
Winckler, Wendy ;
Weir, Barbara A. ;
Beroukhim, Rameen ;
Pellman, David ;
Levine, Douglas A. ;
Lander, Eric S. ;
Meyerson, Matthew ;
Getz, Gad .
NATURE BIOTECHNOLOGY, 2012, 30 (05) :413-+
[6]   Integrating genomic features for non-invasive early lung cancer detection [J].
Chabon, Jacob J. ;
Hamilton, Emily G. ;
Kurtz, David M. ;
Esfahani, Mohammad S. ;
Moding, Everett J. ;
Stehr, Henning ;
Schroers-Martin, Joseph ;
Nabet, Barzin Y. ;
Chen, Binbin ;
Chaudhuri, Aadel A. ;
Liu, Chih Long ;
Hui, Angela B. ;
Jin, Michael C. ;
Azad, Tej D. ;
Almanza, Diego ;
Jeon, Young-Jun ;
Nesselbush, Monica C. ;
Keh, Lyron Co Ting ;
Bonilla, Rene F. ;
Yoo, Christopher H. ;
Ko, Ryan B. ;
Chen, Emily L. ;
Merriott, David J. ;
Massion, Pierre P. ;
Mansfield, Aaron S. ;
Jen, Jin ;
Ren, Hong Z. ;
Lin, Steven H. ;
Costantino, Christina L. ;
Burr, Risa ;
Tibshirani, Robert ;
Gambhir, Sanjiv S. ;
Berry, Gerald J. ;
Jensen, Kristin C. ;
West, Robert B. ;
Neal, Joel W. ;
Wakelee, Heather A. ;
Loo, Billy W., Jr. ;
Kunder, Christian A. ;
Leung, Ann N. ;
Lui, Natalie S. ;
Berry, Mark F. ;
Shrager, Joseph B. ;
Nair, Viswam S. ;
Haber, Daniel A. ;
Sequist, Lecia V. ;
Alizadeh, Ash A. ;
Diehn, Maximilian .
NATURE, 2020, 580 (7802) :245-+
[7]   Dual Polarization Modality Fusion Network for Assisting Pathological Diagnosis [J].
Chen, Yi ;
Dong, Yang ;
Si, Lu ;
Yang, Wenming ;
Du, Shan ;
Tian, Xuewu ;
Li, Chao ;
Liao, Qingmin ;
Ma, Hui .
IEEE TRANSACTIONS ON MEDICAL IMAGING, 2023, 42 (01) :304-316
[8]  
Chen ZT, 2019, Arxiv, DOI arXiv:1804.05298
[9]   A Polarization-Imaging-Based Machine Learning Framework for Quantitative Pathological Diagnosis of Cervical Precancerous Lesions [J].
Dong, Yang ;
Wan, Jiachen ;
Wang, Xingjian ;
Xue, Jing-Hao ;
Zou, Jibin ;
He, Honghui ;
Li, Pengcheng ;
Hou, Anli ;
Ma, Hui .
IEEE TRANSACTIONS ON MEDICAL IMAGING, 2021, 40 (12) :3728-3738
[10]  
Dosovitskiy A, 2021, Arxiv, DOI arXiv:2010.11929