Pancancer outcome prediction via a unified weakly supervised deep learning model

Accurate prognosis prediction is essential for guiding cancer treatment and improving patient outcomes. While recent studies have demonstrated the potential of histopathological images in survival analysis, existing models are typically developed in a cancer-specific manner, lack extensive external validation, and often rely on molecular data that are not routinely available in clinical practice. To address these limitations, we present PROGPATH, a unified model capable of integrating histopathological image features with routinely collected clinical variables to achieve pancancer prognosis prediction. PROGPATH employs a weakly supervised deep learning architecture built upon the foundation model for image encoding. Morphological features are aggregated through an attention-guided multiple instance learning module and fused with clinical information via a cross-attention transformer. A router-based classification strategy further refines the prediction performance. PROGPATH was trained on 7999 whole-slide images (WSIs) from 6,670 patients across 15 cancer types, and extensively validated on 17 external cohorts with a total of 7374 WSIs from 4441 patients, covering 12 cancer types from 8 consortia and institutions across three continents. PROGPATH achieved consistently superior performance compared with state-of-the-art multimodal prognosis prediction models. It demonstrated strong generalizability across cancer types and robustness in stratified subgroups, including early- and advanced-stage patients, treatment cohorts (radiotherapy and pharmaceutical therapy), and biomarker-defined subsets. We further provide model interpretability by identifying pathological patterns critical to PROGPATH's risk predictions, such as the degree of cell differentiation and extent of necrosis. Together, these results highlight the potential of PROGPATH to support pancancer outcome prediction and inform personalized cancer management strategies.

Authors

Institutions

Publication Details

Journal
Signal Transduction and Targeted Therapy
Published
2025-09-02
DOI
https://doi.org/10.1038/s41392-025-02374-w
Citations
6
Primary Topic
AI in cancer detection
Type
article
Field-Weighted Citation Impact
14.89

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

Pancancer outcome prediction via a unified weakly supervised deep learning model

Nathan A. Pennell, Jinxi Xiang, Shlomo A. Koyfman, Kailin Yang et al.
6 citations
Signal Transduction and Targeted Therapy
AI in cancer detection
14.89
article

Pancancer outcome prediction via a unified weakly supervised deep learning model

Nathan A. Pennell, Jinxi Xiang, Shlomo A. Koyfman, Kailin Yang, Anant Madabhushi, Cristian Barrera, Junhan Zhao, Yuanfeng Ji, Shuo Kang, Julie Wu, Germán Corredor, Jing Zhang, Krunal Pandav, Lukas Delasos, Wei Yuan, Xiangde Luo, Biyue Zhu, Sen Yang, Hongping Tang, Prantesh Jain, Xiyue Wang, Kun‐Hsing Yu, Junzhou Huang, Xiaohan Xing, James P. Lewis, Jiayu Zhang, Yuchen Li, Xiao Han, Yijiang Chen, Yufei Zhou, Kangning Zhang, Ning Mao, D. Xiao, Theodoros N. Teknos, Haiyan Zhang, Fang Wang, Himanshu Maurya
article en
6 citations

Abstract

Accurate prognosis prediction is essential for guiding cancer treatment and improving patient outcomes. While recent studies have demonstrated the potential of histopathological images in survival analysis, existing models are typically developed in a cancer-specific manner, lack extensive external validation, and often rely on molecular data that are not routinely available in clinical practice. To address these limitations, we present PROGPATH, a unified model capable of integrating histopathological image features with routinely collected clinical variables to achieve pancancer prognosis prediction. PROGPATH employs a weakly supervised deep learning architecture built upon the foundation model for image encoding. Morphological features are aggregated through an attention-guided multiple instance learning module and fused with clinical information via a cross-attention transformer. A router-based classification strategy further refines the prediction performance. PROGPATH was trained on 7999 whole-slide images (WSIs) from 6,670 patients across 15 cancer types, and extensively validated on 17 external cohorts with a total of 7374 WSIs from 4441 patients, covering 12 cancer types from 8 consortia and institutions across three continents. PROGPATH achieved consistently superior performance compared with state-of-the-art multimodal prognosis prediction models. It demonstrated strong generalizability across cancer types and robustness in stratified subgroups, including early- and advanced-stage patients, treatment cohorts (radiotherapy and pharmaceutical therapy), and biomarker-defined subsets. We further provide model interpretability by identifying pathological patterns critical to PROGPATH's risk predictions, such as the degree of cell differentiation and extent of necrosis. Together, these results highlight the potential of PROGPATH to support pancancer outcome prediction and inform personalized cancer management strategies.

Signal Transduction and Targeted TherapyVol. 10(1)
Qingdao University (CN), University of Iowa (US), Brigham and Women's Hospital (US), Roswell Park Comprehensive Cancer Center (US), Cleveland Clinic (US), United States Department of Veterans Affairs (US), Harvard University (US), Emory University (US), The University of Texas at Arlington (US), Sichuan University (CN), University Hospitals of Cleveland (US), Yuhuangding Hospital (CN), Shenzhen Maternity and Child Healthcare Hospital (CN), Mayo Clinic in Arizona (US), Stanford Medicine (US), Children's Hospital of Chongqing Medical University (CN), Case Western Reserve University (US), Chongqing Medical University (CN), Vanderbilt University Medical Center (US), Stanford University (US)
U.S. Department of Defense, U.S. Department of Veterans Affairs, Eli Lilly and Company, AstraZeneca, National Cancer Center, National Natural Science Foundation of China, Department of Science and Technology of Sichuan Province, Sichuan University, National Institutes of Health, National Heart, Lung, and Blood Institute, National Cancer Institute, DOD Prostate Cancer Research Program, DOD Peer Reviewed Cancer Research Program
Openalex Percentile: Top 1%
AI in cancer detection
14.89
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.