Journal article
Pancancer outcome prediction via a unified weakly supervised deep learning model
Signal transduction and targeted therapy, Vol.10(1), 285
09/03/2025
DOI: 10.1038/s41392-025-02374-w
PMCID: PMC12405520
PMID: 40897689
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.
Details
- Title: Subtitle
- Pancancer outcome prediction via a unified weakly supervised deep learning model
- Creators
- Wei Yuan - Sichuan UniversityYijiang Chen - Stanford UniversityBiyue Zhu - Children's Hospital of Chongqing Medical UniversitySen Yang - Stanford UniversityJiayu Zhang - Sichuan UniversityNing Mao - Yuhuangding HospitalJinxi Xiang - Stanford UniversityYuchen Li - Stanford UniversityYuanfeng Ji - Stanford UniversityXiangde Luo - Stanford UniversityKangning Zhang - Stanford UniversityXiaohan Xing - Stanford UniversityShuo Kang - Children's Hospital of Chongqing Medical UniversityDongyuan Xiao - Children's Hospital of Chongqing Medical UniversityFang Wang - Yuhuangding HospitalJinkun Wu - Qingdao UniversityHaiyan Zhang - Yuhuangding HospitalHongping Tang - Shenzhen Maternity and Child Healthcare HospitalHimanshu Maurya - Emory UniversityGerman Corredor - Emory UniversityCristian Barrera - Emory UniversityYufei Zhou - Case Western Reserve UniversityKrunal Pandav - Emory UniversityJunhan Zhao - Harvard UniversityPrantesh Jain - Roswell Park Comprehensive Cancer CenterLuke Delasos - Cleveland ClinicJunzhou Huang - The University of Texas at ArlingtonKailin Yang - University of IowaTheodoros N. Teknos - University Hospitals of ClevelandJames Lewis - Vanderbilt University Medical CenterShlomo Koyfman - Cleveland ClinicNathan A. Pennell - Cleveland ClinicKun-Hsing Yu - Brigham and Women's HospitalXiao Han - Stanford UniversityJing Zhang - Sichuan UniversityXiyue Wang - Sichuan UniversityAnant Madabhushi - United States Department of Veterans Affairs
- Resource Type
- Journal article
- Publication Details
- Signal transduction and targeted therapy, Vol.10(1), 285
- DOI
- 10.1038/s41392-025-02374-w
- PMID
- 40897689
- PMCID
- PMC12405520
- NLM abbreviation
- Signal Transduct Target Ther
- ISSN
- 2059-3635
- eISSN
- 2059-3635
- Publisher
- Nature Publishing Group UK
- Grant note
- 61571314 / National Natural Science Foundation of China (National Science Foundation of China) (https://doi.org/10.13039/501100001809) R01CA268287A1; U01CA269181; R01CA26820701A1; R01CA249992-01A1; R01CA202752-01A1 / National Cancer Center (https://doi.org/10.13039/100008746) 2023YFS0327-LH / Department of Science and Technology of Sichuan Province (Sichuan Provincial Department of Science and Technology) (https://doi.org/10.13039/501100004829) 1R01HL15127701A1 / U.S. Department of Health & Human Services | NIH | National Heart, Lung, and Blood Institute (NHLBI) (https://doi.org/10.13039/100000050)
- Language
- English
- Date published
- 09/03/2025
- Academic Unit
- Radiation Oncology
- Record Identifier
- 9984958608702771
Metrics
16 Record Views