Weakly supervised computational pathology for prognostic stratification in locally advanced nasopharyngeal carcinoma: a multicenter external validation study
Abstract
IntroductionAccurate prognostic stratification remains challenging in locally advanced nasopharyngeal carcinoma (LA-NPC), as conventional clinicopathological factors provide limited precision for individualized risk assessment. Although hematoxylin and eosin (H&E)-stained whole-slide images (WSIs) contain abundant morphological information, their prognostic potential in LA-NPC remains underexplored using computational pathology approaches.Material and methodsPatients with stage III–IVA LA-NPC from three Chinese tertiary centers were allocated into a training cohort, an internal validation cohort, and an independent external test cohort. A CLAM-inspired attention-based multiple-instance learning (MIL) model was developed using pretreatment H&E-stained WSIs and externally validated. The model was compared with four alternative MIL architectures incorporating different feature extractors and pooling strategies. Prognostic performance for 3-year progression-free survival (PFS) was evaluated using time-dependent receiver operating characteristic analysis, the concordance index (C-index), Kaplan–Meier analysis, decision curve analysis, and Cox regression.ResultsA total of 253 patients contributing 265 WSIs were included: 149 patients/160 WSIs in the training cohort, 52 patients/53 WSIs in the internal validation cohort, and 52 patients/52 WSIs in the independent external test cohort. In the internal validation cohort, the CLAM-inspired Attention MIL model yielded an AUC of 0.883 (95% CI, 0.762–0.969) and a C-index of 0.802 (95% CI, 0.705–0.884) for predicting 3-year PFS. In the external test cohort, the model achieved an AUC of 0.721 (95% CI, 0.573–0.849) and a C-index of 0.710 (95% CI, 0.591–0.819). The model significantly stratified patients into high- and low-risk groups in both cohorts and showed numerically higher AUC and C-index estimates than the four comparative MIL architectures. The risk score was linked to an elevated risk of progression (HR, 1.30; 95% CI, 1.06-1.60; P = 0.0126).ConclusionsA weakly supervised CLAM-inspired Attention MIL framework can extract prognostically relevant information from routine H&E-stained WSIs and provide externally validated risk stratification for patients with LA-NPC. This approach may complement conventional clinicopathological assessment and facilitate individualized prognostic evaluation in precision oncology.
Related
- Supplementary value of primary tumor volume for prognostic stratification in patients with T3-4 nasopharyngeal carcinoma: A retrospective cohort study
- Lymph Node Regression Improves Prognostic Stratification in Esophageal Squamous Cell Carcinoma Receiving Neoadjuvant Chemoimmunotherapy: A Multicenter Retrospective Cohort Study.
- Baseline Predictors, Prognostic Impact, and Risk Stratification of Early Progressive Disease on Atezolizumab-Bevacizumab in Unresectable Hepatocellular Carcinoma: A Multicenter Real-World Study.
- Prognostic value of modified thrombelastometry for thromboembolic risk stratification in neurosurgical patients: protocol for a prospective, multicenter, observational study (TENS).
- Smoking exposure and overall survival in nasopharyngeal carcinoma: a multicenter cohort from non-endemic northern China with an exploratory meta-analysis
BibTeX
@article{jiang2026weakly,
title = {{Weakly supervised computational pathology for prognostic stratification in locally advanced nasopharyngeal carcinoma: a multicenter external validation study}},
author = {Chao Jiang and Wenxuan Lu and Mingjun Ding and Xuanchen Zhou and Butuo Li and Xiong Liu},
journal = {Frontiers in Oncology},
year = {2026},
month = nov,
doi = {10.3389/fonc.2026.1966935},
url = {https://www.frontiersin.org/articles/10.3389/fonc.2026.1966935},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Weakly supervised computational pathology for prognostic stratification in locally advanced nasopharyngeal carcinoma: a multicenter external validation study
AU - Chao Jiang
AU - Wenxuan Lu
AU - Mingjun Ding
AU - Xuanchen Zhou
AU - Butuo Li
AU - Xiong Liu
PY - 2026
DA - 2026/11/09
JO - Frontiers in Oncology
DO - 10.3389/fonc.2026.1966935
UR - https://www.frontiersin.org/articles/10.3389/fonc.2026.1966935
AB - IntroductionAccurate prognostic stratification remains challenging in locally advanced nasopharyngeal carcinoma (LA-NPC), as conventional clinicopathological factors provide limited precision for individualized risk assessment. Although hematoxylin and eosin (H&E)-stained whole-slide images (WSIs) contain abundant morphological information, their prognostic potential in LA-NPC remains underexplored using computational pathology approaches.Material and methodsPatients with stage III–IVA LA-NPC from three Chinese tertiary centers were allocated into a training cohort, an internal validation cohort, and an independent external test cohort. A CLAM-inspired attention-based multiple-instance learning (MIL) model was developed using pretreatment H&E-stained WSIs and externally validated. The model was compared with four alternative MIL architectures incorporating different feature extractors and pooling strategies. Prognostic performance for 3-year progression-free survival (PFS) was evaluated using time-dependent receiver operating characteristic analysis, the concordance index (C-index), Kaplan–Meier analysis, decision curve analysis, and Cox regression.ResultsA total of 253 patients contributing 265 WSIs were included: 149 patients/160 WSIs in the training cohort, 52 patients/53 WSIs in the internal validation cohort, and 52 patients/52 WSIs in the independent external test cohort. In the internal validation cohort, the CLAM-inspired Attention MIL model yielded an AUC of 0.883 (95% CI, 0.762–0.969) and a C-index of 0.802 (95% CI, 0.705–0.884) for predicting 3-year PFS. In the external test cohort, the model achieved an AUC of 0.721 (95% CI, 0.573–0.849) and a C-index of 0.710 (95% CI, 0.591–0.819). The model significantly stratified patients into high- and low-risk groups in both cohorts and showed numerically higher AUC and C-index estimates than the four comparative MIL architectures. The risk score was linked to an elevated risk of progression (HR, 1.30; 95% CI, 1.06-1.60; P = 0.0126).ConclusionsA weakly supervised CLAM-inspired Attention MIL framework can extract prognostically relevant information from routine H&E-stained WSIs and provide externally validated risk stratification for patients with LA-NPC. This approach may complement conventional clinicopathological assessment and facilitate individualized prognostic evaluation in precision oncology.
SN - 2234-943X
ER -