IMG-66. Self-supervised multimodal learning for survival prediction in glioblastoma: a multicenter study from the ReSPOND consortium

Abstract PURPOSE Glioblastoma is the most aggressive adult brain tumor, with a median overall survival of approximately 15 months. It is important to build accurate prognostic models for glioblastoma patients to inform clinical management and trials. This study proposes a self-supervised learning-based approach with multimodal data integration for survival prediction and prognostic stratification of glioblastoma patients on the ReSPOND consortium. METHODS We curated a multi-parametric MRI dataset (T1, T1CE, T2, FLAIR) of 3,119 glioblastoma patients from 22 institutions across 3 continents. Masked autoencoder (MAE) was adapted to pretrain a Vision Transformer (ViT) encoder by reconstructing the masked image patches. The encoder was utilized for extracting patch embeddings for survival tasks, with cross-attention mechanism to incorporate the molecular and clinical information (age, sex, extent of resection, MGMT) to guide imaging feature aggregation. Imaging and clinical embeddings were fused through a multi-layer perceptron (MLP) for log-risk hazard estimation, optimized using Cox partial likelihood. Model performance and generalizability were assessed via k-fold cross-validation on the ReSPOND consortium and the leave-one-site-out validation was performed on 11 institutions comparing with CoxPH, DeepSurv and DeepHit. Prognostic risk stratification via Kaplan-Meier analysis divided the patients into low-, medium- and high-risk subgroups per site. RESULTS Multimodal data integration using the proposed framework achieved the highest C-index (0.674 ± 0.017) on the ReSPOND consortium. Integration of clinical information and MGMT consistently boosted the performance of the proposed model across sites (0.615 ± 0.046 vs. 0.662 ± 0.044). The imaging-based approaches, i.e., radiomics and convolutional neural network (CNN) features performed less robustly. The Kaplan-Meier curves and log-rank tests suggested the proposed framework achieved more separable prognostic subgroups. CONCLUSION The proposed self-supervised multimodal learning framework shows promise for survival prediction and prognostic risk stratification in glioblastoma. It highlights the challenge for clinical model deployment due to the data heterogeneity in multi-institutional cohort.

Authors

Institutions

Publication Details

Journal
Neuro-Oncology
Published
2025-11-01
DOI
https://doi.org/10.1093/neuonc/noaf201.1145
Primary Topic
Glioma Diagnosis and Treatment
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

IMG-66. Self-supervised multimodal learning for survival prediction in glioblastoma: a multicenter study from the ReSPOND consortium

Anahita Fathi Kazerooni, Elias R. Melhem, Susan Chang, Thomas C. Booth et al.
Neuro-Oncology
Glioma Diagnosis and Treatment
article

IMG-66. Self-supervised multimodal learning for survival prediction in glioblastoma: a multicenter study from the ReSPOND consortium

Anahita Fathi Kazerooni, Elias R. Melhem, Susan Chang, Thomas C. Booth, Yoon Seong Choi, Pratik Chaudhari, Stephen Bagley, Yuji Matsumoto, Jaume Capellades, Santiago Cepeda, MacLean Nasrallah, Adam E. Flanders, Daniel S. Marcus, Rajan Jain, Robert A. Lustig, Ujjwal Baid, Christos A. Davatzikos, Sung Soo Ahn, Joshua D. Palmer, Kristin Waite, Evan Calabrese, Pallavi Tiwari, Spyridon Bakas, Steven Brem, Adam P. Dicker, Javier Villanueva‐Meyer, Jill S. Barnholtz‐Sloan, Carmen Balañá, Wenyin Shi, Pamela LaMontagne, Chaitra Badve, Brent Griffith, Yu Tian, Donald M. O’Rourke, Laila Poisson, Suyash Mohan, Jeffrey D. Rudie, Josep Puig, Hamed Akbari, Fanyang Yu, Rivka R. Colen, Graeme F. Woodworth, Mariam Aboian, Andrew Edward Sloan, José García, Jun Guo, Han Tianyu, Arnab Chakravarti, Rubo Xing, Benedikt Wiestler
article en

Abstract

Abstract PURPOSE Glioblastoma is the most aggressive adult brain tumor, with a median overall survival of approximately 15 months. It is important to build accurate prognostic models for glioblastoma patients to inform clinical management and trials. This study proposes a self-supervised learning-based approach with multimodal data integration for survival prediction and prognostic stratification of glioblastoma patients on the ReSPOND consortium. METHODS We curated a multi-parametric MRI dataset (T1, T1CE, T2, FLAIR) of 3,119 glioblastoma patients from 22 institutions across 3 continents. Masked autoencoder (MAE) was adapted to pretrain a Vision Transformer (ViT) encoder by reconstructing the masked image patches. The encoder was utilized for extracting patch embeddings for survival tasks, with cross-attention mechanism to incorporate the molecular and clinical information (age, sex, extent of resection, MGMT) to guide imaging feature aggregation. Imaging and clinical embeddings were fused through a multi-layer perceptron (MLP) for log-risk hazard estimation, optimized using Cox partial likelihood. Model performance and generalizability were assessed via k-fold cross-validation on the ReSPOND consortium and the leave-one-site-out validation was performed on 11 institutions comparing with CoxPH, DeepSurv and DeepHit. Prognostic risk stratification via Kaplan-Meier analysis divided the patients into low-, medium- and high-risk subgroups per site. RESULTS Multimodal data integration using the proposed framework achieved the highest C-index (0.674 ± 0.017) on the ReSPOND consortium. Integration of clinical information and MGMT consistently boosted the performance of the proposed model across sites (0.615 ± 0.046 vs. 0.662 ± 0.044). The imaging-based approaches, i.e., radiomics and convolutional neural network (CNN) features performed less robustly. The Kaplan-Meier curves and log-rank tests suggested the proposed framework achieved more separable prognostic subgroups. CONCLUSION The proposed self-supervised multimodal learning framework shows promise for survival prediction and prognostic risk stratification in glioblastoma. It highlights the challenge for clinical model deployment due to the data heterogeneity in multi-institutional cohort.

Neuro-OncologyVol. 27(Supplement_5)
University of Maryland, Baltimore (US), University of Wisconsin–Madison (US), Thomas Jefferson University (US), Emory University (US), Henry Ford Health System (US), Santa Clara University (US), National University of Singapore (SG), University of Pittsburgh (US), Duke University (US), University of California, San Francisco (US), King's College London (GB), Yonsei University (KR), Washington University in St. Louis (US), University of California San Diego (US), NYU Langone Health (US), Institut Català d'Oncologia (ES), Hospital Clínic de Barcelona (ES), Premier Community Health (US), University Hospitals Seidman Cancer Center (US), National Cancer Institute (MY), Badalona Serveis Assistencials (ES), National Cancer Institute (US), William S. Middleton Memorial Veterans Hospital (US), Hospital Universitario Río Hortega (ES), Consorci Institut D'Investigacions Biomediques August Pi I Sunyer (ES), Indiana University School of Medicine, The Ohio State University (US), Indiana University – Purdue University Indianapolis (US), Technical University of Munich (DE), University of Pennsylvania (US), Philadelphia University (US)
Openalex Percentile: Top 52%
Glioma Diagnosis and Treatment
DOISourceOpenAlex
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.