ESTRO 2025 - Abstract Book

S1052

Clinical – Head & neck

ESTRO 2025

Conclusion: This study highlights inter-center differences in OS and LC for patients with HPV-negative head and neck cancers. Chemotherapy and PS were key factors influencing outcomes, whereas the RT regimen showed no impact. These real-world findings underscore the importance of developing standardized treatment protocols and personalized care approaches to optimize patient outcomes. This work was supported by the Ministry of Health of the Czech Republic, grant no. NU22-03-00435.

Keywords: radiochemotherapy, HPV negative, survival

2983

Digital Poster Contrastive Learning Guided by Clinical Risk Scores Enhances WSI-Based Progression-Free Survival Prediction in Head and Neck Cancer Juan Duran 1 , Yujing Zou 1 , Laya Rafiee 1 , Elliot Wadge 1 , Khalil Sultanem 2 , Martin Vallières 3 , Shirin Abbasinejad Enger 1,4 1 Medical Physics Unit, McGill university, Montreal, Canada. 2 Department of Oncology, McGill university, Montreal, Canada. 3 Department of computer science, Sherbrooke university, Sherbrooke, Canada. 4 Oncology, Lady Davis institute for medical research, Montreal, Canada Purpose/Objective: Accurate progression-free survival (PFS) prediction is important for effective patient stratification and improved treatment decisions. Using deep learning (DL) and classical machine learning (ML) models, we developed a pipeline that refines whole slide image (WSI) embeddings through contrastive learning (CL) guided by risk scores from a survival module trained on clinical variables (ClinVars) in a head and neck (H&N) cancer cohort. This integration of WSI with ClinVars improves PFS predictions.