Less is More: Efficient PET/CT Segmentation and Multimodal Prediction of Recurrence-Free Survival and HPV Status in Head and Neck Cancer

Lishan Cai, Xinglong Liang, Tianyu Zhang, Jiaju Huang, Tao Tan, Yunchao Yin

Keywords: Medical Image Analysis，HECKTOR challenge，Deep Learning，Segmentation，Recurrence-Free Survival，HPV Status

TL;DR: In this study, Team MEDAI presents our solutions for the three HECKTOR 2025 tasks: segmentation, recurrence-free survival prediction, and HPV status classification.

Abstract:

Accurate segmentation of head and neck tumors and metastatic
lymph nodes on PET/CT is essential for radiotherapy planning and prog-
nosis. HECKTOR 2025 addresses this need by benchmarking multimodal
approaches that integrate imaging and clinical data for primary gross tu-
mor volume and involved lymph node segmentation, recurrence-free sur-
vival prediction, and HPV status classification. In this study, we (Team
MEDAI) present our solutions for all three tasks. For automated tumor
segmentation, an ensemble of 10 lightweight STU-Net (small) models
was employed, yielding efficient and accurate delineation of both pri-
mary tumors and metastatic lymph nodes. For recurrence-free survival
prediction and HPV status classification, a multimodal framework was
developed that integrates volumetric PET/CT, lesion masks from the
segmentation models, and structured clinical features. Code is available
at https://github.com/Liiiii2101/HECKTOR2025-MEDAI. Team: MEDAI

Submission Number: 6