HEad and neCK TumOR (HECKTOR) 2025: Benchmark of Segmentation, Diagnosis, and Prognosis in Multimodal PET/CT
Accurate segmentation of head and neck tumors and non-invasive imaging-based prediction of recurrence risk and HPV status remain challenging. This study establishes the first comprehensive benchmark integrating multimodal PET/CT scans and electronic health records from over 1,100 patients across multiple centers, simultaneously advancing three core tasks: automated tumor segmentation, survival prognosis prediction, and HPV status classification. The proposed deep learning framework fuses imaging and clinical data, achieving a Dice coefficient of 0.75 for segmentation, a concordance index (C-index) of 0.66 for survival prediction, and a balanced accuracy of 0.56 for HPV classification on the test set. These results lay a robust foundation for clinically translatable artificial intelligence systems in head and neck oncology.