HEad and neCK TumOR (HECKTOR) 2025: Benchmark of Segmentation, Diagnosis, and Prognosis in Multimodal PET/CT

The HECKTOR 2025 challenge has produced a benchmark for automated head and neck cancer analysis, using multimodal PET/CT imaging and electronic health records from more than 1,100 patients across 10 centers worldwide [1][2]. Head and neck cancers carry a significant global health burden, and precise tumor delineation is critical for effective radiotherapy planning [1]. Manual segmentation of the oropharyngeal anatomy is time-intensive and prone to inter-observer variability, while predicting recurrence-free survival and classifying human papillomavirus (HPV) status from noninvasive imaging remain difficult clinical objectives [1][2]. The challenge, detailed in a paper posted to the arXiv preprint repository on June 18, 2026, builds on earlier editions held in 2020 and 2022 [1][2]. arXiv, founded in 1991, is an open-access repository that hosts electronic preprints across physics, computer science, and related fields, and now receives roughly 24,000 submissions per month [6]. Thirty-five teams registered for HECKTOR 2025, and 15 submitted final algorithms for evaluation on a held-out test set [1][2]. Participants addressed three tasks: segmenting primary gross tumor volumes and metastatic lymph nodes, predicting recurrence-free survival, and classifying HPV status [1][2]. The top-performing segmentation algorithms achieved a mean Dice similarity coefficient of 0.75 [1][2]. For survival prediction, the best models reached a concordance index of 0.66, while HPV classification yielded a balanced accuracy of 0.56 [1][2]. The challenge organizers analyzed submitted methodologies across different lesion characteristics and discussed implications for clinical translation into automated oncology workflows and decision support systems [1][2]. The paper appears on arXiv as a preprint, meaning it has undergone moderation but not formal peer review [6]. The repository’s Labs framework allows community collaborators to build tools such as bibliographic explorers and code finders that sit alongside article pages, though new Labs proposals are temporarily paused while arXiv migrates its systems to the cloud [3][4][5].