Abstract
We present HERMES (Hybrid Ensemble for Radiotherapy- target segmentation, Malignancy staging, and Event-free Survival), a single containerized algorithm for the three HECKTOR 2026 subtasks: segmentation of the primary tumor (GTVp) and lymph nodes (GTVn), radiological T/N staging, and recurrence-free survival (RFS) prediction, computed from a paired FDG-PET/CT scan and an electronic health record. A 10-fold ensemble of STU-Net Small networks produces the segmentation, and the predicted mask then drives two downstream tasks. Rather than pass a generic radiomics vector to the staging models, we derive from the predicted masks a compact set of geometry features (nodal count, size, and burden) aligned with the size and number axes of AJCC/UICC 7th-edition radiological T/N staging. On internal cross- validation these features raise N-stage balanced accuracy from 0.691 to 0.720 at lower feature dimensionality (a favorable trend whose paired interval includes zero); read from ground-truth masks they reach 0.897 and beat radiomics by a statistically signicant margin, indicating the gain reects mask-derived geometry rather than a mask artifact and is diluted at deployment by imperfect segmentation. For prognosis we combine complementary deep and clinical risk experts in an equal-weight ensemble, one trained with a concordance-tracking survival loss whose value tracks the concordance index during training. Every component is selected on out-of-fold cross-validation under a regularization-oriented protocol. On the HECKTOR 2026 validation set, HERMES attained a Mean Dice of 0.641, T and N balanced accuracies of 0.580 and 0.642, and an RFS C-index of 0.679, and qualied for the testing phase. Team: AMC_HNC.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/HECKTOR_001.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=ZVilrR6nSk
BibTex
@InProceedings{WanKai_HERMES_MICCAISAT2026,
author = { Wang, Kai AND Chen, Meixu AND Nasr, Elie AND Lanning, Ryan AND Miften, Moyed},
title = { { HERMES: A Hybrid Ensemble for Head-and-Neck Tumor Segmentation, TN Staging, and Recurrence-Free Survival on PET/CT } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17271},
month = {pending},
page = {pending}
}
