面向分割、诊断与预后预测的多模态和多中心头颈癌数据集
计算机视觉与模式识别
2025-09-23 v3
摘要
我们提供一个公开的多模态数据集用于头颈癌研究，包含来自 10 个国际医学中心经组织学确认病例的 1123 例标注的正电导影/CT (PET/CT) 检查。所有检查均包含经过配准的 PET/CT 扫描，采用各异的获取方案，这反映了长期随时间积累的多机构回顾性收集所呈现的真实临床多样性。主要实体肿瘤体积 (Gross Tumor Volume, GTVp) 和涉及的淋巴结 (GTVn) 均由经验丰富的放射治疗专家和放射科医生遵循既定指南进行手动分割。我们提供脱敏的 NifTI 文件、专家标注的分割掩码、详尽的临床元数据以及部分患者的放射治疗剂量分布。元数据包括 TNM 分期、HPV 状态、人口学信息、长期随访结果、生存时间、 censoring 指示器以及治疗信息。为展示数据集的实用性，我们使用 UNet、SegResNet 等最新深度学习模型对其进行基准测试，涵盖自动化肿瘤分割、无复发生存预测和 HPV 状态分类三个关键临床任务。
引用
@article{arxiv.2509.00367,
title = {A Multimodal and Multi-centric Head and Neck Cancer Dataset for Segmentation, Diagnosis and Outcome Prediction},
author = {Numan Saeed and Salma Hassan and Shahad Hardan and Ahmed Aly and Darya Taratynova and Umair Nawaz and Ufaq Khan and Muhammad Ridzuan and Vincent Andrearczyk and Adrien Depeursinge and Yutong Xie and Thomas Eugene and Raphaël Metz and Mélanie Dore and Gregory Delpon and Vijay Ram Kumar Papineni and Kareem Wahid and Cem Dede and Alaa Mohamed Shawky Ali and Carlos Sjogreen and Mohamed Naser and Clifton D. Fuller and Valentin Oreiller and Mario Jreige and John O. Prior and Catherine Cheze Le Rest and Olena Tankyevych and Pierre Decazes and Su Ruan and Stephanie Tanadini-Lang and Martin Vallières and Hesham Elhalawani and Ronan Abgral and Romain Floch and Kevin Kerleguer and Ulrike Schick and Maelle Mauguen and David Bourhis and Jean-Christophe Leclere and Amandine Sambourg and Arman Rahmim and Mathieu Hatt and Mohammad Yaqub},
journal= {arXiv preprint arXiv:2509.00367},
year = {2025}
}
备注
10 pages, 5 figures. Numan Saeed is the corresponding author. Numan Saeed, Salma Hassan and Shahad Hardan contributed equally to this work. Project page: https://hecktor25.grand-challenge.org/