Mediastinal Lymph Node Quantification (LNQ): Segmentation of Heterogeneous CT Data

Accurate lymph node size estimation is critical for staging cancer patients, initial therapeutic management, and in longitudinal scans, assessing response to therapy. Machine learning models for medical image segmentation have shown remarkable progress in recent years, but we know of no tools available to fully quantify lymph nodes. Lymph nodes are often within the intensity profile of normal soft tissue and have ill-defined borders. Additionally, their presentation across subjects can vary significantly, making it difficult to scale from small datasets to a robust tool.

This dataset contains chest CT scans from 540 patients acquired during treatment for various cancer types, curated for the LNQ2023 MICCAI Challenge (https://lnq2023.grand-challenge.org/ ) to help develop new segmentation tools from weakly annotated cases. The dataset contains both partially annotated (i.e., one mediastinal lymph node out of several in the case are segmented) as well as fully annotated (i.e., all mediastinal lymph nodes are segmented) cases and clinical data (gender/cancer type).

准确的淋巴结尺寸估算对于癌症患者的临床分期、初始治疗管理，以及在纵向随访扫描中评估治疗响应均具有关键意义。近年来，面向医学图像分割的机器学习模型已取得显著进展，但目前尚无能够实现淋巴结全量化分析的可用工具。淋巴结通常处于正常软组织的强度分布范围内，且边界模糊不清。此外，不同受试者的淋巴结表现差异显著，这使得从小规模数据集扩展至鲁棒性强的应用工具极具挑战。
本数据集包含540名罹患各类癌症的患者在治疗期间获取的胸部CT扫描影像，系为LNQ2023 国际医学图像计算与计算机辅助干预（MICCAI）挑战赛（https://lnq2023.grand-challenge.org/）精心整理而成，旨在助力从弱标注（weakly annotated）样本中开发新型分割工具。该数据集涵盖部分标注样本（即仅分割了病例中多个纵隔淋巴结中的单个淋巴结）与全标注样本（即完成了所有纵隔淋巴结的分割），同时附带临床数据（性别、癌症类型）。