Breast Metastases to Axillary Lymph Nodes

The detection of breast cancer metastases to lymph nodes is of great prognostic value for patient treatment. Using machine learning to detect metastatic breast cancer to lymph nodes can increase efficiency of pathologist diagnosis and ultimately ensure patients are accurately staged for prospective treatment. This dataset allows for the objective comparison of breast cancer metastases detection algorithms.The dataset consists of 130 de-identified whole slide images of H&E stained axillary lymph node specimens from 78 patients. Metastatic breast carcinoma is present in 36 of the WSI from 27 patients. No patient inclusion/exclusion criteria were followed. No slide inclusion/exclusion criteria were followed. The slides were scanned at Memorial Sloan Kettering Cancer Center (MSKCC) with Leica Aperio AT2 scanners at 20x equivalent magnification (0.5 microns per pixel). Together with the slides, the class label of each slide, either positive or negative for breast carcinoma, is given. The slide class label was obtained from the pathology report of the respective case.

检测腋窝淋巴结乳腺癌转移灶，对患者的治疗方案制定与预后评估均具有重要临床价值。借助机器学习技术实现淋巴结乳腺癌转移灶检测，可提升病理医师的诊断效率，最终确保患者能获得精准的临床分期以开展针对性后续治疗。本数据集可用于乳腺癌转移灶检测算法的客观性能对比。
本数据集包含78例患者的130份去标识化全玻片图像（whole slide image, WSI），均为苏木精-伊红（H&E）染色的腋窝淋巴结组织标本。其中27例患者的36份全玻片图像中存在乳腺癌转移灶。本数据集未设置任何患者入组/排除标准，亦未对玻片设定入组/排除规则。
所有玻片均由纪念斯隆凯特琳癌症中心（Memorial Sloan Kettering Cancer Center, MSKCC）采用徕卡Aperio AT2扫描仪进行扫描，扫描参数为等效20倍放大倍率（每像素对应0.5微米）。随数据集一同提供的还有每份玻片的分类标签，即乳腺癌转移灶检测结果为阳性或阴性，该标签源自对应病例的病理报告。

搜集汇总

背景与挑战

背景概述

该数据集专注于乳腺癌转移至腋窝淋巴结的检测，包含130张H&E染色全切片图像，来自78名患者，其中36张图像显示转移性乳腺癌，用于机器学习算法开发以辅助病理诊断和患者分期。图像以20倍放大扫描，并附带基于病理报告的阳性或阴性标签，旨在客观比较检测算法的性能。