ccRCC_WHO_data

Clear cell renal cell carcinoma is the most prevalent and aggressive subtype of kidney cancer. It is characterized by malignant proliferation of epithelial cells in the proximal convoluted tubules and is driven by specific genetic alterations. Therefore, the histological analysis of pathological images plays a critical role in enabling definitive diagnosis, accurate severity grading, identification of sarcomatoid features, and guidance for treatment strategies such as targeted therapy and immunotherapy. To support such analyses, we introduce a comprehensive pathology image dataset consisting of 2248 whole-slide images from 403 patients, along with an Excel sheet containing seven key clinical parameters including patient ID, gender, age, WHO/ISUP grade, necrosis status, lymphovascular invasion, and renal confinement. We further evaluated the utility of this dataset using twelve mainstream deep learning models for the classification of the images into four WHO/ISUP nuclear grade categories.

透明细胞肾细胞癌（clear cell renal cell carcinoma）是肾癌中发病率最高且侵袭性最强的亚型，其特征为近端曲小管上皮细胞发生恶性增殖，且由特定的遗传改变驱动。因此，病理图像的组织学分析对于实现明确诊断、精准病情分级、识别肉瘤样特征以及指导靶向治疗、免疫治疗等治疗策略至关重要。为支持此类分析，本研究构建了一套全面的病理图像数据集，该数据集包含来自403例患者的2248张全切片图像（whole-slide images），并附带一份Excel表格，其中涵盖7项关键临床参数，包括患者ID、性别、年龄、WHO/ISUP分级、坏死状态、脉管浸润情况及肾局限状态。本研究进一步采用12种主流深度学习模型，对该数据集的应用价值进行了评估，以实现将图像分类为4种WHO/ISUP细胞核分级类别的任务。