Angelou0516/PI-CAI
收藏资源简介:
PI-CAI数据集(前列腺影像-癌症AI挑战公共训练与开发数据集)包含来自1,476名患者的1,500个双参数MRI(bpMRI)研究,这些数据采集于2012至2021年间的四个荷兰中心(RUMC、ZGT、PCNN、UMCG)。该数据集主要用于临床显著性前列腺癌(csPCa,ISUP ≥ 2)的检测和分割。数据模态包括轴向/冠状/矢状T2W、轴向高b值(≥ 1000 s/mm²)DWI和轴向ADC。数据集包含1,075例良性(ISUP ≤ 1)和425例csPCa阳性(ISUP ≥ 2)病例,总大小约为26 GB(图像)和139 MB(标签)。数据集结构包括图像和标签,标签部分包含人类专家和AI生成的前列腺癌病灶分割掩模、全前列腺腺体分割掩模以及临床信息。需要注意的是,数据集中的模态未进行配准,ADC强度未标准化,且没有固定的训练/验证/测试集划分。数据集采用CC BY-NC 4.0许可。
The PI-CAI (Prostate Imaging - Cancer AI Challenge) Public Training & Development dataset contains 1,500 biparametric MRI (bpMRI) studies from 1,476 patients acquired at four Dutch centers (RUMC, ZGT, PCNN, UMCG) between 2012 and 2021. The dataset is designed for clinically significant prostate cancer (csPCa, ISUP ≥ 2) detection and segmentation. It includes axial/coronal/sagittal T2W, axial high-b-value (≥ 1000 s/mm²) DWI, and axial ADC modalities. The dataset comprises 1,075 benign (ISUP ≤ 1) and 425 csPCa-positive (ISUP ≥ 2) cases, with a total size of approximately 26 GB (images) and 139 MB (labels). The data structure includes images and labels, with the latter containing human expert and AI-generated csPCa lesion segmentation masks, whole-prostate-gland segmentation masks, and clinical information. Important caveats include the lack of co-registration across modalities, non-standardized ADC absolute intensities across centers, and no fixed train/val/test split. The dataset is licensed under CC BY-NC 4.0.




