Prostate MRI clinically significant cancer

This data set is part of the public development data for the 2023 Automated Universal Classification Challenge (AUC23). The data set concerns clinically significant prostate cancer classification on bi-parametric magnetic resonance imaging (bpMRI) and was derived from the Prostate Imaging: Cancer AI (PI-CAI) challenge's public training and development dataset (v2.0). The data set was originally introduced and described by Saha et al. (2022), and no images or patient information were added. Data was restructured in compliance with the AUC23 challenge format. The data set contains 1500 anonymized prostate bpMRI scans from 1476 patients acquired between 2012-2021 at three centers (Radboud University Medical Center, University Medical Center Groningen, Ziekenhuis Groep Twente) based in The Netherlands.

Images are 4D tensors:

0: 3D Axial T2-weighted imaging (T2W)

1: 3D Axial apparent diffusion coefficient maps (ADC)

2: 3D Axial high b-value (≥ 1000 s/mm2) diffusion-weighted imaging (HBV)

Classification labels:

0: Benign or indolent prostate cancer

1: Clinically significant prostate cancer (csPCa)

Folder structure:

imagesTr (root folder with all patients and studies)
    ├── 10417_1000424_0000.mha  (axial T2W imaging for study 1000424)
    ├── 10417_1000424_0001.mha  (axial ADC imaging for study 1000424)
    ├── 10417_1000424_0002.mha  (axial HBV imaging for study 1000424)
    ├── ...
    ├── 11251_1001274_0000.mha  (axial T2W imaging for study 1001274)
    ├── 11251_1001274_0001.mha  (axial ADC imaging for study 1001274)
    ├── 11251_1001274_0002.mha  (axial HBV imaging for study 1001274)
    ├── ...

Please cite the following article if you are using the PI-CAI: Public Training and Development Dataset:

A. Saha, J. J. Twilt, J. S. Bosma, B. van Ginneken, D. Yakar, M. Elschot, J. Veltman, J. J. Fütterer, M. de Rooij, H. Huisman, "Artificial Intelligence and Radiologists at Prostate Cancer Detection in MRI: The PI-CAI Challenge (Study Protocol)", DOI: 10.5281/zenodo.6522364