I2CVB prostate mp-MRI
I2CVB multi-parametric MRI dataset of the prostate
Multi-parametric 3 T prostate MRI (T2-weighted, DCE, DWI with ADC maps, MR spectroscopic imaging) of 19 patients with raised PSA and biopsy results, with radiologist masks of the prostate, its zones and the cancer. Files on Zenodo are restricted.
Overview
The I2CVB prostate dataset is a multi-parametric MRI collection built by the Initiative for Collaborative Computer Vision Benchmarking, a group from the Universitat de Girona and the Université de Bourgogne. It was put together to develop and compare computer-aided detection and diagnosis of prostate cancer, and it was used in the group's PhD thesis and papers on DCE-MRI normalization and multi-parametric classification. The images are deposited on Zenodo, where the record carries a CC BY 4.0 label but its files are restricted.
Composition
The thesis describes the 3 T set as 19 patients referred for a raised prostate-specific antigen level, all of whom had a guided biopsy. 17 had biopsy-proven cancer: 12 in the peripheral zone, 3 in the central gland and 2 in both. The other 2 had negative biopsies and are treated as healthy. Each patient has T2-weighted MRI, dynamic contrast-enhanced MRI, diffusion-weighted MRI with an ADC map, and MR spectroscopic imaging. The website states that images are shared as DICOM and the Siemens spectroscopy as RDA files.
Acquisition
All 3 T scans come from one Siemens Magnetom Trio TIM. T2-weighted imaging uses 3D fast spin echo in an oblique axial plane with 1.25 mm slices. DCE-MRI is a fat-suppressed 3D T1 VIBE with 16 partitions of 3.5 mm, one volume every 6 s for about 5 minutes after a Gd-DTPA bolus. DWI is single-shot spin-echo EPI at b = 100 and 800 s/mm², with the ADC map made on the scanner workstation. Spectroscopy uses a PRESS sequence tuned for choline and citrate. The thesis also describes a 1.5 T GE Signa protocol with an endorectal coil, which the website lists as coming soon.
Annotations
An experienced radiologist outlined the prostate on T2-weighted, DCE and ADC images, and the peripheral zone, central gland and cancer on the T2-weighted images. The website names four label classes: prostate, peripheral zone, central gland and cancer.
Known limitations
- Files are not downloadable from Zenodo without permission from the record owner, and no request procedure is published.
- Very small, single-scanner cohort with only 2 biopsy-negative patients.
- Cohort size differs between sources: the thesis reports 19 patients, while a DCE-MRI preprint on a subset from the same scanner reports 20.
- The 1.5 T GE part has no published patient count.
- No age or other demographic data are published.
Cohort
Aggregate numbers from the sources below. Bars are relative to the 19 subjects.
Condition
subjects, values can overlap
- Prostate cancer 17 89%
- Healthy control 2 11%
Scanner vendor
subjects
- Siemens Healthineers 19 100%
Field strength
subjects
- 3 T 19 100%
License and access
Our reading of the license, not legal advice. Before you use the data, read the original license and confirm that your use is allowed. We take no responsibility for how you use a dataset. Full disclaimer
Ask the authors or the data holder. No published process or terms, and access is at their discretion
The Zenodo record is public but its files are restricted and the record does not accept access requests, so the data have to be requested from the authors. The record is labelled CC BY 4.0. A 1.5 T GE subset is announced on the website as coming soon.
Creative Commons Attribution 4.0 International
Use, share and adapt the data for any purpose, including commercial use, as long as you credit the creators.
What you can do
- Yes
- Yes
- Yes
- Yes
What you can share
- Yes
- Yes
- Yes
What you must do
- Yes
- Share alike No
- No
- No
- Manuscript review No
- Release code No
- Return results No
- Delete after use No
Limits
- No
- Location limits No
Citation
G. Lemaitre, R. Marti, J. Freixenet, J. C. Vilanova, P. M. Walker, and F. Meriaudeau, Computer-Aided Detection and Diagnosis for prostate cancer based on mono and multi-parametric MRI: A Review, Computers in Biology and Medicine, vol. 60, pp. 8-31, 2015. doi:10.1016/j.compbiomed.2015.02.009
All numbers
Every number on this page, as stored in stats.csv, with its source.
| Measure | Breakdown | Value | Source |
|---|---|---|---|
| Subjects | total 3 T Siemens set; a DCE-MRI preprint by the same group describes 20 patients from the same scanner | 19 | lemaitre2016-thesis Section 4.2.2, p. 96 |
| Subjects | condition=prostate_cancer biopsy-proven; 12 peripheral zone, 3 central gland, 2 both | 17 | lemaitre2016-thesis Section 4.2.2, p. 96 |
| Subjects | condition=healthy negative biopsy | 2 | lemaitre2016-thesis Section 4.2.2, p. 96 |
| Subjects | field_strength=3 | 19 | lemaitre2016-thesis Section 4.2.2, p. 96 |
| Subjects | vendor=siemens Siemens Magnetom Trio TIM | 19 | lemaitre2016-thesis Section 4.2.2, p. 96 |
Sources
The keys used in the table above.
- i2cvb-website I2CVB website, Prostate section website
- zenodo-162231 Original multi-parametric MRI images of prostate, Zenodo record 162231 website
- lemaitre2016-thesis Lemaitre 2016, PhD thesis, Computer-aided diagnosis for prostate cancer using multi-parametric MRI paper
- lemaitre2017-dce Lemaitre et al., Automatic prostate cancer detection through DCE-MRI images, preprint (Feb 2017) paper