PROSTATEx
SPIE-AAPM-NCI PROSTATEx Challenge Data
Multiparametric prostate MRI (T2w, proton density, DCE and DWI with ADC) of 346 men from Radboudumc on Siemens 3T scanners, with lesion positions and clinical significance labels from the PROSTATEx and PROSTATEx-2 challenges.
Overview
PROSTATEx is a retrospective collection of multiparametric prostate MRI from Radboud University Medical Center in Nijmegen, the Netherlands, curated for computer-aided diagnosis research and hosted by The Cancer Imaging Archive. It was the data of two challenges: PROSTATEx (SPIE Medical Imaging 2017), which asked whether a marked lesion is clinically significant cancer, and PROSTATEx-2 (AAPM 2017), which asked for the Gleason Grade Group of a lesion. TCIA marks the collection as superseded by PI-CAI, whose public training set contains all PROSTATEx cases.
Composition
The collection holds 346 subjects with 349 studies, 18,321 DICOM series and 309,251 images, about 16.7 GB in total. The PROSTATEx challenge split 204 subjects into a training cohort and 140 into a test cohort. The PROSTATEx-2 subset has 112 training and 70 test findings. Besides the DICOM images, each case has a Ktrans map in MetaImage format, derived from the DCE series, plus thumbnails of the marked lesions. In the findings tables, the training cohort lists 330 findings, 76 of them clinically significant, and the test reference standard lists 208 findings, 48 of them clinically significant.
Acquisition
All studies include T2-weighted turbo spin echo images (about 0.5 mm in plane, 3.6 mm slices), a proton density weighted image, a 3D turbo FLASH DCE series (about 1.5 mm in plane, 4 mm slices, 3.5 s temporal resolution) and single-shot echo planar DWI with b-values 50, 400 and 800 in three directions, with an ADC map from the scanner software. Two Siemens 3T scanners were used, the MAGNETOM Trio and the Skyra, without an endorectal coil.
Annotations
Each finding is given as a point in scanner coordinates with its prostate zone. For PROSTATEx, a finding is clinically significant when the biopsy Gleason score was 7 or higher; PI-RADS 2 findings were not biopsied and count as not significant. PROSTATEx-2 labels each finding with its Gleason Grade Group. There are no voxel segmentations in the collection itself.
Known limitations
- All data come from one center and one vendor.
- Labels are lesion points, not outlines, and unbiopsied PI-RADS 2 findings are assumed benign.
- No age or other clinical variables are released.
Cohort
Aggregate numbers from the sources below. Bars are relative to the 346 subjects.
Contrast combinations
How many subjects have exactly each set of contrasts.
| T2w | PDw | dwi | ADC | dce | Subjects with exactly this set |
|---|---|---|---|---|---|
346 |
Contrast / sequence
subjects, values can overlap
- T2-weighted 346 100%
- Proton density weighted 346 100%
- DCE perfusion 346 100%
- Diffusion-weighted 346 100%
- ADC map 346 100%
Scanner vendor
subjects
- Siemens Healthineers 346 100%
Field strength
subjects
- 3 T 346 100%
Country
subjects
- Netherlands 346 100%
Split
subjects
- Training 204 59%
- Test 140 40%
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
Download without an account
DICOM images download with the TCIA Data Retriever or through the NCI Imaging Data Commons. Ktrans maps, lesion information and thumbnails download through IBM Aspera or per challenge cohort from the collection page.
Creative Commons Attribution 3.0 Unported
Older version of CC BY. 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
Litjens G, Debats O, Barentsz J, Karssemeijer N, Huisman H. (2017). SPIE-AAPM PROSTATEx Challenge Data (Version 2) [dataset]. The Cancer Imaging Archive. https://doi.org/10.7937/K9TCIA.2017.MURS5CL. Publication: Litjens G, Debats O, Barentsz J, Karssemeijer N, Huisman H. Computer-Aided Detection of Prostate Cancer in MRI. IEEE Transactions on Medical Imaging 33(5):1083-1092 (2014). https://doi.org/10.1109/tmi.2014.2303821
All numbers
Every number on this page, as stored in stats.csv, with its source.
| Measure | Breakdown | Value | Source |
|---|---|---|---|
| Subjects | total | 346 | tcia-prostatex Data Access table |
| Studies | total | 349 | tcia-prostatex Data Access table |
| Scans | total DICOM series | 18,321 | tcia-prostatex Data Access table |
| Images | total DICOM images (slices) | 309,251 | tcia-prostatex Data Access table |
| Subjects | split=train PROSTATEx training cohort | 204 | tcia-prostatex Detailed Description, PROSTATEx Challenge |
| Subjects | split=test PROSTATEx test cohort; reference standard released in version 2 | 140 | tcia-prostatex Detailed Description, PROSTATEx Challenge |
| Subjects | contrast=T2w the page states all studies include T2w | 346 | tcia-prostatex Image Acquisition Details |
| Subjects | contrast=PDw the page states all studies include proton density weighted imaging | 346 | tcia-prostatex Image Acquisition Details |
| Subjects | contrast=dce the page states all studies include DCE | 346 | tcia-prostatex Image Acquisition Details |
| Subjects | contrast=dwi the page states all studies include DWI (b = 50, 400, 800) | 346 | tcia-prostatex Image Acquisition Details |
| Subjects | contrast=ADC ADC map computed by the scanner software from the DWI | 346 | tcia-prostatex Image Acquisition Details |
| Subjects | contrast_set=ADC+dce+dwi+PDw+T2w | 346 | tcia-prostatex Image Acquisition Details |
| Subjects | field_strength=3 Siemens MAGNETOM Trio and Skyra | 346 | tcia-prostatex Image Acquisition Details |
| Subjects | vendor=siemens | 346 | tcia-prostatex Image Acquisition Details |
| Subjects | country=NL all imaging at Radboudumc in Nijmegen | 346 | tcia-prostatex Acknowledgements |
Sources
The keys used in the table above.
- tcia-prostatex TCIA PROSTATEx collection page (version 2, updated 2022/07/05) website