Datasets:

Angelou0516
/

PI-CAI

patient_id stringlengths 5 5	study_id stringlengths 7 7	case stringlengths 13 13	center stringclasses 3 values	patient_age float32 35 92	psa float32 0.1 224 ⌀	psad float32 0.01 7 ⌀	prostate_volume float32 4 308 ⌀	case_ISUP int32 0 5	case_csPCa stringclasses 2 values	lesion_GS stringclasses 112 values	histopath_type stringclasses 5 values	num_slices_t2w int32 17 45
10000	1000000	10000_1000000	PCNN	73	7.7	null	55	0	NO	0+0	MRBx	31
10001	1000001	10001_1000001	RUMC	64	8.7	0.09	102	0	NO			21
10002	1000002	10002_1000002	ZGT	58	4.2	0.06	74	0	NO			22
10003	1000003	10003_1000003	ZGT	72	13	null	71.5	0	NO	0+0	SysBx	23
10004	1000004	10004_1000004	RUMC	67	8	0.1	78	0	NO	0+0,0+0	SysBx+MRBx	21
10005	1000005	10005_1000005	RUMC	64	12.1	0.24	51	3	YES	4+3,0+0	MRBx	19
10006	1000006	10006_1000006	ZGT	73	6.2	0.23	27	1	NO	0+0,3+3	SysBx+MRBx	21
10007	1000007	10007_1000007	RUMC	68	3.83	0.09	41	1	NO	3+3	SysBx+MRBx	19
10008	1000008	10008_1000008	ZGT	81	11.1	0.2	56	3	YES	4+3,3+4	SysBx+MRBx	21
10009	1000009	10009_1000009	PCNN	65	24	null	120	0	NO	0+0	SysBx	27
10010	1000010	10010_1000010	RUMC	67	6.19	0.08	76	0	NO			19
10011	1000011	10011_1000011	PCNN	60	7.1	null	34	1	NO	3+3,0+0	MRBx	27
10012	1000012	10012_1000012	RUMC	64	9.9	0.14	70	2	YES	3+4	MRBx	23
10013	1000013	10013_1000013	ZGT	64	4.5	0.1	52	2	YES	3+4	RP	24
10014	1000014	10014_1000014	RUMC	48	null	0.07	62	0	NO			22
10015	1000015	10015_1000015	PCNN	64	23	null	50	0	NO	0+0	MRBx	27
10016	1000016	10016_1000016	RUMC	60	8.5	0.13	71	0	NO			19
10017	1000017	10017_1000017	RUMC	67	5.5	null	null	0	NO			19
10018	1000018	10018_1000018	PCNN	70	8.8	null	35	1	NO	3+3	MRBx	23
10019	1000019	10019_1000019	RUMC	61	10	0.24	42	2	YES	N/A,N/A,N/A,3+4	MRBx	19
10020	1000020	10020_1000020	RUMC	43	4.6	0.11	47	0	NO			21
10021	1000021	10021_1000021	RUMC	61	23	0.39	60	3	YES	4+3	MRBx	19
10022	1000022	10022_1000022	PCNN	56	5	null	25	1	NO	3+3	MRBx	27
10023	1000023	10023_1000023	RUMC	62	1.5	0.03	37	0	NO	0+0	MRBx	19
10024	1000024	10024_1000024	RUMC	65	5.5	0.17	35	0	NO			19
10025	1000025	10025_1000025	RUMC	37	4.1	0.17	24	0	NO			19
10026	1000026	10026_1000026	RUMC	54	4.06	0.11	35	1	NO	3+3	SysBx+MRBx	19
10027	1000027	10027_1000027	PCNN	63	7.4	null	37	0	NO	0+0	MRBx	25
10028	1000028	10028_1000028	RUMC	64	9.99	0.09	117	0	NO			23
10029	1000029	10029_1000029	RUMC	64	8.9	0.2	46	5	YES	3+4,4+5,3+3	SysBx+MRBx	19
10030	1000030	10030_1000030	ZGT	70	6	0.08	72	0	NO			21
10031	1000031	10031_1000031	RUMC	67	11.4	0.15	78	0	NO			19
10032	1000032	10032_1000032	RUMC	73	19.18	0.34	57	3	YES	4+3,0+0	MRBx	19
10033	1000033	10033_1000033	RUMC	61	9.3	0.24	38	0	NO			19
10034	1000034	10034_1000034	ZGT	66	4.6	0.11	47	1	NO	3+3,0+0,0+0	SysBx	19
10035	1000035	10035_1000035	PCNN	70	3.59	null	19	1	NO	3+3	MRBx	27
10036	1000036	10036_1000036	RUMC	69	7	0.17	40.490002	2	YES	3+4,3+3	MRBx	19
10037	1000037	10037_1000037	RUMC	38	2.9	0.11	26	0	NO			19
10038	1000038	10038_1000038	RUMC	62	9.9	0.2	50	1	NO	3+3	MRBx	21
10039	1000039	10039_1000039	RUMC	66	10	0.09	109	0	NO			19
10040	1000040	10040_1000040	RUMC	78	55	1.25	44	2	YES	3+4,N/A,N/A	MRBx	21
10041	1000041	10041_1000041	PCNN	74	22	null	30	1	NO	3+3	SysBx	27
10042	1000042	10042_1000042	ZGT	67	4.8	0.08	60	1	NO	3+3,0+0	SysBx	22
10043	1000043	10043_1000043	RUMC	77	38	null	48	3	YES	4+3	MRBx	21
10044	1000044	10044_1000044	ZGT	71	6.4	0.41	16	4	YES	3+4,4+4	SysBx+MRBx	21
10045	1000045	10045_1000045	ZGT	51	7.9	null	67	0	NO			23
10046	1000046	10046_1000046	RUMC	60	null	null	37.700001	0	NO			21
10047	1000047	10047_1000047	ZGT	74	7.4	null	164	0	NO			29
10048	1000048	10048_1000048	RUMC	64	20	0.65	31	3	YES	4+3	MRBx	19
10049	1000049	10049_1000049	RUMC	68	18.200001	null	157	0	NO			23
10050	1000050	10050_1000050	RUMC	65	6	0.41	31	2	YES	3+4	MRBx	19
10051	1000051	10051_1000051	ZGT	68	4.8	0.07	72	0	NO	0+0	SysBx	25
10052	1000052	10052_1000052	RUMC	69	5.1	0.07	97	1	NO	2+3	MRBx	21
10053	1000053	10053_1000053	ZGT	64	4.3	0.1	48	2	YES	3+4	RP	21
10054	1000054	10054_1000054	RUMC	68	23	0.26	88	0	NO			25
10055	1000055	10055_1000055	RUMC	60	2.3	null	47	0	NO			21
10056	1000056	10056_1000056	ZGT	69	6.6	0.13	50	0	NO	0+0	SysBx+MRBx	23
10057	1000057	10057_1000057	RUMC	62	7.1	0.15	47	0	NO			19
10058	1000058	10058_1000058	PCNN	69	9	null	60	0	NO	0+0	SysBx	27
10059	1000059	10059_1000059	RUMC	69	52	0.85	61	2	YES	3+4	MRBx	23
10060	1000060	10060_1000060	RUMC	64	7.2	0.1	67	0	NO	0+0	MRBx	21
10061	1000061	10061_1000061	RUMC	62	4.9	0.08	60	0	NO			21
10062	1000062	10062_1000062	ZGT	71	15.2	0.14	112	1	NO	3+3	SysBx+MRBx	26
10063	1000063	10063_1000063	RUMC	70	18.4	0.26	69	1	NO	3+3,N/A	MRBx	21
10064	1000064	10064_1000064	RUMC	66	13	0.12	104	0	NO	0+0	SysBx+MRBx	23
10065	1000065	10065_1000065	PCNN	60	14	null	85	1	NO	3+3	SysBx	27
10066	1000066	10066_1000066	ZGT	64	6.5	0.19	33	0	NO			23
10067	1000067	10067_1000067	PCNN	63	7.2	null	110	0	NO	0+0	SysBx	27
10068	1000068	10068_1000068	PCNN	60	7.4	null	65	0	NO	0+0	SysBx	27
10069	1000069	10069_1000069	RUMC	68	6.3	0.05	124	0	NO	0+0	MRBx	19
10070	1000070	10070_1000070	RUMC	58	5.8	0.08	74	0	NO			19
10071	1000071	10071_1000071	ZGT	50	9.9	0.18	55	0	NO	0+0	SysBx	21
10072	1000072	10072_1000072	RUMC	70	8.8	0.09	98	0	NO	0+0,0+0	MRBx	23
10073	1000073	10073_1000073	RUMC	73	3.5	0.01	99	1	NO	3+3	MRBx	23
10074	1000074	10074_1000074	PCNN	60	11	null	45	2	YES	3+4	MRBx	27
10075	1000075	10075_1000075	RUMC	68	21	0.27	79	0	NO	0+0	MRBx	19
10076	1000076	10076_1000076	RUMC	61	14	0.2	69	0	NO			21
10077	1000077	10077_1000077	RUMC	69	7.98	0.22	37	0	NO			19
10078	1000078	10078_1000078	RUMC	51	7.6	0.14	56	4	YES	4+4	MRBx	19
10079	1000079	10079_1000079	PCNN	76	6.7	null	28	3	YES	4+3	MRBx	25
10080	1000080	10080_1000080	PCNN	64	9.4	null	69	1	NO	0+0,3+3	MRBx	25
10081	1000081	10081_1000081	RUMC	49	8.9	0.08	114	0	NO			23
10082	1000082	10082_1000082	RUMC	65	null	null	63	1	NO	3+3	MRBx	19
10083	1000083	10083_1000083	ZGT	72	3.7	null	33	0	NO	0+0	SysBx	19
10084	1000084	10084_1000084	ZGT	48	4	null	33	0	NO	0+0	SysBx	21
10085	1000085	10085_1000085	PCNN	59	8.3	null	65	3	YES	4+3	MRBx	27
10086	1000086	10086_1000086	RUMC	52	7.4	0.25	28	0	NO	0+0	SysBx+MRBx	19
10087	1000087	10087_1000087	RUMC	67	7.3	0.09	77	0	NO			21
10088	1000088	10088_1000088	RUMC	62	4.1	0.03	130	0	NO			25
10089	1000089	10089_1000089	RUMC	58	4	0.09	44.779999	0	NO			21
10090	1000090	10090_1000090	ZGT	65	5.2	null	41	1	NO	0+0,3+3	SysBx+MRBx	27
10091	1000091	10091_1000091	ZGT	59	4.7	0.1	47	0	NO			23
10092	1000092	10092_1000092	ZGT	62	6.7	0.18	37	1	NO	0+0,3+3	SysBx+MRBx	17
10093	1000093	10093_1000093	PCNN	75	7.4	null	35	1	NO	3+3	MRBx	25
10094	1000094	10094_1000094	RUMC	62	10.24	0.27	38	5	YES	4+5,0+0	SysBx+MRBx	19
10095	1000095	10095_1000095	PCNN	65	9.2	null	60	0	NO	0+0	MRBx	27
10096	1000096	10096_1000096	RUMC	64	22.33	0.66	34	0	NO	0+0	MRBx	19
10097	1000097	10097_1000097	PCNN	66	15.01	null	32	2	YES	3+4	MRBx	27
10098	1000098	10098_1000098	RUMC	69	12	0.17	69	0	NO			19
10099	1000099	10099_1000099	RUMC	64	5.5	0.05	120	0	NO			23

End of preview. Expand in Data Studio

PI-CAI: Prostate Imaging - Cancer AI Challenge (Public Training & Development)

The PI-CAI 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 challenge targets clinically significant prostate cancer (csPCa, ISUP ≥ 2) detection and segmentation.

Dataset Summary

Field	Details
Modality	Biparametric MRI: axial/coronal/sagittal T2W + axial high-b-value (≥ 1000 s/mm²) DWI + axial ADC
Body Part	Prostate (whole gland + csPCa lesions)
Cases	1,500 studies / 1,476 patients
Class composition	1,075 benign (ISUP ≤ 1) / 425 csPCa-positive (ISUP ≥ 2)
Total Size	~26 GB (images) + 139 MB (labels)
License	CC BY-NC 4.0

Data Structure

images/
  {patient_id}/
    {patient_id}_{study_id}_t2w.mha   # axial T2W
    {patient_id}_{study_id}_cor.mha   # coronal T2W (occasionally missing)
    {patient_id}_{study_id}_sag.mha   # sagittal T2W (occasionally missing)
    {patient_id}_{study_id}_adc.mha   # axial ADC
    {patient_id}_{study_id}_hbv.mha   # axial high-b-value DWI
labels/
  csPCa_lesion_delineations/
    human_expert/
      original/         # native-resolution expert csPCa lesion masks (1,295 cases)
      resampled/        # SAME masks resampled to T2W axial geometry (1,295 cases)
      Pooch25/          # 2025 expert annotations for the 205 originally-only-AI-labeled cases
    AI/
      Bosma22a/         # AI-derived csPCa lesion masks (binary, 1,500 cases)
  anatomical_delineations/
    whole_gland/
      AI/Bosma22b/      # AI whole-prostate-gland masks (binary, 1,500 cases) — only gland labels
      AI/Guerbet23/     # alternate AI gland masks
    zonal_pz_tz/
      AI/HeviAI23/      # AI zonal masks (peripheral zone / transition zone)
      AI/Yuan23/        # alternate AI zonal masks
  clinical_information/
    marksheet.csv       # per-case clinical info (PSA, ISUP, center, etc.)
  additional_resources/
    ProstateX-mapping.json
marksheet.csv           # convenience copy at root

Recommended Ground Truth

For csPCa lesion segmentation, use:

labels/csPCa_lesion_delineations/human_expert/resampled/   # 1,295 cases
labels/csPCa_lesion_delineations/human_expert/Pooch25/     # +205 cases (= 1,500 total expert masks)

These are produced by trained investigators under three expert radiologists' supervision, carry granular multi-class ISUP labels (0/2/3/4/5), and are aligned to T2W axial geometry. The PI-CAI organizers explicitly recommend them over the AI-derived Bosma22a masks, which exist primarily as a complementary semi-supervised label source for the originally unannotated 205 positive cases (now superseded by Pooch25 for those cases).

For whole-prostate-gland segmentation, only AI-derived masks are available (anatomical_delineations/whole_gland/AI/Bosma22b/); the README in the picai_labels repo cautions these "can be susceptible to errors or faulty segmentations."

Important Caveats

Modalities are NOT co-registered. T2W (axial/cor/sag), DWI, and ADC are acquired at different resolutions and not co-registered across sequences (only 54 / 9,107 cases were manually co-registered in the full corpus). Models must handle this.
ADC absolute intensities are NOT standardized across centers — do not treat as universal.
Splits: The 1,500 public cases are physically distributed across 5 cross-validation folds (fold0–fold4) in the original Zenodo release; here they are merged into a single images/ directory. There is no fixed train/val/test split.

Citation

@article{saha2024picai,
  title   = {Artificial intelligence and radiologists in prostate cancer detection on MRI (PI-CAI): an international, paired, non-inferiority, confirmatory study},
  author  = {Saha, Anindo and Bosma, Joeran S and Twilt, Jasper J and van Ginneken, Bram and Bjartell, Anders and Padhani, Anwar R and others},
  journal = {The Lancet Oncology},
  volume  = {25},
  number  = {7},
  pages   = {879--887},
  year    = {2024},
  doi     = {10.1016/S1470-2045(24)00220-1}
}

@dataset{picai_public_training_2022,
  title     = {The PI-CAI Challenge: Public Training and Development Dataset},
  author    = {Saha, Anindo and Twilt, Jasper J and Bosma, Joeran S and others},
  year      = 2022,
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.6624726},
  url       = {https://doi.org/10.5281/zenodo.6624726}
}

Sources

Original imaging: https://zenodo.org/records/6624726 (Zenodo, CC BY-NC 4.0)
Original labels: https://github.com/DIAGNijmegen/picai_labels
Challenge: https://pi-cai.grand-challenge.org/
Preprocessing toolkit: https://pypi.org/project/picai-prep/
Baseline code: https://github.com/DIAGNijmegen/picai_baseline

Downloads last month: 48