Datasets:

MedOtter
/

SOOP

subject_id stringlengths 5 8	split stringclasses 1 value	annotation_status stringclasses 2 values	tier_shown stringclasses 2 values	slice_rule stringclasses 2 values	slice_idx int64 2 30	n_slices int64 21 32	grid listlengths 3 3	spacing listlengths 3 3	n_fg_tier_shown int64 9 129k ⌀	evaluable_lesionAcute bool 2 classes	adc_usable bool 2 classes	qc_flags listlengths 0 3	in_participants_tsv bool 2 classes	sex stringclasses 2 values	age stringclasses 68 values	race stringclasses 2 values	acuteischaemicstroke stringclasses 1 value	priorstroke stringclasses 2 values	bmi stringlengths 2 6 ⌀	nihss stringclasses 35 values	gs_rankin_6isdeath stringclasses 7 values	etiology stringclasses 5 values
sub-1	train	annotated	lesionAcute	max_lesion_area	17	26	[ 256, 256, 26 ]	[ 0.9375, 0.9375, 6 ]	3,724	true	true	[]	false	null	null	null	null	null	null	null	null	null
sub-2	train	annotated	lesionAcute	max_lesion_area	12	26	[ 432, 432, 26 ]	[ 0.5556, 0.5556, 6 ]	44,765	true	true	[]	true	M	78	w	1	0	22.84	17	null	2: Cardioembolism (e.g., atrial fibrillation/flutter, prosthetic heart valve, recent MI)
sub-3	train	annotated	lesionAcute	max_lesion_area	16	26	[ 256, 256, 26 ]	[ 0.8984, 0.8984, 6 ]	32,330	true	true	[]	true	F	87	w	1	1	19.23	15	6	5: Cryptogenic Stroke
sub-4	train	annotated	lesionAcute	max_lesion_area	11	28	[ 224, 224, 28 ]	[ 1.1058, 1.1058, 6 ]	440	true	true	[]	false	null	null	null	null	null	null	null	null	null
sub-5	train	annotated	lesionAcute	max_lesion_area	17	25	[ 256, 256, 25 ]	[ 0.8984, 0.8984, 6 ]	171	true	true	[]	true	M	58	w	1	0	37.29	1	null	3: Small-vessel disease (e.g., Subcortical or brain stem lacunar infarction <1.5 cm)
sub-6	train	unannotated	null	middle	13	26	[ 240, 240, 26 ]	[ 1.0417, 1.0417, 6 ]	null	false	true	[]	true	null	null	null	null	null	null	null	null	null
sub-7	train	annotated	lesionAcute	max_lesion_area	13	25	[ 224, 224, 25 ]	[ 1.0224, 1.0224, 6 ]	1,360	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-8	train	annotated	lesionAcute	max_lesion_area	19	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	12,714	true	true	[]	true	F	34	b	1	0	23.14	19	null	5: Cryptogenic Stroke
sub-9	train	annotated	lesionAcute	max_lesion_area	13	25	[ 384, 384, 25 ]	[ 0.599, 0.599, 6 ]	15,950	true	true	[]	true	F	70	w	1	0	26.89	18	4	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-10	train	annotated	lesionAcute	max_lesion_area	15	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	1,661	true	true	[]	false	null	null	null	null	null	null	null	null	null
sub-11	train	annotated	lesion	max_lesion_area	12	26	[ 224, 224, 26 ]	[ 1.1058, 1.1058, 6 ]	705	false	true	[]	true	null	null	null	null	null	null	null	null	null
sub-12	train	unannotated	null	middle	12	24	[ 256, 256, 24 ]	[ 0.8984, 0.8984, 6 ]	null	false	true	[]	false	null	null	null	null	null	null	null	null	null
sub-13	train	annotated	lesionAcute	max_lesion_area	18	26	[ 256, 256, 26 ]	[ 0.9766, 0.9766, 6 ]	189	true	true	[]	true	M	73	w	1	0	33.77	0	0	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-14	train	annotated	lesionAcute	max_lesion_area	13	26	[ 224, 224, 26 ]	[ 1.1058, 1.1058, 6 ]	61	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-15	train	annotated	lesionAcute	max_lesion_area	13	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	804	true	true	[]	true	F	70	w	1	0	26.56	0	null	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-16	train	annotated	lesionAcute	max_lesion_area	13	24	[ 240, 240, 24 ]	[ 1.0417, 1.0417, 6 ]	5,321	true	true	[]	false	null	null	null	null	null	null	null	null	null
sub-17	train	annotated	lesionAcute	max_lesion_area	13	26	[ 240, 240, 26 ]	[ 1.0714, 1.0714, 6 ]	552	true	true	[]	true	M	72	w	1	0	30.48	0	2	3: Small-vessel disease (e.g., Subcortical or brain stem lacunar infarction <1.5 cm)
sub-18	train	unannotated	null	middle	13	26	[ 256, 256, 26 ]	[ 0.8984, 0.8984, 6 ]	null	false	true	[]	true	M	44	null	1	0	22.99	3	2	3: Small-vessel disease (e.g., Subcortical or brain stem lacunar infarction <1.5 cm)
sub-19	train	annotated	lesionAcute	max_lesion_area	14	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	38	true	true	[]	true	F	43	w	1	1	20.96	0	0	5: Cryptogenic Stroke
sub-20	train	annotated	lesionAcute	max_lesion_area	17	26	[ 256, 256, 26 ]	[ 0.9375, 0.9375, 6 ]	162	true	true	[]	true	M	77	w	1	0	25	6	null	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-21	train	annotated	lesionAcute	max_lesion_area	12	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	2,006	true	true	[]	true	F	52	null	1	0	31.3	14	null	null
sub-22	train	unannotated	null	middle	12	25	[ 288, 288, 25 ]	[ 0.8214, 0.8214, 6 ]	null	false	true	[]	true	F	78	w	1	0	41.41	9	0	5: Cryptogenic Stroke
sub-23	train	annotated	lesionAcute	max_lesion_area	14	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	597	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-24	train	annotated	lesionAcute	max_lesion_area	17	27	[ 224, 224, 27 ]	[ 1.1058, 1.1058, 6 ]	918	true	true	[]	true	F	52	b	1	0	36.89	0	null	5: Cryptogenic Stroke
sub-25	train	annotated	lesionAcute	max_lesion_area	14	24	[ 240, 240, 24 ]	[ 1.0417, 1.0417, 6 ]	8,016	true	true	[]	false	null	null	null	null	null	null	null	null	null
sub-26	train	annotated	lesionAcute	max_lesion_area	13	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	489	true	true	[]	true	F	67	w	1	0	31.64	2	0	5: Cryptogenic Stroke
sub-27	train	annotated	lesionAcute	max_lesion_area	16	26	[ 224, 224, 26 ]	[ 1.1058, 1.1058, 6 ]	4,047	true	true	[]	false	null	null	null	null	null	null	null	null	null
sub-28	train	annotated	lesionAcute	max_lesion_area	15	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	1,170	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-29	train	unannotated	null	middle	13	26	[ 256, 256, 26 ]	[ 0.8984, 0.8984, 6 ]	null	false	true	[]	true	F	60	b	1	0	30.85	2	4	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-30	train	annotated	lesionAcute	max_lesion_area	10	26	[ 256, 256, 26 ]	[ 0.8984, 0.8984, 6 ]	586	true	true	[]	true	F	59	w	1	0	null	1	null	5: Cryptogenic Stroke
sub-32	train	unannotated	null	middle	13	26	[ 240, 240, 26 ]	[ 1.0417, 1.0417, 6 ]	null	false	true	[]	true	F	52	w	1	0	21.91	1	2	4: Stroke of other determined etiology
sub-33	train	annotated	lesionAcute	max_lesion_area	14	26	[ 240, 240, 26 ]	[ 1.0417, 1.0417, 6 ]	9,425	true	true	[]	true	M	41	w	1	0	23.67	14	2	5: Cryptogenic Stroke
sub-34	train	annotated	lesionAcute	max_lesion_area	13	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	109	true	true	[]	true	M	66	w	1	0	40.76	3	null	3: Small-vessel disease (e.g., Subcortical or brain stem lacunar infarction <1.5 cm)
sub-35	train	annotated	lesionAcute	max_lesion_area	21	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	13,697	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-36	train	annotated	lesionAcute	max_lesion_area	13	24	[ 240, 240, 24 ]	[ 1.0417, 1.0417, 6 ]	4,121	true	true	[]	true	F	85	w	1	0	22.72	15	3	2: Cardioembolism (e.g., atrial fibrillation/flutter, prosthetic heart valve, recent MI)
sub-37	train	unannotated	null	middle	12	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	null	false	true	[]	true	F	79	w	1	0	29.34	6	4	3: Small-vessel disease (e.g., Subcortical or brain stem lacunar infarction <1.5 cm)
sub-38	train	annotated	lesionAcute	max_lesion_area	14	26	[ 240, 240, 26 ]	[ 1.0417, 1.0417, 6 ]	336	true	true	[]	true	F	89+	w	1	0	null	12	null	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-39	train	annotated	lesionAcute	max_lesion_area	13	28	[ 224, 224, 28 ]	[ 1.0224, 1.0224, 6 ]	3,390	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-40	train	unannotated	null	middle	13	26	[ 384, 384, 26 ]	[ 0.599, 0.599, 6 ]	null	false	true	[]	true	null	null	null	null	null	null	null	null	null
sub-41	train	annotated	lesionAcute	max_lesion_area	9	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	1,494	true	true	[]	true	M	76	w	1	0	30.04	1	1	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-42	train	annotated	lesionAcute	max_lesion_area	9	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	21,011	true	true	[]	true	M	81	w	1	0	23.72	21	4	2: Cardioembolism (e.g., atrial fibrillation/flutter, prosthetic heart valve, recent MI)
sub-43	train	annotated	lesionAcute	max_lesion_area	17	28	[ 384, 384, 28 ]	[ 0.599, 0.599, 6 ]	941	true	true	[]	true	M	77	w	1	0	25.62	1	1	2: Cardioembolism (e.g., atrial fibrillation/flutter, prosthetic heart valve, recent MI)
sub-44	train	annotated	lesionAcute	max_lesion_area	12	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	183	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-45	train	annotated	lesionAcute	max_lesion_area	12	27	[ 240, 240, 27 ]	[ 1.0417, 1.0417, 6 ]	345	true	true	[]	true	M	60	w	1	0	38.27	0	null	3: Small-vessel disease (e.g., Subcortical or brain stem lacunar infarction <1.5 cm)
sub-46	train	annotated	lesionAcute	max_lesion_area	16	26	[ 224, 224, 26 ]	[ 1.1058, 1.1058, 6 ]	560	true	true	[]	true	F	50	b	1	0	39.9	3	null	5: Cryptogenic Stroke
sub-47	train	annotated	lesionAcute	max_lesion_area	15	26	[ 256, 256, 26 ]	[ 0.9375, 0.9375, 6 ]	67,855	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-48	train	unannotated	null	middle	13	27	[ 288, 288, 27 ]	[ 0.9328, 0.9328, 6 ]	null	false	true	[]	true	M	48	w	1	0	34.04	12	0	4: Stroke of other determined etiology
sub-49	train	annotated	lesionAcute	max_lesion_area	14	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	2,865	true	true	[]	true	M	83	b	1	0	24.39	0	null	2: Cardioembolism (e.g., atrial fibrillation/flutter, prosthetic heart valve, recent MI)
sub-50	train	annotated	lesionAcute	max_lesion_area	17	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	4,699	true	true	[]	true	M	58	w	1	0	26.28	18	0	4: Stroke of other determined etiology
sub-51	train	annotated	lesionAcute	max_lesion_area	13	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	1,247	true	true	[]	true	M	75	b	1	0	25.93	1	null	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-52	train	annotated	lesionAcute	max_lesion_area	14	25	[ 256, 256, 25 ]	[ 0.8984, 0.8984, 6 ]	456	true	true	[]	true	F	83	w	1	0	null	9	null	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-53	train	annotated	lesionAcute	max_lesion_area	17	26	[ 256, 256, 26 ]	[ 0.8984, 0.8984, 6 ]	25,817	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-54	train	unannotated	null	middle	12	25	[ 224, 224, 25 ]	[ 1.0224, 1.0224, 6 ]	null	false	true	[]	true	null	null	null	null	null	null	null	null	null
sub-55	train	unannotated	null	middle	13	27	[ 256, 256, 27 ]	[ 0.9375, 0.9375, 6 ]	null	false	true	[]	true	null	null	null	null	null	null	null	null	null
sub-56	train	annotated	lesionAcute	max_lesion_area	18	24	[ 384, 384, 24 ]	[ 0.599, 0.599, 6 ]	84,070	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-57	train	unannotated	null	middle	13	26	[ 384, 384, 26 ]	[ 0.599, 0.599, 6 ]	null	false	true	[]	true	M	50	w	1	0	21.87	4	null	null
sub-58	train	annotated	lesionAcute	max_lesion_area	19	26	[ 288, 288, 26 ]	[ 0.9328, 0.9328, 6 ]	2,876	true	true	[]	true	M	58	b	1	0	29.39	4	2	5: Cryptogenic Stroke
sub-59	train	annotated	lesionAcute	max_lesion_area	11	25	[ 288, 288, 25 ]	[ 0.9328, 0.9328, 6 ]	365	true	true	[]	false	null	null	null	null	null	null	null	null	null
sub-60	train	annotated	lesionAcute	max_lesion_area	11	26	[ 256, 256, 26 ]	[ 0.8984, 0.8984, 6 ]	699	true	true	[]	false	null	null	null	null	null	null	null	null	null
sub-61	train	unannotated	null	middle	13	26	[ 288, 288, 26 ]	[ 0.8333, 0.8333, 6 ]	null	false	true	[]	true	F	62	w	1	0	29.76	11	4	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-62	train	annotated	lesionAcute	max_lesion_area	13	26	[ 224, 224, 26 ]	[ 1.1058, 1.1058, 6 ]	3,330	true	true	[]	true	M	61	b	1	0	27.78	8	null	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-63	train	annotated	lesionAcute	max_lesion_area	15	26	[ 224, 224, 26 ]	[ 1.1058, 1.1058, 6 ]	410	true	true	[]	false	null	null	null	null	null	null	null	null	null
sub-64	train	annotated	lesionAcute	max_lesion_area	15	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	865	true	true	[]	true	M	58	b	1	0	19.44	12	null	null
sub-65	train	unannotated	null	middle	13	26	[ 240, 240, 26 ]	[ 0.9989, 0.9989, 6 ]	null	false	true	[]	true	F	72	w	1	0	30.84	0	null	5: Cryptogenic Stroke
sub-66	train	annotated	lesionAcute	max_lesion_area	16	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	1,571	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-67	train	annotated	lesionAcute	max_lesion_area	9	24	[ 480, 480, 24 ]	[ 0.5271, 0.5271, 6 ]	23,174	true	true	[]	true	M	66	null	1	0	31.05	8	3	5: Cryptogenic Stroke
sub-68	train	annotated	lesionAcute	max_lesion_area	11	26	[ 256, 256, 26 ]	[ 0.9386, 0.9386, 6 ]	12,301	true	true	[]	true	M	53	b	1	0	21.45	3	4	5: Cryptogenic Stroke
sub-69	train	annotated	lesionAcute	max_lesion_area	15	24	[ 256, 256, 24 ]	[ 0.8984, 0.8984, 6 ]	4,641	true	true	[]	true	F	40	b	1	0	null	0	0	4: Stroke of other determined etiology
sub-70	train	annotated	lesionAcute	max_lesion_area	9	25	[ 256, 256, 25 ]	[ 0.8984, 0.8984, 6 ]	2,471	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-71	train	annotated	lesionAcute	max_lesion_area	17	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	160	true	true	[]	true	M	64	w	1	0	16.6	1	null	5: Cryptogenic Stroke
sub-72	train	unannotated	null	middle	13	26	[ 224, 224, 26 ]	[ 1.1058, 1.1058, 6 ]	null	false	true	[]	true	F	76	w	1	0	19.88	4	null	3: Small-vessel disease (e.g., Subcortical or brain stem lacunar infarction <1.5 cm)
sub-73	train	annotated	lesionAcute	max_lesion_area	17	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	272	true	true	[]	true	M	62	w	1	0	24.24	4	0	5: Cryptogenic Stroke
sub-74	train	unannotated	null	middle	12	24	[ 240, 240, 24 ]	[ 1.0417, 1.0417, 6 ]	null	false	true	[]	true	M	73	b	1	0	20.83	1	0	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-75	train	annotated	lesionAcute	max_lesion_area	2	24	[ 256, 256, 24 ]	[ 0.9766, 0.9766, 6 ]	11,771	true	true	[]	true	F	89+	w	1	0	22.07	29	6	null
sub-76	train	unannotated	null	middle	13	26	[ 256, 256, 26 ]	[ 1.0547, 1.0547, 6 ]	null	false	true	[]	true	M	66	w	1	0	24.1	4	2	3: Small-vessel disease (e.g., Subcortical or brain stem lacunar infarction <1.5 cm)
sub-77	train	annotated	lesionAcute	max_lesion_area	3	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	1,874	true	true	[]	true	M	60	w	1	0	35.59	0	4	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-78	train	unannotated	null	middle	13	26	[ 224, 224, 26 ]	[ 1.1058, 1.1058, 6 ]	null	false	true	[]	true	M	63	w	1	0	null	2	1	3: Small-vessel disease (e.g., Subcortical or brain stem lacunar infarction <1.5 cm)
sub-79	train	annotated	lesionAcute	max_lesion_area	15	26	[ 256, 256, 26 ]	[ 0.9375, 0.9375, 6 ]	14,985	true	true	[]	true	M	68	b	1	0	18.69	11	4	2: Cardioembolism (e.g., atrial fibrillation/flutter, prosthetic heart valve, recent MI)
sub-80	train	annotated	lesionAcute	max_lesion_area	15	26	[ 256, 256, 26 ]	[ 0.8984, 0.8984, 6 ]	2,976	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-81	train	annotated	lesionAcute	max_lesion_area	18	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	480	true	true	[]	true	F	74	b	1	0	27.46	9	2	5: Cryptogenic Stroke
sub-82	train	annotated	lesionAcute	max_lesion_area	13	26	[ 256, 256, 26 ]	[ 1.0156, 1.0156, 6 ]	13,382	true	true	[]	true	M	61	w	1	0	22.68	10	0	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-83	train	annotated	lesionAcute	max_lesion_area	16	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	372	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-84	train	annotated	lesionAcute	max_lesion_area	16	24	[ 384, 384, 24 ]	[ 0.599, 0.599, 6 ]	1,857	true	true	[]	true	F	75	w	1	0	35.16	1	1	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-85	train	unannotated	null	middle	13	26	[ 224, 224, 26 ]	[ 1.1058, 1.1058, 6 ]	null	false	true	[]	true	null	null	null	null	null	null	null	null	null
sub-86	train	annotated	lesionAcute	max_lesion_area	13	24	[ 240, 240, 24 ]	[ 1.0417, 1.0417, 6 ]	1,167	true	true	[]	true	F	53	b	1	0	24.61	13	0	4: Stroke of other determined etiology
sub-87	train	annotated	lesionAcute	max_lesion_area	12	27	[ 256, 256, 27 ]	[ 0.8984, 0.8984, 6 ]	1,256	true	true	[]	true	F	72	w	1	0	27.4	6	null	5: Cryptogenic Stroke
sub-88	train	unannotated	null	middle	13	26	[ 224, 224, 26 ]	[ 1.0224, 1.0224, 6 ]	null	false	true	[]	true	F	89+	b	1	0	30.08	22	3	5: Cryptogenic Stroke
sub-89	train	annotated	lesionAcute	max_lesion_area	17	29	[ 256, 256, 29 ]	[ 0.9704, 0.9704, 6 ]	1,338	true	true	[]	false	null	null	null	null	null	null	null	null	null
sub-90	train	annotated	lesionAcute	max_lesion_area	8	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	740	true	true	[]	true	M	74	null	1	0	22.22	0	null	5: Cryptogenic Stroke
sub-91	train	annotated	lesionAcute	max_lesion_area	16	26	[ 384, 384, 26 ]	[ 0.599, 0.599, 6 ]	514	true	true	[]	true	F	68	w	1	0	25.82	1	2	5: Cryptogenic Stroke
sub-92	train	annotated	lesionAcute	max_lesion_area	15	26	[ 240, 240, 26 ]	[ 1.0417, 1.0417, 6 ]	728	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-93	train	unannotated	null	middle	13	26	[ 384, 384, 26 ]	[ 0.599, 0.599, 6 ]	null	false	true	[]	true	M	74	w	1	0	21.66	2	4	5: Cryptogenic Stroke
sub-94	train	annotated	lesionAcute	max_lesion_area	12	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	1,141	true	true	[]	true	M	55	b	1	0	35.19	18	0	5: Cryptogenic Stroke
sub-95	train	annotated	lesionAcute	max_lesion_area	15	27	[ 256, 256, 27 ]	[ 0.8984, 0.8984, 6 ]	29,433	true	true	[]	true	M	64	w	1	0	37.04	28	5	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-96	train	annotated	lesionAcute	max_lesion_area	4	26	[ 224, 224, 26 ]	[ 1.0148, 1.0148, 6 ]	4,583	true	true	[]	true	M	77	w	1	0	28.01	7	3	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-97	train	annotated	lesionAcute	max_lesion_area	12	25	[ 224, 224, 25 ]	[ 1.1058, 1.1058, 6 ]	1,000	true	true	[]	true	M	60	w	1	0	25.01	1	2	5: Cryptogenic Stroke
sub-98	train	annotated	lesionAcute	max_lesion_area	18	26	[ 256, 256, 26 ]	[ 0.8984, 0.8984, 6 ]	1,961	true	true	[]	true	null	null	null	null	null	null	null	null	null
sub-99	train	annotated	lesionAcute	max_lesion_area	12	27	[ 256, 256, 27 ]	[ 0.8984, 0.8984, 6 ]	584	true	true	[]	true	M	79	b	1	0	29.07	5	null	1: Large-artery atherosclerosis (e.g., carotid or basilar artery stenosis)
sub-100	train	annotated	lesionAcute	max_lesion_area	16	26	[ 256, 256, 26 ]	[ 0.8984, 0.8984, 6 ]	10,699	true	true	[]	true	F	87	b	1	0	39.9	14	4	5: Cryptogenic Stroke
sub-101	train	annotated	lesionAcute	max_lesion_area	16	24	[ 256, 256, 24 ]	[ 0.8984, 0.8984, 6 ]	2,704	true	true	[]	false	null	null	null	null	null	null	null	null	null

End of preview. Expand in Data Studio

SOOP — Stroke Outcome Optimization Project (DWI subset)

Acute ischemic stroke lesion segmentation on clinical brain MRI from a single comprehensive stroke center (Prisma Health-Upstate, Greenville, South Carolina).

Mirrored from OpenNeuro ds004889 v1.1.2 (doi:10.18112/openneuro.ds004889.v1.1.2).

⚠️ This mirror is a documented DWI subset

The source release carries four modalities per subject. This mirror carries two:

Modality	Size in source	Mirrored
DWI TRACE (`rec-TRACE`)	1.19 GiB	✅
DWI ADC (`rec-ADC`)	1.41 GiB	✅
T1w	41.63 GiB	❌ excluded
T2w-FLAIR	22.73 GiB	❌ excluded

Why: every lesion mask is defined in space-TRACE. TRACE and ADC come from the same acquisition and share the mask's voxel grid exactly, so image and label are aligned with no resampling. T1w and FLAIR are independent clinical acquisitions in their own geometry — pairing them with a mask requires coregistration, which this mirror does not perform. The published nnU-Net benchmark on SOOP reports median Dice 0.777 for DWI+FLAIR vs 0.774 for DWI alone, so the excluded 64 GiB buys ~0.003 Dice at the cost of a registration step. Users needing T1w/FLAIR should pull them from OpenNeuro directly.

Everything mirrored is byte-identical to OpenNeuro: each volume was verified against the SHA256 embedded in its git-annex key at download time.

Cohort and annotation

	Subjects
Imaged (TRACE + ADC complete)	1,715
Annotated (≥1 mask tier)	1,456
Unannotated (no mask of any tier)	259
Covered by `participants.tsv`	1,505

Mask tiers — all three are shipped

The three tiers are anatomically distinct targets, not competing raters:

Tier	Subjects	Meaning
`lesionAcute`	1,451	recent infarct — DWI-hyperintense
`lesionChronic`	203	pre-existing injury — not DWI-hyperintense
`lesion`	1,455	union of the two

All masks are manual (three trained raters, MRIcroGL, traced slice-by-slice on the TRACE/ADC diffusion images). No automated pre-segmentation and no multi-rater consensus; no inter-rater reliability is reported by the source publication.

Default target: lesionAcute — it matches the DWI input. Scoring the combined lesion mask against a diffusion image injects label noise on the 203 subjects carrying chronic lesions, which are not DWI-bright.

Absent ≠ empty ≠ negative

This mirror resolves absent-mask semantics explicitly, because the source release does not:

259 subjects have no mask of any tier. They are unannotated, not lesion-free controls. Nothing in the release supports reading them as true negatives: acuteischaemicstroke takes only 1 or n/a (never 0), and participants.tsv covers just 1,505 of 1,715 subjects. They are excluded from the evaluable set rather than zero-filled.
5 subjects (sub-11, sub-251, sub-1311, sub-1613, sub-1724) have a chronic mask but no acute mask. Under the default lesionAcute target they are unannotated for that target and must be skipped, never zero-filled.
1 subject (sub-235) has acute + chronic tiers but no combined tier — a missing union, not an empty one.
A mask that is present but all-zero is a genuine empty annotation and is kept, flagged by the empty_<tier> column. lesionAcute has no empty masks; the combined lesion tier has 3 (see hazard 3 below).

Evaluable subjects under the default target: 1,451.

The source paper's abstract claims 1,461 acute ischemic stroke cases. The released data contains 1,451 acute masks (1,456 subjects carry a mask of some tier). The 10-subject difference is not explained in the release; 1,451 is what actually ships and is the number this mirror evaluates on. A stale figure of 1,449 also circulates in the authors' demo repository.

Files

sub-<N>/dwi/sub-<N>_rec-TRACE_dwi.nii.gz    (+ .json sidecar)
sub-<N>/dwi/sub-<N>_rec-ADC_dwi.nii.gz      (+ .json sidecar)
derivatives/lesion_masks/sub-<N>/dwi/sub-<N>_space-TRACE_desc-<tier>_mask.nii.gz
subjects.jsonl        one row per subject (see below)
cohort_summary.json   aggregate counts, including every number quoted in this card
participants.tsv      source clinical table (partial coverage)
SOURCE_README.md      the original OpenNeuro README, verbatim
dataset_description.json, CHANGES, dwi.bval, dwi.bvec, participants.json

subjects.jsonl columns: subject_id, patient_id, split, image_trace, image_adc, mask_lesion / mask_lesionAcute / mask_lesionChronic (nullable — null means that tier was never annotated), has_<tier>, n_fg_<tier>, empty_<tier>, evaluable_<tier>, annotation_status, is_lesion_free_control (always false — see above), grid, spacing, orientation, trace_ndim, trace_dtype, adc_dtype, adc_affine_matches_trace, adc_usable, lesion_equals_union, qc_flags, in_participants_tsv, plus the clinical columns (sex, age, race, acuteischaemicstroke, priorstroke, bmi, nihss, gs_rankin_6isdeath, etiology), all nullable.

Loading hazards

⚠️ Masks are NOT all binary — binarize with > 0, never == 1. 11 subjects (sub-503, sub-504, sub-505, sub-507, sub-508, sub-509, sub-510, sub-1252, sub-1253, sub-1255, sub-1257) carry mask labels encoded as {0, 2}, {0, 3} or {0, 1, 4} instead of {0, 1}. A mask == 1 test yields a silently empty mask for almost all of them. The n_fg_<tier> column in subjects.jsonl is computed as > 0 and is the reference count.
DWI volumes are 4-D (X, Y, Z, 1); masks are 3-D (X, Y, Z). Squeeze the trailing singleton axis before pairing.
⚠️ 3 subjects have a present-but-all-zero combined lesion mask that contradicts a non-empty lesionAcute (sub-335, sub-539, sub-1550). For these the union tier is broken, not merely empty — they are the only 3 violations of lesion == lesionAcute ∪ lesionChronic in the cohort (which otherwise holds exactly, checked on all 1,455 combined masks). The default lesionAcute target is unaffected.
⚠️ 2 subjects have an ADC affine that disagrees with TRACE (sub-235, sub-1233) — by ~11 mm and ~237 mm respectively. Because masks are defined in space-TRACE, the labels do not align to those subjects' ADC volumes. Skip them when using modality='adc'; the adc_usable column marks them. The other 1,713 subjects have ADC and TRACE affines matching exactly.
Per-subject geometry varies — there is no fixed shape. 55 distinct voxel grids and 88 distinct spacings across the cohort (clinical protocols differ by subject; the source README says so explicitly). Slice count ranges 21–32 (thick-slice 2-D EPI, ~5 mm thick / 6 mm apart). All volumes are LAS.
Subject IDs are non-contiguous: sub-1 … sub-1737 with 1,715 present and 22 gaps. Never assume a dense index.
Intensity scaling differs between the two modalities. Both are int16 on disk. TRACE carries no effective scaling (values 0–4095). ADC carries a NIfTI scl_slope/scl_inter that must be applied — read via np.asanyarray(img.dataobj) or get_fdata(), which returns float in roughly 0–50,000. Reading ADC as raw ints yields negative values and is wrong.
acuteischaemicstroke does not identify the stroke cohort — it is 1 or n/a, never 0. Use mask presence instead.
participants.json documents a lesion_size_cc column that is absent from participants.tsv.
priorstroke undercounts chronic lesions — far fewer subjects are flagged than carry a lesionChronic mask.

QC-flagged subjects

Detected by a full-cohort scan of all 1,715 subjects (mask label values, grid and affine agreement against TRACE, and the union property lesion == lesionAcute ∪ lesionChronic) — not by sampling:

Subject	Flags
`sub-235`	adc_affine_mismatch
`sub-335`	empty_lesion; union: +0/-34762
`sub-503`	nonbinary_lesion:[0, 2]; nonbinary_lesionAcute:[0, 2]
`sub-504`	nonbinary_lesion:[0, 2]; nonbinary_lesionAcute:[0, 2]
`sub-505`	nonbinary_lesion:[0, 2]; nonbinary_lesionAcute:[0, 2]
`sub-507`	nonbinary_lesion:[0, 1, 4]; nonbinary_lesionAcute:[0, 2]; nonbinary_lesionChronic:[0, 2]
`sub-508`	nonbinary_lesion:[0, 2]; nonbinary_lesionAcute:[0, 2]
`sub-509`	nonbinary_lesion:[0, 2]; nonbinary_lesionAcute:[0, 2]
`sub-510`	nonbinary_lesionChronic:[0, 2]
`sub-539`	empty_lesion; union: +0/-16377
`sub-1233`	adc_affine_mismatch
`sub-1252`	nonbinary_lesion:[0, 2]; nonbinary_lesionAcute:[0, 2]
`sub-1253`	nonbinary_lesion:[0, 2]; nonbinary_lesionAcute:[0, 2]
`sub-1255`	nonbinary_lesion:[0, 3]; nonbinary_lesionAcute:[0, 3]
`sub-1257`	nonbinary_lesion:[0, 3]; nonbinary_lesionAcute:[0, 3]
`sub-1550`	empty_lesion; union: +0/-3495

16 of 1,715 subjects carry at least one flag. The remaining 1,699 have binary {0, 1} uint8 masks sharing the TRACE grid and affine exactly. Every mask in the cohort matches its TRACE volume's grid and affine (no grid mismatches were found at any tier).

License

CC0 1.0 (public domain) — as declared in the source dataset_description.json and by the OpenNeuro API. Redistribution and re-hosting are permitted.

⚠️ The PMC page for the associated Scientific Data article shows CC BY-NC-ND 4.0. That governs the article, not the data. The dataset is CC0.

Provenance

Official author upload (Rorden, Absher, Newman-Norlund) to OpenNeuro — not a third-party re-host. Version 1.1.2 (2024-05-08).

Cross-dataset overlap: no shared patients with the BraTS family, Medical Segmentation Decathlon, TCIA collections, or the ISLES editions (which are European cohorts; ATLAS is chronic stroke). The companion Aphasia Recovery Cohort (ARC, ds004512) shares investigators and a center grant but covers chronic aphasia at a different site. No cross-reference ID column exists in either release, so overlap is undocumented and cannot be verified — treat as low risk but not formally excluded.

Citation

@article{absher2024soop,
  title={The stroke outcome optimization project: Acute ischemic strokes from a comprehensive stroke center},
  author={Absher, John and Goncher, Sarah and Newman-Norlund, Roger and others},
  journal={Scientific Data}, volume={11}, pages={839}, year={2024},
  doi={10.1038/s41597-024-03667-5}
}

Dataset: Rorden C., Absher J., Newman-Norlund R. Stroke Outcome Optimization Project (SOOP). OpenNeuro, 2024. doi:10.18112/openneuro.ds004889.v1.1.2

Downloads last month: 31