Detecting MRI-Invisible Prostate Cancers Using a Weakly Supervised Deep Learning Model

Table 1

Demographic and clinical characteristics of 777 included men.

Variables	Training cohort ()	Testing cohort ()	value
Age	65 (61-71)	65 (61-70)	0.779
PSA	6.5 (4.4-10.1)	6.6 (4.1-9.8)	0.630
No. of core per examination	15 (12-17)	15 (12-17)	0.645
No. of systematic core per examination	10 (3-11)	10 (3-12)	0.709
No. of targeted core per examination	6 (4-8)	6 (4-8)	0.820
Per examination maximum Gleason score			>0.99
Total	903	261
Gleason	464	141
No prostate cancer	216	58
	248	83
Gleason	361	99
	270	71
	91	28
Gleason	78	21
	9	0
	35	8
	26	10
	0	1
	7	2
	1	0
Per systematic core Gleason score			0.119
Total	7083	2025
Gleason	6750	1907
No prostate cancer	6155	1719
	595	188
Gleason	275	98
	217	69
	58	29
Gleason	58	20
	5	4
	30	2
	16	8
	0	0
	6	6
	1	0

Data in parentheses are the interquartile range.