Skip to main content

Table 1 Characteristics of patients and X-ray images assigned to the training and test datasets

From: Computer-aided diagnosis with a convolutional neural network algorithm for automated detection of urinary tract stones on plain X-ray

  Training dataset Test dataset P value
Number of patients 827 190  
Gender, n (%)    0.132
 Male 537 (64.9) 143 (75.3)  
 Female 290 (35.1) 47 (24.7)  
Age, median (range), years 58 (17–89) 56 (14–87) 0.038
Number of labeled lesions per image, n (%)    0.486
 One 656 (79.4) 144 (75.8)  
 Two 112 (13.5) 32 (16.8)  
 More than two 59 (7.1) 14 (7.4)  
Location of urinary tract stone, n (%)    
 Kidney 428 (51.8) 106 (55.8) 0.895
 Proximal ureter 334 (40.4) 72 (37.9) 0.582
 Mid-ureter 75 (9.1) 27 (14.2) 0.046
 Distal ureter 184 (22.2) 18 (9.5) < 0.001
Staghorn calculus, n (%)    0.553
 Yes 17 (2.1) 2 (1.1)  
 No 810 (97.9) 188 (98.9)  
Artificial foreign body in image, n (%)    0.672
 Yes 53 (6.4) 10 (5.3)  
 No 774 (93.6) 180 (94.7)