Skip to main content

Table 1 Patient and tumor baseline characteristics

From: Deep learning models of ultrasonography significantly improved the differential diagnosis performance for superficial soft-tissue masses: a retrospective multicenter study

 

Internal cohort

Test cohort A

Test cohort B

P

Training cohort

Validation cohort

Subject

618

154

156

123

 

Malignant

Age

 

49.1 (± 16.1)

51.0 (± 12.2)

47.4 (± 12.0)

49.1 (± 16.1)

0.91

Gender

Male

17 (53.1%)

4 (50.0%)

2 (40.0%)

4 (40.0%)

0.431

 

Total

32

8

5

10

 

Benign

Age

33.9 (± 9.0)

32.7 (± 7.9)

32 (± 7.5)

33.9 (± 9.0)

0.02

Gender

Male

288 (49.2%)

69 (47.3%)

80 (53.0%)

59 (52.2%)

 < 0.001

Type

0.028

 

Lipomyoma

121 (20.6%)

30 (20.5%)

40 (26.5%)

28 (24.8%)

 
 

Hemangioma

120 (20.5%)

30 (20.5%)

16 (10.5%)

20 (17.7%)

 
 

Neurinoma

120 (20.5%)

30 (20.5%)

25 (16.6%)

18 (15.9%)

 
 

Epidermal cyst

160 (27.3%)

40 (27.4%)

45 (29.8%)

27 (23.9%)

 
 

Calcifying epithelioma

65 (11.1%)

16 (11.1%)

25 (16.6%)

20 (17.7%)

 
 

Total

586

146

151

113

 
  1. Data are presented as n (%) or mean ± SD