Skip to main content

Table 1 Comparison of patients’ characteristics in the development and testing datasets

From: Predicting tumor deposits in rectal cancer: a combined deep learning model using T2-MR imaging and clinical features

Clinical variables

Development dataset

Independent testing dataset

Non-TDs (n = 143)

TDs (n = 103)

p-value

Non-TDs (n = 47)

TDs (n = 34)

p-value

Gender (female/male)

95/48

61/42

0.247

34/13

26/8

0.676

Age (mean ± SD)

57.6 ± 12.6

57.2 ± 12.3

0.798

59.0 ± 10.8

58.6 ± 12.3

0.906

MR T-stage (T1–2/T3/T4)

26/112/5

10/83/10

0.034

5/42/0

4/29/1

0.486

MR N-stage (N0/N1/N2)

75/53/15

14/60/29

 < 0.001

26/18/3

5/18/11

 < 0.001

MR M-stage (M0/M1)

137/6

99/4

0.903

47/0

33/1

0.237

Lesion location (bottom/mid/top/bottom-mid/mid-top/all)

2/13/59/14/53/2

25/16/17/19/23/3

 < 0.001

2/5/18/2/20/0

3/4/11/12/4/0

0.001

Morphologic type (mass/infiltrate)

24/119

34/69

0.003

8/39

16/18

0.004

CRM (− / +)

72/71

60/43

0.220

23/24

27/7

0.005

EMVI(− / +)

139/4

73/30

 < 0.001

46/1

22/12

 < 0.001

ADC value (mean ± SD)

0.914 ± 0.181

0.915 ± 0.136

0.961

0.905 ± 0.114

0.924 ± 0.135

0.493

CEA (− / +)

92/51

56/47

0.115

26/21

16/18

0.463

CA19-9 (− / +)

119/24

70/33

0.005

40/7

23/11

0.062

CA125 (− / +)

139/4

101/2

0.668

45/2

30/4

0.203

  1. Abbreviations: CRM circumferential resection margin, EMVI extramural vascular invasion; CEA ( +), > 5 ng/m; CA19-9 ( +), > 35 u/ml; CA125 ( +), > 35 u/ml