Table 4 Multivariate logistic regression analysis of the training set.

From: A nomogram for predicting non-curative resection in patients with early gastric cancer based on white light imaging

Variables

CR (n = 385)

NCR (n = 70)

Total (n = 455)

p value

Odds ratio

95% CI

Ulceration

39 (10.1%)

22 (31.4%)

61 (13.4%)

< 0.01

3.779

1.861–7.672

Fold concentration

16 (4.2%)

9 (12.9%)

25 (5.5%)

< 0.01

3.934

1.468–10.544

Remarkable redness

26 (6.7%)

15 (21.4%)

41 (9.0%)

< 0.01

3.453

1.556–7.664

Marginal elevation

37 (9.6%)

15 (21.4%)

52 (11.4%)

< 0.05

2.333

1.058–5.146

Spontaneous bleeding

11 (2.9%)

5 (7.1%)

16 (3.5%)

> 0.05

  

Whitish mucosal change

21 (5.5%)

9 (12.9%)

30 (6.6%)

< 0.05

2.850

1.077–7.539

Size (mm)

 ≤ 20

192 (49.9%)

17 (24.3%)

209 (45.9%)

< 0.01

1

 

 20–30

131 (34.0%)

28 (40.0%)

159 (34.9%)

2.463

1.203–5.043

 > 30

62 (16.1%)

25 (35.7%)

87 (19.1%)

3.368

1.543–7.351

Tumor location

 Lower 1/3 of stomach

206 (53.5%)

25 (35.7%)

231 (50.8%)

< 0.01

1

 

 Upper 1/3 of stomach

23 (6.0%)

14 (20.0%)

37 (8.1%)

1.792

0.941–3.410

 Mid 1/3 of stomach

156 (40.5%)

31 (44.3%)

187 (41.1%)

8.609

3.488–21.247

 Hp infection

52 (13.5%)

16 (22.9%)

68 (14.9%)

< 0.05

2.175

1.053–4.492

  1. CR curative resection, NCR non-curative resection, Hp Helicobacter pylori.