Skip to main content

Table 7 Multivariate regression analysis with TRG as dependent variable

From: CEA clearance pattern as a predictor of tumor response to neoadjuvant treatment in rectal cancer: a post-hoc analysis of FOWARC trial

Variable

TRG 0–1

TRG 2–3

P

Odds Ratio (95%CI)

Training cohort

 Tumor length

  

0.037

3.19 (1.07–9.46)

  4 cm

19 (42%)

11 (71%)

  

   > 4 cm

14 (58%)

27 (29%)

  

 Circumferential extent

  

0.146

2.65 (0.71–9.86)

  50%

12 (36%)

5 (13%)

  

   > 50%

21 (64%)

33 (87%)

  

 Normalization of post- treatment CEA Level

  

0.082

2.81 (0.88–9.01)

  Normal (< 5 ng/mL)

26 (79%)

21 (55%)

  

 Elevated (5 ng/mL)

7 (21%)

17 (45%)

  

 CEA clearance pattern

  

0.161

2.18 (0.73–6.44)

  Exponential decrease (R20.9)

19 (58%)

12 (32%)

  

  Non-exponential decrease (R2 < 0.9)

14 (42%)

26 (68%)

  

Validation cohort

 Tumor length

  

0.009

4.23 (1.43–12.50)

  4 cm

19 (58%)

12 (29%)

  

   > 4 cm

14 (42%)

30 (64%)

  

 Circumferential extent

  

0.147

0.42 (0.13–1.35)

  50%

9 (27%)

18 (43%)

  

   > 50%

24 (73%)

24 (57%)

  

 Normalization of post- treatment CEA Level

  

0.884

0.92 (0.29–2.92)

  Normal (< 5 ng/mL)

23 (70%)

23 (55%)

  

  Elevated (5 ng/mL)

10 (30%)

19 (45%)

  

 CEA clearance pattern

  

0.009

4.37 (1.44–13.23)

  Exponential decrease (R20.9)

21 (64%)

15 (36%)

  

  Non-exponential decrease (R2 < 0.9)

12 (36%)

27 (64%)

  
  1. Abbreviation: TRG tumor regression grade
  2. The p value in boldface means statistically significant, that is, less than 0.05