Skip to main content
Insights into Imaging

Table 3 Univariable and multivariable Cox regression and concordance analyses for the prediction of progression-free survival by MRI variables in 357 endometrial cancer patients

From: MRI-assessed tumor-free distance to serosa predicts deep myometrial invasion and poor outcome in endometrial cancer

Imaging variables

Univariable model

Multivariable model α

HR (95%CI) [p value]*

Concordance

HR (95% CI) [p value]* α

Concordance

iTFD (< 6 mm)

2.36 (1.48–3.76) p < 0.001

0.60

1.96 (1.20–3.18) p = 0.007

0.72

DOI (≥ 5 mm)

2.19 (1.26–3.80) p = 0.005

0.58

2.10 (1.21–3.66) p = 0.009

0.72

APD (≥ 17 mm)

2.77 (1.69–4.55) p < 0.001

0.63

2.17 (1.31–3.59) p = 0.003

0.72

iDMI

2.95 (1.81–4.82) p < 0.001

0.63

2.74 (1.66–4.51) p < 0.001

0.73

Patient age (yrs)

1.04 (1.01–1.06) p = 0.003

0.61

  

Preoperative high-risk histology β

4.14 (2.62–6.53) p < 0.001

0.66

  
  1. APD, anteroposterior tumor diameter; CI, confidence interval; DOI, depth of invasion; HR, hazard ratio; iDMI, deep (≥ 50%) myometrial invasion based on standard imaging reading; iTFD, tumor-free distance to serosa based on imaging findings
  2. *Cox proportional hazard model, p value refers to log-rank test
  3. αMultivariable analyses for each imaging variable after adjusting for patient age and preoperative high-risk histology
  4. βHigh-risk histology (endometrioid grade 3/non-endometrioid histology) based on preoperative curettage/biopsy
  5. All significant p values are given in boldface
Back to article page