Whole-lesion apparent diffusion coefficient (ADC) histogram as a quantitative biomarker to preoperatively differentiate stage IA endometrial carcinoma from benign endometrial lesions

Table 6 Results of multivariate logistic regression models for differentiating stage IA EC from BELs

Models	Coefficients	Odds ratio	95% CI	p value
Clinical model
Age at diagnosis, years
≤ 50		Reference
51–64	1.412	4.106	2.269–7.429	< 0.001
≥ 65	0.574	1.776	0.563–5.604	0.328
Nulliparity
No		Reference
Yes	1.433	4.191	1.066–16.472	0.040
Long-term Tamoxifen therapy
No		Reference
Yes	− 1.963	0.14	0.038–0.524	0.003
ADC histogram model
ADC_10th	− 3.270	0.038	0.016–0.092	< 0.001
rMAD	− 0.710	0.492	0.271–0.893	0.020
Total energy	1.025	2.788	1.416–5.490	0.003
Skewness	0.605	1.832	1.155–2.907	0.010
Combined model
ADC-score*	0.971	2.641	2.045–3.411	< 0.001
Age at diagnosis, years
≤ 50		Reference
51–64	0.711	2.036	0.821–5.046	0.125
≥ 65	1.623	5.068	0.905–28.383	0.065
Nulliparity
No		Reference
Yes	0.446	1.562	0.252–9.670	0.631
Long-term Tamoxifen therapy
No		Reference
Yes	− 0.846	0.429	0.061–2.996	0.393

rMAD, robust mean absolute deviation
*ADC-score = 0.257 + − 3.270 × ADC_10th + − 0.710 × rMAD + 1.025 × Total_Energy + 0.605 × Sknewness. All ADC histogram features were standardized by Z-score normalization

ISSN: 1471-2342