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

Insights into Imaging

Table 2 Univariate and multivariate regression analysis of the clinical variables in the development dataset

Clinical variables	Univariate regression analysis			Multivariate regression analysis
Clinical variables	Odds ratio	95% CI	P	Odds ratio	95% CI	P
Gender	1.363	0.807–2.302	0.247
Age	0.997	0.975–1.019	0.797
MR T-stage	2.172	1.177–4.006	0.013	1.497	0.687–3.261	0.310
MR N-stage	3.428	2.255–5.211	< 0.001	3.549	2.153–5.852	< 0.001
MR M-stage	0.923	0.254–3.356	0.903
EMVI	14.281	4.845–42.098	< 0.001	11.456	3.366–38.984	< 0.001
Tumor location	0.687	0.565–0.835	< 0.001	0.661	0.516–0.846	0.001
CRM	0.727	0.436–1.211	0.041	0.404	0.198–0.825	0.013
Morphologic type	0.409	0.224–0.746	0.004	0.707	0.335–1.493	0.363
ADC value	1.049	0.157–7.021	0.960
Serum CEA level	1.514	0.903–2.539	0.116
Serum CA125 level	0.6881	0.124–3.830	0.670
Serum CA19-9 level	2.3375	1.279–4.272	0.006	2.356	1.139–4.874	0.021

CRM circumferential resection margin, EMVI extramural vascular invasion, ADC apparent diffusion coefficient