. 2020 Apr 20;15:86. doi: 10.1186/s13014-020-01502-w

Table 2.

Univariate and multivariate analyses of potential prognostic factors for local control in the validation dataset. Statistically significant differences were found according to the imaging feature-based risk scores in the univariate and multivariate analyses

Variables		n	3-year LC (%)	P value		HR	95% CI
Variables		n	3-year LC (%)	univariate	multivariate	HR	95% CI
Imaging feature-based LC risk score^a	< 1.1	20	89.1	0.023^b	0.044^b	7.46	1.06–52.58
Imaging feature-based LC risk score^a	≥1.1	12	66.7	0.023^b	0.044^b	7.46	1.06–52.58
Age	< 50	15	60.6	0.160	0.161	5.50	0.50–59.52
Age	≥50	17	92.3	0.160	0.161	5.50	0.50–59.52
Pathology	SCC	30	84.9	0.706	0.741	0.65	0.05–8.19
Pathology	Non-SCC	2	50.0	0.706	0.741	0.65	0.05–8.19
FIGO stage^c	IIB	25	81.9	0.213	1.00	< 0.01	0- ∞
FIGO stage^c	IIIA, IIIB	7	85.7	0.213	1.00	< 0.01	0- ∞
Extent of nodal involvement	Pelvic only Pelvic + para-aortic	25	82.1	0.692	0.248	3.40	0.43-27.15
Extent of nodal involvement	Pelvic only Pelvic + para-aortic	7	85.7	0.692	0.248	3.40	0.43-27.15
Primary tumor size (mm)	< 50	16	79.1	0.808	0.284	0.28	0.03–2.85
Primary tumor size (mm)	≥50	16	87.5	0.808	0.284	0.28	0.03–2.85

Abbreviations: LC local control, HR hazard ratio, CI confidence interval, SCC squamous cell carcinoma

^aImaging features of primary cervical tumor

^bSignificant P values

^cThe 2009 International Federation of Gynecology and Obstetrics (FIGO) staging system