MRI-based radiomics analysis improves preoperative diagnostic performance for the depth of stromal invasion in patients with early stage cervical cancer

Insights into Imaging

Table 1 Characteristics of patients in training and validation cohorts

	Training cohort (n = 188)			Validation cohort (n = 46)			P*
	Superficial stromal invasion	Middle or deep stromal invasion	p value	Superficial stromal invasion	Middle or deep stromal invasion	p value	P*
	(n = 57)	(n = 131)		(n = 13)	(n = 33)
Age, mean ± SD, years	43.81 ± 9.88	44.98 ± 9.50	0.441	46.00 ± 9.49	46.79 ± 10.82	0.819	0.229
Menstrual status (N. %)			0.327			1.000	0.279
Premenopausal	44	92		8	21
Postmenopausal	13	39		5	12
FIGO stage (N. %)			< 0.001			0.028	0.985
IB1	41	36		10	10
IB2	15	81		3	20
IB3	0	6		0	1
IIA1	1	8		0	2
Biopsy histological type (N. %)			0.553			0.862	0.729
SCC	37	92		9	24
AC	19	34		4	7
ASC	1	5		0	2
MTD on MRI, mean ± SD, mm	17.63 ± 7.03	25.82 ± 9.13	< 0.001	14.65 ± 6.16	25.06 ± 8.39	< 0.001	0.469

P is derived from the univariable association analysis of each clinical variable between superficial stromal invasion patients and middle or deep stromal invasion patients in the training and validation cohort, respectively. P^* represents the difference of each clinical variable between the training and validation cohorts
SD, standard deviation; FIGO, Federation International of Gynecology and Obstetrics; SCC, squamous cell carcinoma; AC, adenocarcinoma; ASC, adenosquamous carcinoma; MTD, maximal tumor diameter