Prediction of lymph node metastasis in operable cervical cancer using clinical parameters and deep learning with MRI data: a multicentre study

Insights into Imaging

Table 2 Clinical characteristics of the training and internal validation cohorts and external validation cohort in other centres

Parameters	Training cohort (n = 225)			Internal validation cohort (n = 98)			external validation cohort (n = 69)
Parameters	LNM group (n = 75)	Non-LNM group (n = 150)	p	LNM group (n = 33)	Non-LNM group (n = 65)	p	LNM group (n = 20)	Non-LNM group (n = 49)	p
Age (years)			0.920			0.72			0.140
≥ 50	48	94		17	36		12	38
< 50	27	56		16	29		8	11
Menstrual state			0.700			0.73			0.770
Menopause	49	48		19	35		13	37
Premenopausal	26	102		14	30		7	12
SCC-Ag (ng/L)			< 0.001			0.02			0.020
≥ 1.5	48	47		18	20		13	17
< 1.5	27	103		15	45		7	32
Pathological type			0.460			0.22			0.880
Squamous cell carcinoma	68	131		30	53		18	44
Adenocarcinoma	7	19		3	12		2	5
Histologic grade			0.510			0.16			0.400
High/middle	40	87		26	58		14	39
Low	35	63		7	7		6	10
LN_max short diameter (mm)			< 0.001			0.008			0.004
≥ 10	18	5		23	6		8	5
< 10	57	145		10	59		12	44
Maximum tumour diameter (mm)			< 0.001			0.030			0.052
≥ 40	38	26		14	14		7	7
< 40	37	124		19	51		13	42
Tumour volume (cm³)			< 0.001			0.005			0.020
≥ 15	31	35		13	20		10	11
< 15	44	115		20	45		10	38