Applicability of a pathological complete response magnetic resonance-based radiomics model for locally advanced rectal cancer in intercontinental cohort

Table 2 Clinical and treatment characteristics of the patient’s cohort

	Training^a [6]	Validation^b (present study)	Difference significance (p value)
Number	162	59
Age—^c yr
Median	65.0	56.0	0.67
Range	28.0–83.0	34.0–75.0
Sex—no. (%)
Male	123 (75.9)	47 (79.7)	0.56
Female	39 (24.1)	12 (20.3)
T stage—^d no. (%)
T2	15 (9.3)	6 (10.2)	0.98
T3	95 (58.6)	34 (57.6)
T4	52 (32.1)	19 (32.2)
N stage—^d no. (%)
N0	9 (5.6)	25 (42.4)	< 0.05
N1	58 (35.8)	24 (40.7)
N2	95 (58.6)	10 (16.9)
MR scanner strength^e
1.5 T no (%)	162 (100.0)	32 (54.2)	< 0.05
3.0 T no (%)	–	27 (45.8)	< 0.05
Interval MRI^e and start CRT^f-mo^e
Median (range)	1.4 (0.0–10.0)	0.9 (0.7–1.0)	< 0.05
Interval end CRT^f and surgery-mo^g
RT short course: median/range	–	0.3 (0.3–0.5)	< 0.05
RT long course: median/range	2.6 (1.1)	1.9 (1.0–2.7)	< 0.05
RT course—no (%)^h
Short (5fr × 5 Gy)	–	19 (32.2)	< 0.05

Statistical tests results investigating significant differences are reported in the last column: chi-square test was performed for categorical variables, Wilcoxon Mann Whitney for continuous ones
^aCohort 1: European Cohort, used for the training and first validation of the model
^bCohort 2: Intercontinental cohort
^cyr years
^dno. number
^eMR/MRI magnetic resonance/magnetic resonance imaging
^fCRT chemoradiotherapy
^gmo. months
^hRT radiotherapy

ISSN: 1748-717X