A T2-weighted MRI-based radiomic signature for disease-free survival in locally advanced cervical cancer following chemoradiation: An international, multicentre study

•An MRI-based radiomic signature prognostic of DFS in cervical cancer was developed.•The radiomic signature was validated using an external cohort.•Gene set enrichment analysis revealed the biological basis of the radiomic signature. To develop and validate a T2-weighted magnetic resonance imaging (MRI)-based radiomic signature associated with disease-free survival (DFS) in locally advanced cervical cancer. The study comprised a training dataset of 132 patients (93 Norwegian; 39 The Cancer Imaging Archive (TCIA) and an independent validation Canadian dataset of 199 patients with FIGO stage IB-IVA cervical cancer treated with chemoradiation. Radiomic features were extracted using PyRadiomics. A radiomic signature was developed based on a multivariable radiomic prognostic model for DFS built using the training dataset, with minimal redundancy maximum relevancy feature selection method. Univariate and multivariable Cox regression analyses were then conducted to examine the association of the derived radiomic signature with DFS. A radiomic signature was prognostic for DFS in the training cohort (Norwegian hazard ratio [HR] 5.54, p = 0.002; TCIA HR 3.59, p = 0.04). The radiomic signature remained independently associated with DFS (HR 3.70, p = 0.004) when adjusted for stage and tumor volume. The radiomic signature was also prognostic for DFS in the validation cohort, both on univariate analysis (HR 2.22, p = 0.003), and multivariable analysis adjusted for stage and tumor volume (HR 1.84, p = 0.04). The 4-year DFS rates of patients with radiomic signature score > 0 vs ≤ 0 were 48.2 % vs 87.9 %, and 56.4 % vs 80.8 % for training and validation cohorts respectively. An MRI-based radiomic signature can be used as a prognostic biomarker for DFS in patients with locally advanced cervical cancer undergoing chemoradiation.