Assessment of magnetic resonance image compilation (MAGiC) abilities of therapeutic selection and prediction on recurrence risk factors and short-term treatment efficacy in cervical cancer.

Abstract

Objectives: To investigate the feasibility of MAGiC (hereafter, sy-T2WI; T1, T2, and PD maps) in determination of treatment plan and prediction of recurrence risk factors (RRF) and short-term treatment efficacy (STE) in patients with cervical cancer (CC) using hr-T2WI and DWI as reference standards.

Methods: 119 consecutive CC patients who underwent MAGiC, hr-T2WI and DWI were prospectively recruited from October 2021 to March 2024. The subjective evaluation of image quality and tumor staging using sy-T2WI and hr-T2WI was conducted. The accuracy, sensitivity and specificity of sy-T2WI were analyzed for selection of a treatment strategy (stage IB-IIA: surgical operation; stage IIB-IVA: CCRT). RRF and STE were evaluated in staging IB-IIA and IIB-IVA CC patients respectively. Then, the area under the curve (AUC) was used to objectively predict RRF and STE using the quantitative T1, T2, PD maps, their combinations, and apparent diffusion coefficient (ADC).

Results: There was no significant difference of image quality (all P > 0.05), but a strong agreement on tumor staging (Kappa value = 0.935; p < 0.001) between sy-T2WI and hr-T2WI. The accuracy, sensitivity, and specificity of sy-T2WI in deciding treatment strategies were 0.908, 0.908, and 0.999, respectively. Furthermore, the combination of T1 and T2 values was superior to ADC values for predicting RRF (AUC: 0.980 vs. 0.776; p = 0.005) in staging IB-IIA and STE (AUC: 0.982 vs. 0.737; p < 0.001) in IIB-IVA CC subjects.

Conclusions: MAGiC is a promising technique for determination on treatment selection, RRF prediction and STE prognosis in CC patients as its performance is equivalent and even superior to hr-T2WI and DWI.