Prediction of early response to concurrent chemoradiotherapy in cervical cancer: Value of multi-parameter MRI combined with clinical prognostic factors

Abstract

PurposeThis study aimed to investigate the prediction of early response to concurrent chemoradiotherapy (CCRT) through a combination of pretreatment multi-parametric magnetic resonance imaging (MRI) with clinical prognostic factors (CPF) in cervical cancer patients. MethodsEighty-five patients with pathologically confirmed cervical cancer underwent conventional MRI, intravoxel incoherent motion diffusion weighted imaging (IVIM-DWI), and dynamic contrast-enhanced MRI (DCE-MRI) before CCRT. The patients were divided into non- and residual tumor groups according to post-treatment MRI. Univariable and multivariable analyses were performed to pretreatment MRI parameters and CPF between the two groups, and optimal thresholds and predictive performance for post-treatment residual tumor occurrence were estimated by drawing the receiver operating characteristic (ROC) curve. ResultsThere were 52 patients in non- and 33 in residual group. The residual group showed a lower perfusion fraction (f) value and volume transfer constant (Ktrans) value, a higher apparent diffusion coefficient (ADC) value, diffusion coefficient (D) value and volume fraction of extravascular extracellular space (Ve) value, and a higher stage than the non-residual tumor group (all P < .05). D, Ktrans, Ve and stage were independent prognostic factors. The combination of D, Ktrans and Ve improved the diagnostic performance compared with individual MRI parameters. A further combination of these three MRI parameters with stage exhibited the highest predictive performance. ConclusionsPretreatment D, Ktrans, Ve and stage were independent prognostic factors for cervical cancer. The predictive capacity of multi-parametric MRI was superior to individual MRI parameters. The combination of multi-parametric MRI with CPF further improved the predictive performance.