Early prediction of therapeutic response by functional assessments with magnetic resonance simulator in nasopharyngeal carcinoma

Abstract

Nasopharyngeal carcinoma (NPC) is a prevalent malignancy in Southern China and Southeast Asia. Concurrent chemo-radiation therapy (CCRT) which lasts for about seven weeks is the standard treatment for locally advanced NPC patients. Post-treatment surveillance with conventional morphological imaging can only be done months after the treatments when the therapeutic side effects have subsided and the tumour shrinkage has settled. Additional boost RT treatments might be considered giving to those poor-responders with detected residual tumours. While they had treatment breaks for months, their aggressive tumours might have repopulated which affected the efficacy of boost RT treatments. Functional magnetic resonance imaging (MRI) detects tumour physiological changes which happen much earlier than morphological changes. This study aimed to investigate the values of functional MRI in the early prediction of therapeutic responses for NPC patients receiving CCRT treatments.

A cohort of 27 patients were recruited from the HKSH Cancer Centre (Island East). Functional MRI scans including dynamic contrast-enhanced (DCE-) MRI, diffusion weighted (DW-) MRI and proton magnetic resonance spectroscopy (1H-MRS) were acquired weekly during the CCRT treatments with a MR Simulator dedicated for these patients. Patients were classified as responders (N=23) and poor-responders (N=4) according to their post-treatment nasopharyngeal tumour shrinkage. The Tofts compartment model was employed in the quantitative analyses of DCE-MRI perfusion parameters. The influx rate constant (Ktrans), the reflux rate constant (Kep), the volume of extra-cellular extra-vascular space per unit volume (Ve) and the integrated area-under-curve for the first 60 seconds (iAUC60) of the acquired time-intensity curve (TIC) were studied. Apparent diffusion coefficient (ADC) was used in the DW-MRI analyses. Metabolite intensity peaks were studied in the spectra acquired from 1H-MRS.

In DCE-MRI studies, the iAUC60 values at week 2 of the CCRT were found significantly higher in the poor-responders than the responders (p=0.012). The differences in the percentage increase of iAUC60 from week 1 to week 2 were found borderline statistically significant between the groups (p=0.052). In DW-MRI studies, the percentage increase in ADC values from pre-treatment to week 1 were found significantly higher in the responders than the poor-responders (p=0.027). In 1H-MRS studies, the interpretable rate of MR spectra acquired by the MR Simulator for the nasopharyngeal tumours was 65.8% in our cohort. No significant difference could be found in choline peaks between the two groups.

In conclusion, early physiological changes occurred in the nasopharyngeal tumours during the CCRT. The tumour vascular permeability changes at week 2 and the tumour diffusability changes at week 1 were significantly different between the responders and poor-responders. They might be used as the biomarkers to have early therapeutic response prediction of the CCRT in NPC patients. The multi-parametric maps from functional MRI scans contain physiological information of the tumours which might be feasible to use in personalized adaptive RT for the NPC patients.