A feasibility study of diffusion weighted imaging and parametric response map analysis for treatment response prediction in nasopharyngeal cancer

Abstract

Diffusion-weighted imaging (DWI) is an MRI technique which provides functional information of tissue by detecting microscopic motion of water molecules. The change of apparent diffusion coefficient (ADC) derived from DWI was used as an imaging biomarker for treatment response prediction in cancers. However, it was based on whole-tumor analysis which did not reflect heterogeneity within the tumor. To overcome this limitation, a new method called parametric response map (PRM) analysis was proposed to evaluate response by quantifying voxel-wise changes in ADC. Here we investigated the use of PRM analysis on ADC from DWI as an imaging biomarker for treatment response prediction in nasopharyngeal carcinoma (NPC) patients. We collected twenty-six patient datasets including twenty complete response (CR) patients and six partial response (PR) patients at King Chulalongkorn Memorial Hospital where one patient dataset consisted of DWI and ADC data acquired before (i.e. pre-treatment) and at five weeks after initiation of chemoradiation therapy (i.e. mid-treatment). For each dataset, we compared pre-treatment ADC image with co-registered mid-treatment ADC image, and calculated PRM+ which was defined as the percentage of voxels with increased ADC values with respect to total voxels within the tumor ROI. To validate the feasibility of the PRM biomarker, we computed the mean and standard deviation (SD) of percentage change in tumor volume (%ΔVol) and in ADC (%ΔADC) and PRM+ across CR and PR patients classified by RECIST1.1 guideline at 6 months. We determined if each of the three biomarker yielded difference between the two patients groups using t-test. To evaluate outcome prediction performance for each biomarker, we constructed the receiver operating characteristic (ROC) and compared with random guessing using Mann-Whitney’s U-test. The results showed that no significant difference in %∆Vol and in %∆ADC between CR an PR groups. In contrast, PRM+ was significantly different between CR and PR groups ( 80.5±8.5% in CR vs 70.2±7.1% in PR, p < 0.05). In terms of prediction performance, PRM+ has higher AUC value than both %ΔADC and %ΔVol (0.817, 0.633, and 0.417 for PRM+, %ΔADC and %ΔVol, respectively). Only PRM+ was significantly different from random guessing (p < 0.05). Our results implied that the proposed PRM+ from ADC could be a potential biomarker for early treatment response prediction in NPC patients.