Anal Squamous Cell Carcinoma (ASCC) is radically treated by chemoradiotherapy (CRT). The use of DW-MRI as a predictive marker of outcome would enable early individualisation of treatment. We aimed to quantify the changes in mean apparent-diffusion-coefficient (ADCMean) between a DW-MRI at diagnosis and on fraction 8-10 of CRT as a biomarker for cellularity, and correlate these with outcome.
This prospective study recruited patients with ASCC between October 2014 and November 2017. DW-MRI was performed at diagnosis and after fraction 8-10 of radical CRT. A region of interest (ROI) was delineated for all primary tumours and any lymph nodes >2 cm on high-resolution T2 –weighted images and propagated to the ADC map. A pre-defined cut-off for percentage change in ADCMean (ΔADC) was set at < 20%. Complete response (CR) was assessed 3 months following completion of CRT. Routine clinical follow up data from patients were collected from NHS electronic systems.
Twenty-three of 29 recruited patients underwent paired DW-MRI scans. The median (range) tumour volume was 13.6 cm3 (2.8 cm3 to 84.9 cm3). The median ΔADCMean between scans was 20.7% (95% CI: 12.7%, 34.1%). Nine of 23 (43%) patients had a change of ADCMean <20%. Two patients failed to achieve a CR, with ΔADCMean of 14.6% and 20.3%, respectively. On routine follow up, 2 further patients had relapsed, with ΔADCMean of -1.2% and 6.8%.
In a subset of patients, <20% ΔADCMean on DW-MRI between diagnosis and day 8-10 of CRT was observed. Four patients in the study demonstrated persistent or recurrent disease, all of whom demonstrated an ADC on or below the pre-specified cut-off. Further investigation of the predictive merit of DW-MRI change, and the optimum cut-off is needed in larger cohorts.
NCT02145416.
University of Oxford.
CRUK EPSRC Cancer Imaging Centre in Oxford, the CRUK Oxford Centre and the NIHR.M Hawkins is supported by Medical Research Council grant MC_UU_00001/2.
R. Muirhead: Honoraria (self), Fee for lecture: Boehringer Ingelheim. V. Goh: Research grant / Funding (self): Siemens. All other authors have declared no conflicts of interest.