ESTRO 2021 Abstract Book

S292

ESTRO 2021

University Hospital, Department of Radiology, Lørenskog, Norway; 4 Akershus University Hospital, Department of Pathology, Lørenskog, Norway; 5 Oslo University Hospital, Department of Diagnostic Physics, Division of Radiology and Nuclear Medicine, Oslo , Norway; 6 Sunnmøre MRI-Clinic, Institute of Radiology, Ålesund, Norway; 7 Akershus University Hospital , Department of Oncology, Lørenskog, Norway; 8 Norwegian Institute of Science and Technology, Department of Physics, Trondheim, Norway Purpose or Objective Functional MRI to assess tumour microenvironmental characteristics in rectal cancer has shown promise as a tool for treatment stratification. In rectal cancer patients given neoadjuvant radiation, we aimed to identify high- and low risk cases by multi-echo dynamic contrast MRI and diffusion-weighted (DW) MRI before and after treatment. Materials and Methods 24 rectal cancer patients received standard neoadjuvant treatment, consisting of either short-course radiotherapy (5 Gy x 5) (n = 6) or long-course chemoradiotherapy (2 Gy x 25) (n = 18). Patients were scanned with a multi-echo dynamic contrast MRI sequence, a DW sequence and the standard diagnostic MRI before and after the radiation. The multi-echo sequence was used to estimate both the T 1 - and the T 2 * -curve during contrast agent passage (Dotarem® 279.3 mg/ml, Guerbet Roissy, France). This allowed us to both perform Tofts kinetic modelling of T 1 -weighted dynamic contrast enhanced curves to estimate K trans , k ep and v p , and to analyse T 2 * -weighted dynamic susceptibility contrast curves with a standard deconvolution approach to estimate blood flow (BF) and area under the contrast enhancement curve. In addition, we estimated the apparent diffusion coefficient from the DW data (b-values = 0, 25, 50, 100, 500 and 1000 s/mm 2 ). Scanning was performed on a Philips Achieva 1.5 Tesla system (Philips Healthcare, Best, The Netherlands). Tumour delineations were done by two radiologists on T 2 -weighted images with the DW sequence as an extra guidance. The tumour regions were then semi-automatically co-registered to the parametrical images before the median tumour values were extracted. From 11 surgical specimens, mid-tumour sections were stained with the endothelial marker CD34, and an automated Matlab program estimated the percentage of CD34 staining each. ypT-status and tumour regression grade (TRG) were determined by a pathologist. The statistical tests used were univariable Cox regression with continuous variables for overall survival (OS), Mann Whitney U-test for ypT-status and TRG, and Spearman’s rank correlation for correlation to CD34 percentage. Median follow-up time was 50 months, range (37-74). Results Patients who had no or minor decrease in BF during the radiation were at higher risk of shortened OS (Hazard ratio = 1.02 [1.01-1.04], p-value = 0.015). Figure 1 shows a Kaplan-Meier plot where patients with less than 20% decrease in BF had poor OS. Example images are given in Figure 2. No other MRI parameters were related to OS and no MRI parameters were significantly related to the surrogate markers for OS, TRG and ypT status. BF after radiation was correlated to percentage of CD34 staining in the tumour specimen (rho = 0.89, p < 0.001).