ESTRO 2022 - Abstract Book

S1408

Abstract book

ESTRO 2022

fraction, emplying quantitative imaging as MR diffusion weighted (DWI). The aim of this study is to evaluate the use of DWI in a low tesla MR-Linac system to assess any tissue variation during the treatment. Materials and Methods Six patients enrolled in the clinical trail (simultaneous integrated boost with two levels: 45 Gy to the pelvis and 55 Gy to Gross Tumor Volume (GTV) plus the corresponding mesorectum) have been included in this preliminary evaluation. This trial foresees a primary tumor boost (60.1 Gy) from the third week of the treatment (from fraction number 11) if patient shows an Early Regression Index (ERI) higher than 13.1, which correspond to a prediction of not complete response. DWI is acquired after treatment delivery on the central GTV slice during the pretreatment simulation and each five fractions. DWI sequences has been optimized for low tesla MR scanner accurately selecting parameters as: direction of the diffusion (only slice direction), number of measurements, artifact noise threshold reduction and number of b-value employed for the calculation of Apparent Diffusion Coefficient (ADC) map. ADC map is then generated using ImageJ software using a mono- exponential model. Results have been analysed in terms of mean ADC values variation along the fractions for each patients. ADC values has been computed not only on the GTV but also on the right femoral head (FH) as reference for ADC constancy. Results Among the six patients enrolled in this preliminary analysis, only two result to be not responder (figure 1, continuous line). DWI were acquired within a maximum time of seven minutes, with quite good compliance of the patients. ADC maps have been computed off-line and analysed for each fraction (figure 1) in terms of ADC values. GTV ADC value standard deviation (SD) is 0.6: 0.4 and 0.8 for “responder” and “not responder”, respectively. Similarly, if considering right femoral head, the SD is 0.08, while for “responder” and “not responder” is 0.09 and 0.05 rispectively. Figure 2 shows the ADC percentage difference of the GTV with respect to the simulation ADC value. ADC percentage difference between fraction 10 and 15 is always within 66% for “responder” and higher than 99% for “not responder”.

Conclusion Even obtained using a reduced number of cases, these preliminary results seem to highlight a different behaviour of the GTV mean ADC values between “responder” and ”not responder” patients included in this clinical trial. ADC values can potentially detect and quantify GTV tissues variations caused to the ongoing treatment in a low field MR-Linac system. Further analysis of the remaining 56 patients that will be enrolled in the clinical trail, could confirm this promising preliminary results.