Dosimetric impact of intrafraction motion during moderate hypo-fractionated prostate cancer radiotherapy treatment

Purpose: To investigate the impact on dose distribution of intra-fraction motion occurred during moderate hypo-fractionated prostate cancer treatment and to predict minimum anisotropic treatment margins required to take into account the motion. Materials and Method: Prostate intra-fraction 3D translations were evaluated in 2400 sessions from 96 prostate cancer patients treated by radiation therapy in our institution with intra-fraction motion monitoring thanks to the Clarity transperineal ultrasound (TP-US) probe (Elekta AB, Stockholm). All patients underwent volumetric modulated arc therapy. Treatment durations ranged between 200-600s. The prescribed dose was 60 Gy delivered in 20 fractions to the Clinical-Target-Volume (CTV) prostate and the treatment goal was that 100% of the prescribed dose must cover 99% of the CTV prostate. For 19 randomly selected patients of the cohort (380 sessions) treatment plans were recomputed using 0 mm CTV-to-PTV margins, with an auto-planning optimization algorithm implemented in a research version of Monaco TPS (version 5.59.11, Elekta AB, Stockholm). This enabled assessing the impact of intrafraction motion on CTV coverage without having operator variability in the planning process which could bias the conclusions. Then, the voxel shifting method was used to move the CTV prostate structure every second of treatment, according to the movements retrieved by the TP-US. The global shifting was then applied to the dose distribution maps. Finally, Dose-Volume Histograms (DVH) of the perturbed dose distributions were computed. Results: Mean intra-fraction prostate displacements observed were (0.02±0.39) mm, (0.38±0.74) mm and (-0.71±0.64) mm in left-right, supero-infero and antero-posterior directions, respectively. Mean displacements >3, 5 and 7 mm were observed for 16%, 5%, and 1% of the total number of fractions. Rare large deviations >8 mm were detected but the frequency was less than 1%. On average, the largest displacements were observed in inferior and posterior directions. After applying the voxel shifting method, the prescription constraint on the CTV prostate was achieved for ≤ 57% of the fractions for half of the cohort, whereas for 5 patients, the constraints were respected for more than 94% of the fractions. DVH showed that the motion influence the dose amount to the organs at risk: increment in rectum and bladder Dmean of 3.0 Gy and 1.5 Gy were respectively registered. Conclusions: Prostate movements impact the dose distribution and are more important in posterior and inferior directions. Hence, anisotropic margins would be required to optimally take into account intrafraction motion. Further work is in progress to evaluate these margins.