Abstract

Youssef Ez-zyouy^1,3, Akos Gulyban¹, Christelle Bouchart², Nick Reynaert¹

¹Institut Jules Bordet, Medical Physics, Brussel, Belgium; ²Institut Jules Bordet, Radiation Oncology, Brussel, Belgium; ³Université Toulouse, Faculté des Sciences et d’Ingénierie, Toulouse, France

Stereotactic ablative radiotherapy (SABR) for pancreatic cancer is challenging due to the proximity of dose limiting organs at risks (OARs). In this study we aimed to 1) determine the daily residual error 2) evaluate anatomical variations 3) recalculate the daily dose distribution based on the residual error and 4) estimate the worst case using deformable registration and dose accumulation on our first series of pancreatic SABR cohort.

Thirteen consecutive patients with pancreatic adenocarcinoma were included in this study. SABR prescription was isotoxic 35 Gy in 5 fractions up to 50 Gy SIB following our local protocol with a 3 mm safety margin. Residual error was determined using the CBCT with the applied correction and were rigidly registered to the planning CT using global (Body) and planning target volume (PTV) masked (Mask) approach with MICE Toolkit (version 1.0.6, NONPI Medical AB, Stockholm, Sweden). This was followed by displacing the isocenter with the corresponding residual error (for both Body and Mask) and recalculating the dose on the planning CT using Monaco TPS (v5.51.2, Elekta AB, Stockholm, Sweden). After a B-spline image based deformation (corrected CBCTs to CT), anatomical changes were determined by visual inspection and quantitative measurements by the Jacobian per each PTV and OAR volumes. Daily and accumulated total DVHs were calculated for Body and Mask based approaches. Worst-case scenarios were assessed by daily dose matrices (up-scaled to full course) and were compared to the planning and accumulated DVHs for each contoured volume. Relevant DVH parameters were compared between the two matching strategies using paired t-test at p<0.05 statistically significant level.

A total of 65 fractions were evaluated. Average residual errors were 2.83 (SD: 2.63) mm and 0.62 (SD: 0.88) mm with the Body and Mask approaches. A total of 130 plan recalculations were performed. Following the deformation, visual inspection showed good overall results and the Jacobian remained within reasonable limits (Figure 1) confirming a plausible deformation for dose deformation and accumulation.

Figure 1. Average and SD of the Jacobians for PTVs and OARs.

Daily dose coverage for the target volumes were maintained. The relevant OARs constraints were also respected without major deviation from the planning values, while no significant differences were found between the Body or Mask based processing (Figure 2). Accumulated dose remained comparable to the planned one as well. 

Figure 2. Daily DVH parameters using Mask and Body based deformation.

We demonstrated the feasibility of daily and accumulated dose determination for pancreas SABR using residual error assessment and deformable registration. The clinical plausibility is granted with visual and quantitative evaluation of deformation. As the accumulated DVHs also met our dose constrains, our results reassuring the safe administration of pancreas SABR using our current margins of 3mm.