316 Background: Due to the low dose tolerance of the organs at risk (OARs) in the abdomen the tumor dose for pancreatic cancer patient is restricted to 50-60 Gy in 1.8-2.0 Gy fractions when combined with chemotherapy. The goal of this study was to develop a system that can determine the minimal radiation dose to the OARs of each individual patient that is achievable while maintaining adequate tumor coverage. This could guide treatment planners to spare the OARs to the fullest extent. When the minimal doses to the OAR are achieved, the total plan can be upscaled until the normal tissue dose constraints are met, allowing for an increase in tumor dose without increased normal tissue toxicity. Methods: The minimal achievable dose to the OARs depends on its proximity to the planning target volume (PTV). The overlap volume histogram (OVH) was used to describe the spatial relation of each OAR to the PTV. A database of 33 patients, treated with IMRT, was queried to find the lowest achieved dose to an organ for any of the prior patients with less favorable PTV-OAR configurations than the current patient. This minimal dose must also be achievable for the OAR of the new patient. For 25 randomly chosen patients the lowest achievable dose to the liver and kidneys was predicted this way. Then the patients were replanned to verify if this dose could be achieved. The new plans were compared to the original clinical plans. Results: After replanning the predicted achievable dose to the liver was realized within 1 and 2 Gy for more than 86% and 96% of the patients respectively. For the kidneys these numbers were 83% and 96%. The average improvement in terms of mean dose was 1.4 Gy (range 0 – 4.6 Gy) for the liver and 1.7 Gy (range 0 – 6.3 Gy) for the kidneys. This would have allowed an increase in PTV dose of on average 5 Gy (range 0-13 Gy) based on the liver and 8.5 Gy (range 0-38 Gy) based on the kidneys compared to the original plan, without an increase in dose to the bowel, cord, and stomach. Conclusions: The lowest achievable dose to the OARs could accurately be predicted for pancreatic cancer patients within seconds. This can guide dosimetrists to spare the OARs or increase the PTV dose by 5 Gy without increased toxicity. [Table: see text]
Normal tissue complication probability (NTCP) modelling using spatial dose metrics and machine learning methods for severe acute oral mucositis resulting from head and neck radiotherapy
