Room: Exhibit Hall | Forum 2
Purpose: MRI-guided radiotherapy is rapidly increasing globally. Of interest is real-time, on-table adaptive radiotherapy (ROAR) to improve the sparing of organs at risk (OARs) while providing adequate planning target volume (PTV) coverage. We determined the PTV coverage for SBRT pancreas patients treated with ROAR on the MRIdian MR-linac.
Methods: Five patients were selected, each receiving 5 fractions. Structures, except PTV, were modified based on daily MR imaging. The original plan was re-calculated based on the modified OARs and then re-optimized to always meet OAR constraints, potentially sacrificing PTV coverage. Three metrics were evaluated: 95% coverage of the PTV and GTV by the prescription dose (V95) and PTV mean dose. Metrics were quantified two ways: A dose summation of all 5 fractions per patient in VelocityAI, rigidly registering only to the PTV, and evaluating the composite target DVHs; and tabulating the statistics of the delivered adaptive plans. These metrics were compared between the two methods and to the original plan.
Results: Patients 1-3 had all metrics agreeing to within 8.3% with the original plan. Patients 4-5 had large degrees of OAR shape changes compared to their respective original plans, causing us to substantially sacrifice PTV coverage to meet OAR constraints. Patients 1-4 had all metrics agreeing to within 6.8% when comparing the two methods used to calculate metrics. Patient 5 had larger discrepancies due to difficulty in visualizing the target for the image registration in VelocityAI.
Conclusion: Despite fractional dose variations, PTV metrics remained reasonable compared to the original plan for 3 of 5 patients. This demonstrates that for these patients ROAR successfully spared OARs while meeting original PTV metrics. The discrepancies between the two methods of obtaining the metrics are such that one could forgo the more laborious dose summation, and use the statistics provided in the delivered adaptive plans.