Determination of the Detectability and the Dosimetric Impact of Plan Delivery Errors in Volumetric Radiotherapy Treatments

Abstract

Purpose: To determine the detectability and the dosimetric impact of plan delivery errors for radiotherapy volumetric modulated arc (VMAT) treatments. Methods: Radiotherapy plans from five anatomical sites (larynx, prostate, prostate and nodes, cervix, SABR lung) for two levels of complexity (average and high) were modified to include simulated errors in MLC position, collimator and gantry angle, and dosimetric output. Clinically likely combinations of these errors were also considered to simulate more real-world deliveries. The original and modified plans were delivered to the Delta4+ patient specific verification system for dosimetric comparison using the gamma method, therefore allowing the minimum detectable errors to be determined. These just detectable errors were then simulated in the treatment planning system, and hence their dosimetric impact could be assessed using the concept of equivalent uniform dose (EUD). Results: At the gamma metric of 3%/3mm, the average clinical minimum detectable error (averaged over all anatomical sites and complexities) was found to be 0.8mm and 2.3mm for MLC field size changes and field offset respectively. At this gamma criterion, the minimum detectable error in collimator angle, gantry angle, and dose offset were found to be 2.4°, 4.0°, and 2.4%. The detectability of clinically likely combinations of errors was found to be poor for all gamma metrics. The dosimetric impact of the just detectable, independent delivery errors was found to be clinically significant for a considerable proportion of target structures, with a change in EUD of greater than 3% found in 24% of PTV volumes. The impact to OAR structures was found to be less profound, with a change in EUD of greater than 2Gy found for only 3% of structures. The dosimetric impact of just detectable combinations of clinically likely errors was also found to be clinically significant, with an even higher proportion of structures showing a change in EUD beyond the clinical threshold, corresponding to 36% and 8% for PTVs and OARs respectively. Conclusions: The Delta4+ patient specific verification system demonstrates poor sensitivity to both independent and combinations of clinically likely simulated delivery errors. The system was not sufficiently sensitive to detect potentially clinically significant delivery errors.

Date of Award	1 Aug 2023
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Nicki Mcgrath (Supervisor) & Geoffrey Budgell (Supervisor)