Advantages of Implementing Sliding Window Gamma Thresholding to Traditional QA Methods

Abstract

Recent recommendations on gamma passing criteria from the AAPM's TG-218/TG-219 for patient quality assurance are crucial in our endeavor for standardization. The need for standardization is important as demonstrated by the TG 218's review of different reported criteria in dose volume comparisons. However, the need for a standardized use of a low dose threshold leaves an important question to be investigated to ensure a high sensitivity for error detection and an increasingly robust analysis of patient quality assurance. The objective of this work is to provide evidence that a multi-level dose threshold gamma evaluation should be further investigated and that implementing the suggested 10% threshold could conceivably wash out failing points in areas of higher importance. Five patients plans computed with collapsed cone in the TPS were exported and computed with an independent calculation-based dose verification system also using a collapsed cone algorithm. In addition, these patients were also measured with a surrogate quality assurance detector. For the independent calculation, a 2%/2mm gamma criteria was used. For the detector measurements a 3%/3mm criteria was used. Both methods were evaluated with a sliding threshold incremented in several steps of maximum dose from 0-98%. Four of the five patients plans evaluated had possible issues with QA performance. Three (60%) of the patient plans, one of which is included in the table below, demonstrated that utilizing a multi-level dose threshold gamma evaluation proved to be a promising tool in identifying plans with potential discrepancies in areas of importance, such as the high dose areas which include the target and abutting OARs. Further investigation and consideration is needed into how the low dose thresholds are utilized when performing gamma comparisons between dose volumes. This work proposes a simple to implement method that will greatly improve how potential errors in treatment planning can be identified. There is promise in future works in progress through our continued work, the work of AAPM TG-360 and requests for vendors to continuously improve the available tools and technology.