Purpose: To evaluate the suggestions that QA passing rates decrease as a function of plan complexity and the subsequent need for site-specific tolerance limits.
Methods: Patient-specific ionchamber array-based delivery QA analysis for 250 previously treated, non-SBRT VMAT patients analyzed at 3%, 3mm and a threshold of 10% was repeated with gamma criteria of 3%, 2mm and a threshold of 10%. Cases were divided into H&N (n=76), Thorax (40), Abdomen (61), and Pelvis (73) with passing rates plotted as a function of an approximate modulation scaling factor (MSFapprox.). Statistical Process Control (SPC) methods were used to compare the action limit of each subset to the universal action limit (ALuniv) of 90%.
Results: The site-specific action limits differ from the ALuniv by 0.9, -1.1, -2.3 and -4.4% for the Abdomen, H&N, Pelvis and Thorax respectively with MSFapprox. and passing rates being significant between the groups at the 1% and 5% levels, respectively. The passing rates did not reach the same level of significance due to the lower number of thorax cases available for evaluation. The correlation of passing rate as a function of MSFapprox. was “low” for all sites with pelvis having the highest r value at -0.35. The tolerance limits (TL) for the above body sites were 92.8, 91.0, 89.3 and 88.7%, respectively.
Conclusion: Both site-specific TL and AL for our system show some variation. The site order presented above demonstrated difference from ALuniv as well as general increasing MSFapprox.. It is interesting that the Pelvis (GI, GU, GYN) site demonstrated higher plan complexity and lower AL than H&N at our institution which is contrary to what is generally thought. Additionally, correlation with plan complexity being “low” for all suggests that the need for site-specific action levels may be unwarranted.
Not Applicable / None Entered.