Stereotactic Ablative Body Radiotherapy (SABR), when delivered to the lung, provides several unique challenges from a quality assurance (QA) standpoint. To date, there has been little research performed on the efficacy of QA at detecting significant dosimetric errors in organs at risk (OARs). In addition, the feasibility of using the MapCHECK phantom housed in the MapPHAN accessory for SABR QA is not fully appreciated. The aim of this study is to investigate the sensitivity and specificity of MapCHECK in MapPHAN to introduced errors in conformal arc SABR plans, with a specific focus on OAR dosimetry. Multi-Leaf Collimator (MLC) class shift errors up to 2 mm, and isocentre shift errors up to 1 mm, were introduced to 26 simulated 6 MV X-ray conformal arc lung SABR plans on Raystation. 10 of these plans had errors introduced, and were delivered on a Varian Clinac iX linear accelerator to the MapCHECK phantom housed in MapPHAN. In order to commission the phantom for this project, the inherent angular dependency of the MapCHECK diodes had to be corrected to within ± 2%. Gamma analysis was used to compare the measurements on MapCHECK to the simulated Raystation plans, using 1%/1 mm, 2%/1 mm, 2%/2 mm, 3%/2 mm, 3%/3 mm and 5%/1 mm gamma criteria. Based on these results, Receiver Operating Characteristic (ROC) curves were generated to determine the sensitivity and specificity of MapCHECK in MapPHAN to the introduced errors. These ROC curves were then used to determine an optimal set of gamma criteria for use in QA.

It was found that the angular dependency of the diodes in MapPHAN can be reduced to within ± 2% for the central section of the phantom, making it viable for absolute dosimetric measurements of SABR plans. In general, the sensitivity and specificity of the phantom to introduced errors was highest when using 5%/1 mm gamma criteria, although other criteria provided superior accuracy at certain gamma thresholds, with 2%/1 mm being the best alternative criteria. The phantom was most sensitive to introduced MLC errors, and exhibited poor sensitivity and specificity to introduced isocentre shift errors. This was exacerbated when automatic shifts were introduced to correct for setup errors during gamma analysis.

In summary, MapCHECK in MapPHAN is a viable alternative to film for patient-specific QA of lung SABR plans. It provides sufficient resolution to take measurements in high dose gradient regions, and allows absolute dosimetry to be performed following angular dependency corrections. Based on the results obtained here, a combination of 5%/1 mm with 95% threshold and 2%/1 mm with 90% threshold would provide good sensitivity and specificity for QA of conformal arc SABR plans. It is probable that placing restrictions on the shifts included in SNC Patient, or removing them entirely, would significantly improve detection of misalignment errors.