2426 - An Investigation on the Relationships of Target Irregularities, Plan Complexities and Patient-Specific Quality Assurance Pre- and during Treatment

X. Tan¹, H. Luo¹, F. Jin², and Y. Wang¹; ¹Radiation Physics Center, Chongqing University Cancer Hospital, Chongqing, China, ²Department of Radiation Oncology, Chongqing University Cancer Hospital & Chongqing Cancer Institute & Chongqing Cancer Hospital, Chongqing, China

Purpose/Objective(s):: Both target irregularities (TIs) and plan modulation can influence plan complexities (PCs), subsequently affecting quality assurance (QA), so it’s important to investigate the correlations among TIs, PCs, and QA pre- and during treatment. Materials/

Methods: 203 radiotherapy plans were analyzed, including fifty-six nasopharynx, thirty-nine breast, forty esophagus, thirty-six lung, and thirty-two cervix cases. TIs were assessed by a surface area-to-volume ratio, and PCs were delineated using eighty complexity metrics established previously. Dose was calculated by anisotropic analytical algorithm and compared with collapsed cone convolution (i.e., CQA). Measurement-based dose verification (MQA) utilized an "O"-shaped 3D diode array before treatment. An in-vivo QA (ProQA) compared planning doses with ones reconstructed from planning CT images and three fractional log files selected randomly during treatment. Relationships among TIs, PCs, and QAs were analyzed through Spearman correlation and principal component linear regression.
Results: TIs primarily influenced MLC aperture and control points-related PCs, and MU and MLC aperture-related PCs significantly influenced QAs. Most of dominant PCs decreased in lung and esophagus but increased in nasopharynx with rising TIs, and no notable pattern was observed in others. Also, PCs impacts on QA varied with dose difference and distance-to-agreement. There were positive correlations between TIs and QAs in lung, esophagus, and cervix, and a negative correlation in nasopharynx, as well as no significant correlation in breast. Meanwhile, the maximum correlation coefficients for TIs with CQA, MQA, and ProQA were 0.574, 0.310, and 0.536, respectively. ProQA demonstrated close correlations with most metrics, particularly MLC aperture-related metrics, and outperformed CQA and MQA when predicted by PCs (maximum R²: 0.848 vs. 0.637, 0.431) or PCs+TIs (maximum R²: 0.845 vs. 0.636, 0.433).
Conclusion: TIs have demonstrated effectiveness in assessing PCs and QAs. Futhermore, ProQA exhibits superior accuracy in reflecting both TIs and PCs when compared to CQA and MQA.