Study on the Impact of Tumor Characteristics on Acuros Xb Dose Calculation Algorithm Accuracy for Lung Stereotactic Body Radiation Therapy (SABR)
Abstract
Purpose
To validate the Acuros XB dose calculation algorithm for lung Stereotactic Body Radiation Therapy (SBRT) by comparing it with the Anisotropic Analytical Algorithm (AAA) and Monte Carlo calculation, and to assess the influence of tumor features and location on the accuracy of algorithm.
Methods
Twenty-five lung SBRT patients chosen with key diversity metrics such as tumor size, location, mean density and dose prescription were retrospectively analyzed. Clinical plans made with AAA were recalculated using Acuros XB in Eclipse (version 18) treatment planning system. Plans were evaluated under two conditions: same monitor units and renormalized to planning target volume (PTV) D95. Dose metrics for the PTV and OARs, were extracted using custom C# scripts. Conformality metrics (V50, D2cm), dose coverage (D95, D99), and MU differences were analyzed. 3D Monte Carlo calculation was used as second check, and the reference point doses were compared. Statistical analysis was performed using t-tests to assess significance.
Results
Acuros XB demonstrated systematic differences compared to AAA, with lower predicted target doses and higher MUs. Mean differences between algorithms for renormalized plans were 4.66% for V50, 1.97% for D2cm, and -1.11% for D99. Same MU plans showed reduced differences. Acuros plans exhibited a higher frequency of unmet chest wall constraints in renormalized cases. The average MU difference between algorithms was 1.81%. Compared with Monte Carlo, Acuros showed closer agreement than AAA for both primary (-0.31% vs. -1.54%) and secondary reference point doses. Tumor volume had a statistically significant effect on dose differences, while mean HU had minimal influence.
Conclusion
Acuros XB demonstrated improved dosimetric accuracy and closer agreement with Monte Carlo calculations compared to AAA for lung SBRT. Influence of tumor features and location on the algorithm was studied, and tumor volume was found to be a significant factor on dose calculation accuracy.