Radiobiological assessment of dose-to-medium or dose-to-water with Acuros XB algorithm compared with Anisotropic Analytical Algorithm for lung cancer radiotherapy- What should we know to manage the transition?

Abstract

Purpose: To track the dosimetric changes for similar dose prescriptions, when dose calculation algorithms are upgraded in the treatment planning system (TPS). Clinically significant representations of the treatment outcomes are used to provide interpretable data for radiation oncologists, as the equivalent uniform dose (EUD), the tumor control probability (TCP), the late toxicity as normal tissue complication probability (NTCP) and the uncomplicated tumor control probability (UTCP) scores. Results are presented and discussed in a clinical perspective. Methods: Ten lung cancer patients were included in this study. For each patient, five treatment plans were generated. The doses were calculated using Anisotropic Analytical Algorithm (AAA) and both Acuros XB (AXB) dose reporting modes: dose-to-medium AXB D(m,m) and dose-to-water AXB D(w,m). In plans 1, 2 and 3, the doses were calculated respectively with AAA, AXB D(m,m) and AXB D(w,m) using exactly the same prescription dose and beam set-up. The doses in plans 4 and 5 were calculated using both AXB dose reporting modes using, as input, the same number of monitor units (MUs) as yielded by AAA, with the same beam set-up. The EUD, TCP and NTCP were computed using the assumed radiobiological parameters from literature. The Wilcoxon paired test was used to calculate p-values. Results: Using the same prescription dose, TCP values were higher with AXB than with AAA, and corresponding UTCP scores were 1-2% better with p < 0.05. In addition, absolute NTCP values were slightly increased with AXB. Both AXB dose reporting modes yielded comparable lower TCP and NTCP values (again in the order of 1-2%) than with AAA, when using same MU numbers as with AAA. Conclusion: Compared to AAA, taken as reference, both AXB dose reporting modes yielded better results. AAA showed very close values to AXB D(w,m), but it is difficult to give recommendation between D(w,m) and D(m,m) yet, due to the lack of recommended radiobiological parameters associated with these dose reporting modes. We suggest doing experimental and modelling studies to determine the real radiobiological effects in both targets and organs at risks. Should the differences be substantial in some conditions and relevant to clinical practice, discussions regarding dose prescription and optimization of the tolerance doses to OAR should be undertaken between medical physicists and radiation oncologists.