Session Item

To evaluate differences in dosimetry and to apply and compare Normal Tissue Control Probability (NTCP) models in rival photon Volumetric Modulated Arc Therapy (VMAT) and Pencil-Beam Scanning Intensity Modulated Proton Therapy (PBS-IMPT) plans for left-sided breast cancer (LSBC) patients (pts).

A dosimetric analysis on VMAT and  IMPT plans was performed. Both plans were generated on the simulation CT scans, acquired in  Deep Inspiration Breath Hold (DIBH).  Relevant dose metrics for organs at risk (OARs) and targets were retrieved from dose-volume histograms (DVHs).  Target coverage was evaluated in terms of the volume receiving 95% of the prescription dose (V95 %). VMAT plans were optimized on PTV, elaborated on Eclipse treatment planning system (TPS) and calculated using the Analytical Anisotropic Algorithm. IMPT plans were optimized on CTV with Raystation TPS and calculated using Monte Carlo algorithm, applying robustness settings to the target to account for set-up and range uncertainties (5 mm isotropic and 3-5% range, respectively). Relative Biological Effectiveness (RBE) was assumed to be 1.1. Ten NTCP models for selected toxicities were then applied to both rival plans for all pts. Paired, 2-tailed Wilcoxon signed rank test was employed to evaluate median differences in OARs dosimetry and NTCP values in VMAT versus IMPT plans (∆NTCP PH-PT).

Twenty-two CT scans of LSBC pts were performed from April 2022 to July 2022. Target volume consisted of the breast gland in 13 patients (59%) and the chest wall, with or without reconstruction, in 9 patients (41%). Loco-regional irradiation was delivered in 9 patients (41%). Prescribed dose range was 40.05-42.4 Gy/15-16 fractions . Tumour bed boost was delivered in 7 patients, specifically a simultaneous boost (up to 48 Gy[RBE]) in 5 patients (71 %) and a sequential boost in 2 patients (29%) which was up to 16 Gy[RBE] due to positive margins. V95% was equal to or higher than 95% of the target volume in all plans. Median differences in dosimetric parameters for selected OARs were all statistically significant and in favour of protons (P <0.05) except for thyroid V20 (absolute volume of gland receiving a dose ≤ 20 Gy[RBE]) which was comparable in both rival plans. All median ∆NTCP PH-PT for the studied endpoints  were statistically significant and in support of  IMPT (P <0.05), with the exception of ∆NTCP for hypothyroidism (P >0.05) and radiation dermatitis ≥ grade 2, which was significantly higher for IMPT (median value 43,27 %  versus  11,25 % for VMAT). Median ∆NTCP PH-PT are shown in Table 1. 

In clinical practice, VMAT DIBH for LSBC pts is a satisfactory option to spare OARs while ensuring optimal target coverage. In this analysis, IMPT plans provided a statistically relevant reduction of OARs doses leading to significant differences for almost all the NTCP model considered. Nevertheless, clinical validation in a bigger dataset of pts is warranted to confirm these results.