Near-Surface Dose Correlates With Toxicity in Patients Receiving Pencil Beam Scanning Intensity Modulated Proton Beam Breast Irradiation

Abstract

Moist desquamation (MD) and breast/chest wall (CW) pain are common, important acute toxicities associated with breast irradiation (RT). In this study, we correlate near-surface dose (NSD) with acute toxicity in patients receiving pencil beam scanning intensity modulated proton breast RT (PBSIMPT).We reviewed thirty-one patients at a single institution who received PBSIMPT for breast cancer. All patients received PBSIMPT for the initial chest wall or whole breast treatment; all but two patients received regional nodal irradiation. 25 patients were treated to 50 Gy in 25 fractions and 6 patients were treated to 42.56 Gy in 16 fractions. Boosts to the lumpectomy cavity or mastectomy scar were delivered with proton, electron, or photon RT. A near-surface rind volume (SR3), which was bound by the skin surface and 3 mm and limited by the breast or chest wall PTV, was generated to be used prospectively during treatment planning to limit near-surface dose while maintaining appropriate PTV coverage. For this analysis, we retrospectively generated two additional near-surface rind volumes, bound by either the skin surface and 5 mm (SR5), or by 5 mm and 10 mm (SR10) from the skin surface. We determined mean SR doses of each patient's composite plan. Skin rind doses were converted to percent of pre-boost prescription dose to account for differences in prescription dose. The presence of any breast/CW MD, and grade 2/3 ("moderate" or "severe") breast/CW pain during RT and within the first two months of RT were recorded. Mann-Whitney U tests were performed to determine correlations between mean SR doses and toxicity endpoints.The mean composite SR3 dose was 4484 cGY (range 4030-5267) for conventionally fractionated patients and 3983 cGY (range 3568-4040) for hypofractionated patients. Five patients (16.1%) experienced MD, and 6 patients (19.3%) experienced grade 2/3 breast/CW pain. Mean composite SR3 dose converted to percent of prescription dose was 94.6% for patients with grade 2 or higher breast or CW pain, compared to 89.6% for those without (P = 0.036). This difference persisted for mean SR5 dose (96.1% versus 91.1%, P = 0.023) and mean SR10 dose (102.4% versus 99.9%, P = 0.012). Patients with MD had a mean SR3 dose of 93.5% versus 91.0% for those without (P = 0.7), and this trend weakened with mean SR5 dose (94.7% versus 92.9%, P = 0.78) and mean SR10 dose (100.7% versus 100.3%, P = 1.0).Higher NSD significantly correlated with increased breast/CW pain. Further analysis with a larger sample size is warranted to explore correlations between NSD and other toxicity endpoints. Optimization of NSD may reduce toxicity in patients receiving PBS IMPT for breast cancer.