Avoiding Highly Functional Ventilation Zones as Defined by 4D Simulation Decreases the Risk of Radiation Pneumonitis During Proton Radiation Therapy for Lung Cancer

Abstract

Locally-advanced non-small cell lung cancer (LA-NSCLC) patients are treated with chemoradiotherapy (CRT); however, approximately 15-20% of these patients develop significant morbidity with radiation pneumonitis (RP). Proton beam therapy (PBT) may reduce this risk. Currently, the most acceptable dosimetric predictors assume homogenous underlying lung function. We hypothesize that the risk of RP is correlated with dose to more highly ventilated portions of the lung, and PBT could be used to protect these areas of higher function. This study correlates the radiation dose in the low and high ventilation portions of the lung with RP outcome for patients treated with proton radiotherapy. This was a case-control retrospective study. 30 patients with LA-NSCLC treated between 2011 and 2016 with CRT were retrospectively identified, of whom 15 exhibited RP (graded using CTCAE v4.0) after PBT, and 15 were negative controls (matched based on age, smoking , tumor location and total dose). None received prior lung radiation and all satisfied institutional dosimetric constraints (V5<60%, V20Gy <37%, Mean lung dose < 20 Gy ). The inhale and exhale simulation CT scans were deformed using clinically validated deformable registration Advanced Normalization Tools . The 3D lung ventilation maps were derived from the deformation matrix and partitioned into low, medium and high ventilation zones for dosimetric analysis. A two-tailed student’s t test was used to correlate dose in each ventilation zone with risk of RP. The typical dose parameters V5 (37.2 ± 6.1 Gy vs 32.6 ± 7.3 Gy) and V20 (30.4 ± 5.7 Gy vs 27.2 ± 5.4 Gy) were higher for pneumonitis patients, though not statistically significant (p = 0.073, 0.127, respectively). In the low ventilation portion of the lung, similar dose distribution was observed for V20 between the RP (9.6± 2.7 %) and non-RP (9.7± 2.3 %) groups (p=0.914). V5 was also not significantly different (11.9±2.8 % vs 12.1± 2.7 %, p = 0.856). However, in the high-ventilation portions of the lung, RP group had higher V5 (12.5 ± 2.4 % vs. 9.3 ± 4.0 %, p = 0.012) and V20 (10.1 ± 2.4 % vs. 7.3 ± 3.6 %, p = 0.018). Furthermore, the mean dose in the high-ventilation portions of the lung of the RP patients was significantly higher than that of the non-RP patients (16.8 ± 4.4 Gy vs. 11.7 ± 6.1 Gy, p=0.015), while the mean dose in the low-ventilation portion of the lung did not show such differentiation (16.3 ± 5.0 Gy vs. 15.6 ± 4.2 Gy, p = 0.683). There is accruing data demonstrating that radiation dose to the functional high ventilated lung might serve as a predictor of RP. Our data supports these findings and is additionally the first study to demonstrate this in a cohort of patients treated with PBT. This is particularly promising given the greater ability to control dose to low versus high-ventilated portions of the lung using PBT compared to photon based RT . Further work will be needed to validate these findings in a larger cohort and compare these results in patients treated with photons.