Packer, D. L.

Background: Cardiac radioablation (CRA) is emerging as a treatment modality for refractory ventricular tachycardia (VT). This study aimed to evaluate the effects of radiation on the coronary arteries in a swine model of proton beam CRA. Methods: Eighteen domestic swine underwent 30–40 Gy pencil-beam scanning proton therapy in a single-fraction targeting the left ventricle with no attempt to spare coronary arteries and were euthanized 12-40 weeks later. Contrast-enhanced cardiac computed tomography (CT) was performed at baseline for treatment planning and at study end. Using dose-volume histograms, the maximum point (D max ), mean (D mean ) doses and the minimal doses received by the highest irradiated volumes of 0.01 cc (D 0.01cc ) for the left main, left anterior descending, circumflex, and right coronary artery, as well as the large diagonal and obtuse marginal branches were calculated. The D max location for each artery was mapped from the baseline planning CT onto the follow-up CT using deformable registration. Coronary artery segments were harvested from the D max sites for histological analysis. Receiver operating characteristic (ROC) analysis was performed to identify dose thresholds for predicting histologically severe coronary stenoses (≥75%). Results: Ninety-six coronary artery segments were analyzed. No stenoses were observed by CT imaging before irradiation. By histological analysis post-euthanasia, 25/96 (26%) coronary segments sampled at the Dmax site per artery had ≥75% stenosis. The median D max was 4.7 Gy for the <75% stenosis group and 29.7 Gy for the ≥75% stenosis group (p<0.001). The AUC-ROC for the association between D max and severe stenosis was 92.2%. A D max value of 20.1 Gy in a single fraction best predicted severe stenosis, with sensitivity of 92.3% and specificity of 87.1%. The AUC-ROC for the association between severe stenosis versus D mean and D 0.01cc was 84.8% and 91.6%, respectively. Based on histopathologic analysis, intimal hyperplasia was the most common coronary artery abnormality at the D max sites. Conclusion: In this pre-clinical model of proton beam CRA, coronary stenoses occurred in a dose-dependent manner, with D max being a better dosimetric predictor of stenosis than D mean and D 0.01cc . These data provide for the first time radiation dose constraint information for the epicardial coronary arteries that can be considered in treatment planning for CRA and radiation for thoracic malignancies.

Background: Proton beam irradiation to the ventricular myocardium may enable non-invasive targeting of ventricular tachycardia. Previous work has demonstrated the feasibility of imaging ablation lesion with late gadolinium enhanced (LGE) cardiac magnetic resonance imaging (MRI). Hypothesis: We hypothesize that proton beam irradiation will lead to progressively increasing density of LGE corresponding to lesion formation during follow-up after treatment delivery. Aim: The aim of the current study was to determine the time course of LGE intensity in ventricular myocardium following proton beam therapy. Methods: Eleven swine were irradiated with 40 Gy proton beam in the anterior or lateral left ventricular (LV) followed by LGE cardiac MRI every 4 weeks for up to 24 weeks. After manual contouring of the endocardial and epicardial myocardium, a region of interest (ROI) in the area irradiated with 0-5 Gy was contoured (remote ROI). The mean and standard deviation (SD) of the signal intensity (SI) of the remote ROI were calculated. Light LGE was defined as voxels with SI > mean+2SD but < mean+3SD of the remote ROI, while voxels with SI > mean+3SD of the remote ROI were classified as having dense LGE. Voxels identified as bright due to image artifacts or noise were manually removed. The analysis was conducted in a fashion blinded to swine ID and time point after irradiation. Results: A total of 33 time points were analyzed (3.1 ± 1.4 MRI studies per swine). Figure 1A shows a representative time course of LGE in the same swine, progressing from diffuse light LGE at 8 weeks to dense LGE at 16-24 weeks. Figure 1B demonstrates the proportion of light versus dense LGE across timepoints in irradiated myocardial regions. At 8 weeks, light and dense LGE areas were evenly distributed. At 12 to 24 weeks, dense LGE increased to approximately 75%. The proportion of dense versus light LGE was significantly higher at 12, 16, and 24 weeks, as compared to 8 weeks. Conclusion: In ventricular myocardium, diffuse light LGE evident at 8 weeks progresses to denser LGE at 12-24 weeks following proton beam irradiation corresponding with lesion formation.