An<sup>18</sup>F-FDG PET/CT and Mean Lung Dose Model to Predict Early Radiation Pneumonitis in Stage III Non–Small Cell Lung Cancer Patients Treated with Chemoradiation and Immunotherapy
Maria Thor, Chen Lee, Lian Sun, Purvi Patel, Aditya Apte, Milan Grkovski, Annemarie F. Shepherd, Daphna Y. Gelblum, Abraham J. Wu, Charles B. Simone, Jamie E. Chaft, Andreas Rimner, Daniel R. Gomez, Joseph O. Deasy, Narek Shaverdian
Abstract
Radiation pneumonitis (RP) that develops early (i.e., within 3 mo) (RP<sub>Early</sub>) after completion of concurrent chemoradiation (cCRT) leads to treatment discontinuation and poorer survival for patients with stage III non–small cell lung cancer. Since no RP<sub>Early</sub> risk model exists, we explored whether published RP models and pretreatment <sup>18</sup>F-FDG PET/CT–derived features predict RP<sub>Early</sub>. <b>Methods:</b> One hundred sixty patients with stage III non–small cell lung cancer treated with cCRT and consolidative immunotherapy were analyzed for RP<sub>Early</sub>. Three published RP models that included the mean lung dose (MLD) and patient characteristics were examined. Pretreatment <sup>18</sup>F-FDG PET/CT normal-lung SUV featured included the following: 10th percentile of SUV (SUV<sub>P10</sub>), 90th percentile of SUV (SUV<sub>P90</sub>), SUV<sub>max</sub>, SUV<sub>mean</sub>, minimum SUV, and SD. Associations between models/features and RP<sub>Early</sub> were assessed using area under the receiver-operating characteristic curve (AUC), <i>P</i> values, and the Hosmer–Lemeshow test (pHL). The cohort was randomly split, with similar RP<sub>Early</sub> rates, into a 70%/30% derivation/internal validation subset. <b>Results:</b> Twenty (13%) patients developed RP<sub>Early</sub>. Predictors for RP<sub>Early</sub> were MLD alone (AUC, 0.72; <i>P</i> = 0.02; pHL, 0.87), SUV<sub>P10</sub>, SUV<sub>P90,</sub> and SUV<sub>mean</sub> (AUC, 0.70–0.74; <i>P</i> = 0.003–0.006; pHL, 0.67–0.70). The combined MLD and SUV<sub>P90</sub> model generalized in the validation subset and was deemed the final RP<sub>Early</sub> model (RP<sub>Early</sub> risk = 1/[1+e<sup>(−</sup><i><sup>x</sup></i><sup>)</sup>]; <i>x</i> = −6.08 + [0.17 × MLD] + [1.63 × SUV<sub>P90</sub>]). The final model refitted in the 160 patients indicated improvement over the published MLD-alone model (AUC, 0.77 vs. 0.72; <i>P</i> = 0.0001 vs. 0.02; pHL, 0.65 vs. 0.87). <b>Conclusion:</b> Patients at risk for RP<sub>Early</sub> can be detected with high certainty by combining the normal lung’s MLD and pretreatment <sup>18</sup>F-FDG PET/CT SUV<sub>P90</sub>. This refined model can be used to identify patients at an elevated risk for premature immunotherapy discontinuation due to RP<sub>Early</sub> and could allow for interventions to improve treatment outcomes.