Purpose: To investigate changes in imaging biomarkers as precursors to atelectasis post-RT in a swine model and develop a dose-response relationship for bronchial stenosis.
Methods: A cohort of three swine were analyzed. The swine underwent a research course of 60 Gy in 5 fractions delivered to a targeted airway in the left inferior lung. The right lung was used as control as it was left unirradiated. Airway segmentation was performed on the pre- and three months post- radiation therapy (RT) maximum inhale phase of the four dimensional (4D) computed tomography (CT) scans. Segmentation repeatability was assessed by calculating the percent relative standard deviation of the inner cross-sectional area (Aᵢ) for each airway. A decrease in Aᵢ was assumed to be a surrogate of bronchial stenosis.
Results: Repeatability analysis showed good repeatability for both timepoints (r>0.99, small deviations <5%). Airways (n=30) in the right (unirradiated) lung for all swine showed no significant changes in Aᵢ post-RT compared to pre-RT. Airways (n=28) in the left (irradiated) lung of all swine were found to have a decreased Aᵢ post-RT compared to pre-RT, which was significantly (p<<0.001) correlated (Adjusted R² = 0.95) with dose, and independent of pre-RT inner radius. Additionally, while the square root of wall area (i.e. thickness) decreased significantly (p=0.0001) with dose, the wall area percent increased significantly (p=0.04) with dose.
Conclusion: This work shows there is a significant correlation between radiation dose and bronchial stenosis as well as wall area percent. Airway toxicities beyond the proximal bronchial tree have mostly been ignored in treatment planning. These results support the hypothesis that airway dose should be considered during treatment planning in order to preserve functional lung and reduce lung toxicities. Further work should be performed in order to verify these effects in humans and the consequences on pulmonary function.
Funding Support, Disclosures, and Conflict of Interest: This work was supported by the National Institutes of Health grant CA166703.