Preliminary Results of Developing Imaging Complexity Biomarkers for the Incidence of Severe Radiation Pneumonitis Following Radiotherapy in Non-Small Cell Lung Cancer Patients with Underlying Idiopathic Pulmonary Fibrosis

Abstract

Background: Idiopathic pulmonary fibrosis (IPF) has the potential to cause fatal pulmonary toxicity after radiotherapy and can increase the morbidity and mortality of non-small-cell lung cancer (NSCLC) patients. In this context, we aimed to develop imaging complexity biomarkers to predict the incidence of severe pulmonary toxicity in patients with NSCLC who have underlying IPF and are treated with radiotherapy. Methods: We retrospectively reviewed the medical records of 19 patients with NSCLC who had underlying IPF and were treated with radiotherapy at the Korea University Guro Hospital between March 2018 and December 2022. To quantify the morphometric complexity of the lung parenchyma, box-counting fractal dimensions and lacunarity analyses were performed on pre-radiotherapy simulation chest computed tomography scans. Results: Of the 19 patients, the incidence of grade 3 or higher radiation pneumonitis was observed in 8 (42.1%). After adjusting for age, sex, smoking status, histology, and diffusing capacity of the lung for carbon monoxide, eight patients with a lower fractal dimension showed a significantly higher hazard ratio of 7.755 (1.168–51.51) for grade 3 or higher pneumonitis than those with a higher fractal dimension. Patients with lower lacunarity exhibited significantly lower hazards in all models, both with and without adjustments. The lower-than-median lacunarity group also showed significantly lower incidence curves for all models built in this study. Conclusions: We devised a technique for quantifying morphometric complexity in NSCLC patients with IPF on radiotherapy and discovered lacunarity as a potential imaging biomarker for grade 3 or higher pneumonitis.

Keywords

• non-small cell lung cancer
• idiopathic pulmonary fibrosis
• radiotherapy
• complication
• biomarker