Clinicopathological predictors of the presence of blood circulating tumor DNA in early-stage non-small cell lung cancers.

The implementation of lung cancer screening programs across the world has drawn considerable attention to improve early-stage lung cancer detection and prognostication. Several blood-based assays detecting circulating tumor DNA (ctDNA) recently emerged as non-invasive methods to detect malignancies. However, their limited sensitivity and predictive value remain a hurdle to their clinical use. Our objective was to evaluate the association between clinicopathological parameters and pre-surgical ctDNA detection in clinical stage I non-small cell lung cancer patients to further understand the ctDNA shedding biology. The cohort included 180 adenocarcinomas (LUAD) and 80 squamous cell carcinomas (LUSC) stage I patients who underwent lung cancer resection. Patients' clinical and pathological features were collected. A multi-cancer early detection test (GRAIL LLC) was used to detect ctDNA using targeted methylation patterns. Association between the cell-free DNA tumor methylated fraction (TMeF) and the clinicopathological predictors was evaluated using univariate and multivariate modeling. LUSC was associated with a higher TMeF than LUAD. Pathological stage, tumor grade and tumor volume were key determinants of ctDNA detection in both LUSC and LUAD. In LUAD, ctDNA detection also correlated with histological pattern composition, necrosis, acute inflammation, and, to a lesser degree, with spread through alveolar spaces (STAS) and lymphovascular invasion. Based on our results, we propose classification methods for both LUAD (using histological pattern composition) and LUSC (using tumor grade and pathological stage) to identify patients likely to have high ctDNA levels. These results confirm previous findings and suggest that previously unidentified factors, including histological pattern composition and acute inflammation, influence ctDNA levels. These results will help in understanding the ctDNA shedding process and may allow identification of patients eligible for ctDNA-detection-based follow-up.