Direct Comparison of Tumor-Informed and Tumor-Naive Circulating Tumor DNA Assays for Recurrence Detection in Early-Stage NSCLC.

Background: Circulating tumor DNA (ctDNA) is a promising prognostic biomarker in early-stage NSCLC. The typical tumor-informed method of ctDNA testing requires tissue specimens of high quality, which is a challenge in developing countries. Tumor-naive approach is an alternative, but the performance of the two methods has not been directly compared using the same samples and platform. Methods: We retrospectively analyzed tumor and blood samples of patients with early-stage NSCLC enrolled in our published study. For analytical performance, pretreatment samples of 42 patients and 50 healthy donors were used to assess the accuracy of ctDNA detection. For clinical performance, 176 postsurgical blood samples of 76 patients with NSCLC were analyzed to compare the ctDNA status with recorded clinical recurrence. The tumor-informed method evaluated personalized mutations and a fixed 500-hotspot panel, whereas the tumor-naive method combined predesigned mutation panels and nonmutation genome-wide features of ctDNA. Results: The tumor-informed assay had 66.7% sensitivity and 99.3% specificity for detecting ctDNA in early-stage NSCLC, higher than the tumor-naive assay with 52.6% sensitivity and 95.7% specificity. Postsurgical ctDNA status determined by both methods had significant prognostic value to predict recurrence ahead of clinical diagnosis (hazard ratio >100, Conclusions: Both methods exhibited high accuracy in detecting residual cancer in NSCLC. The tumor-naive approach is a reliable alternative when high-quality tissue specimens are unavailable.