Fig. 3

Bar graph of the ctDNA-positive rate at baseline (upper row) and during serial follow-up (lower row) according to clinicopathologic and ctDNA assay-related factors. The recur site (1st column, ‘Recur Site’) had significant association with ctDNA-positive rate. The tumor subtype of the primary tumor (2nd column, ‘Initial Subtype’) and disease-free interval (3rd column, ‘Disease-Free Interval’), the interval between primary diagnosis and recurrence, were not related to ctDNA-positive rate. Assays that applied more anchor mutations (4th column, ‘Number of Anchor Mutations’) presented with higher ctDNA-positive rate. Using only primary tumor tissue vs both primary and metastatic tissue to select anchor mutations (5th column, ‘Primary vs Paired Tissue’) was also not related to ctDNA-positive rate. HR + HER2− hormone receptor-positive, HER2-negative, HER2+ HER2-positive, DFI disease-free interval, F/U follow-up, De novo de novo metastatic cancer.