Additional file 1 of PRIME: an interpretable artificial intelligence model based on liquid biopsy improves prediction of progression risk in non-small cell lung cancer

Additional file 1. Table S1 Baseline characteristics of patients from all cohorts (n = 493). Table S2 Clinical, treatment and ctDNA data for patients in the NCC-1 cohort (n = 105). Table S3 Clinical, treatment and ctDNA data for patients in the NCC-2 cohort (n = 103). Table S4 Clinical, treatment and ctDNA data for patients in the Stanford cohort (n = 37). Table S5 Clinical, treatment and ctDNA data patients in the LUCID cohort (n = 100). Table S6 Clinical, treatment and ctDNA data for patients in the TRACERx cohort (n = 100). Table S7 Clinical, treatment, and ctDNA data for patients in the Moding et al. cohort (n = 48). Table S8 Mutational data of patients in the NCC-1 cohort (n = 105). Table S9 Mutational data of patients in the NCC-2 cohort (n = 103). Table S10 Comparison of baseline characteristics of patients in the training set and validation set. Table S11 The VIF for each feature in the logistic regression analysis. Table S12 Baseline information and variants data for patients in the TCGA WES/WGS validation cohort (n = 430). Table S13 Missing data summary in the training set (n = 345). Table S14 F1-score, accuracy, precision, and recall for each model. Table S15 Generalized linear mixedeffects model of disease progression across different ctDNA detection techniques.