Comprehensive ctDNA profiling reveals molecular and prognostic landscapes in Korean pan-cancer patients

Liquid biopsy offers a non-invasive approach for capturing essential molecular profiles to make informed decisions within a clinical framework. However, large-scale, population-specific studies evaluating its clinical utility remain scarce. To address this unmet need, we established a Korean advanced pan-cancer circulating tumor DNA (ctDNA) atlas and systematically evaluated its clinical feasibility. We also assessed concordance with tissue-based sequencing, explored East Asian-specific molecular disparities, and quantified the prognostic impact of genomic alterations using a newly developed prognostic framework. We analyzed 1,243 ctDNA samples across 17 tumor types, integrating genomic alterations with long-term clinical follow-up. Somatic mutations were identified and clinically interpreted using the OncoKB database, and compared with the MSK-ACCESS cohort to identify population-level genomic differences. To quantify the degree of prognostic significance, we developed the Molecular-Prognostic Index (MPI), a unified metric capturing the magnitude and significance of survival effects for individual genomic alterations at both pan-cancer and individual tumor levels. Using this framework, we identified key alterations and pathways shaping patient survival and therapeutic sensitivities. East Asian cancer patients demonstrated distinctive mutational profiles compared with MSK-ACCESS, including a higher prevalence of MTOR , and FGFR3 mutations in NSCLC and enriched BRCA2 alteration in breast cancer. Comparison with tissue-based sequencing revealed higher TP53 mutations detected in tissue, whereas clinically actionable ESR1 E380Q and Y537S (OncoKB Level 1), were more prevalent in ctDNA. Pathway-level enrichment analysis exhibited strong concordance between tissue and ctDNA profiles in intestinal and prostate cancers, but marked discrepancies in lung and breast cancers, underscoring tumor type-specific biological and shedding differences. Across multiple tumor types, increased tumor mutation burden (>= 3) was significantly associated with poor clinical outcomes. BRAF V600E predicted an unfavorable prognosis in patients diagnosed with NSCLC and colorectal cancer. MPI analysis further identified PIK3R1 alterations as high-risk predictors in breast cancer and TP53 Y220C as a poor prognostic marker in prostate cancer, while NOTCH1 mutations emerged as a favorable biomarker in intestinal tumors. This large-scale ctDNA-based analysis in parallel with long-term clinical follow-up of East Asian solid tumors provides a comprehensive reference for understanding tumor propagation and therapeutic vulnerabilities. We demonstrated that ctDNA recapitulates key tumor characteristics and reveals prognostic insights. The newly developed MPI framework stratifies patient risk based on ctDNA features, and supports liquid biopsy for precision oncology.