Key studies of bile-based liquid biopsy in CCA and BTC.
| Author (year) | Design | n | Tumour type | Biomarker/Method | Key finding | Bile vs. plasma | Main limitation |
|---|---|---|---|---|---|---|---|
| Arechederra et al. (2022) [9] | Prospective | 68 | CCA, PDAC (biliary strictures) | cfDNA/NGS (Pan-Cancer panel; Bilemut assay) | Overall sensitivity 96.4%, specificity 69.2%; 100% sensitivity in initially benign/indeterminate strictures | Bile cfDNA outperformed initial pathological diagnosis; plasma not directly compared | Pilot study; mixed biliopancreatic stricture cohort; CCA not analysed as an independent validation subgroup; no external validation cohort |
| Liu et al. (2025) [11] | Prospective | 10 | Resectable CCA | cfDNA/NGS | Mutations in 70% bile cfDNA vs. 10% plasma; higher maxAF and TMB in bile | Bile showed higher mutation detection than plasma for local molecular characterisation in this cohort | Very small cohort (n = 10); single centre; limited to resectable disease |
| Gou et al. (2021) [12] | Retrospective | 42 | BTC (mixed) | cfDNA/NGS (520-gene panel) | Mutations in 71.4% bile vs. 53.6% plasma; higher maxAF and tissue concordance in bile | Bile cfDNA showed higher mutation detection and stronger tissue concordance than plasma in this cohort | Mixed BTC population; retrospective; limited matched plasma data |
| Li et al. (2022) [13] | Retrospective | 13 | BTC (mixed) | cfDNA/NGS (425 genes) | Bile cfDNA 1,918 ng/mL vs. 63.1 ng/mL plasma; more alterations per sample in bile | Bile supernatant and pellet detected additional alterations not identified in matched tissue in this cohort | Small cohort; mixed BTC; no prospective validation |
| He et al. (2023) [10] | Retrospective + validation | Not stated for CCA alone | Pancreatobiliary cancer | cfDNA mutations + methylation/BileScreen assay | High diagnostic performance; outperformed CA19-9 in specificity | Bile not directly compared to plasma in same cohort | Not CCA-specific; pancreatobiliary mixed population; methylation markers not independently validated in CCA |
| Shen et al. (2019) [33] | Retrospective | 10 | BTC (mixed) | cfDNA/targeted deep sequencing (150 genes) | SNV/Indel sensitivity 94.7%; CNV sensitivity 75%; Lower CNV sensitivity reported in AJCC stage II and distal bile duct tumours, though cohort size limits interpretation | First study to demonstrate bile cfDNA-based somatic variant detection in BTC using a 150-gene panel | Very small cohort; mixed BTC; n = 1 distal CCA |
| Vedeld et al. (2022) [19] | Retrospective | 273 | CCA, PSC-associated CCA, PSC | DNA methylation/ddPCR (CDO1, CNRIP1, SEPT9, VIM) | 100% sensitivity, 90% specificity for PSC-associated CCA when diagnosed within 12 months of bile collection | Bile not compared to plasma in same cohort | Retrospective; PSC-focused; markers not validated for sporadic CCA; no prospective surveillance pathway |
BTC: biliary tract cancer; CCA: cholangiocarcinoma; cfDNA: cell-free DNA; ctDNA: circulating tumour DNA; maxAF: maximum allele fraction; NGS: next-generation sequencing; PDAC: pancreatic ductal adenocarcinoma; PSC: primary sclerosing cholangitis; TMB: tumour mutational burden; ddPCR: droplet digital polymerase chain reaction; SNV: single nucleotide variant.