Quality of clinical practice guidelines on precision oncology in biliary tract cancers: a systematic review using AGREE II and AGREE-REX

Abstract Background Precision oncology in biliary tract cancers (BTCs) requires more than identifying actionable biomarkers: clinical practice guidelines must connect who and how to test with the treatment or trial action that follows a positive result. We systematically evaluated contemporary BTC guidelines for methodological quality, recommendation quality, and completeness of this testing-to-treatment pathway. Methods Eight databases and 13 guideline repositories or society websites were searched for guidance published or updated from January 2016 to June 2026. Linked reports were grouped into guideline families. Three appraisers independently applied AGREE II and separate AGREE-REX clusters for molecular Testing and biomarker-directed Treatment. A prespecified 12-topic matrix mapped cross-guideline support; explicit or conditional recommendations counted as primary support, and clinical-trial-only recommendations were included in sensitivity analyses. Historical reliability used original pre-discussion ratings; a separate revision-stage sensitivity included de novo Italian Cluster A ratings. Results Thirty reports formed 27 guideline families. Of 1,107 appraisal targets, 54 were structurally not applicable. AGREE II medians were highest for clarity of presentation (98.1%) and scope and purpose (88.9%), while patient/public involvement, updating, monitoring/audit, and external review were weaker. Values and preferences was the lowest AGREE-REX domain in both clusters. Broad molecular profiling/next-generation sequencing Testing was supported by 18/27 families. Treatment support was highest for MSI-H/dMMR (20/27), FGFR2 (17/27), and IDH1 (16/27); HER2 was 12/26 after one unclassifiable unit was excluded. Testing and Treatment support diverged for several topics. Historical ICC(A,3) values ranged from 0.918 to 0.960; the 22-family Cluster A revision-stage sensitivity estimate was 0.925. Exploratory subgroup estimates were imprecise. Conclusions Contemporary BTC guidelines often provide clear individual recommendations but do not consistently specify complete precision-oncology pathways. Separating Testing from Treatment reveals coverage differences, which do not themselves indicate lower recommendation quality. Future guidance should link testing indication, timing, and method to post-test clinical action while strengthening patient involvement, evidence-to-decision reporting, resource-sensitive implementation, audit indicators, conflict management, and prospective updating. PROSPERO registration CRD420261438683.