Anatomic and Biologic Distinctions
Cholangiocarcinoma (CCA) is a heterogeneous group of biliary tract epithelial malignancies classified anatomically as intrahepatic cholangiocarcinoma (iCCA), arising from bile ducts within the liver parenchyma, and extrahepatic cholangiocarcinoma (eCCA), which encompasses perihilar cholangiocarcinoma (pCCA, Klatskin tumors) and distal cholangiocarcinoma (dCCA). This anatomic distinction is far more than descriptive: it reflects fundamentally divergent etiologic backgrounds, cellular origins, and molecular driver landscapes that translate directly into differential eligibility for targeted therapies and necessitate distinct molecular testing strategies.
Intrahepatic CCA is increasingly associated with metabolic liver disease, including nonalcoholic fatty liver disease (NAFLD), viral hepatitis, and cirrhosis, as well as hepatolithiasis—particularly in East Asian populations 27. Extrahepatic subtypes arise more frequently in the context of primary sclerosing cholangitis (PSC), choledochal cysts, and choledocholithiasis 2. A landmark whole-genome and epigenomic study of 489 CCA cases from 10 countries demonstrated that liver-fluke–associated tumors (predominantly eCCA) exhibit enrichment in ERBB2 amplifications and TP53 mutations, whereas fluke-negative tumors (predominantly iCCA) demonstrate epigenetic mutations (IDH1/2, BAP1) and FGFR-related rearrangements 27. These etiologically driven molecular divergences underscore why a "one-size-fits-all" testing algorithm is suboptimal for CCA and why anatomic subtype should anchor molecular profiling decisions.
Frequency Disparities for Key Actionable Alterations
FGFR2 Fusions and Rearrangements
Fusions and rearrangements of fibroblast growth factor receptor 2 (FGFR2) represent among the most clinically significant iCCA-enriched alterations. Across large genomic profiling cohorts—including the MSK-IMPACT series, Foundation Medicine–based studies, and the large ANITA Italian multicenter cohort—FGFR2 fusions occur in approximately 10–20% of iCCA, with particular enrichment in the small-duct histological subtype 3724. In the MSK-IMPACT cohort of 195 CCA patients, FGFR2 fusions were detected in 14% of intrahepatic cases 24. The Tempus clinicogenomic database analysis of 454 BTC samples reported FGFR2 fusions in 8.7% overall, with markedly higher enrichment in iCCA compared with eCCA and gallbladder cancer 2. In contrast, FGFR2 fusions are exceedingly rare to absent in eCCA, a finding corroborated across multiple profiling platforms 726. This dramatic anatomic disparity reflects the distinct molecular pathogenesis of iCCA and makes FGFR2 testing essentially an iCCA-specific endeavor of minimal clinical yield in eCCA.
More than 150 distinct FGFR2 fusion partners have been described, including BICC1, PPHLN1, AHCYL1, and TACC3, with partner diversity creating a significant assay-design challenge 318.
IDH1 Mutations
Mutations in isocitrate dehydrogenase 1 (IDH1)—predominantly at the R132 hotspot—are similarly restricted to iCCA. In the MSK-IMPACT cohort, IDH1 mutations were identified in 30% of iCCA 24; other large cohorts, including the Tempus series and ctDNA-based profiling studies, report frequencies of 13–20% 217. The Chinese population study reported markedly lower frequencies (7.14% IDH1/2 combined), highlighting that geographic and etiologic variation substantially influences observed mutation rates 4. IDH1 mutations are mutually exclusive with FGFR2 fusions and are rare in eCCA, aligning with the distinct molecular pathogenesis of hepatic-origin tumors 24. The clinical relevance of IDH1 mutation detection at diagnosis is underscored by the phase III ClarIDHy trial demonstrating that ivosidenib significantly prolonged progression-free survival (PFS) versus placebo (median 2.7 vs. 1.4 months; HR 0.37) in IDH1-mutant iCCA 10.
HER2/ERBB2 Amplification and Overexpression
In striking contrast to FGFR2 and IDH1, HER2/ERBB2 amplification and overexpression exhibit an eCCA-enriched distribution. Among large European surgical cohorts applying strict IHC and dual-color chromogenic in situ hybridization (dc-CISH) criteria, HER2 amplification was identified in only 0.6% of iCCA, 1.3% of pCCA, and 2.4% of dCCA, with a combined rate of 1.4% overall 21. A Japanese-American collaborative cohort using tissue and plasma next-generation sequencing (NGS) reported higher rates of HER2 amplification (3% in iCCA, 5% in eCCA, and 27% in gallbladder cancer), reflecting differences in assay sensitivity and tumor stage 22. Comprehensive molecular profiling of 1,502 biliary tract cancers confirmed that eCCA demonstrates higher rates of ERBB2 amplification than iCCA, while gallbladder carcinoma carries the highest HER2 burden 26. HER2-directed therapies are therefore of greatest relevance in eCCA and gallbladder carcinoma. The bispecific HER2-directed antibody zanidatamab received FDA accelerated approval in November 2024 for previously treated, unresectable or metastatic HER2-positive (IHC 3+) biliary tract cancer, based on the HERIZON-BTC-01 study reporting a confirmed objective response rate (ORR) of 41% (approximately 52% in IHC 3+ tumors), median duration of response of 14.9 months, and median overall survival (OS) of approximately 15.5 months 10.
BRAF V600E Mutations
BRAF V600E mutations are present in approximately 3–6% of CCA overall, without the strong anatomic enrichment characteristic of FGFR2 or IDH1 alterations, though the Genomic Landscape study identified BRAF among iCCA-enriched significantly mutated genes 23. The MSK-IMPACT cohort and large profiling studies report similar modest rates across subtypes 242. BRAF V600E is reliably detected by standard DNA-based NGS hotspot assays, and dual BRAF/MEK inhibition with dabrafenib plus trametinib has demonstrated clinically meaningful activity in BRAF V600E–mutant cholangiocarcinoma and other solid tumors, with regulatory approvals in several jurisdictions supporting tumor-agnostic use for eligible BRAF V600E–mutant cancers 10. Testing is warranted in both iCCA and eCCA.
Assay-Platform Implications
FGFR2 Fusion Detection
The extraordinary diversity of FGFR2 fusion partners (>150 described) fundamentally constrains FISH-based detection: break-apart FISH identifies rearrangements but cannot distinguish among fusion partners, misses intragenic deletions, and cannot detect 3' UTR deletions as a novel upregulation mechanism 27. Single-assay sensitivity for FGFR2-BICC1 by certain platforms has been reported as low as 2% versus 58% by more sensitive approaches, a dramatic disparity with direct consequences for patient treatment access 1. RNA-based fusion testing—using hybrid-capture or amplicon-based chemistry with RNA sequencing—provides partner-agnostic detection, superior sensitivity, and direct evidence of expressed fusion transcripts, and is the recommended first-line approach for FGFR2 in iCCA 101. Combined DNA-RNA panels that simultaneously detect point mutations, copy-number alterations, and gene fusions represent the most comprehensive solution for iCCA. Break-apart FISH remains a salvage option for tissue-limited cases but should not substitute for RNA-based or combined profiling.
HER2 Testing
HER2 assessment requires a multimodal sequential strategy. IHC provides initial semi-quantitative scoring (0/1+/2+/3+); IHC 3+ defines HER2 overexpression actionable with zanidatamab. IHC 2+ cases require confirmatory FISH or CISH for ERBB2 amplification; IHC 2+ without amplification is not a therapeutic threshold in BTC. An important caution from the Chinese iCCA dataset is that FGFR2 translocation and NTRK1 amplification did not reliably predict protein overexpression, whereas HER2, MDM2, and MET gene amplification did predict high protein expression 4. DNA-NGS can detect ERBB2 copy-number gains and rare mutations but does not replace IHC/FISH for therapeutic selection under current companion diagnostic requirements 10. The FDA simultaneously approved VENTANA PATHWAY anti-HER-2/neu (4B5) as a companion diagnostic for zanidatamab.
Pre-Analytic and Analytic Challenges
Tissue scarcity is the dominant pre-analytic challenge in CCA: biliary sampling frequently yields low tumor cellularity material, hampering NGS analysis and increasing failure rates 1. Tumor purity, fixation artifacts, decalcification (for bone biopsies), and intratumoral heterogeneity further confound molecular profiling. Biopsy site selection matters: the ANITA study and others have demonstrated that molecularly matched targeted therapies provide markedly improved OS (HR 0.34–0.49), making pre-treatment molecular profiling a clinical imperative 15. Comparative genomic analysis of 1,632 iCCA patients demonstrated that potentially actionable alterations were found in 52% of primary biopsies, 34% of metastatic biopsies, and 35% of liquid biopsies 8, highlighting that primary tumor tissue is the preferred source when available.
Circulating tumor DNA (ctDNA) testing offers minimally invasive profiling in tissue-inadequate cases and has been used to detect FGFR2 fusions, IDH1 mutations, and ERBB2 amplifications. In a prospective ctDNA/tissue-DNA concordance study, overall concordance between the two modalities was 68% for TP53, 80% for KRAS, and 90% for PIK3CA, with ctDNA numerically more concordant with metastatic than primary tumor DNA 16. However, ctDNA abundance is low particularly in early-stage disease, methodological standardization is lacking, and existing studies are limited by small cohorts and heterogeneous collection protocols 17. Current evidence supports ctDNA as a complementary tool—especially in the second-line or tissue-exhausted setting—rather than a routine upfront replacement for tissue-based profiling.
Practical Recommendations for Molecular Testing by Anatomic Subtype
For unresectable or metastatic iCCA, comprehensive molecular profiling should be initiated at diagnosis using DNA-RNA hybrid-capture NGS as the preferred platform, enabling simultaneous detection of FGFR2 fusions, IDH1 mutations, BRAF V600E, HER2 amplification, MSI-high/mismatch repair deficiency (MMRd), NTRK fusions, and KRAS/PIK3CA alterations 71. RNA-based fusion profiling is mandatory or strongly preferred for fusion detection, given the diversity of FGFR2 partners and the risk of false negatives with DNA-only panels. FISH serves as a fallback when RNA quality is insufficient. For unresectable or metastatic eCCA, the testing algorithm should prioritize HER2/ERBB2 assessment (IHC with reflex FISH/CISH and NGS confirmation), KRAS mutation testing, BRAF V600E, and MSI-high/MMRd, while FGFR2 and IDH1 testing are lower yield but still informative in the context of comprehensive profiling. HER2 IHC 3+ in eCCA/gallbladder cancer is now directly actionable with zanidatamab per FDA-approved companion diagnostic criteria.
All expert groups—including Italian consensus guidelines and ESMO-aligned recommendations—advocate early, comprehensive profiling during first-line therapy, given that many patients do not survive to access second-line molecular profiling 7115. Molecular tumor boards should integrate anatomic subtype, etiologic background, assay platform sensitivity, tissue quality, and regional prevalence data when interpreting results. Geographic variation in alteration frequencies is substantial—Chinese iCCA cohorts show lower FGFR2 translocation and IDH1 rates but higher NTRK1 amplification compared to Western cohorts 4—and counseling should account for local epidemiology.
Summary Table: Actionable Molecular Alterations in Cholangiocarcinoma by Anatomic Subtype
| Alteration | iCCA Frequency | eCCA Frequency (pCCA/dCCA) | Enrichment | Preferred Assay(s) | Key Testing Pitfalls | Approved/Actionable Therapy |
|---|---|---|---|---|---|---|
| FGFR2 fusions/rearrangements | ~10–20% (small-duct enriched; 14% MSK-IMPACT; 8.7% Tempus overall) | Rare/absent | Strongly iCCA | RNA-based fusion NGS or DNA-RNA hybrid-capture; FISH as fallback only | >150 fusion partners; DNA-only panels miss fusions; FISH cannot identify partners; FFPE RNA degradation; platform sensitivity ranges 2–58% 13 | Pemigatinib (FIGHT-202; ORR 37%, OS 21.1 mo); Futibatinib (FOENIX-CCA2; ORR 42%, PFS 9.0 mo, OS 21.7 mo); iCCA-specific FDA approvals 10 |
| IDH1 mutations | ~13–30% (30% MSK-IMPACT; 13–20% ctDNA/other cohorts; 7.14% Chinese cohort) | Rare/absent | Strongly iCCA | DNA-based NGS (hotspot; R132 variants); no RNA assay required | Geographic/etiologic variation; benefit is disease stabilization, not radiographic response 424 | Ivosidenib (ClarIDHy phase III; PFS HR 0.37; OS benefit after crossover adjustment) 10 |
| HER2/ERBB2 amplification/overexpression | ~0.6–5% (0.6% European surgical cohort; ~3% Japan-US NGS cohort) | ~2.4–27% (dCCA ~2.4–5%; gallbladder ~27%; pCCA ~1.3–5%) | eCCA/GBC-enriched | IHC (score 0–3+); FISH/dc-CISH if IHC 2+; DNA-NGS for amplification support | IHC 2+ without amplification not actionable; strict scoring criteria required to avoid overestimation; amplification more reliably predicts protein expression than FGFR2/NTRK 21224 | Zanidatamab (FDA Nov 2024; ORR 41%, DOR 14.9 mo; HER2 IHC 3+ eCCA/GBC); T-DXd, tucatinib+trastuzumab 1022 |
| BRAF V600E | ~3–6% (iCCA-enriched significantly mutated gene per genomic landscape study) | ~3–6% (no strong anatomic enrichment) | No strong subtype enrichment; modestly iCCA-enriched per some cohorts | DNA-based NGS (hotspot) | Low incidence overall; avoid misclassifying non-V600E BRAF alterations 2310 | Dabrafenib + trametinib (supported by ROAR and other basket-study data; regulatory status varies by jurisdiction, and BRAF V600E tumor-agnostic approvals may apply where available) 10 |
| MSI-high/dMMR | ~1–2% | ~1–2% | Rare across all subtypes | IHC (MLH1, MSH2, MSH6, PMS2) or PCR/NGS-based MSI | Very low prevalence; IHC requires expert interpretation; confirm discordances | Pembrolizumab (tumor-agnostic; KEYNOTE-158) 10 |
| NTRK fusions | <1% | <1% | Exceedingly rare | RNA-based hybrid-capture NGS; confirmatory NGS if pan-TRK IHC used | Pan-TRK IHC limited specificity; FISH amplification not equivalent to fusion (NTRK1 FISH signal in Chinese cohort did not confirm fusion by NGS) 4 | Larotrectinib, entrectinib (tumor-agnostic) 10 |
Frequency ranges reflect cohort composition (surgical vs. advanced disease), assay platform, tumor cellularity, geographic/etiologic variation, and classification of pCCA/dCCA. Cross-cohort comparisons should be made with caution. GBC: gallbladder carcinoma; DOR: duration of response; T-DXd: trastuzumab deruxtecan.
Conclusion
Intrahepatic and extrahepatic cholangiocarcinomas represent molecularly distinct entities whose divergent landscapes demand anatomically informed testing strategies. FGFR2 fusions and IDH1 mutations are strongly enriched in iCCA and are now linked to FDA-approved therapies; HER2/ERBB2 amplification is most clinically relevant in eCCA and gallbladder carcinoma following zanidatamab approval; BRAF V600E warrants testing in both subtypes. Comprehensive molecular profiling using combined DNA-RNA hybrid-capture NGS, with IHC and FISH applied as complementary or confirmatory tools, is the current standard for all patients with unresectable or metastatic CCA. Pre-analytic challenges—especially tissue scarcity and low tumor purity—demand proactive multidisciplinary biopsy planning, reflex ctDNA testing when tissue is exhausted, and close attention to assay-platform sensitivity, which varies dramatically for fusion detection 1157. Timely molecular profiling at diagnosis is essential, as the ANITA cohort demonstrates that matched targeted therapy provides substantially improved overall survival (HR 0.34–0.49) compared with unmatched treatment or no profiling 15. Integration of molecular tumor board review, anatomic subtype awareness, and regional epidemiology will be critical to realizing the full precision oncology potential of this rapidly evolving therapeutic landscape.