Biliary Tract · BTC

ACTICCA-1 is the ongoing landmark adjuvant trial testing GemCis doublet (the same backbone that proved superior in advanced BTC via ABC-02) in the resectable setting. The control arm was pragmatically amended from observation to capecitabine after BILCAP's per-protocol positive result, making it the first adjuvant trial to directly test whether GemCis outperforms capecitabine mono. Results are awaited with high interest; if GemCis fails to beat capecitabine, it would suggest the advanced-disease doublet advantage does not translate to the adjuvant setting. Primary publication pending.

Population: Resected biliary tract cancer (ICC, ECC, GBC) and muscle-invasive gallbladder carcinoma, R0 or R1 resection, European multicenter (Germany-led); target accrual 781 patients in original protocol; subsequently amended to ~783 with control arm changed to capecitabine after BILCAP

SWOG S0809 demonstrated that sequential chemo-CRT is feasible and active after resection of ECC and GBC, with 2-year OS exceeding the pre-specified 60% threshold. The trial is single-arm phase 2 and cannot establish superiority over chemotherapy alone, but remains the largest prospective dataset for adjuvant CRT in extrahepatic BTC. R1 resection patients showed numerically comparable outcomes to R0, suggesting CRT may compensate for microscopic residual disease. Toxicity was manageable, with grade 3-4 events in ~52% but limited grade 4.

Population: Resected extrahepatic cholangiocarcinoma (ECC) and gallbladder carcinoma (GBC), stages T2-T4 or node-positive; US multicenter (SWOG intergroup); ECC ~54%, GBC ~46%

BCAT unambiguously failed: gemcitabine monotherapy provided zero OS benefit over observation in resected ECC in a well-conducted Japanese phase 3 trial. Notably, BCAT enrolled only extrahepatic bile duct cancers, making it the most subtype-specific adjuvant RCT in BTC. The complete absence of any trend toward benefit (HR 1.01) definitively excludes gemcitabine mono from adjuvant recommendations for this subtype. This null result, combined with PRODIGE-12, led to the hypothesis that more potent regimens (doublet platinum-based, or the JCOG1202 oral fluoropyrimidine approach) might be needed.

Population: Resected extrahepatic bile duct cancer (ECC only — excludes GBC and ICC), R0 resection, Japan multicenter; all patients are Japanese; median age ~68 years

BILCAP narrowly missed its ITT primary OS endpoint (p=0.097), but the pre-specified per-protocol analysis showed a significant 25% reduction in death risk with capecitabine. The ITT/per-protocol discordance reflects protocol deviations (dose reductions, early stops) in a real-world chemotherapy-tolerability context. Most major guidelines (ESMO, NCCN, CSCO) adopted capecitabine adjuvant therapy as standard of care based on the per-protocol result and absence of better data, making BILCAP the most practice-changing adjuvant BTC trial until ASCOT. The trial enrolled all BTC subtypes, precluding subtype-specific conclusions.

Population: Resected biliary tract cancer (intrahepatic and extrahepatic cholangiocarcinoma + gallbladder cancer), R0 or R1 resection, mixed subtype but predominantly intrahepatic/extrahepatic cholangiocarcinoma (~65%) and GBC (~30%), UK multicenter

PRODIGE-12 failed to demonstrate a significant improvement in RFS or OS with adjuvant GEMOX versus observation. This was likely underpowered (target 190 events, actual 170) and suffered from slower accrual. The trial is important as a counterpoint to BILCAP: it shows that not all adjuvant doublets are effective, and that GEMOX (the regimen used in many advanced BTC patients at the time) does not translate to an adjuvant benefit. The individual patient data meta-analysis combining BCAT and PRODIGE-12 (PMID 35182925) similarly showed no RFS/OS benefit for gemcitabine-based adjuvant therapy across both trials.

Population: Resected biliary tract cancer (intrahepatic + extrahepatic cholangiocarcinoma + GBC), R0 or R1 resection, mixed BTC subtypes, France multicenter (UNICANCER/FFCD)

ASCOT is the cleanest positive adjuvant BTC phase 3 trial to date, demonstrating a statistically significant 31% reduction in death risk with oral S-1 versus observation. Unlike BILCAP, ASCOT hit its ITT primary endpoint without controversy. The absolute benefit translates to approximately 11 months median OS improvement. Subgroup analyses suggested benefit across ICC, ECC, and GBC, though GBC numbers were smaller. The trial is limited to Japanese patients and S-1 PK/tolerability differs by ethnicity (East Asian populations tolerate S-1 better due to lower DPD activity). Whether results extend to capecitabine (a related fluoropyrimidine) in Western populations is inferred but not proven.

Population: Resected biliary tract cancer (ICC ~30%, ECC ~49%, GBC ~22%), R0 resection, Japan multicenter (JCOG); all Japanese patients; Stage I-III disease

This is the first randomized trial to show a significant DFS benefit with adjuvant GemCis followed by CRT specifically in high-risk GBC — a subtype with historically dismal outcomes after resection. The 14-month DFS improvement is clinically meaningful. The trial is unique in its sequential chemo-then-CRT design and its GBC-only population, which is highly relevant given India's high GBC incidence. Limitations include single-center design and an Indian-specific patient population; generalizability to other regions is uncertain. OS data at time of primary publication showed a trend but were immature.

Population: Resected gallbladder carcinoma (GBC only) with high-risk features (T3-T4 or node-positive or R1), India single-institution (Tata Memorial Hospital, Mumbai); all Indian patients; GBC is endemic in Northern/Eastern India

ABC-02 is the foundational 1L advanced BTC trial that established GemCis as the standard of care for over a decade. The 3.6-month median OS improvement and HR of 0.64 were highly significant and reproducible (BT22 Japanese confirmatory trial also positive). Tolerability was broadly acceptable. The trial set the backbone that TOPAZ-1, KEYNOTE-966, and SWOG-1815 would all build upon. One key caveat: the survival was still only 11.7 months median, illustrating the unmet need that drove the subsequent IO combination era.

Population: Advanced (unresectable or metastatic) biliary tract cancer — intrahepatic cholangiocarcinoma ~39%, extrahepatic cholangiocarcinoma ~20%, gallbladder cancer ~38%, ampullary carcinoma ~3%; UK multicenter; treatment-naive

AMEBICA/PRODIGE-38 showed that mFOLFIRINOX is not superior to GemCis as 1L therapy for advanced BTC. Neither PFS nor OS favored the triplet, and toxicity was numerically higher with mFOLFIRINOX (Grade 3-4 adverse events ~72% vs ~60%). This is a phase 2 RCT designed as a selection/screening trial rather than a definitive superiority trial. The negative result is consistent with SWOG-1815's finding that adding cytotoxic complexity to the GemCis backbone is not beneficial in unselected BTC. However, mFOLFIRINOX remains of interest for specific high-volume disease contexts and as a salvage backbone.

Population: Advanced BTC (unresectable/metastatic), 1L treatment-naive; ICC ~55%, ECC ~20%, GBC ~22%, ampullary ~3%; France multicenter (PRODIGE/FFCD); predominantly Caucasian

TOPAZ-1 was the first phase 3 trial to show a statistically significant OS benefit from adding a PD-L1 inhibitor (durvalumab) to GemCis in advanced BTC. The primary OS HR of 0.80 is statistically significant but clinically modest — median OS improvement is only 1.3 months. However, the 24-month OS landmark (25% vs 10%) demonstrates a meaningful tail enrichment, suggesting durable benefit in a subset. The benefit was consistent regardless of PD-L1 expression status, arguing against biomarker stratification for treatment selection. TOPAZ-1 led to durvalumab + GemCis approval by regulatory agencies in multiple countries.

Population: Advanced BTC (unresectable or metastatic), 1L treatment-naive; intrahepatic cholangiocarcinoma ~56%, extrahepatic cholangiocarcinoma ~20%, gallbladder cancer ~19%, ampullary carcinoma ~5%; global (Asia ~54%, Europe ~30%, Americas ~16%); all PD-L1 unselected

KEYNOTE-966 independently confirmed the class effect of IO + GemCis in advanced BTC, with a HR of 0.83 (p=0.0034). As with TOPAZ-1, the median OS gain is modest (1.8 months) but highly statistically significant in the largest BTC phase 3 trial to date (N=1069). The higher GBC enrollment (~30%) is notable; prior studies were ICC-enriched. A post-hoc analysis suggested GBC subtype may derive less benefit, though this was underpowered. PD-L1 CPS ≥10 did not predict enhanced benefit, consistent with TOPAZ-1 findings. KEYNOTE-966 independently drove pembrolizumab + GemCis approval.

Population: Advanced BTC (unresectable/metastatic), 1L treatment-naive; ICC ~43%, ECC ~17%, GBC ~30%, ampullary ~10%; global enrollment including Asia, Europe, Americas; PD-L1 unselected; largest of the IO+GemCis phase 3 trials by sample size

KHBO1401-MITSUBA tested adding S-1 to GemCis as a Japan-specific triplet (GCS) in advanced BTC. Like BILCAP in the adjuvant setting, it narrowly missed its ITT primary endpoint but showed a borderline significant result in per-protocol analysis. The 0.9-month median OS advantage is small but the per-protocol HR of 0.73 suggests potential benefit in patients tolerating full-dose therapy. Japan-specific tolerability of S-1 (East Asian DPD pharmacogenomics) may limit extrapolation. Note: candidate list referenced 'KHBO1202' but that trial registration name corresponds to a different protocol; the principal published phase 3 GCS vs GemCis result is KHBO1401-MITSUBA (Ioka T, J Hepatobiliary Pancreat Sci 2023).

Population: Advanced BTC (unresectable/metastatic), 1L treatment-naive; ICC ~35%, ECC ~35%, GBC ~30%; Japan multicenter (Kansai Hepatobiliary Oncology Group, KHBO); all Japanese patients

SWOG S1815 definitively failed to show OS benefit from adding nab-paclitaxel to GemCis in advanced BTC. The HR of 0.93 crossed the null with a p of 0.52. Subgroup analyses showed no enriched population benefiting from the triplet. The trial is an important negative result because nab-paclitaxel + GemCis had shown activity in phase 2 (Shroff 2019) and enthusiasm was high. The failure highlights the difficulty of improving on GemCis with cytotoxic agents alone in an unselected BTC population, and reinforces the field's shift toward IO-based combinations. SWOG-1815 was originally designed pre-IO era; by the time it completed, IO+GemCis had already become standard.

Population: Advanced BTC (unresectable/metastatic), 1L treatment-naive; ICC ~55%, ECC ~15%, GBC ~28%, ampullary ~2%; US multicenter (SWOG intergroup); mixed GBC/CCA enrollment reflecting North American epidemiology

Regorafenib showed modest single-agent activity in refractory BTC with a 56% DCR but only 11% ORR and a short median PFS of 2 months. The 7.4-month median OS is consistent with 2L BTC historical controls. Grade 3-4 adverse events occurred in ~60% (hand-foot syndrome, fatigue, hypertension — typical regorafenib toxicity profile). The results are exploratory and the trial was not powered for formal efficacy testing. Biomarker correlates (RAS/RAF mutations, VEGFR expression) were explored but no predictive biomarker emerged. This trial did not lead to regulatory approval for regorafenib in BTC.

Population: Advanced BTC (unresectable/metastatic) refractory to ≥1 prior gemcitabine-based chemotherapy; ICC ~53%, ECC ~23%, GBC ~24%; US multicenter (University of Pittsburgh and collaborators); ECOG PS 0-1; all prior GemCis or GemOx

Kim 2020 demonstrated that nivolumab monotherapy has meaningful single-agent activity in refractory BTC with a 22% ORR — noteworthy given historical single-agent ORRs of 5-10% in chemotherapy-refractory BTC. The 14.2-month median OS in a refractory population is surprisingly favorable, likely reflecting selection bias toward good-PS patients. Notably, responses appeared enriched in MMR-deficient/MSI-H tumors (2 patients) and those with high TMB. This study informed the design of IO combination trials and highlighted that a subset of BTC patients (MSI-H, high TMB) may be IO-sensitive, fueling the biomarker-stratified precision approach (now chapter 4 scope).

Population: Advanced BTC (unresectable/metastatic) refractory to ≥1 prior therapy (predominantly GemCis-refractory); ICC ~61%, ECC ~19%, GBC ~20%; US multicenter; all patients received prior GemCis or GemCis-based therapy; PD-L1 unselected

ABC-06 established modified FOLFOX as the first evidence-based 2L standard for advanced BTC after GemCis failure, showing a statistically significant but modest OS benefit (0.9 months median, HR 0.69). The 12-month OS landmark difference (26% vs 11%) is more clinically impactful than the median. The trial is notable for its comparator (ASC alone), which is ethically appropriate given the absence of proven 2L therapy at trial initiation. Toxicity was manageable; FOLFOX is well-tolerated relative to triplet regimens. ABC-06 is now the reference comparator for all subsequent 2L BTC trials.

Population: Advanced BTC (unresectable/metastatic) previously treated with GemCis (progressed after 1L); ICC ~57%, ECC ~17%, GBC ~20%, ampullary ~6%; UK multicenter; ECOG PS 0-1

NIFTY showed a highly significant PFS benefit (HR 0.56, p=0.002) for nal-IRI+5FU/LV vs 5FU/LV alone in previously treated BTC. The median PFS advantage of 5.7 months is the largest 2L PFS delta in any BTC RCT to date. However, OS was not the primary endpoint and the OS HR was 0.82 (95% CI 0.57-1.17, p=0.27) — not significant. The trial uses 5FU/LV alone as comparator, which is weaker than FOLFOX (ABC-06 standard), limiting direct comparison to current standard of care. Important distinction from NALIRICC: NIFTY is POSITIVE for PFS, NALIRICC is NEGATIVE for OS.

Population: Metastatic BTC (unresectable/metastatic) previously treated with GemCis ± targeted therapy; ICC ~62%, ECC ~15%, GBC ~20%, ampullary ~3%; Korea multicenter; all Korean patients; ECOG PS 0-1

NALIRICC failed to meet its primary OS endpoint (HR 0.68, p=0.074 — borderline but below significance threshold). This is a phase 2 trial with only 100 patients and limited statistical power; it is arguably underpowered to detect a modest OS difference reliably. The HR of 0.68 is numerically similar to ABC-06 FOLFOX effect, but given the 5FU/LV-only comparator (not FOLFOX), and the small sample size, the trial is inconclusive rather than definitively negative. The divergent results between NIFTY (positive PFS) and NALIRICC (negative OS) are best explained by: (1) different primary endpoints, (2) different populations (Korean vs German), and (3) statistical power differences (N=174 vs N=100).

Population: Advanced BTC (cholangiocarcinoma and gallbladder carcinoma) previously treated with gemcitabine-based therapy; ICC ~55%, ECC ~20%, GBC ~25%; Germany multicenter (AIO cooperative group); predominantly Caucasian patients

Larotrectinib received the first-ever tumor-agnostic FDA approval (November 2018) based on NTRK fusion biomarker regardless of tumor histology — a landmark regulatory precedent. The 75% overall ORR is exceptional, driven by pediatric tumors with high NTRK fusion rates; BTC patients respond when NTRK fusions are present, though NTRK fusion prevalence in BTC is <1%, making BTC-specific outcome data sparse in published cohorts. The histology-independent approval was enabled by the mechanistic (driver fusion) argument and the consistency of response across tumor types.

Population: TRK (NTRK1/2/3) fusion-positive advanced solid tumors; pan-cancer pooled analysis of 3 trials. BTC/CCA cohort: small (n=2-3 of 55 in primary publication); subsequent expanded analyses report BTC patients. NTRK fusions occur in <1% of all BTC. TRK fusions detected by RNA-based NGS (preferred) or DNA-based fusion panels. Drilon A, NEJM 2018 is the pivotal FDA-registration paper.

First positive phase 3 RCT for any targeted therapy in CCA, and the only phase 3 biomarker-selected CCA trial with a placebo comparator. Ivosidenib received FDA approval in August 2021 for previously treated IDH1-mutant CCA. PFS benefit was clear (HR 0.37), but the OS analysis was confounded by ~70% crossover from placebo arm; after adjusting for crossover, OS HR favored ivosidenib. The modest ORR (2.4%) highlights that IDH1 inhibition primarily stabilizes disease rather than inducing tumor shrinkage, consistent with mutant IDH1's role in epigenetic reprogramming rather than direct oncogenic signaling.

Population: Previously treated (≥1 prior therapy) unresectable or metastatic CCA with IDH1 mutation (R132 hotspot) by central testing; ICC-enriched (~90% ICC, ~10% ECC/GBC) given IDH1 mutation biology. Randomized 2:1 ivosidenib:placebo. Primary pub: Abou-Alfa GK, Lancet Oncol 2020 (PMID 32416072). OS update: Zhu AX, JAMA Oncol 2021 (PMID 34554208).

First FDA-approved targeted therapy for any CCA molecular subset (approved April 2020, accelerated approval). Established FGFR2 fusions as a therapeutically actionable biomarker in ICC. Cohort A ORR 35.5% against a single-digit historical chemo benchmark validated the molecular-matching principle. Cohort B (non-fusion FGFR alterations, n=20) showed 0% ORR, confirming that fusion/rearrangement specificity is required. Major toxicities: hyperphosphatemia (94%), alopecia, dysgeusia — all on-target FGFR inhibition effects.

Population: Previously treated unresectable or metastatic CCA; Cohort A (n=107) FGFR2 fusion or rearrangement; Cohort B (n=20) other FGFR alterations; Cohort C (n=18) no FGFR alterations. ICC-dominant (>95% of enrolled patients, consistent with FGFR2 fusion biology exclusive to ICC). Prior chemo median 1 line (range 1-6). FGFR2 status by local next-generation sequencing or FISH.

FIGHT-302 is the phase 3 confirmatory trial for first-line pemigatinib in FGFR2-rearranged CCA. ESMO 2024 data presented by Bekaii-Saab showed significant improvement in PFS vs GemCis; full PFS and OS data with confidence intervals are not yet in a peer-reviewed publication as of April 2026. If positive, this will establish pemigatinib as a biomarker-selected 1L standard and potentially replace GemCis for FGFR2+ ICC. The pmid cited here is the protocol publication (Future Oncol 2020); the primary results PMID should be updated once published.

Population: First-line unresectable or metastatic CCA with FGFR2 fusion or rearrangement by central testing; ICC-dominant by definition given FGFR2 biology. PMID 32677452 is the protocol/design paper (Future Oncol 2020); primary efficacy results had not been published in a peer-reviewed journal as of April 2026 (ESMO 2024 oral presentation showed positive PFS; full JCO/LO publication pending).

KEYNOTE-158 Cohort K provided the BTC-specific MSI-H efficacy data that contributed to the pan-tumor FDA accelerated approval of pembrolizumab for MSI-H/dMMR solid tumors (June 2020). The BTC cohort ORR of 40.9% (n=22) is limited by small sample size but directionally consistent with pembrolizumab activity across MSI-H tumor types. The Maio 2022 Ann Oncol update (PMID 35680043) extended follow-up and confirmed durable responses in MSI-H responders regardless of histology. MSI-H prevalence in BTC is ~2-3%, making universal MSI testing in BTC critical for a small but potentially highly responsive subgroup.

Population: MSI-H or dMMR advanced non-colorectal solid tumors after ≥1 prior therapy across 10 tumor cohorts; Cohort K = biliary tract cancer (n=22 of 233 total BTC in MSI-H analysis; MSI-H prevalence in BTC ~2-3%). Primary pub: Marabelle A, JCO 2020 (PMID 31682550) — reports overall non-CRC MSI-H ORR and tumor-cohort breakdown. Long-term update: Maio M, Ann Oncol 2022 (PMID 35680043) — extended follow-up. BTC cohort: predominantly ICC; MSI-H detected by PCR or NGS.

NCI-MATCH subprotocol H independently confirmed tumor-agnostic BRAF V600E targeting with dab+tram in non-melanoma, non-NSCLC solid tumors. The BTC/CCA patients in this trial showed responses consistent with ROAR BTC data, providing cross-study validation. The small BTC subgroup (n=7-8) in NCI-MATCH H limits definitive conclusions for BTC alone, but collectively with ROAR (n=43), the evidence base for dab+tram in BRAF V600E CCA is robust.

Population: BRAF V600E-mutated advanced solid tumors excluding melanoma and NSCLC (those were covered by dedicated arms); pan-cancer basket including CCA/BTC among several tumor types. BTC/CCA subgroup n=7-8 of 35 total; exact BTC N reported in primary publication. All enrolled after ≥1 prior therapy. BRAF V600E detected by central CLIA-certified molecular testing (NCI-MATCH trial molecular screening).

Dabrafenib + trametinib achieved ORR 51% in BRAF V600E-mutated BTC — among the highest response rates observed in any BTC biomarker cohort. FDA granted accelerated approval for dab+tram in BRAF V600E-mutated solid tumors (tumor-agnostic, 2022) covering BTC. BRAF V600E mutation is rare (~3-5% ICC) but the high ORR means this signal strongly justifies testing. The combination is well-established from melanoma and NSCLC, facilitating rapid adoption once the mutation is identified. Pyrexia and rash are class-effect toxicities.

Population: BRAF V600E-mutated unresectable or metastatic BTC; BTC cohort n=43 (ICC predominantly ~85%, with ECC ~10%, GBC ~5%). BRAF V600E occurs in ~3-5% ICC, rare in other BTC subtypes. All patients had ≥1 prior therapy. BRAF V600E detected by local molecular testing. ROAR is a multi-cohort basket trial; the BTC cohort was published as part of the primary BTC analysis (Subbiah V, Lancet Oncol 2020).

Entrectinib received FDA tumor-agnostic approval for NTRK fusion-positive solid tumors in August 2019, with CNS activity as a distinguishing feature vs larotrectinib (given entrectinib's CNS penetrance). The overall ORR (57%) is lower than larotrectinib (75%), likely reflecting less-selective kinase binding and broader patient populations with lower fusion-driver dependency. For BTC patients with NTRK fusions (rare, <1%), both larotrectinib and entrectinib are options; CNS-active entrectinib may be preferred in cases with brain metastases.

Population: NTRK fusion-positive advanced solid tumors; integrated analysis of 3 trials (Doebele RC, Lancet Oncol 2020). BTC/CCA subgroup: not specifically reported in primary analysis due to small N (<3% of enrolled); entrectinib's tumor-agnostic registration included heterogeneous histologies. NTRK fusions detected by RNA-based NGS or DNA-based platforms.

MyPathway was the first prospective evidence for dual anti-HER2 blockade (pertuzumab + trastuzumab) in BTC, preceding later dedicated BTC trials. ORR 23.1% established proof-of-concept for HER2 targeting in BTC but was lower than subsequently reported for zanidatamab or tucatinib combinations. The IHC 2+/FISH+ inclusion (not just IHC 3+) may have diluted the signal. This basket cohort data contributed to the clinical rationale for larger dedicated BTC HER2 trials.

Population: HER2-positive (IHC 3+ or IHC 2+/FISH+) previously treated advanced BTC; BTC cohort n=39 of 251 total MyPathway patients. All BTC subtypes; GBC and ECC more prevalent than ICC given HER2 biology. Median 2 prior lines. Published as dedicated BTC-cohort analysis (Javle M, Lancet Oncol 2021).

Pralsetinib demonstrated 57% ORR in RET fusion-positive pan-cancer basket with durable responses (12-month DoR rate 81%), supporting FDA tumor-agnostic accelerated approval. Like selpercatinib (LIBRETTO-001), BTC cohort data are too sparse to report separately due to ultra-low RET fusion prevalence in BTC. Both pralsetinib and selpercatinib are clinically regarded as equivalent precision options for RET fusion-positive BTC, though no head-to-head comparison exists. Pneumonitis and hypertension are class-effect toxicities.

Population: RET fusion-positive solid tumors (non-lung, non-thyroid, or post-standard therapy) in a pan-cancer ARROW expansion cohort; n=29 (Subbiah V, Nat Med 2022). BTC/CCA: subset of the pan-cancer cohort; RET fusions in BTC <1%. RET fusions detected by central RNA-based NGS or DNA panel. Pralsetinib received FDA tumor-agnostic accelerated approval for RET fusion-positive solid tumors (August 2022).

Selpercatinib demonstrated durable responses (median DoR 24.5 mo) in RET fusion-positive non-lung/non-thyroid tumors, supporting tumor-agnostic FDA approval. The BTC subgroup (n=7) showed ORR consistent with the overall cohort, though the ultra-small N precludes definitive BTC-specific conclusions. RET fusions in BTC are very rare (<1%), detectable by comprehensive NGS panels; when present, they represent a highly actionable target given the strong efficacy signal from the broader RET fusion evidence base. CNS activity of selpercatinib is relevant if BTC patients have brain metastases.

Population: RET fusion-positive advanced solid tumors excluding lung and thyroid (those had dedicated larger LIBRETTO-001 cohorts); pan-tumor basket n=45 (Subbiah V, Lancet Oncol 2022). BTC/CCA subgroup: n=7 of 45 non-lung/non-thyroid patients; RET fusions occur in <1% BTC. RET fusion detected by RNA-based NGS or DNA NGS. Selpercatinib received FDA tumor-agnostic approval for RET fusion-positive solid tumors (September 2022).

André 2023 reports the expanded efficacy data for dostarlimab in dMMR solid tumors, providing the third distinct anti-PD-1 agent (after pembrolizumab and nivolumab) with pan-tumor dMMR approval. For BTC, dostarlimab offers a regulatory and mechanistically equivalent alternative to pembrolizumab in dMMR-positive cases, though no head-to-head comparison exists. The durable DoR (24-month DoR rate 69%) supports the clinical narrative that dMMR/MSI-H tumors have deep, lasting IO responses. BTC-specific subgroup data not reported separately due to small N.

Population: dMMR advanced solid tumors after ≥1 prior therapy (pan-tumor basket); multiple cohorts including endometrial, CRC, and non-CRC/non-endometrial cohorts. BTC/CCA subgroup: within the 'other solid tumors' dMMR cohort (n not specifically segregated for BTC in primary publication); dMMR BTC estimated <1-2% of any dMMR pan-tumor basket. André T, JAMA Netw Open 2023. Dostarlimab received FDA approval for dMMR solid tumors (August 2021) based on this and the GARNET trial.

Futibatinib received FDA accelerated approval in September 2022 for previously treated FGFR2-rearranged ICC, making it the second FGFR2-targeted therapy approved for CCA. The covalent (irreversible) binding mechanism was hypothesized to overcome certain acquired resistance mutations that arise on reversible inhibitors like pemigatinib. ORR 41.7% is numerically higher than FIGHT-202 Cohort A (35.5%), though cross-trial comparison is confounded by patient selection and testing platform differences. Both trials validated FGFR2 rearrangement (not just amplification) as the actionable molecular event.

Population: Previously treated (≥1 prior systemic therapy) unresectable or metastatic ICC with FGFR2 fusions or other rearrangements by central next-generation sequencing (FoundationOne CDx). 100% ICC by design (FGFR2 rearrangements are ICC-exclusive). 67% had ≥2 prior lines.

Zanidatamab received FDA accelerated approval in August 2024 for previously treated HER2-amplified BTC, making HER2 the third approved biomarker target in this disease (after FGFR2 and IDH1). The bispecific mechanism (simultaneous HER2 domain 2 + domain 4 engagement) induces receptor clustering, internalization, and ADCC — conceptually distinct from single-agent trastuzumab. ORR 41.3% with durable median DoR of 12.9 months is clinically meaningful in the 2L+ setting. Tolerability favorable: diarrhea (45%) was the most common treatment-related adverse event.

Population: Previously treated (≥1 prior therapy) unresectable or metastatic BTC with HER2 amplification by central IHC/ISH (IHC 3+ or IHC 2+/ISH+); all BTC subtypes enrolled: GBC (~40%), ICC (~35%), ECC (~25%). HER2 amplification highest in GBC (~20-30%) vs lower ICC/ECC rates. Median 2 prior lines.

Adagrasib demonstrated anti-tumor activity in KRAS G12C-mutated BTC (ORR 33.3%, n=12), providing proof-of-concept for KRAS G12C targeting in this disease. BTC was the first GI tumor type beyond NSCLC/CRC where KRAS G12C targeting showed activity. The small BTC N (n=12) produces wide confidence intervals but the disease control rate of 83.3% signals meaningful biological activity. Adagrasib received FDA accelerated approval for KRAS G12C-mutated solid tumors (June 2024, tumor-agnostic). Combination strategies (adagrasib + cetuximab, or with SOS1 inhibitors) are being explored to enhance efficacy in KRAS G12C tumors.

Population: KRAS G12C-mutated advanced solid tumors excluding NSCLC and CRC (those had dedicated KRYSTAL-1 cohorts); pan-tumor basket. BTC/CCA cohort: n=12 of 112 pan-tumor non-NSCLC/CRC patients (Bekaii-Saab TS, JCO 2023). KRAS G12C mutation occurs in ~1-2% BTC, predominantly ICC. KRAS G12C detected by local or central solid tumor NGS panel.

RAGNAR supported FDA's tumor-agnostic accelerated approval of erdafitinib (September 2024) for FGFR-altered solid tumors after chemotherapy. The BTC cohort (n=27) showed responses consistent with the overall basket, validating FGFR2 targeting in CCA within a broader tumor-agnostic framework. Erdafitinib differs from pemigatinib/futibatinib in its pan-FGFR activity (covers FGFR1-4 and also addresses mutations in addition to fusions). Hyperphosphatemia was the key class-effect toxicity; ocular toxicity (central serous retinopathy) required monitoring.

Population: Advanced solid tumors with FGFR1-4 alterations (fusions, mutations, or amplifications) after ≥1 prior therapy; tumor-agnostic design spanning ≥15 tumor types. BTC cohort: n=27 of 217 total (CCA/BTC subgroup); FGFR2 fusions the dominant alteration in BTC patients consistent with ICC biology. FGFR alterations detected by central or local NGS.

Tucatinib (oral HER2-selective TKI) + trastuzumab demonstrated ORR of 46.7% in HER2+ BTC, numerically the highest single-arm ORR among HER2-directed doublets in BTC at time of publication. The combination leverages complementary mechanisms (kinase inhibition + antibody-dependent cellular cytotoxicity). The small cohort (n=30) limits precision of estimates. Tucatinib's CNS penetrance — established in breast cancer (HER2CLIMB) — has not been evaluated in BTC, though BTC brain metastases are rare.

Population: Previously treated (≥1 prior line) HER2-positive (IHC 3+ or IHC 2+/FISH+) advanced BTC; BTC cohort n=30. HER2-positive BTC includes GBC-enriched population; specific ICC/ECC/GBC split not reported. Median 1 prior line. SGNTUC-019 is a basket trial (HER2+ solid tumors) with the BTC cohort published separately (Nakamura Y, JCO 2023).

CheckMate-848 demonstrated that TMB-H is an actionable biomarker for dual checkpoint blockade (nivo+ipi), with ORR 43.3% vs 26.0% for monotherapy — establishing combination IO as superior in TMB-H tumors. BTC-specific outcomes are not separately reported in the primary publication; given the small expected BTC fraction in a pan-tumor TMB-H basket, BTC-specific conclusions require caution. The trial indirectly supports routine TMB testing in BTC as part of comprehensive biomarker profiling, since TMB-H BTC patients may derive greater benefit from IO combination approaches.

Population: TMB-H (≥10 mut/Mb by FoundationOne CDx) advanced solid tumors after ≥1 prior therapy; pan-tumor randomized basket. BTC cohort within TMB-H enrollment: not specifically segregated in primary publication; BTC proportion in pan-tumor TMB-H basket estimated ~3-5% of enrolled patients. TMB-H is distinct from MSI-H (overlap ~40-50%); BTC can be TMB-H without MSI-H. Schenker M, J Immunother Cancer 2024.

T-DXd demonstrated substantial activity in HER2-overexpressing BTC, with a strong IHC 3+ signal (ORR 56.3%) contrasting with a modest IHC 2+ signal (20.0%), establishing HER2 IHC score as a response predictor. T-DXd received FDA tumor-agnostic accelerated approval in April 2024 for HER2-overexpressing (IHC 3+) solid tumors after prior therapy — covering BTC. Interstitial lung disease (ILD)/pneumonitis occurred in 5.6% of BTC cohort (grade ≥3 in 1.9%), representing the key safety concern requiring monitoring. The ADC mechanism broadens HER2 targeting to patients with HER2 overexpression (not requiring amplification).

Population: HER2-expressing (IHC 3+ or IHC 2+) advanced solid tumors across 7 tumor cohorts; BTC cohort: n=41 of 267 total. BTC cohort included ICC (~70%), GBC (~20%), ECC (~10%). HER2 overexpression (not amplification) the entry criterion; IHC 3+ and IHC 2+ subgroups analyzed separately. All patients had ≥1 prior therapy.

Zenocutuzumab (Bizengri) received FDA accelerated approval in August 2024 for NRG1 fusion-positive NSCLC and pancreatic cancer (the two largest cohorts), with the broader tumor-agnostic data supporting use in rare NRG1 fusion-positive histologies including BTC. The CCA-specific analysis (Schram 2025 MCT abstract) showed responses in the BTC/CCA subgroup consistent with overall efficacy. NRG1 fusions differ mechanistically from most other oncogenic fusions — they create a ligand (NRG1) that activates HER2-HER3 signaling rather than encoding a constitutively active fusion kinase, which is why zenocutuzumab's anti-HER3 mechanism is specifically suited (blocking neuregulin-1 from binding HER3). PMID 39908431 (NEJM 2025) is PubMed-verified.

Population: NRG1 fusion-positive advanced solid tumors after ≥1 prior therapy; pan-tumor basket spanning multiple histologies. Pancreatic cancer and lung cancer are most common NRG1 fusion-harboring types; BTC/CCA is represented with estimated n=3-5 of 64 in primary pub. NRG1 fusions occur in <1% BTC, comparable to RET and NTRK. Primary pub: Schram AM, NEJM 2025 (PMID 39908431). A CCA-specific analysis was reported as a conference abstract (Schram AM 2025, Mol Cancer Ther) — note abstract data, not peer-reviewed.

Repotrectinib (Augtyro, BMS) received FDA approval in November 2023 for NTRK fusion-positive solid tumors, including both TRK-inhibitor-naive and -pretreated patients. Its compact macrocyclic structure enables accommodation of resistance mutations that arise on first-generation TRK inhibitors (notably NTRK G595R solvent-front mutation), providing a sequential treatment option. For BTC patients progressing on larotrectinib/entrectinib with NTRK resistance mutations, repotrectinib is the designated next-line precision option. Note: PMID 41639379 (Nat Med 2026) is a later updated analysis; initial FDA registration was based on earlier data.

Population: NTRK fusion-positive advanced solid tumors in both TRK inhibitor-naive and TRK inhibitor-pretreated patients; pan-tumor design (Besse B, Nat Med 2026). BTC/CCA subgroup: small, not separately quantified in primary publication due to low NTRK prevalence in BTC (<1%). Designed to address resistance emerging after larotrectinib/entrectinib therapy.