Optimizing Patient Pathways in Advanced Biliary Tract Cancers: Recent Advances and a French Perspective

Only 30–50% [25] of patients diagnosed with CCA are suitable for resection, which is the only potentially curative treatment [27]. For patients with resectable pCCA, biliary drainage may be used to flush the bile ducts in order to alleviate cholestasis, and this can be performed prior to surgical resection, but is not without risk of complications such as infection or the seeding of metastatic cells [28, 29]. Biliary drainage may also be performed in patients with unresectable CCA and is beneficial for patients who are deteriorating and cannot be treated with chemotherapy. In these patients, biliary drainage may improve the patient’s quality of life (QoL) and allows the initiation of chemotherapy [30].

For the minority of patients who are able to undergo surgical resection, postoperative relapses are frequent and 5-year OS after surgery is only 25–35% [27, 31, 32]. Some of the main prognostic factors in patients with resected BTC are lymph node metastasis, depth of tumor invasion, and positive surgical margins [33‐35].

Adjuvant therapy has become a widely used option for patients who have undergone surgical resection, although a standard treatment approach in this setting has not yet been established [36]. No benefit with gemcitabine plus oxaliplatin (GEMOX) adjuvant therapy was observed in a French phase 3 trial (PRODIGE 12-ACCORD 18) in resected patients with BTC compared with surgery alone [37]. This finding was supported by a Japanese phase 3 trial (BCAT), which also did not show benefit of gemcitabine monotherapy over surveillance in resected eCCA [38].

More favorable results were obtained with adjuvant capecitabine (a fluoropyrimidine pro-drug) in a British phase 3 trial (BILCAP) that included 447 patients with BTC after curative-intent resection [39, 40]. Patients were randomized to either an observation group or capecitabine treatment for 24 weeks [39]. Although a survival improvement was noted in the capecitabine group relative to the observation group (median OS of 49.6 vs. 36.1 months), the difference between active treatment and the observation group did not reach statistical significance (HR: 0.84; 95% CI: 0.67–1.06). However, the OS benefit became statistically significant following adjustment for prognostic factors including tumor grade and stage (HR: 0.74; 95% CI: 0.59–0.94) [40], and a recurrence-free survival (RFS) benefit was shown in the first 24 months post-randomization (HR: 0.74; 95% CI: 0.57–0.96) [39]. Despite the lack of significance for the primary endpoint in this study, 6 months of adjuvant chemotherapy with capecitabine has been recommended in international [41] and French [42] clinical practice guidelines in patients with surgically resected BTC. Recently, the results of two randomized trials presented in congress (but not yet published) have reinforced the place of fluoropyrimidines in the adjuvant treatment of BTC. A South Korean randomized phase 2 trial (STAMP) in 101 patients with resected lymph node-positive eCCA failed to demonstrate a benefit of adjuvant cisplatin plus gemcitabine (CisGem regimen) over capecitabine [43]. Median disease-free survival (primary endpoint) was 14.3 months in the CisGem arm and 11.1 months in the capecitabine arm (HR: 0.96; 95% CI: 0.71–1.30; P = 0.86), and median OS was 35.7 months in both arms (HR: 1.08; 95% CI: 0.72–1.64; P = 0.81). A Japanese phase 3 trial (ASCOT) showed an OS benefit (HR: 0.69; 95% CI: 0.51–0.94) over surveillance alone after curative-intent resection of BTC with S-1, an oral fluoropyrimidine combination consisting of tegafur, gimeracil, and oteracil, which is widely used in Asia but not validated in Western populations [44]. However, RFS was not significantly improved (HR: 0.80; 95% CI: 0.61–1.04). The results of this study, although not transposable to the European population, reinforce the rationale for using a fluoropyrimidine in adjuvant therapy.

To date, data from real-world clinical practice settings in France on the use of adjuvant therapy, especially capecitabine, are lacking. In our experience, a relatively high proportion of patients may receive adjuvant chemotherapy; however, there is no overall information on exposure and dose/intensity of adjuvant treatment, and this gap in knowledge needs to be addressed.

For patients with unresectable iCCA, which generally presents as a liver-only or liver-predominant disease, loco-regional treatments are a promising therapeutic option that may be used in addition to chemotherapy [45]. Radioembolization (also known as selective internal radiotherapy (SIRT)) involves the administration of Yttrium-90 (⁹⁰Y)–labelled microspheres into the hepatic arteries to treat both primary liver tumors and metastases. A phase 2 trial in 41 patients with locally advanced, unresectable iCCA receiving SIRT combined with CisGem reported a centrally confirmed best overall response rate (ORR) of 39% at 3 months (90% CI: 26−53%) and a disease control rate (DCR) of 98% [46]. After a median follow-up of 36 months, median progression-free survival (PFS) was 14 (95% CI: 8–17) months, suggesting that the combination of chemotherapy with SIRT was beneficial for a significant proportion of patients. Furthermore, some patients were able to be downstaged, allowing for curative-intent secondary resection. Glass radiolabeled microspheres for the treatment of iCCA are now reimbursed in France [47].

Hepatic arterial infusion of chemotherapy (HAI) involves the delivery of chemotherapy directly into the liver. In a phase 2 trial, 38 patients with unresectable iCCA received HAI with floxuridine (FUDR, a fluoropyrimidine) plus intravenous gemcitabine-oxaliplatin combination chemotherapy (GEMOX regimen) [48]. The 6-month PFS rate was 80%. After a median follow-up of 30.5 months, partial radiographic response was observed in 58% of patients and DCR was 84%. In France, floxuridine is no longer used for HAI, which rather relies on oxaliplatin via implantable catheters and ports in patients with liver metastases as well as iCCA. An ongoing French phase 2 trial (GEMOXIA-02; NCT03364530) is evaluating HAI gemcitabine plus oxaliplatin in patients with iCCA who failed prior first-line systemic chemotherapy. A US phase 2 trial (NCT04251715) is currently recruiting patients with liver-dominant, unresectable iCCA to investigate the efficacy and safety of induction systemic chemotherapy with modified FOLFIRINOX (mFOLFIRINOX; oxaliplatin, irinotecan, 5-fluorouracil (5-FU), and leucovorin (LV)), followed by HAI with floxuridine and dexamethasone given concurrently with systemic modified FOLFIRI (irinotecan, 5-FU, and LV).

Other loco-regional treatments include external beam radiation therapy (EBRT), including stereotactic body radiotherapy, which delivers high-dose radiation beams to a tumor with high precision while limiting toxicity to surrounding tissue [49]. However, chemoradiotherapy with EBRT (50 Gy) plus 5-FU/cisplatin failed to show an improvement in PFS compared with GEMOX in a French randomized phase 2 trial of patients with locally advanced BTC, with a median PFS of 5.8 versus 11.0 months (HR: 0.65; 95% CI: 0.32–1.33) [50]; this trial was stopped early due to low patient recruitment. Transarterial chemoembolization (TACE) allows for the selective delivery of antitumor agents via the hepatic artery and has also been reported to limit toxicity in healthy tissue [51]. TACE using irinotecan-loaded beads was compared with CisGem in a randomized trial and showed an improved median OS of 33.7 versus 12.6 months, respectively (P = 0.048) [45].

Radiologically guided ablative procedures (e.g., radiofrequency ablation) have been advocated for the treatment of small (< 3 cm) single lesions when surgical resection is not an option [27]. Data from a meta-analysis that investigated the efficacy of locoregional therapies showed good outcomes for patients undergoing ablation, with a complete response rate of 93.3% and a median OS of 30.2 months [45]. Due to the quality of the studies investigated within the meta-analysis, ablation was the only strategy that received a strong recommendation. For EBRT and intra-arterial therapies, the pooled ORR from the meta-analysis was 23.4–41.3%, with a pooled mean OS of 14.1–21.3 months.

While evidence supporting the use of loco-regional therapies is mounting, there is still a relatively limited amount of robust data. French guidelines recommend that these therapies should be discussed on a case-by-case basis in patients with liver-only (or liver-dominant) iCCA, in combination with first-line systemic chemotherapy or as monotherapy following first-line treatment failure [42]. The results of two trials are currently awaited: a phase 3 trial (SIRCCA; NCT02807181) investigating first-line SIRT in patients with unresectable iCCA, and a randomized phase 2 trial (ABC-07; ISRCTN identifier 10639376) investigating EBRT.

Gemcitabine monotherapy has previously been used to treat patients with advanced BTC. An analysis of 100 cases (of whom 23 had iCCA and 25 had eCCA) demonstrated modest efficacy with an ORR of 7% (95% CI: 2.9–13.9) and median OS of 7.3 (95% CI: 5.4–9.2) months [52].

The gold-standard, first-line therapy for BTC is a CisGem regimen for 6 months, which has been shown to be more effective than gemcitabine alone in patients with locally advanced or metastatic BTC [53, 54], and is now the chemotherapy backbone for any new regimens being investigated in this setting. A pivotal UK phase 3 trial (ABC-02) demonstrated a significant PFS and OS benefit of CisGem over gemcitabine, regardless of the primary tumor site (Table 1) [53]. Of note, patients with an altered ECOG PS of 2 did not seem to benefit from the doublet chemotherapy to the same extent as those with a preserved ECOG PS of 0 or 1 [55]. Similarly, a Japanese randomized phase 2 study (BT-22) showed higher 1-year OS rate (39.0% vs. 31.0%) and median OS (11.2 vs. 7.7 months) with CisGem compared with gemcitabine [54].

Trials investigating alternative combinations for first-line treatment of BTC are also summarized in Table 1 [56‐62]. A single-center phase 3 study conducted in India suggested that gemcitabine and oxaliplatin doublet (a modified GEMOX regimen) provided similar outcomes compared with CisGem in patients with unresectable GBC [61]. A systematic review of 33 studies (1470 patients) compared CisGem and GEMOX for the treatment of advanced BTC [63]. The weighted median OS was 9.7 months for the CisGem group and 9.5 months for the GEMOX group. However, the CisGem combination was associated with higher toxicity compared with GEMOX [63]. A South Korean phase 3 study showed non-inferiority of capecitabine plus oxaliplatin (CAPOX) versus GEMOX in terms of 6-month PFS, and no significant differences in OS and ORR [57]. In a Japanese phase 3 trial (FUGA-BT), gemcitabine plus S-1 was non-inferior to CisGem [58]. Recently, a German randomized phase 2 study (NIFE) showed similar results with CisGem and the combination of nanoliposomal-irinotecan (Nal-IRI) plus 5-FU and LV in patients with advanced BTC (Table 1) [59]. Subgroup analyses suggested that the latter combination was of particular benefit in patients with eCCA, but this should be interpreted with caution due to the small number of patients.

Triplet chemotherapy regimens have also been evaluated in the treatment of patients with advanced BTC. All randomized trials that had evaluated a triple-agent therapy combining cisplatin-platinum with another molecule (mostly epidermal growth factor receptor inhibitors [64, 65] or antiangiogenic agents [66]) had been negative so far (Table 1). For instance, a French randomized phase 2 trial (PRODIGE 38 AMEBICA) in patients with advanced BTC and an ECOG PS of 0 or 1 failed to demonstrate a benefit of mFOLFIRINOX over CisGem in terms of 6-month PFS rate (the primary endpoint) or OS (median: 11.7 months in the mFOLFIRINOX group and 13.8 months in the CisGem group) [60]. Of note, exploratory subgroup analyses suggested that CisGem was superior to mFOLFIRINOX for patients with iCCA (HR: 0.58; 95% CI: 0.39–0.85; P < 0.1). The only exception to date came from Japan, where the cisplatin, gemcitabine, and S-1 triplet regimen improved OS compared with CisGem (median: 13.5 vs. 12.6 months; HR: 0.791; 90% CI: 0.620–0.996; P = 0.046) [67]. Among a number of non-randomized phase 2 studies that have assessed triplet chemotherapy regimens, oxaliplatin, irinotecan, and S-1 provided encouraging results, with an ORR of 50% and median PFS and OS of 6.8 months and 12.5 months, respectively [68]. In a US study, cisplatin, gemcitabine, and nanoparticle albumin-bound–paclitaxel provided an ORR of 45%, a median PFS of 11.8 months, and a median OS of 19.2 months (95% CI: 13.2 to not estimable) [62]. This regimen is currently being compared with CisGem in the US SWOG-1815 phase 3 study (NCT03768414).

GEMOX was developed in France and has historically been the preferred regimen in our country for the treatment of advanced BTC after the results of a phase 2 study [56]; however, CisGem is increasingly used in routine practice [60, 69‐73]. As a result, GEMOX may be used preferentially for patients who could be susceptible to hearing loss or abnormal renal function (with creatinine clearance > 30 mL/min).

The optimal duration of CisGem has not been clearly established. The ABC-02 trial showed a benefit for treatment with CisGem for up to 24 weeks, but median OS was not markedly different in the BT-22 trial in patients given CisGem for twice the duration (48 weeks); median OS was 11.4 months for the CisGem group in the BT-22 trial and 11.2 months in the ABC-02 trial [53, 54]. A South Korean study retrospectively analyzed data from 120 patients with advanced BTC who continued gemcitabine after six to eight CisGem cycles (maintenance group) compared with 111 patients who received no further maintenance chemotherapy after CisGem [74]. The median OS in the maintenance group was not significantly different from that in the observation group (22.4 months; 95% CI: 17.0–27.8 and 20.5 months; 95% CI: 15.4–25.6, respectively; P = 0.162); nevertheless, maintenance gemcitabine monotherapy is frequently administered in routine practice in France and elsewhere.

According to French [42] and European Society of Medical Oncology (ESMO) clinical practice guidelines [55] for BTC, systemic chemotherapy is the treatment of choice for patients with advanced disease; combination chemotherapy is recommended for patients with an ECOG PS of 0–1 and gemcitabine monotherapy for patients with an ECOG PS of 2. The recommended combination regimen for patients with an ECOG PS of 0–1 is CisGem.

FOLFOX (oxaliplatin, 5-FU and LV) is the reference second-line chemotherapy regimen for patients with advanced BTC, based on the findings of a UK phase 3 study (ABC-06) (Table 2) [75]. This study compared FOLFOX plus active symptom control (ASC, which included analgesia, biliary drainage, anti-emetics, antibiotics, steroids, other palliative treatment for symptom control, palliative radiotherapy, or blood transfusion) with ASC alone in patients with advanced BTC who had progressed on or after treatment with CisGem. This study demonstrated a significant, though modest, improvement in OS (median OS gain < 1 month) with FOLFOX plus ASC versus ASC alone, along with an ORR of 5% and a median PFS of 4 months. However, the 12-month OS rate was more than doubled in the FOLFOX group compared with the control group (25.9% vs. 11.4%), and FOLFOX was active even in patients resistant to CisGem. Interestingly, FOLFOX seemed less beneficial in patients with eCCA than in those with iCCA, with HRs of 0.84 (95% CI: 0.45–1.57) and 0.64 (95% CI: 0.38–1.06), respectively. Despite its modest efficacy results, the ABC-06 trial may have far-reaching implications for clinical practice. Its impact may be reminiscent of the effects of the study by Burris and colleagues in 1997, which showed a modest, albeit undisputed, benefit for more than a decade of gemcitabine over 5-FU as first-line treatment for advanced pancreatic cancer [76]. In France, the recommendation of FOLFOX as the second-line standard-of-care for BTC [42] may have hastened the shift from GEMOX to CisGem as first-line therapy.

Nal-IRI in combination with 5-FU and LV was investigated in South Korea in patients with metastatic BTC after progression on CisGem in a randomized phase 2 study (NIFTY) (Table 2) [77]. The Nal-IRI-5-FU-LV combination significantly improved PFS (7.1 vs. 1.4 months; HR: 0.56; 95% CI: 0.39–0.81; P = 0.0019) and OS (8.6 vs. 5.5 months; HR: 0.68; 95% CI: 0.48–0.98; P = 0.035) compared with 5-FU-LV [77]. Of note, the median OS in the 5-FU-LV group was close to that in the FOLFOX group in the ABC-06 study [75] (5.5 and 6.2 months, respectively), raising the question of whether the ABC-06 trial would have been positive if the comparator had been 5-FU-LV instead of ASC. The NIFTY results suggest that the Nal-IRI-5-FU-LV regimen is a promising second-line treatment option; however, these findings were not confirmed in a Western population. NALIRICC (NCT03043547) was a randomized phase 2 study that compared the efficacy and safety of Nal-IRI-5-FU-LV with that of 5-FU-LV in German patients with metastatic BTC who had progressed after first-line gemcitabine-based therapy [78]. The results of this study (presented at the 2022 ESMO conference) showed that, despite an improved ORR in the Nal-IRI-5-FU-LV arm (14.3% vs. 3.9%), there was no difference in median PFS (the primary endpoint) between patients receiving Nal-IRI-5-FU-LV and those receiving 5-FU-LV (2.8 vs. 2.3 months). Moreover, median OS tended to be longer in the group receiving 5-FU-LV than in those receiving Nal-IRI-5-FU-LV (8.2 vs. 6.9 months) [78].

To date, no randomized trial of chemotherapy has been reported for patients with BTC whose disease progressed after two systemic treatment lines. As a consequence, no recommendation has been made in this setting in French [42] and European [70] guidelines.

A cancer registry study in France found that, during the 5 years between 2001 and 2005, 14.2% of BTC patients underwent curative-intent surgery, 21.6% had palliative radiation therapy and/or chemotherapy, and 60.3% had BSC [7]. Between 1996 and 2005, only 15.9% of patients received adjuvant therapy. Interestingly, the proportion of patients receiving BSC decreased between 1976 and 2005, and the proportion of patients receiving palliative chemotherapy and/or radiotherapy, as well as curative-intent resection, increased over this period [7]. However, this shift has somewhat plateaued more recently, as attested by the abovementioned French nation-wide hospital database study of 3650 newly diagnosed iCCA cases (assessed between 2014 and 2015), in which 11.2% of patients underwent surgery, 23.8% received chemotherapy, and 65.0% received BSC [23]. Patient care was mainly provided by general hospitals (59.8%), rather than university hospitals (15.0%) and private (19.2%) or cancer centers (6.0%). In all, 28.4% of patients were admitted via the emergency room. A palliative care code was associated with the first hospital stay in 25% of patients and with a subsequent hospital stay in 60%. Of note, a total of 655 centers were involved in the patient’s first hospital stay and 28.8% of patients received care in low-volume hospitals (less than five patients with iCCA during the 2014–15 period), which represented 446 (68%) of the 655 centers, with a small proportion of patients being referred to high-volume centers, except for surgery. These findings reflect the marked decentralization of the French healthcare system. Comparatively, a retrospective US study of patients with iCCA who underwent surgical resection between 2000 and 2014 showed that 17% of patients received radiation therapy and 40% received chemotherapy, with the use of chemotherapy significantly increasing over time [79].

Under normal circumstances, the IDH1 and IDH2 enzymes catalyze the conversion of isocitrate to α-ketoglutarate; however, the mutated forms instead catalyze the nicotinamide adenine dinucleotide phosphate (NADPH)-dependent conversion of α-ketoglutarate to D-2-hydroxyglutarate (D-2HG), which is an oncometabolite (Fig. 2) [95‐97]. Numerous pro-tumor effects are seen as the α-ketoglutarate-dependent enzymes are competitively inhibited by D-2HG. These effects include a decrease in NADPH and glutathione, resulting in increased reactive oxygen species, DNA damage, inhibition of DNA repair mechanisms, and epigenetic dysregulation [97].

IDH mutations have been reported across other tumor types besides iCCA, including glioma, chondrosarcoma, and acute myelogenous leukemia. Both IDH1 and IDH2 mutations are the predominant gene alterations in iCCA [83]. Recent genomic profiling data from iCCA samples have indicated that IDH1 and IDH2 mutations have various genetic locations, and are mutually exclusive, suggesting that IDH1 and IDH2 mutations are driver oncogenes in iCCA [98]. Of note, a meta-analysis of 45 studies determined that IDH1 mutations were more prevalent than IDH2 mutations in CCA, and that IDH1 mutations were significantly more frequent in non-Asian than in Asian patients (weighted mean, 16.5% vs. 8.8%; P < 0.001) [99]. In patients with advanced iCCA, IDH mutations have not been found to be associated with patient prognosis [100].

Ivosidenib is a potent targeted inhibitor of the IDH1 mutant enzyme and was approved in the USA in August 2021 for use in patients with CCA, based on the findings from a double-blind, placebo-controlled phase 3 trial (ClarIDHy) (Table 3) [85, 94]. The ClarIDHy trial was conducted across six countries, including France, and demonstrated that in addition to being well tolerated, ivosidenib improved PFS (the primary endpoint) compared with placebo in chemorefractory, IDH1-mutated CCA (Fig. 3) [85]. Patients taking ivosidenib (n = 124) had significantly improved PFS compared with those in the placebo group (n = 61), with a median PFS of 2.7 (95% CI: 1.6–4.2) months versus 1.4 (95% CI: 1.4–1.6) months, respectively (HR: 0.37; 95% CI: 0.25–0.54; P < 0.0001 one-sided). The PFS rate was 32% (95% CI: 23–42) at 6 months and 22% (95% CI: 13–32) at 12 months for ivosidenib, whereas no patients in the placebo group were free from progression at 6 months or later [85]. Median OS was 10.3 (95% CI: 7.8–12.4) months with ivosidenib (n = 126) and 7.5 (95% CI: 4.8–11.1) months with placebo (n = 61; HR: 0.79; 95% CI: 0.56–1.12; P = 0.09 one-sided). In an OS analysis adjusted for crossover to ivosidenib (70% of patients in the placebo group), the placebo group had an adjusted median OS of 5.1 (95% CI: 3.8–7.6) months (HR: 0.49; 95% CI: 0.34–0.70; P < 0.001 one-sided) [94]. While long-term follow-up in ClarIDHy of patient health-related QoL (HRQoL) was limited by small sample size, initial results suggested that patients treated with ivosidenib were likely to maintain their baseline HRQoL scores, and none experienced a clinically meaningful deterioration in the domain of physical functioning [101]. In the latest French and ESMO guidelines [42, 55], ivosidenib is recommended for patients with CCA and IDH1 mutations who have progressed after at least one prior line of systemic chemotherapy.

Fibroblast growth factors (FGFs) control a range of cellular processes, including the regulation of cellular survival, proliferation, migration, and differentiation, through their high-affinity binding to four tyrosine kinase receptors (FGFRs) [102]. Genomic alterations in FGFR genes, including chromosomal translocations, amplifications, and gain-of-function mutations, have been associated with a range of cancers, including BTC [103, 104]. Interestingly, FGFR2 alterations are generally—but not absolutely—mutually exclusive with IDH mutations [82]. Occurring almost exclusively in iCCA, translocations of FGFR2 are present in about 13.6% of iCCA patients, and result in activation of mitogen-activated protein kinase-promoting tumorigenesis, allowing anchorage-independent growth [103].

Pemigatinib is a selective, potent, oral inhibitor of FGFR1, 2, and 3 [105]. The FIGHT-202 study, a multicenter, open-label, phase 2 study, examined the effects of pemigatinib on outcomes in chemorefractory patients with iCCA, most of whom had FGFR2 fusions/rearrangements (Table 3) [86]. Among the 107 patients with FGFR2 fusions or rearrangements, 38 had an objective response (three with complete response and 35 with partial response (Fig. 4)) for an ORR (the primary study endpoint) of 35.5% (95% CI: 26.5–45.4%). The median duration of response was 7.5 months, and the median OS was 21.1 months. Of 103 patients with FGFR2 fusions or rearrangements and post-baseline tumor measurements, 91 (88%) had a decrease from baseline in target lesion size. For patients with FGFR2 fusions or rearrangements, objective responses were seen across all demographic and disease subgroups assessed. None of the patients with other FGF/FGFR alterations (n = 20) nor those without an FGF/FGFR alteration (n = 18) had an objective response, but stable disease was reported in eight patients with another FGF/FGFR alteration, as well as in four patients without an FGF/FGFR alteration [86].

Several other FGFR inhibitors are currently under investigation for use in patients with iCCA. Infigratinib has shown a manageable safety profile and preliminary clinical activity against iCCA tumors with FGFR alterations in a phase 2 study [106]. Patients with confirmed FGFR2 or other FGFR alterations had an ORR of 23.1% (95% CI: 15.6–32.2), with one complete response and 24 partial responses seen [106]. Derazantinib is another multi-kinase inhibitor currently undergoing phase 2 assessment [107]. Data showed that, in addition to a manageable safety profile, derazantinib showed good antitumor activity in patients with FGFR2 fusion-positive, advanced iCCA who had progressed on first-line chemotherapy, with an ORR of 20.7% and a DCR of 82.8%. Futibatinib, an FGFR 1–4 inhibitor, has been investigated in the phase 2 FOENIX-CCA2 trial, the results of which were presented at the 2022 American Society of Clinical Oncology conference [108]. In 103 patients with iCCA and FGFR2 alterations, the ORR (the primary study endpoint) was 41.7%, DCR was 82.5%, median PFS was 8.9 months, and median OS was 20 months, with 12-month PFS and OS rates of 35.4% and 73.1%, respectively [108]. Futibatinib had a tolerable safety profile, with hyperphosphatemia being the most common adverse event (85%). Another FGFR inhibitor, erdafitinib, is currently under investigation in a phase 2a study to assess its efficacy and safety in Asian patients with advanced CCA who have FGFR alterations [109]. Interim findings showed, in addition to an acceptable safety profile, six partial responses and four stable disease responses in the 12 evaluable patients treated with erdafitinib (ORR of 50% and DCR of 83.3%). Of the ten patients who were FGFR-positive, the ORR was 60% and the DCR was 100%. Several ongoing phase 3 studies are comparing FGFR inhibition with pemigatinib (FIGHT-302; NCT03656536), infigratinib (PROOF; NCT03773302), or futibatinib (FOENIX-CCA3; NCT04093362) to CisGem as first-line therapy for patients with advanced iCCA harboring FGFR2 rearrangements. The 2022 French and ESMO guidelines now recommend FGFR inhibitors (e.g., pemigatinib, futibatinib, infigratinib) for patients with CCA and FGFR2 fusions who have progressed after at least one prior line of systemic chemotherapy [42, 55].

Next-generation FGFR2 inhibitors are also in development. Preliminary results from the phase 1/2 REFOCUS study of RLY-4008 (presented at the 2022 ESMO conference) reported a confirmed ORR of 82.4% and DCR of 100% at the recommended phase 2 dose (n = 17 patients), with 88.2% of patients remaining on treatment with this dose [110]. Further data are awaited with interest.

Data are also needed on the optimal sequencing of FGFR inhibitors. A retrospective US study suggested that patients who progress after one FGFR inhibitor can receive treatment with another FGFR inhibitor, with similar outcomes to patients receiving post-FGFR inhibitor chemotherapy [111]. However, only about 50% of patients are eligible for a second course of FGFR inhibitor treatment after disease progression.

The DNA mismatch repair (MMR) system is one of the cellular DNA repair mechanisms that maintain genomic integrity during cell division [112]. A proficient system is denoted as pMMR. A deficient MMR system (dMMR phenotype) is unable to repair replication errors, which are particularly frequent in repeated sequences such as microsatellites. These recurring sequences of DNA are about 10–60 base pairs long, each made of repeating units of one to five base pair motifs [112]. Microsatellite instability (MSI) can result from MMR malfunction, leading to uncontrolled microsatellite lengths. The MSI genotype has been observed in multiple cancers. dMMR tumors present with a high MSI (MSI-H), due to varying lengths in microsatellite sequences and frameshift mutations resulting in truncated or abnormal (highly immunogenic) proteins, as well as the activation of oncogenes [112].

The proportion of MSI-H status among BTC patients is controversial. A whole exome-sequencing analysis found dMMR or MSI-H status in 36% of 260 patients with BTC [80]. A systematic review estimated that 10%, 5–13%, and 5% of iCCA, eCCA, and GBC cases, respectively, harbor dMMR and/or MSI-H [113]. Even lower proportions of 1–2% have been reported [112, 114, 115].

dMMR/MSI-H tumors are sensitive to immune checkpoint blockade with anti-programmed death 1 receptor (PD-1) antibodies, regardless of the cancer’s tissue of origin [116, 117]. Pembrolizumab, a monoclonal antibody that binds to PD-1, was shown in the KEYNOTE-158 study (Table 3) to have durable antitumor activity, as well as an acceptable safety profile, in patients with CCA who were not responding to other therapies, with two complete responses and seven partial responses among 22 patients (ORR: 40.9%) [93]. Based on these data, the 2022 French and ESMO guidelines recommend pembrolizumab in patients with MSI-H/dMMR who show progression after, or intolerance to, prior treatments [42, 55].

Neurotrophin receptor tyrosine kinase (NTRK) gene fusions are found in > 90% of cases of several rare tumors (e.g., infantile fibrosarcoma, secretory breast carcinoma, and mammary analogue secretory carcinoma), but with a low frequency (< 1–5%) in most other tumor types, and are considered to be tumor-agnostic alterations. The NTRK1, NTRK2, and NTRK3 genes encode the tropomyosin receptor kinase (TRK) proteins, which are members of the tyrosine kinase family. Following a fusion event, these overexpressed chimeric proteins can drive oncogenic transformation via ligand-independent activation of the TRKs [118, 119]. NTRK gene fusions are reported to account for approximately 2% of all genetic alterations in advanced BTC [120]. A recent retrospective study assessing next-generation sequencing (NGS) results of 29 patients with various cancers that were NTRK gene fusion-positive, determined that the frequency of co-occurrence of NTRK fusion with MSI-H was 17.6% and co-occurrence of NTRK fusion with a high tumor mutation burden (> 20 mutations/Mb) biomarker was 20.7% [121]. While this result has yet to be validated in patients with BTC, it may inform future diagnostic and therapeutic strategies.

Both larotrectinib and entrectinib target NTRK fusions. In a clinical study, patients treated with larotrectinib showed a significant objective response across a range of NTRK fusion-positive cancers (Table 3). Only two patients in that study had CCA: one had a decrease in tumor size and the other had progressive disease [87]. Data to date from the STARTRK-2 study showed that patients with a range of advanced NTRK fusion-positive solid tumors (including one patient with CCA) also had a high response rate with entrectinib [122]. Despite the lack of data, the French and ESMO guidelines recommend NTRK inhibitors for BTC patients with NTRK fusions who have progressed after, or are intolerant to, prior therapy [42, 55].

One of the other targets of interest in the second-line setting is human epidermal growth factor receptor (HER2) [123]. HER2 (ERBB2) amplification, overexpression, or both are observed in approximately 15–30% of GBC, eCCA, and ampullary carcinomas, and in 4–5% of iCCA cases [6, 82, 124‐126].

Recently published were the findings from the MyPathway trial, which investigated the efficacy of a dual anti-HER2 regimen, pertuzumab + trastuzumab, in previously treated patients with metastatic BTC who had either or both HER2 amplification and overexpression [127]. In this non-randomized, multicenter, open-label phase 2a trial, nine of the 39 enrolled patients achieved a partial response (ORR: 23%; 95% CI: 11–39).

A phase 2 study investigated the efficacy of FOLFOX plus trastuzumab as second-line or third-line treatment for patients with HER2-positive BTC that had progressed despite chemotherapy with gemcitabine and cisplatin [128]. Of the 34 patients enrolled, ten patients had a partial response and 17 had stable disease, to give an ORR (primary study endpoint) of 29.4% (95% CI: 16.7–46.3) and a DCR of 79.4% (95% CI: 62.9–89.9). Median PFS was 5.1 (95% CI: 3.6–6.7) months and median OS was 10.7 (95% CI: 7.9–not reached) months.

Zanidatamab is a bispecific HER2-targeted antibody that is accessible in France via an early access program. A phase 1 study, investigating the safety and preliminary efficacy of zanidatamab in BTC patients with centrally confirmed HER2 overexpression has demonstrated acceptable tolerability and promising and durable anti-tumor activity with this agent [129]. In this non-randomized, multicenter study, eight of the 17 patients evaluable for response had a response (ORR: 47%; 95% CI: 23–72); further randomized clinical trials are ongoing in this population (NCT04466891).

The HERB study investigated the efficacy and tolerability of trastuzumab deruxtecan, an antibody-drug conjugate consisting of the humanized monoclonal antibody trastuzumab covalently linked to the topoisomerase I inhibitor, deruxtecan, in patients with HER2-expressing BTC who were refractory or intolerant to a gemcitabine-containing regimen [130]. In this investigator-initiated, multicenter, single-arm phase 2 trial, two and six of the 22 enrolled HER2-positive patients achieved a complete and partial response, respectively (ORR: 36.4%; 90% CI: 19.6–56.1; P = 0.01) and the DCR was 81.8%. Median OS was 7.1 months. Interestingly, this study reported promising results even in the subgroup of HER2-low patients (n = 8), who had a DCR of 75.0% and a median OS of 8.9 months [130].

While these results are promising, physicians need to be aware of accuracy problems inherent in HER2 testing, with high rates of discordance (20–25%) between central and local readings observed in breast cancer studies [131]. Moreover, the pattern of HER2 staining on immunohistochemistry in breast cancer is different from that in digestive tract cancers, and the optimal testing modality and criteria for HER2 overexpression in BTC have yet to be defined [132]. Studies to date in BTC have used the HER2 expression criteria defined for gastroesophageal cancer [125, 126].

The 2022 French and ESMO guidelines state that HER2-directed therapies “can be considered” (rather than recommended) in patients with progressive advanced disease and relevant genetic alterations [42, 55], noting that none of these treatments have yet been approved for use in BTC by the European Medicines Agency or the US Food and Drug Administration [55].

The prevalence of v-raf murine sarcoma viral oncogene homolog B1 (BRAF) mutations in patients with BTC is relatively low, with the vast majority being BRAFV600E (~ 3%) [133].

The combination of dabrafenib (a BRAF inhibitor) and trametinib (a mitogen-activated protein kinase (MEK) inhibitor) is under investigation in the Rare Oncology Agnostic Research (ROAR; NCT02034110) basket trial as a potential regimen for patients with BRAFV600E-mutated, advanced BTC. This is a phase 2, open-label, single-arm, multicenter trial. An investigator-assessed objective response was observed in 22 of the 43 evaluable patients to date, with an ORR of 51% (95% CI: 36–67%) [134]. Of note, these results support “classical” BRAF therapeutic manipulation in BRAFV600E-mutated BTC, i.e., BRAF inhibition potentiated by downstream MEK inhibition without the need for negative feedback loop inhibition via an epidermal growth factor receptor inhibitor such as in advanced colorectal cancer.

At the time of writing, dabrafenib-trametinib is approved for BTC in the USA, but not in Europe. Nevertheless, the current French and ESMO guidelines recommend this treatment for patients with progressive BTC after systemic therapy and BRAFV600E mutations [42, 55].

To assist with the determination of the best available therapy based on genomic markers, ESMO developed the ESMO Scale for Clinical Actionability of molecular Targets (ESCAT) system [135]. This system ranks and matches genomic alterations with targeted therapies according to their clinical utility, as shown in Table 4 [135].

The ESMO Precision Medicine Working Group currently recommends NGS for a range of genomic alterations with actionable targets (Table 5) [120]. IDH1 mutations in advanced CCA have received an IA rating [120], indicating that the evidence thoroughly supports the use of the IDH1-targeted therapy for patients with this mutation. FGFR2 fusions received an IB rating in advanced CCA based on results from the FIGHT-202 trial of pemigatinib [86]. This rating supports the use of targeted therapy in this group, albeit without the same level of evidence as ivosidenib. Other rankings for less common mutations include MSI-H and NTRK (both with a rating of IC), as well as BRAF, which has a ranking of IIB. Certain genomic alterations, such as HER2, were not listed, but could be now considered as IIB due to specific trial data [127]. This highlights the importance of genetic sequencing of tumor samples, which is recommended by the French and ESMO guidelines [42, 55, 120], to identify patients who may benefit from targeted therapy, to screen patients for inclusion in clinical trials, and to further drug research and capture valuable data. An improved understanding of genetic alterations and their possible associations will lead to better treatments.

Numerous trials are currently underway to investigate targeted therapies for first-line use in patients with BTC (Table 6). In addition to the targets already recommended by the ESMO Precision Working Group, targets such as programmed cell death ligand 1 (PD-L1) are under active investigation. Monoclonal antibodies targeting immune regulatory checkpoint molecules, such as PD-1 and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), as well as associated ligands, have been shown to be effective against various tumors by mediating immune recognition of the tumor [136‐138]. Consequently, agents targeting PD-1/PD-L1 and CTLA-4 are being increasingly studied, despite reports that an effective antitumor response may be induced in < 30% of patients [139].

An international phase 3 study (TOPAZ-1) compared CisGem in combination with the anti-PD-L1 monoclonal antibody durvalumab (n = 341) with CisGem + placebo (n = 344) in patients with advanced BTC. The addition of durvalumab significantly prolonged OS (median: 12.8 vs. 11.5 months; HR: 0.80; 95% CI: 0.66–0.97; P = 0.021), and PFS (median: 7.2 vs. 5.7 months; HR: 0.75; 95% CI: 0.64–0.89; P = 0.001); ORR were 26.7% versus 18.7%, respectively [140, 141]. Although durvalumab is approved in the USA for BTC and was approved very recently in Europe (December 2022), it is not currently reimbursed in France for this indication (although in the latter, an early access program has been recently implemented). Another phase 3 study (KEYNOTE 966) investigating an anti-PD-1 monoclonal antibody (pembrolizumab) in patients with previously untreated BTC is currently underway. That double-blind trial is expected to enroll 1,048 patients, who will be randomized to receive either CisGem + pembrolizumab or CisGem + placebo. The primary endpoints are PFS and OS [142].

Numerous trials are also underway to evaluate targeted therapies in the second-line setting (Table 7). These include therapies targeting FGFR2 and vascular endothelial growth factor (VEGF). VEGF is a common cancer target and has been found to be overexpressed in BTC – one study reported VEGF overexpression in 53.8% of iCCA cases and 59.2% of eCCA cases [1]. Treatment with VEGF inhibitors helps to prevent angiogenesis and VEGF-related tumor metastasis. Single-arm studies investigating therapies for iCCA include the MATCH Screening trial (NCT02465060), which is investigating the multi-targeted receptor tyrosine kinase inhibitor sunitinib, among other targeted therapies, as a possible treatment for patients with advanced, refractory solid tumors, including iCCA.

Kirsten rat sarcoma viral oncogene homolog (KRAS) mutations occur with widely varying frequencies among solid cancers from 0% to up to 90% of pancreatic cancers, but only a minority are KRAS G12C mutations. In the phase 2 KRYSTAL-1 trial (NCT05162443), four (50%) of the eight pretreated patients with BTC experienced a partial response with adagrasib, a KRAS G12C-selective inhibitor [143]. Enrollment in this study continues, and an early access program has been initiated in patient populations with KRAS G12C–mutant solid tumors. In addition, other studies are starting to evaluate drugs that target the other KRAS mutations in pancreatic and other cancers, starting with KRAS G12D (MRTX1133), which is the most common alteration.

As shown in Table 8, various trials are now underway investigating the use of combination of targeted therapies, mostly with immune checkpoint inhibitors.