*2.2. Metabolic Regulator (IDH Inhibitors)*

Reprogramming of cancer cells' metabolism has been defined as one of the hallmarks of cancer [15] and represents a possible target for precision medicine. Genomic and transcriptomic studies [16] have demonstrated that isocitrate dehydrogenase 1 and 2 (IDH1, IDH2) mutations occur in 13–25% of iCCA. These enzymes are involved in tricarboxylic acid cycle (TCA), β-oxidation of unsaturated fatty acids, response to oxidative stress, and expression of chromatin remodelers. In IDH1/2-mutated cells, the oncometabolite D-2 dihydroxyglutarate (2-HG) accumulates, leading to metabolic and epigenetic changes, enhanced proliferation, and susceptibility to DNA damage. This pathway may be hampered by inhibitors of IDH1 (AG120) and IDH2 (AG221), such as ivosidenib and enasidinib (NCT02273739), with encouraging results in randomized control trials (RCTs). Patients with IDH1-mutated iCCA who had progressed on previous therapy [17] showed a significant response to ivosidenib when compared to placebo-administered patients in the ClarIDHy phase III double-blind clinical trial (Table 1), in terms of both progression-free survival (2–7 vs. 1–4 months) and overall survival (10–8 vs. 9–7 months). Based on these results, ivosidenib has been recently approved by the FDA for locally advanced and metastatic cholangiocarcinoma with IDH1 mutations. IDH1 inhibitors are currently under investigation also in combination with other treatments. A phase Ib/II basket trial is evaluating Olutasidenib (FT-2102) alone, in combination with azacitidine, nivolumab, or gemcitabine and cisplatin in 200 patients with different solid tumors harboring the same IDH1 mutations (NCT03684811).

#### *2.3. Tyrosine Kinase Inhibitors*

Mutations of epidermal growth factor receptors play a pivotal role in different cancers [18], and several drugs are already approved for specific subsets of malignancies, i.e., EGFR-mutated non-small cell lung cancer [19] and colorectal cancer [20]. Nevertheless, convincing evidence of their efficacy in CCA is still lacking.

In the PiCCA phase II randomized clinical trial [21], panitumumab, a monoclonal anti-EGFR1 antibody, was administered in combination with gemcitabine and cisplatin in KRAS-wild-type patients versus gemcitabine and cisplatin alone, but it failed to improve ORR, PFS, and OS. Similar results were obtained in a phase II study in chemotherapy-naive patients with advanced BTC, treated with panitumumab and GEMOX and GEMOX alone. Despite the attempt of selecting patients by IHC, PCR, and Sanger sequencing for KRAS, BRAF, and PI3KCA, no significant survival differences were observed. Nevertheless, it needs to be underlined that the cohorts of these two studies were not specifically tested for enrichment in EGFR alterations [22]. In addition, a phase II clinical trial studied the

efficacy of cetuximab combined with GEMOX vs. GEMOX alone in advanced BTC patients; KRAS, NRAS, and BRAF mutations and EGFR expression, were the criteria selected to stratify these patients. Despite a significant difference in progression-free survival, the study did not reach the primary endpoint (ORR) nor demonstrated a higher OS in the cetuximab arm. However, other genetic alterations involved in the EGFR pathway, i.e., ROS1, ALK, or c-MET [23], were not specifically investigated and might have a role in explaining anti-EGFR resistance.

The EGFR inhibitor erlotinib (Table 1) was studied in combination with chemotherapy regimens [24] and bevacizumab [25], but no clear survival benefits were observed when compared to current standard of care. Varlitinib, a competitive inhibitor of the tyrosine kinases EGRF and HER 2–4, is currently under investigation in monotherapy (phase II, NCT02609958) and in combination with capecitabine in advanced BTC patients (phase II/III, NCT03093870) (Table 1).

As far as the HER family is concerned, molecular profiling studies [26] have underlined the frequency of ERRB2 aberrations in p/dCCA, but evidence about the efficacy of anti-HER2 drugs in CCA has not supported their use in clinical practice so far [1]. On these bases, the feasibility of this treatment has already been demonstrated [27], and several phase II clinical trials are currently evaluating the efficacy of combination treatments with trastuzumab and tucatinib (NCT04579380) and with chemotherapy (NCT04430738).

Combination treatments with bevacizumab and gemcitabine or capecitabine have been tested in a multicenter phase II trial, given the high prevalence of VEGF overexpression in CCA [28]. Nevertheless, the patients were not selected based on their mutational profile, and this may be responsible for the poor outcome of the study.

The lack of patients' stratification may have also affected the results of different clinical trials that evaluated the multikinase inhibitor sorafenib, also targeting VEGFR2 and 3 [29]. Adding sorafenib to GEM–CIS in biliary tract cancer showed increased treatment toxicity without simultaneous clinical benefits in a phase II RCT [30] including biliary adenocarcinomas of all subtypes without taking into account histological and molecular differences. Sun et al. [31] have shown that regorafenib improved PFS of (15.6 weeks) and OS (31.8 weeks) in advanced BTC patients with disease progression after first-line therapy. Targeting neurotrophic tyrosine kinase receptor (NTKR) fusions has seemed promising, too [32]. Two phase II basket trials have investigated entrectinib [33] and larotrectinib [34]. FDA and EMA have approved larotrectinib and entrectinib as "wildcard" drugs that can be used in every kind of malignancy harboring this genetic alteration, regardless of the anatomical origin. Unfortunately, NTKR fusions are rarely detected in CCA [35].

#### *2.4. Proteasome Inhibitors*

Mutations/deletions of the PTEN gene were observed in approximately 5% of iCCAs associated with poor prognosis [6]. It was also observed that PTEN mutation/deletion is also associated with increased activity of proteasomes in iCCAs. On these bases, a phase III study (Table 1) is actually evaluating the efficacy of Bortezomib, a proteasome inhibitor, in patients with advanced iCCA who have progressed after at least two cycles of systemic chemotherapy (NCT03345303).

#### **3. Immunotherapy**

Since 2010, immunotherapy has been one of the most important strategies in the treatment of malignancies, together with surgery, chemotherapy, radiotherapy, and targeted therapy, even if its efficacy is very variable, and only a percentage of patients obtain a durable response [36]. The mechanism of immunotherapy is to enhance the anti-tumor immune response, including both adaptative cells (B and T cells) and innate cells such as macrophages, neutrophils, natural killers. Immunotherapy includes immune checkpoint inhibitors (ICIs) targeting programmed death 1 (PD-1), programmed death-ligand 1 (PD-L1), and cytotoxic T lymphocyte antigen-4 (CTLA-4), cancer vaccines, and adoptive cell transfer (ACT). Several factors can influence the effect of immunotherapy-based treat-

ments: the environment of tumor and immune cells, vascularization, extracellular matrix, and molecular signaling pathway [37]. Several therapeutic options in patients affected by biliary tract cancers are under investigation, such as immunotherapeutic strategies with checkpoint inhibitors, peptide- and dendritic cell-based vaccines, and adoptive cell therapy, in monotherapy or in combination with targeted therapy and/or chemotherapy. Nowadays, scientific evidence on the use of immunotherapy in CCA are limited, although different trials are currently investigating the role of anti-CTLA-4 monoclonal antibodies, the targeting of PD-L1 or its receptor, PD-1, and chimeric antigen receptor T (CAR-T) cell immunotherapy. Unfortunately, checkpoint inhibitor monotherapy has shown low efficacy in CCA patients. Indeed, Pembrolizumab, a PD-L1 inhibitor, demonstrated a median progression-free survival of 1.8 months in patients affected by CCA in the phase Ib basket trial KEYNOTE 028 [38]. Checkpoint inhibitors showed encouraging results in patients with microsatellite instability or DNA mismatch repair in the KEYNOTE 158 trial [39], even if only a small percentage of patients with a positive response to this kind of treatment reported a better clinical response [40]. Pembrolizumab demonstrated good efficacy in a recent Korean study that retrospectively analyzed 51 patients with PD-L1-positive CisGem-refractory biliary tract cancer. In PD-L1-positive patients, pembrolizumab showed durable efficacy, with a 9.8% response rate with manageable adverse events. Ongoing studies and clinical trials are currently exploring combined immunotherapeutic approaches targeting both the innate and the adaptive immune system, and/or combined strategies also involving chemotherapy or radiation.

Particularly, there are many ongoing phase I–III trials exploring the role of targeting PD-L1, its receptor PD-1, anti CTL-A4 with monoclonal antibodies in monotherapy or in combination with chemotherapy, targeted therapy, local ablative therapy, and the role of CAR-T cell immunotherapy in biliary tract cancer (Tables 2 and 3). In particular, KEYNOTE-028 and KEYNOTE-158, two multicentric, non-randomized, open-label, phase IB and II trials, showed a durable antitumor activity of Pembrolizumab in 6–13% of patients with advanced BTC. In KEYNOTE-158, they observed a median progression free survival (PFS) of 2.0 months and a Median overall survival (OS) of 7.4 months; adverse events were mainly mild to moderate in severity [39]. Another immunotherapeutic agent, Nivolumab showed a response rate of 22% and a disease control rate of 59% in a Phase II multi-institutional study including 46 patients affected by advanced biliary tract cancer in second-line therapy [41].

The combination of immunotherapy and chemotherapy looks promising. Two Phase III trials are evaluating the efficacy and safety of KN035 plus Gemcitabine–Oxaliplatin compared to standard of care Gemcitabine–Oxaliplatin therapy (NCT03478488) and the association of Durvalumab and Gemcitabine plus cisplatin (NCT03875235). BilT-01, a multicenter randomized Phase II trial, described a prolonged PFS six months after the addition of nivolumab to gemcitabine and cisplatin (NCT02829918) [42]. LEAP 005 demonstrated a promising antitumor activity and manageable toxicity of Pembrolizumab in combination with Lenvatinib in 31 patients affected by BTC [43].

Regarding Adoptive Cell Therapy (ACT), a phase III, non-randomized trial is studying the role of cytokine-induced killer cells in association with radiofrequency ablation in 50 patients with CCA (NCT02482454).


**Table 2.** Ongoing immunotherapy trials of biliary tract cancers.


**Table 2.** *Cont*.

ACT: adoptive cellular therapy, BTC: biliary tract cancer, CAR-T cell: chimeric antigen receptor T cell, CCA: cholangiocarcinoma, FOLFOX: folinic acid (leucovorin) + 5-fluorouracil + oxaliplatin, GEM: gemcitabine, GEMCIS: gemcitabine + cisplatin, HCC: hepatocellular carcinoma, iCC: intrahepatic cholangiocarcinoma, ICI: immune-checkpoint inhibitors, MUC-1: mucin 1, PDAC: pancreatic ductal adenocarcinoma, RFA: radiofrequency ablation, RT: radiotherapy, SBRT: stereotactic body radiation therapy, TACE: trans-arterial chemo embolization, TME: tumor microenvironment.


**Table 3.** Ongoing immunotherapy trials for BTC with preliminary results.

#### **4. Clinical-Pathological and Radiomic Monotherapy Susceptibility in Patients with Cholangiocarcinoma**

Within the CCA clinical-pathological spectrum, the pattern of tumor growth has been correlated with specific histological features, e.g., small-bile duct iCCAs and cholangiolocarcinoma (CLC) showed a mass-forming growth pattern, while large-bile duct iCCAs showed both a mass-forming growth pattern and a combination of a mass-forming growth pattern with a periductal infiltrative growth pattern, the latter being the typical pattern of growth of pCCA [44]. Mass-forming iCCAs showed more heterogeneous clinical-pathological characteristics than other gross types [45]. Radiologically, at dynamic contrast-enhanced imaging, all large-bile duct iCCAs showed concentric filling at the venous phase, whereas small-bile duct iCCAs/CLCs showed washout in various patterns, in a clinical-pathological study including correlates with magnetic resonance imaging [44].

The USA Food and Drug Administration approved the use of pembrolizumab for patients with advanced solid tumors lacking the expression of mismatch repair (MMR) proteins (MLH1, MSH2, MSH6, and PMS2) or having high microsatellite instability (MSI-H) [46]. MMR proteins can be inactivated through somatic or germline mutations or they can be silenced through promoter hypermethylation, e.g., of the MLH1 gene [47]. These alterations culminate to hypermutation during DNA replication (MSI) and may lead to the development of malignancies [48]. Interestingly, such molecular alterations predispose to an increase of the neoantigen load of the tumor, promoting susceptibility to immunotherapies targeting the PD-1 pathway because of the increased inflammation surrounding these tumors [40].

Given the potential for immunotherapy in patients with CCA, authors studied the expression of PD-L1/PD-1 and evaluated the presence of associated genetic alterations. For example, in 652 biliary tract cancers that comprised 77 p/dCCA, 372 iCCA, and 203 gallbladder cancer (GBC), 8.6% tumors were PD-L1-positive [GBC 12.3% (25/203), iCCA 7.3% (27/372), and p/dCCA 5.2% (4/77)]. Interestingly, there was an increase in BRAF, BRCA2, RNF43, and TP53 mutations in the PD-L1-positive group with respect to the PD-L1 negative one. Furthermore, there was an association between PD-L1 expression and certain biomarkers (TOP2A, TMB high, MSI-H). As noted by the authors, the aforementioned combinations of molecular alterations might direct the use of rational combination strategies and clinical trial development [49]. On the same line, Ju et al. analyzed 96 cases of CCA for morphology using H&E staining and for mutations of MMR genes using immunohistochemical staining. The authors found that 6% of the samples showed MMR deficiency (MMR-d). Divided by location, 10% (3 of 31) of iCCA and 5% (3 of 65) of p/dCCA were MMR-d. The best predictive factor for MMR-d was a nontypical infiltrating pattern of invasion [50].

The increasing awareness of CCA heterogeneity at the morphological and molecular levels, together with the advent of radiomic, artificial intelligence (AI), and machine learning, has revitalized the study of radiological correlates. For example, it has been shown that the magnetic resonance imaging texture signature, including three wavelets and one 3D feature, has the ability to discriminate inflamed from non-inflamed immunophenotypes based on the density of CD8+ T cells. This may be a surrogate of the response to immune checkpoint blockade [51]. The preoperative prediction of PD-1/PD-L1 expression and outcome in iCCA patients using magnetic resonance biomarkers and a machine learning approach has been attempted [52]. Utilizing qualitative and quantitative imaging traits, reasonable accuracy in predicting tumor grade and higher AJCC stage in iCCA has been shown [53].

#### **5. Conclusions**

The role of targeted therapy and immunotherapy in the treatment of CCA is currently under investigation. These options might improve survival outcomes (OS and PFS), as shown by the promising results of several clinical trials illustrated in the present review. This is even more important considering the poor therapeutic options in the management of CCA. The co-presence of driver mutations and markers of susceptibility to immunotherapy may lead to rational therapeutic combination strategies and clinical trial development. The combination of new therapeutic strategies, such as targeted therapy and immunotherapy, with conventional chemotherapy and/or locoregional treatments could be the next frontier for the treatment of advanced CCA. The evaluation of innovative strategies for the prediction of immunotherapy susceptibility, such as multi omics, preferably within longitudinal clinical trials, and the use of systems of data analysis based on the precepts of AI, may circumvent the lack of therapeutic biomarkers for immunotherapy. A better understanding of immunological-based therapeutic weapons is needed, which will lead to a form of a precision medicine strategy capable of alleviating the clinical aggressiveness and to improve the prognosis of CCA.

**Author Contributions:** All authors: writing—review and editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

