Gallbladder Cancer: Current Insights in Genetic Alterations and Their Possible Therapeutic Implications

Simple Summary Knowledge of genetic alterations in gallbladder cancer (GBC) continues to increase. This systematic review provides an overview of frequently occurring genetic alterations in GBC and describes their possible therapeutic implications. We detected three frequently (>5%) altered genes (ATM, ERBB2 and PIK3CA) for which targeted therapies are available in other cancer types. For solid cancers with microsatellite instability or a high tumor mutational burden pembrolizumab is FDA-approved. Altogether, these five biomarkers might be used in future molecular panels to enable precision medicine for patients with GBC. We found only nine clinical trials evaluating targeted therapies in GBC directed at frequently altered genes (ERBB2, ARID1A, ATM and KRAS). This underlines the challenges to perform such clinical trials in this rare, heterogeneous cancer type and emphasizes the need for multicenter clinical trials. Abstract Due to the fast progression in molecular technologies such as next-generation sequencing, knowledge of genetic alterations in gallbladder cancer (GBC) increases. This systematic review provides an overview of frequently occurring genetic alterations occurring in GBC and their possible therapeutic implications. A literature search was performed utilizing PubMed, EMBASE, Cochrane Library, and Web of Science. Only studies reporting genetic alterations in human GBC were included. In total, data were extracted from 62 articles, describing a total of 3893 GBC samples. Frequently detected genetic alterations (>5% in >5 samples across all studies) in GBC for which targeted therapies are available in other cancer types included mutations in ATM, ERBB2, and PIK3CA, and ERBB2 amplifications. High tumor mutational burden (TMB-H) and microsatellite instability (MSI-H) were infrequently observed in GBC (1.7% and 3.5%, respectively). For solid cancers with TMB-H or MSI-H pembrolizumab is FDA-approved and shows an objective response rates of 50% for TMB-H GBC and 41% for MSI-H biliary tract cancer. Only nine clinical trials evaluated targeted therapies in GBC directed at frequently altered genes (ERBB2, ARID1A, ATM, and KRAS). This underlines the challenges to perform such clinical trials in this rare, heterogeneous cancer type and emphasizes the need for multicenter clinical trials.


Introduction
Gallbladder cancer (GBC) is a relatively uncommon malignancy, with a worldwide incidence of less than 2 per 100,000 [1]. However, GBC shows a broad geographical and ethnic distribution, with low incidence rates in developed countries and higher incidence rates in South American countries, India, Pakistan, Japan and Korea. Higher incidence rates are also observed among Mexican Americans, Indian Americas, and Eastern Europeans [1][2][3]. Well-established risk factors include age, obesity, female gender, family history, cholelithiasis, and anomalous junction of the pancreatobiliary duct [4][5][6].
At present, radical resection by cholecystectomy with lymphadenectomy of the hepatoduodenal ligament and wedge or segment resection of the liver is the only treatment with curative intent [7]. Unfortunately, most patients present at an advanced stage and are unresectable [8,9]. Even after resection, the five-year survival rate ranges from 18% to 34% [10,11]. Palliative systemic chemotherapy for patients with GBC has shown limited efficacy. In patients with locally advanced or metastatic biliary tract cancer (BTC), including GBC, the ABC-02 trial reported a median overall survival of 11.7 months in the BTC group treated with gemcitabine plus cisplatin and 8.1 months in the BTC group treated with gemcitabine monotherapy [12].
With the rapid developments in next-generation sequencing (NGS), our knowledge of genetic alterations occurring in GBC has increased over the past decade [13]. Consequently, the number of preclinical studies and clinical trials evaluating therapies targeting these genetic alterations is slowly increasing [14]. Several reviews on genetic alterations in biliary tract cancer have been published. However, a systematic review specifically focusing on genetic alterations in GBC and their therapeutic implications is lacking.

Literature Search
The protocol for this systematic review was prospectively registered in the PROSPERO registry (CRD42021265246). The study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [15].
A literature search was performed until 26 March 2021 in PubMed, EMBASE, Cochrane Library, and Web of Science. Keywords or medical subject headings (MeSH) used for the search are provided in Table A1. Only full-text articles published in English in 2000 or afterwards were selected.

Study Selection
Citations were deduplicated by using tools in Endnote and Rayyan (rayyan.ai). Subsequently, all duplications were manually verified. Titles and abstracts of all retrieved records were then independently screened for eligibility by two investigators (H.K. and T.J.J.d.B.) using the Rayyan platform for systematic reviews. Discordance was re-evaluated by both investigators and resolved by consensus after discussion. Full papers were obtained for records that were considered potentially eligible by both investigators. In the case of studies with overlapping data, the article reporting the largest cohort was selected.
Only studies were included that identified somatic genetic alterations in human GBC with Polymerase Chain Reaction (PCR) with Sanger sequencing, NGS (targeted, whole-exome, and whole-genome sequencing), MassArray, or SNaPshot. Studies using immunohistochemistry or FISH techniques were excluded. Case reports, retracted articles, and preliminary results were excluded, as well as studies focusing on gene expression, genetic alterations in cell lines, cell-free DNA from serum, and mitochondrial DNA.
This review focused on adenocarcinoma of the gallbladder. Other histologies, i.e., adenosquamous, squamous, neuroendocrine, and sarcomatoid carcinoma, were excluded. In case no histological type was reported, tumors were assumed to represent adenocarcinomas since only 5% of GBC have histology other than adenocarcinoma [16,17].

Data Extraction
Data extracted included first author, year of publication, study population, sequencing technique, number of samples, and reported genetic alterations and their frequencies.
Included alterations were non-synonymous mutations and copy number aberrations (DNA amplifications and deletions). In addition, frequencies of high tumor mutational burden (TMB) and the presence of microsatellite instability (MSI) were extracted.
Only frequently occurring genetic alterations (>5% [18,19] across all studies and in >5 of all included GBC samples) were included. Per genetic alteration, weighted average of reported frequencies was calculated by using the number of samples analyzed in a study as the weight.
A column scatter plot was constructed for all frequently occurring genetic alterations. Bars displayed the minimum and maximum reported frequency. Dots represented the frequency per study. Diamonds represented the weighted averages.
No risk of bias was assessed since no methods are available that could assess confounders such as inter-population diversity and variations in DNA techniques [20,21]. Nevertheless, to facilitate the reader to assess the studies' quality and risk of bias, details of each included study were displayed in tables.

Therapeutic Implications
Actionable alterations were identified by comparing the frequently altered genes detected in this study with the actionable genes of the OncoKB database (accessed 21 July 2021) [22]. Only actionable genes with level 1 and 2 therapeutic evidence in solid tumors were analyzed. Level 1 evidence was defined as a U.S. Food and Drug Administration (FDA)-recognized biomarker predictive of response to FDA-approved drugs for a specific indication and level 2 as a standard care biomarker predictive of response to FDA-approved drugs for another indication. Data extracted included targetable alteration, drug, level of evidence, and cancer type.
Second, clinicaltrials.gov and clinicaltrialsregister.eu were examined for clinical trials targeting frequently occurring genetic alterations detected in this study in patients with GBC or BTC (accessed 26 July 2021).

Literature Search and Study Selection
In total, 4324 records were retrieved from all databases. After removing duplicate records, 2159 records were screened for eligibility based on title and abstract ( Figure 1). A total of 92 articles underwent full-text reading. Out of 92 articles, 30 were excluded due to: no data for adenocarcinoma of the gallbladder available (N = 12), overlapping cohorts (N = 8), the wrong type of article (N = 5), no genetic alteration frequency data available (N = 3), wrong study population (N = 1), and not written in English (N = 1). A total of 62 articles were included for final data extraction, describing a total of 3893 GBC samples from individual patients [18,19,.  Table A2 shows the results for all frequent gene mutations. TP53 was the most frequently mutated gene in GBC (weighted average 57%), followed by SHH (weighted average 20%), ELF3 (weighted average 19%), ARID1A (weighted average 14%), and SMAD4 (weighted average 13%). Cancers 2021, 13, x FOR PEER REVIEW 4 of 22  Figure 2A represents all frequently occurring gene mutations in GBC. Appendix Table A2 shows the results for all frequent gene mutations. TP53 was the most frequently mutated gene in GBC (weighted average 57%), followed by SHH (weighted average 20%), ELF3 (weighted average 19%), ARID1A (weighted average 14%), and SMAD4 (weighted average 13%).

Microsatellite Instability (MSI)
Table 3 outlines all studies (N = 13) reporting on MSI status in 1162 patients with GBC. This table shows that a high diversity of marker panels and MSI definitions were used. A broad range of MSI-high frequencies were reported, although most studies reported a low incidence. The weighted average for MSI-high frequency was 3.5%.   3.6. Possible Therapeutic Implications 3.6.1. Targetable Alterations in Other Malignancies Currently, no FDA-approved therapies targeting gene mutations or copy number aberrations are available for GBC. However, several targeted therapies directed at actionable alterations that were also frequently observed in GBC are FDA-approved in other cancers (Table 4). These alterations include mutations in ATM, PIK3CA and ERBB2, and amplifications in ERBB2. For solid cancers with MSI-H or TMB-H, pembrolizumab is approved by the FDA.   Most studies evaluate targeted therapies directed at ERBB2 alterations and related signal pathway components. One trial assesses the safety and efficacy of ELI-002 immunotherapy, a novel amphiphile therapeutic vaccine targeting KRAS-driven cancers, for patients with KRAS or NRAS mutations (G12D or G12R) in various solid tumor types including GBC (NCT04853017). Moreover, a phase 2 trial analyzes the efficacy and toxicity of olaparib in patients with metastatic BTC with an DNA repair gene mutation, including ARID1A and ATM (NCT04042831). The NCT02091141 trial evaluates six treatment regimens based on molecular testing in patients with advanced solid tumors, including BTC. For example, patients with tumors demonstrating elevated TMB (≥10 mutations/Mb) will receive atezolizumab.

Discussion
In this systematic review of 62 articles assessing genetic alterations in GBC, 3893 GBC samples were analyzed. Frequently occurring genetic alterations included mutations in 17 genes and amplifications/deletions in nine genes. Since no targeted therapy is currently available for GBC, frequent genetic alterations detected in this study were matched to actionable genetic alterations in other solid tumors. Only mutations in ATM, ERBB2 and PIK3CA and amplifications in ERBB2 are currently targetable with FDA-approved drugs in other solid tumors. Therefore, these alterations are potential targets in GBC and might be included in future molecular testing panels for personalized treatment decisions.
In GBC, the prevalence of TMB-H was low. Previous studies have shown that TMB-H is correlated with clinical response to immunotherapy in several tumors, including BTC, and that patients with TMB-H tumors can benefit from immune checkpoint inhibitors such as pembrolizumab [108][109][110]. In patients with GBC receiving immunotherapy, a higher objective response rate was observed for TMB-high tumors compared to TMB-low tumors (50% vs. 25%) [47]. Pembrolizumab has been FDA-approved for patients with TMB-H advanced solid cancers.
Another biomarker associated with response to immune checkpoint inhibitors is MSI, which was observed in 3.5% of patients in the present study. For patients with MSI-H cancers, pembrolizumab is also approved by the FDA, as well as by the Pharmaceuticals and Medical Devices Agency (PMDA, Japan). The European Medicines Agency (EMA, European Union) and the National Medical Products Administration (NMPA, China) have not yet given approval for this agent in patients with MSI-H (or TMB-H). In patients with cholangiocarcinoma who received pembrolizumab, an objective response rate of 41% was seen [111]. Therefore, although only beneficial in a small proportion of patients, both TMB-H and MSI-H are interesting biomarkers in precision medicine for GBC.
A broad range was observed in genetic alteration frequencies in GBC across all studies. In part, this could be attributed to differences in sequencing technologies and assessed gene regions. For example, NGS has a slightly lower sensitivity for mutation detection compared to some other technologies like real-time PCR testing, although detecting lowfrequency genetic alterations through (ultra-) deep sequencing with a higher coverage can partly overcome this issue [112][113][114]. Moreover, geographical differences in mutation patterns might be present, as previous studies have shown [35,46,73]. However, no one-to-one regional comparisons of genetic alteration frequencies could be made due to different sequencing techniques used throughout all studies. Substantial differences were also observed regarding the frequency of MSI. A large variation in marker panels and MSI definitions, and some missing MSI definitions, complicated comparison of all MSI frequencies. Although a weighted average of MSI frequency was calculated to minimalize the influence of sample size on the average frequency, our results should be taken with caution due to the heterogeneity in MSI panels and definitions. Usage of the National Cancer Institute (NCI) panel of microsatellites and definitions in future research would facilitate MSI comparison among studies [115].
Since some genetic alterations might impede the efficacy of therapies targeting other genetic alterations, co-existing alterations might be an important determinant for therapeu-tic sensitivity and resistance. For example, patients with solid stage IV cancer types harboring amplification in the MDM2 family or EGFR aberrations who received anti-PD1/PDL1 immunotherapy showed increased tumor growth [116]. Unfortunately, co-existing alterations could not be assessed in this study due to differences in sequencing techniques and missing data in some studies.
Only nine clinical trials which evaluated targeted therapies directed at frequently occurring genetic alterations in GBC were identified. Though we assessed only trials targeting these genetic alterations specifically, these limited results underline that GBC remains a relatively under-investigated cancer type regarding therapeutic agents. Unfortunately, initiation of clinical trials including patients with GBC is logistically challenging since GBC is a rare cancer type in most Western countries. The large inter-tumor heterogeneity of GBC poses another challenge for targeted therapies. However, basket trials, a clinical trial that tests agents in different cancer types with the same genetic alteration, could provide a solution to this challenge in future studies.
This study has several limitations. First, histological types and subtypes of GBC samples could not always be retrieved. It is important to report histological (sub)types to be able to determine whether differences in genetic alterations are associated with different histology. Second, a publication bias might exist. Therefore, actual overall alteration frequencies might be over-or underestimated.

Conclusions
In conclusion, GBC has a diverse mutational landscape. Frequently occurring genetic alterations in GBC that are actionable in other solid tumors are rare, including mutations in ATM, ERBB2, PIK3CA and amplifications in ERBB2. For all solid tumors with MSI-high or TMB-high, including GBC, the immune checkpoint inhibitor pembrolizumab is FDAapproved. Few clinical trials targeting the frequently altered genes in GBC are performed, emphasizing the need for multicenter clinical basket trials.

Conflicts of Interest:
The authors declare no conflict of interest.      WA: weighted average (calculated by using the number of samples analyzed in a study as the weight); N.A.: not available; AC: adenocarcinoma; NGS: next-generation sequencing; PCR: polymerase chain reaction; Seq: sequencing; DS: direct sequencing; WES: whole-exome sequencing; WGS: whole-genome sequencing; SSCP: single-strand conformation polymorphism; RFLP: restriction fragment length polymorphism; TS: targeted sequencing.