Brain Metastases from Biliary Tract Cancers: A Case Series and Review of the Literature in the Genomic Era

Abstract

Background

Biliary tract cancers (BTCs) are highly fatal malignancies that make up less than 1% of all cancers. BTC is often diagnosed at an unresectable stage; surgical resection remains the only definitive treatment. Brain metastases (BMs) from BTC are extremely rare, and few studies on patients with BMs from BTC exist. The aim of this study was to identify clinical characteristics associated with poor prognosis for patients with BMs from BTC.

Materials and Methods

We performed a retrospective review of electronic medical records for patients with BMs from BTC managed at Mount Sinai Hospital from 2000 to 2017. Data on patient characteristics, magnetic resonance imaging findings, treatment regimens, and clinical outcomes were analyzed.

Results

We identified 1,910 patients with BTC. Nine patients developed BMs, with an incidence of 0.47%. Of these nine patients, six had intrahepatic cholangiocarcinoma, two had extrahepatic cholangiocarcinoma, and one had gallbladder cancer. Six (66.7%) patients had one BM, one (11.1%) patient had two BMs, and two (22.2%) patients had three or more BMs. Four (44.4%) patients underwent BM resection, and seven (77.8%) received BM radiation. Median overall survival from time of BM diagnosis was 3.8 months (95% confidence interval 0.1–16.9).

Conclusion

Development of BMs from BTC is rare; however, prognosis is less than 4 months. BM diagnosis can occur within 2 years of primary diagnosis. As targeted therapeutics emerge, future studies ought to focus on identifying genomic BM markers associated with BTC subtypes.

Implications for Practice

In the largest retrospective study of biliary tract cancer brain metastases, the clinical presentation and outcomes are reported of nine patients with an extremely rare clinical entity. The genomic literature and potential therapeutic targets for these patients with limited treatment options is comprehensively and exhaustively discussed.

Introduction

Biliary tract cancers (BTCs) are highly fatal malignancies that make up less than 1% of all cancers globally [1]. These cancers are generally of the adenocarcinoma histologic subtype and include extrahepatic and intrahepatic cholangiocarcinoma (CC), gallbladder cancer (GBC), and cancer of the ampulla and papilla of Vater [2, 3]. The incidence of BTC varies widely by region of the world, with the lowest incidence in Western countries and the highest in Asia and Latin America [4].

Owing to nonspecific presenting symptoms, BTC is often diagnosed at an advanced, unresectable stage; radical surgical resection remains the only definitive treatment option for BTC and portends the best prognosis [5]. Thus, prognosis for BTC is poor, with 5‐year survival rates of 10% for CC and 5% for GBC [6, 7]. Chemotherapy can improve survival in patients with advanced BTC and is the current standard of care for patients with unresectable disease [3]. First‐line chemotherapy is gemcitabine combined with a platinum agent. Other regimens are not well‐established despite many clinical trials for new chemotherapeutic regimens and targeted therapies [1, 8].

BTC commonly invades adjacent lymph nodes and the liver and less frequently metastasizes to distant sites including lungs and bone. Metastasis to the brain is extremely rare. The incidence of brain metastases (BMs) from BTC is largely unknown. Only a few case reports and two case series have characterized BMs from BTC [9–14]. The most recent case series reports the incidence of BMs from BTC to be 1.4% [10].

As such, there is an unmet need to understand clinical outcomes of patients with BTC who develop BMs. Our study aims to expand the existing literature by further describing patient outcomes for those who develop BMs from BTC within the context of the genomic era.

Materials and Methods

Patient Cohort

We performed a retrospective review of electronic medical records for patients with BMs from BTC managed at Mount Sinai Hospital from 2000 to 2017. Inclusion criteria were (a) presence of primary CC or GBC confirmed by a board‐certified pathologist and (b) presence of at least one metastatic brain lesion confirmed by computed tomography (CT) or magnetic resonance (MR) imaging. Nine patients met these criteria and were included for analysis. Data on patient characteristics, treatment regimens, and clinical outcomes were extracted and are presented in Table 1. The study was reviewed and approved by the Mount Sinai Institutional Review Board and was in compliance with Health Insurance Portability and Accountability Act guidelines.

Imaging Analysis

Axial T1 postcontrast gadolinium‐enhanced MR images were reviewed using OsiriX software. Tumor cross‐sectional area (CSA) was calculated for each case by drawing perpendicular lines on the axial slice that visualized the most prominent tumor size. Presence of vasogenic edema was assessed using axial T2 fluid‐attenuated inversion recovery (FLAIR) images. Contrast‐enhanced CT images were analyzed in cases where an MR was not available. For patients with more than one BM, CSA was calculated for the largest lesion.

Statistical Analysis

Statistical analysis was conducted using statistical database software (JMP version Pro 14, SAS Institute Inc, Cary, NC). Overall survival was calculated as time from BM imaging diagnosis to date of death or last follow‐up using Kaplan‐Meier and chi‐square proportional hazards methodologies; p ≤ .05 was used to denote significance.

Results

There were 1,910 patients with BTC treated at our institution between 2000 and 2017. Of these, nine patients (0.47%) developed BMs (Fig. 1). Six patients (66.7%) were female and three were male (33.3%). Smoking status was reported as never smoker (33.3%), former smoker (33.3%), current smoker (22.2%), and unknown (11.1%). The most common medical comorbidities included cardiac disease (55.6%), type 2 diabetes (33.3%), pulmonary disease (11.1%), and cirrhosis (11.1%). One patient had systemic lupus erythematosus.

Primary tumor diagnosis included six patients (66.7%) with intrahepatic CC (ICC), two (22.2%) with extrahepatic CC (ECC), and one (11.1%) with GBC. Median age at primary diagnosis was 63.7 years (range, 47.4–68.1 years). Six patients (66.7%) underwent resection of their primary disease site, with five of six patients (83.3%) having negative margins upon histological evaluation. Additionally, seven patients (77.8%) received chemotherapy and three (33.3%) received radiation prior to BM diagnosis.

Metastatic sites prior to BM diagnosis included lung in 66.7% as well as liver, chest wall, and bone. All patients underwent an abdominal/pelvic CT prior to BM diagnosis to assess for the presence of vertebral metastases. Median time from abdominal/pelvic CT to BM diagnosis was 46 days (95% confidence interval [CI] 4–113). One (11.1%) patient had vertebral metastases with sclerotic lesions present in the pelvic and iliac bones, sacrum, and L4 vertebral body.

The most common presenting symptoms at BM presentation included seizure (33.3%), dizziness (33.3%), altered mental status (33.3%), weakness (33.3%), headache (22.2%), and cranial neuropathy (11.1%). Median age at BM diagnosis was 63.7 years (range, 48.8–72.4 years). Median time from primary diagnosis to BM diagnosis was 16.7 months (range, 0.7–66.7 months).

Six patients (66.7%) had one BM, one patient (11.1%) had two BMs, and two patients (22.2%) had three or more BMs upon imaging evaluation at initial presentation. Five (55.6%) patients had a supratentorial lesion, one patient (11.1%) had an infratentorial lesion (Fig. 2), and the remaining three (33.3%) patients had both supratentorial and infratentorial lesions. The most common parenchymal lobe with metastatic disease was frontal (n = 5, 55.6%), followed by parietal (n = 4, 44.4%), cerebellar (n = 4, 44.4%), temporal (n = 2, 22.2%), and occipital (n = 1, 11.1%). Hydrocephalus was not appreciated on any scan. Vasogenic edema was visualized on T2 FLAIR sequence in all cases. Mean tumor CSA was 6.9 cm² ± 7.1 (Fig. 1).

Four (44.4%) patients underwent definitive surgical resection, and seven (77.8%) received BM radiation. Radiation modalities included stereotactic radiosurgery (SRS) in two patients, whole‐brain radiation therapy (WBRT) in three patients, and both SRS and WBRT in three patients. Median overall survival from time of BM diagnosis was 3.8 months (95% CI 0.1–16.9). Patients who received BM radiotherapy had significantly improved overall survival of 10.2 months (95% CI 3.8–16.9) compared with 0.8 months (95% CI 0.1–1.5; p = .0039, log‐rank; RR = 0.38; p = .01, chi‐square).

Discussion

Case Series

Limited studies have evaluated brain metastases from biliary tract cancers. Beyond the few case reports of patients with BMs from BTC, two case series of eight [9] patients from Thailand and six [10] patients from Italy have been reported in the literature. In our series of nine patients, the median age at BM diagnosis was 63.7 years, which is similar to that reported in the two case series of 60 years and 68 years, respectively. The incidence of BM in our cohort was 0.47%, compared with 0.15% and 1.4%, respectively. Additionally, we report a median survival of 3.8 months for patients with BMs from BTC, which is similar to 9.5 weeks and 3.7 months, respectively. When compared with BM from other primary cancers, median survival for BTC is similar to the survival seen in non‐small cell lung cancer (4 months) and head/neck cancer (5 months) and shorter than the survival seen in melanoma (6 months) and breast cancer (12 months) [15].

In our cohort, median time from primary diagnosis to BM diagnosis was 16.7 months, compared with 8 months [9] and 13.6 months [10], respectively. The longer time to BM diagnosis in our cohort compared with that seen in the other two case series may reflect differences in clinical outcomes due to wide genomic variability seen in BTC and varying responses to treatment [16, 17].

We identified a significant survival difference between patients who received BM radiotherapy (10.2 months) compared with those who did not receive BM radiotherapy (0.8 months); however, we acknowledge the limitation that only two patients did not receive radiation. Moreover, although there was no statistical difference between patients who received neurosurgical resection compared with those who did not (data not shown), given the underpowered sample, we caution against clinical interpretation of these findings. Indeed, further institutional case series and a meta‐analysis of previously reported cases are warranted to identify prognostic factors that are associated with survival in patients with BMs from BTC.

Genomics of BTC

Treatment of BTC has traditionally relied on gemcitabine‐based chemotherapy. Recently, genomic characterization of patients with BTC has led to the identification of clinically actionable mutations and subsequent development of targeted therapies. Early efforts to characterize the genomics of BTC relied on genotyping of hotspot oncogenes, leading to the identification of recurrent mutations in known oncogenic pathways, such as the epidermal growth factor receptor (EGFR) pathway, the mitogen‐activated protein kinase (MAPK) pathway, and the phosphoinositide‐3‐kinase (PI3K) pathway [18].

Genomic differences between subtypes of BTC have been identified using next‐generation sequencing. For example, mutations in isocitrate dehydrogenase 1 or 2 (IDH1/2) and FGFR2 gene fusions are more likely to be seen in ICC; KRAS and human epidermal growth receptor 2 (HER2) mutations are more likely to be seen in ECC; mutations in EGFR, HER2/neu, and PIK3CA are more likely to be seen in GBC [19].

Furthermore, within ICC, genomic differences have been reported based on ICC etiology. Whole‐exome sequencing of 15 CC cases revealed that non–liver‐fluke ICC are more likely to harbor BAP1 and IDH mutations, whereas liver‐fluke ICC are more likely to harbor TP53 and SMAD4 mutations [20]. Of note, 30% of ICC have FGFR, IDH, or BRAF mutations, all of which have targeted therapies currently in clinical trials [18].

Genomics of Brain Metastases from BTC

To our knowledge, there have been no efforts to characterize the genomic landscape of BMs from BTC, likely because BMs from BTC is such a rare event. Furthermore, patients with BMs are often ineligible from inclusion in clinical trials, which is a major reason why there is a lack of information on how to treat patients with BMs. Given the dismal prognosis of patients with BMs from BTC, it is important to investigate the genomic events that support metastasis to the brain so that more advanced therapeutic options may be elucidated.

Because BMs from BTC have not yet been genomically characterized, the importance of targeted therapies in the treatment of BMs from BTC remains unclear. However, there is evidence to support targeting of known genomic alterations in patients with BMs from other primary cancers, including EGFR inhibitors in lung cancer [21–25], PI3K inhibitors in BMs from breast cancer [26], BRAF inhibitors in BMs from melanoma [27], and vascular endothelial growth factor receptor tyrosine kinase inhibitors in BMs from renal cell carcinoma [28]. The role of targeting genomic alterations in patients with BMs is currently under study (Alliance A071701; NCT03994796) in a national cooperative group phase II clinical trial.

One limitation in treatment of BMs is poor penetration of systemic therapies through the blood–brain barrier (BBB), due to factors that include low passive permeability and efflux pumps. However, advanced cancer may alter BBB permeability, thereby supporting susceptibility to certain targeted agents [29]. Furthermore, BBB permeability is heterogeneous and associated with vascular remodeling [30]. Therefore, more work is needed to further explore the ability of targeted agents to penetrate the BBB and elucidate their potential role in treating BMs from BTC.

Clinical Trials in BTC

A number of clinical trials targeting mutations in BTC have been initiated in recent years. These trials and the mutations that they target are summarized below. Many of the trials focus on the treatment of primary BTC and often do not include patients with BM.

MAPK Pathway

Efforts to target the MAPK pathway in patients with BTC have shown mixed results. A phase II study of the MEK inhibitor selumetinib in patients with advanced BTC reported an 80% disease control rate and a median overall survival of 9.8 months for patients receiving selumetinib as first‐line therapy [31]. More recently, in a phase Ib trial of selumetinib (ABC‐04) in combination with gemcitabine and cisplatin, selumetinib did not show clinical benefit in the treatment of advanced BTC [32]. Trials of the multikinase inhibitor sorafenib have shown little success as well [33–36]. Selumetinib may have the potential to penetrate the BBB given efficacy in the treatment of pediatric low‐grade glioma [37].

FGFR2 Fusions

Fusions have been observed between the receptor tyrosine kinase FGFR2 and other genes including AHCYL1, BICC1, PARK2, MGEA5, TACC3, KIAA1598, KCTD1, and TXLNA [38–41]. A phase II trial of the FGFR inhibitor BGJ398 showed efficacy in controlling disease in patients with advanced or metastatic cholangiocarcinoma harboring FGFR2 fusions or other alterations [42]. Furthermore, a phase II trial of the nonselective FGFR inhibitor ponatinib is ongoing (NCT02272998). Although there are no data regarding BGJ398 or ponatinib and the BBB, the FGFR inhibitor dovitinib is able to cross the BBB with limited efficacy in the treatment of glioblastoma [43].

IDH

There are multiple clinical trials ongoing that target mutations in IDH1/2. Most notably, the IDH1 inhibitor AG‐120 has shown efficacy in stabilizing disease progression in patients with IDH1‐mutant ICC [44], and a phase III trial for patients with IDH1‐mutant cholangiocarcinoma is currently underway (NCT02989857). Certain patients with IDH‐mutant ICC are sensitive to the multikinase inhibitor dasatinib [45]. Finally, the IDH inhibitors BAY1436032 (NCT02746081), IDH305 (NCT02381886), and AG‐881 (NCT02481154) are currently in phase I testing. Emerging preclinical data support the use of the IDH1 inhibitor BAY 1436032 for the treatment of IDH1 mutant astrocytoma in vivo [46]. Clinical studies are underway (NCT03030066, NCT02481154).

HER2/neu

An estimated 15% of GBC have HER2/neu amplification and overexpression [47], which can be targeted with HER2 antagonists including trastuzumab, afatinib, and lapatinib; however, early trials of HER2 antagonists combined with chemotherapy to treat advanced BTC were not impressive [48–50]. A phase II trial of trastuzumab for a select group of patients with HER2‐positive BTC is ongoing (NCT02999672). With regard to the BBB, trastuzumab has been shown to cross the BBB [51].

PIK3CA

The PI3K/AKT/mTOR pathway is a target in many cancers, but data are limited on the role of PIK3CA inhibitors in the treatment of BTC. One study of the mTOR inhibitor everolimus in 27 patients with advanced BTC demonstrated a median overall survival of 9.5 months [52]. Further studies are needed to elucidate the role of PI3K pathway inhibitors in patients with BTC. Importantly, PI3K and mTOR inhibitors have shown good BBB penetration [53].

Immunotherapy for BTC

Recently, great advances have occurred in the field of immunotherapy with the rise of checkpoint inhibitors. Checkpoint inhibitors have efficacy in the treatment of untreated or progressive brain metastases from patients with melanoma and non‐small cell lung cancer [54]. However, few studies have evaluated the role of immunotherapy in patients with BTC.

In one study, a unique whole‐exome sequencing approach was used to determine that tumor‐infiltrating lymphocytes (TILs) from a patient with metastatic cholangiocarcinoma contained CD4+ Helper 1 (TH1) cells that recognized a mutation in erbb2 interacting protein (ERBB2IP). Tumor regression was achieved by using adoptive transfer of TILs containing mutation‐specific polyfunctional TH1 cells. The patient was re‐treated with mutation‐reactive TH1 cells upon disease progression, and tumor regression was again seen [55]. Additionally, in a recent study of pembrolizumab in pretreated BTC with programmed death‐ligand 1 expression, the objective response rate was 17% [56].

Conclusion

Development of BMs from BTC is an extremely rare event associated with dismal prognosis, and as such, few case series of patients with BMs from BTC exist. Future prospective cohort studies are necessary to fully understand prognostic factors that may be associated with survival in patients with BMs from BTC as well as to drive clinical treatment paradigms. As our understanding of the underlying genomics of BM improves, clinical trials that target actionable mutations in patients with BM are encouraged.

Acknowledgments

P.K.B. receives grant funding from Damon Runyon Cancer Research Foundation, Breast Cancer Research Foundation, Susan G. Komen, and NIH.

Author Contributions

Conception/design: Megan R. D'Andrea, Corey M. Gill, Priscilla K. Brastianos, Raj K. Shrivastava

Provision of study material or patients: Megan R. D'Andrea, Corey M. Gill, Joshua B. Bederson, Raj K. Shrivastava

Collection and/or assembly of data: Megan R. D'Andrea, Corey M. Gill, Melissa Umphlett, Nadejda M. Tsankova, Mary Fowkes, Joshua B. Bederson, Raj K. Shrivastava

Data analysis and interpretation: Megan R. D'Andrea, Corey M. Gill, Priscilla K. Brastianos, Raj K. Shrivastava

Manuscript writing: Megan R. D'Andrea, Corey M. Gill, Priscilla K. Brastianos, Raj K. Shrivastava

Final approval of manuscript: Megan R. D'Andrea, Corey M. Gill, Melissa Umphlett, Nadejda M. Tsankova, Mary Fowkes, Joshua B. Bederson, Priscilla K. Brastianos, Raj K. Shrivastava

Disclosures

Priscilla K. Brastianos: Eli Lilly and Company, Tesaro, Angiochem, Merck (C/A), Merck, Pfizer (RF), Genentech‐Roche (H). The other authors indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board

References

Author notes