A phase I trial investigating pulsatile erlotinib in combination with gemcitabine and oxaliplatin in advanced biliary tract cancers

Approximately 7500 new cases of advanced biliary tract cancer (ABTC) will be diagnosed each year in the United States [1]. These are generally divided into intrahepatic cholangiocarcinoma, extrahepatic cholangiocarcinoma, and cancer of the gallbladder. In addition, ampulla of Vater cancer is variably included with biliary cancers. These cancers frequently present in a stage too advanced for surgical resection, and a majority of the patients with operable disease will have recurrence after complete resection. Several chemotherapy agents have been evaluated in biliary tract cancer patients, with a gemcitabine backbone emerging as a standard [2]. A phase II trial of single-agent gemcitabine observed an objective response rate of 22 % and a disease control rate of 50 % [3]. In a phase II study of first-line therapy, the combination of gemcitabine and oxaliplatin (GEMOX) was well tolerated and resulted in an objective response rate of 35 % and stable disease in 26 % [4]. Despite the modest success with gemcitabine combinations, biliary cancers remain among the deadliest malignancies [5, 6]. Therefore, the development of new therapeutic regimens to improve treatment efficacy in this patient population is warranted.

The epidermal growth factor receptor (EGFR) is part of a complex series of cellular signaling pathways that lead to increased cell proliferation, motility, survival, and angiogenesis, all of which can lead to tumor growth and progression [7]. EGFR expression is increased in a majority of bile duct cancers along with one of its ligands, TGF-alpha [8]. Additionally, HER2, which can dimerize with EGFR thus activating downstream signaling pathways, is overexpressed (immunohistochemical staining ≥2) in 15–43 % of biliary cancers [9‐11]. Thus, there has been hope for the development of EGFR-directed therapies for the treatment of ABTC [12‐14]. Erlotinib, an orally active tyrosine kinase inhibitor (TKI) of EGFR, is currently FDA approved for use in non-small cell lung cancer, as well as in pancreatic cancer when administered in combination with gemcitabine. In a study of 42 patients with either unresectable or metastatic biliary cancer being treated with daily oral doses of erlotinib, 7 % achieved an overall confirmed response, 43 % achieved stable disease, and the median OS was 7.5 months [14].

Preclinical data from mouse xenograft models suggested that pulsed gefitinib (another EGFR TKI) before paclitaxel caused significantly more tumor regression than continuous gefitinib dosing in combination with paclitaxel [15]. When pancreas, gastric, and colon cancer cell lines were treated with combinations of gemcitabine with flavopiridol, an agent that induces G1/S cell cycle arrest, maximal antitumor effect was observed with the combination of gemcitabine followed by flavopiridol, whereas the reverse sequence showed no synergy [16]. As erlotinib also induces G1/S arrest, we hypothesized that erlotinib may exhibit sequence-specific synergy with gemcitabine in a similar manner, such that pulsatile dosing after chemotherapy with adequate washout prior to repeat dosing would achieve improved efficacy. Therefore, our primary objective in this phase Ib study was to determine the maximum tolerated dose (MTD) and recommended phase II dose (RP2D) of pulsatile erlotinib in combination with GEMOX. A secondary objective was to describe any anti-tumor activity associated with treatment in patients with ABTC. Exploratory objectives included an investigation of the relationship between clinical response and the expression levels of the cytoskeleton protein vimentin and the cell-cell adhesion protein E-cadherin. Furthermore, given the evidence that patients with KRAS mutations have a poorer response to EGFR-directed therapies [17, 18] and that biliary tract cancers have a spectrum of mutations in EGFR and its downstream signaling pathways, including KRAS and PIK3CA [19‐23], we evaluated the potential relationship between mutational status and clinical outcome.

Therapeutic regimens combining chemotherapy with EGFR-targeted agents are an active area of research. Despite the high level of EGFR expression in solid tumors, combining EGFR inhibitors with cytotoxic chemotherapy has not been exceptionally successful in large-scale clinical trials. For example, in the TALENT lung cancer trial, erlotinib was added to gemcitabine and cisplatin without an improvement in OS [25]. Evaluation of 274 patient tumors from the TRIBUTE trial, in which patients with non-small cell lung cancer were treated with carboplatin and paclitaxel with or without erlotinib, showed a trend towards an erlotinib benefit on time to progression but not improved survival [26]. More recently, the BINGO trial in ABTC evaluated GEMOX with and without cetuximab but did not observe an improvement in efficacy [27]. Based on the preclinical evidence suggesting that EGFR tyrosine kinase inhibitors result in a cell cycle arrest rather than inducing apoptosis [28], it is reasonable to hypothesize that antagonist effects with a cytotoxic agent such as chemotherapy may occur. However, based on the preclinical observation that a pulsatile regimen of EGFR-targeted therapy in combination with chemotherapy was more effective [15], we designed the current study to test the hypothesis that erlotinib administered at periodic time points during standard chemotherapy would be a synergistic treatment strategy for patients with ABTC.

The primary objective of this study was to determine the MTD of GEMOX administered in combination with erlotinib. The MTD was determined (and verified in the dose expansion cohort) to be 800 mg/m² GEM, 85 mg/m² OX, and 150 mg erlotinib. When the current study was being developed, we were unaware of the concurrent phase III trial in South Korea that was investigating continuous erlotinib in combination with GEMOX compared to GEMOX alone in ABTC patients [13]. However, despite the slightly different treatment regimens (phase III study: 100 mg erlotinib administered daily, 1000 mg/m² GEM, and 100 mg/m² OX), the toxicity profiles were similar between the two studies suggesting that erlotinib in combination with chemotherapy is well tolerated in patients with ABTC.

It is well known that skin rash is a major side effect of erlotinib, and perhaps pulsatile dosing reduces the incidence of erlotinib-associated skin rash. A meta-analysis of 2911 patients with a variety of solid tumors observed that skin rash associated with single agent erlotinib (150 mg daily) occurred in 75 % of patients [29]. A phase II study that investigated daily erlotinib (150 mg) in combination with bevacizumab (5 mg/kg; Day 1 and 15 of a 28-day cycle) in ABTC reported that 76 % of patients experienced erlotinib-associated skin rash [30]. The total incidence of skin rash in our study was much lower at 57 %, and only 29 % of patients treated at 150 mg developed erlotinib-associated skin rash. Thus, pulsatile dosing (Day 3–8 of a 14-day cycle) of erlotinib could be an alternative treatment strategy to reduce the incidence of rash. Although interval dosing of a TKI, such as erlotinib, is attractive from the context of decreasing the frequency of specific adverse events, the possibility of the “disease flare” phenomenon that occurs after stopping a TKI should not be overlooked. A retrospective study in non-small cell lung cancer investigated the time of development of disease flare after stopping EGFR TKI (either erlotinib or gefitinib) in patients that acquired clinical resistance. This study observed a 23 % flare rate with a median time to disease flare of eight days (range: 3–21 days) [31]. Thus, this phenomenon should be considered when designing future treatment regimens that include interval dosing of erlotinib.

As a secondary objective, we sought to describe any anti-tumor activity associated with the combination of pulsatile erlotinib and GEMOX. As this study completed enrollment, the phase III trial of daily erlotinib plus GEMOX was presented. The South Korean study demonstrated a non-significant improvement in progression-free survival in the cohort treated with erlotinib plus GEMOX compared to GEMOX alone [13]. This lack of significance could be in part a result of erlotinib inducing G1/S cell cycle arrest, instead of apoptosis, which theoretically blocks the subsequent effects of cytotoxic chemotherapy [28]. We observed an ORR of 29 % in the ABTC cohort, which is similar to the 30 % ORR observed by Lee et al. [13]. However, the combination chemotherapy plus pulsatile erlotinib tested in the current study achieved a 94 % DCR compared to 66 % in the phase III study. Furthermore, we observed a provocative OS of 18 months in our ABTC cohort. While these two studies differed in treatment doses, patient demographics, and number of allowed prior therapies, our data suggest that a therapeutic regimen combining chemotherapy with pulsatile dosing of erlotinib may be a better treatment strategy. However, due to the nature of our study design (i.e., phase Ib, 3 + 3 dose-escalation with expansion at the MTD), caution should be taken when interpreting the observed clinical activity due to the small number of patients evaluated in this study. Further prospective trials with larger sample sizes would be needed to confirm activity.

The choice of chemotherapy (e.g., GEMOX) for this study was based on preclinical clinical data showing synergy and clinical efficacy observed in ABTC [4, 32, 33]. Subsequent to the development of our study, gemcitabine plus cisplatin rather than oxaliplatin has emerged as the chemotherapy standard in ABTC. The phase III ABC-02 trial observed an improvement in both ORR (26 % versus 15 %) and DCR (81 % versus 72 %) with gemcitabine plus cisplatin compared to single agent gemcitabine with an excellent toxicity profile [6]. Therefore, one might hypothesize that investigating pulsatile dosing of erlotinib with gemcitabine and cisplatin could be more beneficial. However, the lack of a survival advantage seen by adding erlotinib in the Phase III study with GEMOX limits enthusiasm for exploring this combination with an alternate chemotherapy backbone at the current time. Ideally, as understanding of the complex landscape of the heterogeneous biology of ABTC improves, this could be a future direction, but likely only in a subset of patients.

We had hoped to find a molecularly defined subset of biliary cancers from this study that would be particularly sensitive to EGFR inhibition. Activating mutations in EGFR are well-described predictors of response to erlotinib in lung cancer [34‐37]. We found no EGFR mutations in our patients while the South Korean group found two patients with exon 20 mutations out of 116 total (1.6 %). Although KRAS mutational status is an established negative predictor of response to anti-EGFR therapies in colorectal cancer [38, 39], it is currently unknown if this status predicts response in patients with ABTC. The South Korean phase III study suggested a survival benefit with the addition of erlotinib to GEMOX in patients with wild-type KRAS [19]. Of the 11 ABTC patients in our study that had available tissue for KRAS testing, two with wild-type KRAS achieved a partial response, whereas the two ABTC patients with a KRAS mutation had stable disease. Thus, our results are in concordance with the previous phase III study suggesting a trend towards a survival advantage when chemotherapy is combined with anti-EGFR therapy in ABTC patients with wild-type KRAS. However, contradictory to the observed trends in survival for wild-type KRAS patients, three of the six patients with a KRAS mutation in the South Korean phase III study still responded to the combination therapy [19]. Furthermore, Gruenberger et al. investigated GEMOX in combination with cetuximab and reported that two of the three patients with KRAS mutations achieved a partial response [12]. These contradictory findings between KRAS mutation and response suggest that anti-EGFR therapies or, more likely, GEMOX might be beneficial irrespective of KRAS mutation in ABTC; however, further investigation is still needed to fully characterize the prognostic indication of a KRAS mutation in advanced biliary cancers. The TCGA analysis of biliary cancers demonstrates that there is extensive molecular heterogeneity of these tumors [40]. At the present time it is unclear what role, if any, EGFR inhibitors such as erlotinib may have in emerging subsets such as FGFR mutant, IDH mutant or microsatellite instability high tumors where more specific targeted or immunotherapy has demonstrated response [41‐43].

In conclusion, pulsatile erlotinib with GEMOX administered at the established MTD was well tolerated with an acceptable toxicity profile in patients with advanced pancreatic and biliary tract cancers. Additionally, this treatment combination resulted in encouraging anti-tumor activity as evidenced by a high disease control rate and longer median OS in an ABTC cohort.