Patients with resected, histologically confirmed pancreatic ductal adenocarcinoma who had not previously received adjuvant or neoadjuvant therapy were eligible for this study. Eligibility criteria also included age 18 years or older, Eastern Cooperative Oncology Group performance status of 0 or 1, and adequate organ function, including bone marrow (absolute neutrophil count ≥ 1500/µl, hemoglobin ≥ 9 g/dl, platelets ≥ 100,000/µl), liver (total bilirubin = ≤ mg/dl, aspartate aminotransferase/alanine aminotransferase < 5x upper limit of reference range, prothrombin time < 2 seconds more than upper limit of reference range, activated partial thromboplastin time < 40 seconds), and kidney (serum creatinine ≤ 2 mg/dl). Exclusion criteria included any metastasis, other malignancies diagnosed within 5 years, previous abdominal radiation, incomplete healing from previous oncological or other major surgery, inability to take oral medication, pregnancy, clinically active interstitial lung disease, uncontrolled medical illnesses, and hypersensitivity to erlotinib or capecitabine. The study protocol was approved by the institutional review board at Johns Hopkins, and patients were required to provide written informed consent before study enrollment.
The primary objective of this phase 1 study was to determine the recommended phase 2 dose for the combination of erlotinib and capecitabine when administered concurrently with IMRTas adjuvant therapy in resected pancreatic cancer patients. The secondary objectives were to determine the toxicity and pharmacokinetic profile of erlotinib and capecitabine when administered with IMRT. Adjuvant treatment was initiated 6 to 8 weeks after surgery. After completion of chemoradiation, enrolled patients began four cycles of systemic therapy consisting of concurrent gemcitabine and elotinib after 4 to 8 weeks of rest.
The original plan was to enroll a cohort of six patients to receive adjuvant treatment with concurrent erlotinib 150 mg once daily, capecitabine 800 mg/m2
twice a day (total 1600 mg/m2
a day), and IMRTat doses of 50.4 Gy in 28 fractions for 5.5 weeks (). On the basis of our own institutional experience, IMRT resulted in less gastrointestinal toxicity than three-dimensional conformal RT, and we therefore opted to use IMRTas part of the protocol. After pancreaticoduodenectomy, patients received 45 Gy to the tumor bed (as defined from preoperative imaging and clips) and adjacent lymph nodes as described previously [22
]. This was denoted planning treatment volume one (PTV1). The final cone down volume PTV2 (tumor bed plus 1–1.5 cm) received an additional 5.4Gy. After a distal or total pancreatectomy, again the tumor bed and adjacent lymph nodes received 45 Gy to PTV1 with a cone down (5.4 Gy) to the tumor bed alone plus 1 to 1.5 cm (PTV2). Dose-limiting structures included the liver 50% with less than 30 Gy, 66.67% of one kidney with less than 18 Gy, and spinal cord with less than 45 Gy, and every effort was made to limit radiation dose to the small bowel and stomach. The goal was to ensure that the PTV was covered by the 95% isodose line and avoid any hot spots greater than 10% of the prescribed dose. A preplanned analysis of the first cohort was performed before proceeding to complete enrollment for the phase 2 portion. A dose level would be considered intolerable if 33.33% of patients experience dose-limiting toxicity. During the analysis, the starting dose level (dose level 1) was determined to be intolerable by this definition. The doses were modified to erlotinib 100 mg daily, capecitabine 800 mg/m2
twice daily (Monday to Friday, thereby providing a break during weekends) to 50.4 Gy in 28 fractions (dose level -1).
Dose Levels of Erlotinib and Capecitabine during Combined Chemoradiation.
Toxicity was assessed weekly during chemoradiation and for up to 8 weeks after the completion of chemoradiation or the start of systemic adjuvant chemotherapy, whichever came first, using National Cancer Institute Common Terminology Criteria for Adverse Events version 3. Dose-limiting toxicity was defined as treatment-related grade 3 non-hematological toxicity or worse despite maximal supportive treatment, grade 4 neutropenia for 5 days or longer, grade 3 or 4 febrile neutropenia (absolute neutrophil count < 1.0 x 109), fever with a temperature of 38.5°C or higher, grade 4 thrombocytopenia (<25,000/mm3), and grade 3 or worse cutaneous toxicity. The recommended phase 2 dose was defined as the dose level at which fewer than 2 patients of six experience treatment-related dose-limiting toxicity.
Pharmacokinetic (PK) Sampling and Data Analysis
Pharmacokinetic (PK) studies were performed after a single dose administration of capecitabine and erlotinib during cycle 1 day 1. Serial sampling of venous blood for PK assessment was performed before and after treatment at 24 hours for erlotinib and at 8 hours for capecitabine during cycle 1 day 1. Blood samples were processed by centrifugation at 1000g
at 4°C for 10 minutes. For capecitabine samples, tetrahydrouridine, a cytidine deaminase inhibitor, was added at a final concentration of 400 nM to increase the stability of capecitabine and metabolites in plasma during storage [23–25
]. Plasma samples were stored at -20°C or below until analysis.
Erlotinib and the OSI-420 metabolite and capecitabine and the 5′ DFCR, 5′DFUR, and 5-FU metaboliteswere quantified using a validated liquid chromatography-tandem mass spectrometry analytical method [23,26
]. Individual pharmacokinetic parameters for capecitabine were estimated by standard noncompartmental analysis using WinNonlin version 5.0 (Pharsight, Mountain View, CA) [27
]. The capecitabine data were fit to a one-compartment linear model using weighted least squares regression. The pharmacokinetics of erlotinib was assessed by compartments analysis using ADAPT II [28
]. The erlotinib data were fit to a one-compartment linear model using weighted least squares regression and an iterative two-stage approach as previously described [29
]. For several patients, one or more blood samples were missed or obtained improperly and were not usable for analysis. Any sample that was documented not to be a pretreatment sample (i.e., within 3 hours before the next dose or after a dose) was not used in subsequent analysis. After excluding concentrations from these samples, the total number of concentration-time observations per individual was relatively small compared with the number of parameters estimated. Therefore, an iterative two-stage approach was implemented to estimate pharmacokinetic parameters for all patients. Parameter estimates from singledose data from a previous trial conducted at our institution [29
] were used to establish Bayesian priors for the structural model parameters, which included volume of the central compartment (Vc
), absorption rate constant (Ka
), and elimination rate constant (Ke
). An iterative two-stage approach was used in each iteration updating the Bayesian priors, until the mean estimates of all parameters differed by less than 5% from the previous mean estimate, which was our arbitrarily predefined stopping point. In the final model fit, data from all patients, including those with incomplete observations (four to six observations; time of last plasma concentration = 8 to 24 hours), were analyzed using a Bayesian algorithm to estimate individual structural parameters. Calculated secondary pharmacokinetic parameters included half-life (T 1/2
) and apparent systemic clearance (Cls
). Area under the curve (AUC) was calculated as dose divided by Cls
. The maximum drug concentration (Cmax
) and time maximum plasma concentration achieved (Tmax
) values were obtained from the observed values for both erlotinib and OSI-420.
All outcome parameters were summarized using descriptive statistics. AWilcoxon signed rank test was used to compare PK parameters to erlotinib single-agent administration obtained from patients treated on a previous clinical trial at our institution [29
]. The a priori
level of significance was set at P
< .05. These tests were performed using SPSS version 15 (SPSS, Chicago, IL) or JMP Statistical Discovery software (version 4.0.4; SAS Institute, Cary, NC).
Fourteen patients were enrolled to the study from March 2006 to April 2008. Patients were observed for up to 8 weeks for toxicity on completion of chemoradiation or removal from study. The characteristics of the patients are shown in . The median age was 60 years (range = 47–73 years) and 62% were females. Eight (85%) underwent a pancreaticoduodenectomy, 4 (31%) underwent a distal pancreatectomy, and 1 (8%) underwent a total pancreatectomy. After surgery, 10 had pathologic stage 2, 2 had stage 1, and 1 had stage 3 disease.
Patient Characteristics (n = 13).
Toxicity of Chemoradiation and Determination of Recommended Phase 2 Dose
Eight patients were enrolled to dose level 1. One patient was replaced after developing ischemic bowel from adhesions requiring emergent surgery. A second patient had treatment-unrelated wound dehiscence and was not evaluable for dose-limiting toxicity assessment. This patient was replaced, although the toxicity data are reported here. Four (57%) of the six evaluable patients completed the planned chemoradiation, although all required dose modification or interruption. Four patients developed dose-limiting toxicities. The first patient experienced grade 3 neutropenia for more than 5 days, and the second patient had uncontrolled grade 3 diarrhea that required hospitalization. Both did not complete the planned chemoradiation. The third patient developed grade 3 sensory neuropathy requiring dose reduction, and the fourth patient had grade 3 rash. Capecitabine dosing was interrupted in five patients (range days 8–29) and erlotinib in three patients (range days 15–31). Dose level 1 was therefore determined as intolerable at a planned interim analysis, and additional six patients were enrolled to dose level -1.
All six patients enrolled to dose level -1 completed the planned chemoradiation. Four patients developed elevated liver enzymes, which improved when capecitabine and/or erlotinib were held for fewer than 7 days. All patients had grade 1 rash or worse and two had grade 1 diarrhea or worse. Dose level -1 was therefore considered clinically feasible and declared as the recommended phase 2 dose. No grade 4 toxicity or worse was observed at both dose levels. The most common (>30%) grade 2 or worse treatment-related toxicities were fatigue (46%), elevated alanine aminotransferase (46%), and anorexia (31%) ().
Grade 2 and above Treatment-Related Toxicity during Chemoradiation.
Chronic gastrointestinal toxicity was seen in two patients. One patient (dose level 1) developed a bile duct stricture at the surgical anastomosis 1 year after the Whipple operation. A second patient (dose level -1) developed radiation enteritis of the colon (within the radiation field) 27 months after a distal pancreatectomy. In an attempt to dilate the stricture, the bowel was perforated, and the patient required surgery. At the time of resection, the patient was found to have biopsy-proven metastatic disease to the liver. No other late effects have been reported from this cohort of patients.
Erlotinib and capecitabine pharmacokinetic data was evaluable for all patients. Erlotinib and OSI-420 Cmax
, AUC, and T1/2
are presented in . The Cmax
of erlotinib was observed at 5.04 hours after oral administration with a mean terminal half-life of 14.7 hours. Erlotinib Cmax
and AUC were comparable to data from an erlotinib single-agent administration from another clinical trial at Johns Hopkins () and published historical controls [29,30
]. There was no difference in the Cmax
, AUC, T1/2
of capecitabine, 5′DFCR, 5′DFUR, and 5-FU with either dose of daily erlotinib (100 or 150 mg; P
> .05). Therefore, the data are combined and presented in . High interpatient variability was observed for capecitabine and metabolite pharmacokinetic parameters, which is consistent with previous reports in the literature [31–33
Summary of Pharmacokinetic Parameters for Erlotinib, Capecitabine, and Metabolites.
Erlotinib Cmax (ng/ml) (A) and AUC (B) after a single dose of erlotinib and capecitabine (closed symbols) compared with erlotinib alone (open symbols). Data for erlotinib alone were obtained from single-agent clinical trials at our institution.
Concurrent chemoradiation preceding or followed by systemic chemotherapy remains one of the standards in the adjuvant treatment of pancreatic cancer. The National Comprehensive Cancer Network recommends the use of fluoropyrimidines, including capecitabine, as a radiosensitizer in the adjuvant setting [34
]. The recent European Study Group for Pancreatic Cancer 3 trial (ESPAC-3) study, randomizing resected pancreatic cancer patients to gemcitabine or fluorouracil/folinic acid, reported no significant survival difference between the two arms, indicating that fluoropyrimidine may be as efficacious as gemcitabine in the adjuvant setting [35
]. Efforts to integrate EGFR inhibitors into adjuvant therapy are of great interest after the positive result of erlotinib in advanced disease. However, progress has been slowed by the lack of a clinically feasible regimen.
At the inception of this protocol, therewere limited preclinical studies to guide how to optimally combine erlotinib, capecitabine, and radiation [8,10
]. Nevertheless, we hypothesized that combining an inhibitor of the EGFR pathway (erlotinib) with capecitabine and radiation may have a potent and selective antitumor activity against pancreatic cancer.
Czito et al. [19
] evaluated concurrent conformal (three or four fields) external-beam radiation therapy (50.4 Gy) with gefitinib 250 mg daily and capecitabine twice daily (7 days/wk) at the doses of 650 and 825 mg in 10 pancreatic cancer patients. Six patients (60%) developed dose-limiting toxicity, including diarrhea and arterial thrombi. The authors concluded that the combination resulted in excessive toxicities, and the recommended phase 2 dose was not determined. In comparison, the starting dose level in our study (capecitabine 800 mg/m2
, 7 days/wk) also resulted in excessive toxicities, and the dose-limiting toxicity included diarrhea, neutropenia, and rash. After dose de-escalation, the combination of concurrent IMRTwith capecitabine 800 mg/m2
twice daily (Monday to Friday) and erlotinib 100 mg daily resulted in less toxicity and has been found to be clinically feasible and determined to be the recommended phase 2 dose. We hypothesize that the lower frequency of capecitabine (5 vs
7 days/wk) improved the tolerability of this regimen. This coupled with the use of IMRT as opposed to conformal irradiation may have resulted in less small bowel irradiation, thus decreasing gastrointestinal toxicity. Although unclear, reported anorexia was likely due to a combination of radiation and difficulty taking oral chemotherapy (erlotinib and capecitabine) after surgical resection.
The toxicity profile of the study regimen is comparable to that previously known to be associated with the two agents given independently or with radiation. In a study of capecitabine with concurrent radiation, Dunst et al. [36
] reported 825 mg/m2
twice a day on a continuous basis with concurrent radiation as clinically feasible in rectal cancer. The dose-limiting toxicity at 1000-mg/m2
twice-a-day level was hand-foot syndrome, and non-dose-limiting grade 3 toxicities or worse were diarrhea and skin reactions. The recommended dose for capecitabine in our study is comparable to that reported by Saif et al. [9,37
], at the level of 800 mg/m2
twice a day (Monday to Friday) with concurrent radiation in locally advanced pancreatic cancer patients.
The frequency of non-dose-limiting liver enzyme elevation at the recommended phase 2 dose in our study seems to be higher than previously published reports. In the National Cancer Institute of Canada BR.21 trial, liver enzyme abnormalities were transient and occurred in 4% and less than 1% of erlotinib- and placebo-treated patients, respectively [38,39
]. There was no grade 3 or worse elevation in erlotinib-treated patients. Saif et al. [9
] reported grade 3 hyperbilirubinemia in 5% of pancreatic cancer patients who received concomitant radiation with capecitabine 800mg/m2
twice a day (Monday to Friday) for 6 weeks, but there was no grade 3 or worse alanine aminotransferase, aspartate aminotransferase, or alkaline phosphatase elevation. Liver enzyme abnormalities were reported in less than 10%of patients receiving combinations containing capecitabine and erlotinib [32,33,40
]. Our institutional standard adjuvant therapy regimen (off protocol) is capecitabine 800 mg/m2
(Monday to Friday) with IMRT. These patients rarely experience elevated liver function tests; therefore, we believe that this elevation is more likely due to the addition of erlotinib. Given these, we speculate that the higher frequency of liver enzyme elevation may be a result of a longer continuous concurrent exposure to erlotinib with capecitabine, and this should be monitored closely in future adjuvant studies.
The recommended phase 2 dose of concurrent capecitabine and erlotinib in our study is comparable to the efficacious dose in non-radiation-containing regimens. Kulke et al. [40
] reported a phase 2 study of a combination of erlotinib 150 mg daily continuously and capecitabine 1000 mg/m2
twice daily for 2 weeks followed by 1 week of rest in patients with gemcitabine-refractory metastatic pancreatic cancer. The progression-free survival and median survival were comparable to published second-line studies at 3.4 and 6.5 months, respectively [41–44
Drug-drug interaction was not expected between capecitabine and erlotinib because their metabolic pathways are theoretically distinct. In our study, the pharmacokinetic profile of erlotinib is not significantly different in the presence of capecitabine when compared with an erlotinib single-agent study [29
]. This is consistent with previously published trials of erlotinib in combination with capecitabine and docetaxel or in combination with capecitabine and oxaliplatin [32,33
]. High interpatient variability was observed for capecitabine and metabolite pharmacokinetic parameters, which is consistent with previous literature, making it difficult to determine whether a drug-drug interaction occurred [31–33
]. In addition, there are conflicting data in the literature regarding drug-drug interaction between erlotinib and capecitabine. In a trial in breast cancer patients, no alterations in capecitabine pharmacokinetics were noted when 825 mg/m2
twice a day capecitabine was combined with erlotinib (100mg/d) and docetaxel (60–75 mg/m2
intravenously every three weeks) [32
]. However, there was a trend toward decreased capecitabine and metabolite exposure when capecitabine (825–1000 mg/m2
/d) was combined with erlotinib (100mg/d) and oxaliplatin (130 mg/m2
intravenously every three weeks) in metastatic colorectal cancer patients [33
]. In retrospect, monotherapy and combination pharmacokinetic assessments should have been performed in this trial to definitively confirm the presence or absence of drug-drug interaction when involving a highly variable drug like capecitabine. Regardless of the drug-drug interaction, in both trials where capecitabine was combined with erlotinib, the maximum tolerated dose (MTD) of erlotinib was 100 mg/d. This is consistent with the MTD observed in the current trial. Whether a drug-drug interaction exists and influences patient outcomes is unclear, and future studies should incorporate appropriate pharmacokinetic studies to evaluate this relationship definitively.
The role of erlotinib in combination with gemcitabine-containing radiation regimens has been explored in pancreas cancer but in locally advanced disease: Iannitti et al. [45
] reported erlotinib 50 mg daily as the MTD in combination with weekly gemcitabine (75 mg/m2
), paclitaxel (40 mg/m2
), and radiation (50.4 Gy), whereas Duffy et al. [46
] reported erlotinib 100 mg daily as the MTD in combination with gemcitabine (40 mg/m2
twice weekly) and radiation 1.8 Gy daily. Recently, the results from ESPAC-3 support the use of a fluoropyrimidine in the adjuvant setting instead of gemcitabine, given the higher frequency of hematological toxicities observed in the gemcitabine arm [35
]. The new intergroup adjuvant trial (Radiation Therapy Oncology Group 0848) is evaluating six cycles of gemcitabine with or without erlotinib followed by a randomization for patients to receive or not to receive 5-FU-based chemoradiation. Unlike the regimen described herein, the intergroup trial does not include erlotinib during the chemoradiation. As such, this trial is unique, as it is the first tolerable EGFR inhibitor-containing adjuvant pancreas cancer regimen given concurrently with 5-FU-based chemoradiation. We are currently enrolling patients to the phase 2 cohort of the study, and we anticipate completion within the next year.