Introduction

Cholangiocarcinoma (intrahepatic, extrahepatic, ampulla of Vater) and gallbladder carcinoma are uncommon cancers. When they are diagnosed as locally advanced and not amenable to combined chemotherapy and radiation or in the setting of metastatic disease, chemotherapy has served as the primary therapy. A variety of chemotherapy drugs have been evaluated, but in general the response to these drugs has been limited. Several older studies evaluated 5-fluorouracil (5-FU) alone or in combination with other forms of chemotherapy and showed mixed results. In general, 5-FU as a single agent produces few responses and provides an overall survival of less than 6 months [1, 2]. However, several recent small trials have suggested high-dose 5-FU or 5-FU in combination with other agents may produce partial responses in up to one third of patients [3–7]. Despite improved response rates, the duration of response is generally short and little increase in overall survival with 5-FU is seen.

More recently, trials have focused on gemcitabine after several case reports suggested that gemcitabine may have activity [8, 9]. Subsequent phase II trials have now been published [10–14]. The appropriate dose and schedule of gemcitabine continues to be evaluated. In one of these phase II trials, two different schedules were evaluated [14]. Gemcitabine 1,200 mg/m2 given weekly for 3 weeks, followed by a 2-week rest period, resulted in four of 24 patients (17%) achieving a partial response. The median survival was 6.8 months and the time to progression was 3.5 months. In the second arm of this study, gemcitabine 2,200 mg/m2 was given every 2 weeks. Four of 14 patients (29%) achieved a partial response. The median survival with this schedule was 10.5 months and the median time to progression was 4.8 months. In 2002, on review of the available data, the Food and Drug Administration approved a diagnosis of cholangiocarcinoma as an indication for the use of gemcitabine. Phase III trials, to establish the efficacy of gemcitabine, have yet to be reported.

Gemcitabine combined with other agents have produced responses ranging from 25% to 64% with median survivals of 10 to 15 months in several phase II trials [15–18]. However, no phase III trials and few appropriately powered phase II trials have been published to confirm these apparent promising rates. It therefore remains unclear if multiagent therapy, including gemcitabine, has any benefit over gemcitabine alone. The potential added toxicity of a second agent has also not been assessed in comparison to gemcitabine.

The selection of agents to evaluate in patients with cholangiocarcinoma is limited by the frequent occurrence of underlying hepatic dysfunction. As such potential therapeutic options need to be tolerated at clinically active doses that can be given in patients with moderate hepatic dysfunction as measured by transaminase and bilirubin values. Pemetrexed offers one potential option. Pemetrexed is a folate antimetabolite that acts primarily through inhibition of the enzymes thymidylate synthase, dihydrofolate reductase, and glycinamide ribonucleotide formyl transferase [19]. Pemetrexed’s mechanism of action and tolerability in patients with hepatic dysfunction also makes it a potential agent to evaluate in patients with cholangiocarcinoma.

This report provides the final results of a combined phase I and II clinical trial of pemetrexed and gemcitabine for patients with previously untreated biliary tract or gallbladder carcinoma. Preclinical studies have suggested a synergistic activity when pemetrexed and gemcitabine are combined [20]. A biweekly schedule was used to try to maximize the doses of each of the two drugs while limiting the toxicity.

Methods

Eligibility Criteria

Patients with a histologic diagnosis of carcinoma of the biliary tract (intrahepatic, extrahepatic, ampulla of Vater) or gallbladder with either locally advanced unresectable or metastatic disease that was not amenable to combined chemotherapy and radiation were potentially eligible for enrollment in this trial. Patients were required to be at least 18 years and to have an Eastern Cooperative Oncology Group (ECOG) performance score of 0–2. Hematologic and chemistry parameters were to be in the following ranges: absolute neutrophil count ≥ 1.5 × 109 per liter, platelets ≥ 100 × 109 per liter, total bilirubin ≤ 3.0 times the institutional upper normal limit (UNL), AST ≤ 5.0 times the UNL, and creatinine ≤ 1.5 times the UNL.

Prior embolization, chemoembolization, or radiofrequency ablation was permitted if completed at least 4 weeks prior to enrollment and new tumor growth was present. Adjuvant chemoradiotherapy or chemoradiotherapy for locally advanced disease using a drug other than gemcitabine was permitted if at least 6 months prior to enrollment. Prior radiation to ≥25% of bone marrow was not allowed. Patients were not eligible if pregnant or lactating and were required to use adequate contraception methods to prevent pregnancy during treatment. Other contraindications included a history of brain or other central nervous system metastases not amenable to local therapy. Any history of prior malignancy diagnosed within 5 years was not allowed, with the exception of basal or squamous cell carcinoma of the skin and cervical carcinoma in situ.

This study was approved by the Mayo Institutional Review Board. A signed written informed consent was obtained from all patients prior to initiating therapy. The trial was supported by a grant from Lilly Pharmaceuticals.

Treatment

Lilly Pharmaceuticals provided an investigational supply of pemetrexed (Alimta) while a commercial supply of gemcitabine (Gemzar) was used. Following registration, patients initially received folic acid 350–600 μg (commercial supply) by mouth and vitamin B12 1,000 μg (commercial supply) by IM injection 7–14 days prior to starting pemetrexed. The folic acid was continued as a daily dose until 4 weeks after the last dose of pemetrexed. Vitamin B12 injections were repeated every 12 weeks until 4 weeks after the last dose of pemetrexed. Dexamethasone 4 mg was given by mouth twice daily the day before, the day of, and the day after treatment with pemetrexed and gemcitabine. Pemetrexed, diluted into an infusion solution of 100 mL of 0.9% sodium chloride, was given as a 10-min IV infusion. At the end of the infusion, the IV line was flushed with up to 50 mL of dextrose 5% in water. Gemcitabine was reconstituted by adding 5 mL (for 200-mg vials) or 25 mL (for 1,000-mg vials) of 0.9% sodium chloride to yield a 38-mg/mL solution. The appropriate amount of the 38-mg/mL solution was then diluted in an infusion solution of 250 mL. The gemcitabine was given as a fixed-dose-rate infusion of 1,000 mg/m2 at a rate of 10 mg/min immediately after the completion of the pemetrexed infusion. Treatment with pemetrexed and gemcitabine was repeated every 2 weeks. One cycle consisted of two every-2-weeks treatments.

Evaluation, Disease Assessment, and Follow-up

Prior to each treatment, patients underwent standard clinical assessments, evaluation for adverse events, and assessment of selected chemistries including liver function tests. During treatment, patients also had weekly complete blood counts obtained. Disease status was assessed by chest X-ray, magnetic resonance imaging (MRI), or computed tomography (CT; spiral and conventional) prior to every other cycle. Following the discontinuation of treatment, patients having achieved a complete response (CR) had imaging performed every 12 weeks.

Patients with stable disease, CR, or partial response (PR) were permitted to continue treatment. Patients discontinuing treatment for reasons other than disease progression began the event-monitoring phase of the trial with reporting of status every 3 months until disease progression and every 6 months thereafter. Patients having achieved a CR were to be followed in an observation phase with an office evaluation every 12 weeks until disease progression, thereafter followed in event monitoring every 6 months. Follow-up continued until 5 years, post-registration or death, whichever was earlier.

Disease status was evaluated using Response Evaluation Criteria in Solid Tumors [21]. Measurable disease was defined as at least one lesion whose longest diameter could be accurately measured in at least one dimension as ≥2.0 cm with conventional techniques or as ≥1.0 with spiral CT scan. Use of conventional CT or MRI required a contiguous slice thickness of 1.0 cm or less, and spiral CT was performed using a 5-mm contiguous reconstruction algorithm for tumors of the chest, abdomen, and pelvis. Lesions found by physical exam were only considered measurable if they were superficial (e.g., skin nodules and palpable lymph nodes). Lesions on chest X-ray were considered measurable as long as they were clearly defined and surrounded by aerated lung although CT scanning was preferred.

Repeat tumor measurements were required at least 4 weeks following a PR or CR. A CR was defined as the disappearance of target lesions. A PR was defined as at least a 30% decrease in the sum of the longest diameter (LD) of the target lesions, relative to the sum of the baseline LD. Progressive disease (PD) was defined as at least a 20% increase in the sum of the LD of the target lesions, relative to the smallest sum LD recorded since the treatment started, or the appearance of any new lesion(s). Stable disease was defined as having occurred when patients did not achieve a CR or PR and did not manifest PD.

Duration of response was defined as the time that elapsed from the first documentation of a CR or PR to the date of progression. Time to progression (survival) was defined as the time from registration to disease progression (last contact or death). Patients lost to follow-up were censored for progression and survival at their date of last disease assessment and contact, respectively. Patients having died without documentation of their disease status at death were considered to have had progressive disease at the date of death. Survival was defined as the time from registration to death due to any cause.

Phase I Dose Escalation and Dose-Limiting Toxicity Criteria

A standard cohort of three designs was used for the phase I portion of the trial, where the maximum tolerated dose (MTD) was defined as the highest safely tolerated dose where at most one patient out of six experienced a dose-limiting toxicity (DLT) with the next higher dose having at least two out of six patients experiencing a DLT [22]. Three patients were initially treated at each dose level (Table 1) and observed for a minimum of 1 week before new patients were treated. Doses were not escalated in any individual patient. Using the National Cancer Institute Common Toxicity Criteria (NCI CTC v 2.0), a DLT for this study was defined as any of the following: grade 4 febrile neutropenia; grade 4 neutropenia or thrombocytopenia lasting ≥5 days; serum creatinine ≥2 times baseline or upper limit of normal, whichever was less; grade 4 nausea and vomiting despite maximal antiemetic treatment; grade 4 diarrhea in spite of intensive loperamide therapy; or ≥grade 3 other nonhematological toxicity. In addition, in order to qualify as a DLT, the treating physician had to have attributed the cause of the event as either possibly, probably, or definitely related to study treatment.

Table 1 Dose levels for the phase I portion of the trial
Full size table

Statistical Considerations

Primary End Point for the Phase I Portion

If DLT was seen in 1 of 3 patients treated at a given dose level, 3 additional patients were entered at the same dose level. If DLT was not evident in any of the 3 additional patients, doses were escalated. If DLT was seen in 1 or more of the 3 additional patients (i.e. 2 or more of the 6 patients entered at that dose level), the maximum tolerated dose (MTD) was considered to have been exceeded, further accrual at that dose level ceased and all further accrual occurred at the preceding dose to more fully assess the toxicities associated with the MTD. If a patient failed to complete the initial course of therapy (defined as drug administration and 1 week observation) for reasons other than toxicity, the patient was regarded as non-evaluable and was replaced with an additional patient who was treated at the same dose level. However, all toxicity information was utilized in the analysis. For patients who developed DLT a dose reduction schema allowed retreatment at one of the prior dose levels.

Primary End Point for the Phase II Portion

The primary end point for the phase II portion of the trial was 6-month survival, calculated as the number of evaluable patients alive at least 6 months post-registration, divided by the total number of evaluable patients. A one-stage design with an interim analysis was used to test whether there was sufficient evidence to determine that the 6-month survival rate was at least 65% (i.e., clinically promising) versus at most 45% (i.e., clinically inactive). This study had 92% power to detect a 6-month survival rate of 65%, with a 0.05 level of significance.

The decision rules for the design are as follows. If 12 or more of the first 24 evaluable patients lived for at least 6 months from study entry, the study would continue to a full accrual of 59 patients. If at least 33 of these 59 evaluable patients lived for 6 months or more, this would be considered adequate evidence of promising activity and would warrant further testing of this regimen in subsequent studies. Otherwise, if 11 or fewer of the first 24 evaluable patients or 32 or fewer of the first 59 evaluable patients lived for at least 6 months, we would conclude that this treatment does not warrant further study in this patient population. A confidence interval for the 6-month survival rate was calculated using the Duffy and Santner [23] approach.

Secondary End Points

Secondary end points included adverse events, the confirmed response rate, time to disease progression, overall survival, duration of response, and laboratory correlates. Adverse events are summarized in a tabular manner as the maximum grade for a given type of event for each patient. The commonly occurring grade 3+ adverse events (regardless of attribution) are also reported. Kaplan–Meier methodology [24] was used to describe the distribution of time-to-disease progression, overall survival, and duration of response. The laboratory correlates (i.e., genetic polymorphisms) are summarized via simple statistics (e.g., frequency, percentage) and correlated with the study end points of survival and progression via Cox proportional hazards modeling [25].

Results

Phase I Portion

Twelve patients were accrued to the phase I portion of the trial. Six patients were accrued to the starting dose level of pemetrexed 500 mg/m2 and gemcitabine 1,000 mg/m2 on days 1 and 15 of an every-4-weeks cycle. Since one patient canceled prior to receiving any treatment, five patients were evaluable for adverse events at this dose level. Of these five patients, four (80%) experienced at least one grade 3+ adverse event and three (60%) experienced at least one grade 4+ adverse event. Two of the five patients experienced a DLT including one patient with a grade 5 renal failure that was considered possibly related to the study medication and one with grade 3 hyperglycemia that was felt to be possibly related to the study medication. In addition to the DLTs mentioned above, the following additional grade 4 adverse events also occurred: neutropenia (60%), leukopenia (40%), thrombocytopenia (20%), and hyponatremia (20%).

Due to the toxicity observed at the starting dose level, the dose level was dropped to pemetrexed 500 mg/m2 and gemcitabine 800 mg/m2 on days 1 and 15 of an every-4-weeks cycle. Six patients were accrued to this dose level during the phase I portion, and only one patient experienced a DLT (grade 3 vasovagal episode). This dose level was therefore selected as the MTD per protocol. It was decided to include the MTD patients from the phase I portion with the phase II patients, and the summary of all these patients is below.

Phase II Summary

Baseline Characteristics

Between October of 2004 and August of 2006, 58 eligible patients were enrolled through the NCCTG, which excludes four patients that canceled prior to receiving any treatment (Table 2). At study registration, 33 (57%) of the patients were female; 53 (91%) had metastatic disease; 54 (93%) had gallbladder or biliary primaries; four (7%) received prior radiation therapy; four (7%) received prior chemotherapy, and the median age was 61 years (range 40–81). In addition, most of the patients had an ECOG performance score of 0 or 1 (93%).

Table 2 Phase II baseline patient characteristics (N = 58)—excludes four cancellations
Full size table

Outcome Measures

Fifty-eight patients were evaluable for the outcome measures of survival, time to progression, and response (Table 3). Of all 58 patients, 54 (93%) have died and 55 (95%) have progressed, with a median follow-up time of 18.2 months (range 14.2 to 34.9) for the four patients still alive. Thirty-two (55%) of the patients survived ≥6 months (95% confidence interval (CI) 39% to 74%), which nearly met our predefined criteria for success. The median survival (Table 3; Fig. 1) was 6.6 months (95% CI 5.4 to 8.7 months) and the median time to progression (Table 3; Fig. 2) was 3.8 months (95% CI 2.4 to 5.4 months). The confirmed response rate was 10% (95% CI 4% to 21%), which consisted of five partial responses and one complete response. Two of these confirmed responders had primary tumors located in the gallbladder (one CR, one PR); three responses (all PRs) were from patients’ with biliary tract tumors, and the other responder had a primary tumor located in the ampulla of Vater. All six of the patients responding to treatment have subsequently progressed, with a median duration of response of 7.1 months (95% CI 5.8 to 8.0 months).

Fig. 1
figure 1

Median survival

Full size image
Fig. 2
figure 2

Median time to progression

Full size image
Table 3 Phase II patient outcome measures (58 evaluable patients)
Full size table

Dose Intensity

A median of three cycles of therapy was given (range 1–10). Over the first six cycles of treatment, the median percent of targeted dose varied from 66% to 100% for both gemcitabine and pemetrexed, with the highest percent of targeted doses administered during the first two cycles. Of the 58 patients that received treatment, only 15 (26%) were able to receive six or more cycles of treatment. Twenty-eight (48%) patients had at least one dose reduction during the first six cycles. In addition, 21 (36%) patients had at least one cycle of treatment delayed in the first six cycles. Finally, 24 (41%) patients had at least one dose omission on day 15 in the first six cycles. All patients have ended active treatment. A majority (32/58) discontinued therapy due to disease progression; the remainder went off therapy due to refusal (11/58), adverse events (11/58), death on study (three of 58), or other medical problems (one of 58; Table 4).

Table 4 Phase II treatment information
Full size table

Adverse Events

Fifty-eight patients were evaluable for adverse events (Table 5). Of these 58 patients, 55 (95%) experienced at least one grade 3 or worse adverse event. In addition, 32 (55%) patients experienced at least one grade 4 or worse adverse event. Two patients died from grade 5 adverse events that were unrelated to study treatment. One patient had a fall and developed a hemorrhage, which led to his or her death. Another patient died from a grade 5 pulmonary event that was felt to be disease-related. Five patients (9%) died within 60 days of registration, but all five of these deaths were due to disease progression. Nearly half of the patients experienced grade 4 neutropenia (25/58; 43%), which was by far the most commonly occurring grade 3+ adverse event. Other grade 4 events that occurred in at least two patients consisted of: thrombosis (3%) and abdominal pain (3%). See Table 5 for all the details.

Table 5 Phase II commonly occurring maximum severity (grade 3+) adverse events across all cycles of treatment (regardless of attribution), including all grade 4/5 AEs
Full size table

Discussion

In this phase I and II clinical trial for patients with biliary tract or gallbladder carcinoma, the activity of the regimen of gemcitabine and pemetrexed was assessed. Gemcitabine has been the recent standard of therapy for patients with biliary tract carcinomas. Given evidence of activity and tolerability in biliary tract tumors, it was served as a platform for developing multiagent regimens with the intent of enhancing the activity of gemcitabine. However, the frequent presence of underlying liver dysfunction has been one of the major obstacles to evaluating new drugs in this setting. Pemetrexed mechanism of action and tolerability in patients with hepatic dysfunction make it a potential drug to add to gemcitabine. Finally, evidence of activity was seen in a phase I trial of this combination [26].

In this trial, a phase I biweekly schedule was established at an MTD of pemetrexed 500 mg/m2 and gemcitabine 800 mg/m2 on days 1 and 15. The dose established in this trial was lower than that established in a separate phase I trial for patients with solid tumors [27]. In that phase I trial, an MTD of pemetrexed 450 mg/m2 and gemcitabine 1,750 mg/m2 on days 1 and 15 was reached. It is likely that the presence of underlying liver dysfunction limited our ability to reach a higher MTD. Both trials indicate that the biweekly schedule is tolerable but is primarily limited by neutropenia. A weekly schedule of pemetrexed and gemcitabine was also evaluated in separate phase I trials using a day 1 and 8 administration every 3 weeks. Base on that trial, an MTD of pemetrexed 500 mg/m2 on day 8 and gemcitabine 1,250 mg/m2 on days 1 and 8 was established [28]. This schedule also produced neutropenia as the most common severe toxicity.

The lower dose of gemcitabine achieved in the phase I portion of this trial may also reflect the manner in which the gemcitabine was given. A fixed-dose-rate infusion was used based on phase I and II data. Based on the pharmacokinetic properties of gemcitabine, a higher level of the active metabolite (gemcitabine triphosphate) is achieved through a prolonged infusion. In a phase I trial using a fixed rate of infusion at 10 mg/m2, an MTD of 1,500 mg/m2 was established [29]. The dose-limiting toxicity was granulocytopenia and thrombocytopenia. In a subsequent randomized phase II trial comparing a 30-min infusion to the fixed-dose-rate infusion in patients with pancreatic cancer, the fixed-dose-rate infusion appeared to provide a superior median overall survival [30]. However, patients also experienced a higher rate of hematologic toxicity compared to the 30-min infusion. In a phase III trial in pancreatic cancer, the fixed dose rate did not produce a significantly better outcome and did result in greater toxicity [31]. Based on the findings of these three trials, it is possible that the fixed-dose-rate infusion limited the ability to give a higher and potentially more efficacious dose. In the phase II portion of the trial, there was a greater reduction of gemcitabine which occurred compared to pemetrexed, at least through cycle 3 (Table 4).

Fifty-eight patients were evaluated in the phase II portion of this trial, including 38 patients with biliary tract cancer, 16 with gallbladder cancer, and four with ampulla of Vater cancer. A response rate of 10%, including one complete response, occurred. This response rate and the median overall survival of 6.6 months observed were not superior to what would be expected with gemcitabine alone. As such, it does not appear that this combination, in the biweekly schedule used, has meaningful activity in biliary tract tumors. Use of this combination in other tumor types has provided mixed results. A phase III trial in pancreatic cancer showed no benefit over gemcitabine alone [32]. A phase II trial in transitional cell carcinoma of the urothelium also failed to show activity greater than would be expected with gemcitabine alone [33]. However, a trial in previously treated metastatic breast cancer did show a promising response rate and median overall survival [34].

In conclusion, the use of pemetrexed and a fixed-dose-rate infusion of gemcitabine on an every-2-weeks schedule did not provide meaningful benefit to patients with biliary tract carcinoma or gallbladder carcinoma. The potential benefit of pemetrexed alone or in combination with other drugs has not been assessed in this group of cancers. Trials with gemcitabine continue including a trial of gemcitabine and cisplatin versus gemcitabine alone.