- Email: [email protected]
Low-dose trimetrexate glucuronate and protracted5-fluorouracil infusion in previously untreated patients with advanced pancreatic cancer
Annals of Oncology 13: 582–588, 2002 DOI: 10.1093/annonc/mdf090
Original article
Low-dose trimetrexate glucuronate and protracted 5-fluorouracil infusion in previously untreated patients with advanced pancreatic cancer R. G. Amado*, L. S. Rosen, J. R. Hecht, L.-S. Lin & P. J. Rosen Division of Hematology/Oncology and Jonsson Comprehensive Cancer Center, University of California, Los Angeles, CA, USA Received 18 June 2001; revised 26 September 2001; accepted 16 October 2001
Background: 5-Fluorouracil (5-FU)-based regimens have not been shown to prolong survival or provide clinical benefit in patients with advanced pancreatic cancer. The purpose of this study was to determine the tolerability of protracted venous infusion (PVI) of 5-FU, modulated by a low dose of the synthetic antifolate trimetrexate, in patients with advanced pancreatic cancer. Patients and methods: Twenty-three chemotherapy-naïve patients were evaluated. Patients were enrolled in four consecutive cohorts in which the weekly dose of trimetrexate was escalated in 10 mg/m 2 increments, from 20 to 50 mg/m2. PVI 5-FU was administered at a fixed dose of 225 mg/m2/day. Treatment was administered for 6 successive weeks, every 8 weeks. Results: Twenty-two patients were assessable. The maximum tolerated dose of trimetrexate was 40 mg/m2. The most common grade 3 and 4 toxicity was diarrhea. There were no treatment-related deaths. Preliminary analysis of activity revealed a response rate of 9%, with 41% of the patients having stable disease for a median duration of 3.8 months. The median survival for the entire group was 6.9 months (range 1–29 months). A clinical benefit response was experienced by 27.2% of patients. Conclusions: Low-dose trimetrexate can be safely administered in combination with PVI 5-FU. This treatment is well tolerated and is associated with palliative activity in advanced pancreatic cancer. Key words: antifolates, 5-fluorouracil, 5-FU modulation, infusional 5-FU, pancreatic cancer, trimetrexate
Introduction Pancreatic cancer is a lethal disease, for which the current treatment is of limited value. Over 80% of patients present with locally advanced or metastatic disease, a situation where current therapy results in only modest clinical benefit, without a major impact on survival [1, 2]. 5-Fluorouracil (5-FU) administered in bolus form is the most extensively studied drug. Objective responses to bolus 5-FU occur in <10% of patients, without an impact on overall survival or quality of life [3–9]. Protracted venous infusion (PVI) of 5-FU allows for higher dose intensity than bolus administration, is associated with a different toxicity profile and seems to have a higher response rate and a marginal survival benefit in colorectal cancer patients [10]. Doses of 200–300 mg/m2/day of 5-FU have been employed in advanced pancreatic cancer as a single agent or in combinations [11–16]. Response rates to PVI 5-FU have ranged from 8 to 19%.
*Correspondence to: Dr R. G. Amado, UCLA School of Medicine, Division of Hematology/Oncology, 11-964 Factor Building, Los Angeles, CA 90095-1678, USA. Tel: +1-310-825-8419; Fax: +1-310-825-6192; E-mail: [email protected] © 2002 European Society for Medical Oncology
Trimetrexate is a synthetic folate analog, which potently inhibits dihydrofolate reductase (DHFR) [17–19]. As a result of inhibition of purine synthesis, trimetrexate increases intracellular levels of phosphoribosyl pyrophosphate (PRPP), a cofactor for the enzyme orotate phosphoribosyltransferase, which is required for the intracellular conversion of 5-FU to active fluorinated nucleotides. These fluoronucleotides are incorporated into RNA, thus potentiating 5-FU cytotoxicity [20, 21]. Trimetrexate has been reported to enhance the cytotoxicity of 5-FU and folinic acid in vitro [22] in a scheduledependent manner [20, 23, 24]. Based on these preclinical results, sequential administration of trimetrexate has been studied in patients with colorectal cancer using schedules that employ bolus 5-FU [21, 25]. It is unknown whether trimetrexate-mediated 5-FU modulation occurs using PVI 5-FU. As indicated, increased levels of PRPP mediate RNA and protein synthesis inhibition. However, the putative primary mechanism of action of PVI 5-FU is inhibition of thymidylate synthase (TS), rather than RNA synthesis [26–28]. Several lines of evidence lend support to the possibility of synergy between trimetrexate and PVI 5-FU, based on potentiation of TS inhibition in the absence of folinic
583 acid. First, elevation of intracellular PRPP results in increased levels of 5-fluorodeoxyuridine monophosphate (FdUMP), the 5-FU nucleotide involved in TS inhibition. Secondly, inhibition of DHFR by methotrexate is associated with an increase in intracellular levels of dihydrofolate polyglutamates [29]. These metabolites, like folinic acid [30], can complex with TS and FdUMP to enhance TS inhibition [31]. Based on the activity and favorable toxicity profile of PVI 5-FU in phase II studies, which suggests synergy between trimetrexate and 5-FU, and on the above rationale for a potential modulatory effect on PVI 5-FU by trimetrexate, we conducted this dose escalation trial in previously untreated patients with advanced pancreatic cancer. The main objectives of this study were to assess toxicity and to determine maximum tolerated dose. Secondary objectives were response rate, time to treatment failure, survival and clinical benefit.
Patients and methods Patients Patients ≥18 years of age, with histologically confirmed adenocarcinoma of the pancreas, either locally advanced (unresectable stages II and III) or metastatic (stage IV) were eligible. A Karnofsky performance status (KPS) of ≥60% was required. Patients who had lost >10% of their weight within the previous 2 months were ineligible. Patients who had received previous chemotherapy for advanced metastatic pancreatic cancer were also excluded, but adjuvant chemoradiotherapy was allowed. All patients had bi-dimensionally measurable disease. Patients who had received previous radiotherapy were required to have measurable disease outside the radiation port, or measurable progressive disease in the radiation port. An adequate baseline bone marrow function was necessary, defined as an absolute neutrophil count (ANC) ≥1500/mm3 and a platelet count ≥100 000/mm3. Adequate renal function (defined as serum creatinine ≤2.0 mg/dl) and hepatic function (defined as serum bilirubin ≤2.0 mg/dl, serum albumin ≥2.5 mg/dl and AST ≤3 × the upper limit of normal in the absence of liver metastasis, or ≤5 × the upper limit of normal in the presence of liver metastasis) were also required. Informed consent was obtained before study entry. This study was performed with the approval of the local human subject protection committee.
Treatment All patients were treated on an outpatient basis. Patients underwent placement of a central venous catheter for PVI 5-FU infusion. Patients were entered into one of four cohorts to receive trimetrexate glucuronate at 20, 30, 40 or 50 mg/m2 weekly. As folinic acid can exacerbate the toxicity associated with PVI 5-FU, we used lower doses of trimetrexate than those previously employed [21, 25] to avoid the need for folinic acid rescue. We performed dose escalation after treatment of five patients in a prior cohort had been completed without excessive toxicity. In the event that three patients in any given cohort experienced grade 3 or 4 toxicity, dose escalation was terminated and all remaining patients were to be enrolled at the preceding dose level. All patients remained in their cohort of entry until discontinuation from the study. A treatment cycle was defined as 8 weeks and included 6 weeks of chemotherapy followed by a 2-week rest period. Treatment cycles were repeated until either disease progression or unacceptable toxicity supervened.
Trimetrexate glucuronate was supplied by MedImmune Oncology (Gaithersburg, MD, USA) in lyophilized powder (25 mg/vial). The drug was reconstituted and diluted, with 5% Dextrose Water for Injection USP, to obtain concentrations of 0.25 mg/ml to 2 mg/ml. Patients received 20–50 mg/m2 of trimetrexate i.v. over 60 min on days 1, 8, 15, 22, 29 and 36 of every 8-week cycle. Patients were to receive folinic acid rescue in the event of grade 4 hematological toxicity. After completion of trimetrexate infusion on day 1, patients received treatment with 5-FU at a dose of 225 mg/m2/day administered as a continuous i.v. infusion during the first 6 weeks of every 8-week cycle.
Dose modifications All toxicities were graded using the NCI Common Toxicity Criteria version 1. Patients were evaluated weekly for toxicity. In the event of a grade 2 hematological toxicity, the dose of trimetrexate was reduced by 25%. For grade 3 or 4 hematological toxicity, both trimetrexate and 5-FU were withheld until blood counts returned to ANC >2000/mm3 and platelets >100 000/mm3, then the dose of trimetrexate was reduced by 25% for grade 3 and 50% for grade 4, and 5-FU was resumed at a dose of 225 mg/m2. For grade 2, or greater, mucosal ulceration or diarrhea, 5-FU and trimetrexate were discontinued until symptom resolution, and then the doses were resumed as follows: unmodified for grade 2; reduced by 10% for 5-FU and 25% for trimetrexate in the event of grade 3; and by 20 and 50%, respectively, in the event of grade 4. For hand–foot syndrome (including pain, erythema, swelling and tenderness in hands and/or feet), both agents were discontinued and treatment was resumed after resolution of symptoms with full dose trimetrexate, 90% dose-reduced 5-FU for grade 3 and 80% for grade 4. The duration of the cycles was not prolonged to compensate for toxicity-related treatment interruptions. Patients whose doses were withheld for 3 consecutive weeks because of any toxicity were terminated from the study.
Assessment of efficacy Standard radiographic criteria for response were used. Survival and time to treatment failure were assessed prospectively. Clinical benefit analysis was performed using a modified version of methods previously described [4], in which weight was considered a primary measure along with pain and functional status, measured using the KPS scale. Data on these parameters were collected weekly. The pain parameter was a composite of information obtained on pain score and analgesic consumption. Information on pain score was obtained weekly from a brief pain inventory. Records of analgesic consumption were normalized to milligrams of oral morphine sulfate over a 1-week period. KPS was assessed by two observers. Patients were classified as positive, stable or negative for each of the three parameters. To be considered as positive or negative in any parameter, improvements or deteriorations had to be sustained for at least 28 days. Pain was scored as positive if it improved ≥50% from baseline (assuming an initial minimum pain score of ≥2 on a 0–10 scale), as negative if it worsened over baseline and as stable for all other results. Analgesic consumption was scored as positive if it decreased ≥50% from baseline (assuming an initial minimum of ≥5 mg of morphine), as negative if it increased from baseline and as stable if any other result. If both pain score and analgesic consumption were positive, or if one was positive and the other remained stable, the patient was classified as positive for clinical pain score. If either or both were negative, the patient was classified as a non-responder and if both remained unchanged the patient was classified as stable. Weight parameter was scored as positive if dry weight increased by ≥7% from baseline, as negative if it decreased by ≥4% and stable for all other results. KPS score was considered as positive if it increased by
584 ≥20%, as negative if it decreased by ≥20% and as stable for all other results. Patients were classified as positive for overall clinical benefit if at least one of the parameters improved without deterioration of the others and as non-responder if all three parameters remained stable or if any deteriorated.
Statistical methods Patient survival and time to disease progression were measured in days from study entry and data were summarized by Kaplan–Meier survivor function estimates and associated confidence intervals. Cox proportional hazards regression was used to assess the effects of trimetrexate dose, treatment time and baseline patient characteristics, on survival and time to progressive disease. Differential incidence of toxicity events across dose levels was analyzed by the χ2 test of equivalent incidence, and associations between trimetrexate dose and toxicity grade were analyzed by the χ2 test of association.
Results Patient demographics
Table 1. Patient characteristics Characteristics
n
Patients enrolled
23
Assessable for response/toxicity
22
95
Age (years) Median
66
Range
40–80
Gender Male Female
16
70
7
30
1
5
Karnofsky performance status 60% 70%
0
0
80%
11
50
90%
8
36
100%
2
8
1
4
Stage III
4
17
Stage IV
19
83
Pancreas
20
87
Liver
14
61
3
13
Peritoneum
3
13
Lungs
2
9
Other
1
4
1
8
35
2
11
48
3
4
17
Well differentiated
4
17
Moderately differentiated
6
26
Poorly differentiated
8
35
Unknown
5
22
20
5
22
30
5
22
40
6
26
50
7
30
Prior adjuvant therapy
From February 1998 to August 1999, 23 patients were enrolled in this university-based multicenter study in which nine community oncology sites participated. One patient enrolled in the 50 mg/m2 trimetrexate dose cohort withdrew before receiving treatment and is not included in the analysis. All other patients were included in an intention-to-treat analysis, including one patient, enrolled in the 40 mg/m 2 cohort, who voluntarily discontinued therapy after 1 week on the study. The total number of cycles administered was 45 (median 1, range 1–8) over four dose levels of trimetrexate. Patient characteristics and treatment groups are shown in Table 1.
% patients
Disease stage
Disease localization
Lymph nodes
Number of sites involves
Toxicity Seven patients required dose modifications because of toxicity; in four of them toxicity was grade 3 or 4. One patient (a radiographic and clinical benefit responder) had to be withdrawn from the study because of repeated dose reductions and treatment discontinuations for grade 2 mucositis and hand– foot syndrome. This patient was enrolled on the trimetrexate cohort receiving 20 mg/m2. No other toxicity-related study discontinuations occurred. Out of 230 administered doses of trimetrexate and 5-FU in this study, 6% of the doses were withheld because of toxicity, and dose modifications were implemented in 7.3% of the trimetrexate doses and 3.5% of the 5-FU doses. Dose reductions or discontinuations occurred in 21% of patients in cohorts receiving 20–40 mg/m2 trimetrexate, and 50% of patients in the 50 mg/m2 cohort. There were no treatment-related deaths. Grade 3 and 4 toxicity is shown in Table 2. The most common toxicity event leading to drug modification or discontinuation was diarrhea. Diarrhea was also the most common grade 3 and 4 toxicity overall (23% of all patients). Three patients, or 50% of the cohort treated at the 50 mg/m2 dose, experienced grade 4
Pathology
2
Trimetrexate dose (mg/m )
diarrhea. Therefore, diarrhea was considered the dose-limiting toxicity. One patient in the 50 mg/m2 group experienced a non-neutropenic septic episode that resolved with antibiotic treatment. There were no catheter-related thrombotic epi-
585 Table 2. Grade 3 and 4 treatment-related toxicity Toxicity
Trimetrexate dose level (mg/m2/week) 20 n
30 % cohort
40
% total
n
% cohort
% total
n
50 % cohort
% total
n
% cohort
% total
Asthenia Grade 3
0
0
0
1
20
5
0
0
0
1
17
5
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Nausea and vomiting Grade 3
0
0
0
0
0
0
1
17
5
1
17
5
Grade 4
1
20
5
0
0
0
0
0
0
0
0
0
Grade 3
0
0
0
0
0
0
1
17
5
0
0
0
Grade 4
1
20
5
0
0
0
0
0
0
3
50
14
Grade 3
0
0
0
0
0
0
0
0
0
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
1
17
5
Grade 3
0
0
0
0
0
0
0
0
0
1
17
5
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Grade 3
0
0
0
1
20
0
0
0
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
1
17
5
Grade 3
0
0
0
0
0
0
1
17
5
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Diarrhea
Abdominal pain
Anorexia
Infection
Cutaneous
Neutropenia Grade 3
0
0
0
0
0
0
1
17
5
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Grade 3
0
0
0
0
0
0
1
17
5
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Grade 3
0
0
0
0
0
0
0
0
0
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Anemia
Thrombocytopenia
sodes, grades 3 or 4, stomatitis, hand–foot syndrome or hypersensitivity reactions. The most common grade 1 and 2 nonhematological toxicities were nausea and vomiting (55%), cutaneous (50%), stomatitis (45%), diarrhea (32%) and hand– foot syndrome (27%). With regard to hematological toxicity, grades 1 and 2 neutropenia, anemia and thrombocytopenia occurred in 9, 27 and 27% of patients, respectively, across all four cohorts. There were no grade 4 hematological toxicity events. One patient in the 40 mg/m2 cohort experienced grade 3 anemia and neutropenia. When all toxicity categories were analyzed together their grade was found to increase significantly with the dose of
trimetrexate (P = 0.008). In addition, the total number of toxic events, irrespective of their grade, tended to increase with the dose of trimetrexate (P = 0.062). The maximum tolerated and recommended dose of trimetrexate for future investigation using this combination schema dose was determined to be 40 mg/m2.
Activity Twenty-two patients were assessable for radiographic response. Two patients experienced a partial response (response rate 9%) and nine patients (41%) had stable disease for at least 8 weeks. The two responses lasted 3 and 3.9
586
Figure 1. Time to disease progression for all patients and for the metastatic disease patient subgroup.
Figure 2. Kaplan–Meier survival curves for all patients and for the metastatic disease patient subgroup.
months, occurring in one patient with metastatic disease and in another with locally advanced disease, in the 20 mg/m2 and 40 mg/m2 cohorts, respectively. The median duration of stable disease was 3.8 months. The median time to disease progression for the entire group was 2 months (Figure 1). The median survival for the entire group was 6.9 months (range 1–29 months) and the 1-year survival was 27% (Figure 2). For patients with stage IV disease, median and 1-year survivals were 6.7 months and 26%, respectively. No correlation was observed between baseline patient characteristics or trimetrexate dose, and time to treatment failure or survival. At the time of this analysis, the two responders remain alive at 32 and 25 months from initiation of therapy. Six patients fulfilled the criteria for clinical benefit response, for a total response rate of 27.2%. Three patients fulfilled these criteria on the basis of improvements in pain scores with stable weight and KPS scores, one had an improvement in pain with the other two parameters remaining stable during one cycle, followed on the subsequent cycle by an improvement in weight with stable pain and KPS scores, and two patients qualified for a response on the basis of an improvement in weight with stable pain and KPS scores. Of the six patients, one was classified as a partial radiographic responder and the remaining five patients had stable disease. The second radiographic responder did not qualify for a clinical benefit response as all three parameters remained stable throughout the duration of therapy.
dose and biochemical evidence of modulation exists for relatively low doses of this drug [21], we have chosen 40 mg/m2 as the dose of trimetrexate for future investigations. The relative paucity of severe toxicity using doses of trimetrexate below 50 mg/m2 contrasts with that of previously reported results from phase I and II studies using higher doses of trimetrexate as a single agent. In those studies, the major toxicity of trimetrexate was myelosuppression [17, 32, 33]. The toxicity profile of the current study also compares favorably with that of previously reported double modulation 5-FU schedules with folinic acid rescue [21, 25]. The current treatment schema is simpler to administer, as treatments are scheduled only once weekly and do not require oral folinic acid to rescue trimetrexate toxicity. In the present study, response seemed to correlate with survival; the two patients that experienced a radiographic response (one with metastatic and one with locally advanced disease) remain alive at 32 and 25 months from initiation of therapy. As a secondary objective of this study, we sought to obtain preliminary evidence of clinical improvement, because symptom control is an important endpoint in the treatment of advanced pancreatic cancer. Our analysis showed a clinical benefit response rate of 27%. However, because the main endpoint of this study was dose finding, with patients treated across a range of trimetrexate doses, these efficacy results should be considered preliminary. It should also be noted that, although the study was opened to community oncologists, reducing the possibility of selection bias, extrapolating these toxicity and activity results to routine clinical practice should take into account that most patients had well-preserved performance status, a relatively uncommon situation in patients presenting with advanced pancreatic cancer. It is not known whether trimetrexate-mediated 5-FU modulation resulted in enhanced activity over that obtained with unmodulated PVI 5-FU. A randomized study would be required to establish the role of trimetrexate in this treatment schedule. Any benefit of the combination of trimetrexate and 5-FU is unlikely to be derived from a direct cytotoxic action of trimetrexate, as previous studies indicate that trimetrexate, as a single agent, is inactive against pancreatic cancer [34, 35]. Given that 5-FU and gemcitabine do not exhibit cross-
Discussion We undertook this study to determine the maximum tolerated dose and the toxicity of the combination of low-dose trimetrexate and PVI 5-FU without folinic acid in patients with advanced pancreatic cancer. To our knowledge, this study represents the first report of PVI 5-FU modulation using an antifolate. Doses of 40 and 225 mg/m2 of trimetrexate and 5-FU, respectively, were well tolerated, without incidence of grade 4 toxicity. Diarrhea was the most common severe toxicity, with grades 3 and 4 occurring in 23% of all patients, 60% of these in the cohort receiving trimetrexate at a dose of 50 mg/m2. Given that the toxicity of this regimen correlates with trimetrexate
587 resistance in this disease [36], it is conceivable that the regimen described here combined with gemcitabine may result in additive or synergistic activity. The reproducible activity of regimens containing un-modulated PVI 5-FU [12–15, 37] and the results reported herein, support further exploration of PVI 5-FU combinations in pancreatic cancer.
Acknowledgements This work was supported in part by a grant from MedImmune Oncology, Gaithersburg, MD, USA.
References 1. Warshaw AL, Fernandez-del Castillo C. Pancreatic carcinoma. N Engl J Med 1992; 326: 455–465. 2. Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1999. CA Cancer J Clin 1999; 49: 8–31. 3. Ahlgren JD. Chemotherapy for pancreatic carcinoma. Cancer 1996; 78: 654–663. 4. Burris HA III, Moore MJ, Andersen J et al. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 1997; 15: 2403–2413. 5. Pazdur R, Ajani JJ, Abbruzzese JL et al. Phase II evaluation of fluorouracil and recombinant alpha-2a-interferon in previously untreated patients with pancreatic adenocarcinoma. Cancer 1992; 70: 2073– 2076. 6. Scheithauer W, Pfeffel F, Kornek G et al. A phase II trial of 5-fluorouracil, leucovorin, and recombinant alpha-2b-interferon in advanced adenocarcinoma of the pancreas. Cancer 1992; 70: 1864– 1866. 7. DeCaprio JA, Mayer RJ, Gonin R, Arbuck SG. Fluorouracil and high-dose leucovorin in previously untreated patients with advanced adenocarcinoma of the pancreas: results of a phase II trial. J Clin Oncol 1991; 9: 2128–2133. 8. Crown J, Casper ES, Botet J et al. Lack of efficacy of high-dose leucovorin and fluorouracil in patients with advanced pancreatic adenocarcinoma. J Clin Oncol 1991; 9: 1682–1686. 9. Cascinu S, Fedeli A, Fedeli SL, Catalano G. 5-Fluorouracil, leucovorin and interferon alpha 2b in advanced pancreatic cancer: a pilot study. Ann Oncol 1993; 4: 83–84. 10. Efficacy of intravenous continuous infusion of fluorouracil compared with bolus administration in advanced colorectal cancer. Metaanalysis Group In Cancer. J Clin Oncol 1998; 16: 301–308. 11. Sparano JA, Lipsitz S, Wadler S et al. Phase II trial of prolonged continuous infusion of 5-fluorouracil and interferon-alpha in patients with advanced pancreatic cancer. Eastern Cooperative Oncology Group Protocol 3292. Am J Clin Oncol 1996; 19: 546–551. 12. Maisey N, Chau I, Norman A et al. A randomized trial of protracted venous infusion 5-fluorouracil with or without mitomycin-C in advanced pancreatic cancer. Proc Am Soc Clin Oncol 2001; 20: 128a (Abstr 507). 13. Auerbach M, Wampler GL, Lokich JJ et al. Treatment of advanced pancreatic carcinoma with a combination of protracted infusional 5-fluorouracil and weekly carboplatin: a Mid-Atlantic Oncology Program Study. Ann Oncol 1997; 8: 439–444.
14. Nicolson M, Webb A, Cunningham D et al. Cisplatin and protracted venous infusion 5-fluorouracil (CF)—good symptom relief with low toxicity in advanced pancreatic carcinoma. Ann Oncol 1995; 6: 801– 804. 15. Hansen R, Quebbeman E, Ritch P et al. Continuous 5-fluorouracil (5FU) infusion in carcinoma of the pancreas: a phase II study. Am J Med Sci 1988; 295: 91–93. 16. Hidalgo M, Castellano D, Paz-Ares L et al. Phase I–II study of gemcitabine and fluorouracil as a continuous infusion in patients with pancreatic cancer. J Clin Oncol 1999; 17: 585–592. 17. Lin JT, Bertino JR. Update on trimetrexate, a folate antagonist with antineoplastic and antiprotozoal properties. Cancer Invest 1991; 9: 159–172. 18. Fulton B, Wagstaff AJ, McTavish D. Trimetrexate. A review of its pharmacodynamic and pharmacokinetic properties and therapeutic potential in the treatment of Pneumocystis carinii pneumonia. Drugs 1995; 49: 563–576. 19. Kamen B. Folate and antifolate pharmacology. Semin Oncol 1997; 24 (Suppl 18): 30–39. 20. Elliott WL, Howard CT, Dykes DJ, Leopold WR. Sequence and schedule-dependent synergy of trimetrexate in combination with 5-fluorouracil in vitro and in mice. Cancer Res 1989; 49: 5586–5590. 21. Conti JA, Kemeny N, Seiter K et al. Trial of sequential trimetrexate, fluorouracil, and high-dose leucovorin in previously treated patients with gastrointestinal carcinoma. J Clin Oncol 1994; 12: 695–700. 22. Romanini A, Li WW, Colofiore JR, Bertino JR. Leucovorin enhances cytotoxicity of trimetrexate/fluorouracil, but not methotrexate/ fluorouracil, in CCRF-CEM cells. J Natl Cancer Inst 1992; 84: 1033– 1038. 23. Leopold WR, Dykes DJ, Griswold DP Jr. Therapeutic synergy of trimetrexate (CI-898) in combination with doxorubicin, vincristine, cytoxan, 6-thioguanine, cisplatin, or 5-fluorouracil against intraperitoneally implanted P388 leukemia. NCI Monogr 1987; 5: 99– 104. 24. Sobrero A, Romanini A, Russello O et al. Sequence-dependent enhancement of HCT-8 cell kill by trimetrexate and fluoropyrimidines: implications for the mechanism of this interaction. Eur J Cancer Clin Oncol 1989; 25: 977–982. 25. Blanke CD, Kasimis B, Schein P et al. Phase II study of trimetrexate, fluorouracil, and leucovorin for advanced colorectal cancer. J Clin Oncol 1997; 15: 915–920. 26. Aschele C, Sobrero A, Faderan MA, Bertino JR. Novel mechanism(s) of resistance to 5-fluorouracil in human colon cancer (HCT-8) sublines following exposure to two different clinically relevant dose schedules. Cancer Res 1992; 52: 1855–1864. 27. Sobrero AF, Aschele C, Guglielmi AP et al. Synergism and lack of cross-resistance between short-term and continuous exposure to fluorouracil in human colon adenocarcinoma cells. J Natl Cancer Inst 1993; 85: 1937–1944. 28. Sobrero AF, Aschele C, Bertino JR. Fluorouracil in colorectal cancer—a tale of two drugs: implications for biochemical modulation. J Clin Oncol 1997; 15: 368–381. 29. Danenberg PV, Danenberg KD. Effect of 5, 10-methylenetetrahydrofolate on the dissociation of 5-fluoro-2′-deoxyuridylate from thymidylate synthetase: evidence for an ordered mechanism. Biochemistry 1978; 17: 4018–4024. 30. Romanini A, Lin JT, Niedzwiecki D et al. Role of folylpolyglutamates in biochemical modulation of fluoropyrimidines by leucovorin. Cancer Res 1991; 51: 789–793.
588 31. Fernandes DJ, Bertino JR. 5-Fluorouracil-methotrexate synergy: enhancement of 5-fluorodeoxyridylate binding to thymidylate synthase by dihydropteroylpolyglutamates. Proc Natl Acad Sci USA 1980; 77: 5663–5667. 32. Fanucchi MP, Walsh TD, Fleisher M et al. Phase I and clinical pharmacology study of trimetrexate administered weekly for three weeks. Cancer Res 1987; 47: 3303–3308. 33. Lin JT, Cashmore AR, Baker M et al. Phase I studies with trimetrexate: clinical pharmacology, analytical methodology, and pharmacokinetics. Cancer Res 1987; 47: 609–616. 34. Carlson RW, Doroshow JH, Odujinrin OO et al. Trimetrexate in locally advanced or metastatic adenocarcinoma of the pancreas. A
phase II study of the Northern California Oncology Group. Invest New Drugs 1990; 8: 387–389. 35. Shiomoto GM, Kuo S, Wade J et al. Phase II study of trimetrexate glucuronate in the treatment of advanced carcinoma of the pancreas. Proc Am Soc Clin Oncol 1992; 11: 184 (Abstr 549). 36. Rothenberg ML, Moore MJ, Cripps MC et al. A phase II trial of gemcitabine in patients with 5-FU-refractory pancreas cancer. Ann Oncol 1996; 7: 347–353. 37. Rothman H, Cantrell JE Jr, Lokich J et al. Continuous infusion 5-fluorouracil plus weekly cisplatin for pancreatic carcinoma. A Mid-Atlantic Oncology Program study. Cancer 1991; 68: 264–268.
Original article
Low-dose trimetrexate glucuronate and protracted 5-fluorouracil infusion in previously untreated patients with advanced pancreatic cancer R. G. Amado*, L. S. Rosen, J. R. Hecht, L.-S. Lin & P. J. Rosen Division of Hematology/Oncology and Jonsson Comprehensive Cancer Center, University of California, Los Angeles, CA, USA Received 18 June 2001; revised 26 September 2001; accepted 16 October 2001
Background: 5-Fluorouracil (5-FU)-based regimens have not been shown to prolong survival or provide clinical benefit in patients with advanced pancreatic cancer. The purpose of this study was to determine the tolerability of protracted venous infusion (PVI) of 5-FU, modulated by a low dose of the synthetic antifolate trimetrexate, in patients with advanced pancreatic cancer. Patients and methods: Twenty-three chemotherapy-naïve patients were evaluated. Patients were enrolled in four consecutive cohorts in which the weekly dose of trimetrexate was escalated in 10 mg/m 2 increments, from 20 to 50 mg/m2. PVI 5-FU was administered at a fixed dose of 225 mg/m2/day. Treatment was administered for 6 successive weeks, every 8 weeks. Results: Twenty-two patients were assessable. The maximum tolerated dose of trimetrexate was 40 mg/m2. The most common grade 3 and 4 toxicity was diarrhea. There were no treatment-related deaths. Preliminary analysis of activity revealed a response rate of 9%, with 41% of the patients having stable disease for a median duration of 3.8 months. The median survival for the entire group was 6.9 months (range 1–29 months). A clinical benefit response was experienced by 27.2% of patients. Conclusions: Low-dose trimetrexate can be safely administered in combination with PVI 5-FU. This treatment is well tolerated and is associated with palliative activity in advanced pancreatic cancer. Key words: antifolates, 5-fluorouracil, 5-FU modulation, infusional 5-FU, pancreatic cancer, trimetrexate
Introduction Pancreatic cancer is a lethal disease, for which the current treatment is of limited value. Over 80% of patients present with locally advanced or metastatic disease, a situation where current therapy results in only modest clinical benefit, without a major impact on survival [1, 2]. 5-Fluorouracil (5-FU) administered in bolus form is the most extensively studied drug. Objective responses to bolus 5-FU occur in <10% of patients, without an impact on overall survival or quality of life [3–9]. Protracted venous infusion (PVI) of 5-FU allows for higher dose intensity than bolus administration, is associated with a different toxicity profile and seems to have a higher response rate and a marginal survival benefit in colorectal cancer patients [10]. Doses of 200–300 mg/m2/day of 5-FU have been employed in advanced pancreatic cancer as a single agent or in combinations [11–16]. Response rates to PVI 5-FU have ranged from 8 to 19%.
*Correspondence to: Dr R. G. Amado, UCLA School of Medicine, Division of Hematology/Oncology, 11-964 Factor Building, Los Angeles, CA 90095-1678, USA. Tel: +1-310-825-8419; Fax: +1-310-825-6192; E-mail: [email protected] © 2002 European Society for Medical Oncology
Trimetrexate is a synthetic folate analog, which potently inhibits dihydrofolate reductase (DHFR) [17–19]. As a result of inhibition of purine synthesis, trimetrexate increases intracellular levels of phosphoribosyl pyrophosphate (PRPP), a cofactor for the enzyme orotate phosphoribosyltransferase, which is required for the intracellular conversion of 5-FU to active fluorinated nucleotides. These fluoronucleotides are incorporated into RNA, thus potentiating 5-FU cytotoxicity [20, 21]. Trimetrexate has been reported to enhance the cytotoxicity of 5-FU and folinic acid in vitro [22] in a scheduledependent manner [20, 23, 24]. Based on these preclinical results, sequential administration of trimetrexate has been studied in patients with colorectal cancer using schedules that employ bolus 5-FU [21, 25]. It is unknown whether trimetrexate-mediated 5-FU modulation occurs using PVI 5-FU. As indicated, increased levels of PRPP mediate RNA and protein synthesis inhibition. However, the putative primary mechanism of action of PVI 5-FU is inhibition of thymidylate synthase (TS), rather than RNA synthesis [26–28]. Several lines of evidence lend support to the possibility of synergy between trimetrexate and PVI 5-FU, based on potentiation of TS inhibition in the absence of folinic
583 acid. First, elevation of intracellular PRPP results in increased levels of 5-fluorodeoxyuridine monophosphate (FdUMP), the 5-FU nucleotide involved in TS inhibition. Secondly, inhibition of DHFR by methotrexate is associated with an increase in intracellular levels of dihydrofolate polyglutamates [29]. These metabolites, like folinic acid [30], can complex with TS and FdUMP to enhance TS inhibition [31]. Based on the activity and favorable toxicity profile of PVI 5-FU in phase II studies, which suggests synergy between trimetrexate and 5-FU, and on the above rationale for a potential modulatory effect on PVI 5-FU by trimetrexate, we conducted this dose escalation trial in previously untreated patients with advanced pancreatic cancer. The main objectives of this study were to assess toxicity and to determine maximum tolerated dose. Secondary objectives were response rate, time to treatment failure, survival and clinical benefit.
Patients and methods Patients Patients ≥18 years of age, with histologically confirmed adenocarcinoma of the pancreas, either locally advanced (unresectable stages II and III) or metastatic (stage IV) were eligible. A Karnofsky performance status (KPS) of ≥60% was required. Patients who had lost >10% of their weight within the previous 2 months were ineligible. Patients who had received previous chemotherapy for advanced metastatic pancreatic cancer were also excluded, but adjuvant chemoradiotherapy was allowed. All patients had bi-dimensionally measurable disease. Patients who had received previous radiotherapy were required to have measurable disease outside the radiation port, or measurable progressive disease in the radiation port. An adequate baseline bone marrow function was necessary, defined as an absolute neutrophil count (ANC) ≥1500/mm3 and a platelet count ≥100 000/mm3. Adequate renal function (defined as serum creatinine ≤2.0 mg/dl) and hepatic function (defined as serum bilirubin ≤2.0 mg/dl, serum albumin ≥2.5 mg/dl and AST ≤3 × the upper limit of normal in the absence of liver metastasis, or ≤5 × the upper limit of normal in the presence of liver metastasis) were also required. Informed consent was obtained before study entry. This study was performed with the approval of the local human subject protection committee.
Treatment All patients were treated on an outpatient basis. Patients underwent placement of a central venous catheter for PVI 5-FU infusion. Patients were entered into one of four cohorts to receive trimetrexate glucuronate at 20, 30, 40 or 50 mg/m2 weekly. As folinic acid can exacerbate the toxicity associated with PVI 5-FU, we used lower doses of trimetrexate than those previously employed [21, 25] to avoid the need for folinic acid rescue. We performed dose escalation after treatment of five patients in a prior cohort had been completed without excessive toxicity. In the event that three patients in any given cohort experienced grade 3 or 4 toxicity, dose escalation was terminated and all remaining patients were to be enrolled at the preceding dose level. All patients remained in their cohort of entry until discontinuation from the study. A treatment cycle was defined as 8 weeks and included 6 weeks of chemotherapy followed by a 2-week rest period. Treatment cycles were repeated until either disease progression or unacceptable toxicity supervened.
Trimetrexate glucuronate was supplied by MedImmune Oncology (Gaithersburg, MD, USA) in lyophilized powder (25 mg/vial). The drug was reconstituted and diluted, with 5% Dextrose Water for Injection USP, to obtain concentrations of 0.25 mg/ml to 2 mg/ml. Patients received 20–50 mg/m2 of trimetrexate i.v. over 60 min on days 1, 8, 15, 22, 29 and 36 of every 8-week cycle. Patients were to receive folinic acid rescue in the event of grade 4 hematological toxicity. After completion of trimetrexate infusion on day 1, patients received treatment with 5-FU at a dose of 225 mg/m2/day administered as a continuous i.v. infusion during the first 6 weeks of every 8-week cycle.
Dose modifications All toxicities were graded using the NCI Common Toxicity Criteria version 1. Patients were evaluated weekly for toxicity. In the event of a grade 2 hematological toxicity, the dose of trimetrexate was reduced by 25%. For grade 3 or 4 hematological toxicity, both trimetrexate and 5-FU were withheld until blood counts returned to ANC >2000/mm3 and platelets >100 000/mm3, then the dose of trimetrexate was reduced by 25% for grade 3 and 50% for grade 4, and 5-FU was resumed at a dose of 225 mg/m2. For grade 2, or greater, mucosal ulceration or diarrhea, 5-FU and trimetrexate were discontinued until symptom resolution, and then the doses were resumed as follows: unmodified for grade 2; reduced by 10% for 5-FU and 25% for trimetrexate in the event of grade 3; and by 20 and 50%, respectively, in the event of grade 4. For hand–foot syndrome (including pain, erythema, swelling and tenderness in hands and/or feet), both agents were discontinued and treatment was resumed after resolution of symptoms with full dose trimetrexate, 90% dose-reduced 5-FU for grade 3 and 80% for grade 4. The duration of the cycles was not prolonged to compensate for toxicity-related treatment interruptions. Patients whose doses were withheld for 3 consecutive weeks because of any toxicity were terminated from the study.
Assessment of efficacy Standard radiographic criteria for response were used. Survival and time to treatment failure were assessed prospectively. Clinical benefit analysis was performed using a modified version of methods previously described [4], in which weight was considered a primary measure along with pain and functional status, measured using the KPS scale. Data on these parameters were collected weekly. The pain parameter was a composite of information obtained on pain score and analgesic consumption. Information on pain score was obtained weekly from a brief pain inventory. Records of analgesic consumption were normalized to milligrams of oral morphine sulfate over a 1-week period. KPS was assessed by two observers. Patients were classified as positive, stable or negative for each of the three parameters. To be considered as positive or negative in any parameter, improvements or deteriorations had to be sustained for at least 28 days. Pain was scored as positive if it improved ≥50% from baseline (assuming an initial minimum pain score of ≥2 on a 0–10 scale), as negative if it worsened over baseline and as stable for all other results. Analgesic consumption was scored as positive if it decreased ≥50% from baseline (assuming an initial minimum of ≥5 mg of morphine), as negative if it increased from baseline and as stable if any other result. If both pain score and analgesic consumption were positive, or if one was positive and the other remained stable, the patient was classified as positive for clinical pain score. If either or both were negative, the patient was classified as a non-responder and if both remained unchanged the patient was classified as stable. Weight parameter was scored as positive if dry weight increased by ≥7% from baseline, as negative if it decreased by ≥4% and stable for all other results. KPS score was considered as positive if it increased by
584 ≥20%, as negative if it decreased by ≥20% and as stable for all other results. Patients were classified as positive for overall clinical benefit if at least one of the parameters improved without deterioration of the others and as non-responder if all three parameters remained stable or if any deteriorated.
Statistical methods Patient survival and time to disease progression were measured in days from study entry and data were summarized by Kaplan–Meier survivor function estimates and associated confidence intervals. Cox proportional hazards regression was used to assess the effects of trimetrexate dose, treatment time and baseline patient characteristics, on survival and time to progressive disease. Differential incidence of toxicity events across dose levels was analyzed by the χ2 test of equivalent incidence, and associations between trimetrexate dose and toxicity grade were analyzed by the χ2 test of association.
Results Patient demographics
Table 1. Patient characteristics Characteristics
n
Patients enrolled
23
Assessable for response/toxicity
22
95
Age (years) Median
66
Range
40–80
Gender Male Female
16
70
7
30
1
5
Karnofsky performance status 60% 70%
0
0
80%
11
50
90%
8
36
100%
2
8
1
4
Stage III
4
17
Stage IV
19
83
Pancreas
20
87
Liver
14
61
3
13
Peritoneum
3
13
Lungs
2
9
Other
1
4
1
8
35
2
11
48
3
4
17
Well differentiated
4
17
Moderately differentiated
6
26
Poorly differentiated
8
35
Unknown
5
22
20
5
22
30
5
22
40
6
26
50
7
30
Prior adjuvant therapy
From February 1998 to August 1999, 23 patients were enrolled in this university-based multicenter study in which nine community oncology sites participated. One patient enrolled in the 50 mg/m2 trimetrexate dose cohort withdrew before receiving treatment and is not included in the analysis. All other patients were included in an intention-to-treat analysis, including one patient, enrolled in the 40 mg/m 2 cohort, who voluntarily discontinued therapy after 1 week on the study. The total number of cycles administered was 45 (median 1, range 1–8) over four dose levels of trimetrexate. Patient characteristics and treatment groups are shown in Table 1.
% patients
Disease stage
Disease localization
Lymph nodes
Number of sites involves
Toxicity Seven patients required dose modifications because of toxicity; in four of them toxicity was grade 3 or 4. One patient (a radiographic and clinical benefit responder) had to be withdrawn from the study because of repeated dose reductions and treatment discontinuations for grade 2 mucositis and hand– foot syndrome. This patient was enrolled on the trimetrexate cohort receiving 20 mg/m2. No other toxicity-related study discontinuations occurred. Out of 230 administered doses of trimetrexate and 5-FU in this study, 6% of the doses were withheld because of toxicity, and dose modifications were implemented in 7.3% of the trimetrexate doses and 3.5% of the 5-FU doses. Dose reductions or discontinuations occurred in 21% of patients in cohorts receiving 20–40 mg/m2 trimetrexate, and 50% of patients in the 50 mg/m2 cohort. There were no treatment-related deaths. Grade 3 and 4 toxicity is shown in Table 2. The most common toxicity event leading to drug modification or discontinuation was diarrhea. Diarrhea was also the most common grade 3 and 4 toxicity overall (23% of all patients). Three patients, or 50% of the cohort treated at the 50 mg/m2 dose, experienced grade 4
Pathology
2
Trimetrexate dose (mg/m )
diarrhea. Therefore, diarrhea was considered the dose-limiting toxicity. One patient in the 50 mg/m2 group experienced a non-neutropenic septic episode that resolved with antibiotic treatment. There were no catheter-related thrombotic epi-
585 Table 2. Grade 3 and 4 treatment-related toxicity Toxicity
Trimetrexate dose level (mg/m2/week) 20 n
30 % cohort
40
% total
n
% cohort
% total
n
50 % cohort
% total
n
% cohort
% total
Asthenia Grade 3
0
0
0
1
20
5
0
0
0
1
17
5
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Nausea and vomiting Grade 3
0
0
0
0
0
0
1
17
5
1
17
5
Grade 4
1
20
5
0
0
0
0
0
0
0
0
0
Grade 3
0
0
0
0
0
0
1
17
5
0
0
0
Grade 4
1
20
5
0
0
0
0
0
0
3
50
14
Grade 3
0
0
0
0
0
0
0
0
0
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
1
17
5
Grade 3
0
0
0
0
0
0
0
0
0
1
17
5
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Grade 3
0
0
0
1
20
0
0
0
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
1
17
5
Grade 3
0
0
0
0
0
0
1
17
5
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Diarrhea
Abdominal pain
Anorexia
Infection
Cutaneous
Neutropenia Grade 3
0
0
0
0
0
0
1
17
5
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Grade 3
0
0
0
0
0
0
1
17
5
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Grade 3
0
0
0
0
0
0
0
0
0
0
0
0
Grade 4
0
0
0
0
0
0
0
0
0
0
0
0
Anemia
Thrombocytopenia
sodes, grades 3 or 4, stomatitis, hand–foot syndrome or hypersensitivity reactions. The most common grade 1 and 2 nonhematological toxicities were nausea and vomiting (55%), cutaneous (50%), stomatitis (45%), diarrhea (32%) and hand– foot syndrome (27%). With regard to hematological toxicity, grades 1 and 2 neutropenia, anemia and thrombocytopenia occurred in 9, 27 and 27% of patients, respectively, across all four cohorts. There were no grade 4 hematological toxicity events. One patient in the 40 mg/m2 cohort experienced grade 3 anemia and neutropenia. When all toxicity categories were analyzed together their grade was found to increase significantly with the dose of
trimetrexate (P = 0.008). In addition, the total number of toxic events, irrespective of their grade, tended to increase with the dose of trimetrexate (P = 0.062). The maximum tolerated and recommended dose of trimetrexate for future investigation using this combination schema dose was determined to be 40 mg/m2.
Activity Twenty-two patients were assessable for radiographic response. Two patients experienced a partial response (response rate 9%) and nine patients (41%) had stable disease for at least 8 weeks. The two responses lasted 3 and 3.9
586
Figure 1. Time to disease progression for all patients and for the metastatic disease patient subgroup.
Figure 2. Kaplan–Meier survival curves for all patients and for the metastatic disease patient subgroup.
months, occurring in one patient with metastatic disease and in another with locally advanced disease, in the 20 mg/m2 and 40 mg/m2 cohorts, respectively. The median duration of stable disease was 3.8 months. The median time to disease progression for the entire group was 2 months (Figure 1). The median survival for the entire group was 6.9 months (range 1–29 months) and the 1-year survival was 27% (Figure 2). For patients with stage IV disease, median and 1-year survivals were 6.7 months and 26%, respectively. No correlation was observed between baseline patient characteristics or trimetrexate dose, and time to treatment failure or survival. At the time of this analysis, the two responders remain alive at 32 and 25 months from initiation of therapy. Six patients fulfilled the criteria for clinical benefit response, for a total response rate of 27.2%. Three patients fulfilled these criteria on the basis of improvements in pain scores with stable weight and KPS scores, one had an improvement in pain with the other two parameters remaining stable during one cycle, followed on the subsequent cycle by an improvement in weight with stable pain and KPS scores, and two patients qualified for a response on the basis of an improvement in weight with stable pain and KPS scores. Of the six patients, one was classified as a partial radiographic responder and the remaining five patients had stable disease. The second radiographic responder did not qualify for a clinical benefit response as all three parameters remained stable throughout the duration of therapy.
dose and biochemical evidence of modulation exists for relatively low doses of this drug [21], we have chosen 40 mg/m2 as the dose of trimetrexate for future investigations. The relative paucity of severe toxicity using doses of trimetrexate below 50 mg/m2 contrasts with that of previously reported results from phase I and II studies using higher doses of trimetrexate as a single agent. In those studies, the major toxicity of trimetrexate was myelosuppression [17, 32, 33]. The toxicity profile of the current study also compares favorably with that of previously reported double modulation 5-FU schedules with folinic acid rescue [21, 25]. The current treatment schema is simpler to administer, as treatments are scheduled only once weekly and do not require oral folinic acid to rescue trimetrexate toxicity. In the present study, response seemed to correlate with survival; the two patients that experienced a radiographic response (one with metastatic and one with locally advanced disease) remain alive at 32 and 25 months from initiation of therapy. As a secondary objective of this study, we sought to obtain preliminary evidence of clinical improvement, because symptom control is an important endpoint in the treatment of advanced pancreatic cancer. Our analysis showed a clinical benefit response rate of 27%. However, because the main endpoint of this study was dose finding, with patients treated across a range of trimetrexate doses, these efficacy results should be considered preliminary. It should also be noted that, although the study was opened to community oncologists, reducing the possibility of selection bias, extrapolating these toxicity and activity results to routine clinical practice should take into account that most patients had well-preserved performance status, a relatively uncommon situation in patients presenting with advanced pancreatic cancer. It is not known whether trimetrexate-mediated 5-FU modulation resulted in enhanced activity over that obtained with unmodulated PVI 5-FU. A randomized study would be required to establish the role of trimetrexate in this treatment schedule. Any benefit of the combination of trimetrexate and 5-FU is unlikely to be derived from a direct cytotoxic action of trimetrexate, as previous studies indicate that trimetrexate, as a single agent, is inactive against pancreatic cancer [34, 35]. Given that 5-FU and gemcitabine do not exhibit cross-
Discussion We undertook this study to determine the maximum tolerated dose and the toxicity of the combination of low-dose trimetrexate and PVI 5-FU without folinic acid in patients with advanced pancreatic cancer. To our knowledge, this study represents the first report of PVI 5-FU modulation using an antifolate. Doses of 40 and 225 mg/m2 of trimetrexate and 5-FU, respectively, were well tolerated, without incidence of grade 4 toxicity. Diarrhea was the most common severe toxicity, with grades 3 and 4 occurring in 23% of all patients, 60% of these in the cohort receiving trimetrexate at a dose of 50 mg/m2. Given that the toxicity of this regimen correlates with trimetrexate
587 resistance in this disease [36], it is conceivable that the regimen described here combined with gemcitabine may result in additive or synergistic activity. The reproducible activity of regimens containing un-modulated PVI 5-FU [12–15, 37] and the results reported herein, support further exploration of PVI 5-FU combinations in pancreatic cancer.
Acknowledgements This work was supported in part by a grant from MedImmune Oncology, Gaithersburg, MD, USA.
References 1. Warshaw AL, Fernandez-del Castillo C. Pancreatic carcinoma. N Engl J Med 1992; 326: 455–465. 2. Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1999. CA Cancer J Clin 1999; 49: 8–31. 3. Ahlgren JD. Chemotherapy for pancreatic carcinoma. Cancer 1996; 78: 654–663. 4. Burris HA III, Moore MJ, Andersen J et al. Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: a randomized trial. J Clin Oncol 1997; 15: 2403–2413. 5. Pazdur R, Ajani JJ, Abbruzzese JL et al. Phase II evaluation of fluorouracil and recombinant alpha-2a-interferon in previously untreated patients with pancreatic adenocarcinoma. Cancer 1992; 70: 2073– 2076. 6. Scheithauer W, Pfeffel F, Kornek G et al. A phase II trial of 5-fluorouracil, leucovorin, and recombinant alpha-2b-interferon in advanced adenocarcinoma of the pancreas. Cancer 1992; 70: 1864– 1866. 7. DeCaprio JA, Mayer RJ, Gonin R, Arbuck SG. Fluorouracil and high-dose leucovorin in previously untreated patients with advanced adenocarcinoma of the pancreas: results of a phase II trial. J Clin Oncol 1991; 9: 2128–2133. 8. Crown J, Casper ES, Botet J et al. Lack of efficacy of high-dose leucovorin and fluorouracil in patients with advanced pancreatic adenocarcinoma. J Clin Oncol 1991; 9: 1682–1686. 9. Cascinu S, Fedeli A, Fedeli SL, Catalano G. 5-Fluorouracil, leucovorin and interferon alpha 2b in advanced pancreatic cancer: a pilot study. Ann Oncol 1993; 4: 83–84. 10. Efficacy of intravenous continuous infusion of fluorouracil compared with bolus administration in advanced colorectal cancer. Metaanalysis Group In Cancer. J Clin Oncol 1998; 16: 301–308. 11. Sparano JA, Lipsitz S, Wadler S et al. Phase II trial of prolonged continuous infusion of 5-fluorouracil and interferon-alpha in patients with advanced pancreatic cancer. Eastern Cooperative Oncology Group Protocol 3292. Am J Clin Oncol 1996; 19: 546–551. 12. Maisey N, Chau I, Norman A et al. A randomized trial of protracted venous infusion 5-fluorouracil with or without mitomycin-C in advanced pancreatic cancer. Proc Am Soc Clin Oncol 2001; 20: 128a (Abstr 507). 13. Auerbach M, Wampler GL, Lokich JJ et al. Treatment of advanced pancreatic carcinoma with a combination of protracted infusional 5-fluorouracil and weekly carboplatin: a Mid-Atlantic Oncology Program Study. Ann Oncol 1997; 8: 439–444.
14. Nicolson M, Webb A, Cunningham D et al. Cisplatin and protracted venous infusion 5-fluorouracil (CF)—good symptom relief with low toxicity in advanced pancreatic carcinoma. Ann Oncol 1995; 6: 801– 804. 15. Hansen R, Quebbeman E, Ritch P et al. Continuous 5-fluorouracil (5FU) infusion in carcinoma of the pancreas: a phase II study. Am J Med Sci 1988; 295: 91–93. 16. Hidalgo M, Castellano D, Paz-Ares L et al. Phase I–II study of gemcitabine and fluorouracil as a continuous infusion in patients with pancreatic cancer. J Clin Oncol 1999; 17: 585–592. 17. Lin JT, Bertino JR. Update on trimetrexate, a folate antagonist with antineoplastic and antiprotozoal properties. Cancer Invest 1991; 9: 159–172. 18. Fulton B, Wagstaff AJ, McTavish D. Trimetrexate. A review of its pharmacodynamic and pharmacokinetic properties and therapeutic potential in the treatment of Pneumocystis carinii pneumonia. Drugs 1995; 49: 563–576. 19. Kamen B. Folate and antifolate pharmacology. Semin Oncol 1997; 24 (Suppl 18): 30–39. 20. Elliott WL, Howard CT, Dykes DJ, Leopold WR. Sequence and schedule-dependent synergy of trimetrexate in combination with 5-fluorouracil in vitro and in mice. Cancer Res 1989; 49: 5586–5590. 21. Conti JA, Kemeny N, Seiter K et al. Trial of sequential trimetrexate, fluorouracil, and high-dose leucovorin in previously treated patients with gastrointestinal carcinoma. J Clin Oncol 1994; 12: 695–700. 22. Romanini A, Li WW, Colofiore JR, Bertino JR. Leucovorin enhances cytotoxicity of trimetrexate/fluorouracil, but not methotrexate/ fluorouracil, in CCRF-CEM cells. J Natl Cancer Inst 1992; 84: 1033– 1038. 23. Leopold WR, Dykes DJ, Griswold DP Jr. Therapeutic synergy of trimetrexate (CI-898) in combination with doxorubicin, vincristine, cytoxan, 6-thioguanine, cisplatin, or 5-fluorouracil against intraperitoneally implanted P388 leukemia. NCI Monogr 1987; 5: 99– 104. 24. Sobrero A, Romanini A, Russello O et al. Sequence-dependent enhancement of HCT-8 cell kill by trimetrexate and fluoropyrimidines: implications for the mechanism of this interaction. Eur J Cancer Clin Oncol 1989; 25: 977–982. 25. Blanke CD, Kasimis B, Schein P et al. Phase II study of trimetrexate, fluorouracil, and leucovorin for advanced colorectal cancer. J Clin Oncol 1997; 15: 915–920. 26. Aschele C, Sobrero A, Faderan MA, Bertino JR. Novel mechanism(s) of resistance to 5-fluorouracil in human colon cancer (HCT-8) sublines following exposure to two different clinically relevant dose schedules. Cancer Res 1992; 52: 1855–1864. 27. Sobrero AF, Aschele C, Guglielmi AP et al. Synergism and lack of cross-resistance between short-term and continuous exposure to fluorouracil in human colon adenocarcinoma cells. J Natl Cancer Inst 1993; 85: 1937–1944. 28. Sobrero AF, Aschele C, Bertino JR. Fluorouracil in colorectal cancer—a tale of two drugs: implications for biochemical modulation. J Clin Oncol 1997; 15: 368–381. 29. Danenberg PV, Danenberg KD. Effect of 5, 10-methylenetetrahydrofolate on the dissociation of 5-fluoro-2′-deoxyuridylate from thymidylate synthetase: evidence for an ordered mechanism. Biochemistry 1978; 17: 4018–4024. 30. Romanini A, Lin JT, Niedzwiecki D et al. Role of folylpolyglutamates in biochemical modulation of fluoropyrimidines by leucovorin. Cancer Res 1991; 51: 789–793.
588 31. Fernandes DJ, Bertino JR. 5-Fluorouracil-methotrexate synergy: enhancement of 5-fluorodeoxyridylate binding to thymidylate synthase by dihydropteroylpolyglutamates. Proc Natl Acad Sci USA 1980; 77: 5663–5667. 32. Fanucchi MP, Walsh TD, Fleisher M et al. Phase I and clinical pharmacology study of trimetrexate administered weekly for three weeks. Cancer Res 1987; 47: 3303–3308. 33. Lin JT, Cashmore AR, Baker M et al. Phase I studies with trimetrexate: clinical pharmacology, analytical methodology, and pharmacokinetics. Cancer Res 1987; 47: 609–616. 34. Carlson RW, Doroshow JH, Odujinrin OO et al. Trimetrexate in locally advanced or metastatic adenocarcinoma of the pancreas. A
phase II study of the Northern California Oncology Group. Invest New Drugs 1990; 8: 387–389. 35. Shiomoto GM, Kuo S, Wade J et al. Phase II study of trimetrexate glucuronate in the treatment of advanced carcinoma of the pancreas. Proc Am Soc Clin Oncol 1992; 11: 184 (Abstr 549). 36. Rothenberg ML, Moore MJ, Cripps MC et al. A phase II trial of gemcitabine in patients with 5-FU-refractory pancreas cancer. Ann Oncol 1996; 7: 347–353. 37. Rothman H, Cantrell JE Jr, Lokich J et al. Continuous infusion 5-fluorouracil plus weekly cisplatin for pancreatic carcinoma. A Mid-Atlantic Oncology Program study. Cancer 1991; 68: 264–268.