Schedule-selective biochemical modulation of 5-fluorouracil in advanced colorectal cancer – a phase II study

Background 5-fluorouracil remains the standard therapy for patients with advanced/metastatic colorectal cancer. Pre-clinical studies have demonstrated the biological modulation of 5-fluorouracil by methotrexate and leucovorin. This phase II study was initiated to determine the activity and toxicity of sequential methotrexate – leucovorin and 5-fluorouracil chemotherapy in patients with advanced colorectal cancer. Methods Ninety-seven patients with metastatic colorectal cancer were enrolled onto the study. Methotrexate – 30 mg/m2 was administered every 6 hours for 6 doses followed by a 2 hour infusion of LV – 500 mg/m2. Midway through the leucovorin infusion, patients received 5-fluorouracil – 600 mg/m2. This constituted a cycle of therapy and was repeated every 2 weeks until progression. Results The median age was 64 yrs (34–84) and the Eastern Cooperative Group Oncology performance score was 0 in 37%, 1 in 55% and 2 in 8% of patients. Partial and complete responses were seen in 31% of patients with a median duration of response of 6.4 months. The overall median survival was 13.0 months. The estimated 1-year survival was 53.7%. Grade III and IV toxic effects were modest and included mucositis, nausea and vomiting. Conclusions This phase II study supports previously reported data demonstrating the modest clinical benefit of 5-FU modulation utilizing methotrexate and leucovorin in patients with metastatic colorectal cancer. Ongoing studies evaluating 5-fluorouracil modulation with more novel agents (Irinotecan and/or oxaliplatin) are in progress and may prove encouraging.


Background
Although a number of recently introduced chemotherapeutic agents have demonstrated significant anti-tumor activity in advanced colorectal cancer, [1][2][3][4][5][6][7][8][9] standard therapy for patients with metastatic disease remains 5fluorouracil (5-FU) -based chemotherapy. [10] A vast series of pre-clinical and clinical studies have suggested 5-FU is a more active anti-tumor agent when modulated by a host of compounds including leucovorin, [11][12][13] methotrexate, [14] folinic acid, [15,16] N-phosphonacetyl-L-aspartic acid (PALA) [17][18][19][20][21][22] and recombinant interferon alfa-2a (IFNaα-2a). [23,24] Protracted venous infusion (PVI) 5-FU regimens have also resulted in increased response rates in some studies, but only results in a modest benefit in median survival. [25,26] Oral fluoropyrimidines and oral regimens using prodrugs of 5-FU or inhibitors of dihydropyrimidine dehydrogenase (DPD) which pharmacologically simulate the intravenous continuous infusion administration of 5-FU are currently under clinical evaluation. Although these studies are important and may demonstrate equivalency, it is not clear these oral compounds will result in a significant improvement compared to the intravenous infusion of 5-FU. [27] Potentiation of the anti-tumor activity of 5-FU by methotrexate (MTX) and leucovorin (LV) requires careful scheduling to achieve the most favorable interactions between these 3 drugs. In vitro and in vivo studies have both suggested a synergistic as well as antagonistic effect when 5-FU, MTX and LV are used in combination. [28][29][30][31][32] In this report, two modulators of 5-FU were used in a sequential manner, anticipating an enhancement of the efficacy of 5-FU before the development of 5-FU resistance and subsequent disease progression. The order of administration of the three agents was chosen to optimally exploit the reported differing mechanisms of biochemical modulation. The objectives of this phase II study were to determine the activity and toxicity of MTX/LV/5-FU in patients with advanced colorectal cancer.

Patients and Methods
Between August of 1987 and August of 1988, 97 patients with advanced colorectal cancer were enrolled onto this study. Informed consent was obtained in writing from all patients according to Institutional Review Board guidelines. Eligibility criteria included histologic proof of metastatic colorectal carcinoma and the presence of at least one bidimensionally measurable lesion on computed tomography scan. All patients were required to have an Eastern Cooperative Oncology (ECOG) Group performance score of 0 -2 and have a life expectancy greater than 2 months. Patients also were required to meet the following criteria: adequate bone marrow, renal, and hepatic function as evidenced by a white blood cell count = 4,000 cells/ul, platelet count = 100,000 cells/ul, creatinine = 2.0 mg/dl or creatinine clearance of 60 ml/min and a serum bilirubin = 1.5 mg/dl and no serious intercurrent medical or mental illness. Patients could not have received prior chemotherapy (excluding adjuvant chemotherapy).

Treatment Plan
Consistent with the traditional design of phase II studies, all patients received therapy as outlined in the treatment schema in figure 1. Methotrexate was administered at a dose of 30 mg/m 2 orally every 6 hours on an empty stomach for 6 doses. At hour 36 a 2 hour intravenous infusion of leucovorin was administered at a dose of 500 mg/m 2 . Midway through the leucovorin infusion, 5-FU at a dose of 600 mg/m 2 was given as a 30 minute intravenous bolus infusion. This represented a course of therapy and was repeated every two weeks. Dose modifications for hematologic toxicity were made on the basis of counts on the day of treatment. For a white blood count = 3,500 cells/ul or a platelet count = 100,000 platelets/ul -patients received full doses of all three drugs as described above. For a white blood count between 3,000 and 3,500 cells/ul and a platelet count = 100,000 platelets/ul, a 50% dose reduction of all three drugs was performed. For white blood counts less than 3,000 cells/ul and a platelet count < 100,000 platelets/ul -all therapy was held until the counts had fully recovered (white blood count = 3,500 cells/ul and a platelet count = 100,000). Patients experiencing a white blood count = 2,000 and/or a platelet count = 75,000 any time during therapy subsequently received 50% of the calculated dose even if the counts fully recovered.
Weekly complete blood counts, including platelet counts and white blood count differentials were conducted to determine the level of myelosuppresion. Before every other cycle of treatment (q 4 weeks), a chest x-ray and complete chemistry panel including electrolytes and serum creatinine were obtained. All lesions were measured bidimensionally and assessed for changes by either an imaging study (computed tomography, chest x-ray, ultrasound) or by clinical examination to determine response to therapy. Treatment was continued until disease progression became evident or side effects became intolerable.

Statistical Methods
This study was planned to recruit 50 evaluable patients, allowing us to estimate objective tumor response to within +/-15% with 95% confidence. Additional patients were subsequently accrued to increase the precision of our estimates of response and survival. The trial was conducted with one interim analysis after nine evaluable patients. If no patient had experienced an objective tumor response, the study would have been terminated. The two-stage design used in this study was first proposed by Ed Gehan in 1961 (Gehan E.A. The determination of the number of patients required in a preliminary and a follow-up trial of a new chemotherapeutic agent. J. Chron. Dis. 13:346-353, 1961). The probability of observing no responses was less than 5% if the true response was 30% or greater. Progression-free survival was defined as the time from study entry to the first observation of disease progression or death as a result of any cancer. Survival was defined as the time of study entry until death as a result of any cause. These timeto-event parameters were summarized using Kaplan-Meier product-limit estimates. Log rank tests were used to assess which factors were univariately predictive of the time to progression and survival, and the Cox proportional hazards regression model was used to assess which factors were jointly predictive of these outcomes. The statistical analysis was performed using the SAS package, version 6.12 (SAS Institute, Cary, NC).

Results
From August of 1987 through September of 1988, 97 patients were accrued to this study with two patients subsequently found to be ineligible. Data was not collected for the first patient due to a protocol violation and a second patient was found to have an ovarian primary cancer after enrollment. Patient characteristics for the 95 eligible patients are listed in table 1. The median age was 64. The median ECOG performance score was 1. All patients had a histologic diagnosis of adenocarcinoma. The major sites of metastatic disease were liver, lung, soft tissue and lymph node. None of the patients remain on study; 71 pa-tients progressed on study, 16 patients were removed from study either at the discretion of the treating physician or patient refusal to continue treatment, 3 patients were removed from study due to toxicity, 2 patients expired while receiving therapy. Three patients were removed for other reasons; one patient failed to adhere to the treatment schedule, one relocated, and a third patient never received therapy due to rapid medical decline.

Treatment and Response
A total of 1257 cycles of treatment were delivered during the trial (median 9 cycles; range 0 to 41 cycles). Table 2 reflects the average percent of ideal dose given at each selected cycle of treatment. Even after 25 cycles of therapy, at least 75% of the ideal doses for all drugs were being administered. Ninety-three of the 95 eligible patients enrolled onto the study were assessable for toxicity. Mucositis, nausea/vomiting and diarrhea were the most commonly observed toxicities and are listed in table 3. The hematologic toxicities are also shown in table 3 and in general were infrequently observed. Overall the regimen was well tolerated.

Figure 1 Treatment Schema
Of the 90 patients evaluable for response, ten patients (11%) demonstrated a complete response of their disease (median response duration 14.3 months). Remarkably, 3 of the 10 patients remain alive, 2 without evidence of disease with 51, 73 and 138 months of follow-up, respectively. An additional 18 patients (20%) attained partial responses; 16 have died of progressive disease while 2 patients are alive and without disease progression. Table 4 reflects the overall best response for the entire patient cohort.
The overall estimated median time to progression was 5.6 months. The estimated progression free survival at 12 months was 25.5% (standard error 4.5%). The overall estimated median survival time was 13.0 months. The estimated survival rate at 12 months was 53.7% (standard error 5.1%). Table 5 reflects overall survival for various subgroups of patients. The prognostic factors found to be significant on univariate analysis were ECOG perform-ance score, gender and number of metastatic sites. Female patients, patients with fewer metastatic sites and patients with an ECOG performance 0 had a statistically significant improved survival. Survival by metastatic site is shown in Table 6. Those patients without lung metastasis and those without intra-abdominal metastasis had longer survival times. In the multivariate analysis, presence of lymph nodes, lower LDH levels female sex, better performance status and fewer metastatic sites were statistically predictive of longer survival (Data not shown).

Discussion
The rationale, at least in part, for this phase II study comes from pre-clinical data suggesting MTX and LV potentiate the anti-tumor activity of 5-FU. The cytotoxic effects of 5-FU are mediated through inhibition of DNA and/or RNA synthesis. [33] DNA synthesis inhibition is mediated by the binding of the active 5-FU metabolite 5-fluoro-2'deoxyuridine -5-monophosphate (F-dUMP) to the enzyme    thymidylate synthetase (TS). This binding is enhanced by high concentrations of reduced folate cofactors. [34] Although this ternary complex is covalent, it dissociates with a half-life of 2-3 hours in the absence of excess 5, 10methylene tetrahydrofolate. Thus high levels of 5, 10methylene tetrahydrofolate derived from LV not only allows optimum ternary complex formation, but also prevents subsequent breakdown of the complex. [35] Additional in vitro studies suggest cells may need prolonged exposure to reduced folates in order to obtain maximum inhibition of TS by FdUMP. [36] A second observed means of potentiating the cytotoxic effects of 5-FU is through the pre-treatment of cells with MTX. MTX-induced enhancement of 5-phosphoribosyl-1-pyrophosphate (PRPP) pools with subsequent increased synthesis of 5-FU nucleotides and incorporation of 5-fluoriuridine -5'-triphosphate (FUTP) into RNA has been proposed as the basis for this biochemical modulation. [37][38][39] Subsequent in vivo studies using fluorine 19 magnetic resonance spectroscopy confirmed a three-fold increase in 5fluoronucleotide (FNuct), the active 5-FU metabolite, following the pre-treatment of animals with MTX [32]. These observed changes in 5-FU catabolism resulted in a greater anti-tumor activity than MTX given alone or MTX given after 5-FU. Consistent with these pre-clinical studies, the design of this phase II trial involved a fluorouracil infusion that was initiated one hour into the 2 hour high dose infusion of LV and 36 hours after the initial methotrexate dosing.  Our phase II study provides long-term efficacy and safety data for sequential MTX-LV-5-FU chemotherapy administered on a 14-day cycle. The 31% response rate and median survival duration (median 13.0 months) observed in this trial are slightly improved when compared to the results reported for intravenous 5-FU therapy. Overall, treatment with sequential MTX-LV-5-FU was very welltolerated. Grade III diarrhea was seen in only 8% of patients, grade III and IV mucositis in 8% and 1% of patients respectively and grade III and IV nausea/vomiting in 11% and 1% of patients respectively. The most common hematologic toxicity observed was grade II and III anemia in 18% and 4% of patients respectively.
Other similar studies have shown comparable results in terms of the efficacy and safety of MTX-LV-5-FU for patients with advanced colorectal cancer [44][45][46][47][48]. In a meta analysis of eight randomized trials of 5-FU/MTX versus 5-FU alone, the complete and partial response rates for the 5-FU patient cohort was 10% compared with 19% for patients receiving MTX/5-FU. [14] In addition, the median survival was improved for patients receiving MTX/5-FU, 10.7 months versus 9.1 months for patients receiving 5-FU alone (p = .024). Marsh et al. in a randomized trial, reported a statistical improvement in response, time to progression (9.9 months versus 5.9 months) and median survival (15.3 months versus 11.4 months) when the interval between the MTX and 5-FU infusion was increased from 1 hour to 24 hours.

Conclusions
We suggest our results and the pre-clinical and clinical data discussed indicate the sequence of administration of 5-FU with MTX and LV is important. Future studies combining 5-FU with other cytotoxic / biologic compounds should consider the mechanism of the interaction and incorporate that information into the design of the trial.

Authors' Contributions
DRW and PS were involved in the original design and accrual to this phase II clinical trial. The manuscript was prepared and written by SKT and AWB. The manuscript was then read and revised by members of the Gastrointestinal Research Group at Wake Forest University (SAM and VH). LDC wrote and performed all statistical analysis and preparation of the data presented in this manuscript.