Stop-Start Treatment Strategies

Perhaps the most interesting and important concept that was elaborated on at the 2006 meeting of the American Society of Clinical Oncology (ASCO) was the idea of “treatment holidays,” or “chemotherapy-free intervals.” The concept of treating until treatment failure or unacceptable toxicity has been widely accepted in the United States over the past several decades. More recently, data have suggested that interruptions in therapy—that can be beneficial to the patient, both physically and psychologically—do not interfere with long term outcome.

This concept was explored previously in the initial OPTIMOX trial, published by Tournigand et al. in the Journal of Clinical Oncology.[1] In this OPTIMOX I study, efficacy results were virtually the same with the FOLFOX 4 regimen given continuously versus the FOLFOX 7 regimen given for six cycles, followed by twelve cycles of biweekly 5FU and leucovorin alone, with the oxaliplatin omitted, and with planned reintroduction of the oxaliplatin after that twelve-cycle (24 week) interval. In that particular trial, the incidence of the actual reintroduction of oxaliplatin was relatively low. Nevertheless, the efficacy of the two arms was equivalent and the toxicity was lower with the planned oxaliplatin holiday. It seems clear that in terms of managing oxaliplatin neurotoxicity this sort of stop-start strategy is a necessary innovation, and should be routinely incorporated into standard care when using oxaliplatin in the metastatic setting.

In the OPTIMOX II trial, presented in abstract form at ASCO this year, the idea of a fully chemotherapy-free interval was explored.[2] In the control arm, patients received FOLFOX 7, with a modified reduced dose of oxaliplatin compared to the original FOLFOX 7, for six cycles, followed by biweekly 5FU and leucovorin until progression to baseline size of the tumor, and then reintroduction of oxaliplatin.  In the investigational arm, patients received six cycles of modified FOLFOX 7 followed by no maintenance therapy at all, until progression of disease to baseline measurements, and then reintroduction of modified FOLFOX 7.

The initial planned sample size was 600 patients, but this was necessarily downgraded to 200 when bevacizumab was approved in France. The primary objective of the study was to evaluate the duration of disease control (DDC). No formal statistical hypotheses were explored between the two arms given the limited size. That having been said, the progression-free survival appeared to be modestly superior in the OPTIMOX I arm (8.7 months vs. the OPTIMOX II arm) (6.9 months) P value=0.09. However, the duration of disease control was virtually the same (12.9 months vs. 11.7 months, P value = 0.41).

The authors’ conclusions were that maintenance therapy may improve progression free survival, but does appear to improve duration of disease control (DDC). They note that the median duration of chemotherapy-free interval was 4.6 months, and 8 months in patients with favorable prognostic factors. These results suggest that chemotherapy-free intervals may be considered for incorporation into routine treatment practice. Further ongoing studies are exploring the use of cytotoxics with biologics and evaluate stop start approaches while continuing biologic therapy.

Another interesting study that explored the idea of interrupted therapy was performed by Labianca et al. and the Italian GISCAD group.[3] This study looked at stop-start vs. continuous FOLFIRI. The control arm gave standard FOLFIRI every two weeks, and the investigational arm gave two months of FOLFIRI followed by a two month break followed by two months of FOLFIRI and then two months of break, and so on. The progression-free survival and the overall survival curves of the two arms appeared superimposable, again suggesting that providing chemotherapy-free intervals may be reasonable, and that these treatment breaks do not appear to cause significant detriment to the patient. The potential benefits of chemotherapy free intervals would be both a psychological and physical rest for the patient, as well as decreased expense due to less chemotherapy being administered.

Fluoropyrimidine Schedule of Administration

Another trial presented in the oral session of the 2006 colorectal ASCO program was the so-called BICC-C trial.[4] This trial compared FOLFIRI vs. a modified bolus IFL schedule (2 weeks on, 1 week off, rather than the original 4 weeks on, 2 weeks off) vs. capecitabine plus irinotecan (Cape-IRI), with the second randomization to plus/minus celecoxib. This trial was also cut short secondary to the commercial availability of bevacizumab, and a second phase of the trial compared FOLFIRI plus bevacizumab to modified IFL plus bevacizumab. The important take-home message from this trial was that progression-free survival was superior for FOLFIRI compared to either of the other two arms. The overall survival of FOLFIRI plus bevacizumab was superior to IFL plus bevacizumab with a p value of 0.01. Toxicity favored the FOLFIRI arm. The conclusion of this trial, which supports much previous data, is that the infusional 5FU regimen is superior to bolus 5FU when combined with irinotecan. The FOLFIRI regimen is the unequivocal regimen of choice for irinotecan plus fluoropyrimidine, and the bolus IFL schedule should be regarded as obsolete technology that should not be used. Furthermore, capecitabine in this trial was not an acceptable substitute for infusional 5FU when combined with irinotecan.

Combination Versus Sequential Use of Chemotherapies

Falcone et al. presented an abstract on their “FOLFOXIRI” study, which randomly compared a FOLFIRI control arm to FOLFOXIRI, a regimen administering irinotecan, oxaliplatin, leucovorin, and infusional 5FU (no bolus) all together on a biweekly schedule.[5] The results of this study were intriguing. Diarrhea, vomiting, neurotoxicity, and neutropenia were increased in the FOLFOXIRI arms, but were within tolerable limits, and no toxic deaths were reported on either arm of the trial. The sixty-day all-cause mortality was 1.6% in both arms. The major objective response rate was 60% in the FOLFOXIRI arm vs. 34% in the FOLFIRI arm and this was significant with a p-value of 0.0001. It was also noted that for patients with metastases confined to the liver only, R0 resections were accomplished in 36% of 39 patients, vs. 12% of 42 patients in the FOLFIRI arm. This difference was also statistically significant. The median progression-free survival was 6.9 months in the FOLFIRI arm and 9.8 months in the FOLFOXIRI arm (P = 0.0006) and the overall survival was 16.7 months in the FOLFIRI arm and 22.6 months in the FOLFOXIRI arm (P = 0 0.032). Although this relatively small trial is not sufficient to be practiced-changing, and represents use of cytotoxics prior to the incorporation of biologic agents, it is, nevertheless, intriguing. The favorable response rate and increased resectability rate raises consideration of the use of this approach for patients in whom a definitive surgical resection might be planned. It is noteworthy, however, that the overall toxicity seen on this study appears favorable compared to what might be expected in an American population based on combination experienced with other trials.

Other data shown at this meeting by Haller et al.,[6] indicating that the patients from the United States tolerate fluoropyrimidines more poorly than those from the rest of the world, may explain some of the potential issues. As such, toxicity data from non-American trials must be extrapolated to American patients with caution.

Bevacizumab Dosing Schedules

Several studies presented at the meeting considered the appropriate dose and administration schedule of bevacizumab. Giantonio et al. reported an interesting follow-up analysis of a subset of patients from the ECOG 3200 study.[7] This study, presented previously at ASCO in 2005, was a randomized trial in bevacizumab-naïve patients who had previously failed irinotecan and fluorouracil. The treatment arms were FOLFOX 4 vs. FOLFOX 4 plus bevacizumab vs. bevacizumab alone. The bevacizumab dose in this trial was 10 mg/kg repeated every other week. Of note, as reported in this particular abstract, a substantial number of patients received protocol-specified dose reductions from 10 mg/kg to 5 mg/kg. The outcome for patients undergoing these reductions was virtually the same as for those who did not undergo reductions. This raises important questions regarding the appropriate dose of bevacizumab. Currently 5 mg/kg is the widely accepted dose for first line treatment for metastatic colorectal cancer. Most clinicians thus far have continued to use the same dose in those patients who have not received prior bevacizumab and who are receiving it in second line. No data have yet been presented justifying the use of bevacizumab continuously in multiple lines of therapy and such an approach should be considered investigational at this time (note that the recent revision to the bevacizumab package insert states that the drug is approved for first or second line, not first and second line). Some have advocated that a 10 mg/kg dose should be used routinely in second line. These newly-presented data, however, appear to question whether or not that higher dose is necessary or beneficial. These data are hypothesis-generating data only however, and without randomized comparisons of the two doses, any conclusions would necessarily be conjectural.

Another study, from Saltz et al. at Memorial Sloan-Kettering Cancer Center, explored the administration schedule of bevacizumab.[8] When bevacizumab was first entering clinical development, the degree to which there was a risk of hypersensitivity reactions (HSRs) to this humanized monoclonal antibody was not known. With no direct experience to serve as a guide, the schedule that had been used with other monoclonal antibodies was adopted in first-in-man phase I studies. This caution, appropriate in the initial development of the drug, included a 90 minute infusion time for the first dose, and then decreasing infusion times if no HSRs were seen. Based on these initial Phase I studies, the first dose of bevacizumab had been given over 90 minutes; the second dose over 60 minutes and then subsequent doses over 30 minutes in virtually all clinical trials, and this schedule established in the phase I study and never re-explored, was incorporated into the package insert. 

The experience at Memorial Sloan-Kettering Cancer Center was reviewed and this indicated that no patients were having difficulty progressing to 30 minute infusion times, and that HSRs were not being observed. All non-protocol doses of bevacizumab at 5 mg/kg were therefore then given over 30 minutes, without the 90 minute and 60 minute run-ins. Because no difficulty was seen in 464 sequential patients treated with an initial 30 minute bevacizumab infusion dose, the possibility of shortening this dose further was then explored. It was noted that the standard bevacizumab doses at 15 mg/kg over 30 minutes that were being routinely used in breast cancer and lung cancer patients, as well as in an investigational study in upper GI malignancies, had an infusion rate of 0.5 mg/kg/minute. Since this rate was tolerable without HSRs being encountered, as of November 1, 2005, Memorial Sloan-Kettering Cancer Center adopted this as its standard institutional bevacizumab infusion rate. Thus, a 5 mg/kg dose given at 0.5 mg/kg/minute resulted in the dose of bevacizumab being administered over 10 minutes. Doses of 10 mg/kg were given over 20 minutes, and 15 mg/kg doses were given over 30 minutes. At the time of data cut off for this analysis (February 28, 2006), 259 patients had received a total of 1,059 doses of bevacizumab at 5 mg/kg over 10 minutes. A total of 185 patients had been receiving bevacizumab over 30 minute infusions and were converted to 10 minute infusions, with no hypersensitivity reactions reported. Seventy-four patients began therapy with their first dose over 10 minutes. No reactions were reported in the first dose. Two patients had clinically minor reactions on the second dose, neither of which involved shortness of breath, oxygen desaturation, wheezing, or hypotension, and both patients continued on bevacizumab.

The conclusion of the authors of this study was that a bevacizumab infusion rate of 0.5 mg/kg/minute is logical and remains the current standard practice at Memorial Sloan-Kettering Cancer Center, and in the patients reported showed no serious hypersensitivity reactions.

Cetuximab Studies

Scheduling issues of cetuximab were also explored in trials reported at the 2006 ASCO meeting. Tabrinaro et al. explored the pharmacokinetics involved in giving cetuximab on an every-other-week schedule.[9] These initial studies suggest that doses of 500 mg/m2 given every other week provide pharmacokinetics similar to the standard 250 mg/m2 weekly dosing schedule.

Dr. Alan Venook presented the initial results from the aborted CALGB trial C80203.[10] This trial had been scheduled to be a randomized study of FOLFIRI vs. FOLFOX with the 2X2 randomization plus/minus cetuximab. The trial was cut short due to the commercial availability of bevacizumab, which made accrual to a chemotherapy-only arm inappropriate. Thus, the sample size is much smaller than had been initially planned for, and the study has been redesigned and presented as a randomized Phase II trial. Looking at the response rate in the patients who received cetuximab with their chemotherapy vs. those that did not, Venook noted that chemotherapy plus cetuximab had a 52% response rate vs. chemotherapy alone at 38% (P=0.029). However the overall progression-free survival did not appear to be substantially different between the two arms. At present this study is insufficient to address the question of the appropriateness of front line cetuximab, and use of cetuximab in the front line setting remains investigational at this time.

Conclusion

To summarize then, the significant findings from the 2006 ASCO meeting regarding the management of colorectal cancer:

• Chemotherapy for metastatic colorectal cancer can and probably should be given with planned treatment interval holidays.
• Bolus IFL (irinotecan, fluorouracil, leucovorin) should be regarded as an outdated regimen, and should not be used, and bolus 5FU should rarely if ever be used.
• Bevacizumab does not appear to require routine 90 and 60 minute infusions and appears to be safe at a rate of 0.5 mg/kg/minute for routine use, including the initial dose. 
• The pharmacokinetics of cetuximab given 500 mg/m2 every other week appears to support the use of this dose and schedule as an alternative treatment schedule for this agent. 
• Limited front line cetuximab data do show an improved response rate but do not show an improvement in progression free survival, and the study is both early in terms of data maturation and too small to permit a meaningful assessment of overall survival. Cetuximab front line therapy remains investigational at this time.