The Impact of Targeted Therapy on the Treatment of Colorectal Cancer

The treatment of colorectal cancer has truly undergone a dramatic revolution over the past 4 to 5 years. When one thinks back to a "5-FU-only" world-where the median survival of patients with metastatic colon cancer was on the order of 10 to 12 months and the benefit of adjuvant therapy was modest at best-and compares that to the world today, where we are relatively rich with therapeutic options, it is at times overwhelming. Even for those of us who are focused entirely on the management of gastrointestinal oncology, it is difficult to determine the best options for patients and if there truly are standards of care for patients with colorectal cancer. As we have evolved with the incorporation of new chemotherapy medicines, and now with new effective monoclonal antibodies, we have gone from a world of having single choices to having multiple choices for our patients accompanied by clear success, with median survivals now approaching 3 years. The focus of this article is to summarize some of the key advances of incorporating biologically targeted therapies into the routine management of patients with colorectal cancer. We will review the key findings from the 2005 American Society of Clinical Oncology (ASCO) annual meeting, and comment on our incorporation of these data into the management of our patients on a day-to-day basis. FOLFOX and Bevacizumab(Drug information on bevacizumab): ECOG 3200 When the pivotal trial of bevacizumab (Avastin) was designed for colorectal cancer, the standard of care was the irinotecan (Camptosar)/bolus fluorouracil(Drug information on fluorouracil) (5-FU) regimen known as IFL. However, by the time the trial was reported, the standard of care in the United States and much of the world was the use of either FOLFOX (5-FU, leucovorin, oxaliplatin(Drug information on oxaliplatin) [Eloxatin]) or FOLFIRI (5-FU, leucovorin, irinotecan) as front-line treatment for metastatic colon cancer. Fortunately, the US Food and Drug Administration (FDA) approval of bevacizumab allowed for the incorporation of this agent into any intravenous 5-FU regimen for colorectal cancer. Interestingly, even though there were significant safety data in breast cancer combining bevacizumab with capecitabine(Drug information on capecitabine) (Xeloda), the specific indication was with intravenous 5-FU. Despite the lack of data combining bevacizumab with a FOLFOX regimen, the majority of US oncologists chose to incorporate bevacizumab with FOLFOX in front-line metastatic disease; this was done on the solid assumption that if bevacizumab improves the outcome with IFL, then the incorporation with the superior regimen using infusional 5-FU and either oxaliplatin or irinotecan(Drug information on irinotecan) would offer further improvements. The Eastern Cooperative Oncology Group trial ECOG 3200 is an important randomized clinical trial that demonstrated the benefit of adding bevacizumab to FOLFOX4 chemotherapy.[ 1] However, it is important to recognize that this study was done in second-line metastatic colorectal cancer patients who had previously received an irinotecan-based chemotherapy. As shown in Figure 1, the randomization of this trial was between FOLFOX4, FOLFOX4 plus bevacizumab at 10 mg/kg, and bevacizumab alone at 10 mg/kg. The results of this trial showed a statistically significant and clinically meaningful improvement in overall survival in patients with metastatic colorectal cancer. The toxicity profile, as a combination regimen, was acceptable. The neuropathy that was seen at a higher frequency in the FOLFOX4 plus bevacizumab arm is attributed to the fact that these patients remained on FOLFOX4 for a longer period of time and therefore developed more cumulative neurotoxicity. The interesting survival data from the bevacizumab alone arm is probably best explained by the role of crossover therapy in this arm. This arm was, in fact, closed early due to inferior results compared to the other two arms; this again supports the notion that bevacizumab alone is probably inactive in patients with progressing metastatic colorectal cancer. It is critical to recognize that this trial does not answer one important question: what is the role of bevacizumab in second-line therapy in patients who have already received bevacizumab in the first line? The benefit observed in this trial is in bevacizumab- naive patients. This trial does tell us clearly that bevacizumab does offer benefit to FOLFOX chemotherapy, that it is safe to combine, and that those patients who have not yet been given bevacizumab in the first line should be offered bevacizumab in the second or third line. TREE 1/2 The combination of bevacizumab is now clearly accepted with either irinotecan- based or oxaliplatin-based therapy, but approved with intravenous 5-FU. The question still remains as to whether bevacizumab can also be added to a capecitabine-containing regimen with a similar benefit. Two important trials that were presented at the ASCO 2005 annual meeting helped address this important issue. The first is known as the TREE study, presented by Hochster et al.[2] The TREE 1 study was originally designed without bevacizumab. It was a three-arm randomized clinical trial comparing (1) infusional 5-FU combined with oxaliplatin to (2) bolus 5-FU with oxaliplatin to (3) oral 5-FU with oxaliplatin. This phase II randomized trial was powered for safety and toxicity end points but give us a very solid comparison in terms of response rates as well. TREE 1, without bevacizumab, accrued approximately 150 patients. These were equally distributed across the three arms. It was demonstrated that, at least based on response rate, the infusional 5-FU regimen was superior to the oral 5-FU regimen, which was superior to the bolus 5-FU regimen. The capecitabine arm was found to be too toxic at the 1,000-mg/m² dose used in this original design, and many patients required dose reductions. The trial had to be amended when bevacizumab was approved for frontline use in metastatic colorectal cancer, giving an ideal opportunity to test the incorporation of bevacizumab into each of these arms. TREE 2, as it became known, is essentially the same clinical trial as TREE 1 but with the addition of bevacizumab to each of the arms and the capecitabine dose reduced to 825 mg/m². The results of the TREE 2 study are quite interesting in that it demonstrates that the infusional 5-FU with bevacizumab appears to be roughly the same as the oral 5-FU bevacizumab arm based on response rate and toxicity. The bolus 5-FU regimen was a distant third. In addition, what is important to note is that the benefit seen with the addition of bevacizumab carried over into each of the three arms, supporting the fact that capecitabine could be given with bevacizumab with seemingly similar results as compared to infusional 5-FU. The question as to whether bevacizumab will become FDA-approved with oral 5-FU is unclear. In a single-arm phase II study, Fernando et al combined oxaliplatin, capecitabine, and bevacizumab for front-line metastatic colorectal cancer.[ 3] Patients received oxaliplatin at 85 mg /m² on day 1, capecitabine at 1,000 mg/m² twice a day, days 1 through 5 and 8 through 12, and bevacizumab at 10 mg/kg on day 1. Cycles were repeated every 2 weeks. As in other trials, the dose of capecitabine had to be reduced to 850 mg/m² due to toxicity in the first 27 patients. The results of this trial show not only an acceptable toxicity profile but also an interestingly high response rate of 57%, with one complete response and an additional 37% with stable disease out of the total 30 patients that were treated. Again, these are supporting data to suggest that bevacizumab can be combined with oral 5-FU-containing regimens in metastatic colorectal cancer. The authors do report an 11.9-month median time to progression, but for such a small phase II clinical trial, it is hard to give that number too much weight. The Dosing Controversy What is the proper dose of bevacizumab? The original trial supporting the approval of bevacizumab used a 5- mg/kg dose. That dose comes from a small phase II randomized clinical trial comparing the 5-mg/kg dose to the 10- mg/kg dose in combination with intravenous 5-FU.[4] The 5-mg/kg dose offered a higher response rate and a higher time to progression than the 10- mg/kg dose; therefore, it was selected for the pivotal trial. As you might notice in the previous two clinical trials, ECOG 3200 and the Fernando study, the 10-mg/kg dose was used for patients with metastatic colon cancer. It is still to this day unclear what the proper dosing in colorectal cancer is. In breast cancer, lung cancer, and renal cell cancer, the higher dose of bevacizumab was used, and it may be that the higher dose will win the day in the end. Certainly, the toxicity profile is not significantly different between the high-dose and the low-dose regimens. For colorectal cancer, the approved dose is 5 mg/kg. These studies that I have reviewed do support using the higher dose but do not answer the question of whether the lower dose would be equally active. Further studies will have to be performed to better understand what the appropriate dosing is for patients with metastatic colorectal cancer. Vatalanib Following on the promising results of bevacizumab in metastatic colorectal cancer, vatalanib (PTK 787), an oral vascular epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor, has also been evaluated in metastatic colorectal cancer. The first randomized results with the compound were reported by Hecht.[5] In the CONFIRM- 1 study, 1,168 patients with metastatic colorectal cancer who had not received prior chemotherapy for metastatic disease were randomly assigned to receive oxaliplatin, 5-FU, and leucovorin on the FOLFOX4 schedule, with either vatalanib at 1,250 mg daily or placebo. The primary end point of the study was progression-free survival, and the study was powered to detect a 25% decrease in the risk of progression of disease with the addition of vatalanib. There was no significant difference in either response rate (46% for FOLFOX alone and 42% with the addition of vatalanib) or progression-free survival (7.6 vs 7.7 months, with a hazard ratio of 0.88 that suggests a nonsignificant 12% decrease in the risk of progression). A subset analysis of progression- free survival based on the investigator- reported outcomes suggested a possible benefit with the addition of vatalanib, with a median 7.7 months, compared with 7.5 months in the placebo arm, which was statistically significant, with a hazard ratio of 0.83. The toxicity profile of therapy with the addition of vatalanib to FOLFOX4 chemotherapy was deemed acceptable (Table 1). The severe chemotherapybased toxicities such as neutropenia, diarrhea, and nausea/vomiting were similar between the two arms. Somewhat more patients treated with vatalanib experienced grade 3 hypertension, which required medical intervention: 21% vs 6% with placebo. Severe bleeding episodes were similar and infrequent in the two arms, 2.9% with vatalanib, and 2.8% with placebo. Although there was no significant differences in the frequency of deep-vein thromboses (7% with vatalanib vs 4%), there was a notable difference in the incidence of pulmonary emboli (6% compared with 1%). More episodes of grade 3 dizziness were documented in the patients who received vatalanib (7%) compared to placebo (2%). Because of the potential of an oral agent such as vatalanib, other investigations have also been conducted with vatalanib in combination with chemotherapy. As irinotecan/5-FU/leucovorin combination chemotherapy may have similar antitumor activity to oxaliplatin- based chemotherapy, vatalanib has also been combined with irinotecan with the LV5FU2 schedule (leucovorin at 200 mg/m² IV, 5-FU at 400 mg/m² IV bolus, then 5-FU 600 mg/m² IV over 22 hours on days 1 and 2 every 14 days) of chemotherapy. Trarbach reported on the phase I part of the study that had been conducted, in which the chemotherapy doses were fixed and the dose of vatalanib was escalated.[6] Twenty-one patients were enrolled in this portion of the study. They demonstrated that the 1,250-mg daily dose of vatalanib was safe in combination with irinotecan/ LV5FU2 chemotherapy. Dose-limiting toxicities were reported in one patient each in the 500-mg (fatigue) and 1,000- mg (hypertension) dose levels. No doselimiting toxicities were reported in the four patients enrolled on the 1,250-mg dose level or the three patients treated on the 1,500-mg dose level. The other toxicities that were reported were expected. About 19% of patients had grade 3 diarrhea; an additional 19% had grade 3 hypertension. Four episodes of deepvein thrombosis were diagnosed. An additional four patients also had palmar- plantar syndrome, though only one event was grade 3 in severity. Pharmacokinetic studies were conducted evaluating both vatalanib and irinotecan and SN38. The former was not influenced by chemotherapy, whereas the data that were limited by the small numbers of patients enrolled on the study suggested that vatalanib did decrease the exposure to irinotecan, SN-38, and SN-38G. In the limited evaluation of response possible during a phase I study, this combination appeared to have promising antitumor activity. In the 20 evaluable patients in all dose levels, a 55% response rate was seen, with an additional 35% of patients having stable disease. A better assessment will be available when the data are available from the additional 27 patients that were enrolled on the phase II portion of the study at the 1,250-mg dose. A pharmacokinetic evaluation of vatalanib with irinotecan alone was also reported.[7] A total of 44 patients with metastatic colorectal cancer that had received zero or one prior chemotherapy for metastatic disease were randomly assigned to receive irinotecan at 250 mg/m² IV every 3 weeks, with or without vatalanib at 1,250 mg/m² daily. More severe diarrhea and vomiting were reported in the arm with vatalanib and irinotecan. Again, the irinotecan did not influence the pharmacokinetic profile of the vatalanib, but the exposure to irinotecan and SN-38 was decreased by 28% and 49%, respectively. In this limited study, crossover was allowed. However, the preliminary data suggest that the pooled response rate with the addition of vatalanib to irinotecan (23.5%) was superior to that with irinotecan alone (14.3%). A study of vatalanib in the secondline setting in metastatic colorectal cancer (CONFIRM-2) has also been completed, although the results are not yet available. Whether those results mirror the results of ECOG 3200 will be closely watched, as it may further elucidate the true mechanism of activity of the drug, and the roles of the VEGF targets in colorectal cancer. Epidermal Growth Factor Receptor Inhibitors The epidermal growth factor receptor has also been a fruitful target in metastatic colorectal cancer and other malignancies. In colorectal cancer, cetuximab(Drug information on cetuximab) (Erbitux) has been approved for use in patients whose disease has progressed despite prior therapy with irinotecan-based therapy. The results of the pivotal study demonstrated a response rate that was superadditive when cetuximab was combined with irinotecan (22.9%), in comparison to cetuximab alone (10.8%).[8] As the patients who received cetuximab with irinotecan had disease that had progressed on prior irinotecan, these data suggest that cetuximab, and perhaps anti-EGFR therapy in general, acts at least in part by enhancing tumor sensitivity to chemotherapy. Given the activity of cetuximab in patients who have had multiple prior therapies, this drug has been evaluated in a potentially more chemotherapy- sensitive population, the first-line metastatic disease setting, with intriguing initial results. Diaz-Rubio updated results of a multicenter phase II study of oxaliplatin, 5-FU, and leucovorin (the FOLFOX4 schedule) with cetuximab.[9] A total of 43 patients were enrolled on the study, with objective responses in 81% of the 42 evaluable patients, including 10% complete responses. An additional 23% of patients had stable disease. The median progression-free survival was 54 weeks, with 52% of patients free from progression at 1 year. The toxicity reported with this combination was similar to that which would be expected: 30.2% of patients had grade 3 rash; 25.6% of patients had grade 3 or 4 diarrhea and a similar proportion of patients had grade 3 or 4 neuropathy. Grade 3 or 4 neutropenia was reported in 20.9% of patients. Severe mucositis or stomatitis occurred in 16.3% of patients. However, little randomized data regarding cetuximab in the first-line setting of metastatic colorectal cancer are available at this time to confirm Diaz- Rubio's dramatic findings. The advent of bevacizumab has forced the closure or redesign of studies that would have furthered our knowledge about cetuximab in the first-line setting. One of these studies was presented at ASCO. The EXPLORE study was a randomized phase III study of oxaliplatin, 5- FU, and leucovorin administered on the FOLFOX4 schedule, with or without cetuximab, in the first-line metastatic disease setting.[10] Only 102 of the planned 1,100 patients were enrolled, so clearly no definitive conclusions could be drawn. Nonetheless, the data presented did not corroborate Diaz- Rubio's report. Of the 50 patients who received cetuximab, 10 confirmed responses (20%) were noted, with 22 patients (44%) having stable disease. At the meeting, the median progressionfree survival was reported to be 4.4 months. In comparison, confirmed responses occurred in 4 of 42 patients (10%), with 26 patients (62%) having stable disease. Another monoclonal antibody that targets EGFR, panitumumab (ABXEGF), is also being developed. This humanized antibody has the theoretical advantage of being less likely to produce infusion reactions. Malik presented data from a phase II study at ASCO.[11] A total of 148 patients whose tumors expressed EGFR by immunohistochemical stains and who had received prior chemotherapy were enrolled. Forty-four percent of the patients had received prior 5-FU, irinotecan, and oxaliplatin. Panitumumab was given intravenously at a dose of 2.5 mg/kg weekly. Ten percent of patients had partial responses to therapy; the median time to disease progression was 2.5 months, and the median survival was 9.4 months. Therapy was well tolerated, with only one infusion reaction reported. Ninety-five percent of patients had some rash, but only 7% had grade 3 rash. Other grade 3 toxicities were infrequent, with only fatigue (9%) occurring in more than 3% of patients. This agent is undergoing further investigation in combination with chemotherapy the first-line setting. The other commercially available EGFR inhibitors, the oral tyrosine kinase inhibitors erlotinib (Tarceva) and gefitinib(Drug information on gefitinib) (Iressa), have not demonstrated significant antineoplastic activity in metastatic colorectal cancer. At the 2004 ASCO meeting, Rothenberg reported on ECOG 6200, a clinical trial of gefitinib in patients with metastatic colorectal cancer whose disease had progressed despite prior irinotecan and 5-FU chemotherapy.[12] A total of 115 patients who had not been evaluated for the expression of EGFR were randomly assigned to receive either 250 or 500 mg of gefitinib daily. In the results presented at the meeting, 55 patients in each arm were evaluable for response. Only one partial response was noted, in a patient who was treated at the 500-mg dose level. Thirteen percent of the patients who were treated on the lower dose had stable disease, compared to 24% of those treated at 500 mg. The median progression-free survival was 1.8 months on the 250-mg arm and 2.1 months at the higher dose; median survival was 5.2 and 8.2 months, respectively. The toxicity profile reflected that from prior experience reported with gefitinib. The authors concluded that gefitinib is inactive as a single agent in metastatic colorectal cancer after irinotecan and 5-FU therapy. Oza reported a study of erlotinib at a dose of 150 mg daily in patients with metastatic colorectal cancer who had received at least one prior chemotherapy regimen, with similar results.[ 13] They, too, were not selected for the expression of EGFR. Of the 25 evaluable patients who were reported, no objective responses were reported and 32% had stable disease. However, this year Keilholz reported an ongoing study of erlotinib (150 mg daily) as either second or third line therapy in metastatic colorectal cancer.[14] At the time of presentation, 23 patients received therapy in the second-line setting, with two patients having a partial response (8%) and nine having stable disease. However, in the third-line setting, no objective responses occurred in 28 treated patients, and 9 more patients had stable disease. The toxicities noted in this study were similar to those previously reported. In contrast to the reported experience in non-small-cell lung cancer, though, the EGFR tyrosine kinase inhibitors may have greater antitumor efficacy when used in combination with cytotoxic chemotherapy, rather than as single agents in metastatic colorectal cancer. Kuo reported a phase II trial of oxaliplatin, 5-FU, and leucovorin on the FOLFOX4 schedule with gefitinib at 500 mg daily in patients who had received prior chemotherapy for metastatic disease.[ 15] About three-quarters of patients had received prior irinotecan. Objective responses were reported in 33% of the 27 patients enrolled. The median event-free survival was 5.4 months, and the median survival of patients was 12.0 months. Acknowledging the difficulties comparing a small single-institution phase II study to phase III studies, all of these results compare favorably to the expected results with FOLFOX therapy alone. Severe diarrhea was more frequently noted, in 48% of patients, with the addition of gefitinib. Combined Targeted Therapy The available data suggest that "targeted therapy" such as bevacizumab and cetuximab may be most effective in advanced colorectal cancer when they are combined with other therapies. This would correspond to the known models that demonstrate that unlike chronic myelogenous leukemia and gastrointestinal stromal tumors, which often arise as the result of a single activating mutation, colon cancer often arises as the result of multiple mutations. A natural extension of this knowledge would be to attempt to combine targeted therapies. At this time, there are two ways this may be accomplished. First, one may combine two separate agents. Alternatively, new drugs are being developed that target multiple proteins, including VEGF, EGFR, HER2, c-kit, and platelet derived growth factor receptor (PDGFR). An example of the former type of study was presented by Saltz.[16] The so-called BOND-2 trial was a randomized phase II study in patients who had disease refractory to irinotecan- based chemotherapy, but had no prior exposure to cetuximab or bevacizumab; EGFR expression was not necessary for enrollment on this study. The results for 81 patients were reported at ASCO (Table 2). Patients were randomly assigned to receive bevacizumab at 5 mg/kg every 2 weeks, cetuximab at a 400-mg/m² loading dose, followed by a weekly 250-mg/m² dose, either with or without irinotecan. The irinotecan was administered on the same schedule on which the patient had most recently developed progressive disease. This design mirrored that of the pivotal BOND study of cetuximab, with the addition of bevacizumab. During the first cycle of therapy, patients received cetuximab on day 1, then bevacizumab on day 2. On subsequent treatment cycles, these monoclonal antibodies were administered sequentially on the same day. Responses were reported in both arms of the study: 20% with cetuximab/ bevacizumab, and 37% in the irinotecan/cetuximab/bevacizumab arm. The median time to progression was 5.6 and 7.9 months, respectively. Comparison to the results from the BOND study must be limited, but certainly these results suggest the promising antitumor activity of the combination of these targeted agents. Indeed, given the modest activity of bevacizumab as a single agent in this setting, it is possible that combined inhibition of EGFR and VEGF is synergistic, even in the absence of chemotherapy. The toxicity profile of this combination is reasonable (Table 3), and reflects that expected from the combination of theses agents. The frequency of rash was similar to that which would be expected with cetuximab alone. A severe, extensive grade 3 rash was reported in 20% and 12%, respectively, of patients who had received cetuximab/bevacizumab alone or with bevacizumab. The addition of irinotecan resulted in notably more diarrhea and myelosuppression. The patients receiving cetuximab and irinotecan alone did not experience any grade 3 or 4 neutropenia or diarrhea, compared to 18% and 24% of patients who received the addition of cetuximab. One unexpected finding was that 5% of the patients who were treated with the monoclonal antibodies alone developed a severe, grade 3 headache. This event occurred only during the first cycle. This was attributed to a lack of steroid premedication. Lapatinib is an oral dual tyrosine kinase inhibitor of EGFR and HER2. A single-agent study of this agent was performed in patients who had metastatic colorectal cancer that had progressed despite initial 5-FU-based therapy, either in combination with irinotecan (93%) or oxaliplatin (3%).[17] Patients were not selected based on EGFR or HER2 expression, but these proteins were expressed in 54.2% and 44.1% of patients. The 86 patients that were enrolled on the study received lapatinib at a dose of 1,250 mg daily. Limited antineoplastic efficacy was noted, with only one partial response (1%) and five more patients with "minor responses." A total of 20 patients (23%) had a response or stable disease. The median time to progression was 8 weeks, and the median survival was 42.9 weeks. Therapy was well tolerated, with severe, grade 3 diarrhea in 5% of patients and grade 3 constipation in 3%. No other grade 3 toxicity was reported in more than 2% of patients. As would be expected the most common toxicities were diarrhea, rash, fatigue, nausea, and anorexia. Conclusion In summary, the impact of the new biologic therapies on colorectal cancer has been dramatic. There are other medicines which are likely to enter the therapeutic scene directed toward the same targets, but also newer targets will certainly emerge as active and significant in treating patients with colorectal cancer. The important data that we have reviewed from this year's ASCO annual meeting really help us refine how best to incorporate these new medicines into our treatment of patients with metastatic colorectal cancer, and form the basis for new adjuvant trials now ongoing for patients with stage II and III disease. We are learning that there is probably no appropriate single standard of care for patients with metastatic colorectal cancer, and that treatment choices will be increasingly tailored for patients. We are recognizing that patients should receive multiple lines of therapy and that these multiple lines of therapy result in an improved survival. But we also must recognize that as our patients continue on with chemotherapy for many consecutive years, we need to understand how best to manage their side effects and keep their quality of life as high as possible.