Accomplishments in 2007 in the Treatment of Advanced Colorectal Cancer

Abstract

1. Overview of the Disease

   1. Incidence
   2. Prognosis

      1. Prognostic or Predictive Factors
2. Current General Therapy Standards and Regional Variations

   1. Standards in the US and Europe: Combination Therapy Plus a Biologic

      1. FOCUS and CAIRO 1: Serial Single Agents vs. Combination Therapy
      2. Stop-and-Go Strategy
      3. Limited Availability of Biologics in Some Regions
3. Accomplishments During the Year

   1. Therapy
4. What Needs To Be Done

   1. Controversies and Disagreements

      1. BOND-2 and PACCE: Chemotherapy + Bevacizumab + Cetuximab or Panitumumab
      2. Special Populations (Elderly and Poor Performance Status)
5. Future Directions

   1. Comments on Research
   2. Obstacles to Progress

I. OVERVIEW OF THE DISEASE

I-A. Incidence

Colorectal cancer, even when metastatic, is a highly treatable and potentially curable disease, particularly if disease is limited to a few tumor deposits localized in liver, lung, or the anastomotic site in the bowel from which the primary tumor was initially excised. In the United States, the incidence of colorectal cancer peaked in 1986 at 67 cases per 100,000, declining to 51.6 cases per 100,000 in 2006. Estimates for 2007 projected 112,000 new cases of colon and 41,000 new cases of rectal cancer in the United States.¹ In Europe, 372,000 new cases of colorectal cancer were diagnosed in 2002, the most recent year for which full statistics are available.² In the United States, the median age at diagnosis of colorectal cancer is 71 years.¹

I-B. Prognosis

The prognosis of patients with metastatic colorectal cancer is clearly related to the number, size, and location(s) of the disease and is best for disease confined to a single site. Up to 60% of patients with a surgically removed, single, late manifesting liver metastasis were alive at 5 years in a report of more than 1,000 cases from Memorial Sloan-Kettering Cancer Center.³

Colorectal cancer remains the second most frequent cause of cancer-related mortality in the United States and Europe despite the lower incidence rates and better outcomes that have been reported in recent years.^1,4

I-B. 1. Prognostic or Predictive Factors

Prognostic or predictive factors have been extensively studied over the years. New findings are enhancing the understanding of the biology underlying the observed clinical heterogeneity seen among patients with colorectal adenocarcinomas. Substantive discoveries made in 2007 now permit biomarker-guided therapeutic choices.

• Circulating tumor cells (CTC) are prognostic⁵: Immunomagnetic separation of CTC in blood samples obtained from 456 metastatic CRC (mCRC) patients was performed prior to starting first-, second-, or third-line therapy and at subsequent time intervals. Patients with CTC levels ≥3 in 7.5 mL of peripheral blood at first testing or subsequent evaluations were classified as having an unfavorable prognosis and those with <3 CTC were considered as having a favorable prognosis. The overall survival (OS) of patients who maintained or developed ≥ 3 CTC at follow-up time points was worse than for patients maintaining <3CTC (12.6 months, 7.0 months vs. 21.1 months, respectively, P < .0001). In multivariate analyses including age, Eastern Cooperative Oncology Group (ECOG) performance status, and the line and type of therapy, CTC remained the most significant independent predictor of outcome. A large randomized phase III study to confirm these findings in mCRC patients receiving first-line therapy recently completed accrual (Dutch Colorectal Cancer Group [DCCG] CAIRO-2 trial).

• Predictive markers of sensitivity or resistance to anti-epidermal growth factor receptor (EGFR) monoclonal antibodies^6–8: Patients with tumors that express high levels of the EGFR ligands epiregulin (EREG) and amphiregulin (AREG) are more likely to have disease control with cetuximab (EREG, P = .000015; AREG, P = .000025). In addition, patients whose tumors have wild type as opposed to mutant K-ras have a significantly higher disease control rate (P = .0003). In retrospectively analyzed series, essentially no patient with mutated K-ras has manifested a partial response to the anti-EGFR antibodies cetuximab or panitumumab. Emerging data suggest that K-ras status may also have relevance for sensitivity to other antineoplastic agents.

• EGFR gene copy number (GCN) predicts resistance to targeted monoclonal antibodies⁹: A subset of patients from a phase III trial of panitumumab plus best supportive care (BSC, n = 58) vs. BSC alone (n =34), with tumors distinguishable by fluorescence in-situ hybridization (FISH) analysis, were evaluated for GCN. In panitumumab-treated patients whose tumors had a mean EGFR GCN of less than 2.5/nucleus or when less than 40% of tumor cells displayed chromosome 7 polysomy within the tumor (as compared with patients whose tumors exhibited higher values), the progression-free survival (PFS), (P = .039 and P = .029, respectively) and overall survival (P = .015 and P = .014, respectively) were shorter.

• PTEN loss may be predictive for response to EGFR antibodies¹⁰: In a small study, the loss of PTEN protein expression was associated with unresponsiveness to cetuximab in the 11 patients out of 27 whose tumors exhibited loss of PTEN activity (P < .001).

• Nuclear factor-κ B (NFκ B) status appears to segregate responsive vs. nonresponsive tumors¹¹: In 76 patients determined to have NF-κ B–positive or NF-κ B–negative tumors, the response rate was 10% (4 partial responses [PR]) vs. 48% (12 PR), respectively (P = .0007). Median time to progression (TTP) was 3 months in NF-κ B–positive patients vs. 6.4 months in the remaining patients (P = .021). Median overall survival (OS) was 9.5 months vs. 15.8 months for NF-κ B–positive and NF- κ B–negative patients, respectively (P = .036).

II. CURRENT GENERAL THERAPY STANDARDS AND REGIONAL VARIATIONS

II-A. Standards in the US and Europe: Combination Therapy Plus a Biologic

In the United States and Europe, combination therapy in first-line treatment of metastatic disease is the overwhelming standard of care. Regimens based on either irinotecan or oxaliplatin given with 5-fluorouracil (5-FU) or capecitabine prevail. In most cases bevacizumab is also used. This preference is based on data from randomized studies conducted in prior years.

II-A. 1. FOCUS and CAIRO 1: Serial Single Agents vs. Combination Therapy

Two studies comparing serial single-agent therapy vs. combination therapy (FOCUS and CAIRO 1) were published in a single issue of The Lancet in 2007.^12,13

• The FOCUS study randomly assigned 2,130 patients with advanced colorectal cancer to receive either first-line singleagent treatment or combination therapy.¹² Initially, the serial single- agent group received 5-FU plus leucovorin then singleagent irinotecan. A second cohort began with 5-FU to which irinotecan or oxaliplatin was added upon progression. A third group received combination therapy with 5-FU plus irinotecan or oxaliplatin. The primary end point was OS. Median OS was 13.9 months for sequential single-agent therapy, approximately 15 months when single-agent therapy was followed by the addition of a second agent, and 15 to 16 months for sequential combination treatment. The only significant survival advantage was in the irinotecan combination arm as compared with the serial single-agent approach (P = .01).

• The CAIRO-1 study randomly assigned 820 patients with advanced colorectal cancer to receive either first-line treatment with capecitabine, second-line irinotecan, and third-line capecitabine plus oxaliplatin (sequential treatment, n = 410), or first-line treatment with capecitabine plus irinotecan and second- line capecitabine plus oxaliplatin (combination treatment, n = 410).¹³ The primary end point was OS. Median OS was 16.3 (95% CI: 14.3–18.1) months for sequential treatment and 17.4 (95% CI: 15.2–19.2) months for combination treatment (P = .3281). The hazard ratio (HR) for combination vs. sequential treatment was 0.92 (95% CI: 0.79–1.08, P = .3281). The frequency of grade 3–4 toxicity over all lines of treatment did not differ significantly between the two groups, except for grade 3 hand-foot syndrome (HFS), which occurred more often with sequential than with combination treatment (13%vs. 7%, P = .004).

• While these data were considered as definitive by some clinicians, others think that because the outcomes with regard to median OS were so much shorter than is achieved with best available therapies (ie, now exceeding 26 months), and the falloff between first and later lines of therapy was so dramatic, the studies should not be used to change practice and cause the standard of care to revert back to a serial single-agent approach. In Great Britain and the Netherlands where these studies were conducted, single-agent therapy is perhaps more commonly used than elsewhere in the world. Many European clinicians continue to use combination therapy, often with bevacizumab.

II-A. 2. Stop-and-Go Strategy

A stop-and-go strategy in which treatment holidays are planned after response to initial therapy has been an area of debate. There are two main studies:

• In the first of these trials the authors set out to evaluate whether intermittent irinotecan plus 5-FU (irinotecan 180 mg/m² day 1 + l-folinic acid (LV) 100 mg/m² in 2 hours + 5-FU 400 mg/m² IV bolus + 600 mg/m² via 22-hour infusion, days 1 and 2, every 2 weeks, for 2 months on and 2 months off) (arm A) was at least as effective as continuous therapy with the same regimen (arm B), with treatment administered until disease progression in both arms.¹⁴ Among 336 patients randomized, the response rate (RR) was 29% in arm A and 35% in arm B, with median PFS durations of 8.8 months and 7.3 months, respectively (HR = 1.00, 95% CI: 0.74–1.36). At a median follow-up of 27 months, median OS was 16.9 months in arm A and 17.6 months in arm B (HR = 1.11, 95% CI: 0.83– 1.48). Toxicity was acceptable and similar in the two arms. As a conclusion, this study suggests that an intermittent FOLFIRI treatment schedule leads to the same median survival as continuous treatment, and treatment breaks can reduce both the toxicity and the economic costs of treatment for advanced disease.

• In another trial, 202 patients were randomized between the OPTIMOX1 arm of six cycles of FOLFOX7 (LV, 5-FU, oxaliplatin) followed by infusional 5-FU until progression, at which point FOLFOX7 was reintroduced; vs. the OPTIMOX2 arm of six cycles of FOLFOX7 followed by a complete break from chemotherapy, and reintroduction of FOLFOX7 when disease progressed to baseline measurements.¹⁵ Median PFS differed and was 8.3/6.7 months (P = .04) for OPTIMOX1/OPTIMOX2. Median OS was 24.6/18.9 months respectively, (P = .05). As a consequence, it was recommended to maintain the infusional 5-FU therapy, especially in patients with poor prognosis. These study results were difficult to interpret, in part because permitting responding patients to progress to the point of initial bulk of disease may have been a strategic error, and because the study was stopped early due to the evolution of treatment paradigms to include bevacizumab.

II-A. 3. Limited Availability of Biologics in Some Regions of the World (Japan, Asia, India, Africa, Others)

In many areas of the world, there is limited availability of combination chemotherapy and biologic agents due to regulatory or fiscal restrictions. The high prices of biologics have limited their availability. Discounted drugs are also not generally available due to the complexity of the manufacturing processes involved in their synthesis in countries where patent protections are either not in place or are laxly enforced.

III. ACCOMPLISHMENTS DURING THE YEAR

III-A. Therapy

• Investigators from Italy have developed and tested a regimen that combines the three active classes of chemotherapies for CRC, known as FOLFOXIRI (infusional 5-FU, leucovorin, oxaliplatin, irinotecan).¹⁶ A total of 244 patients were randomly assigned to receive FOLFOXIRI or a 5-FU plus irinotecan regimen. Significant differences in toxicities between arms included an increase of grade 2–3 peripheral neurotoxicity (0% vs. 19%, P = .001) and grade 3–4 neutropenia (28% vs. 50%, P = .001) with FOLFOXIRI, although these were not prohibitive. Confirmed RR was 34% vs. 60% favoring FOLFOXIRI (P = .0001). The R0 secondary resection rate of metastases was greater in the FOLFOXIRI arm (6% vs. 15%, P = .033, among all 244 patients; and 12% vs. 36%, P = .017, among patients with liver metastases only). PFS and OS were both significantly improved in the FOLFOXIRI arm (median PFS, 6.9 vs. 9.8 months, HR 0.63, P = .0006; median OS, 16.7 vs. 22.6 months, HR 0.70, P = .032).

• The CRYSTAL (Cextuximab Combined with Irinotecan in First- Line Therapy for Metastatic CRC) trial examined outcomes and toxicity with the addition of cetuximab to FOLFIRI in untreated patients with metastatic disease.¹⁷ A total of 1,217 patients were randomly assigned to receive FOLFIRI with or without cetuximab every 2 weeks. The primary end point was PFS. Median PFS was significantly longer, albeit by less than 1 month, for the cetuximab group (8.9 vs. 8 months, P = .036). RR was also significantly increased by the addition of cetuximab (46.9% vs. 38.7%, P = .005).

• The OPUS trial used FOLFOX chemotherapy with or without cetuximab with treatment cycles repeated every 2 weeks.¹⁸ Among 337 patients randomized, the best overall confirmed RR was 45.6% in the cetuximab-treated cohort and 35.7% in the FOLFOX-alone group. PFS and OS data are not yet available.

• The BICC-C trial was a complicated study done in several stages. It included a double randomization to celecoxib or placebo as well as a number of irinotecan-containing chemotherapy regimens.¹⁹ A total of 430 previously untreated mCRC patients were randomly assigned to receive FOLFIRI (irinotecan + infusional 5-fluorouracil and leucovorin [5-FU/LV], n = 144), modified IFL (irinotecan + bolus 5-FU/LV, n =141), or CapeIRI (capecitabine + irinotecan, n =145). After a protocol amendment, an additional 117 patients were randomly assigned to receive FOLFIRI plus bevacizumab (FOLFIRI-bev, n = 57) or mIFL plus bevacizumab (mIFL-bev, n= 60), whereas the CapeIRI arm was discontinued due to higher rates of severe vomiting, diarrhea, and dehydration. The primary study end point was PFS. Median PFS was 7.6 months for FOLFIRI, 5.9 months for mIFL (P = .004 for the comparison with FOLFIRI), and 5.8 months for CapeIRI (P = .015). Median OS was 23.1 months for FOLFIRI, 17.6 months for mIFL (P = .09), and 18.9 months for CapeIRI (P = .27). After the amendment to add bevacizumab, the median survival time was not reached for FOLFIRI-bev after a median follow-up of 26 months, and was 19.2 months for mIFL-bev (P = .007). FOLFIRI-bev was associated with a higher rate of grade 3 hypertension than mIFL-bev. FOLFIRI and FOLFIRI-bev had activity superior to that of their comparators and were generally associated with less toxicity. An infusional schedule of 5-FU should be preferred when using irinotecan-based regimens in first-line therapy for metastatic colorectal cancer.

• XELOX ± bevacizumab (NO16966, Spanish TTD, German AIO): In the NO16966 trial, patients were initially randomized to firstline treatment with XELOX (capecitabine/oxaliplatin) or FOLFOX4, and for a second cohort, randomization included those two chemotherapy regiments ± bevacizumab in a 2 × 2 factorial design. XELOX was proven to be “non-inferior” to FOLFOX4 for PFS.^20,21 Results also showed a significant difference in PFS favoring the addition of bevacizumab (9.4 vs. 8.0 months, HR 0.83, P =.0023), and a nonsignificant difference in median OS (21.3 vs. 19.9 months, P = .0769), possibly explained by a higher proportion of treatment discontinuation in the bevacizumab arms due to adverse events (31% vs. 21%). XELOX, XELOX-placebo, and XELOX-bevacizumab were “non-inferior” to FOLFOX4, FOLFOX4-placebo, and FOLFOX4-bevacizumab, respectively, in terms of PFS; the HR was 1.05.

• Two phase III studies, the Spanish Cooperative Group for the Treatment of Digestive Tumors (TTD) trial²² and the Association of Medical Oncology of the German Cancer Society (AIO) trial²³ compared oxaliplatin combinations with either continuous infusion 5-FU ± LV or capecitabine in first-line treatment. A lack of a statistical difference was observed in these trials for median TTP, favoring the infusion-based 5-FU program, with the same trend noted for median OS. This was also the case in another phase III trial comparing XELOX vs. FOLFOX4 ± bevacizumab.²⁴ Grade 3 and 4 toxicity profiles showed less diarrhea and more granulocytopenia and febrile neutropenia with FOLFOX4 as compared with XELOX. The patient preference for oral treatments, which avoid the requirement of implanted vascular access devices (VAD) and their complications, is the most powerful argument favoring capecitabine use. However, this advantage is often abrogated in combination therapy as many patients will need a VAD for the administration of oxaliplatin or other second- or third-line drugs. In some analyses, the estimated comparative costs favor 5-FU infusion regimens over XELOX, due to the wide pricing difference between proprietary capecitabine and generic 5-FU. XELOX treatment costs are almost double that of infusional 5-FU treatments (95.2% more), including hospital visits, pumps, and infusion fees.²⁵ All these facts should be taken into consideration when using an oxaliplatin plus fluoropyrimidine combination. XELOX does provide a “non-inferior” and more convenient treatment for the patient, at a higher cost.²⁵ A phase III study comparing XELOX with FOLFOX4 in metastatic CRC patients who had received previous treatment with irinotecan in combination with 5-FU/LV has demonstrated similar findings.²⁶ FOLFOX has not yet been replaced by XELOX as the standard of care. These findings should be considered together with results from other phase III trials, which are maturing.

• CONCEPT: The Combined Oxaliplatin Neuropathy Prevention Trial (CONCEPT) was designed to evaluate the potential to reduce cumulative oxaliplatin neurotoxicity through intermittent administration of oxaliplatin and the use of calcium and magnesium salts in first-line therapy for metastatic colorectal cancer. ²⁷ The independent data monitoring committee (IDMC) closed the trial based on data suggesting a statistically significant response rate disadvantage for patients randomized to receive calcium and magnesium infusion (17% vs. 34%, P < .04). This analysis was based on investigator determination and unconfirmed response. The sponsor then retained independent radiology assessors to verify all data in a blinded fashion. Data were presented at the American Society of Clinical Oncology (ASCO) 2008 Gastrointestinal Cancers Symposium, suggesting that magnesium and calcium did not appear to reduce the response rate as had been reported to the IDMC.²⁸ The discrepancy is difficult to interpret.

IV. WHAT NEEDS TO BE DONE

IV-A. Controversies and Disagreements

IV-A. 1. BOND-2 and PACCE (Chemotherapy + Bevacizumab + Cetuximab or Panitumumab)

• The phase III PACCE (Panitumumab Advanced Colorectal Cancer Evaluation) study evaluated oxaliplatin- or irinotecanbased chemotherapy plus bevacizumab with and without panitumumab in the first-line treatment of patients with metastatic CRC.²⁹ The trial enrolled 1,054 US patients, 824 of whom received oxaliplatin-based chemotherapy. A decision to discontinue panitumumab treatment in the trial was based on a preplanned interim efficacy analysis scheduled after the first 231 events, which revealed a statistically significant difference in PFS in favor of the control arm. An unplanned analysis of OS also demonstrated a statistically significant difference favoring the control arm.²⁹ Interim safety analysis showed an increased incidence of grade > 3 severe events of diarrhea, dehydration, and infections in the panitumumab-treated patients.

• The BOND-2 study was a randomized phase II trial in patients with irinotecan-refractory colorectal cancer not previously treated with bevacizumab or cetuximab.³⁰ Patients in arm A received irinotecan at the same dose and schedule as last received before study entry, plus cetuximab 400 mg/m² loading dose, then weekly cetuximab 250 mg/m², plus bevacizumab 5 mg/kg administered every other week. Patients in arm B received the same cetuximab and bevacizumab dosages as in arm A but without irinotecan. Toxicities were as expected from the single agent experiences. For the three-drug program, TTP was 7.3 months and the RR was 37%. For the combined biologics arm that did not include chemotherapy, TTP was 4.9 months and the RR was 20%. OS for arm A was 14.5 months and for arm B was 11.4 months. The activity seen with the addition of bevacizumab to cetuximab, or to cetuximab plus irinotecan, seems to be favorable when compared with historic controls of cetuximab or cetuximab/irinotecan in patients who are naïve to bevacizumab. This study provided evidence that the theoretical interaction between EGFR and vascular endothelial growth factor (VEGF) inhibition led to higher-than-expected response rates and TTP as compared with use of the same agents alone.

• This finding contrasts with those from the PACCE study, which used panitumumab rather than cetuximab and also a different chemotherapy that did not include irinotecan. Results of these studies suggest caution in the use of combined anti-EGFR and anti-VEGF agents outside of clinical trials. A recent editorial in the Journal of Clinical Oncology has commented on this point.³¹ Cetuximab is a chimeric immunoglobulin G1 monoclonal antibody with the potential to induce antibody-dependent cellular cytotoxicity, whereas panitumumab is a fully human immunoglobulin G2 monoclonal antibody without any predictable antibody- dependent cellular cytotoxicity activity. An issue related to the type and doses of the chemotherapy regimens used in these two studies may be relevant. In the BOND-2 study, single-agent irinotecan was administered at the dose and schedule shown to be safe in each patient with the previous treatments, whereas in the PACCE study, full doses of oxaliplatin-based chemotherapy (mainly FOLFOX) were administered. Although available safety data do not suggest a different toxicity profile for panitumumab and cetuximab, a specific interaction between panitumumab and bevacizumab, resulting in more adverse effects compared with the combination of cetuximab and bevacizumab, cannot be ruled out. It is also possible that the increased toxicity in the panitumumab arm resulted in a shorter duration of therapy because of early treatment discontinuation, which also would have interfered with PFS and OS. This is supported by the lower dose-intensities for chemotherapy and bevacizumab observed in the panitumumab-containing arm.

• The ongoing Cancer and Leukemia Group B/Southwest Oncology Group 80405 National Cancer Institute Intergroup phase III study of first-line treatment of metastatic CRC allows the physician’s choice of chemotherapy (FOLFOX or FOLFIRI), with patients randomly assigned to receive bevacizumab alone, cetuximab alone, or the combination of both. The primary study objective is to demonstrate an increase in OS in the doubleantibody arm. Because of the PACCE IDMC alert, the Cancer and Leukemia Group B IDMC conducted a safety evaluation of the patients enrolled in the 80405 study, and concluded that the study should continue recruiting patients because no potential safety imbalances among the three arms were observed.³² An early activity analysis is planned as a consequence of the PACCE trial findings.

IV-A. 2. Special Populations (Elderly and Poor Performance Status)

Several recent pooled analyses have addressed the benefits of combination chemotherapy in older and poor performance score patients. Older patients with advanced disease treated with FOLFOX regimens in the first- and second-line settings were found to tolerate these regimens as well as younger patients and to have similar benefits in terms of RR, PFS, and OS as compared with patients younger than age 70 years.³³ A similar analysis showed comparable findings with irinotecan-based combination therapy.³⁴ A pooled analysis of more than 6,000 patients enrolled in nine randomized clinical trials supports the judicious use of combination therapy in patients with PS of 2, who derived a similar magnitude of benefit from the best regimens in each trial as did patients with PS of 0 or 1.³⁵ The incidence of grade ≥ 3 nausea and vomiting is significantly augmented in patients with PS of 2; the rate of early deaths within 60 days of study enrollment is also higher in PS 2 vs. PS 0,1 patients; 12% vs. 3%, respectively.

V. FUTURE DIRECTIONS

V. A. Comments on Research

The management of patients with advanced colorectal cancer has become increasingly complex–to the point that this area of patient management defies the creation and updating of timely, relevant practice guidelines. The most exciting development in 2007 is arguably the discovery that patients whose tumors harbor K-ras mutations do not benefit from currently available EGFR-targeting therapies. The ability to have biologically relevant predictive factors has been elusive until this discovery. Exactly how best to combine the available agents–whether to use them as serial single agents or to exploit them as multiple drug combinations–remains unclear and likely varies depending on the clinical scenario. How and whether to integrate treatment holidays into ongoing management and whether some agents are best continued during a maintenance phase remain to be discerned. The Spanish TTD Cooperative Group is conducting a 475-patient study comparing CapeOx (capecitabine/oxaliplatin) + bevacizumab every 3 weeks vs. six cycles of CapeOx + bevacizumab followed by bevacizumab monotherapy. In both arms, treatment is continued until disease progression is observed. The primary study objective is comparison of PFS.

Other unanswered issues include whether cetuximab and bevacizumab are best partnered with one chemotherapy agent or a combination of agents and what the ideal combinations should be. When and if pharmacogenetics should be assessed in the nonstudy setting for individualization of drug regimens also remains uncertain. Identification of which patients will or will not benefit from surgical management of metastatic disease is a pressing issue. There are many relevant questions to be addressed in ongoing and future trials in patients with advanced colorectal cancer. To access information about active clinical trials listed in the National Cancer Institute data base, go to http://www.nci.nih.gov/clinicaltrials.

V. B. Obstacles to Progress

Drug availability and cost have become major issues for managing colorectal cancer. Restricted availability of some agents and the impact of skyrocketing costs on fixed resources are two concerns creating ethical and fiscal dilemmas worldwide.