The Use of Biomarkers to Guide Cetuximab Therapy

Roy S. Herbst, MD, PhD
The University of Texas MD Anderson Cancer Center, Houston, Texas

Lung cancer is the most significant cause of cancer death in the United States, with only 16% of patients surviving 5 years after diagnosis. Poor survival reflects in part the fact that almost 50% of cases present with advanced-stage disease, for which effective systemic therapies are limited. However, the recent development of novel molecular therapies, such as inhibitors of the epidermal growth factor receptor (EGFR), offer considerable promise for subsets of patients sensitive to these agents. The EGFR tyrosine kinase inhibitors (TKIs) erlotinib and gefitinib were the first molecularly targeted agents to demonstrate benefit in patients with advanced non–small cell lung cancer (NSCLC). Early studies defined patient characteristics of those most likely to respond as: female, adenocarcinoma or bronchioloalveolar cell carcinoma (BAC) histology, never- smoker, and Asian. BR.21, an NCI-Canada trial comparing erlotinib to placebo as second-line therapy in unselected patients, was the first demonstration of survival benefit for any targeted agent in NSCLC.⁴⁵ A similar placebo-controlled trial with gefitinib, the ISEL study, did not show a survival benefit, perhaps due to differences in patient eligibility criteria.¹⁵⁶ Even in BR.21, the response rate to erlotinib was <10% and the overall impact on survival was modest, prompt ing research efforts to define molecular biomarkers predictive of clinical benefit, especially improved survival.⁴⁵ Although a number of potential EGFR biomarkers have been proposed, including EGFR protein expression, mutation status, and gene copy number defined by FISH, EGFR FISH was found among the most effective in predicting survival benefit in the two placebo-controlled randomized trials. A hazard ratio (HR) for EGFR FISH-positive patients of 0.44 was seen in the BR.21 study and 0.61 in ISEL.^155,156,157

The presence of certain EGFR tyrosine kinase domain (TKD) mutations defines an NSCLC patient subset in which dramatic responses to EGFR-TKIs are often seen. Outside of Asian populations this biomarker is not yet clearly associated with prolonged survival from EGFR-TKI therapy, though the recent IPASS trial suggests this might be. This discrepancy perhaps reflects the low incidence of these TKD mutations in Western populations of NSCLC patients. For example, there was no improvement in survival in patients with EGFR mutations in the BR.21 trial, despite an association with improved response rates.^155,156 In terms of prognosis, i.e., independent of therapy, studies performed to date suggest that EGFR mutations are actually associated with a good prognosis and more indolent course. In the TRIBUTE study of chemotherapy with or without erlotinib as first-line therapy, patients with EGFR-mutated tumors had a better survival independent of therapy.⁴⁹ There are at present no published patient data regarding EGFR mutations under the influence of cetuximab-based therapy. However, preclinical data suggest that in contrast to the TKIs, NSCLC cell lines harboring EGFR mutations do not exhibit hypersensitivity to cetuximab.⁴¹

The prognostic and predictive value of EGFR protein expression by immunohistochemistry (IHC) has been studied in a relatively large number of reports, with mixed results. This is perhaps due to heterogeneity in the patient populations studied as well as variability in methodologies. In a meta-analysis by Meert and coworkers using an eight-study subset, EGFR expression by IHC was modestly associated with a poor prognosis (HR 1.13, 95% confidence interval [CI] 1.00-1.28).¹⁵⁸ In BR.21, using the EGFR Pharm Dx(TM) kit (Dako), which defines positivity as staining in >10% of tumor cells, 57% of tumors tested were positive. EGFR protein expression defined in this manner was not a strong prognostic factor.^155,156

However, there was predictive value. In IHC-positive patients receiving erlotinib, the HR for survival was 0.68 (P=0.02), with out a significant interaction coefficient (P=0.25).¹⁵⁵ The ISEL study of gefitinib versus placebo showed a HR of 0.77 (95% CI 0.56-1.08) with a borderline significant interaction coefficient (P=0.05)].¹⁵⁷ In the BR.21 study there was no difference in survival between erlotinib versus placebo in IHC-negative patients, while in ISEL, IHC-negative patients had a tendency for shorter survival with gefitinib (HR 1.57, 95% CI 0.86-2.87). EGFR FISH appears to have both prognostic and predictive implications in NSCLC. The concept of EGFR FISH as a predictive biomarker for EGFR-TKIs originated with SWOG investigators, led by Dr. Fred Hirsch. They developed the assay classification system that has been commonly used.^159-161 EGFR FISH has previously demonstrated predictive value for efficacy of gefitinib and erlotinib in advanced NSCLC in a number of clinical studies, including BR.21 and ISEL. Recent data, however, have further demonstrated the importance of this predictive biomarker for NSCLC.

S0342

S0342 was a recently completed phase II selection design study of chemotherapy (paclitaxel-carboplatin) with either sequential or concurrent cetuximab, conducted to determine the preferred schedule for subsequent phase III testing against chemotherapy alone. Two hundred twenty-nine patients with advanced NSCLC were enrolled in the study. EGFR FISH analysis was performed in 76 patients with available tumor tissue, which was classified as positive (>4 gene copies per cell in >40% of the cells, or EGFR gene amplification) in 59.2% of patients (Figure 6). This FISH-positive patient cohort had no significant differences in clinical characteristics or outcomes when compared with the overall study group. Increased EGFR gene copy number by FISH was associated with clinical benefit from cetuximab plus chemotherapy.¹⁵³ Objective response rate (complete and partial response [CR/PR]) was numerically higher in FISH-positive patients (45%) compared with FISH-negative patients (26%; P=0.14), while disease control rate (CR/PR plus stable disease) was statistically superior (81% vs 55%; P=0.02). Patients with FISH-positive tumors had a median PFS of 6 months versus 3 months for FISH-negative patients (P=0.0008); median survival was 15 and 7 months, respectively (P=0.04) (Figure 7). Furthermore, overall survival favored FISH-positive patients receiving concurrent therapy, and was statistically significant compared with FISH-negative patients only in the concurrent arm.

Figure 6

These results are the first to suggest that EGFR FISH status is a predictive factor for selection of NSCLC patients for cetuximab plus chemotherapy. Due to the relatively small sample size and retrospective nature of the data, at this point observations in S0342 are only hypothesis-generating, and will be validated in a randomized Phase III study S0819.

Even more recently, SWOG completed a phase II trial (S0536) in 99 patients testing the combination of carboplatin, paclitaxel, bevacizumab, and cetuximab. This study, for which the primary endpoint was incidence of hemorrhagic toxicity, has shown this 4-drug regimen to be safe for future study, with an acceptable toxicity profile when compared with either our S0342 results or the E4599 results.¹⁶⁵

The clinical trials summarized above suggest fundamental differences in how chemotherapy interacts with EGFR-TKIs com pared with the monoclonal antibody cetuximab in NSCLC, which can be further exploited by biomarker selection of patients most likely to benefit. EGFR FISH has considerable potential to change practice and impact patient care for patients receiving cetuximab plus chemotherapy. While the FLEX study results can be considered a positive lead, selection was based on IHC positivity only. EGFR FISH offers a potentially better way to select patients for this combination therapy (given the data presented in the background), and is highly applicable to clinical practice. Routine use of FISH technology is widespread, as HER2 FISH is routinely performed on paraffin-embedded material for selection of breast cancer patients for trastuzumab therapy. Adoption of EGFR FISH for NSCLC therapy could similarly become practice-changing.

Figure 7

< < PREV

NEXT > >