Novel Agents in the
Treatment of Lung Cancer

New Approaches to First-Line Therapies

Using Markers to Select Patients for Anti-EGFR Antibody Therapy

Glenwood D. Goss, MD, FCPSA, FRCPC
Ottowa Hospital Cancer Center

In contrast to EGFR TKIs for which there are known clinical predictors of response and survival, it has been more difficult to determine clinical predictors for anti-EGFR antibody therapy. Although a number of these antibodies are still in early development, the most extensive published data are on cetuximab. This IgG1 monoclonal antibody binds to the extracellular domain of EGFR, thus blocking ligand-activated signal transduction and receptor dimerization and promoting receptor internalization and degradation. Cetuximab may also promote antibody-dependent cell-mediated cytotoxicity (ADCC), which could contribute to its anticancer activity.²²

Molecular Predictors of Efficacy

Molecular predictors of efficacy include EGFR mutation status, EGFR copy number, EGFR protein expression, K-ras mutation status, proteomics, plasma ligands, and EGFR SNP analysis. The best studied among these are EGFR mutation status, copy number and protein expression, as well as K-ras mutation status, with K-ras mutations being a negative predictor of efficacy. Ongoing studies are evaluating the predictive value of plasma ligands, protoemics, and EGFR SNP analyses.

All of these biomarkers will be prospectively evaluated in the large phase III biomarker validation study of second-line therapy, in which patients with advanced NSCLC will be randomized to pemetrexed or erlotinib (NC CTG NO723; the MARVEL study). In this trial, all patients will be required to have tissue samples for EGFR marker analysis including protein expression, gene copy number, and mutation status, with stratification according to high or low EGFR copy number. In addition to EGFR markers, pemetrexed pharmacogenomics will be studied. Based on our current understanding of molecular markers, EGFR FISH positivity and high copy number are positive predictors of response, and these may predict for survival in patients with advanced NSCLC. EGFR gain-of-function mutations in codons 19 and 21 of the internal tyrosine kinase domain are positive predictors of response, PFS, and possibly OS.^10,29,30

To date, clinical predictors of response, PFS, and OS have been studied prospectively in the 2 large phase III randomized studies (BMS 099, FLEX) and the meta-analysis of the 4 randomized studies (LUCAS, BMS 099, FLEX, and BMS 100) as discussed above. Prespecified subgroup analyses in FLEX and BMS 099 showed that a benefit was associated with cetuximab therapy independent of gender, performance status, tumor histology, and smoking status, the same clinical features predicting response and survival with EGFR TKIs.^5,8 These observations were confirmed in the subgroup analyses of the larger meta-analysis.²²

Early Rash as a Marker?

Interestingly, rash had been reported to be an important predictive factor in a number of other trials with EGFR inhibitors.^31-34 Consequently, FLEX investigators have recently reported a subgroup analysis of patients who did and did not develop rash within the first cycle (21 days) of treatment.^33,35 Of 557 patients treated with cetuximab, 518 were available for analysis (i.e., were alive at day 21). Those who developed grade ≥1 acneiform rash (56%) had a HR for survival of 0.631 (95% CI, 0.515-0.774; P<0.001) compared with those who did not develop an early rash. Median survival in the group with rash was 15.0 months compared to 8.8 months in the group without rash. In addition, patients who developed rash had higher response rates (44% vs 28%) and longer PFS (5.4 months vs 4.3 months). A retrospective analysis of the BMS 099 study showed similar findings. Of the 338 patients treated with cetuximab, 315 were alive at day 21 and included in the analysis. Forty-one percent did not develop early rash, while 59% did. The group with early development of rash had better survival, with a HR of 0.76 (95% CI, 0.59-0.98), and a median survival time of 10.4 months versus 8.9 months in those without early rash, confirming the findings in the FLEX trial.³⁴ Therefore, to date rash is the only clinical predictor of efficacy for cetuximab in NSCLC.

Molecular Predictors of Response

Attempts to identify molecular markers of response, PFS, and OS in patients with advanced NSCLC treated with cetuximab have been singularly unsuccessful. The identification of K-ras mutations as a predictor of improved outcome in colorectal cancer led to the analysis of samples from 395 of 1125 patients enrolled in the FLEX trial. Nineteen percent of the specimens assayed had K-ras mutations, but mutation status did not predict outcome in this subgroup. Similar findings were seen in the BMS 099 trial, in which 225 of 676 patient samples were available for analysis.⁴ Seventeen percent (35/202) had K-ras mutations, but there was no difference in response or OS (HR 0.97; 95% CI, 0.67-1.30; P=0.93) between the patients receiving chemotherapy only or chemotherapy plus cetuximab. The small sample size may have influenced the results, however.

In an earlier small, randomized study of concurrent versus sequential cetuximab with first-line chemotherapy, high EGFR gene copy number was associated with better outcomes with cetuximab and chemotherapy.³⁶ In 279 samples evaluated in the FLEX trial, 37% had high EGFR gene copy number by FISH analysis, whereas 54/104 (52%) had high gene copy number in BMS 099. In neither trial was EGFR copy number predictive of better outcome with cetuximab. Surprisingly, in BMS 099 survival was longer in EGFR FISH-positive patients treated with chemotherapy alone in this small sample (P=0.03). PFS was longer in those with IHC-negative tumors compared to patients with IHC-positive tumors in both treatment groups (P=0.048). Finally, in FLEX, EGFR mutation status was determined in 293 of 1125 patients, and 15.4% demonstrated an EGFR activating mutation. The presence of an EGFR mutation was associated with better survival compared with EGFR wild-type in both treatment groups but was not predictive of benefit. Thus, EGFR mutations were prognostic in this patient population. Similar findings were seen in the BMS 099 trial, in which EGFR activating mutations were found in 17 of 167 samples (10% of those assayed), but again were not predictive of benefit when cetuximab was added to chemotherapy.^4,35

Conclusion

Early onset of rash appears to be a clinical predictor of benefit when cetuximab is added to chemotherapy compared to chemotherapy alone. Molecular markers, including K-Ras, EGFR gene copy number, and EGFR activating mutations, do not predict for benefit from cetuximab. Further research is required to define the population that benefits most from anti-EGFR monoclonal antibody therapy.

< < PREVIOUS

NEXT > >