EGFR molecular testing in African-American non-small cell lung cancer patients - a review of discrepant data.

Substantial discrepancy in the literature has recently emerged regarding epidermal growth factor receptor (EGFR) mutational frequency in African American (AA) NSCLC. The first wave of tissue profiling studies, including by our group in 2009, consistently observed a significantly lower frequency of EGFR mutation in AA vs. White NSCLC, whereas three recent reports appear to directly contradict these findings. Reasons for this discrepancy are unclear, but one plausible explanation arises from Simpson's paradox, the consequence of aggregating heterogeneous study cohorts (in this case, the proportion of never-smokers in the study cohort). Our review of all prior studies (combined total 386 AA NSCLC cases) underscores the wide variation in the proportion of AA never-smokers among various studies (13-57%), calling inter-study comparisons into question. In parallel, we assessed objective response by RECIST to EGFR targeted therapy for AA NSCLC in the community setting, prior to the advent of routine EGFR testing. We observed a trend toward reduced response for community-based treatment of unselected AA NSCLC (5%; 3/57) as compared to overall response rates of 10% reported by large North American trials of primarily White NSCLC patients, but this was not significant (P=0.223).