Evolution and Clinical Impact of EGFR Mutations in Circulating Free DNA in the BELIEF Trial

Longitudinal evaluation of mutations in blood samples was a prespecified secondary objective in the BELIEF trial of erlotinib and bevacizumab in advanced EGFR-positive NSCLC. Here, we report the testing results and explore the correlation of EGFR status in blood with clinical outcomes.

Blood samples were prospectively collected from patients at baseline, at response evaluation, and at progression and sent to a central laboratory. Circulating free DNA was purified and EGFR mutations were analyzed with a validated real-time quantitative polymerase chain reaction assay.

EGFR exon 19/21 mutations were detected in 55 of 91 baseline blood samples (60.4%) and correlated with a significantly worse progression-free survival: 11.4 months (95% confidence interval [CI]: 9.0-14.8 mo) for the patients who were positive versus 22.9 months (95% CI: 9.5-33.9 mo) for those who were negative (log-rank p = 0.0020). Among the 74 samples at response, exon 19/21 mutations were detected only in three samples (4.1%). In contrast, 29 of 58 patients (50.0%) were exon 19/21 positive at progression and showed a significantly worse median overall survival of 21.7 months (95% CI: 17.0-30.9 mo) compared with 37.4 months (95% CI: 22.6-53.1 mo) for those who were negative (log-rank p = 0.011). Blood samples at the three time points were available for 48 patients. Of those, among 14 exon 19/21 EGFR-negative at presentation, 13 (93%) were persistently negative for the sensitizing mutations after progression and the p.T790M could only be detected in the blood of two patients.

Longitudinal testing of EGFR mutations in blood can offer valuable clinical information. In patients of the BELIEF study, detection of EGFR mutations in circulating free DNA at presentation was associated with shorter progression-free survival, whereas positivity at progression correlated with shorter overall survival. Finally, patients negative in blood at presentation were almost invariably negative at relapse.