Circulating tumor DNA (ctDNA) was found to have utility for EGFR mutation testing in advanced non–small cell lung cancer (NSCLC) in a real-world setting in the diagnostic ASSESS study [NCT01785888], according to Martin Reck, MD, PhD, Lung Clinic Grosshansdorf, Grosshansdorf, Germany.

The study enrolled 1288 eligible patients, with 997 from Europe and 291 from Japan. Overall, 75.8% of the patients were white and 23.0% were Asian; 19.6% were never-smokers; smokers had 40.0 median pack-years; and the majority of patients (84.6%) had stage IV disease.

The majority of the tissue/cytology samples were obtained during the current diagnosis, derived from the primary tumor, and collected via bronchoscopy. Most samples were prepared as paraffin-embedded tissue blocks and fixed with 4% neutral-buffered formalin. The median turnaround time for EGFR mutation testing was 11 days in Europe (95% CI, 14.0 to 17.3) and 8 days in Japan (95% CI, 8.2 to 14.1). The average test success rate was 98.3% in Europe and 99.6% in Japan.

In Japan, the tests used to evaluate tissue/cytology samples and plasma samples for EGFR mutations were Cycleave PCR and PNA LNA clamp PCR. In Europe, for tissue/cytology testing, PNA LNA clamp PCR and the older methods of DNA sequencing and pyrosequencing were used, along with newer, more sensitive methods, including the Roche cobas EGFR Mutation Test and Sequenom; for plasma testing, the QIAGEN Therascreen RGQ PCR kit and Roche cobas EGFR Mutation Test were used.

The overall concordance was 89.1% (1035 of 1162 patients; 95% CI, 87.1 to 90.8) and overall positive predictive value (PPV) was 77.7% (87 of 112; 95% CI, 68.8 to 85.0). In patients in whom the same testing method was used for tissue/cytology and plasma evaluations, the PPV was 92.6% (95% CI, 75.7 to 99.1) compared with 72.9% (95% CI, 62.2 to 82.0) when different testing methods were used for the evaluations. The sensitivity was 46.0% (95% CI, 38.8 to 53.4), specificity was 97.4% (95% CI, 96.2 to 98.3), and the negative predictive value was 90.3% (95% CI, 88.3 to 92.0) in the overall cohort.

The QIAGEN Therascreen RGQ PCR kit had a sensitivity of 72.7%, specificity of 99.1%, and PPV of 94.1% in this trial. A previous trial of white patients, IFUM [Douillard JY et al. Br J Cancer. 2014], used the same kit and reported a sensitivity of 65.7%, specificity of 99.8%, and PPV of 98.6%.

False-positive results, meaning an EGFR mutation-positive plasma sample and an EGFR mutation-negative tissue/cytology sample, were believed to have come from 25 patients. These patients were from multiple sites and countries, indicating no specific laboratory-based issues. Among these patients, 56% of the tumors were tested by DNA sequencing or pyrosequencing (vs 25% of the overall population), 76% of the patients were never-, former-, or light-smokers (vs 45% of the overall population), and 32% of the tumor samples were needle biopsies/cytology (vs 21% of the overall population). The false-positive rate may have been contributed to by possible over-representation of cytology samples, meaning inadequate tumor samples, or by use of the less sensitive DNA sequencing or pyrosequencing methodologies that had inadequate mutation analysis to detect mutation.

Among the 191 patients overall who were EGFR mutation positive, 30.6% of Japanese patients (86 of 281) and 11.6% of European patients (105 of 903) were positive. The exon 19 deletion was found in 51.3% (n = 40) of Japanese patients and in 54.5% (n = 54) of European patients. The L858R mutation only was found in 47.4% (n = 37) of Japanese patients and 28.3% (n = 28) of European patients. EGFR mutation-positive status was significantly correlated with female sex, ADC histology, never-smoking status, and Japanese ethnicity (all P < .001).

EGFR mutation status was the largest driver of therapy choice. The most common first-line treatment decisions for all EGFR mutation-positive patients were gefitinib, erlotinib, and afatinib; EGFR mutation-negative patients most commonly received pemetrexed, radiation therapy, carboplatin, and cisplatin.

Although practices for both tissue/cytology and plasma samples require improvements, these real-world data from the large, observational ASSESS study suggest that ctDNA may be feasible and suitable for analyzing EGFR mutations. The overall concordance of EGFR mutation status was 89%.