Testing for T790M mutations with plasma and urine assays can complement testing done from biopsy tissue in patients with epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI)-resistant non–small-cell lung cancer (NSCLC), according to results of a new study presented at the 2016 American Society of Clinical Oncology (ASCO) Annual Meeting, held June 3–7 in Chicago (abstract 9001). Response to the EGFR inhibitor rociletinib was similar whether T790M mutation status was tested with biopsy or plasma or urine.
Heather A. Wakelee, MD, of Stanford University Medical Center in Palo Alto, California, presented results of the analysis based on the phase I/II TIGER-X study, during an oral abstract session. She noted that rociletinib clinical development has been discontinued, but the mutation testing results are still broadly applicable.
The TIGER-X study of rociletinib included 548 patients in the safety population and 443 patients in the efficacy population. For this analysis of EGFR testing methods, there were 540 tissue samples, 482 plasma samples, and 213 urine samples. All patients in the trial were EGFR mutation–positive, the median age was 63 years, and most patients (67.9%) were women.
In total, 387 patients were T790M-positive by central tissue genotyping, using a real-time polymerase chain reaction (PCR) assay. Of the 482 patients assessed by the BEAMing plasma test, 374 were T790M-positive. Finally, of 169 patients assessed with next-generation sequencing in urine, 136 were T790M-positive.
“The sensitivity [of plasma testing] if you use tissue as the gold standard, is 80.9%,” Wakelee said. Of the 387 positive tissue samples, 313 plasma samples were also positive, and 74 were negative. Wakelee also noted that the plasma test was particularly useful in patients who had tissue samples deemed inadequate to determine mutation status (38 samples).
The sensitivity of the urine analysis was 81.1%. Of 213 patients with both urine and tissue analysis, there were 175 positive results with tissue testing, and 142 of those were also positive on urine analysis. A total of 104 samples (57% of available) were positive by all three sample types.
The response rates to rociletinib were not substantially different based on the sample type. The objective response rate in positive tissue samples was 33.9%, compared with 32.1% in positive plasma samples, and 36.7% in positive urine samples. Other markers of response, such as duration of response and progression-free survival, had similar results, and it did not matter in what sample type the T790M mutation was found in. Wakelee noted that it did appear to be more difficult to find the mutation with plasma and urine testing in patients with M1a/M0 disease—meaning metastases only in the lung—than in those with M1b (distant metastases) disease.
“T790M plasma, tissue, and urine tests complement one another. Each test identifies cases missed by the other tests,” Wakelee said. “EGFR mutation detection from plasma and urine should be considered viable approaches, particularly when tumor tissue is not available.”