Analysis of ctDNA from plasma may enhance, extend, and complement tissue analysis in the identification of critical gene mutations, particularly with respect to detection of mutations conferring resistance to therapeutic modalities. That’s according to data from a study comparing somatic ctDNA alterations in plasma analysis detected in plasma with tissue samples in The Cancer Genome Atlas (TCGA).
The study was presented by Philip C. Mack, PhD, of the UC David Health System, during the Tumor Biology Oral Abstract Session, held on June 7. According to Dr. Mack, the analysis of ctDNA proved highly effective at detecting low-frequency mutant allele frequency, and it demonstrated high sensitivity and positive predictive value.
“Overall, 27% of cases had potentially actionable resistance mutations detected,” Dr. Mack said. “It has been known for decades that cancers, particularly those that are metastatic or invasive, shed tumor DNA into peripheral circulation. The question becomes: Can we can use that tumor-specific information available in the plasma compartment to predict responses to targeted therapies?”
Comparative Analysis of ctDNA and Tissue Samples
For the analysis, DNA was isolated from blood from 17,628 samples of 15,191 patients with stage III or IV cancers; fragmented DNA was barcoded using a high-efficiency electrical tagging approach, with target capture performed to identify regions of interest in 70 key cancer-associated genes. According to Dr. Mack, the sequencing analysis looked for mutations associated with more than 50 known cancer types, and variant calling included missense mutations, small insertions into lesions, amplification events, and a limited number of fusions. Alterations were detected in 83.4% of samples.
The results from the DNA analysis were compared with 9,077 samples from 9,077 patients gathered from the TCGA, in which 94% were identified as having mutations. There were some crucial differences in the reference groups, Dr. Mack said, including that the TCGA dataset included patients with stages I-IV cancer, with isolates collected prior to treatment, whereas, in the ctDNA group, sample collection occurred most typically after a second line of therapy had been attempted—a mean 342 days after initial diagnosis.
Forty-nine percent of those in the ctDNA group were resistant to at least one U.S. Food and Drug Administration (FDA)-approved therapy, but no such finding was noted in the TCGA group.
Overall, half of all variants detected in the ctDNA dataset occurred below 0.4% mutant allele frequency.
“One of the advantages of using next-generation sequencing in plasma is that it reports a fairly precise mutant allele frequency,” Dr. Mack said. “This allows you to easily discriminate the presence of germline polymorphisms from the somatic mutations.”
The analysis showed that ctDNA demonstrated high detection rates across most cancer types, and, overall, average clinical sensitivity was 85%.
The positive predictive value of plasma DNA analysis, analyzed in a subset of cases in which tumor information was available, ranged from 94% to 100% concordance with tissue data; however, the same pattern was not true with subclonal mutations, which are associated with resistance.
Dr. David N. Hayes
Dr. Mack suggested that plasma DNA analysis may offer clinical utility in its ability to detect mutations to FDA-approved therapies and actionable resistance mutations. For example, ctDNA demonstrated the ability to detect T790M mutations on the EGFR gene, which arise late in the disease course and were rarely seen in tissue samples. In addition, Dr. Mack said that a subanalysis of non–small lung cancer samples suggested that use of ctDNA in conjunction with tissue samples would increase biomarker yield by as much as 42%.
Discussing the paper, David N. Hayes, MD, MPH, of the UNC Lineberger Comprehensive Cancer Center, said that testing modalities that utilize next-generation sequencing may someday prove “as important as microscopes and x-rays in how we [treat patients with] cancer,” but there is a need for studies that clearly demonstrate the clinical utility (perhaps via biomarker studies) and concerted efforts to address the inherent regulatory and reimbursement hurdles.
Failure to address these practical questions may affect whether these modalities are available for clinical use. Dr. Hayes said that greater scientific collaboration is needed to move the idea of next-generation sequencing from theoretical benefit to definite clinical tool. He cited the relatively slow adoption of identifying EGFR mutations for their potential to effect efficacy of gefitinib in treatment of lung cancer; it took 9 years between the time the mutant gene was identified until molecular testing for that particular mutation was codified into national consensus statements.
“It took 9 years to recognize EGFR, which is a no-brainer. What happens when you have something more complex, such as next-generation sequencing?” Dr. Hayes asked.
-- Bryan Bechtel