The advent of targeted therapeutics has necessitated the use of genotypic testing for cancer patients, as the presence of certain driver mutations may suggest specific pharmaceutical susceptibilities. One such patient population is those having non-small cell lung cancer (NSCLC). In these patients, 10-40 percent have a driver mutation for the epidermal growth factor receptor (EGFR), such as Exon 19 deletions or L858R mutations, which suggests a positive response to the use of first generation EGFR tyrosine kinase inhibitors (TKI) (e.g., erlotinib and gefitinib).
Another important sub-population of NSCLC patients are those having a fusion of the echinoderm microtubule-associated protein-like 4 (EML4) and the anaplastic lymphoma kinase (ALK) genes. This EML4-ALK fusion occurs primarily in adenocarcinomas and is present almost exclusively in non-smokers. Patients with this gene fusion are recommended for therapy with crizotinib, or more recently, alectinib.
Prolonged therapy (12-24 months) with EGFR TKIs often gives rise to acquired resistance in these patients. A number of mechanistic pathways for this resistance are known, including ERBB2 and MET amplifications, point mutations in EGFR, as well as mutations in a number of other genes that activate alternative or downstream pathways. The most common resistance mechanism in this group of patients is the EGFR T790M mutation, in which a threonine residue in the ATP binding pocket is switched to a methionine. This particular mutation is observed in 50 percent of cases of acquired resistance to EGFR-TKI therapy. For these patients, therapy with the next-generation EGFR-TKI osimertinib is often recommended. Given the number of mutations that are therapeutically relevant to NSCLC patients, genotypic testing is crucial for desired patient outcomes.
Circulating Tumor DNA
Traditionally, genotypic testing has been done utilizing tumor tissue, however, the methods by which these biopsies are done can be quite invasive. Additionally, quite often, DNA of insufficient quality or quantity is obtained for tumor genotyping. To address these potential shortcomings, a study was undertaken by Erica Carpenter, MBA, PhD, Director of the Circulating Tumor Material Center at the University of Pennsylvania's Perelman School of Medicine, Philadelphia. In that study, cell-free circulating tumor DNA, isolated from a patient's blood, was utilized in combination with next-generation sequencing to molecularly profile NSCLC patient tumors.
As Carpenter explained, "Circulating tumor DNA is released into the blood by tumor cells that have undergone necrosis or apoptosis. This method of DNA isolation is much less invasive, as only a needle stick, such as what is routinely done for a blood test in the doctor's office, is required to obtain the sample.
"In many instances, this testing can be done when routine bloodwork is being performed, as it only requires the use of an additional tube when samples are drawn" she revealed. "Having a less intrusive means of sample taking is important for increased patient testing, as many advanced-stage cancer patients who have had a previous biopsy are more likely to refuse a subsequent one because of the discomfort that may be involved," related Carpenter. "This becomes even more important for disease surveillance, as acquired mutations often arise in this patient population while they are on therapy."
Another important feature of this protocol is the potential for obtaining a more comprehensive genotypic survey of the cancer present in the patient. In traditional tumor biopsies, a needle is injected into a tumor site where there may be cells of different genetic makeup present, depending on the tumor architecture. Additionally, if the cancer has metastasized, there are tumors located in different parts of the body that may be different genetically.
Put succinctly, "There can be significant intra- and inter-tumor variability present in a patient, and if one is looking at only a piece of a single tumor, then one may miss the larger picture of what is going on within the patient," Carpenter noted.
The ctDNA testing done utilizes DNA shed from tumors all over the body, thus it is thought to provide a more thorough view of the genetic landscape present in the patient's cancer.
Study Specifics
This study, which was described in a recent Clinical Cancer Research article (DOI: 10.1158/1078-0432.CCR-16-1231), was performed with 102 cancer patients being treated at the University of Pennsylvania's Abramson Cancer Center from February 2015 to March 2016. From these patients, 112 ctDNA tumor samples were obtained; however, tumor tissue biopsies were only successfully obtained for 50 of the same 102 patients. Thus, for more than half of the patients (52), the ctDNA methodology was the only means by which useful genetic data were obtained. Of these 102 patients, two had cancer of undetermined origin, while all others had NSCLC, 68 percent were women, 81 percent displayed adenocarcinoma, and 96 percent had stage IV disease.
For 36 of the patients, ctDNA screening was done using a panel which detects point mutations in 68 genes, fusions in six genes, and insertions-deletions in EGFR, however, the remaining 66 patients were screened using an expanded 70-gene panel. This expanded panel additionally provides full exon coverage for mutation detection in RB1 and TSC1, as well as insertions-deletions for MET and ERBB2. In comparison, 38 patients had tumor tissue DNA (tDNA) screening using a 47-gene panel, while the remaining 12 only had enough tDNA obtained for screening using a smaller 20-gene panel.
Detectible variants were obtained for 86 of the 102 patients using ctDNA, with only 16 patients showing no ctDNA present in their sample. For those without detectable ctDNA, this can mean the tumor is not shedding DNA, there are variants present but not covered by the sequencing panel, the panel variants present are below the 0.1 percent allelic fraction (AF) assay detection limit, or the disease is being therapeutically controlled (the AF for a variant is lowered with efficacious therapy).
EGFR mutations were noted in 20 percent of the ctDNA samples taken, with the most common being EGFR Exon 19 deletions (16 patients) and L858R (10 patients). Ten ERBB2 mutations were present, of which, five were exon 20 insertions which may be targeted with agents such as trastuzamab, afatinib or lapatinib. Of the 32 patients receiving EGFR-TKI therapy, 10 showed the T790M acquired resistance mutation. Of these 10, all still showed the original EGFR activating mutation. Most importantly, eight of these 10 patients only had ctDNA analyses done, as no usable tDNA was obtained via biopsy. Two samples also showed the presence of the EML4-ALK fusion, suggesting the use of an ALK TKI for treatment.
Overall, 70 percent of the patients tested had a relevant clinical trial available for participation. FDA-approved variant-specific therapies were available for 32 of the patients, while potential off-label targeted therapies were available for 56 patients.
Because there was variability in the time between tDNA and ctDNA sampling, an effort was made to correlate the concordance between these samples and the time between when they were taken. The concordance values calculated for time intervals of <=2 weeks, <=2 months, <=6 months and 6 months were 100 percent, 92 percent, 94 percent, and 60 percent, respectively. These results showed a significant correlation for the concordance and the duration between tissue biopsy and ctDNA blood draw (p=0.038).
When asked about future directions for research, Carpenter discussed the possibility of utilizing cell-free DNA (cfDNA) for predicting patient outcome. "In general, there is an association between lower cell free DNA and improved survival."
The mean cfDNA level for patients that died during the study was significantly higher than that of patients who lived through the study (4.0 vs. 1.6 ng/[mu]l; P<0.001).
For metastatic patients having a cfDNA concentration of >=3 ng/[mu]l, a median survival of 24 months from time of metastatic diagnosis was measured, while the value for those with <3 ng/[mu]l was 46 months (log-rank, p<0.01). These figures remained statistically significant even after several factor adjustments were made (age, number of metastatic tumor sites, EGFR status and performance status). Carpenter noted, "A more thorough investigation of this method as a means to gauge patient survival is ongoing."
Richard Simoneaux is a contributing writer.