NEW YORK – Next-generation sequencing test results led oncologists to change their treatment approaches for nearly 16 percent of solid cancer patients, suggesting to researchers from Princess Margaret Cancer Centre in Toronto that the test may have clinical utility.
The retrospective analysis, published in JCO Precision Oncology in June, relied on data collected within the Ontario-wide Cancer Targeted Nucleic Acid Evaluation (OCTANE) study, which enrolled more than 4,500 patients with solid tumors for NGS panel testing to develop a clinical data registry for future research. The Princess Margaret researchers further surveyed oncologists who enrolled patients in the OCTANE trial about their treatment decisions after receiving NGS results to explore the test's clinical utility.
Patients enrolled in the OCTANE trial received NGS using either Thermo Fisher's Oncomine panel, which screens 161 genes, or a custom hybridization capture panel of 555 cancer-relevant genes called Hi5. The two panels were included because the trial switched from the Hi5 555-gene panel to the 161-gene Oncomine panel in 2018.
"We were interested in getting a sense of how the [NGS] information was being used by oncologists who ordered the test," said Philippe Bedard, senior author of the study and a clinician-investigator at Princess Margaret Cancer Centre. "[NGS] tests are available and widely used in many jurisdictions. But particularly in jurisdictions like ours that are publicly funded, sometimes access to these tests is limited by reimbursement because there are still questions about whether these tests are useful in terms of informing treatment decisions."
There have been many efforts to evaluate the clinical utility of NGS testing as health insurers consider reimbursement and health systems weigh how much to invest in these tests. Some advocacy groups and experts argue that making these tests available to cancer patients and insurance coverage of the often-costly NGS panels would improve treatment outcomes.
The Princess Margaret researchers surveyed 21 medical oncologists who enrolled patients in OCTANE about their treatment strategies for 394 patients representing 25 tumor types. Breast cancer was the most common tumor type included in the study, followed by head and neck, bowel, and pancreatic tumors. The oncologists were surveyed in two cohorts: 10 were surveyed between 2017 and 2019, and 14 were surveyed in 2021, with three oncologists participating in both cohorts.
The researchers also sent follow-up surveys to oncologists one year after initial treatment to ask about patients whose tumors harbored actionable mutations but did not have a treatment change reported in the first survey, receiving responses from six physicians about 56 patients.
Across all cases, one or more mutations were detected by NGS in 89 percent of patients. Of these, 188 patients (47.7 percent) had one or more mutations from their NGS results classified as actionable by their medical oncologists and 72.3 percent had an actionable mutation as defined by the precision oncology database OncoKB.
Overall, 62 of 394 patients (15.7 percent) had their treatment influenced by the results of NGS testing. The most common treatment change was enrollment in a clinical trial, making up 60 percent of this group. Thirteen patients (21 percent) received an approved drug, four (6 percent) received an off-label therapy, and eight patients (13 percent) avoided an ineffective treatment based on their NGS results.
Gene alterations that most frequently led to changes in treatment were the presence or absence of RAS genes, including KRAS, which influenced treatment for 3.2 percent of patients. The next most common gene alterations to influence treatment were in PIK3CA, ERBB2, and BRCA2. Oncologists' treatment decisions were most frequently influenced by NGS results for colorectal cancer patients, with 40 percent of patients having a change in treatment. About a quarter each of patients with breast, biliary tract, and lung cancer had a treatment change due to NGS results.
For the 127 patients with actionable mutations whose treatment did not change based on NGS testing, their oncologists cited lack of available therapy, stability on the current regimen, clinical deterioration, or patient decision as the reason for not changing treatment. In the follow-up surveys for 56 patients who had actionable mutations but did not have an initial change in treatment, five patients (9 percent) had a subsequent change in treatment based on NGS results within one year.
Bedard said there are several potential reasons why an oncologist wouldn't change their treatment decision based on actionable NGS results, from access issues to their patient's decision.
"Even though the mutation in a tumor type was considered actionable, there may not have been a therapy that was available and reimbursed for that patient," he said. "Because some of these alterations were in the investigational range within OncoKB, there may not have been a clinical trial that was available for the patient because the patient wasn't eligible or there wasn't a slot available at the time."
He also added that reimbursement for off-label treatment can be difficult at a publicly funded institution like Princess Margaret Cancer Centre.
The study also found potential gaps in knowledge for some oncologists and for precision oncology databases. Among the mutations classified as actionable by the oncologists, only 72 percent were also considered actionable by OncoKB. Similarly, 18 percent of patients had a mutation classified as actionable by OncoKB but that was not considered actionable by their oncologist.
"That finding highlights that there may be certain alterations that are not captured by these knowledgebases that may be relevant particularly for clinical trial enrollment," Bedard explained. "Conversely, there may be certain alterations that are clinically actionable, but maybe the treating physicians don't consider them actionable either because of [gaps in] education or access to treatments."
The researchers also noted that the type of NGS panel used did not appear to affect the proportion of patients who had their treatment changed based on the results. The larger 555-gene panel did not result in a greater likelihood of treatment change than the smaller panel used at the beginning of the OCTANE trial. The researchers wrote this suggests that using a panel that captures a larger number of genes "may not necessarily result in improved clinical utility," but noted patients were not randomized to receive the different NGS tests.
The researchers also explored overall survival outcomes for patients who had a treatment change versus those who did not. They found, however, that there was no difference in survival, with the caveats that median overall survival was not reached in both cohorts and that the study was not powered to detect survival differences.
Overall, the researchers concluded that the proportion of patients who did have a treatment change due to NGS results — 15.7 percent — in this real-world setting did support the clinical utility of NGS testing in cancer care.
"For patients whose treatment does get changed [due to NGS results], it can potentially have huge implications in terms of their treatment pathway and outcomes," Bedard said.
Bedard hopes to continue exploring the utility of NGS testing at his institution. He noted that the center currently has an even broader panel in use, which includes tumor mutational burden and microsatellite instability status alongside the 555-gene panel. He also hopes to look beyond this single institution and explore how oncologists use NGS results in other settings.
"The cost of doing these tests is decreasing significantly as technology has improved and there are more potentially actionable mutations that we can identify as more approved drugs and precision medicine drugs are available," Bedard said. "Within the confines of this study, what we're able to address and what our local context [is], I do think that NGS testing is clinically meaningful."