NEW YORK – A clinical study of lung cancer patients' molecular response to Merck's Keytruda (pembrolizumab) determined that levels of circulating tumor DNA (ctDNA) reliably correlated with radiographic response at eight weeks from beginning treatment.
The study, published Monday in Nature Medicine, identified patients with metastatic non-small cell lung cancer (NSCLC) less likely to benefit from pembrolizumab therapy and bolsters arguments for the clinical utility of ctDNA as an early endpoint for measuring therapy response in clinical trials.
The single-arm, open-label, and multicenter Phase II trial, called BR.36, is led by Johns Hopkins University and the Canadian Cancer Trials Group. It is conducted in two stages, the first of which culminated in the Nature paper.
The primary objectives of stage one were to ascertain ctDNA response, to identify when it happens, and to measure how well it corresponds to imaging in defining clinical response to immunotherapy.
"BR.36 stage one is the first clinical trial that was designed to determine what [the] ctDNA molecular response [is] and what is its predictive value with respect to clinical outcomes with immunotherapy," Valsamo Anagnostou, director of the thoracic oncology biorepository at Johns Hopkins and the study's lead author, said via email.
"Despite smaller, retrospective studies [that] have shown that ctDNA dynamics may serve as an early endpoint of response to immunotherapy," Anagnostou added, "there hasn't been a prospective clinical trial to validate the clinical utility of ctDNA in predicting response to immunotherapy."
Anagnostou and her colleagues recruited 50 participants with advanced NSCLC who had not received prior systemic chemotherapy or immunotherapy and showed PD-L1 expression levels of no more than 1 percent.
The researchers collected blood samples before treatment administration on the first day of each of three treatments cycles, taking place at three-week intervals. Serial liquid biopsy analyses were then performed for quantitative assessment of ctDNA dynamics through next-generation sequencing and the use of the Elio Plasma Resolve 33-gene panel from Personal Genome Diagnostics, a Labcorp company.
Of 35 participants with complete evaluable results at the end of the trial, 15 achieved molecular response, defined as maximal mutant allele fraction clearance at the third cycle of pembrolizumab. These patients showed a median response time of 2.1 months, longer progression-free survival (5.03 months versus 2.6 months), and longer overall survival (not reached among those with molecular response versus 7.23 months).
Overall, sensitivity of ctDNA response for Response Evaluation Criteria in Solid Tumors (RECIST) — a set of rules for defining when tumors improve, stay the same, or worsen during treatment — was 82 percent, with a specificity of 75 percent.
Together, these data showed that the trial met its primary endpoint of determining ctDNA response, as well as optimal timing and concordance with radiologic response.
Anagnostou and her colleagues wrote that the success of this study opens a therapeutic window of opportunity for ramping up treatment for patients with molecular disease progression.
The findings in stage one of the BR.36 trial have informed the design of the study's second stage, which will evaluate the potential clinical benefit of tailoring treatment to ctDNA molecular response and is slated to begin in the fourth quarter of this year.
In stage two, Anagnostou said, "patients with lung cancer who do not clear ctDNA after two cycles of pembrolizumab will be randomized to either continuing pembrolizumab or treatment intensification with the addition of chemotherapy."
Stage two is designed as a Phase II/III definitive trial, opening a path toward regulatory approval of ctDNA molecular response as an early endpoint of immunotherapy response.
Validating ctDNA as an early endpoint to support drug approval has been a goal in the liquid biopsy field for some time, but there has not yet been enough evidence gathered to support the approval of a drug marketing application.
A group of scientists working for the US Food and Drug Administration addressed this topic in an article published in the Journal for Immunotherapy in Cancer, noting that one key hurdle is building a strong association between ctDNA results and long-term clinical outcomes at both patient and trial levels.
Patient-level correlation, they wrote, might involve an association of ctDNA with long-term survival outcomes that provide actionable insights into an individual's prognosis and can be used for patient counseling. Trial-level correlations would involve documenting consistent relationships between the effect size of ctDNA endpoints such as ctDNA clearance and that of survival endpoints across multiple randomized trials.
Last year, the FDA issued draft guidance on the use of ctDNA in drug development. As with the more recent paper, the main concern was the lack of correlations with long-term outcomes.
Building the needed datasets and establishing such connections has gained eager commercial interest.
Last month, for instance, researchers with Foundation Medicine and Natera published a study assessing the ability of distinguishing NSCLC patients with more or less favorable outcomes via ctDNA testing during induction chemo-immunotherapy. The two companies also just launched the jointly developed FoundationOne Tracker ctDNA monitoring assay.
Natera also recently announced a separate clinical trial aimed at evaluating the use of its Signatera minimal residual disease test in monitoring neoadjuvant therapy response among breast cancer patients.