This article has been updated to correctly represent the interim survival analysis for patients who cleared ctDNA at any point after treatment.
NEW YORK – Researchers have demonstrated in a Phase II study that customized circulating tumor DNA (ctDNA) assays designed by genetic testing firm Natera could potentially be useful for monitoring whether patients with a variety of cancers will respond to immunotherapy.
Although some cancer patients derive dramatic and sustained benefit from immunotherapy, most cancer patients do not. The INSPIRE study, published today in Nature Cancer, in which patients had ctDNA quantified using Natera's personalized Signatera assay, suggests that ctDNA monitoring may be useful for identifying best responders to immunotherapy alongside existing biomarkers, such as tumor mutational burden (TMB), PD-L1 expression, and microsatellite instability.
The study, led by Scott Bratman of Princess Margaret Cancer Center in Toronto, enrolled 106 patients with advanced cancer and stratified them into five cohorts based on tumor type. These cohorts included patients with squamous cell head and neck cancer, triple-negative breast cancer, high-grade serous ovarian cancer, and malignant melanoma; the fifth cohort included patients with other solid tumor types.
Tumor specimens were prospectively collected from patients at baseline, and 94 patients' samples were sufficient for whole-exome sequencing required to design customized assays that could detect ctDNA in accordance with the unique genetic signature of each patient's tumor. Investigators selected up to 16 clonal somatic mutations to include in the customized ctDNA assay for each patient. Although Natera designed the customized assays, the company was blinded to the clinical data.
Prior to patients starting treatment with pembrolizumab (Merck's Keytruda), the researchers used the customized tests to assess baseline ctDNA, which they detected in 92 out of 94 patients. Then, samples were longitudinally collected in 73 patients, starting at week six or seven of treatment for each patient and ongoing at every three cycles.
Bratman and colleagues found that change in ctDNA levels early in the course of immunotherapy, after six weeks of treatment, could predict whether patients would benefit from immunotherapy across the five cohorts. Thirty-three patients had lower ctDNA levels at week six compared to baseline, and 14 patients, or 42 percent, saw their tumors shrink on pembrolizumab. In comparison, only one of the 40 patients whose ctDNA levels increased from baseline at week six responded to immunotherapy.
Change in ctDNA levels at six weeks was also associated with a higher clinical benefit rate and favorable overall and progression-free survival in the five cohorts, the authors reported.
One of the current challenges for oncologists is deciding whether to keep giving immunotherapy after patients progress very early in the course of treatment. In this study, researchers wanted to see if ctDNA assessments could help with this decision. After six weeks of pembrolizumab treatment, 37 out of 73 patients experienced disease progression according to RECIST criteria. Testing detected ctDNA increases in 30 patients who had the shortest survival in the study.
However, 11 of these 30 patients continued to receive additional cycles of pembrolizumab despite having early progression. "In retrospect, such continued immune checkpoint blockade treatment may have been unnecessary in this poor prognosis subset," wrote Bratman and colleagues, noting that adding ctDNA assessments to RECIST measurements improved the accuracy of overall survival predictions.
On the other hand, among 12 patients for whom ctDNA clearance was achieved at any point in the course of treatment — that is, ctDNA became undetectable on at least one of the tests — prolonged objective responses to treatment were observed. These dozen patients were all alive beyond their first ctDNA clearance at follow up, which ranged from 10.8 months to 29.5 months in this group.
Additionally, among patients who responded treatment, the authors noted that the ctDNA clearance was detected prior to a visible radiographic response. For example, one patient experienced ctDNA clearance after four months of pembrolizumab treatment, but the clinical responses were not observed until eight months.
Bratman and colleagues also considered the role of ctDNA evaluations alongside predictive immunotherapy biomarkers in clinical use, such as TMB and PD-L1 expression. The use of these biomarkers has been complicated since the cutoffs delineating response and non-response to immunotherapy appear to differ based on tumor type.
In this study, however, lower ctDNA levels after two pembrolizumab cycles and on-treatment ctDNA clearance appeared to identify subsets of patients with good prognosis, independent of tumor type, TMB, or PD-L1 status. "Of note, patients with relatively low TMB were amenable to the highly targeted, personalized ctDNA assay," the authors wrote in the paper. "Thus, our findings demonstrate that serial ctDNA analysis using a bespoke assay could serve as a generalizable monitoring strategy for patients treated with immune checkpoint blockade."
These results, according to the study authors, demonstrated the potential for broad clinical use of ctDNA-based surveillance during immune checkpoint inhibitor treatment.
"This is, to our knowledge, the largest prospective study that has evaluated the value of serial ctDNA analysis in patients prior to and during immune checkpoint blockade therapy," Solomon Moshkevich, general manager of Natera's oncology business, said in a statement. "The study demonstrates the potential for broad clinical utility of Signatera in patients with advanced solid tumors treated with immune checkpoint blockade."