Evidence is increasing that blood-based biomarkers have predictive utility in advanced non–small cell lung cancer. Going further, blood-based next-generation sequencing appears to have clinical utility in selecting targeted treatment in this setting.
Shirish Gadgeel, MD
Evidence is increasing that blood-based biomarkers have predictive utility in advanced non—small cell lung cancer (NSCLC). Going further, blood-based next-generation sequencing (NGS) appears to have clinical utility in selecting targeted treatment in this setting. The performance of blood-based biomarkers in the setting of advanced NSCLC was the subject of several presentations at the 2019 ESMO Congress, 2 of which are summarized here.
In the global phase II/III BFAST study, a cohort of patients with NSCLC who were assigned to alectinib (Alecensa) on the basis of an ALK fusion found on blood-based NGS achieved an investigator-confirmed objective response rate (ORR) of 87.4% and an ORR of 92.0% by independent review facility.1
The median duration of response had not been reached at the median follow-up of 12.6 months. The 6-month rate of event-free survival was 90.4%.
BFAST used blood-based NGS to identify actionable genetic alterations and to assign appropriate targeted therapy in patients with treatment-naïve advanced NSCLC. With a median duration of follow-up of 12.6 months, the investigator-assessed median progression-free survival (PFS) in the ALK-positive cohort had not yet been reached and the 12-month PFS rate was 78.38%, reported Shirish M. Gadgeel, MD, medical oncologist, University of Michigan.
“Tissue-based assays do have some limitations. In about 30% of the patients, the tissue obtained at the time of initial diagnostic biopsy is inadequate to perform these assays. In addition, in about 20% of the patients, a repeat biopsy is not feasible, and even if a repeat biopsy is performed, the tissue obtained at the time of second biopsy is inadequate to perform these tissue-based assays,” said Gadgeel. “Blood-based NGS testing may overcome some of these limitations.”
In BFAST, patients were enrolled into specific treatment cohorts based on identification of genetic alterations using only blood-based NGS. The aim was to establish whether NGS could be used as a stand-alone test (no tissue-based testing). Eligible patients were those with unresectable stage IIIB or IV NSCLC who had not previously received treatment for metastatic disease.
The goal was to demonstrate the consistency of benefit with alectinib in a population selected by blood-based NGS as opposed to tissue-based assay, as was done in the ALEX study2 of patients with untreated ALK-positive NSCLC who were randomized to the alectinib arm. Consistency was defined as achievement of at least 75% of the confirmed ORR with alectinib in ALEX.
A total of 2219 patients were screened to find 87 patients with ALK fusion partners. The 5.4% prevalence of ALK fusions is close to the expected rate of 5%, said Gadgeel. Enrollment was completed within about 12 months of initiating the study. The time from ctDNA extraction to reporting identified alterations to investigators was 10 to 14 calendar days.
Patients with ALK alterations were treated with alectinib at 600 mg twice daily until disease progression, toxicity, withdrawal from the study, or death. The primary endpoint was investigator-confirmed ORR.
Of the 87 patients, 38 (43.7%) had a TP53 mutation. The median blood tumor mutational burden (bTMB) at baseline was 2/Mb (range, 0- 21). Three (3.4%) patients had ≥16 mutations/Mb.
The median patient age in this cohort was 55.0 years, 60% were female, 67% were non-Asian, 94% had an ECOG performance status of 0 or 1, 6% were active smokers and 37% were past smokers, and 94% had stage IV disease. The histology was adenocarcinoma in 93%, and 40% of patients had CNS metastases. The baseline characteristics were similar to those in the ALEX alectinib arm (n = 152).
The 87.4% investigator-assessed ORR in BFAST compares favorably to the 82.9% response rate (71.7% confirmed) observed with alectinib in the ALEX study, noted Gadgeel, and the 12-month PFS rate of 78.38% exceeds that of the 68.4% found in ALEX.
By investigator assessment, the ORR consisted of 0 with a complete response (CR) and 76 (87.4%) with a partial response (PR). Only 1 patient had progressive disease as their best response. As assessed by independent review, 11 (12.6%) patients had a CR, 69 (79.3%) had a PR, and 1 had progressive disease.
“The confirmed ORR was similar in patients with and without baseline CNS metastases, ranging from 84.6% to 91.4%,” he said.
The safety data were consistent with those from other first-line trials of alectinib in patients with ALK-positive NSCLC. Treatment discontinuation was required in 7%, dose reduction in 8%, and dose interruption in 31%.
Other cohorts in BFAST are being assigned to targeted therapies based on identification of RET-positivity (alectinib) and ROS1-positivity (entrectinib [Rozlytrek]), in addition to a high blood tumor mutational burden (atezolizumab [Tecentriq]).
bTMB Is Predictive of Atezolizumab Efficacy
Final analysis of the B-F1RST study reinforces the utility of bTMB as a predictive biomarker for patients with advanced NSCLC who receive first-line atezolizumab monotherapy, reported Mark A. Socinski, MD, from AdventHealth Cancer Institute, Orlando.
Data from the final analysis, with a follow-up period of at least 18 months, confirmed those observed at 6 months. In a population of 152 patients enrolled from regional and community practice sites, the median PFS in those patients with a bTMB ≥16 was 5.0 months compared with 3.5 months in patients with a bTMB <16 mut/Mb (HR, 0.80, P = 35).3
The confirmed ORR in the high bTMB group was 35.7% versus 5.5% in those with bTMB <16 mut/Mb cutoff (P <.0001). The ORR in the ITT population was 17.1% and 12.6% in the 119 patients who formed the biomarker-evaluable population.
Data from the final analysis also showed that the median OS also numerically favored the group with high bTMB versus low bTMB (23.9 vs 13.4 months) although falling short of significance (HR, 0.66; P =.18).
Patients in B-F1RST had stage IIIB-IV locally advanced or metastatic NSCLC of any histology. They were immunotherapy-naïve and unselected for PD-L1 expression. Patients with sensitizing EGFR mutations or ALK rearrangements were excluded. All patients were treated with atezolizumab at 1200 mg every 3 weeks until progressive disease, unacceptable toxicity, or loss of clinical benefit.
Among the 119 patients who formed the biomarker-evaluable population, 91 had a bTMB <16 mut/Mb and 28 had bTMB ≥16 mut/Mb.
Using a bTMB cutoff of 10 mut/Mb, the median PFS in those with a bTMB ≥10 mut/Mb was 2.7 months, and when using a higher cutoff of ≥20 mut/Mb, the median PFS was 6.9 months. At the 20 cutoff, the HR for high versus low bTMB was 0.59 (P =.59).
Median OS by bTMB cutoffs was 14.4 months at the ≥10 mut/Mb cutoff and 23.9 months at the ≥20 mut/Mb cutoff. At the 20 cutoff, the HR for high bTMB versus low bTMB was 0.44 (P = .036).
No new safety signals were observed with atezolizumab.
In reviewing the data on the impact of TMB on outcomes in patients with NSCLC, invited discussant Federico Cappuzzo, MD, from AUSL della Romagna, Ravenna, Italy, noted that in CheckMate-227, the advantage to the combination of nivolumab (Opdivo) and ipilimumab (Yervoy) over chemotherapy on OS was similar in patients with PD-L1 expression ≥50% and a high TMB and the patients with PD-L1 expression <50% and a low TMB.4
“In other words, TMB seems prognostic and is not useful for defining patients who are candidates or not candidates for immunotherapy,” said Cappuzzo.