乌鸦传媒

BARCELONA -- Skeptics of tumor mutational burden (TMB) as a marker of immunotherapy antitumor activity in non-small cell lung cancer (NSCLC) won the day at the World Conference on Lung Cancer (WCLC) as three studies showed no association with key outcomes.

Two studies of pembrolizumab (Keytruda) with or without chemotherapy showed no association between TMB and treatment efficacy. Similarly, a randomized trial of nivolumab (Opdivo) and ipilimumab (Yervoy) combination therapy showed no association with overall survival (OS).

Lung cancer specialists agreed that the collective results indicated that TMB assessment in NSCLC is "not ready for prime time."

Commenting on the studies and others reported at WCLC, Alex Adjei, MD, PhD, of the Mayo Clinic in Rochester, Minnesota, pointed out that different studies have used different cutoff values for TMB, as well as different methods of determining TMB -- next-generation sequencing versus whole-exome sequencing.

"I think therein lies the problem," said Adjei. "We're not even talking about the same thing. Normally when we're talking about a biomarker, such as EGFR mutation, we are talking about the same thing. If you tell me high TMB, I don't know what it is because everyone's definition is different."

The optimal material for determining TMB -- blood or tissue -- also has yet to be determined, said Corey Langer, MD, of the University of Pennsylvania in Philadelphia. Moreover, TMB remains a surrogate for the actual target of the treatment, the PD-1 signaling pathway.

"I think the approach to TMB can be divided into two groups: the TMB 'cultists' and the TMB skeptics," said Langer. "I would put myself in the latter group, although I still think it has a potential role. We just haven't figured it out. It certainly shouldn't be used -- at least, not yet -- in therapeutic decision-making."

"Long story short: not ready for prime time," he added.

Chemo With or Without Pembrolizumab

Investigators in multiple clinical studies have evaluated TMB as a biomarker for immunotherapy in NSCLC. The phase I/II KEYNOTE-021 study was the first to demonstrate the safety and efficacy of a PD-1 inhibitor in combination with chemotherapy in locally advanced or metastatic NSCLC.

Langer and colleagues performed an exploratory analysis of the data to determine whether TMB had value for predicting outcomes with chemotherapy alone or in combination with pembrolizumab.

The post hoc study included a total of 70 patients with tumor tissue available for TMB analysis. Investigators performed one analysis to determine whether TMB as a continuous variable was associated with outcomes for patients treated with chemotherapy alone or in combination with pembrolizumab. A second analysis assessed whether response to the combination differed between patients with TMB-high versus TMB-low tumors.

The investigators found that response rates in the TMB-evaluable subgroup -- 68% for patients treated with the combination and 38% for those who received chemotherapy alone -- were similar to rates in the total patient population (n=145) at 63% versus 31%.

For the TMB-evaluable subgroup, the analysis showed no association between TMB and the combination's efficacy, including objective response (P=0.180), progression-free survival (PFS, P=0.187), or OS (P=0.o81). The same pattern emerged from the analysis of 26 patients treated with chemotherapy alone (P=0.861 to P=0.763). The data also revealed no association between TMB and PD-L1 expression.

Langer described the KEYNOTE-021 data as the "appetizer" for a similar analysis of KEYNOTE-189, which involved 616 patients, 293 of whom had tissue available for TMB assessment by whole-exome sequencing and matched normal DNA. The trial compared pembrolizumab plus pemetrexed-platinum chemotherapy versus the same chemotherapy alone in patients with metastatic NSCLC.

OS, PFS, and response rate were similar in the TMB-evaluable and total patient population, said Marina C. Garassino, MD, of the Fondazione IRCCS Istituto dei Tumori in Milan. The 293 TMB-evaluable patient group comprised 207 treated with the combination and 86 treated with chemotherapy only. For the combination subgroup, TMB did not have a significant association with OS (P=0.174), PFS (P=0.075), or response (P=0.072). Respective P-values for the chemotherapy subgroup were 0.856, 0.055, and 0.434.

Investigators also found no association between TMB and PD-L1 expression in the combination arm (P=0.27) or the chemotherapy subgroup (P=0.92).

Finally, Garassino and colleagues compared outcomes according to whether patients had TMB-high (n=134) or -low (n=159) tumors, using a cutoff of 175 mutations/exome. They found that OS, PFS, and response with the combination or chemotherapy alone were similar in patients with TMB-high and TMB-low tumors.

"Our data suggest that tissue TMB may not help select patients who would have better outcomes with pembrolizumab plus pemetrexed and a platinum, given as first-line therapy for metastatic nonsquamous NSCLC," she concluded.

Nivolumab Alone or With Chemo

Data for the third negative result for TMB analysis came from the phase III S1400I trial of nivolumab plus ipilimumab versus nivolumab alone for previously treated stage IV squamous-cell lung cancer. The of the trial, reported earlier at the 2019 American Society of Clinical Oncology meeting in Chicago, showed no benefit for the combination over nivolumab monotherapy.

An exploratory analysis suggested an OS benefit for the combination in patients with TMB >10 mut/mb and PD-L1 expression ≤5%, said Lyudmila Bazhenova, MD, of the University of California San Diego Moores Cancer Center.

The new analysis evaluated whether TMB predicted improved OS with the combination across the range of PD-L1 expression values from 0% to ≥50%. The analysis included 231 patients evaluable for TMB, 161 evaluable for PD-L1 expression, and 149 with data for both biomarkers. Investigators chose a cutoff of TMB ≥10 mut/mb for the analysis.

Neither of the individual markers had a significant association with improved OS with the combination. Analysis of TMB across the range of PD-L1 expression levels failed to identify any subgroup that had better OS. The data did yield a trend toward interaction between TMB and PD-L1 (P=0.06) favoring the combination, and an analysis of TMB <10 mut/mb favored nivolumab monotherapy.

The investigators used gene array data to evaluate the association between the top 20 altered genes in squamous-cell lung cancer and OS. None of the abnormalities predicted outcomes with the combination versus nivolumab alone.