“We can predict how a patient responds if PD-L1 levels are close to 100% or close to zero. But we cannot say [with certainty] that a patient with PD-L1 of 30% is going to have significantly different outcomes than a patient with PD-L1 of 40%.”
The search for a better biomarker that can predict response to chemoimmunotherapy agents in non–small cell lung cancer (NSCLC) has led investigators to consider plasma-based tumor mutation burden (pTMB) in conjunction with PD-L1 tumor proportion score (TPS) in this setting. Charu Aggarwal, MD, MPH,and colleagues reported findings of a 66-patient study evaluating baseline pTMB associated with responses in patients treated with pembrolizumab (Keytruda)-based therapy.1 To assess pTMB, patients had undergone a 500-gene panel next- generation sequencing. Patients received pembrolizumab alone or in combinatio with chemotherapy for first-line treatment of metastatic NSCLC.
They found that baseline pTMB was significantly associated with progression-free survival (PFS; HR, 0.93; 95% CI, 0.90-0.97; P= .001). Further, to determine pTMB as a binary variable, they identified 16 mutations per megabase (mut/Mb) as the appropriate limit to assess pTMB.
The investigators reported a median PFS of 14.1 versus 4.7 months for pTMB ≥16 mut/Mb versus <16mut/Mb (HR, 0.30; 95% CI, 0.16-0.60; P<.001; FIGURE).1 Median overall survival (OS) for patients with pTMB ≥16 was not reached versus 8.8 months for patients with pTMB<16 mut/Mb (HR, 0.48; 95% CI, 0.22-1.03; P= .061).
Patients were enrolled in the trial March 1, 2017, to October 11, 2018, if they had pathologically confirmed metastatic NSCLC, and received pembrolizumab- based therapy standard of care as firstline treatment (NCT03047616).
Of the 66 patients, 52 (78.8%) were found to be evaluable for pTMB. Among the52 patients, 28 had a pTMB ≥16 mut/Mb.
Thirty-one patients received pembrolizumab monotherapy for a median of 4.1 months (range, 0-29.4 months) and 35 patients receivedpembrolizumab with platinum plus pemetrexed-based chemotherapy for a median of 7.1 months (range, 2.0-21.7 months). The investigators reported that 14 patients could not be evaluated based on low tumor shedding and no samples were found to be MSI-high (microsatellite instability-high).
The median pTMB for the 45 patients who had a complete response (CR) or partial response (PR) at week 9 was 21.5 mut/Mb (range, 7.7-52.5) versus 13.9 mut/Mb (range, 1.9-31.6) among patients with stable disease (SD) or progressive disease (PD; P = .037). A logic regression model showed that pTMB was associated with an odds ratio of 1.09 (95% CI, 1.02-1.08; P = .18) per 1 unit increase in pTMB.
The median pTMB for the 45 patients who had a complete response (CR) or partial response (PR) at week 9 was 21.5 mut/Mb (range, 7.7-52.5) versus 13.9 mut/Mb (range, 1.9-31.6) among patients with stable disease (SD) or progressive disease (PD; P= .037). A logic regression model showed that pTMB was associated with an odds ratio of 1.09 (95% CI, 1.02-1.08; P= .18) per 1 unit increase in pTMB.
“My colleagues and I used a relatively non- invasive method to assess plasma TMB as abiomarker for frontline pembrolizumab-basedtherapy that included chemotherapy,” said Aggarwal, the Leslye M. Heisler Assistant Professor for Lung Cancer Excellence, Division of Hematology/Oncology at the University of Pennsylvania, Abramson Cancer Center, in Philadelphia, Pennsylvania. “This illustrates the unique nature of the study because we have widely seen TMB as a response to immunotherapy, but we have not seen this as a response tochemoimmunotherapy, and additionally have not seen it with the use of plasma,” Aggarwal told Targeted Therapies in Oncology in an interview.
Aggarwal et al reported that there was no significant difference in medianOSas afunctionof pTMB level overall. The median was not reached for pTMB-high versus 8.8 months for pTMB-low tumors (HR, 0.48; 95% CI, 0.22-1.03; P = .061), or among patients receiving pembrolizumab alone (median not reached vs 6.1months; HR, 0.49; 95% CI, 0.17-1.42; P=.187) or pembrolizumabpluschemotherapy (median not reached vs 19.2 months; HR, 0.50; 95% CI, 0.16-1.54; P= .228).
One of the challenges with using PD-L1 as a biomarker is its imprecise and subjective nature, said Aggarwal. PD-L1 levels are based on staining techniques in which the pathologistdeterminesthelevelofstainingassociated with the sample on a scale ranging from below zero to 100. Findings at the extreme ends of the scale, ie, at zero or 100 are easier to interpret than findings across the middle range. Baseline characteristics of the patients are listed in the TABLE.1
“We can predict how a patient responds if PD-L1 levels are close to 100% or close to zero. But we cannot say [with certainty] that a patient with PD-L1 of 30% is going to have significantly different outcomes than a patient with PD-L1 of 40%,” Aggarwal said.
“Mutations in ERBB2 exon 20, STK11, KEAP1, or PTEN were more common in patients with no durable clinical benefit,” said Aggarwal. “We found that patients who had no durable benefit were more likely to have these negative predictive mutations, which was very interesting to us. When this finding was combined into a score with TMB, we found that we could clearly separate patients even for OS,” she said.
The application of these findings is somewhat vague. “I think the general consensus in community oncology is that TMB is a [difficult] biomarker,” Aggarwal said, “especially in light of the CheckMate 227 data.”
CheckMate 227 was the first large, phase III study that demonstrated that PFS was prolonged in patients with high tumor mutational burden(TMB). Patients with nonmetastatic NSCLC, withnoEGFRmutationsorknown ALKalterations, were randomized based on their PD-L1 expression. Patients with PD-L1– positive disease received either thenivolumab and ipilimumab combination,single-agent nivolumab, or chemotherapy.
Patients with PD-L1–negative disease were randomized to receive either nivolumab and ipilimumab, chemotherapy or the combination of nivolumab and chemotherapy.
In the final analysis of part 1 of the trial, investigators reported that the median overall survival in patients with tumor PD-L1 expression ≥1% was 17.1 months for patients who received the immuno-oncology (IO) combination compared with 14.9 months for patients receiving chemotherapy (HR, 0.79; 97.72% CI,0.650.96; P = .007). The nivolumab and ipilimumab combination showed a 2-year overall survival of 40% compared with 33% for patients in the chemotherapy arm.2
Aggarwal is not ready to bury TMB, however. “I would say that TMB is probably not dead, but we need to find a way to better measure it,” she said.
Next steps on the horizon are designing clinical trials that focus on the ability to finetune the delivery of chemotherapy or chemoimmunotherapy, said Aggarwal.
“What we need more of is the ability to define negative and positive prognostic factors forimmunotherapy that can help us avoid giving immunotherapy to a patient who will never benefit from it, or save patients from receiving chemotherapy who don’t need it and who may do just as well from immunotherapy,” Aggarwal concluded.
1. Aggarwal C, Thompson JC, Chien AL, et al. Baseline plasma tumor mutational burden predicts response to pembrolizumab-basedtherapy in patients with metastatic non-small cell lung cancer [Published online February 26, 2020]. Clin Cancer Res. doi: 10.1158/1078-0432.CCR-19-3663.
2. Hellman MD, Paz-Ares L, Caro RB, et al. Nivolumab plus ipilimumab in advanced non–small-cell lung cancer. N Engl J Med. 2019;381:2020-2031. doi: 10.1056/NEJMoa1910231.