For the treatment of patients with lung cancer and low PD-L1 expression, immunotherapy-based combinations have been shown effective through clinical trial research. Stephen Liu, MD, reviewed the studies during a Targeted Oncology Case-Based Peer Perspectives virtual event.
Stephen Liu, MD
For the treatment of patients with lung cancer and low PD-L1 expression, immunotherapy-based combinations have been shown effective through clinical trial research. Stephen Liu, MD, an associate professor of Medicine, director of Thoracic Oncology, and director of Developmental Therapeutics, at Georgetown Lombardi Comprehensive Cancer Center, Georgetown University Medical Center reviewed the studies during a Targeted Oncology Case-Based Peer Perspectives virtual event.
Targeted OncologyTM: What do you think of the results of this poll? What would you recommend for this patient?
LIU: Early lead goes right to carboplatin/pemetrexed [Alimta] /pembrolizumab [Keytruda]. We have a little use of carboplatin/nab-paclitaxel [Abraxane]/atezolizumab [Tecentriq], but it’s carboplatin/pemetrexed/ pembrolizumab primarily. That’s overwhelmingly option 1.
Per National Comprehensive Cancer Network [NCCN] guidelines—and this is going to get increasingly complex—for PD-L1 low, we have the options of platinum/pemetrexed/ pembrolizumab; carboplatin/paclitaxel/bevacizumab [Avastin]/ atezolizumab; carboplatin/nab-paclitaxel/atezolizumab; nivolumab/ipilimumab [Opdivo/Yervoy]/chemotherapy, based on CheckMate 9LA [NCT03215706]; nivolumab/ipilimumab; or pembrolizumab alone.1
What data support the use of immunotherapy-based combinations in patients with lung cancer and low PD-L1 expression?
KEYNOTE-189 [NCT02578680], a phase 3 randomized trial, randomized patients 2:1 to chemotherapy [carboplatin or cisplatin plus pemetrexed] and pembrolizumab versus chemotherapy alone.2
We see the overall survival [OS] in the intention-to-treat [ITT] population, the clear winner. The addition of pembrolizumab to chemotherapy improved OS. The hazard ratio [HR] at 30 months was 0.56 [95% CI, 0.45-0.70]; median OS for the pembrolizumab arm was double that of the chemotherapy arm, 10.7 months [95% CI, 8.7-13.6] versus 22 months [95% CI, 19.5-25.2].3
We saw improvement in OS with pembrolizumab versus chemotherapy alone across all PD-L1 strata: in PD-L1–high patients, with an HR of 0.59 [95% CI, 0.39-0.88], but also in PD-L1–low [HR, 0.62; 95% CI, 0.42-0.92] and in PD-L1–negative [HR, 0.52; 95% CI, 0.36-0.74] patients.
The KEYNOTE-042 study [(NCT02220894) explored the use of] pembrolizumab in PD-L1[–positive] patients, so PD-L1 of 1% or greater, with 1:1 randomization to pembrolizumab or histology- specific chemotherapy. There was no crossover in this study. The primary end point was OS by PD-L1 expression level cutoffs of 1% or greater, 20% or greater, and 50% or greater.4,5
We knew that OS was going to be better in patients with PD-L1 [expression of] 50% or greater, and it was. We knew that from KEYNOTE-024 [NCT02142738]. And because this [concerned patients with expression] greater than 1%, half the patients here [had] greater than 50% [expression]. In KEYNOTE-042, OS was positive for the 50% or greater [HR, 0.69; 95% CI, 0.56-0.85], 20% or greater [HR, 0.77; 95% CI, 0.64-0.92], and 1% or greater [HR, 0.81; 95% CI, 0.71-0.93] cohorts. All of those cohorts included PD-L1 expression of 50% or greater, a group in whom we know pembrolizumab is better than chemotherapy. What we’re interested in from this study is OS in the 1% to 49% cohort.
The OS for that cohort was an exploratory analysis, not a primary end point, so it didn’t make its way into the label. In the PD-L1–low cohort [1%-49%], we saw the HR close to 1 [HR, 0.92; 95% CI, 0.77-1.11] and the survival curves crossed, which makes the HR a little difficult to interpret, but there was no [statistically significant] difference in the median OS. Pembrolizumab was not convincingly better than chemotherapy in the PD-L1–low cohort; so, not necessarily worse but not convincingly better.
The IMpower150 study [(NCT02366143) looked at the regimen of] carboplatin plus paclitaxel plus bevacizumab plus atezolizumab.6 We saw benefit when atezolizumab was added to carboplatin/ paclitaxel/bevacizumab: an improvement in OS, with an HR of 0.78 [95% CI, 0.64-0.96].6,7
How was the nivolumab/ipilimumab combination investigated in this setting?
CheckMate 227 [NCT02477826] had an extremely complex design. OS in selected PD-L1 cohorts was a coprimary end point. Note the [specific] doses for the PD-L1–positive cohort: nivolumab 3 m/kg every 2 weeks and low-dose ipilimumab, 1 mg/kg every 6 weeks.8,9
Looking at OS in the PD-L1–positive cohort, it was better than in the chemotherapy-alone arm [HR, 0.79; 95% CI, 0.67-0.93].10
We saw a difference in median OS of 14.9 [months for the chemotherapy-alone arm] versus 17.1 months for the PD-L1 1% or greater cohort, with an HR of 0.79 as noted previously; so nivolumab/ipilimumab was the winner in the PD-L1– positive group.
If we look at the forest plot, in an exploratory analysis of the PD-L1–low cohort, we saw an HR pretty close to 1 [HR, 0.94; 95% CI, 0.75-1.18]. But what’s interesting is, if you look at the PD-L1–negative group, the HR was the best, at 0.62 [95% CI, 0.49-0.79]. We saw a clear signal that nivolumab/ipilimumab in the PD-L1 50% or greater group was better, and in the PD-L1– low group with 1% to 49% expression maybe not performing quite as well. The study was not powered for that analysis, but it’s interesting.9
Nivolumab/ipilimumab is chemotherapy free, but it is not toxicity free. The rate of grade 3/4 adverse events [(AEs) 32.8%] was similar to that of chemotherapy [36.0%], though of a different flavor. Discontinuation due to AEs was about 1 in 5 for nivolumab/ipilimumab.
CheckMate 9LA garnered FDA approval before we even saw the data.11 CheckMate 9LA looked at nivolumab/ipilimumab, with 2 cycles of histology-specific chemotherapy versus chemotherapy alone. The updated OS was pretty intriguing here—a median of 10.9 months with chemotherapy versus 15.6 months with nivolumab/ipilimumab, with an HR of 0.66 [95% CI, 0.55-0.80].12 CheckMate 9LA was FDA approved based on this OS benefit. The overall response rate [ORR] was also better [in this trial]: 38% for nivolumab/ipilimumab versus 25% for chemotherapy.
The OS benefit of CheckMate 9LA was seen across PD-L1 strata, including PD-L1 low [HR, 0.61; 95% CI, 0.44-0.84].
There was more toxicity, as we’d expect. These are the highest rates of AEs of any regimen, but we don’t see that cumulative myelosuppression that we see with chemotherapy because we’re just doing the 2 cycles of chemotherapy. Still, about 1 in 5 people [18.1%] stopped treatment because of toxicity.
The combination of nivolumab/ipilimumab was approved on May 26, 2020, by the FDA.11 It was presented 5 days later [at the 2020 American Society of Clinical Oncology Virtual Scientific Program]. Atezolizumab was approved based on [findings from] IMpower110 [NCT02409342] for PD-L1 high.13 While the investigators used SP142 and SP263 [assays for detecting PD-L1 expression], they also did validate it with Dako 22C3 [PD-L1 IHC 22C3 pharmDx assay], and the label doesn’t mandate which PD-L1 antibody you use with atezolizumab. It’s approved for PD-L1 high, not PD-L1 low.
In which situation would you use the CheckMate 9LA regimen?
You could, and people will, use it for a renal transplant with the understanding that you are probably going to lose that transplant. And so it may mean going on dialysis. In a liver transplant or heart transplant, I absolutely would not [use it], where rejection is just not compatible with life. But if the patient understands that he will more than likely...lose that transplanted organ, it can still be a reasonable thing to do.
After the patient completes carboplatin, how do you approach maintenance therapy with this regimen?
We have to tailor this. Not everyone can tolerate pemetrexed, and over time you see some cumulative toxicity. It probably depends on the PD-L1 [expression level]. If you have a PD-L1 of 100%, maybe you’re a little more likely to let go of that pemetrexed as the maintenance part but [on an] individualized [basis].
How long do you continue maintenance treatment for patients?
Two years. So we’re going with the 2 years of pembrolizumab. I think that’s pretty reasonable. With pemetrexed, generally people would probably need to stop around then anyway.
Do you tend to treat beyond progression?
I will say that in my own practice, if they are symptomatic, we’re changing treatment. If they’re asymptomatic and it’s mild progression, maybe we’ll push on.
If this patient has progressed on the KEYNOTE-189 regimen, what second-line regimen would you offer?
A weekly docetaxel dose of 30 mg/m2 is a reasonable approach and there are some weak data for it; I think it is definitely better tolerated. Docetaxel at 75 mg/m2 every 3 weeks is tough, especially if people have had a lot of chemotherapy or with people who are a bit older. Weekly docetaxel is used by a lot of people, and I think it is a reasonable approach because we certainly don’t want to make people too sick.
I also use docetaxel 60 mg/m2 a lot, especially if they’re older patients and we’re worried about [toxicity with a higher dose].
What data support the use of this regimen in the second-line setting?
The REVEL trial [NCT01168973] was published 6 years ago, and was a second-line study of docetaxel/placebo versus docetaxel/ ramucirumab. In that study, most patients got docetaxel 75 mg/ m2 with ramucirumab 10 mg/kg, although there were geographic variants in dosing [in the investigational arm].14
The primary end point was OS, and this was a positive study. The OS improved from a median [rate of] 9.1 months for docetaxel versus 10.5 months for docetaxel/ramucirumab [HR, 0.86; 95% CI, 0.75-0.98]. The docetaxel/ramucirumab arm had a better ORR, 23% versus 14% [odds ratio, 1.89; 95% CI, 1.41-2.54].
Patients with aggressive disease and people who progressed quickly were part of an exploratory analysis, but the end result was that, even if patients progressed quickly at the first scan, they still got a benefit from the addition of ramucirumab. We were able to salvage some of those patients with the REVEL regimen. For patients with rapid progression in the ITT population, if first-line treatment was less than a month, they still had benefit, with an HR of 0.4 for OS [95% CI, 0.22-0.73].15
In terms of safety, as we’d expect, we saw more toxicity with the addition of ramucirumab to docetaxel.14,15 Most of the high toxicity was related to hypertension [11% with added ramucirumab versus 5% without].
Do you think there are any patients who benefit more or less from the addition of ramucirumab?
There are some who would argue that ramucirumab is a bit more effective in someone who’s had prior I/O [immuno]therapy. There were some exploratory data out of Martin Reck, MD, PhD, in Germany. But his sense was that the outcomes for people who had received REVEL post I/O were better, something [possibly related to receiving] VEGF in sequence with I/O. I’m pushing people to add it a little more these days, though we don’t have strong data yet. I think what we’d say objectively is that when we add ramucirumab, as the trials showed us, we increase ORR, we increase OS. For some people, it is a modest improvement, but it is an approved regimen. I would say that my experience echoes that, and I see benefit [in my patients] with it. The dose modifications depend on the patient substrate and how well we think they can tolerate those treatments.
I think it’s important that we not necessarily assume our patient would be bothered by [the therapy] because some really wouldn’t. There are lots of things to consider and another reason why lung cancer [doesn’t always correspond with treatment] flow charts.