Back Back
Pembrolizumab Plus CT for NSCLC
Combination Immunotherapy Plus Chemotherapy as First-line Therapy of Advanced NSCLC: Where Are We Now?

Released: April 11, 2017

Expiration: April 10, 2018

Activity Information

Activity

Progress
1
Course Completed
Continue

Pembrolizumab monotherapy is the current standard of care for patients with advanced NSCLC and high PD-L1 tumor expression defined as ≥ 50% PD-L1 tumor cell staining per the IHC 22C3 pharmDx companion diagnostic assay. The FDA approval of pembrolizumab monotherapy in this patient population was based on the strongly positive results of the phase III KEYNOTE-024 trial, which demonstrated a significant improvement in both PFS (HR: 0.50; P < .001) and OS (HR: 0.60; P = .005) for pembrolizumab vs chemotherapy.

But What About the Combination of Immunotherapy and Chemotherapy for Newly Diagnosed Patients With Advanced NSCLC Regardless of PD-L1 Expression?
At this time (April 2017), a supplemental Biologics License Application for pembrolizumab in combination with first-line carboplatin/pemetrexed chemotherapy for patients with advanced, nonsquamous NSCLC regardless of PD-L1 expression is under review by the FDA. At the foundation of this application is the provocative data from cohort G of the phase II KEYNOTE-021 trial, which showed a significantly improved PFS in patients with newly diagnosed advanced, nonsquamous NSCLC with the addition of pembrolizumab to carboplatin/pemetrexed vs carboplatin/pemetrexed alone (13.0 vs 8.9 months, respectively; HR: 0.53; P = .010), as well as a significantly improved ORR (55% vs 29%, respectively; P = .0016) that appeared to be independent of PD-L1 expression level. However, there was no difference in OS over a median follow-up of 10.6 months, with the caveat that these data are not yet mature. Although the findings of KEYNOTE-021G are certainly exciting, with the prospect of being able to administer immunotherapy as early as possible to a large majority of our patients rather than just a selected population with high PD-L1 expression, I think caution should be taken in the immediate application of these data to clinical practice, especially for patients with low PD-L1 tumor expression.

Contribution of Patients With ≥ 50% PD-L1 Tumor Expression to Observed Benefit With Combination Therapy in KEYNOTE-021G
The patient population of KEYNOTE-021G comprised just 123 patients who were randomized to chemotherapy alone (n = 63) or chemotherapy with pembrolizumab (n = 60), with a fairly even distribution of patients with < 1%, 1% to 49%, or ≥ 50% PD-L1 tumor expression among arms (27% to 37% in each arm). Therefore, I think it is important to consider the distinct possibility that patients with ≥ 50% tumor PD-L1 expression may be driving a large amount of the overall benefit observed in the KEYNOTE-021G trial. If the approximately 30% of patients with high tumor PD-L1 expression in each arm were removed, would the results with the combination of pembrolizumab and chemotherapy compared with chemotherapy alone have been so impressive? Patients with ≥ 50% PD-L1 tumor expression had an 80% response rate with the combination (vs 57% and 26% for patients with < 1% and 1% to 49% PD-L1 tumor expression, respectively), so without them, it is not clear that the results for those patients with < 50% PD-L1 expression would be as meaningful vs chemotherapy alone.

In the context of what is currently the standard of care for advanced NSCLC, the more appropriate comparison for patients with ≥ 50% PD-L1 tumor expression would have been pembrolizumab plus chemotherapy vs pembrolizumab monotherapy and, for those with < 50% PD-L1 tumor expression, pembrolizumab plus chemotherapy vs first-line chemotherapy followed by 100% crossover to an immune checkpoint inhibitor in the second line. In KEYNOTE-021G, approximately 70% of patients receiving chemotherapy alone did eventually cross over to immune checkpoint inhibitor therapy, but that falls short of the total number who could potentially have benefitted from an FDA-approved immune checkpoint inhibitor in the second line, including pembrolizumab, which is approved for patients with ≥ 1% PD-L1 tumor expression per the 22C3 assay, and atezolizumab and nivolumab, which are approved regardless of PD-L1 expression level. For patients with < 50% PD-L1 tumor expression, I think the important questions to consider are: Does concurrent chemotherapy and immunotherapy in the frontline improve survival over sequential therapy? Is there a clear incentive to give all of our potentially effective therapies in the same initial 2- or 3-month period?

Increased Toxicity With Combination Immunotherapy and Chemotherapy
Not surprisingly, additive toxicities were observed for combination therapy in the KEYNOTE-021G trial; patients who received pembrolizumab with chemotherapy experienced the toxicities of both treatments. Even if these combined toxicities were not prohibitive, the lack of evidence for a synergistic benefit with the combination brings into question having patients experience increased toxicity if they are only benefiting from one of the modalities and not the other. Furthermore, because the toxicity profiles of immunotherapy and chemotherapy overlap, teasing out which treatment is the greater culprit—and, therefore, should be dose adjusted or eliminated—can be a real challenge. In the face of this ambiguity, we risk additive toxicities leading to the early cessation of treatments that are actually helping and are not even responsible for the toxicity in question. Moreover, for the 28% to 30% of patients with PD-L1 tumor expression ≥ 50%, combining chemotherapy with immunotherapy could represent potential overtreatment, with this patient population potentially experiencing the combined toxicity without even needing the chemotherapy.

It is worth remembering that a big appeal of immunotherapy in oncology is its potential to confer meaningful anticancer benefits while sparing patients the toxicities of conventional chemotherapy. Furthermore, the idea of precision medicine is to identify and administer the optimal and least toxic treatment with as little additional therapy and toxicity as possible. Thus, as clinicians, I think we need to be very thoughtful about what is helping and what is hurting our patients and to be careful not to overtreat. In treating our patients with advanced NSCLC, I think our goal should be to give patients the minimum of what they need to do as well as possible.

Increased Expense With Combination Maintenance Therapy
Following first-line treatment in the KEYNOTE-021G trial, nonprogressing patients in the pembrolizumab plus carboplatin/pemetrexed arm continued to maintenance pemetrexed and pembrolizumab. Both pemetrexed and pembrolizumab are potentially effective maintenance therapies. However, continuing on this therapy as a 2-drug combination is very expensive, especially considering that the vast majority of the patients in this trial are likely benefiting primarily, if not completely, from just one of these 2 drugs, with the other drug only adding cost and cumulative toxicity. A key benefit of a sequential approach is that it both saves on cost and allows us to isolate which treatments are conferring a benefit and which are causing toxicities without being confounded by giving everything at once.

Which Patients Would I Consider for Combination Therapy?
If the combination of pembrolizumab and carboplatin/pemetrexed is approved, there is a subset of patients in whom I think its use would be appropriate to consider. Potential candidates for immediate access to combination therapy would be patients with rapidly progressing disease and a high tumor burden that look like they may decline more quickly than the 2-3 months of ambulatory outpatient treatment needed to pursue sequential first-line chemotherapy and second-line immunotherapy. However, this leaves only a docetaxel-based regimen as the sole evidence-based treatment option with an established survival benefit left to pursue in this patient population in the event of disease progression.

Looking Ahead
Although provocative, in my opinion, the data from the KEYNOTE-021G study are not yet definitive enough to guide treatment recommendations, especially considering the avalanche of data from ongoing phase III clinical trials evaluating the combination of various immune checkpoint inhibitors (atezolizumab, nivolumab, and pembrolizumab) and chemotherapy that will be reported in the imminent future. For example, the phase III KEYNOTE-189 trial of platinum plus pemetrexed chemotherapy with and without pembrolizumab in nearly 600 patients with nonsquamous NSCLC regardless of PD-L1 expression has a scheduled primary completion date of September 2017. Although it is clear that patients with ≥ 50% PD-L1 tumor expression are likely to do quite well with pembrolizumab alone, and thus can be spared the toxicity and intensiveness of concurrent immunotherapy and chemotherapy, for patients who have < 50% PD-L1 tumor expression, we need to learn more about how much or how little they are benefiting from the addition of immunotherapy to chemotherapy in the frontline. Considering that KEYNOTE-021G had only 123 patients and has not yet shown a survival benefit, I think it is prudent to wait the 6-12 months for larger and more mature data sets to determine whether we should broaden the use of first-line immunotherapy to patients with low or no PD-L1 tumor expression.

What are your thoughts on combining chemotherapy with immune checkpoint inhibitors as first-line therapy for patients with metastatic NSCLC? Please leave a comment below.

Continue

Poll

1.
Approximately how many of your patients with metastatic NSCLC are you currently testing at diagnosis for PD-L1 expression using the 22C3 companion assay?
Submit