Immune checkpoint inhibitors (ICIs) have revolutionized the treatment of patients with advanced non–small-cell lung cancer (NSCLC) by demonstrating significantly better outcomes and tolerability than conventional chemotherapy.1 Although immunotherapy is now recognized as the standard of care for NSCLC without actionable drivers, selecting the right immunotherapy remains challenging, in part because head-to-head trials are lacking, meaning reliable comparisons of safety and efficacy are unavailable.2 Although guidelines recommend specific immunotherapy sequences, a recent meta-analysis of more than 10,000 patients with NSCLC concluded that high variability between sequencing strategies and clinical scenarios precludes identification of one most-effective sequence.3 It is therefore necessary to develop a comprehensive understanding of therapeutic options and decision-making strategies to achieve optimal outcomes across a diverse population of patients. This article discusses immunotherapy selection in NSCLC without actionable mutations, including the use of biomarkers, patient factors, and disease factors. Expert commentary is provided by Katie Kerrigan, DO, and Wallace Akerley, MD, from Huntsman Cancer Institute at University of Utah, Salt Lake City.
According to the American Society of Clinical Oncology (ASCO) and Ontario Health (OH), patients with lung cancer should undergo biomarker testing for actionable molecular biomarkers and programmed death-ligand 1 (PD-L1) expression prior to starting first-line therapy.4,5 The actionable biomarker panel should include EGFR, ALK, ROS1, and BRAF; RET, HER2, KRAS, and MET should also be tested as part of a larger testing panel initially or when routine testing of EGFR, ALK, ROS1, and BRAF is negative.5 “If you’re thinking about carcinogenesis, the panel we’re doing right now is the molecular panel,” explained Dr Akerley, “that looks for what the broken genes are that are driving the cancer. The second panel…is the PD-L1 test that we do. It gives you an answer as to how the cancer got around the immune system.”
Patients with actionable molecular biomarkers should receive targeted therapies, whereas patients without actionable drivers should receive immunotherapy, either alone or in combination with other agents.1,4 “Oftentimes, we do see PD-L1 positivity in patients who have oncogenic driver mutations, but those patients should always get first-line targeted therapy with a tyrosine kinase inhibitor [TKI] because retrospective data has shown that patients with, for example, EGFR/ALK mutations, are not particularly sensitive to immunotherapy and they typically have very short progression-free survival and low response rates to single-agent pembrolizumab or nivolumab or something similar,” Dr Kerrigan said. “TKIs should always trump immunotherapy in the first line when known oncogenic drivers with FDA-approved therapies are present on molecular testing.”
As shown in Figure 1, most patients with NSCLC lack actionable alterations, and approximately 20% to 30% of patients have actionable biomarkers with corresponding therapies that are not yet FDA approved.6
In those cases where a patient with NSCLC does not have actionable alterations, the most important biomarker in immunotherapy treatment planning is the PD-L1 tumor proportion score (TPS), as this has been, to date, the most reliable predictor of response to ICIs.7 Although many immunotherapy studies and recommendations refer to PD-L1 cutoff levels (eg, PD-L1 TPS ≥50%), such cutoffs should not be interpreted strictly due to variable and dynamic expression of PD-L1 in lung tumors, which interferes with the correlation between expression and response.8 Intratumoral heterogeneity and heterogeneity between the primary tumor and metastases may also contribute to difficulty determining accurate cutoffs of biomarker expression, where there may be distinct cellular populations within tumors and metastatic lesions that can acquire new mutations.9
With this in mind, leading guidelines recommend that patients with PD-L1 TPS ≥50% should receive ICI monotherapy, whereas patients with PD-L1 TPS ≤0% should receive ICI-based combination therapies.4 Looking beyond the guidelines at available data, combination immunotherapies are a reasonable choice for all patients regardless of PD-L1 expression, including patients with PD-L1 TPS ≥50%, particularly if they have bulky tumor and a high disease burden.4 This concept is supported by both the CheckMate 9LA and IMpower150 trials, which showed increased median overall survival (OS) across all-comers using regimens of nivolumab + ipilimumab + limited chemotherapy and atezolizumab + bevacizumab + chemotherapy, respectively, as first-line therapy in patients with advanced NSCLC (Table).10,11 Although these trials were not powered to determine survival differences between levels of PD-L1 expression, the data strongly suggest that ICI-based combination therapies are efficacious regardless of PD-L1 status, and may have a pronounced effect in patients with low or negative PD-L1 expression.
Dr Kerrigan explained that in her clinic, there is always a shared decision-making process. “Number one, symptom burden; number two, tumor burden; and number three, the patient’s goals and wishes,” she noted. “If the patient is very symptomatic, in a lot of pain, and they need a quick response and a higher likelihood of response, then I’m recommending combination chemoimmunotherapy because we know that those patients typically have a quicker time to response and are more likely to respond. If a patient has a relatively low tumor burden and/or are not symptomatic, I think immunotherapy as monotherapy is appropriate.”
When asked about his strategy for treating patients with PD-L1 TPS <1%, Dr Akerley explained, “While I do use mostly chemoimmunotherapy in those situations, I think this is the place where dual immunotherapy can be beneficial.”
When considering ICI monotherapy versus combination therapy for patients with NSCLC, it is important to weigh increased efficacy versus increased potential for adverse events (AEs).4 For example, consider the KEYNOTE-042 and CheckMate 227 trials, which evaluated first-line pembrolizumab monotherapy and nivolumab + ipilimumab, respectively, in patients with PD-L1 TPS ≥1%; ICI monotherapy led to a median OS of 16.4 months and grade 3 or 4 treatment-related adverse events (TRAEs) in 16% of patients, compared with dual checkpoint blockade (nivolumab + ipilimumab), which achieved a median OS of 17.1 months, albeit with a higher rate of grade 3 or 4 TRAEs, at 33%.12-14 Thus, for every patient with advanced NSCLC, the benefit/risk ratio of combination ICI therapy should be weighed.
Dr Kerrigan noted that less important biomarkers include tumor mutation burden and mismatch repair deficiency. “There have been exploratory analyses that suggest that tumor mutation burden is a predictor of response to immunotherapy, although in other studies that hasn’t always panned out,” she noted. “While I look at that information on the panel, I don’t necessarily use it to make first-line treatment decisions for lung cancer patients.”
While PD-L1 is the central biomarker today, new biomarkers are on the horizon, including DNA methylation patterns and RNA genes (microRNA), as well as others that measure tumor microenvironment remodeling, the extracellular matrix, and metabolic reprogramming.15-18
For all biomarkers, liquid biopsy is a particularly compelling strategy due to its convenience and because 30% of patients do not have tissue available for molecular analysis.18 As Dr Kerrigan explained, “Sometimes there are limitations to tissue-based next-generation sequencing [NGS], so we are routinely ordering blood-based NGS testing to also look for circulating tumor DNA and mutations floating in the blood.” According to Dr Akerley, the challenge for biomarker testing is related to the speed of blood versus tissue testing. “The fancier we’ve gotten on the tissue, the longer the tissue takes to get a result,” he said. “If you wait for tissue markers, you will not uncommonly wait 3 weeks. Even if the patient is healthy, it’s sometimes a very difficult time for the patient to wait.”
“I think patient comorbidities, ECOG performance status, their goals and wishes—certainly, all of these patient-related factors definitely influence our decision-making in terms of which treatment we go with,” Dr Kerrigan explained. “In my mind, patient-related factors are almost equally as important as biomarkers in deciding what treatment to recommend.”
Performance status (PS) should be taken into account during NSCLC treatment planning; however, most phase 3 clinical trials have limited enrollment to patients with an ECOG PS of 0 or 1, so relevant data are lacking, making it challenging to reliably characterize the safety and efficacy of various immunotherapies in patients with higher ECOG PS scores.4,19
Although the ASCO/OH guideline claims that evidence is too scant to offer any treatment recommendations for patients with NSCLC with an ECOG PS of 2, results of an International Expert Panel Meeting by the Italian Association of Thoracic Oncology offer some guidance.20 First-line single-agent immunotherapy for PD-L1–positive patients with a PS of 2 is thought to be feasible and safe, although efficacy will likely be inferior compared with patients with a PS of 0 or 1.20 Although combined chemotherapy plus single-agent immunotherapy and combined chemotherapy plus double-agent immunotherapy are thought to be effective in patients with a PS of 2, these regimens may not be feasible and safe based on limited available data.20
Offering an alternative viewpoint on the role of PS, a recent review by Ramnaraign and colleagues concluded that immunotherapy should not be automatically denied to patients with poor PS or advanced age based on safety concerns.19 The review highlighted a retrospective analysis of 113 patients with various tumor types, including 40 patients with NSCLC who received and discontinued single-agent immunotherapy.21 Most significantly, rates of treatment-limiting AEs did not correlate with ECOG PS or Charlson Comorbidity Index.21 The study showed that immune-related adverse events (irAEs) were actually less common among those patients with poorer PS: patients with a PS of 2 or 3 demonstrated an irAE rate of 5.6%, versus 17% for patients with a PS of 1 and 32.1% for patients with a PS of 0.21 Supporting tolerability, patients with a PS of 2 or 3 more often stopped therapy due to disease progression rather than toxicity, compared with those who had a PS of 1 or 0 (77.8% vs 64.2% vs 46.4%).21 Collectively, these findings suggest that patients with a PS of 2 or 3 can safely receive immunotherapy and may even have a lower rate of irAEs than patients with better PS. Dr Kerrigan explained, “I don’t steer away from offering immunotherapy simply based on chronologic age. I think much more significantly about biologic and physiologic age. I think that’s where your geriatric assessment can come into play in determining safety.”
According to Dr Akerley, a patient with a PS of 3 should not be treated. “I think you cause more harm than benefit,” he said. Dr Kerrigan noted that immunotherapy alone may not work rapidly enough to salvage these patients. “I think that’s really a case-by-case scenario,” she explained. “It’s hard to make a blanket recommendation in those cases.”
Multiple studies suggest that immunotherapy can extend survival in patients with NSCLC regardless of histology. This was demonstrated by both the CheckMate 227 and CheckMate 9LA trials, which evaluated nivolumab + ipilimumab and nivolumab + ipilimumab + limited chemotherapy, respectively, in patients with NSCLC with both nonsquamous and squamous histology.10,13 The CheckMate 227 trial reported an OS hazard ratio of 0.83 (0.70-0.97) for patients with nonsquamous disease and 0.70 (0.54-0.91) for patients with squamous disease compared with those treated with chemotherapy alone.13 The CheckMate 9LA trial reported hazard ratios of 0.78 (0.63-0.96) and 0.63 (0.47-0.85) for nonsquamous and squamous disease, respectively, compared with chemotherapy.10 Similarly, KEYNOTE-189 and KEYNOTE-407, which evaluated pembrolizumab + chemotherapy for populations separated by histology, reported hazard ratios for survival of 0.56 (0.45-0.70) for nonsquamous disease and 0.71 for squamous disease (0.58-0.88) compared with chemotherapy.22,23
Choosing between first-line monotherapy versus combination therapy should primarily depend on PD-L1 level, whereas histology should be considered when selecting a specific combination. For example, the ASCO guideline recommends pembrolizumab + carboplatin + pemetrexed or atezolizumab + carboplatin + paclitaxel + bevacizumab for nonsquamous disease, and pembrolizumab + carboplatin + paclitaxel for squamous disease (PD-L1 ≥1%-49%).4
A recent study by Monaco and colleagues revealed the real-world relationship between tumor burden and survival.24 The study involved 92 patients with NSCLC who were treated with nivolumab, pembrolizumab, or atezolizumab + bevacizumab in the first- to fourth-line settings.24 Patients underwent PET/CT scans to measure metabolic tumor volume (MTV), defined as the volume of metabolically active tumor.24 The analysis showed that OS dropped 22% with each 100-cm3 increase in MTV, and that patients who achieved disease control (complete, partial, or stable) had significantly lower MTV.24 The investigators concluded that MTV predicts response to therapy, so clinicians may be able to use it during treatment planning.24 MTV is typically calculated by available software based on user- or algorithm-defined regions of interest, making it easy to implement into clinical practice.24
In keeping with these findings, ASCO guidelines recommend more aggressive first-line treatment with ICI-based combination therapy in patients with higher disease burden, especially in the presence of visceral tumor involvement; this is particularly relevant in the PD-L1 ≥50% group, who may benefit more from combination therapy than monotherapy.4
Baseline symptom burden in patients with NSCLC may also predict outcomes, including those with higher Edmonton Symptom Assessment Scale scores, as well as in patients with brain metastases, among whom neurologic symptoms are prognostic.25,26 Before immunotherapy became a mainstay of treatment for NSCLC, patients with greater symptom burden were less likely to receive chemotherapy, but immunotherapy offers higher tolerability with greater clinical benefit, providing greater quality-of-life improvement than chemotherapy in the presence of more severe symptoms.25,27 The latest ASCO guidelines support this strategy, as they advise combination regimens of ICIs or ICIs plus chemotherapy instead of monotherapy when more disease-related symptoms are present.4
Not all biomarkers and clinical factors are of equal importance, and some should be considered in sequence. The first factor to consider during treatment planning is previous therapy, as hard rules apply to patients who have already received certain treatments (eg, patients who have received one ICI should not be switched to another ICI). In addition, irAEs to previous ICI therapy should be considered, as a therapy may have been effective but discontinued due to toxicity. After considering previous therapy (if applicable), PD-L1 status should be considered, then histology. Finally, other clinical factors, such as tumor burden, PS/comorbidities, and disease symptoms, can be considered in parallel, ultimately leading to immunotherapy selection. Figure 2 offers a simple roadmap for treatment planning that conceptualizes the process and may serve as a visual aid when engaging in shared decision-making with patients and other stakeholders.
“I think we should better understand the patient, their comorbidities, medications, and performance status, and combine all of that information when deciding on immunotherapy alone, combination immunotherapy, or chemoimmunotherapy,” Dr Kerrigan explained. “We need to consider numerous factors before coming to a decision about what’s best for each patient. Then, most importantly, asking the patient and understanding their goals and wishes and coming to a shared decision about which way to go is appropriate because there are a lot of options now for treating lung cancer and the patient’s voice certainly should be heard.”
Treatment planning for patients with advanced NSCLC lacking actionable alterations is complicated by a lack of head-to-head trials and PD-L1 as the one reliable biomarker. For now, PD-L1 TPS remains the most valuable tool for selecting immunotherapy, especially for patients with high expression, who may benefit from the simplicity and relatively low AE profile of ICI monotherapy. Still, this rule is not set in stone. Patients with high PD-L1 expression may still elect a combination approach to further increase efficacy, particularly in the presence of other negative prognostic factors, such as greater symptom or tumor burden.
Although it is unclear which biomarkers will join or replace PD-L1 at the forefront of treatment selection in patients with NSCLC, new tools are coming, as well as new treatment options that may alter the current clinical paradigm. Dr Akerley noted that a potentially emerging biomarker that is independent of PD-L1 provides information about the host ability to mount an immune response. Dr Kerrigan explained that there is interest in how the gut microbiome may influence response to immunotherapy and called for improved understanding of tumor signatures, the associated microenvironment, and co-occurring mutations that may be predictive of lack of benefit from single-agent immunotherapy. Dr Kerrigan explained, “Maybe we’ll identify more situations that we see on our routine panels that we can actually target with a drug.”
According to Dr Akerley, immunotherapy is a huge breakthrough that has made conversations with patients more difficult, but he also noted that there is much more hope. “If they do have a targetable mutation, they can live a near-normal life, return to work, and make dramatic recoveries, but the immunotherapies can do exactly the same thing.”
To sign up for our newsletter or print publications, please enter your contact information below.
Subscribe to recieve the free, monthly TON print publication and TON weekly e‑newsletter.