Advances in the development of targeted therapies and the identification of emerging biomarkers of response to treatment have led to an improved ability to predict and enhance responses to immunotherapeutic agents in non–small cell lung cancer.
Suresh S. Ramalingam, MD
Advances in the development of targeted therapies and the identification of emerging biomarkers of response to treatment have led to an improved ability to predict and enhance responses to immunotherapeutic agents in nonsmall cell lung cancer (NSCLC).
Available agents that have contributed to this transformation in recent years include nivolumab (Opdivo) and pembrolizumab (Keytruda), albeit with mixed results. In the first randomized phase III trials to report a 5-year outcomes for a PD-1 inhibitor in the setting of previously treated advanced or metastatic NSCLC, CheckMate 017 and CheckMate 057, data demonstrated a 5-year pooled overall survival (OS) rate of 13.4% with nivolumab versus 2.6% with docetaxel and a 5-year pooled progression-free survival (PFS) rate of 8.0% and 0%, respectively. Following treatment with nivolumab, those patients who did not exhibit disease progression at 2 and 3 years exhibited a 60% and 78% chance, respectively, of remaining progression free at 5 years. Furthermore, 10% were no longer receiving nivolumab at 5 years, had not progressed, and had not received subsequent therapy.1
Another landmark phase III trial, KEYNOTE-024, included patients with metastatic NSCLC and a PD-L1 tumor proportion score (TPS) of ≥50% who received first-line treatment with pembrolizumab or platinum-based chemotherapy. After at least 3 years of follow-up, patients treated with pembrolizumab continued to exhibit significantly increased OS despite 65% of patients assigned to the chemotherapy arm being allowed to cross over to the pembrolizumab arm. Furthermore, pembrolizumab was associated with less toxicity than chemotherapy, with a manageable long-term safety profile. A durable clinical benefit was observed in patients who completed 35 cycles (2 years) of pembrolizumab, with most patients remaining alive at data cutoff.2
Additional emerging biomarkers of treatment response and targetable tumor receptors are under investigation, and recent study results presented at 2019 World Conference on Lung Cancer held in Barcelona, Spain, spotlighted the most promising novel agents and strategies for therapy selection that are under development.Originally identified in human leukemia as a transforming gene,3,4AXL, a receptor tyrosine kinase, has been suggested to be involved in cell cycle reentry, cell survival, and cell transformation.5,6,7AXL expression has also been shown to be involved in cancer invasion, with the potential for the use of protein expression analysis as a biologic marker of tumor progression in patients with lung cancer.8
Bemcentinib is a first-in-class small molecule inhibitor of AXL kinase that has been shown to target the tumorimmune interface to dampen the aggressiveness of cancer cells in vitro and enhance chemotherapy efficacy in murine models.9
In a phase II study of patients with advanced NSCLC and adenocarcinoma histology who were immunotherapy naïve and progressed on platinum-based chemotherapy in the first line, patients were given bemcentinib 200 mg daily plus pembrolizumab 200 mg every 3 weeks to assess safety and efficacy of the combination. Among 46 patients who received at least 1 dose of the combination, 58% of 33 evaluable patients were AXL positive. Among 38 patients evaluable for PD-L1 status, 8% had a TPS ≥50% (strongly positive), 39% had a TPS of 1% to 49% (positive), and 53% were PD-L1 negative (FIGURE 1).10
Among the 35 patients evaluable for response, 10 achieved partial response (PR), 12 had stable disease (SD), and 13 had progressive disease, for an objective response rate (ORR) of 29% and a clinical benefit rate (CBR) of 63%. Among the 15 AXL-positive patients, the ORR was 40% and the CBR was 67%, whereas the 13 AXL-negative patients hadanORRof15%andaCBRof54%.
Of the 2 patients with strongly-positive PD-L1 results, 1 had a PR and 1 had progressive disease, for an ORR and a CBR of 50% each. Among the 12 patients who were PD-L1 positive, 3 achieved a PR and 5 had progressive disease, resulting in an ORR of 25% and a CBR of 58%. Among the 15 PD-L1negative patients, 4 achieved a PR and 6 had progressive disease, resulting in an ORR of 27% and a CBR of 60%.10
The median OS was 12.2 months (95% CI, 6.2not reached), and the 12-month OS rate was 54.2%.9 Among AXL-positive patients, the median PFS was 5.9 months compared with 3.3 months in AXL-negative patients. Adverse events (AEs) leading to treatment discontinuation included transaminitis, fatigue, asthenia, pneumonia, and aspartate aminotransferase increase.10
Overall, the trial investigators concluded that these results suggest that the combination of bemcentinib and pembrolizumab is well tolerated and leads to promising clinical activity, particularly in patients with AXL-positive tumors with low or no PD-L1 expression.9
The AXL-specific human immunoglobulin G1 antibody conjugated with monomethyl auristatin E (MMAE), enapotamab vedotin (ENA V)11 has demonstrated a primary mechanism of action based on MMAE-mediated killing of AXL-positive tumor cells and AXL-negative bystander tumor cells.12
Results for patients with NSCLC without a known activatingEGFRmutation orALKalteration from a phase I/IIa dose-escalation/dose-expansion trial of ENA V were presented at the meeting and showed encouraging preliminary activity in those who had previously failed on immune checkpoint inhibitors and platinum based-chemotherapy.
“This was a phase I study with a built-in expansion cohort for patients with advanced-stage NSCLC,”Suresh S. Ramalingam, MD, professor in the Department of Hematology and Medical Oncology at Emory University School of Medicine, director of the Lung Cancer Program at Winship Cancer Institute of Emory University, both in Atlanta, Georgia, and lead author of the study, toldTargeted Therapies in Oncologyin an interview. “After the recommended dose of ENA V was established, we proceeded to enroll patients with advanced-stage NSCLC that had progressed on prior standard therapies. The intent was to evaluate the efficacy of ENA V in this patient population.”
Among 26 patients, the median treatment duration was 6 weeks (range, 2.0-24.7 weeks), and 25 patients discontinued treatment because of progressive disease (38%), adverse events (AEs; 38%), death (12%), or other (8%).11
“The safety profile is acceptable, with constipation, neuropathy, and fatigue as salient AEs,” Ramalingam said.
The ORR was 19%, and the disease control rate was 50%, with a median time to onset of response of 11 weeks (95% CI, 5-15) and a median duration of response of 18.1 weeks.11
“This is a group of patients for whom there is an urgent need to develop novel treatment approaches. We view these as promising results, given that the patients enrolled had progressed on standard therapies,” Ramalingam said. “ENA V is associated with single-agent activity in advanced NSCLC.”Previous studies have demonstrated a synergistic effect of antiangiogenic therapy and immunotherapy. A recent phase I trial was the first designed to evaluate a chemotherapy-free regimen combining the PD-1 inhibitor sintilimab with the angiogenesis modulator anlotinib in treatment-naïve patients with advanced NSCLC.13
Among 22 enrolled patients, 16 achieved a confirmed PR, for an ORR of 72.7% and a disease control rate of 100%. Responses were consistent by subgroup evaluation based on tumor PD-L1 expression and tumor mutational burden (TMB) status (F I G U R E 2).13The 6-month PFS rate was 93.8%, and no unexpected toxicities were observed. Overall, these results suggest that the combination of sintilimab and anlotinib may be a potential first-line treatment option for patients with advanced NSCLC. The investigators of the trial concluded that based on these results, further confirmation in larger studies is warranted.13
Within the nivolumab arm, investigators confirmed an association betweenTP53mutations and poor outcomes; however, molecular abnormalities were unable to predict outcomes for the nivolumab/ ipilimumab arm.14Recently, relevant clinical trials of immunotherapy agents have yielded studies evaluating the association between mutations and sensitivity or resistance to immunotherapy for better-informed treatment decision making.
In the open-label, randomized phase III MYSTIC study, 1118 immunotherapy- and chemotherapy-naïve patients with stage IV NSCLC were stratified by PD-L1 expression (<25% vs ≥25%) and histology and randomized to receive durvalumab (Imfinzi), durvalumab plus tremelimumab, or platinum-based chemotherapy. Results of the trial were presented in late 2018 and showed that the primary endpoints of OS and PFS superiority with durvalumab/tremelimumab versus chemotherapy were not met, but further analyses into patient subgroups that may benefit from this regimen are ongoing.15
Regardless of treatment, OS was shorter for patients with mutations inSTK11orKEAP1compared with patients with wild-type disease. Patients with mutations inARID1Aexhibited improved OS and ORR following treatment with durvalumab plus tremelimumab compared with chemotherapy. Mutations inARID1AandKEAP1were positively associated with blood TMB ≥20 mutations per megabase, whereas mutations inSTK11were positively associated with PD-L1 expression <25%. These results provide information on the impact of specific mutations on immunotherapy response, but further validation is needed.16
In a retrospective study of patients with metastatic nonsquamous NSCLC,STK11andKEAP1genomic alterations were associated with shorter PFS with frontline chemoimmunotherapy or chemotherapy treatment. Furthermore,STK11genomic alterations were predictive of inferior clinical outcomes with chemoimmunotherapy in patients with PD-L1positive tumors.
Over 75% of patients with progressive disease as the best objective response to chemoimmunotherapy hadSTK11and/orKEAP1mutations. Patients with PD-L1positive disease withSTK11and/orKEAP1mutations experienced shorter PFS and a lack of benefit from the addition of pembrolizumab to chemotherapy. These results suggest that patients with NSCLC andSTK11and/orKEAP1mutations do not derive the same benefit with standard-of-care chemoimmunotherapy regimens and represent an unmet need for novel treatment strategies that contribute to effective antitumor immunity.17
Despite heavy interest in its use for predicting outcomes for immunotherapy, TMB has become an elusive biomarker for assessing treatment response in NSCLC, with many investigators striving to incorporate its use into practice with mixed results.
In the phase III CheckMate 227 trial, patients with advanced NSCLC and high TMB exhibited prolonged PFS with nivolumab plus ipilimumab compared with chemotherapy, irrespective of PD-L1 expression.17 However, updated survival data as of October 2018 demonstrated superior results with nivolumab plus ipilimumab compared with chemotherapy, irrespective of TMB level.18
It is important to remember that not all non-synonymous somatic mutations are equally immunogenic, Mari Mino-Kenudson, MD, director of pulmonary pathology at Massachusetts General Hospital in Boston, told an audience at the WCLC.19Because few somatic mutations ultimately prove immunogenic, the odds of T-cellmediated killing are increased by maximizing the number of different mutations as well as both their length (with frame-shift and nonsense preferable to missense mutations) and prevalence (with early-arising, widespread truncal mutations preferable to the later-arising, rarer branch mutations).20The nonsynonymous mutational burden has been associated with the clinical benefit of antiPD-1 therapy in patients with NSCLC.21
The Lung-MAP substudy S1400I was a randomized phase III trial designed to assess nivolumab versus the combination of nivolumab plus ipilimumab as second-line therapy in patients with metastatic squamous cell lung cancer with no matching biomarker.13When PD-L1 or TMB was used as a biomarker, outcomes between nivolumab monotherapy and the combination of nivolumab/ipilimumab showed no statistically significant differences. However, exploratory analyses of the combination of PD-L1 and TMB suggest that for patients with 0% PD-L1 expression, TMB ≥10 significantly improved outcomes with the immunotherapy combination versus nivolumab (HR, 0.39; 95% CI, 0.16-0.93;P= .004), whereas it was detrimental to patients with TMB <10 (HR, 2.52; 95% CI, 1.03-6.13;P= .042).14
Mino-Kenudson reviewed several issues that have been identified regarding the use of TMB for predicting response to immune checkpoint inhibitor therapy. Questions surround the choice of the best cutoff values for classification of high TMB; the use of whole-exome sequencing (WES) or next-generation sequencing (NGS); the bioinformatics process; and biological variables, including tumor cell content and intratumoral heterogeneity; within NGS, questions arise regarding the size of the panel and genetic region covered, the depth of coverage, and the gene selection biased |toward frequently mutated genes associated with cancer.19
Mino-Kenudson further reviewed proposed recommendations for consistent assessment of TMB. Regarding sample processing, protocols should be standardized and interlaboratory variability should be minimized. Gene panels should be selected to screen for actionable mutations or biomarkers, and these panels should have larger genome coverage, ideally, of approximately 1 megabase or more. For standardization of workflow, panel-derived TMB values should be aligned to a WES analysisderived reference standard to ensure consistency, irrespective of the assay used. Bioinformatics algorithms used for mutation calling and filtering should be standardized. To calibrate outputs, reporting consistency should be ensured via development of templates for clinically meaningful reporting, and calibration of results from different studies should be allowed.19,22
TMB efficiency is also attenuated based on the inclusion of adverse mutations, such as those inEGFR. To combat this, the tumor mutation score was developed as an improved biomarker for response of immune checkpoint inhibitors in 240 patients with NSCLC treated with antiPD-1/ L1 therapy with or without anti–CTLA-4 from the MSK-IMPACT cohort. The tumor mutation score was defined as the number of genes with nonsynonymous somatic mutations. Significant associations with prolonged PFS and improved responses were observed with 12 genes and 11 genes, respectively. The tumor mutation score of 18 favorable genes (TMS18) was significantly correlated with overall TMB. Differences among patients with complete response, PR, SD, and progressive disease were more pronounced using TMS18 than TMB or PD-L1 expression. The combination of TMS18 and PD-L1 expression had a higher area under the curve for prediction of response; however, this difference was not significantly better than TMS18 alone.23
Overall, these findings suggest that the selective TMS18 may be a better biomarker for response to immunotherapy than TMB in patients with NSCLC. Additionally, the combination of TMS18 and PD-L1 expression may contribute to improved efficiency for predicting response to immune checkpoint inhibitors; however, validation in larger cohorts is required.23
These advances have contributed to an improved ability for predicting and enhancing responses to immunotherapeutic agents; however, additional questions remain regarding standardization of biomarker measurement and identification of optimal patient subgroups for treatment. Patient subgroups identified as having unmet needs will still require novel therapies and strategies for improved responses to therapy and overall patient outcomes.