Tumor Immunology and Immunotherapy of Non-Small-Cell Lung Cancer

Abstract

Historically, non-small-cell lung cancer (NSCLC) has been regarded as a nonimmunogenic tumor; however, recent studies have shown that NSCLCs are among the most responsive cancers to monoclonal antibody immune checkpoint inhibitors (ICIs). ICIs have dramatically improved clinical outcomes for a subset of patients (∼20%) with locally advanced and metastatic NSCLC, and they have also demonstrated promise as neoadjuvant therapy for early-stage resectable disease. Nevertheless, the majority of patients with NSCLC are refractory to ICIs for reasons that are poorly understood. Thus, major questions are: how do we initially identify the patients most likely to derive significant clinical benefit from these therapies; how can we increase the number of patients benefiting; what are the mechanisms of primary and acquired resistance to immune-based therapies; are there additional immune checkpoints besides PD-1/PD-L1 and CTLA-4 that can be targeted to provide greater clinical benefit to patients; and how do we best combine ICI therapy with surgery, radiotherapy, chemotherapy, and targeted therapy? To answer these questions, we need to deploy the latest technologies to study tumors and their microenvironment and how they interact with components of the innate and adaptive immune systems. There is also a need for new preclinical model systems to investigate the molecular mechanisms of resistance to treatment and identify novel therapeutic targets. Recent advances in technology are beginning to shed new light on the immune landscape of NSCLC that may uncover biomarkers of response and maximize the clinical benefit of immune-based therapies. Identification of the mechanisms of resistance should lead to the identification of novel targets and the generation of new therapeutic strategies that improve outcomes for a greater number of patients. In the sections below, we discuss the results of studies examining the immune microenvironment in NSCLC, summarize the clinical experience with immunotherapy for NSCLC, and review candidate biomarkers of response to these agents in NSCLC.

Lung cancer is the most common cause of cancer-related mortality worldwide and accounts for more than 1.7 million deaths each year (Bray et al. 2018). Approximately 85% of lung cancers are characterized as non-small-cell lung cancer (NSCLC), which includes three major subtypes: adenocarcinoma (40%), squamous cell carcinoma (30%), and large-cell carcinoma (10%) (Herbst et al. 2018). For many years, the major pillars of NSCLC care consisted of surgery, chemotherapy (CT), and radiation therapy (RT). Recently, immunotherapy was added as a fourth pillar of the cancer treatment plan and provided the first opportunity to treat patients for a potential cure. Now, treatment is guided by molecular testing, and some patients whose tumors harbor actionable molecular aberrations receive personalized therapy that includes small molecule inhibitors (Chen et al. 2014; Garinet et al. 2018). Whereas these agents have been shown to provide a survival benefit for some patients with NSCLC (Kris et al. 2014), not all lung cancers harbor actionable genomic targets.

Immune checkpoint inhibitors (ICIs) may be the best treatment option for a subset of patients with NSCLC. Current clinical guidelines recommend that all newly diagnosed NSCLC patients have their tumors submitted for molecular analyses to complement standard histologic evaluation. After complete staging, decisions are made about whether the patient should undergo a curative surgical resection, with or without neo-adjuvant (pre and/or postoperative) platinum-based CT, potentially followed by adjuvant osimertinib if tumors harbor an epidermal growth factor receptor (EGFR) mutation (mut) (Wu et al. 2020), or if an approach with chemoradiotherapy (CRT) with consolidation immunotherapy (Gray et al. 2020) should be considered. If a patient has metastatic disease and their tumor has an actionable genomic alteration (e.g., EGFR mutations, ALK [anaplastic lymphoma kinase] rearrangements), the patient will receive a targeted therapy directed against that aberration. If not, the patient will receive first-line therapy with ICI with or without CT, and with or without another biologic agent (e.g., bevacizumab) (Chiang and Herbst 2020). The molecular analyses also provide information on the tumor mutation burden (TMB), mutations in key oncogenes/tumor suppressor genes, and standard tumor immunohistochemistry (IHC) information on the expression of PD-L1 (programmed death-ligand 1), which may be used to estimate the likelihood of response to the ICI therapy. ICIs are generally better tolerated than CT and have the potential to provide durable responses by imparting immunologic memory (Brahmer 2013). However, the results generated from recent clinical studies suggest that the majority of unselected patients with lung cancer are refractory to ICIs for reasons that remain unclear. Consequently, there in an urgent need to better understand the cellular and molecular mechanisms that govern the antitumor immune response to identify novel therapeutic targets and develop biomarkers predictive of response to immune-based therapies.

The mechanisms that mediate tumor resistance to ICIs are not completely understood and are the subject of much investigative effort. Tumor genomic aberrations recently shown to be associated with lack of responses to ICIs in NSCLC include STK11 mutations (Skoulidis et al. 2018), EGFR and ALK alterations (Gainor et al. 2016; Gavralidis and Gainor 2020). While some of these genetic alterations are useful biomarkers routinely obtained in patients newly diagnosed with NSCLC to guide therapy, we need to better understand how they contribute to ICI resistance. The examination of the tumor immune contexture, which refers to the spatial orientation, density, and functional role of the different immune cell populations in tumors (Fridman et al. 2012), can yield invaluable information regarding patient prognosis and therapeutic response (Fridman et al. 2017). Analyses of the immune microenvironment have been aided by a variety of techniques, including NanoString technology, RNA-sequencing, cytometry by time-of-flight, intracellular staining flow cytometry, barcoding antibody-based protein arrays (Lyons et al. 2017), and single-cell RNA sequencing (Shalek et al. 2013; Sinjab et al. 2021). In the next sections, we review characteristics of the immune microenvironment in NSCLC and the mechanisms responsible for immune editing of tumors. We also discuss the results from select studies that have led to the use of ICIs for the treatment of NSCLC and the role of biomarkers of therapeutic response.

TUMOR IMMUNE MICROENVIRONMENT IN NSCLC

In health, the innate and adaptive immune systems (Fig. 1) engage in a cooperative series of interactions to protect the lung from pathogens and particulate matter that may enter the respiratory tract (Chen and Kolls 2013). Recent studies suggest that the ability of the immune system to ensure pulmonary homeostasis may be regulated, in part, by the composition of bacteria residing in the gut and lung (Budden et al. 2017). It has also become increasingly clear that the pulmonary immune system may undergo dramatic changes in response to aging (Sharma et al. 2009; Lloyd and Marsland 2017) and lung disease (Sharma et al. 2009). How these changes in immune function affect the development and immunogenicity of NSCLC is not fully understood.

Figure 1.

The cellular mediators of the innate and adaptive immune response. Hematopoietic stem cells differentiate into lymphoid and myeloid progenitors that further branch out into specific cell types mediating innate and adaptive immune response. Cells of the innate immune system provide initial response against non-self antigens, while adaptive immune response is highly specific and is mediated by the activation of lymphocytes. The cellular components of innate immunity include mast cells, basophils, eosinophils, and phagocytic cells, including dendritic cells, macrophages, and neutrophils. Soluble molecules, including complement, acute phase proteins, chemokines, and cytokines make up the humoral component of innate immunity. The cellular components of adaptive immunity include T lymphocytes (CD8⁺ and CD4⁺ [TH1, TH2, TH17, and iTREG]), B lymphocytes, and natural killer (NK) cells while immunoglobulins, chemokines, and cytokines make up the humoral component of the adaptive immune system. NK cells, neutrophils, and iTREG cell subsets overlap between both arms of immune responses. (CD) Cluster of differentiation, (TH) T helper cells, (iTREG) induced T regulatory cells.

Reports comparing the immune content of the uninvolved lung with the cancerous lung suggest that the two tissues exhibit marked differences in their immune infiltrates. Studies examining the spatial distribution of tumor-infiltrating immune cells indicate that the cells have a tendency to localize in the stromal compartment and invasive margin of early-stage NSCLC (Lavin et al. 2017). Neutrophils are among the most prevalent immune cells found in NSCLC (Kargl et al. 2017; Stankovic et al. 2019). When compared with blood neutrophils, tumor-associated neutrophils (TANs) exhibit an activated phenotype (CD62L^loCD54^hi) and express a distinct pattern of chemokine receptors, including up-regulation of C-C motif chemokine receptor (CCR) CCR5, CCR7 and C-X-C motif chemokine receptor CXCR3, and CXCR4 (Eruslanov et al. 2014). TANs isolated from early-stage NSCLCs were found to produce proinflammatory cytokines (MCP-1 [monocyte chemoattractant protein-1], IL-8, MIP-1α [macrophage inflammatory protein-1α], IL-6) and stimulate T-cell proliferation and IFN-γ (interferon γ) secretion (Eruslanov et al. 2014), and to cross-present antigens and amplify effector T-cell responses (Singhal et al. 2016). Whereas the results of these studies advocate a cytotoxic role for TANs in early-stage NSCLC, examinations of experimental lung cancer models suggest that TANs may acquire protumorigenic properties as the disease progresses (Mishalian et al. 2013).

Monocytes traffic to tumors in response to a variety of chemotactic factors in the tumor microenvironment, including MCP-1, TGF-β (transforming growth factor β), and VEGF (vascular endothelial growth factor) (Mantovani et al. 1986; Wahl et al. 1987; Barleon et al. 1996). Tissue-resident alveolar macrophages may contribute to the pool of tumor-associated macrophages (TAMs), but reports indicate that the majority of TAMs are derived from monocyte recruitment from the periphery (Yang and Zhang 2017). TAMs comprise only 5% of the immune infiltrate in NSCLC (Stankovic et al. 2019), and yet are one of the key effectors in mediating cancer-related inflammation (Sica 2010). Macrophage content in early-stage NSCLC is comparable to that in uninvolved lung, but TAMs express higher levels of PD-L1 (Lavin et al. 2017; Liu et al. 2020), a protein that suppresses the adaptive arm of the immune system. In general, TAMs are associated with an M2 phenotype, oriented toward promoting tumor growth, tissue remodeling, and suppressing adaptive immunity (Mantovani et al. 2008). TAMs may also modulate tumor progression through their promotion of tumor angiogenesis by producing potent endothelial cell mitogens, including VEGF and fibroblast growth factor 2 (FGF-2) (Baird et al. 1985). VEGF may also interfere with immune cell recruitment by reducing lymphocyte adhesion to activated endothelial cells (Bouzin et al. 2007).

T lymphocytes are enriched in early-stage NSCLCs when compared to matched uninvolved lung specimens (Ganesan et al. 2013; Lavin et al. 2017; Stankovic et al. 2019). In an examination of resected tumors from 68 patients with localized NSCLC, the most common T cells in tumors were CD4⁺ cells (26%) followed by CD8⁺ cells (22%) (Stankovic et al. 2019). Results from single-cell analysis of 32 resected NSCLCs and matched uninvolved lung tissues indicate that tumors contain higher frequencies of nonfunctional T cells and Forkhead box P3 (Foxp3)-expressing CD4⁺ T regulatory (TREG) cells than the uninvolved lung (Lavin et al. 2017). Regulatory CD4⁺CD25⁺ T cells that were isolated from stage I–II NSCLCs inhibited the proliferation of autologous peripheral blood T cells (Woo et al. 2002), suggesting that immune suppression may be an early event in the disease process. Increased TREG counts are associated with poor outcomes in NSCLC (Shimizu et al. 2010; Tao et al. 2012), whereas high densities of CD8⁺ T cells are an independent positive prognostic predictor for resectable NSCLC (Al-Shibli et al. 2008). Predominant infiltration of CD8⁺ T cells and macrophages into the cancer compartment as opposed to the stromal compartment has also been shown to be an independent predictor of survival in patients with advanced NSCLC (Kawai et al. 2008). Studies have shown that natural killer (NK) cells are reduced in NSCLC compared to noncancerous lung tissue, and that tumor-infiltrating NK cells have diminished cytotoxic function (Platonova et al. 2011; Lavin et al. 2017; Stankovic et al. 2019).

B lymphocytes are one of the most prevalent immune cells in NSCLC (Del Mar Valenzuela-Membrives et al. 2016; Stankovic et al. 2019), and are commonly found in tertiary lymphoid structures (TLS) located at the interface between tumor and stroma. The organization of these structures resembles that of secondary lymphoid organs. TLS contain a T-cell zone with lysosome-associated membrane glycoprotein 3-positive mature dendritic cells (DCs) and a segregated follicular zone where CD20⁺ B cells actively proliferate and differentiate into plasma cells (Germain et al. 2014; Fridman et al. 2017). A high density of follicular B cells correlates with long-term survival in patients with NSCLC (Germain et al. 2014). Immunohistochemical studies of resected NSCLCs suggest that the molecular mechanisms that mediate T-cell recruitment to TLS are similar to those that govern T-cell trafficking to lymph nodes and involve adhesive interactions between CD62L (L selectin) expressed on T cells and its corresponding counter-receptors located on the surface of high endothelial venules (de Chaisemartin et al. 2011). In addition to CD20⁺ B cells, the follicular zone also contains CD3⁺ T cells, a network of CD21⁺ DCs, CD68⁺ macrophages, and shows evidence of an ongoing humoral response (Germain et al. 2014). B cells play an important role in regulating NSCLC progression. Several studies have demonstrated an antitumor role for B cells in NSCLC by virtue of their ability to produce tumor-specific antibodies (Germain et al. 2014) and enhance effector T-cell responses (Wang et al. 2019a). Tumor-infiltrating B cells also present antigen to CD4⁺ T cells and can modulate the CD4⁺ T-cell phenotype (Bruno et al. 2017). Alternatively, B regulatory cells have the potential to impair an antitumor immune response by producing immunosuppressive cytokines, such as interleukin (IL)-10 and TGF-β (Wang et al. 2019a). This dual ability of the immune system to both restrain and promote tumor growth forms the basis for the cancer immunoediting hypothesis, which is discussed below.

IMMUNOEDITING AND MECHANISMS OF IMMUNOSUPPRESSION IN NSCLC

The cancer immunoediting hypothesis proposes that the immune system provides protection against tumor formation and shapes tumor immunogenicity through a series of three related processes: elimination, equilibrium, and escape (Schreiber et al. 2011). The hypothesis provides a framework for studying the complex cellular and molecular interactions between tumor and immune cells. During the elimination phase, the innate and adaptive immune system cooperate to recognize transformed cells that have escaped intrinsic tumor suppression and eliminate the cells before a tumor becomes clinically detectable (O'Donnell et al. 2019). If the cancer is completely cleared during the elimination phase, the process is complete. However, cancer cells that are able to survive the elimination phase may pass into the equilibrium phase. During the equilibrium phase, net growth is restricted and cellular immunogenicity is edited by the adaptive immune system (Koebel et al. 2007; O'Donnell et al. 2019). Tumor variants possessing either dampened immunity or the capacity to attenuate immune responses may enter the escape phase and give rise to clinically apparent cancers (Koebel et al. 2007).

T cells play a critical role during the equilibrium process by virtue of their ability to recognize tumor antigens and restrict tumor growth. Antigens capable of eliciting a T-cell response are typically placed into three categories: viral antigens, antigens that are the product of a gene mutation or rearrangement, and antigens that are encoded by cancer-germline genes (Coulie et al. 2014). Smoking-related NSCLCs possess one of the highest TMBs among all cancers (Alexandrov et al. 2013). However, studies in melanoma, which have a TMB comparable to that of NSCLC (>10 somatic mutations [mut] per megabase [mb] of coding DNA [deoxyribonucleic acid]), suggest that only a small fraction of nonsynonymous mutations actually results in neoantigens capable of eliciting tumor-infiltrating lymphocyte (TIL) reactivity (Linnemann et al. 2015; Schumacher and Schreiber 2015). The diminished number of immunogenic neoantigens in NSCLC may result from deficiencies in the antigen-presenting machinery (Korkolopoulou et al. 1996). Promoter hypermethylation has also been identified as a mechanism of transcriptomic neoantigen depletion (Rosenthal et al. 2019).

Studies have also shown that soluble factors in the tumor immune microenvironment may contribute to tumor escape. DCs play a critical role in mediating an antitumor immune response by acquiring, processing, and presenting tumor-associated antigens on major histocompatibility complex (MHC) molecules and providing the costimulatory and soluble factors that shape the T-cell response (Wculek et al. 2020). Cell-based studies suggest that cross talk between NSCLC cells and DCs may contribute to an impaired immune response via secretion of immunosuppressive cytokines, such as TGF-β1 from DCs (Dumitriu et al. 2009).

Tumor metabolism may also play a critical role in cancer development and therapeutic resistance (Chang et al. 2015; Altman et al. 2016; Nakazawa et al. 2016), and is now regarded as a hallmark of cancer (Hanahan and Weinberg 2011). Cancer cells rewire their metabolism to promote their growth and ensure their survival through increased glucose metabolism even in the presence of functioning mitochondria (i.e., the Warburg effect) (Liberti and Locasale 2016). The up-regulation of cancer cell glycolysis leads to accumulation of lactate in the tumor microenvironment, which, in turn, leads to the polarization of TAMs to an M2 phenotype (Colegio et al. 2014). Lactate also suppresses the activity of T cells and NK cells (Renner et al. 2017). Increased tumor glycolysis limits T-cell infiltration of NSCLCs and melanomas (Cascone et al. 2018). Melanoma-derived lactic acid impaired effector T-cell function in vitro, and inhibition of lactic acid rendered patient-derived melanoma cells susceptible to autologous TIL-mediated killing (Cascone et al. 2018). In a retrospective analysis of tumor preoperative positron emission tomography (PET) ¹⁸F-FDG uptake in 59 resected NSCLCs, we found that each ¹⁸F-FDG PET parameter was positively correlated with tumor bulk expression of glycolysis-related genes and increased FDG (fluorodeoxyglucose) uptake was associated with an immunosuppressive and poorly immune infiltrated tumor microenvironment (Mitchell et al. 2020). Other molecules, such as ectoenzymes, are up-regulated on tumor and tumor-infiltrating cells and lead to immunosuppressive conditions. A good example that has been identified as a mediator of tumor immunosuppression contributing to ICI-resistance in NSCLC is CD38, which is induced by all-trans retinoic acid and IFN-β in the tumor microenvironment. CD38 drives adenosine production and modulates the intermediates of the adenosine pathway, producing inhibition of CD8⁺ T-cell function (Chen et al. 2018; Konen et al. 2020).

IMMUNE CHECKPOINT PROTEINS

Another way in which tumors limit an immune response is by exploiting immune checkpoint proteins (Fig. 2). The programmed cell death protein 1 (PD-1) immune checkpoint is expressed on the surface of T cells, B cells, NK cells, and some myeloid cells (Agata et al. 1996). PD-1 has two known ligands, PD-L1 and PD-L2 (programmed death-ligand 2), and engagement of PD-1 with its ligands inhibits T-cell receptor (TCR)-mediated lymphocyte proliferation and cytokine secretion (Freeman et al. 2000). PD-L1 is constitutively expressed on several immune cell types (Ishida et al. 2002), and expression is inducible on cancer cells in response to IFN-γ (Dong et al. 2002). EGFR mutations may also up-regulate PD-L1 on NSCLC cells (Akbay et al. 2013). PD-L2 has a more restricted expression pattern and is inducible on DCs and macrophages in response to IFN-γ, granulocyte-macrophage colony-stimulating factor (GM-CSF), and IL-4 (Yamazaki et al. 2002). In NSCLC, PD-L1 expression is particularly abundant in macrophages in both the tumor and stromal compartments (Liu et al. 2020). PD-L1-expressing tumor cells and antigen-presenting cells can suppress tumor immunity through multiple mechanisms, including induction of T-cell apoptosis, anergy, functional exhaustion, and IL-10 production (Dong et al. 1999, 2002; Selenko-Gebauer et al. 2003; Chen and Han 2015).

Figure 2.

Immune checkpoint receptors expressed on immune cells and their respective ligands on antigen-presenting cells (APCs) and/or tumor cells highlighting select targetable receptors. (PD-1) Programmed cell death protein 1, (PD-L1) programmed cell death-ligand 1, (CTLA4) cytotoxic T-lymphocyte-associated protein 4, (LAG3) lymphocyte activation gene 3, (MHC-II) major histocompatibility complex-class II, (TIM3) T-cell immunoglobulin and mucin-domain containing 3.

The cytotoxic T lymphocyte-associated protein 4 (CTLA-4) immune checkpoint protein functions to down-regulate immune responses. CTLA-4 attenuates TCR signaling through competition with the costimulatory molecule CD28 for the B7 ligands CD80 and CD86 (Wei et al. 2018). CTLA-4 also suppresses T-cell responses by blocking IL-2 production, IL-2 receptor expression, and cell-cycle progression of activated T cells (Walunas et al. 1996). The physiologic function of CTLA-4 is to suppress T-cell responses to self-antigens by controlling TREG activity (Chen and Han 2015). Consequently, blocking CTLA-4 signaling can result in autoimmune damage to different organ systems (Tarhini 2013). Studies in melanoma indicate that autoimmunity correlates with tumor regression in patients treated with anti-CTLA-4 inhibitors (Attia et al. 2005).

The lymphocyte activation gene-3 (LAG-3) (CD223) immune checkpoint protein is ∼20% identical to CD4 (Triebel et al. 1990). LAG-3 is expressed on activated CD4⁺ and CD8⁺ T cells, TREGs, a subpopulation of NK cells, B cells, and plasmacytoid DCs (Qin et al. 2019). MHC-II is a canonical LAG-3 ligand and galectin-3, LSECtin, a-synuclein and fibrinogen-like protein 1 (FGL1) also interact with LAG-3 (Qin et al. 2019). LAG-3 signaling plays a negative regulatory role in T helper 1 (TH1) cell activation, proliferation, and cytokine secretion (Workman et al. 2002). Dual blockade of LAG-3 and PD-1 has been shown to enhance antitumor immunity and suppress tumor growth by reinvigorating CD8⁺ TILs and decreasing TREGs in the tumor microenvironment (Huang et al. 2015a).

T-cell immunoglobulin and mucin-domain containing-3 (TIM-3) is an immune checkpoint protein expressed on CD4⁺ TH1 and CD8⁺ T cytotoxic cells and on other immune cells, including B cells, TREGs, NK cells, DCs, monocytes, and macrophages (Anderson 2012). Four distinct ligands have been reported to bind to the IgV domain of TIM-3, including galectin-9, high-mobility group protein B1 (HMGB1), carcinoembryonic antigen cell adhesion molecule 1 (Ceacam-1), and phosphatidyl serine (PtdSer) (Qin et al. 2019). The binding of TIM-3 to its ligands induces intracellular calcium flux of TH1 cells, leading to cell death and negative regulation of T-cell responses (Huang et al. 2015b).

In the sections below, we summarize some of the key findings from completed and ongoing clinical studies of cellular therapy and ICIs for the treatment of patients with NSCLC.

CLINICAL EXPERIENCE WITH ADOPTIVE CELL THERAPY WITH TILs IN METASTATIC NSCLC

Adoptive T-cell transfer therapy in metastatic melanoma patients has produced durable complete response rates of 10%–15% (Rosenberg and Restifo 2015). A phase 1 trial recently reported the initial results of TIL transfer as a therapeutic approach in patients with metastatic NSCLC (Creelan et al. 2020). Autologous TILs cultured and expanded from a metastatic lesion were administered as a single infusion in patients who progressed after four doses of nivolumab, followed by IL-2 intermediate treatment, and nivolumab maintenance treatment for 1 year. TIL therapy led to a best objective response rate (ORR) of 25% with an overall manageable toxicity profile and most patients exhibited low or no tumor PD-L1 expression as well as low TMB. Interestingly, infused TILs were reactive to autologous tumor and, in ∼10% of cases, to predicted neoepitopes. Analyses of the TCR clonotype and phenotype indicated good persistence of the transferred TILs, which, in some cases, correlated with clinical benefit (Creelan et al. 2020). These results are encouraging and lend support for continuing to test and optimize TIL therapy for patients with NSCLC. An ongoing phase 2 trial is evaluating a single dose of patient's tumor-derived TILs followed by either a high or low dose of aldesleukin (a recombinant form of human IL-2) in subjects with metastatic NSCLC (NCT02133196).

CLINICAL TRIALS OF IMMUNE CHECKPOINT INHIBITORS FOR ADVANCED/METASTATIC NSCLC

First-Line Therapy

First-line therapy with nivolumab, a fully humanized IgG4 (immunoglobulin G4) antibody directed against PD-1, or atezolizumab, a fully humanized IgG1 monoclonal antibody that targets PD-L1, in patients with PD-L1 positive metastatic NSCLC did not produce survival benefit compared with CT in either the CheckMate 026 (Carbone et al. 2017) or IMpower110 (Spigel et al. 2019) trials, respectively. A significant survival improvement was only noted with atezolizumab in tumors with high PD-L1 expression. In the phase 3 KEYNOTE-024 trial comparing pembrolizumab, a fully humanized IgG4 antibody targeting PD-1, with CT in patients with advanced NSCLC, patients whose tumors expressed PD-L1 ≥ 50% and were treated with pembrolizumab, had longer progression-free survival (PFS) and overall survival (OS) than patients treated with CT (Reck et al. 2016). In the phase 3 KEYNOTE-042 trial comparing pembrolizumab with CT for patients with metastatic NSCLC, the OS benefit with pembrolizumab was largely due to patients with high PD-L1 expression (Mok et al. 2019; Morgensztern 2019). A recent retrospective analysis revealed that survival outcomes with first-line pembrolizumab treatment are improved in patients with high PD-L1 expression, especially for those whose tumors have a PD-L1 ≥ 90% (Aguilar et al. 2019).

The CheckMate 227 study is a multipart phase 3 trial that compared first-line nivolumab plus ipilimumab (CTLA-4 inhibitor) with CT in patients with advanced/metastatic NSCLC (Hellmann et al. 2018). PFS was significantly longer with nivolumab plus ipilimumab than with CT (7.2 vs. 5.5 mo) in patients with a high TMB (≥10 mut/mb), irrespective of PD-L1 expression. There was no PFS difference noted in patients with low TMB (<10 mut/mb), and no differences in OS benefit between both tumor TMB cutoffs assessed. In the portion of the study that reported the coprimary end point OS in patients with PD-L1 ≥ 1%, nivolumab plus ipilimumab was associated with a significant OS improvement compared to CT (17.2 vs. 12.2 mo). This benefit occurred for populations with PD-L1 ≥ 1% (HR 0.79) and PD-L1 < 1% (HR 0.62) (Hellmann et al. 2019). These results led to Food and Drug Administration (FDA) approval of the combination as first-line treatment for patients with metastatic NSCLC with no EGFR and ALK genomic aberrations and whose tumors express PD-L1 (≥1%), as determined by an FDA-approved test. In the phase 3 MYSTIC trial, durvalumab (PD-L1 inhibitor) alone or in combination with tremelimumab (CTLA-4 inhibitor) did not induce significant survival improvements in patients with PD-L1 > 25% when compared to CT (Rizvi et al. 2020).

CT can potentially synergize with ICIs through several different mechanisms, including killing tumor cells and improving the T cell-to-tumor ratio, reducing immunosuppressive molecules, and by releasing antigens for presentation, thereby broadening the antitumor response (Herbst and Sznol 2016). CT has also been shown to increase PD-L1 expression in NSCLC cells (Parra et al. 2018). In the phase 3 trial KEYNOTE-189 comparing first-line pembrolizumab plus CT versus CT alone in metastatic nonsquamous NSCLC without EGFR or ALK aberrations, the combination significantly increased both 12-mo OS (69.2% vs. 49.4%) and median PFS (8.8 vs. 4.9 mo) compared with CT (Gandhi et al. 2018). In the phase 3 KEYNOTE-407 trial that investigated the addition of pembrolizumab to first-line CT in patients with squamous NSCLC, pembrolizumab plus CT produced a significant increase in OS compared with CT (15.9 vs. 11.3 mo) regardless of PD-L1 expression (Paz-Ares et al. 2018). Other phase 3 trials tested atezolizumab combined with CT (IMpower132 and IMpower130) (Barlesi et al. 2018; West et al. 2019), and atezolizumab with carboplatin, paclitaxel, and bevacizumab (ABCP) in the three-arm phase 3 trial IMpower150 (Socinski et al. 2018). In this study, treatment with ABCP improved the PFS and OS compared with BCP (bevacizumab + carboplatin + paclitaxel), but ACP (atezolizumab + carboplatin + paclitaxel) did not produce a significant survival benefit compared with BCP. In patients with squamous histology, the addition of atezolizumab to CT did not improve OS in the IMpower131 trial, with the exception of patients with high tumor PD-L1 expression (Jotte et al. 2019). The phase 3 CheckMate 9LA trial testing the addition of nivolumab and ipilimumab to two cycles of CT compared to up to four cycles of CT alone in the first-line setting in patients with metastatic NSCLC showed that OS, PFS, and ORR were improved with combination chemoimmunotherapy compared with CT alone, irrespective of PD-L1 expression (Reck et al. 2020). At the 2021 American Society for Clinical Oncology (ASCO) Annual Meeting, updated data demonstrated continued benefit from the combination strategy: at a minimum follow-up of 2 yr, the median OS was 15.8 mo with chemoimmunotherapy versus 11.0 mo with CT alone and the 2-yr OS rates were 38% versus 26%, respectively (Reck et al. 2021). These studies provide evidence that the chemoimmunotherapy combination is superior to CT alone as the first-line treatment for patients with advanced NSCLC, regardless of tumor PD-L1 expression status. However, at this time, there is no clear evidence that the efficacy of CT plus ICI is superior to that of ICIs alone in patients with high tumor PD-L1 expression (Remon et al. 2020).

Second-Line Therapy

The phase 3 CheckMate 017 trial compared the safety and efficacy of nivolumab with docetaxel in patients with advanced squamous NSCLC who had progressed on first-line CT (Brahmer et al. 2015). The median OS was 9.2 and 6.0 mo for nivolumab and docetaxel, respectively. Similar results were obtained in the CheckMate 057 trial comparing second-line nivolumab with docetaxel in patients with nonsquamous NSCLC (Borghaei et al. 2015), with a median OS of 12.2 and 9.4 mo with nivolumab and docetaxel, respectively. Consequently, nivolumab was approved by the FDA for treatment of patients with metastatic squamous and nonsquamous NSCLC with progression on or after failure with CT.

In the phase 2/3 KEYNOTE-10 study comparing 2 mg/kg pembrolizumab, 10 mg/kg pembrolizumab, and docetaxel in patients with previously treated, PD-L1-positive, advanced NSCLC, the median OS was 10.4 mo, 12.7 mo, and 8.5 mo, respectively (Herbst et al. 2016). Among patients with tumor PD-L1 ≥ 50%, OS was significantly longer with pembrolizumab 2 mg/kg and 10 mg/kg versus docetaxel. Updated results of KEYNOTE-010 demonstrated that pembrolizumab treatment improved OS over docetaxel in the PD-L1 expression ≥ 50% and ≥ 1% groups after a median follow up time of 42.6 mo (Herbst et al. 2020). In the phase 2 POPLAR trial (Fehrenbacher et al. 2016), OS was 12.6 mo for atezolizumab versus 9.7 mo for docetaxel in patients with NSCLC who progressed on CT. Improvement correlated with PD-L1 expression on tumor and tumor-infiltrating immune cells, suggesting that PD-L1 expression is predictive for atezolizumab benefit. Similar results were reported in the OAK randomized phase 3 study (Rittmeyer et al. 2017).

ICI Therapy in Patients with Actionable Tumor Molecular Aberrations

Recent evidence suggests that the clinical activity and efficacy of ICIs are limited in patients with NSCLC harboring aberrations in EGFR and ALK. In a retrospective study evaluating the activity of PD-1 and PD-L1 inhibitors in 58 patients with NSCLC, responses were noted in 3.6% (1/28) of tumors harboring EGFR mutations or ALK rearrangements compared with 23.3% (7/30) of tumors characterized as EGFR wild-type and ALK-negative/unknown (Gainor et al. 2016). In a meta-analysis of more than 3000 patients with advanced NSCLC treated with a PD-1 or PD-L1 inhibitor or docetaxel, ICI therapy prolonged OS versus docetaxel in the EGFR wild-type subgroup, but not in the EGFR-mutant subgroup (Lee et al. 2018).

Recent studies have investigated the use of ICIs in combination with additional therapies for patients whose tumors harbor EGFR and ALK genomic alterations. The phase 3 IMpower150 study allowed enrollment of patients with EGFR and ALK alterations and demonstrated that ABCP treatment prolonged PFS in the overall population and in this subgroup of patients. Ongoing phase 3 trials are testing the clinical efficacy of ICI-CT combination in patients with EGFR-mutant NSCLC after progression on EGFR tyrosine kinase inhibitors (TKIs) (NCT03515837). Initial reports suggest that the combination of ICIs and TKIs induces higher frequency of toxicities compared to single agents without increased clinical activity (Antonia et al. 2014; Ma et al. 2016; Liang et al. 2018; Creelan et al. 2021). A retrospective analysis of patients harboring EGFR-mutant NSCLC demonstrated that sequential treatment with ICI therapy followed by osimertinib treatment was associated with onset of severe immune-related toxicities, particularly in patients who received osimertinib within 3 mo of last ICI dose (Schoenfeld et al. 2019). These results show the limited applicability of ICIs in patients with NSCLC harboring EGFR and ALK alterations and emphasize the importance of carefully considering the administration of these agents to these subgroups of patients.

CLINICAL TRIALS OF ICIs IN LOCALLY ADVANCED/UNRESECTABLE NSCLC

Tumor RT promotes immune effects by inducing proinflammatory signals that facilitate the recruitment of immune cells to the tumor microenvironment and by increasing neoantigen exposure and availability (Barker et al. 2015). Consequently, there is considerable interest in examining therapeutic effects of combined immunotherapy and RT. The randomized phase 3 PACIFIC study compared the anti-PD-L1 antibody durvalumab as consolidation therapy with placebo in patients with stage III NSCLC who did not have disease progression after two or more cycles of platinum-based CRT. Consolidation therapy with durvalumab produced significant improvements compared with placebo in both PFS (17.3 vs. 5.6 mo) and OS (not reached vs. 28.7 mo) (Antonia et al. 2018). These results established the PACIFIC regimen as the standard of care in this population. Subgroup analyses indicated that PFS and OS were superior if durvalumab was delivered fewer than 14 d after the completion of RT-CT. The 3-yr OS rates were 57% with durvalumab versus 43.5% with placebo (Gray et al. 2020). Updated results from the PACIFIC study were presented at the ASCO Annual Meeting 2021 (Spigel et al. 2021) and demonstrated that, at a median follow-up duration of 34.2 mo in all patients, OS (median 47.5 vs. 29.1 mo) and PFS (median 16.9 vs. 5.6 mo) remained consistent with the results from the primary analyses. The 60-mo OS rates were 42.9% and 33.4% with durvalumab and placebo, respectively, and 60-mo PFS rates were 33.1% and 19.0%, respectively (Spigel et al. 2021).

Several studies are investigating concurrent administration of ICIs and RT for unresectable stage III NSCLC. The ETOP NICOLAS trial is evaluating the safety and efficacy of nivolumab concurrent with CT-RT in 82 patients. In the first 21 patients, no grade 3 or higher pneumonitis was observed by the end of the 3-mo post-RT follow-up period and no unexpected or increased toxicities were reported (Peters et al. 2019). The DETERRED trial investigated concurrent atezolizumab with CT-RT followed by CT-atezolizumab and maintenance atezolizumab and reported that concurrent atezolizumab and CT-RT is safe without increased toxicities (Lin et al. 2019). The KEYNOTE-799 phase 2 nonrandomized 2-cohort study investigated pembrolizumab plus concurrent CRT (cCRT) in 112 NSCLC patients with locally advanced and untreated NSCLC and demonstrated promising antitumor activity of pembrolizumab plus cCRT and an overall manageable safety and toxicity profile (Jabbour et al. 2021).

In addition to promoting antitumor activity at the primary tumor site, RT can also induce systemic effects outside of the irradiation field (i.e., abscopal effects) (Demaria and Formenti 2020). Whereas the precise biological mechanisms underlying an abscopal response remain unclear, the effect has been connected to mechanisms involving the immune system and there is growing interest in combining immunotherapy with RT as a therapeutic strategy to enhance abscopal response rates (Ngwa et al. 2018). There is some evidence to support this supposition. A recent study found that combined RT and CTLA-4 blockade induced systemic antitumor T cells in CT-refractory NSCLC, whereas anti-CTLA-4 antibodies had failed to demonstrate efficacy alone or in combination with CT (Formenti et al. 2018). The investigators reported an ORR of 18% with two complete and five partial responses. These findings demonstrate that RT can synergize with ICIs to produce control of both primary NSCLC and distal metastases.

CLINICAL TRIALS OF ICIs FOR EARLY-STAGE NSCLC

Adjuvant Immunotherapy Trials

Surgery is the preferred treatment approach for patients with operable early-stage NSCLC (Howington et al. 2013). However, more than 50% of patients experience tumor recurrence after surgery alone (Martin et al. 2002). The use of perioperative (neoadjuvant or adjuvant) CT as a means to prevent disease recurrence after surgery provides only a modest (∼5%) improvement in 5-yr OS and is also a source of considerable toxicity (Pignon et al. 2008; NSCLC Meta-Analysis Collaborative Group 2014). ICIs are being evaluated in the adjuvant setting with the goal to eradicate micrometastatic disease and reduce the risk of relapse, based on the rationale that the postsurgical stress response is characterized by physiological cellular mediators that may impair innate and acquired immune responses by generating an immunosuppressive state that favors tumor progression (Bakos et al. 2018). Several phase 3 trials are investigating adjuvant ICIs for resectable stage IB–IIIA NSCLC. The randomized phase 3 open-label IMpower010 trial evaluated adjuvant atezolizumab compared to best supportive care (BSC) after adjuvant platinum-based CT in patients with early-stage resected NSCLC (Wakelee et al. 2021). The primary end point of disease-free survival (DFS) was analyzed in three subgroups: patients with stage II–IIIA tumors and PD-L1 ≥ 1%; all randomly assigned patients with stage II–IIIA NSCLC; and intention-to-treat (ITT) population with stage IB–IIIA NSCLC. In patients with stage II–IIIA resected NSCLC and PD-L1 tumor composite score ≥ 1%, adjuvant atezolizumab reduced the risk of disease progression or death by 34% versus BSC. Median DFS was not reached in the atezolizumab-treated group versus 35.3 mo for patients who received BSC and no new safety concerns were noted. These results are promising to implement an ICI agent in the adjuvant setting for some patients with resected NSCLC; however, several challenges remain in this disease setting, including the lack of established predictive biomarkers of therapeutic benefit, the unclear ideal length of adjuvant treatment and the role of DFS used as a primary end point as compared to the use of OS.

Neoadjuvant Immunotherapy

The results from the clinical trials of ICIs in patients with metastatic NSCLC provided the impetus for testing checkpoint inhibitors in resectable, early-stage NSCLC. Administering ICIs in the neoadjuvant setting provides an opportunity to prime the antitumor immune response and eradicate micrometastases at an early time point when tumor-specific antigens may possess less heterogeneity and induce greater antitumor control. Neoadjuvant treatment also allows rapid assessment of tumor sensitivity or resistance to therapy, particularly when there are evaluable surrogate end points at the time of surgery that can correlate with hard end points of clinical efficacy. Clinical trials using OS as a primary end point require a decade or longer to complete (Hellmann et al. 2014). Major pathologic response (MPR), defined as ≤10% viable tumor cells in resected tumor specimens, provides an objective criterion of response to neoadjuvant CT and correlates with long-term outcomes (Pataer et al. 2012). MPR has been incorporated as a candidate surrogate of survival benefit in neoadjuvant trials testing new therapeutic strategies for patients with early-stage NSCLC.

In the first feasibility study, neoadjuvant nivolumab had a tolerable toxicity profile, did not delay surgery, and resulted in an MPR rate of 45% (Forde et al. 2018). TMB, but not pretreatment PD-L1 expression, was associated with MPR. In the Lung Cancer Mutation Consortium 3 (LCMC3) study, neoadjuvant atezolizumab produced a 21% MPR rate, tumor PD-L1 ≥ 50% expression was associated with MPR to therapy, and a weak correlation was noted between MPR and TMB with a trend toward less pathological regression in STK11-mutated tumors (Carbone et al. 2021). In the phase 2 randomized NEOSTAR trial, which tested neoadjuvant nivolumab as monotherapy or in combination with ipilimumab in 44 patients with resectable NSCLC (Cascone et al. 2021), nivolumab produced a 22% MPR rate, whereas the combination therapy produced a 38% MPR rate in the ITT population (Cascone et al. 2021). Tumor PD-L1 expression at baseline was associated with radiographic responses and greater tumor regression; however, responses were also observed in some patients lacking tumor PD-L1 expression pretherapy (Cascone et al. 2021). In a phase 1b study, the neoadjuvant PD-1 inhibitor sintilimab produced a 40.5% MPR rate in 37 Chinese patients with resected NSCLC (Gao et al. 2020). The NADIM phase 2 trial tested neoadjuvant nivolumab plus CT followed by surgical resection and adjuvant nivolumab treatment for 1 yr in 46 patients with resectable stage IIIA N2 NSCLC. The MPR response was 83% (34/41) in the resected patients or 73% (34/46) in the ITT patients and 24 mo OS was 89.9% (Provencio et al. 2020). Neoadjuvant atezolizumab plus CT induced MPR rates of 57% in treated patients with resectable NSCLC, and no significant associations between MPR and PD-L1 expression were observed (Shu et al. 2020). The initial results of the phase 3 randomized study CheckMate 816 investigating neoadjuvant platinum doublet CT plus nivolumab versus CT in patients with operable NSCLC without EGFR or ALK genomic aberrations and using pathological complete response (pCR) and event-free survival (EFS) as primary end points were recently reported (Forde et al. 2021). The pCR rate was 24% with chemoimmunotherapy as compared to 2.2% with CT alone in the ITT population. The MPR rates were 36.9% with combination therapy versus 8.9% with CT alone. This study is the first phase 3 trial in the neoadjuvant immunotherapy setting to show the benefit of an ICI plus platinum-doublet CT over CT alone. Several ongoing phase 3 trials are evaluating perioperative combinations of ICIs with CT, including IMpower030 (atezolizumab), KEYNOTE-671 (pembrolizumab), AEGEAN (durvalumab), and CheckMate 77T (nivolumab) trials. Much investigative effort is directed toward the identification of several emerging tissue-, blood-, and host-based biomarkers that could be used to maximize the clinical benefit of ICIs (Fig. 3). In the next section, we discuss the evidence supporting the role for some of these biomarkers, including tumor PD-L1, TMB, circulating tumor DNA (ctDNA), and gut microbiome.

Figure 3.

Emerging tissue-, blood-, and host-based biomarkers of potential clinical benefit or therapeutic resistance to immune checkpoint inhibitors (ICIs) under evaluation in NSCLC. (PD-L1) Programmed death-ligand 1, (TILs) tumor-infiltrating lymphocytes, (TLS) tertiary lymphoid structures, (TMB) tumor mutation burden, (TCR) T-cell receptor, (ITH) intratumoral heterogeneity, (NKs) natural killer cells, (PD-1) programmed cell death protein 1, (NLR) neutrophil-to-lymphocyte ratio, (M:L) myeloid-to-lymphoid ratio, (PLR) platelet-to-lymphocyte ratio, (ctDNA) circulating tumor DNA, (bTMB) blood tumor mutation burden.

PD-L1 EXPRESSION AS A BIOMARKER FOR ICIs IN NSCLC

PD-L1 expression has been evaluated as a prospective selection marker for ICI therapy in both first-line and second-line settings in patients with NSCLC. PD-L1 expression is determined by standard immunohistochemical staining of tissue sections that are then read by pathologists as the tumor proportion score (TPS), which represents the best estimated percentage (0%–100%) of tumor cells showing partial or complete membranous PD-L1 staining (Lantuejoul et al. 2020) and percentage of immune cells with similar expression (immune cell proportion score [ICPS]). Despite the fact that only pembrolizumab and nivolumab plus ipilimumab requires a PD-L1 IHC assay as a companion diagnostic to determine patient eligibility for therapy, PD-L1 IHC has also been established as a complementary diagnostic for other PD-L1 inhibitors to determine patient eligibility (Büttner et al. 2017). Recently, two large studies demonstrated that three of the five IHC assays for assessing PD-L1 expression were closely aligned on tumor cell staining (22C3, 28-8, and SP263 assays), whereas the SP142 assay exhibited fewer stained tumor cells overall, and the 73–10 assay had higher sensitivity (Rimm et al. 2017; Tsao et al. 2018). These studies also showed that there is reliability among pathologists in tumor cell PD-L1 scoring with all assays, but poor reliability in immune cell PD-L1 scoring (Tsao et al. 2018).

The results of the clinical studies thus far indicate that there is considerable variability in PD-L1 as a biomarker of response to ICIs. While patients whose tumors express elevated PD-L1 levels may exhibit improved responses to treatment, patients whose tumors lack PD-L1 expression may also receive benefit. PD-L1 expression remains an imperfect marker of response for several reasons, including tumor heterogeneity, the fact that it is assessed as a continuous variable, and different conditions and therapies may regulate its expression (Lantuejoul et al. 2020). Therefore, the use of PD-L1 IHC expression, at this time, should be dedicated to select patients with advanced NSCLC eligible to treatment with ICIs.

TMB AS A BIOMARKER FOR ICIs IN NSCLC

TMB refers to the number of somatic mutations within the coding region of a tumor. It is reasonable to assume that a tumor with a high TMB would be more likely to express a greater number of neoantigens as compared to a tumor with a low TMB and thus be more responsive to ICIs. The relationship between TMB and clinical efficacy of ICIs has been investigated. The CheckMate 026 study demonstrated that patients with PD-L1 positive tumors and high TMB treated with first-line nivolumab had higher response rates (47% vs. 28%) and improved PFS (9.7 vs. 5.8 mo) versus CT (Carbone et al. 2017). However, there were no differences in OS between the nivolumab and CT group. In the CheckMate 227 phase 3 multipart trial, PFS was found significantly longer with first-line nivolumab plus ipilimumab therapy than with chemotherapy among patients with metastatic NSCLC and with a high TMB (≥10 mut/mb), irrespective of PD-L1 expression level (Hellmann et al. 2018). However, in a subsequent report, a similar OS benefit was noted in NSCLC patients treated with nivolumab plus ipilimumab, regardless of high or low TMB (≥10 vs. <10 mut/mb, respectively) (Hellmann et al. 2019). In a retrospective analysis of two large randomized trials of previously treated, metastatic NSCLC patients receiving ICIs, and in a separate nontrial cohort, blood TMB (bTMB) identified patients with improved response rates and PFS from ICI therapy (Gandara et al. 2018; Wang et al. 2019b). In the phase 3 MYSTIC trial (Rizvi et al. 2020), exploratory analyses showed that a bTMB threshold of ≥20 mut/mb was associated with improved OS in patients treated with durvalumab plus tremelimumab versus CT. Nevertheless, the role of TMB as a predictive biomarker remains unclear in the early-stage and metastatic disease setting and challenging to implement in practice due to lack of a standard cutoff for high TMB, variabilities in the assays used to determine TMB, and absence of prospective validation for survival benefit.

ctDNA AS A BIOMARKER FOR ICIs IN NSCLC

Results from recent studies suggest that analyzing the fragments of tumor cell DNA from tumor cells that are released in the circulation may represent a valuable approach for monitoring disease burden in NSCLC patients who have undergone surgery or been treated with ICIs. DNA fragments have been shown to gain access to the circulation and other bodily fluid compartments through processes of apoptosis, necrosis, or active secretion (Heitzer et al. 2020). A recent report in patients with stage I–III NSCLC has shown that ctDNA can serve as a biomarker to detect molecular residual disease (MRD) following surgery and for defining the clonality of relapsing disease (Abbosh et al. 2020). The investigators reported that of the 45 patients that had recurrence of their primary NSCLC, ctDNA was detected at or before clinical relapse in 37 of these patients (Abbosh et al. 2020). In a study of 40 patients with stage I–III lung cancer who were treated with curative-intent therapies, ctDNA was detected in the first blood sample collected posttherapy in 94% of patients who experienced disease recurrence (Chaudhuri et al. 2017). ctDNA analysis also demonstrated the potential to inform consolidation ICI therapy after CRT in patients with locally advanced, unresectable NSCLC (Moding et al. 2020). Patients with MRD after CRT who received consolidation ICI experienced improved outcomes compared to those who did not receive consolidation ICI, suggesting the utility of ctDNA for the personalization of ICI therapy while avoiding unnecessary toxicities (Moding et al. 2020). In an analysis of 87 patients enrolled in the CheckMate 816 trial, ctDNA was more likely to clear when nivolumab was administered in combination with neoadjuvant CT versus CT alone (56% vs. 34%, respectively), and pCR was observed in 46% of patients with ctDNA clearance as compared to 13% of those without (Forde et al. 2021).

GUT MICROBIOME AS A BIOMARKER FOR ICI THERAPY IN NSCLC

Recent evidence suggests that the composition of bacteria residing in the gut may play a key role in determining tumor responses to ICIs and that microbial-derived metabolites may be a central regulator of systemic immune function (Rooks and Garrett 2016). Melanoma patients who responded to anti-PD-1 therapy could be distinguished from nonresponders by the diversity and composition of their gut microbiome (Gopalakrishnan et al. 2018). This effect was mediated by elevated antigen presentation and improved effector T-cell function (Gopalakrishnan et al. 2018). In patients with advanced NSCLC treated with ICIs, concomitant antibiotic administration altered gut microbiome and led to reduced tumor response to ICIs when compared to patients without antibiotic treatment (Routy et al. 2018). In a study involving a cohort of Chinese patients with advanced NSCLC treated with anti-PD-1 therapy, 16S rRNA gene sequencing revealed higher diversity of gut microbiome at baseline followed by stable microbiome population during treatment in responders to ICIs (Jin et al. 2019). In the phase 2 randomized NEOSTAR trial, we found that administration of neoadjuvant ICIs had no significant impact on the diversity, structure, and composition of microbiomes as assessed by 16S sequencing (Cascone et al. 2021). Our exploratory analyses also demonstrated that a relative higher abundance of Ruminococcus and Akkermansia spp. in pretherapy samples was associated with MPR to neoadjuvant nivolumab plus ipilimumab therapy (Cascone et al. 2021). Studies in melanoma also reported that higher diversity and relative abundance of fecal Ruminococcaceae are associated with response to ICIs (Gopalakrishnan et al. 2018), and metagenomics of baseline stool samples in patients with advanced NSCLC and kidney cancer that were refractory to ICIs revealed low levels of Akkermansia muciniphila (Routy et al. 2018). The ability of distinct commensal bacteria to impact tumor immunogenicity and increase the efficacy of ICIs provides an extraordinary opportunity to improve clinical outcomes for patients with resectable NSCLC. However, while encouraging, these results were exploratory in nature and will require validation in larger cohorts.

CONCLUSIONS

ICIs have rapidly changed the treatment landscape for patients with locally advanced/unresectable and metastatic NSCLC. New ICI combinations or regimens with ICI therapy combined with CT or RT are being explored in metastatic disease and the results from those studies are informing new trials for patients with early-stage disease in the perioperative setting. Additional understanding of the immune biology of lung tumors is critical for defining the role of the immune system in controlling tumor formation, progression, and metastasis. Technological advances are providing a deeper understanding of the immune biology of NSCLC that will likely be translated into new therapeutic strategies to improve clinical outcomes. These include tumor molecular profiling (sequencing, expression, epigenome, proteomic, and metabolomic); multiparameter IHC profiling and single-cell RNA sequencing analyses of the tumor and immune cells; identification of tumor neoantigens and individual patient T- and B-cell repertories before and after ICI therapy; and, finally, correlation of these findings with clinicopathological characteristics and therapeutic outcome. This understanding will also yield biomarkers that optimize patient selection and identify those patients most likely to benefit from treatment, thereby maximizing the clinical effectiveness of ICIs.

COMPETING INTEREST STATEMENT

T.C. reports speaker's fees from the Society for the Immunotherapy of Cancer, Bristol Myers Squibb and Roche; advisory/consulting fees from MedImmune, AstraZeneca, Bristol Myers Squibb, EMD Serono, Arrowhead Pharmaceuticals, Genentech and Merck & Co.; clinical research funding to The University of Texas MD Anderson Cancer Center from Boehringer Ingelheim, MedImmune, AstraZeneca, Bristol Myers Squibb and EMD Serono. D.L.G. reports honoraria for scientific advisory boards from AstraZeneca, Sanofi, Alethia Biotherapeutics and Lilly, research support from Janssen, Takeda, Ribon Therapeutics, Astellas, NGM Biopharmaceuticals and AstraZeneca.