Targeted Therapies, Novel Antibodies, and Immunotherapies in Advanced Non–Small Cell Lung Cancer: Clinical Evidence and Drug Approval Patterns

Non–small cell lung cancer (NSCLC) is one of the leading causes for cancer-related mortality worldwide, with an estimated annual global death toll of more than 1.5 million (1). For advanced NSCLC, systemic treatment is the standard therapeutic modality (2, 3). For approximately 30 years, platinum-based chemotherapy constituted the backbone of anticancer therapy (4), yet response rates have plateaued at around 30%, with 5-year overall survival (OS) rates not exceeding 5% (2, 5). The discovery of oncogene driver alterations amenable to targeted therapy (TT), along with the recognition of the immunogenic susceptibility of NSCLC to immune checkpoint inhibitors (ICI), has widened the armamentarium available to clinicians for intervention. The pharmacogenetic identification of mutations affecting intracellular signaling pathways, such as the EGFR, and the Nobel Prize–awarded identification of the programmed cell death 1 (PD-1) / programmed death ligand 1 (PD-L1) axis, as a promising target for immunotherapy, have been breakthroughs in therapeutic innovation (6). The suppression of oncogene signaling by TTs and the activation of the humoral and intercellular immune responses by ICI have changed the treatment landscape of NSCLC (7–11), as evidenced by the remarkable number of approvals for both new drug classes, not only in NSCLC but also in other cancers (12). Considering these underlying molecular mechanisms, clinical practice guidelines today recommend that all new patients with NSCLC should undergo molecular profiling using suitable validated diagnostic testing methods to determine the most adequate therapy (7–11).

In this review, we describe the pivotal clinical evidence from a regulatory science perspective that was provided for the first US approval action of novel therapeutics in NSCLC, including small-molecule inhibitors, mAbs, and antibody–drug conjugates (ADC). The review also outlines and discusses differences in endpoint patterns among their clinical development settings and highlights differences between the regulatory reviews by the US FDA and the European Medicines Agency (EMA) of new NSCLC drugs. Furthermore, this analysis examines the fulfillment of postmarketing requirements (PMR) for “accelerated approval” (AA) in the United States.

Since 2011, the US FDA has approved 30 new drugs for their first use in NSCLC (Fig. 1), many of which were new molecular entities or new biological entities not previously approved for any other indication. For nivolumab, atezolizumab, and tremelimumab, which were initially approved for other tumor types, the first NSCLC approval was granted based on independently filed biologics license applications (BLA), each following an initial approval for a different cancer indication. For seven drugs, either a supplemental BLA or a supplemental new drug application (NDA) was filed to the FDA, thus constituting an extension-of-indication (for previously approved BLA or NDA (in italics in Fig. 1A). Except for durvalumab, all drugs were initially approved for use in stage IIIB and/or IV NSCLC. Supplementary Table S1 displays the differing mechanisms of action of these 30 drugs, which can be categorized—on the basis of their established pharmacologic class (13)—into TT (i.e., small-molecule kinase and RAS GTPase inhibitors), ICI, and mAbs (including ADC) acting on cell surface receptors or acting intracellularly (Fig. 1B; Supplementary Table S1).

The year 2011 is the starting point for this review, marking the approval of the first-ever ICI in a cancer indication with the authorization of the anti-CTLA4 antibody ipilimumab in the spring of 2011 in both the United States and European Union for the treatment of advanced melanoma. In 2011, novel small-molecule inhibitors targeting driver alterations also began to enter clinical practice, with vemurafenib for BRAF^V600E-mutated advanced melanoma and crizotinib for anaplastic lymphoma kinase (ALK)–rearranged advanced NSCLC (6). Not in the focus of this review is the secondary extension-of-indication within NSCLC, although the subsequent label extensions of each drug in NSCLC by supplemental BLAs/NDAs are displayed in Panel B in Fig. 1.

Current American Society of Clinical Oncology and National Comprehensive Cancer Network guidelines for advanced NSCLC, similar to the European Society of Medical Oncology (ESMO) guidelines, recommend conducting up-front biomarker testing consistent with the histologic diagnosis. Following such up-front testing, patients are categorized as (i) those with NSCLC carrying driver alterations (7, 9, 10) and (ii) those for whom biomarker testing did not indicate the presence of driver alterations or testing was not performed or feasible (8, 9, 11). This differentiation is mirrored in current drug development and approvals. Both paths in clinical development are associated with a distinct pattern of primary endpoints based on the characteristics of the pivotal regulatory clinical trials in NSCLC (Fig. 2).

For drugs targeting a driver alteration, the overall response rate (ORR) has served as the primary endpoint for AA, supported by the duration of response. This applies regardless of whether the first approval in NSCLC was granted for first-line use (without or after prior chemotherapy) or subsequent to a previous TT. Of note, the FDA, like the EMA, widely accepted single-arm trials (SAT) or multicohort trials (MCT) as pivotal evidence for eligibility for expedited regulatory approval (14, 15).

The acceptance of ORR as a surrogate endpoint requires a substantial magnitude of effect that is associated with a likely clinical survival benefit in oncogene-addicted NSCLC. For advanced NSCLC, the acceptability of SAT designs as pivotal evidence by the FDA and EMA relies on the experience with EGFR-mutated NSCLC. The phase III LUX-Lung 3 trial of afatinib, the phase III EURTAC trial of erlotinib, and the phase III IPASS trial investigating gefitinib in EGFR-mutated NSCLC all demonstrated statistical superiority and remarkable gains in progression-free survival (PFS) for EGFR tyrosine kinase inhibitors (TKI) compared with standard platinum-based chemotherapy (16–18). In these trials, HRs for the median PFS, serving as the primary endpoint, ranged from 0.58 [95% confidence interval (CI), 0.43–0.78; P = 0.001] in the LUX-Lung 3 to 0.48 (95% CI, 0.36–0.64; P < 0.001) in IPASS and 0.37 (95% CI, 0.25–0.54; P < 0.0001) in EURTAC, representing unprecedented clinical benefit in chemotherapy-naïve advanced NSCLC. These proof-of-principle trials for TTs in EGFR-mutated NSCLC were further backed by the phase III ARCHER 1050 trial of dacomitinib, which compared this irreversible second-generation EGFR TKI with gefitinib monotherapy (19). Compared with gefitinib, dacomitinib demonstrated a gain of 5.5 months in PFS (HR, 0.59; 95% CI, 0.47–0.74; P < 0.0001). ORRs of these open-label randomized trials, ranging from 56% (afatinib, LUX-Lung 3), 58% (erlotinib, EURTAC), and 71% (gefitinib, IPASS) up to 75% for dacomitinib and 72% for gefitinib in the ARCHER 1050 trial, set response benchmarks for other TTs in advanced NSCLC. Due to the clinical setting (first-line) and the open-label design, it was, however, difficult to discern OS advantages because of patient cross-over and multiple subsequent therapies.

The third-generation EGFR TKI osimertinib was specifically developed to overcome resistance to first- and second-generation EGFR TKIs, caused by the EGFR^T790M resistance mutation (20). With an ORR of 59% (95% CI, 54–64), osimertinib became the first FDA-approved TKI to re-treat oncogene-addicted NSCLC, by targeting a mechanism of resistance to prior EGFR TKIs (Fig. 2).

The FDA granted “fast-track” status, “breakthrough therapy designation,” “priority review” (PR), and “AA” (Fig. 1). In Europe, osimertinib was the first-ever cancer drug approved through a “conditional marketing authorisation” (CMA) in line with a shortened EMA review period (“accelerated assessment”). The clinical evidence package for osimertinib included pooled data from two SATs, the first-in-human AURA trial and the phase II AURA2 trial, enrolling a total of 411 patients (21, 22).

The clinical benefit of EGFR TKIs led to an increase in approvals on the basis of SATs—via the AA pathway—for new drugs targeting oncogene-addicted NSCLC. The use of SATs is underpinned by the expectation that drug development should yield a substantial effect size in small, biomarker-selected populations. In advanced NSCLC, AAs have been granted based on SATs demonstrate an ORR of 40% and higher, notably higher than the approximately 30% ORR observed with platinum-based chemotherapy as the historical control (Table 1; ref. 5).

An ORR below the 30% threshold, associated with platinum-based chemotherapy, was accepted only in the case of mobocertinib, a TKI that selectively targets EGFR exon 20 insertions (EGFRex20ins; ref. 38). EGFRex20ins are characterized by an altered active site that sterically hinders TKI binding, resulting in ORRs below 10% with previously approved EGFR TKIs (44). However, in October 2023, the marketing authorization holder announced the withdrawal of mobocertinib from the US market because the phase III, confirmatory EXCLAIM-2 trial (NCT04129502) did not meet its primary PFS endpoint and thus did not fulfill the confirmatory data requirements of the FDA for AA. In the EU, the company had already withdrawn its CMA filing in July 2022 because of concerns raised by the agency during review (54).

Amivantamab, a bispecific antibody targeting EGFR and MET, and trastuzumab deruxtecan (T-DXd), an ADC, also obtained approvals for oncogene-addicted NSCLC (Fig. 2; Table 1). Unlike small-molecule kinase inhibitors that target the intracellular tyrosine kinase domain, these drugs bind to the extracellular domain of their respective receptors. Amivantamab, specifically developed for use in patients with EGFRex20ins mutations, achieved a 40% ORR during clinical exploration in this patient population (44).

T-DXd is the first ADC approved in advanced NSCLC. Initially investigated in previously treated HER2-overexpressing (IHC 3+/IHC 2+) and HER2-mutated NSCLC in the phase II, single-arm DESTINY-Lung01 trial, it was the HER2-mutated cohort that demonstrated promising clinical activity with an ORR of 55% (95% CI, 44–65), compared with the HER2-overexpressing cohort, which showed an ORR of 24.5% (95% CI, 13–39; refs. 45, 55). Subsequent results from the phase II DESTINY-Lung02 trial, which compared two doses of T-DXd in previously treated HER2-mutated NSCLC, confirmed the DESTINY-Lung01 results, demonstrating an ORR of 58% (95% CI, 43–71; Table 1; ref. 56). Prior to AA in August 2023, FDA granted breakthrough therapy and PR to T-DXd (Fig. 1), which received conditional approval in the European Union in November 2022. In the United States, the approval was one of the first taking into account advice from the FDA Oncology Center of Excellence (OCE) “Project Optimus” (57)—an initiative aiming to reform the dose optimization and dose selection paradigm—for conducting dose randomization in the DESTINY-Lung02 trial, leading to the approval of a lower dose compared with initial phase II testing.

Since 2011, in contrast to TT which mainly received AA based on SATs, drugs not targeting driver alterations have provided proof of efficacy primarily generated through randomized controlled trials (RCT) in accordance with the statutory provisions of the FDA (14). OS and, for first-line approvals, OS plus PFS served as (coprimary) endpoints in the pivotal trials of novel ICI or tumor-associated antigen-targeting antibodies in advanced NSCLC (Fig. 1).

For the approval of the programmed cell death 1 inhibitor nivolumab in March 2015, for the treatment of advanced squamous NSCLC after platinum-based chemotherapy (Table 1), the clinical evidence package included data from the pivotal CheckMate 017 RCT (46), supported by data from 117 patients in the phase II CheckMate 063 SAT (58). In April 2014, faced with the modest ORR data (15%; 95% CI, 9–22) from the CheckMate 063 trial, the FDA determined that further data from the ongoing second-line phase III CheckMate 017 trial in advanced squamous NSCLC should be provided. By mid-January 2015, when convincing OS data became available (Table 1), the FDA committed to a PR of the subsequently filed application. This resulted in approval on March 4, 2015, effectively “saving 6 months by not waiting for formal preparation of data by the sponsor and 2.5 months by expediting review” (59).

The AA of pembrolizumab by the FDA on October 2, 2015, for patients with advanced, previously treated NSCLC, regardless of histology, represents an exception to the requirement for proof of efficacy through RCTs for drugs not targeting driver alterations. The clinical evidence package of pembrolizumab included data from an extension cohort of the KEYNOTE-001 MCT, which focused on both squamous and nonsquamous NSCLC (47), reporting an ORR of 41% (95% CI, 28.6–54.3) in 61 patients with PD-L1 tumor proportion score (TPS) ≥50%. These data were sufficient to fulfill the unmet medical need criterion of the FDA for AA, considering that the earlier approval of nivolumab was limited to squamous NSCLC (46).

Like for pembrolizumab, the testing of PD-L1 expression to optimize the benefit–risk profile has supported the regular approval of cemiplimab as monotherapy for the first-line treatment of patients with advanced EGFR/ALK wild-type NSCLC with high PD-L1 expression (TPS ≥50%; Figs. 1, 2). The FDA approval as first-line monotherapy for this biomarker-selected population, as studied in the EMPOWER-Lung 1 trial compared with platinum-based chemotherapy, was based on its meaningful prolongation of OS (22.1 vs. 14.3 months), resulting in a HR of 0.68 (95% CI, 0.53–0.87; P = 0.002; Table 1; ref. 52). The FDA noted that the OS effect size in the study closely matched data from the KEYNOTE-024 trial of pembrolizumab [HR, 0.60 (95% CI, 0.41–0.89); P = 0.005] and the IMpower110 trial of atezolizumab [HR, 0.59 (95% CI, 0.40–0.89); P = 0.01]. Consistently, these last two trials served to extend the label for ICI first-line monotherapy to PD-L1 high–expressing patients (TPS ≥50%; refs. 60, 61).

In advanced NSCLC, the identification and targeting of driver alterations, many of which occur rarely, are key to achieving a clinically convincing effect size (ORR and duration of response) as suitable pivotal evidence for AA, as outlined in this review. A good understanding of the mechanism of action (Supplementary Table S1) combined with reliable biomarker testing and careful dose selection may trigger high response rates in such selected populations. This, in turn, could increase the acceptance of SATs or MCTs for generating pivotal clinical evidence and facilitate early, expedited approval (62–64). In addition, this approach may significantly reduce development timelines and minimize patient exposure to experimental therapeutics prior to established effect. Such approach facilitates early patient access to innovative treatment options (Fig. 3A).

However, the frequency of oncogenic driver alterations is a key differentiator that determines the relative size of the patient population and directly impacts on trial type and feasibility. Frequent driver mutations in NSCLC like EGFR or KRAS^G12C represent 17% and 12%, respectively, of the overall NSCLC population, whereas ROS1 or NTRK 1/2/3 are found in ≤1% of patients with NSCLC (6). Incidence data from 2010 to 2017 for the United States indicate that in rare subpopulations like NTRK, ROS1, or RET, only a few 1,000 stage IV cases are diagnosed together annually (65), limiting the availability of patients for pre- or postmarketing clinical trials drastically.

For ICI development, PD-L1 positivity served for population stratification or hierarchical endpoint testing in numerous pivotal trials, both for first approvals and line extensions within NSCLC (66). The predictive value of the PD-L1 biomarker in NSCLC continues to be a subject of debate, and its true prognostic value in other solid tumors is still under investigation (67). Although the majority of PD-L1 expression–restricted ICI indications in NSCLC are similar across the United States and European Union, some decisions diverged, with the EMA opting for a more conservative approval approach (Fig. 3B). The labeling limitation of durvalumab in stage III NSCLC by the EMA has been subjected to criticism by clinical experts (68).

Since 2011, the FDA granted orphan drug (OD) designation to TTs in NSCLC (Fig. 1). However, neither the FDA nor the EMA granted OD status for the bispecific antibody amivantamab and the ADC T-DXd (Fig. 3C). Amivantamab likely did not get OD designation in the United States—in case it had been requested—because it does not exclusively target a single biomarker-defined population. In contrast to the United States, no OD designation incentives at all for NSCLC were granted in the European Union due to the current legislative framework (Fig. 3C). Under the EU regulatory framework, obtaining an orphan designation for a drug targeting a specific subpopulation within a broader indication is more challenging (69).

In Europe, the applicant must demonstrate a significant benefit over other available treatments, and a medicine can only be approved if the new therapy is not similar to other ODs approved to treat the same rare disease. In addition, the European Union requires additional evidence demonstrating that the drug is ineffective in the remainder of the population suffering from the same condition (70). As a result, OD approval patterns between the United States and European Union often differ (71).

Subpart H of the US Code of Federal Regulations Title 21 Part 314 stipulates the granting of AA based on a surrogate endpoint or based on an effect on a clinical endpoint other than survival or irreversible morbidity. Its use is limited to the treatment of serious or life-threatening illnesses and cases, in which a new drug provides meaningful therapeutic benefit to patients over existing treatments. For biologics, similar AA provisions are stipulated in Title 21 of the US Code of Federal Regulations Part 601 in Subpart E.

Since 2011, this regulatory pathway served to expedite the approval of most new TTs in NSCLC (Fig. 1). In line with the OD status incentives, the granting of AA supported the development of novel drugs in small subpopulations, representing in some cases only ≤1% of the entire NSCLC population. However, the granting of AA demands drug developers to provide postmarketing confirmatory clinical evidence to confirm clinical benefit (72, 73). The prevalence of the driver alteration impacts the nature of PMR.

Table 2 outlines the US PMRs associated with the fulfillment of the Subpart H requirements for the 21 new drugs approved for oncogene-addicted NSCLC; 15 thereof were granted AA (Figs. 1A, 3D). Eleven of these 15 AAs have completed PMRs and converted to full approval. The median time for AA conversion into regular approval was 27.1 months (range 16.2–36.6); in FDA’s cancer-wide analysis of AA, the median time to verify clinical benefit was 3.4 years (range 0.5–12.6 years; ref. 73). For the checkpoint inhibitor pembrolizumab, the initially granted AA in NSCLC (Fig. 1A) was converted to regular approval after only 0.7 months once final results from the KEYNOTE-010 trial became available (78).

In case of rare driver alterations such as RET or NTRK rearrangements and METex14 skipping alterations, for which clinical data for AA could be established through SATs, the FDA did not always omit the conduct of RCT as PMR. More recently, as response to delayed completion of some PMR due to the rarity of the population, FDA’s OCE aimed to stimulate alternative strategies for more timely generation of confirmatory evidence for AAs in oncology (72). Development strategies should consider interim analyses of the pivotal RCT as evidence base for AA with final data to be available for conversion to regular approval, or to confirmatory RCTs in an earlier setting (79). Confirmatory studies are required to be well underway at the time of AA (72).

In terms of the granting of EU CMA in NSCLC, the decision-making patterns of EMA for drugs targeting subpopulations with driver alterations mirror FDA’s approach. In the cases which alternate (Fig. 3D), more comprehensive data from additional independent cohorts or more advanced study data were available at the time of the decision-making in the European Union. Specific in-depth tutorials describe similarities and differences between the EU and US regulatory frameworks in more detail (80, 81).

Progress has been made to develop new drugs targeting molecularly defined subtypes of NSCLC available to patients as early as possible using the AA pathways. In line with current development paradigms to test novel cancer therapies in patients whose established treatment options are exhausted, clinical trials for novel drugs emerge in late lines with fewer patients available for evaluation. FDA’s “Project FrontRunner” is aiming to facilitate such development in earlier, nonrefractory therapy lines in the advanced or metastatic disease stage. The project intends to stimulate the conduct of RCTs with interim analyses as basis for AA, whereas confirmatory efficacy and safety evidence is being generated with final endpoint analyses (82, 83).

FDA’s new project has the potential to achieve two beneficial effects: on the one hand, in larger patient collectives, uncertainties about clinical meaningful outcomes such as OS and toxicity evaluation could be resolved more easily, allowing earlier, more robust benefit–risk assessments. On the other hand, the drug development path toward adjuvant and neoadjuvant use could be shortened. By the end of 2023, only 4 and 2 of the 30 novel NSCLC drugs discussed in this review had reached approval in the adjuvant and neoadjuvant settings, respectively (Fig. 1B). Earlier investigation of novel drugs in early-stage NSCLC would ease the clinical investigation of combination strategies including multimodal treatment approaches.

Since 2011, treatment options for patients with NSCLC have significantly improved with a clear distinction emerging between therapies for oncogene-addicted and non–oncogene-addicted NSCLC. Therapies targeting driver alterations in small subpopulations have changed the prognosis and the clinical practices. These therapies have demonstrated excellent clinical efficacy and have primarily been approved based on evidence obtained from SATs or MCTs, frequently supported by real-world evidence for contextualization. This approach contrasts with the rigorous RCTs conducted for therapies for non–oncogene-addicted NSCLC in larger populations. This distinction contributes to the evolving complexity of the assessment and regulatory decision-making in NSCLC.

Biomarker identification plays a crucial role in advancing NSCLC therapies. Additionally, the identification of new molecules that can prolong OS and delay tumor progression is imperative in the field. Oncogene-driven stratification divides NSCLC into small subpopulations, each with its specific outstanding effect sizes—through SATs, MCTs, or RCTs—to counterbalance the uncertainty of conducting trials in populations that are difficult to recruit. This shift toward personalized medicine introduces significant changes in the drug approval landscape.

To address the challenge of increasingly small subpopulations, particularly in later lines of therapy for advanced NSCLC, the FDA’s OCE is implementing Project “FrontRunner” which aims to stimulate initial clinical development of appropriate new drug candidates in earlier clinical settings, with potentially larger patient populations facilitating drug evaluation in RCTs against the standard of care. Similarly, both the FDA and EMA emphasize the importance of timely postmarketing evidence generation. This is crucial to rapidly confirm the longer-term efficacy and safety of new treatments and transition accelerated or conditional to regular approval. Although this review focuses on efficacy, it is acknowledged that evidence of treatment safety plays an equally important role in decision-making.

The dynamic field of drug development in NSCLC presents both challenges and opportunities for all stakeholders. Bridging the gaps among scientific innovation, regulatory frameworks, and clinical application will be crucial in enhancing patient outcomes and fostering more effective, personalized treatment strategies in NSCLC.

M.U. Koban reports personal fees from The Healthcare Business of Merck KGaA, Darmstadt, Germany, during the conduct of the study. M. Hartmann reports personal fees from Merck KGaA during the conduct of the study as well as personal fees from Bristol Myers Squibb, CatalYm, MorphoSys, Novartis, Pierre Fabre Pharma, Sanofi, and Western German Cancer Center outside the submitted work. G. Amexis reports personal fees from Merck KGaA during the conduct of the study. P. Franco reports personal fees and other support from Merck Serono Limited, Feltham, United Kingdom, during the conduct of the study. L. Huggins reports personal fees from Global Regulatory Affairs Oncology, The Healthcare Business of Merck KGaA, Darmstadt, Germany, during the conduct of the study. I. Shah reports personal fees from EMD Serono during the conduct of the study. N. Karachaliou reports employment with The Healthcare Business of Merck KGaA, Darmstadt, Germany.

The views expressed are those of the authors and should not be understood or quoted as being on behalf of organizations with which the authors are affiliated.

The authors thank Chris Bray (Global Regulatory Affairs Oncology, The Healthcare Business of Merck KGaA, Feltham, United Kingdom), Christopher Stroh (Clinical Biomarkers and Companion Diagnostics, The Healthcare Business of Merck KGaA, Darmstadt, Germany), and Stephanie Choi (EMD Serono Inc., Washington DC, United States) for helpful comments on biomarker terminology issues and the granting of breakthrough therapy designation. This study was sponsored by The Healthcare Business of Merck KGaA, Darmstadt, Germany (CrossRef Funder ID: 10.13039/100009945).