FormalPara Key Summary Points

More than half of newly diagnosed patients with lung cancer have advanced or metastatic disease, 10–26% present with brain metastases at the time of diagnosis, and another 30% will develop brain metastases over the course of their disease.

Current treatment options, particularly in later lines of therapy, are limited in their ability to pass through the blood–brain barrier, leaving a continuing treatment need in patients with non-small cell lung cancer (NSCLC) who have or develop brain metastases.

This study reviewed the current global landscape of clinical management used for patients with NSCLC, brain metastases, and actionable genomic alterations to gain a better understanding of treatment needs and how emerging therapies can fill those gaps.

For patients with NSCLC, brain metastases, and actionable genomic alterations, the current standard of care is suboptimal, and even with targeted therapies and local therapies (e.g., radiotherapies), prognosis is generally poor, regardless of the therapeutic regimen.

These findings emphasize the need for new therapies and therapeutic approaches that can improve clinical outcomes for this patient population.

Introduction

Lung cancer is the most common cause of cancer mortality worldwide, with an estimated 1.80 million deaths annually [1]. Non-small cell lung cancer (NSCLC) accounts for 81% of all lung cancers [2]. More than half of newly diagnosed patients with lung cancer have advanced or metastatic disease [3]; 10–26% present with brain metastases at diagnosis, and another 30% will develop brain metastases over the course of their disease [4,5,6]. Although various treatments are available, including surgery, stereotactic radiotherapy (SRT), stereotactic radiosurgery (SRS), whole-brain radiotherapy (WBRT), systemic therapy, and supportive care, survival rates for patients with NSCLC and brain metastases remain low, with overall poor patient outcomes and prognosis [7].

In the past, treatment strategies had been based primarily on the stage of disease or histologic appearance (squamous vs. non-squamous). Now, in addition to staging and histology, an improved understanding of tumor biology, along with overall advancements in treatment for NSCLC, has facilitated personalization of clinical management. Research has demonstrated that the presence or absence of actionable genomic alterations (AGAs), e.g., alterations in EGFR, ALK, ROS1, and other less common alterations, is a key determinant of optimal treatment selection, which aims to maximize responses and minimize toxicities [8]. To test for AGAs, advanced polymerase chain reaction (PCR)-based methods, such as quantitative PCR (qPCR) and reverse transcriptase PCR (RT-PCR), are used. Since brain metastases genetically diverge from the main tumor in NSCLC, evaluating for AGAs is important to determine therapy. Circulating tumor DNA (ctDNA) from cerebrospinal fluid seems to provide a better representation of brain metastases profiling compared to plasma ctDNA [9]. Tissue sampling of brain metastases poses a particular challenge, as many patients are not candidates for brain resections or have tumors in inaccessible sites. The low availability of tissue samples makes designing comprehensive studies problematic [10]. Still, AGAs that predict response to targeted treatment, including tyrosine kinase inhibitors (TKIs), are only present in approximately 30% of patients with NSCLC [8]. Even with recent advancements in earlier-line treatment, e.g., third-generation epidermal growth factor receptor (EGFR) TKIs, patients with both NSCLC and brain metastases have limited therapeutic options in later lines as many current treatments are unlikely to cross the blood–brain barrier (BBB) because of their molecular size [11].

In addition, patients often develop resistance to treatments, and therapeutic options may be associated with adverse events due to off-target drug activity [12]. For the development of new therapeutic options, such as fourth-generation EGFR TKIs and antibody–drug conjugates (ADCs), it will be important to understand their comparative activity in relation to the current treatments used for patients with NSCLC and brain metastases.

While the overall objective of the systematic literature review (SLR) was to understand the current landscape of clinical characteristics and clinical management for patients with brain metastases in NSCLC, this current review focused on summarizing the subset of studies with patients whose NSCLC harbored AGAs (e.g., EGFR, ALK, ROS1, NTRK, BRAF, MET, RET, KRAS, HER2).

Methods

An SLR was performed following standard methods outlined in the updated Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) 2020 and Cochrane guidelines [13, 14]. This article is based on previously published scientific studies and does not contain any new studies with human participants or animals performed by any of the authors.

Eligibility Criteria

The criteria are presented according to the PICOTS (Population, Intervention, Comparators, Outcomes, Timing, and Study design) format (Table 1). The inclusion and exclusion criteria are presented in Table 2.

Table 1 PICOTS eligibility criteria for overall systematic literature review
Table 2 Eligibility criteria amendments

During the course of the SLR, amendments were made (December 5, 2022) to the protocol to refine the eligibility criteria and focus on the most relevant and robust information available. All amendments were made prior to the data extraction phase and were applied universally across all records (Table 2).

Databases Searched

The search was conducted in the following databases using the OvidSP® platform:

  • EMBASE

  • MEDLINE® Epub Ahead of Print, In-Process and Other Non-Indexed Citations, Medline® Daily, Medline and Versions®

The search strategy was based on a combination of free-text words, indexing terms (e.g., Excerpta Medica database [EMBASE] subject heading [EMTREE] or Medical Subject Headings [MESH] terms) and their relationship using Boolean terms (e.g., and, or, not). Full strategies (including search dates) for all sources searched are included in Tables S1 and S2.

Screening Process

Publications identified through the systematic literature search were evaluated in a stepwise process to assess whether they should be included for data extraction.

Step 1—Title and abstract review: All unique records identified from the searches were reviewed on the basis of the predefined PICOTS criteria described in Table 1. Two reviewers independently screened titles and abstracts and classified each record as either (1) exclude or (2) continue to full-text review. Any discrepancy between reviewers was resolved by a third reviewer, who also confirmed the classifications for all studies marked for full-text review and from a sample of excluded abstracts. Furthermore, artificial intelligence technology was used to screen all excluded records and assign each a probability of likelihood for inclusion. Any study with a probability ranking over 85% was rescreened.

On the basis of the large number of potentially relevant studies identified after title and abstract screening (> 800), the eligibility criteria were amended to prioritize the most relevant and applicable evidence available that would address the research questions of interest. The amendments made following title and abstract screening are shown in Table 2.

Step 2—Full-text review: Full-texts of publications included after title and abstract review, and meeting the amended eligibility criteria, were screened by two reviewers on the basis of the amended PICOTS criteria. A third reviewer resolved any discrepancies.

Following full-text screening, additional amendments were made to better refine the project scope and identify the most relevant studies. The amendments made following full-text screening are shown in Table 2. Studies that met the amended PICOTS criteria after full-text review were included in the SLR.

Records that were excluded after review of the full-text report were documented, along with a clear justification for their exclusion. All references included after completion of the full-text review were retained for quality assessment and data extraction.

Data Extraction

Extraction of data from the included studies was conducted using a standardized Excel-based data extraction template. For each included study and methodological characteristics, selection criteria, study population/patient characteristics, and results were collected. If results for the same study were reported in more than one publication, the relevant records were grouped per study. Data extraction from included sources was conducted by two investigators independently, with discrepancies resolved by a third reviewer.

Quality Assessment

The quality assessment analyzes the strength and robustness of the available evidence with the aim of evaluating the applicability and internal and external validity of studies.

A quality assessment of individual papers was performed according to the study design. The Revised Cochrane Risk of Bias Tool for Randomized Trials (RoB 2) [15] was used for randomized controlled trials (RCTs), and the Newcastle–Ottawa Scale (NOS) [16] was used for non-randomized studies. The quality assessment results were recorded in a tabular format in the data extraction file. The quality assessment for studies included in this NSCLC, brain metastases, and AGA subgroup analysis are available in Tables S4 and S5.

Supplemental Search Results

Although not eligible for inclusion in the SLR, the most recent meetings of five conferences (Table S6) and two major clinical trial registries were searched to provide a current view of the evidence landscape (see supplementary materials).

Results

The search and screening process in the SLR is reported in accordance with the PRISMA flow diagram (Fig. 1).

Fig. 1
figure 1

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) flow diagram of literature search results. Figure shows the flow of the study identification and selection process. In total, 7884 records were originally identified. After removal of duplicates, 3815 records were screened. Several records were excluded throughout the process for reasons such as the population being out of scope, not having results because of an ongoing study, or the study design being out of scope. Ultimately, 179 studies were included in this review

Overall Systematic Search Output

The database and registry searches identified a total of 7884 records. Following deduplication, 3815 records underwent title and abstract screening, of which 901 records were classified as potentially relevant according to the original PICOTS criteria. After the PICOTS criteria were amended, 394 records were excluded and not sought for full-text review; the full texts of the remaining 507 records were retrieved and reviewed.

After full-text review and amendments to the PICOTS criteria, 432 records were excluded, and 75 records meeting the eligibility criteria were included. An additional 150 records were identified for inclusion, as were an additional 12 records identified from reviewing bibliographies of relevant review articles, based on a concurrent SLR on overall unmet needs in NSCLC with similar eligibility criteria. Overall, 237 publications reporting data on 179 studies were included in the review. The 179 studies included 33 RCTs, two single-arm trials, and 144 observational studies. Most RCTs (n = 17, 52%) were multiregional, whereas most observational studies (n = 75, 52%) were conducted in Asia.

Studies were further characterized as follows:

Studies of patients with AGAs: Of the 80 studies that included patients with AGAs, 19 were RCTs, two were single-arm trials, and 59 were observational studies. These studies are summarized in Tables 3 and S3 [17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98].

Table 3 Summary of studies for subgroup of patients with NSCLC, brain metastases, and actionable genomic alterations

Studies of patients without AGAs: Of the 18 studies of patients with no actionable mutations, nine were RCTs and nine were observational studies.

Studies of patients with or without AGAs: Of the 51 studies of patients with or without actionable mutations, five were RCTs and 46 were observational studies.

Studies of patients with not specified/unknown alterations: All 30 studies of patients with mutation status not specified or unknown were observational studies.

Results for Studies of Subgroup of Patients with NSCLC, Brain Metastases, and AGAs

Of the 19 RCTs, 10 were multiregional and nine were conducted in Asia (China, Japan, South Korea, Hong Kong, Malaysia, Taiwan, and Thailand). Only three trials enrolled fewer than 200 participants, 15 enrolled 200–500 participants, and one enrolled more than 500 participants (median 296 patients; range 119–556 patients). In the four RCTs reporting age and sex for the brain metastasis population, median age was 58.5 years (range 56–63), and median percentage of female patients was 60.5% (range 54–62%).

Of the two single-arm trials, one was multiregional and included 479 participants with EGFR mutations treated with targeted therapy (afatinib) for first-line or later-line therapy. The other study was conducted in multiple countries in Asia and included 541 participants with EGFR mutations treated with targeted therapy (afatinib) for first-line or later-line therapy. Of the 59 observational studies, 75% (n = 44) were conducted in Asia (China, Japan, South Korea, Taiwan, Thailand, and Turkey), eight in North America (USA and Canada), and seven in Europe (France, Germany, the Netherlands, and multi-country). Twenty-seven studies included fewer than 200 patients, 22 included between 200 and 500 patients, and 10 included more than 500 patients (median 208 patients; range 100 to 1081 patients). In the 34 observational studies that reporting age and sex for the brain metastasis population, median age was 58.5 years (range 54–68), and median percentage of female patients was 60.5% (range 37–73%).

Clinical Characteristics

Across the 80 studies that included patients with AGAs, 64 reported data for EGFR mutations (with the majority when reported being exon 19 deletions or exon 21 L858R), 14 reported data for patients with ALK alterations, and two reported data for both EGFR mutations and ALK alterations (Table 3). Still, across all 179 of the publications reviewed in the SLR, only a minority of patients with NSCLC and brain metastases (20–30%) had any actionable mutation. No study reported biomarkers specific to brain metastases.

For patients who had NSCLC and brain metastases, most RCTs included only those patients who were asymptomatic and/or neurologically stable at baseline. Similarly, on the basis of 23 observational studies, a range of 12.3–81.5% patients were reported to have asymptomatic brain metastases at baseline (Table 3). Among the three observational studies that reported symptoms, headache, nausea, and mental changes were the most frequently reported [30, 46, 64]. Patients who were asymptomatic were more likely to have been treated with EGFR TKIs only or EGFR TKIs plus SRS compared with patients who were treated with WBRT alone or in combination with another type of therapy.

Brain metastases were more often multisite than single site. The majority of brain metastases reported were located at the cerebral hemispheres and cerebellum. Few studies reported the median time interval between the diagnosis of NSCLC and brain metastases; among those studies that did, the average time to diagnosis was between 1 and 2 years (this average does not include patients who had brain metastases at NSCLC diagnosis) [45, 82, 88]. One Japanese retrospective study noted that the rate and frequency of developing brain metastases were rapid and higher in patients with EGFR mutations than in patients without EGFR mutations [99]. Similarly, in a Canadian cohort study, patients with EGFR mutations were reported to be at higher risk of developing brain metastases than patients without EGFR mutations [100].

Clinical Management

Overall, the clinical management reported by studies included in this SLR typically followed respective clinical practice guidelines. In brief, of the 80 total studies, 76 evaluated targeted therapy; 15 evaluated chemotherapy; six evaluated immune checkpoint inhibitors (ICIs)/monoclonal antibodies (mAbs), including bevacizumab, sintilimab, and denosumab; and 29 evaluated radiotherapy. Table 3 provides details on therapeutic regimens evaluated in each study.

Nine RCTs reported data for EGFR mutations only, eight of which evaluated EGFR TKIs. Six RCTs evaluated first-line targeted therapy, including aumolertinib, furmonertinib, osimertinib, apatinib, icotinib, gefitinib, and erlotinib plus bevacizumab. One RCT evaluated first- and second-line therapies, including icotinib versus WBRT with or without chemotherapy, and one evaluated second-line therapy with osimertinib versus platinum and pemetrexed-based chemotherapy. One RCT evaluated sintilimab plus a bevacizumab biosimilar plus chemotherapy versus chemotherapy only in patients who had unsuccessful treatment with an EGFR TKI.

The remaining 10 RCTs evaluated patients with ALK alterations. Eight RCTs evaluated first-line targeted therapies (alectinib, brigatinib, ensartinib, lorlatinib, crizotinib, and ceritinib). One RCT evaluated second-line and later-line therapy with ceritinib versus chemotherapy, and one RCT evaluated alectinib versus chemotherapy for third-line therapy.

Among the 59 observational studies, 22 reported first-line therapy, nine reported first-line or second-line therapy, two reported second-line or later-line therapy, 14 reported first-line or later-line therapy, and 10 did not report the line of therapy. For first-line therapy, treatments included targeted therapy (EGFR and ALK TKIs), chemotherapy (platinum- and pemetrexed-based), and radiotherapy (WBRT, SRS, and gamma knife radiotherapy). For second-line therapy, treatments included targeted therapy (gefitinib, erlotinib, and osimertinib). For other lines of therapy and studies that did not report line of therapy, treatments included targeted therapy (EGFR and ALK TKIs), chemotherapy (platinum- and pemetrexed-based), radiotherapy (WBRT, SRS, and gamma knife radiotherapy), and surgery.

CNS Clinical Outcomes

Median central nervous system–progression-free survival (CNS-PFS) and intracranial response (ICR) rates were reported in a minority of studies (n = 24). Studies of EGFR-mutated NSCLC and brain metastases continued to assess first- and second-generation EGFR TKIs, often in combination with other agents or radiotherapy. In studies of first- and second-generation EGFR TKI monotherapy, where reported, CNS-PFS and ICR rate did not vary greatly across agents within each study [18, 25, 43]. In the first-line setting, treatment with upfront WBRT with or without concomitant TKIs resulted in the more favorable clinical outcomes compared with treatment with TKIs only or upfront TKIs followed by WBRT. Three observational studies found that median CNS-PFS was longer in patients who had received earlier or upfront versus no or delayed radiotherapy [32, 68, 88]. Additional observational studies found that EGFR TKIs in combination or sequenced with radiotherapy (WBRT and/or SRS) had longer median CNS-PFS than with EGFR TKI monotherapy [24, 27, 30, 34, 38, 82, 98]. Several of these combination studies were utilizing first- or second-generation EGFR TKIs. One study by Yu et al., which compared osimertinib with and without upfront radiotherapy, also resulted in the combination having a longer median CNS-PFS [91].

In one RCT, second-generation icotinib resulted in a CNS-PFS of 10 months compared with 4.8 months with WBRT plus chemotherapy [86]. In one RCT, CNS-PFS with first-line use of third-generation furmonertinib was 20.8 months versus 9.8 months with first-generation TKIs [73]. In another RCT, second-line osimertinib resulted in a CNS-PFS of 11.7 months versus 5.6 months with chemotherapy [63]. In a first-line RCT, the ICR rate with osimertinib was 66% compared with 43% with first-generation TKIs [77]. Similarly, in a first-line observational study, the ICR rate with osimertinib was 68% compared with 50% with first-generation TKIs [96]. In a second-line RCT, the ICR rate with osimertinib was 40% versus 17% with chemotherapy [63].

Eight RCTs evaluated treatment for patients with NSCLC, brain metastases, and ALK alterations. Crizotinib continues to be the comparator for the second- and third-generation ALK TKIs. Where reported, these second- and third-generation ALK TKIs consistently demonstrated higher CNS-PFS and ICR than did crizotinib or chemotherapy. CNS-PFS with brigatinib was 24 months compared with 5.6 months with crizotinib [22], and 9.6 months with alectrinib compared with 1.4 months with chemotherapy [81]. ICR rates reached 73% (range 46–73%) with second- and third-generation ALK TKIs versus up to 22% with crizotinib and 21% with chemotherapy [71, 75, 97].

Discussion

Trends in Clinical Management

In terms of clinical management for patients with NSCLC, brain metastases, and AGAs, TKIs were described as potentially exhibiting higher penetration rates through the BBB as they are small molecules and have a good lipid–water partition coefficient. For patients with EGFR-mutated NSCLC, EGFR TKIs are the established first-line standard of care. In the treatment of brain metastases, while some countries may continue to rely on first- or even second-generation EGFR TKIs for first-line therapy, evidence has demonstrated that afatinib and osimertinib, as well as other third-generation EGFR TKIs, have better CNS penetration and superior CNS efficacy compared with first-generation options. Similarly, second- and third-generation ALK TKIs are also showing significantly improved CNS efficacy over the previous standard of care, crizotinib. The CNS-PFS and ICR results from both RCTs and observational studies related to these TKIs are continually assessed and reflected in updates across practice guidelines and recommendations globally.

Radiotherapy was found to positively affect the BBB by increasing permeability and the concentration of TKIs in cerebrospinal fluid. Adjuvant radiotherapy, when administered with TKIs, facilitated the TKIs’ capacity to cross the BBB, and thus demonstrated favorable anticancer effect. Additionally, patients who were asymptomatic were more likely to have been treated with EGFR TKIs only or EGFR TKIs plus SRS compared with patients who were treated with WBRT alone or in combination with another type of therapy.

Continuing Need for Optimal Management of NSCLC and Brain Metastases

Optimal management of brain metastases in NSCLC remains a high priority with continuing unmet needs. There were limited actionable targets evaluated among the included studies (79 of 80 studies evaluated EGFR or ALK, and one study evaluated EGFR, ALK, RET, MET, or ROS1). Still, results favored targeted therapy, as well as a combination of localized therapy and targeted therapy, over chemotherapy. Although there is evidence of the effectiveness of systemic therapies and targeted therapies for treatment of brain metastases, many studies of potentially effective anticancer therapies continue to exclude patients with active brain metastases [101].

Though some benefit was observed, the WBRT studies reviewed in this SLR are likely reflective of outdated practice patterns. In the current treatment landscape, conventional WBRT is generally avoided because it causes more neurocognitive problems than SRS [102, 103]. WBRT is frequently reported as an independent prognostic factor of overall survival along with extracranial metastases and performance score. Overall, WBRT is associated with serious harm, does not prolong survival, and yields poorer quality of life; thus, SRS or SRT has been suggested for treating patients with brain metastases when feasible. Lower incidence of radiotherapy-induced brain damage with SRS versus WBRT can be attributed to SRS’s ability to target the high-dose radioactive ray directly at the metastatic brain lesion, resulting in less damage to surrounding normal brain tissues and mitigating the radiotherapy-induced adverse reactions [41, 52]. Of note, in developing countries, first-generation EGFR TKIs and WBRT remain the primary treatment in patients with NSCLC and brain metastases, further highlighting the need for a consistent standard of care for this population [52].

In a retrospective study by Rakshit et al., the authors noted that patients with NSCLC with driver mutations had a high incidence of brain metastases at diagnosis; however, no statistically significant differences in survival outcomes were observed between patients with brain metastases and those without brain metastases [104]. These favorable outcomes for patients with brain metastases were surmised to be related to the use of potent active targeted therapies with good CNS penetration for patients with AGAs. For example, osimertinib has exhibited a protective effect against developing brain metastases, demonstrating an advancement over its first- and second-generation EGFR TKI predecessors [105].

Another study by Julian et al. found that patients with KRAS G12C-positive NSCLC had a higher prevalence of brain metastases compared with patients with KRAS wild-type tumors. This finding suggests that more research should be performed to evaluate whether KRAS G12C inhibitors can be beneficial for patients with brain metastases [106]. A systematic review concluded that TKI alone resulted in superior results in comparison with TKI plus radiotherapy in patients with NSCLC and brain metastases [107].

Emerging Therapies

Although the BBB remains the primary focus of emerging therapies for patients with NSCLC and brain metastases, it must be acknowledged that primary and metastatic brain tumors can disrupt the structure of the BBB and form a blood–tumor barrier (BTB) [108, 109]. This BTB permeability appears to aid in the successful transport of not only targeted therapies but also some chemotherapies. Thus, emerging therapies for patients with NSCLC and brain metastases with or without AGAs focus on maximizing opportunities to cross the BBB and BTB.

As such, EGFR TKIs were reported to exhibit higher penetration rates than other systemic therapies. Some studies also suggested that radiotherapy, such as WBRT, demonstrated favorable effects in increasing the permeability and concentration of TKIs in cerebrospinal fluid. Current emerging targeted therapies being evaluated in patients with brain metastases include almonertinib, anlotinib, apatinib, dacomitinib, icotinib, lazertinib, lenvatinib, neratinib, osimertinib, zorifertinib, D-0316 (InventisBio), and TY-9591. Emerging therapies for other actionable alterations include alectinib, crizotinib, ensartinib, and lorlatinib for patients with ALK alterations; crizotinib and entrectinib for patients with ROS1 mutations; sotorasib and adagrasib for patients with KRAS mutations; tepotinib and capmatinib for patients with MET exon 14 mutations; pralsetinib and selpercatinib for patients with RET fusions; and dabrafenib plus vemurafenib and dabrafenib plus trametinib for patients with BRAF-V600E mutations.

While targeted therapies continue to emerge for those with de novo alterations, as patients move into later lines, therapies no longer work for these AGAs. Thus, it is also important to consider how to treat patients with NSCLC and brain metastases who no longer harbor AGAs. For patients with NSCLC, brain metastases, and no AGAs, immunotherapy has emerged as a new first-line standard of care mostly in combination with or following platinum-based chemotherapy. Emerging therapies for patients without AGAs include immunotherapies targeting programmed death cell (ligand) 1 (PD-1/PD-L1), which are thought to be able to penetrate the BBB, including atezolizumab, camrelizumab, cemiplimab, nivolumab, pembrolizumab, sintilimab, tislelizumab, and zimberelimab. Other emerging treatments noted for patients without AGAs include datopotamab deruxtecan (Dato-DXd; a trophoblast cell-surface antigen 2 [TROP2]-directed antibody–drug conjugate [ADC]), bevacizumab, Endostar (an endostatin), ipilimumab (a CTLA-4 inhibitor), temozolomide, 4-demethyl-4-cholesteryloxycarbonyl-penclomedine, OSE-2101 (a neoepitope vaccine restricted to HLA-A2-positive patients), and patritumab deruxtecan (HER3-DXd; a HER3-targeted ADC).

ADCs are emerging as an effective AGA-agnostic therapeutic option across tumor types and treatment lines. By synergistically combining the specificity of mAbs with the antitumor activity of cytotoxic agents, ADCs selectively bind to cancer cells and deliver their cytotoxic payload directly into cancer cells. Along with US Food and Drug Administration (FDA)-approved T-DXd, which targets HER2 in breast cancer and NSCLC, telisotuzumab-vedotin, Dato-DXd, and HER3-DXd are among several ADCs being investigated in NSCLC. Eight patients with brain metastases experienced a best overall intracranial response of partial or complete response. These findings demonstrate the potential of ADCs to effectively treat patients with brain metastases in later lines of therapy. Further recent data suggest that HER3 may be more abundantly expressed in brain metastases in patients with NSCLC than in extracranial metastases [110]. On the basis of these data and positive results of ADCs in extracranial disease, brain metastases-specific trials with HER3-targeting agents are warranted.

Generalizability

Most studies included in this review were observational studies, which effectively represent the NSCLC population in a real-world setting and reflect the generalizability of the study population. Additionally, studies in all mutation status subgroups were eligible for inclusion in this review, and results were summarized by patient subgroups, type of therapy, and line of therapy. There were no restrictions on interventions or geography. The outcomes included in this review covered a wide range of topics; as such, the findings could provide a comprehensive understanding of the current landscape of clinical characteristics, clinical management, and emerging therapies for patients with NSCLC and brain metastases with and without AGAs.

Strengths and Limitations

The strengths of this review include following the PRISMA and Cochrane guidelines, using independent reviewers with a process for resolving discrepancies, and utilizing artificial intelligence technology to screen excluded records. The inclusion criteria related to interventions and comparators were left broad to increase generalizability. Furthermore, amendments were made to the protocol in order to focus on the most relevant and robust information available. Most studies focused on EGFR mutations or ALK alterations, and few on other actionable driver alterations; however, these are the most common AGAs in this patient population. Although the search period started only in 2017, the purpose was to summarize and interpret the most recent findings on this topic on the basis of the latest treatment landscape. While the trials and trial designs differed between studies, all trials included followed the PICOTS eligibility criteria. Overall, this review provides a comprehensive overview of the clinical characteristics, clinical management, and emerging therapies for patients with NSCLC and brain metastases.

Of note, in the studies included in this SLR, a majority of patients were treated with first- and second-generation EGFR TKIs in the first-line setting. With osimertinib as the current standard of care in first-line EGFR-mutated NSCLC with brain metastases, this may be an important confounder, given its superior CNS activity in comparison to first-generation TKIs. On the basis of the quality assessment, although the majority of all studies were at low risk of bias, it must be acknowledged that eight of the 19 RCTs included in this subset review were categorized as high risk, and all for deviations from the intended interventions (e.g., non-protocol interventions, non-adherence by patient to assigned intervention) (Table S4). Many of the results from all types of studies supported current practice guidelines and continued to highlight the key treatment gaps for patients with NSCLC and brain metastases.

Conclusion

Brain metastases are a poor prognostic factor and are common in NSCLC. This review underscores the continued needs of patients with brain metastases in NSCLC, even in those who have AGAs, likely due to the lack of clear understanding regarding effective transport of therapeutic agents across the BBB. The results of this SLR emphasize the need for therapies that can improve clinical outcomes for this patient population. More data are still needed to confirm these findings, given the differences in trial designs of the trials evaluated in this SLR.

Given the recent advancement in targeted therapies, such as fourth-generation EGFR TKIs, new options may continue to improve CNS-related outcomes. Similarly, with an array of ADCs demonstrating their ability to deliver cytotoxic payload to tumors bearing the target antigen, it may be valid to hypothesize that ADCs might have strong activity in the CNS. Furthermore, brain metastases may increase the permeability of the BBB, allowing a more efficient passage of these drugs into the brain. It is important to evaluate therapies in patients with active, untreated brain metastases as these patients are often excluded from phase 3 trials because of logistical challenges and higher risks for toxicities [111]. Several approaches are being evaluated to overcome the challenges of the BBB [112]. Further clinical validation and transfer of these strategies to ADCs is planned. Aside from NSCLC, tumor regressions and prolongation of survival have been observed with ado-trastuzumab emtansine (T-DM1) in preclinical mouse models of HER2-positive breast cancer and brain metastases. Given the success of these agents in other tumor types, it is hypothesized that they may also prove to be successful in NSCLC.