The role of immunotherapy in fusion-driven lung cancer

KEYWORDS:

• Checkpoint inhibition
• gene fusion
• gene rearrangement
• immunotherapy
• non-small cell lung cancer

1. Introduction

Our understanding of non-small cell lung cancer (NSCLC) has transformed with the recent discovery of therapeutically targetable oncogenic drivers, including gene fusions or rearrangements [1]. This includes gene fusions in ALK present in approximately 5% of advanced NSCLC patients, RET in 1–2%, ROS1 in 1% and NTRK in <1% [2–5]. Although individually rare, collectively they comprise a significant proportion of NSCLC. Tyrosine kinase inhibitors (TKIs) are now established treatment options for advanced fusion-driven NSCLC patients; however, drug resistance inevitably ensues [2,6]. More recently, immune checkpoint inhibitor therapy has dramatically changed the treatment landscape for NSCLC in both the treatment naïve and resistant settings [7]. Immune checkpoint inhibition is now standard of care in the first-line setting for advanced disease, either alone or in combination with chemotherapy, in patients without EGFR or ALK genomic aberrations [8,9]. Nevertheless, NSCLC remains one of the leading causes of cancer death with many unanswered questions, including the efficacy for immune checkpoint inhibition in fusion-driven NSCLC. Importantly, there is mounting evidence suggesting extremely poor efficacy for immunotherapy in fusion-driven NSCLC [10].

1.1. Biological underpinnings in fusion-driven NSCLC

The fundamental underlying biology of fusion-driven cancers may ultimately be responsible for this poor response. Fusion-driven cancers are known to generally exhibit lower tumor mutation burden (TMB) [11,12]. This is of heightened relevance given the increasing evidence for TMB as a biomarker for immunotherapy in NSCLC [13]. Moreover, the immune microenvironment and infiltrating immune cell populations likely play a significant role. Targetable gene fusions are more frequently identified in never smokers and younger patients, potentially influencing both the immune microenvironment and TMB. Although there have been limited studies on PD-L1 expression, ALK rearrangements may be associated with PD-L1 overexpression [14–16]. However, this may be accompanied by low rates of concurrent PD-L1 expression and CD8+ tumor-infiltrating lymphocytes (TILs) [17]. These issues are further compounded by challenges in testing and interpretation of both TMB and PD-L1 status [18]. Even so, results from the IMpower150 study discussed in further details below, demonstrated that anti-angiogenic therapy with bevacizumab in combination with chemotherapy and anti-PD-1 inhibition may potentially exert immunomodulatory effects on the tumor immune microenvironment to enhance the efficacy of immunotherapy [19]. Therefore, this suggests strategies for immunotherapy in fusion-driven NSCLC may need to harness rational combination therapies to overcome biological limitations.

1.2. Evidence for immune checkpoint inhibition in fusion-driven NSCLC

The IMMUNOTARGET registry, demonstrated low objective response rates (ORR) in small cohorts of ALK (ORR = 0%), RET (ORR = 6%) and ROS1 (ORR = 17%) rearranged patients [20]. In addition, numerous other retrospective studies have supported these findings with low ORRs and short progression-free survival (PFS) in ALK [21,22] and RET [23,24] rearranged NSCLC. In particular, responses to anti-PD-1 monotherapy are rare. Accordingly, a recent systematic review and meta-analysis pooled data from numerous retrospective studies, highlighting the limited efficacy of immunotherapy in ALK, RET and ROS1 fusion-driven NSCLC [25]. There remains limited prospective clinical trial data concerning the efficacy of immune checkpoint inhibitors in fusion-driven NSCLC (Table 1). The IMpower150 study, an open-label phase III trial evaluating the use of atezolizumab in combination with bevacizumab, carboplatin and paclitaxel, included a small cohort of ALK rearranged patients (n = 34) [26]. This study demonstrated an improvement in PFS with the addition of atezolizumab, when ALK rearranged patients were analyzed together with EGFR mutant patients. The ATLANTIC study, a single-arm phase II trial evaluating durvalumab also included a cohort of ALK rearranged patients (n = 16) [27]. However, none of the ALK rearranged patients had an objective response.

Table 1. Immune checkpoint inhibitor efficacy in fusion-driven NSCLC from prospective clinical trials*

Study	Treatment	Fusion-driven NSCLC Cohort	Efficacy Outcomes
ATLANTIC [27,28]	Durvalumab	ALK rearranged NSCLC after prior ALK TKI and platinum-based chemotherapy (n = 16)	ORR 0%
		TC≥25% subgroup (n = 12)	Median PFS 1.8 months (95%CI 0.5–1.9) Median OS 6.3 months (95%CI 0.9-NR) 6-month OS rate 57.1% (95%CI 25.4–79.6) 12-month OS rate 35.7% (95%CI 9.8–63.3)
IMpower150 [19,26]	Atezolizumab plus bevacizumab and chemotherapy (ABCP) versus bevacizumab plus chemotherapy (BCP)	ALK rearranged NSCLC after prior ALK TKI (n = 34) ABCP group (n = 13) vs BCP group (n = 21)	Outcomes of ALK rearranged subgroup not reported separately, outcomes combined with EGFR mutant patients^# Median PFS 9.7 months (ABCP) vs 6.1 months (BCP), HR 0.59 (95%CI 0.37–0.91)
IMpower130 [29]	Atezolizumab plus chemotherapy versus chemotherapy (carboplatin plus nab-paclitaxel) alone	ALK rearranged NSCLC after prior ALK TKI ALK rearranged subgroup was reported combined with EGFR mutant subgroup (n = 44)	Outcomes of ALK rearranged subgroup not reported separately, outcomes combined with EGFR mutant patients Median PFS 7.0 months (atezolizumab plus chemotherapy) vs 6.0 months (chemotherapy alone), HR 0.75 (95%CI 0.36–1.54)
CheckMate 057 [30]	Nivolumab versus docetaxel	ALK rearranged NSCLC after prior platinum-based chemotherapy, prior ALK TKI allowed (n = 21)	Outcomes of ALK rearranged subgroup not reported

NR – not reached; NSCLC – non-small cell lung cancer; ORR – objective response rate; OS – overall survival; PFS – progression-free survival; TKI – tyrosine kinase inhibitor.

*Only studies with 10 or more fusion-driven NSCLC patients are included.

^#Outcomes for the EGFR mutant subgroup have been reported separately [19].

1.3. Cost-effectiveness of immune checkpoint inhibition in fusion-driven NSCLC

Despite the growing evidence for lack of efficacy, in the majority of large registration trials of immune checkpoint inhibitors, whilst ALK rearranged patients were excluded other fusion-driven NSCLC patients remained eligible. Furthermore, in addition to ALK, RET and ROS1, there are several emerging targetable fusion drivers, albeit rare such as BRAF, FGFR, NRG1 and NTRK [31]. Diagnostic testing for the range of fusion drivers is also evolving, and internationally in many centers, standard practice may only include ALK and ROS1 testing [32]. Collectively, this means there is a significant cohort of fusion-driven NSCLC patients, that are likely to be exposed to immune checkpoint inhibition whether as first-line therapy or in the resistant setting. Recent cost-effectiveness analyses have questioned the value of immune checkpoint inhibition in various diverse healthcare settings [33–35]. In a cohort of patients in whom there is limited efficacy, exposing patients to ineffective and expensive therapies that are not without toxicity is also of concern. As increasing numbers of targeted therapies for fusion-driven NSCLC are approved, the role for immune checkpoint inhibition in the targeted therapy resistance setting becomes the crucial clinical question to be addressed.

1.4. Important treatment considerations and future directions

The identification of oncogenic gene fusions in advanced NSCLC is increasingly important to match patients to targeted therapies. The potential to exclude patients from costly and ineffective therapies however, is an underappreciated consequence. With an expanding list of biomarkers not limited to fusions, molecular profiling with multiplex assays such as next-generation sequencing (NGS) panels is paramount. Nevertheless, more robust evidence is needed to determine what influence the presence of gene fusions should have on treatment decisions for immune checkpoint inhibitors. Whilst TKI therapies are now standard of care for ALK, NTRK, RET and ROS1 rearranged NSCLC, the role of immunotherapy after targeted therapy resistance or in the treatment-naïve setting in other fusion-driven NSCLC remains to be completely elucidated. Furthermore, as immunotherapy enters standard clinical practice for early stage and locally advanced NSCLC, understanding of its efficacy becomes increasingly important. This highlights the need for more high-quality data, particularly subgroup analyses from prospective clinical trials to learn about the true efficacy in this cohort. Future trials in unselected NSCLC populations may then also consider whether molecular subgroups such as fusion-driven NSCLC should be stratified or even excluded from trials, as commonly occurs with ALK rearranged patients. Ongoing phase III trials of RET TKIs (selpercatinib and pralsetinib) compared with chemotherapy with or without pembrolizumab will also provide important evidence in understanding the role for immune checkpoint inhibition. Lessons may also be learnt from our understanding of EGFR mutated NSCLC, in which there is a greater wealth of published clinical trial data and translational studies. Finally, real-world evidence from patient registries and pooled analyses are of heightened relevance given the relative rarity of many fusion-driven NSCLC subsets.

Given the increasing evidence for TKI therapy in the first-line setting for oncogene-driven NSCLC, decisions for immune checkpoint inhibitor therapy are likely to be on resistance [10]. Commonly in clinical practice, combination therapy with chemotherapy may be considered. This emphasizes the importance of understanding the additive or synergistic effects of combination therapy, as has been demonstrated with the IMpower150 combination regimen. Furthermore, numerous studies of TKI in combination with immune checkpoint inhibitors are ongoing [36]. Tumor evolution after resistance to therapy is also crucial to appreciate, and there is evidence to suggest biomarkers such as PD-L1 may change over time [37]. Additionally, risk of toxicity depending on the combination or sequencing of therapy must be considered. Higher rates of hepatic toxicity for example, have been reported with sequential ALK TKI after immune checkpoint inhibitor therapy [38] and the combination of nivolumab and crizotinib [39].

A deeper understanding of tumor biology with comprehensive immunoprofiling of fusion-driven tumors including the immune microenvironment is also essential. Correlations with immune checkpoint inhibitor response data will further elucidate potential biomarkers for treatment selection, and even allow for rational targets for combination therapy approaches. Immunogenicity with tumor-specific neoantigens and gene expression profiling for example are potential avenues for exploration to determine candidate targets [11]. Interestingly, a functioning immune system is needed for tumor clearance after molecular targeted therapies [40], and this may indicate that future research may find a way toward combinations of targeted and immunotherapy in this population.

2. Conclusion and expert summary

There is a fundamental need to improve our understanding and better characterize the role for immune checkpoint inhibition in fusion-driven NSCLC. There remains a distinct lack of evidence, particularly from prospective studies and largely owing to the relative rarity of these subgroups. As the molecular classification of NSCLC becomes more diverse and treatment options and sequencing more complex, improved patient selection for therapy incorporating clinical and molecular characteristics is crucial. This will have implications not only for patient outcomes and survival but also for drug development and clinical trial design.

Declaration of interest

AC Tan has received honoraria from ThermoFisher and consultancy/advisory fees from Amgen. J Chan has received honoraria from Pfizer. M Khasraw has received consultancy/advisory fees from Janssen, AbbVie, Ipsen, Pfizer Roche, and Jackson Laboratory for Genomic Medicine; research funding from AbbVie, Bristol-Myers Squibb, and Specialized Therapeutics.

Reviewer disclosures

A reviewer on this manuscript has received honoraria from Takeda, Genentech, served in an advisory/consultancy role for; AstraZeneca, Blueprint Medicines, Gerson Lehrman Group, Hengrui Pharmaceuticals, Jazz Pharmaceuticals, Novartis, Pfizer, Takeda, Research America and Equinox Group, has received a research grant/funding from Pfizer, and has received travel/accommodation/expenses AstraZeneca, Novartis, Pfizer, Takeda, Daichii Sanyko, Checkmate Pharmaceuticals. Other reviewers on this manuscript have no other relevant financial relationships or otherwise to disclose.

Additional information

Funding

This paper was not funded.