Although the majority of patients diagnosed with extensive-stage small cell lung cancer (ES-SCLC) respond to first-line chemotherapy, relapse invariably occurs and only 5% of patients are alive two years after initial diagnosis [1–3]. In the past several decades, very little progress has been made in developing effective systemic therapies for SCLC [4]. Programmed death (PD)-1 inhibitors, either alone or in combination with cytotoxic T-cell lymphocyte 4 (CTLA-4) inhibitors have shown promising antitumor activity in a subset of patients with previously-treated SCLC. In the CheckMate 032 phase I/II trial [5], the objective response rate (ORR) to nivolumab monotherapy and nivolumab plus ipilimumab was 11 and 23%, and the two-year overall survival rates were 14 and 26%, respectively [6]. Based on these results, single-agent nivolumab was granted accelerated FDA approval for patients with SCLC with disease progression following platinum-based chemotherapy and one other line of therapy. Additionally, among 24 patients with PD-L1 positive SCLC treated with the PD-1 inhibitor pembrolizumab in the KEYNOTE-028 phase 1b study, the ORR was 33% [7]. Recently, the phase I/III IMpower 133 trial demonstrated an overall survival benefit when the PD-L1 inhibitor atezolizumab was added to platinum/etoposide chemotherapy for the initial treatment of ES-SCLC [8], although why only a subset of patients benefitted from this combination therapy is not currently known.

Unfortunately, the identification of predictive biomarkers of efficacy of immune checkpoint inhibitors (ICIs) in SCLC has been challenging. In contrast to the ~ 60% of non-small cell lung cancers (NSCLCs) which are positive for expression of the programmed death ligand 1 (PD-L1) [9], only approximately 18–32% of SCLC cases are PD-L1 positive [5, 7]. Furthermore, responses to nivolumab alone or in combination with ipilimumab do not appear to correlate with PD-L1 expression, which argues against the use of PD-L1 as predictive biomarker for immunotherapy in SCLC [5, 6], and highlights the need to identify novel biomarkers in this disease.

In several tumor types, such as NSCLC, melanoma, and urothelial carcinomas, cancers with a high number of non-synonymous somatic mutations, and therefore a greater neoantigen load which be recognized and targeted by immune cells tend to have higher response rates to immune checkpoint inhibitors than cancers with a low tumor mutational burden (TMB) [10–16]. Although mechanisms underlying the association between TMB and benefit from ICIs are not fully understood, tumor-specific neoantigens resulting from somatic nonsynonymous mutations may elicit neoantigen-specific T-cell responses that direct anti-tumor immunity [17]. SCLC, which is almost invariably associated with smoking, has among the highest mutational loads across cancer types, likely owing to tobacco-induced mutagenesis, which is characterized by a high transversion/transition ratio and increased genomic instability [18–21]. A recent exploratory analysis of the CheckMate 032 study using whole exome sequencing (WES) with paired germline sequencing to quantify tumor somatic mutational load found that the estimated one-year progression-free survival rates were higher in the high TMB group (21.2 and 30.0% for nivolumab monotherapy and nivolumab plus ipilimumab, respectively) compared with the low (not calculable and 6.2%, respectively) or medium (3.1 and 8.0%, respectively) TMB groups. Similarly, within each treatment group, the estimated one-year overall survival rate was higher in the high TMB group (35.2 and 62.4% for nivolumab monotherapy and nivolumab plus ipilimumab, respectively) than in the low (22.1 and 23.4%, respectively) or medium (26.0 and 19.6%, respectively) tumor mutational burden groups [22]. By contrast, exploratory subgroup analyses of the IMpower 133 showed no clear suggestion that blood-based TMB is associated with clinical outcome in patients receiving chemotherapy plus atezolizumab [8].

While WES may be the best-established technique for quantifying mutations in the coding genome, this technique is not readily available to most practicing clinicians since it requires significant informatics expertise and relies on sequencing of paired normal samples to filter out germline variants. Targeted next generation sequencing (NGS) is a relatively fast, cost-effective, clinically-available tool for estimating TMB, and there is generally good correlation between NGS and WES for determining TMB [23–26]. Whether TMB as assessed by targeted NGS is associated with improved efficacy of ICIs in patients with advanced SCLC is still unknown.

In the present study we investigate the feasibility of using targeted NGS to quantify TMB in SCLC and determine if patients with SCLC and a high TMB are more likely to benefit from treatment with immune checkpoint inhibitors than in patients with SCLC and a low TMB.

Although ICIs can provide a substantial clinical benefit in a small proportion of patients with SCLC, the lack of clinically-accessible predictive biomarkers makes it challenging to identify patients who are more likely to respond to ICIs. Recent evidence using WES with paired germline sequencing has shown that high-TMB SCLCs are more likely to benefit from treatment with nivolumab ± ipilimumab [22]. However, whether TMB as assessed by targeted NGS is associated with immunotherapy efficacy in patients with SCLC is unknown. To address this, we conducted a retrospective study using targeted sequencing data to evaluate the impact of TMB on ICI efficacy in a cohort of patients with SCLC.

We found that patients with SCLC and an elevated TMB had significantly better clinical outcomes after immunotherapy treatment compared to those with lower TMB. Highlighting the continuous nature of TMB as a predictive biomarker, we also demonstrated that 1-year PFS and OS rates improved with increasing mutational load when TMB was divided into tertiles. Importantly, supporting the hypothesis that TMB is predictive of immunotherapy benefit, we found no association between TMB and outcomes in patients treated only with chemotherapy. Limitations of this study include that this was a retrospective analysis on a relatively small sample size of patients treated both on clinical trials as well as on commercial immunotherapy, and there was also heterogeneity of treatment with different PD-1 inhibitors with or without combined CTLA-4 inhibition.

In the context of available literature, our data provide the first evidence for the use of targeted NGS to assess TMB status for the prediction of efficacy of ICIs in SCLC. In contrast to WES, TMB can be easily assessed using targeted NGS profiling panels that are already in routine clinical use. Several reports have recently sequenced the same tumors with both WES and targeted NGS and found that TMB determined by WES closely correlated with TMB determined by NGS in different tumor types, including in SCLC [20, 23, 24]. However, not all sequencing panels can accurately estimate TMB, especially those with low genomic coverage < 0.5 Mb [23].

Whether TMB is also predictive in patients with SCLC treated with a combination of chemotherapy plus immunotherapy is unclear. An exploratory subgroup analysis of the IMpower 133 SCLC study of platinum/etoposide ± atezolizumab showed no clear evidence that high blood-based TMB (bTMB) levels were associated with improved clinical outcomes [8], but TMB from tumor tissue was not reported in this study. Other recent analyses have shown that high bTMB may identify patients who derive a clinical benefit from atezolizumab in previously-treated NSCLC [27]. Additional prospective analyses on the role of blood- versus tissue-based mutational load will be needed to identify the optimal technique for biomarker assessment in SCLC and other cancers.

How TMB will be incorporated into clinical decision-making for SCLC at this time is in need of further study, particularly because there is no clearly-established TMB cutoff for patient selection. Given the very limited treatment options currently available for patients with SCLC, immunotherapy should not be withheld from patients with SCLC and a low TMB. As more effective treatment options hopefully become available for patients with SCLC, TMB might be a useful biomarker in determining the order in which therapies are administered. Given the potential for unprecedented, durable responses to ICIs in patients with SCLC, use of targeted NGS to identify high-TMB tumors can rapidly identify patients who should be treated with immunotherapy without delay.