ctDNA Monitoring for Small Cell Lung Cancer: Ready for Prime Time?

Summary:

Small-cell lung cancer (SCLC) is the deadliest form of lung cancer and has few precision medicine approaches available. A recent study analyzed circulating tumor DNA (ctDNA) in 33 patients with extensive-stage SCLC and showed that ctDNA levels and response patterns correlate strongly with clinical response and survival outcomes.

In this issue of Clinical Cancer Research, Sivapalan and colleagues correlated ctDNA dynamics with clinical response and survival outcomes in patients with extensive-stage small cell lung cancer (ES-SCLC) (1). The authors performed hybrid-capture next-generation sequencing (NGS) of plasma cell-free DNA (cfDNA) using targeted error-correction sequencing (TEC-Seq) (2) in 33 patients with ES-SCLC, along with matched peripheral blood mononuclear cell (PBMC) sequencing from 32 of these patients. The authors identified tumor-specific single-nucleotide variants (SNVs) and insertions/deletions (indels) by identifying hotspot mutations in plasma with ≥25 occurrences in the Catalog of Somatic Mutations in Cancer (3), and for variants that did not meet this criteria, required that they not be present in matched PBMCs or the DNMT3A gene. They additionally performed genome-wide copy number alteration (CNA) analysis of plasma using CNVkit (4).

The authors hypothesized that the combination of targeted SNV/indel and genome-wide CNA data, which they called cell-free tumor load (cfTL), would increase the sensitivity for detecting ctDNA molecular responses. They analyzed plasma samples longitudinally in at least three timepoints per patient (pre-treatment, during treatment, and at the time of clinical progression). Then, they classified ctDNA dynamics into three groups: molecular response (sustained complete elimination of cfTL), molecular response followed by recrudescence (initial elimination of cfTL followed by an increase at the final timepoint), and molecular progression (persistence of cfTL across all timepoints). Over a median follow-up time of 11 months, they found that cfTL dynamics correlated with progression-free survival (PFS), overall survival (OS), durable clinical benefit, and anticipated radiologic findings, suggesting that cfTL may be valuable as a dynamic liquid biopsy biomarker to molecularly measure responses earlier than standard radiographic imaging.

The authors also highlighted the importance of matched PBMC DNA sequencing to enhance the specificity of their ctDNA analysis approach which does not require paired tumor tissue sequencing. In their cohort, 35% of patients had PBMC-derived variants detected by TEC-seq that would have been misclassified as tumor-derived. In fact, ctDNA analysis did not effectively stratify survival outcomes when these PBMC-derived variants were not filtered out, similar to what was observed earlier in gastric cancer patients (5). Thus, it is important to filter out these clonal hematopoiesis-related mutations from ctDNA results by deeply sequencing matched PBMCs in order to minimize false positive results and maximize prognostic power.

To further put the current study into context, it is important to examine previous work on ctDNA in SCLC. Our prior work published in Cancer Discovery in 2017 focused primarily on non-small-cell lung cancer, but included three patients with limited-stage SCLC, two of which had detectable ctDNA after definitive-intent chemoradiation, which predicted disease progression (6). The remaining SCLC patient had ctDNA that went from detectable pre-treatment to undetectable immediately post-treatment, which predicted long-term disease-free survival (6). In 2018, Almodovar and colleagues published clinical vignettes of SCLC patients suggesting that ctDNA changes can precede clinical progression, adjudicate radiographic mixed responses, enable early identification of treatment-refractory disease, and correspond to disease remission (7). More recently, Nong et al. and Herbreteau et al. showed that SCLC patients with elevated pre-treatment ctDNA levels (above the cohort median level) had worse PFS and OS than those with lower ctDNA levels (8,9).

Like the current study by Sivapalan et al. (1), these earlier studies were based on detection of SNVs and indels in cfDNA using targeted hybrid-capture panels. An innovation, however, of the current study is that it integrated genome-wide CNA status, which enabled molecular response assessment in six patients who did not have identifiable tumor-specific mutations in plasma. Additionally, similar to the four previous studies (6-9), Sivapalan et al. employed a tumor-agnostic approach to ctDNA analysis, which would make prospective studies using this technology more feasible in SCLC patients.

The current study also differentiates itself from the earlier ones by adding a third ctDNA response category – initial response followed by recrudescence – that was only possible by analyzing ctDNA from multiple serial timepoints. Interestingly, patients that fell in this intermediate response category also had survival outcomes that were intermediate in nature – falling in-between their molecular response and molecular progression counterparts. It will be interesting to corroborate these findings in the future not only in SCLC, but potentially in other solid tumor types as well. Future clinical trials should determine if treatment could be further customized for these patients, which if successful, would be a major clinical contribution.

Although this study’s findings are promising, it is important to note that some issues remain unanswered including the lack of standardization of timepoints for ctDNA analysis which were not matched across patients or imaging timepoints, and the abnormally low rate of RB1 mutations identified using the TEC-Seq approach. Indeed, RB1 has been reported to be mutated in 65-90% of all SCLC patients (10,11), yet the RB1 mutation rate in this cohort measured from plasma cfDNA was only 15%. The authors state that this is likely related to incomplete tiling of the RB1 gene, where targeted hybrid-capture probes covered only 31% of the coding sequence. Still, it will be important to optimize the gene panel so that it better captures the typical genomic characteristics of SCLC, including the high rate of RB1 mutations, prior to clinical application.

Assuming the gene panel can be further optimized, it will be important to consider how this cfDNA technology could be combined with clinical practice. The authors suggest that ctDNA monitoring can offer an opportunity for timely intervention, however, will this information truly help us make clinical decisions? It is important to note that we currently have a limited therapeutic armamentarium for ES-SCLC, and knowing whether a patient is responding or not to the available standard of care treatment prior to radiographic progression may not necessarily change treatment decisions. It will thus be important to conduct a follow-up study to determine whether ctDNA is predictive and can inform personalized treatment decision-making to improve survival outcomes for this deadly disease.

Overall, the study by Sivapalan et al. contributes valuable data for ctDNA monitoring of ES-SCLC, building upon prior studies which did not include genome-wide CNAs in their analytical approaches. However, while highly compelling, given the lack of predictive data shown here or in the previously published papers in this space, without clear guidance regarding how to utilize ctDNA data clinically, we believe that ctDNA is not ready yet for prime time use in SCLC. However, we share the authors’ vision that ctDNA has significant potential in SCLC, and that this study paves the way toward future investigation in clinical trials where ctDNA-based molecular progression could facilitate treatment escalation or adjustment, and ctDNA-based molecular response could facilitate treatment breaks and close surveillance in selected patients (Figure 1). Future studies should compare ctDNA-guided treatment paradigms to standard of care treatment, which if successful, could revolutionize clinical management and survival outcomes for this deadly disease.

Figure 1. Proposed clinical trial design for extensive-stage small cell lung cancer using ctDNA to personalize treatment decision-making.

Patients treated with first-line chemotherapy + immunotherapy that have persistence of ctDNA will be randomized to standard maintenance immunotherapy vs. initiating approved second-line therapies or novel investigational agents. Patients who become ctDNA(+) after initial elimination (molecular response followed by recrudescence) during the 2 years of standard of care ICI maintenance will be randomized to continue standard immunotherapy vs. initiate approved second-line therapies or novel investigational agents. Patients that sustain complete elimination of ctDNA (molecular response) will be randomized to continue standard of care immunotherapy for up to 2 years vs. undergo close surveillance. ChemoIO, chemotherapy + immunotherapy; ctDNA, circulating tumor DNA; ES-SCLC, extensive-stage small-cell lung cancer; ICI, immune check point inhibitor; OS, overall survival; PFS, progression-free survival; SOC, standard of care. (Figure created with BioRender.com.)