Epigenomic Profiling of Circulating Cell-Free DNA (cfDNA) to Characterize the Dynamic Evolution of Molecular Subtypes in Extensive-Stage Small Cell Lung Cancer During First-Line Chemoimmunotherapy (EPICIRC-SCLC)

Lung cancer is the leading cause of cancer-related deaths in France and worldwide. It is divided into two main types: non-small cell lung cancer (NSCLC), which makes up about 85% of cases, and small cell lung cancer (SCLC), which accounts for the remaining 15%. Treatment options for SCLC have changed very little over the past 40 years. The recent addition of immunotherapy to standard platinum-based chemotherapy for extensive-stage SCLC has improved overall survival by only 2-3 months. As a result, the prognosis remains extremely poor, with a median survival of about 12 months.

A new molecular classification of SCLC has emerged based on gene-expression profiling. It identifies four subgroups: SCLC-A, SCLC-N, SCLC-P, and SCLC-I. The first three groups are defined by the high activity of a specific transcription factor (ASCL1, NEUROD1, or POU2F3). The SCLC-I group is characterized by low activity of these transcription factors and high expression of inflammation-related genes, and may respond better to immunotherapy. Pre-clinical studies also suggest that each subgroup has specific sensitivities to different chemotherapies and to certain targeted drugs such as PARP or AURKA inhibitors.

SCLC also shows strong heterogeneity within individual tumors, with cancer cells able to shift from one transcriptional state to another over time. This plasticity suggests that the dominant molecular subtype at diagnosis may change during disease progression or in response to treatment. However, how these subtypes evolve during chemo-immunotherapy has never been studied directly in patients, mainly because new tumor biopsies are rarely performed after the initial diagnosis.

Recent work has shown that circulating cell-free DNA (cfDNA) methylation profiling can distinguish SCLC molecular subtypes from a blood sample. In addition, new epigenomic approaches based on cfDNA fragmentation patterns derived from low-pass whole-genome sequencing (lpWGS) can provide information about gene regulation and transcription factor activity using plasma samples. These techniques make it possible to capture both gene-level signals and broader epigenomic patterns, potentially overcoming the lack of tumor tissue available for analysis.

Hypothesis: We hypothesize that changes in SCLC molecular subtypes during first-line chemoimmunotherapy contribute to treatment resistance and progression, and can be identified by cfDNA epigenomic profiling. These changes may also limit the predictive utility of molecular characterization performed at diagnosis for guiding second-line therapeutic strategies.

Therefore, EPICIRC SCLC project aims in patients with an ES SCLC:

• To assess the evolution of the four molecular subtypes during first-line chemoimmunotherapy using epigenomic analyses on patient plasma samples.
• To identify subtype changes associated with tumor progression and treatment resistance.
• To find new therapeutic targets based on epigenomic data at baseline and after first line treatment failure.

This project is a collaborative study based on the BioLung SCLC cohort (NCT03387865, French ethics approval 2016-A01383-48), launched at Lille University Hospital in September 2023. The cohort will be expanded with patients from GHU Paris Centre (Cochin Hospital, Pr M. Wislez; and HEGP, Dr E. Fabre), whose ethics approval is currently being finalized.

The BioLung SCLC cohort includes consecutive patients who meet the following criteria:

A confirmed diagnosis of previously untreated small cell lung cancer.

Extensive-stage disease based on the Veteran's Administration Lung Cancer Group classification.

A multidisciplinary tumor board decision to start first-line chemo-immunotherapy (a platinum-etoposide regimen combined with a PD-L1 inhibitor).

Treatment delivered in the Thoracic Oncology Department of Lille University Hospital.

Health insurance coverage.

Signed informed consent.

The cohort is active and will continue enrolling patients until September 1st, 2027. For each participant, plasma samples are collected at three key time points (Figure 1):

At diagnosis, before treatment begins.

After four cycles of first-line chemo-immunotherapy.

At disease progression under first-line treatment.

Plasma samples are stored at -80°C in the Lille University Hospital Biological Resource Center, and diagnostic tumor samples (FFPE blocks) are archived in the pathology department.

For each enrolled patient, demographic data, tumor characteristics, treatments, follow-up information, and patient-reported outcomes (FACT-L, FACT-G, HADS, PEC, CARE) are recorded in an electronic case-report form.

So far, 24 patients have been enrolled. With an expected recruitment rate of about 20 patients per year, we anticipate including around 80 patients over four years, resulting in roughly 240 plasma samples and 80 tumor samples. Additional patients recruited at GHU Paris Centre following the same criteria should contribute ~15 patients per year.

Epigenomic Profiling of Tumor Samples First, diagnostic FFPE tumor samples will undergo bulk 3' RNA sequencing (ICM sequencing platform, Pitié-Salpêtrière Hospital) to classify tumors into the four known SCLC molecular subtypes using established methods.

We will then generate an integrated epigenomic signature for each subtype by combining:

Tumor DNA methylation profiling using MeDIP-seq (methylated DNA immunoprecipitation followed by sequencing).

Additional chromatin-based analyses aimed at identifying active and inactive genomic regions and transcription factor-associated patterns derived from sequencing-based assays.

The aim is to define, for each SCLC subtype, a set of genomic regions associated with active or repressed gene programs, as well as characteristic patterns linked to the activity of the key transcription factors ASCL1, NEUROD1, and POU2F3. Integrating these elements should provide a robust framework to distinguish SCLC subtypes based on their regulatory landscapes.

Epigenomic Profiling of Plasma Samples

Next, we will attempt to detect these tumor-derived regulatory signatures in blood, using the following cfDNA-based approaches:

cf-MeDIP-seq, to assess methylation patterns in circulating DNA.

Fragmentation-based assays, including low-pass whole-genome sequencing (lpWGS), to analyze cfDNA fragmentation patterns that reflect chromatin structure, transcription factor activity, and global regulatory states.

Bioinformatic Analyses

Tumor RNA-seq: Reads will be aligned with the STAR pipeline, and gene-expression matrices generated using RSEM.

Tumor epigenomic assays: Sequencing data will be processed using established pipelines (such as ChiLin) for alignment, quality control, and identification of enriched genomic regions. Differential analyses will rely on standard tools (e.g., deepTools, DiffBind) to define thousands of subtype-specific regulatory regions.

Plasma epigenomic assays: Data from cf-MeDIP-seq and fragmentation-based approaches will be analyzed with similar pipelines. A consensus genomic reference will be created, and coverage profiles across all samples will be compared to detect differences linked to molecular subtype or treatment response.

lpWGS: Tumor fraction will be estimated using the ichorCNA pipeline, which detects copy-number alterations in plasma. Chromatin accessibility and transcription factor activity will be inferred from nucleosome-based cfDNA fragmentation patterns using tools such as Griffin.

Previous work has shown that specific epigenomic patterns detected in circulating DNA can distinguish different tumor types-for example, separating adenocarcinoma from neuroendocrine prostate cancer, or differentiating SCLC from NSCLC using combined cfDNA methylation and chromatin-based signals. However, the molecular evolution of SCLC during first-line chemo-immunotherapy has never been studied in patients.

In this project, we aim to determine for the first time whether SCLC molecular subtypes change over the course of treatment, and whether specific epigenomic alterations in plasma are associated with the development of resistance. If successful, this work could:

Provide a detailed map of how regulatory programs evolve from diagnosis to progression, enabling future research on therapeutic vulnerabilities.

Support more personalized second-line treatment decisions based on dynamic molecular changes.

Reveal new therapeutic targets-such as cell-surface proteins-by characterizing how their regulatory pathways shift during treatment.

More broadly, this project could accelerate the development of innovative plasma-based epigenomic technologies. The approaches used here, including advanced cfDNA profiling methods recently published by only a small number of laboratories worldwide (including the Harvard group with whom S. Garinet previously collaborated), could be integrated across multiple liquid biopsy studies within our team and the Institute, helping establish a new standard for non-invasive tumor monitoring.