Recurrence-linked eccDNA patterns offer new prognostic insights for lung adenocarcinoma

Lung adenocarcinoma, the most common histological form of lung cancer, can appear to have been completely removed after surgery while microscopic disease remains capable of returning years later. A new multi-omics study has identified distinctive features of extrachromosomal circular DNA, or eccDNA, that may help explain why some early-stage tumors recur after surgery. The findings, published in Precision Clinical Medicine on 6 July 2026, combine circular DNA sequencing with genomic, transcriptomic, epigenomic, immune, and clinical data to build a seven-gene model for estimating recurrence risk.

The study focused on a persistent challenge in lung cancer care. Surgery offers the possibility of cure for many patients diagnosed at an early stage, yet approximately 30%–50% eventually experience recurrence or aggressive metastasis within five years. Conventional clinical indicators do not always reveal which patients carry the highest risk. The researchers therefore investigated whether eccDNA could provide a molecular record of tumor behavior that is detectable not only in tumor tissue, but also in blood plasma.

Unlike the linear DNA packaged into chromosomes, eccDNA consists of closed, circular, double-stranded DNA molecules that exist independently of the normal chromosome structure. These circles can arise when parts of the genome are rearranged, amplified, or excised during tumor development. Because eccDNA can carry complete or partial genes and regulatory elements, it may increase the activity of cancer-promoting genes and contribute to genetic diversity within a tumor. That flexibility allows cancer cells to adapt rapidly to environmental stress, immune pressure, or treatment.

Researchers from Southwest Jiaotong University, The Third People’s Hospital of Chengdu, Chengdu University of Traditional Chinese Medicine, and Deyang People’s Hospital enrolled 90 treatment-naïve patients with early-stage lung adenocarcinoma. Tumor tissue, matched adjacent non-tumor tissue, and plasma were collected from each participant. Patients were followed for two years, and the final analysis included individuals with recurrence information, non-degraded samples, and complete follow-up data. The investigators then compared eccDNA patterns in patients whose disease returned with those in patients who remained recurrence-free.

To enrich the circular DNA, the team first extracted total DNA from the clinical samples and used exonuclease digestion to remove linear DNA. The remaining circular molecules were amplified and purified through rolling circle amplification, a technique that copies circular templates repeatedly to generate sufficient material for sequencing. The enriched DNA was analyzed using 150-base-pair paired-end sequencing on the NovaSeq 6000 platform. A computational method known as Circle-Map was then used to identify eccDNA by detecting sequencing reads that cross the junction where the two ends of a linear genomic segment have been joined into a circle.

The resulting profiles revealed that eccDNA from recurrent tumors had higher GC content and stronger split-read signals than eccDNA from non-recurrent tumors. GC content describes the proportion of guanine and cytosine bases in a DNA sequence, while split-read signals provide evidence that a sequencing read spans an eccDNA junction. Together, these features suggest that recurrence-associated circles may arise from particular genomic regions and may be present in structurally distinctive or more abundant forms.

The recurrence-associated eccDNAs were preferentially derived from transcriptionally active portions of the genome, including exons, CpG islands, and regions known as A compartments. In three-dimensional genome organization, A compartments are generally open, accessible chromatin domains with active gene transcription. The circles also showed links to active histone modifications, chemical marks on DNA-packaging proteins that help maintain genes in an accessible state. These observations indicate that eccDNA in recurrent tumors is not randomly distributed: it appears to be connected to genomic regions already primed for gene activity.

Several genes carried by recurrence-associated eccDNA were involved in pathways central to tumor progression, including mTOR, Notch, and Ras signaling. These pathways regulate cell growth, survival, differentiation, metabolism, and communication between tumor cells and their surroundings. The analysis also detected differences in the expression of immune-related factors and their receptors, raising the possibility that eccDNA contributes to altered immune signaling in recurrent lung adenocarcinoma. Although the study does not establish that these circles directly cause immune escape, the molecular associations provide a basis for investigating how circular DNA may influence the tumor microenvironment.

The researchers next examined plasma, where eccDNA could potentially be measured through a minimally invasive blood test. They identified 2,387 eccDNAs commonly increased in lung cancer and recurrence-associated samples. Genes connected to these plasma eccDNA signals were compared with public transcriptomic datasets and survival information. This integration produced a seven-gene signature consisting of AFAP1L2, LHX8, IL20RB, SLC12A8, EGLN3, CDH3, and PLTP. The resulting model consistently separated patients with different disease-free survival outcomes in both training and validation cohorts, suggesting that the signature may help classify postoperative recurrence risk.

The findings present eccDNA as more than a by-product of genomic instability. In lung adenocarcinoma, these circular molecules may act as mobile genetic platforms that preserve and amplify cancer-relevant information outside the chromosomes. Their presence in plasma also creates the possibility of monitoring tumor biology without repeatedly sampling tissue. However, the seven-gene model and the broader eccDNA patterns still require testing in larger, prospective, and geographically diverse patient groups. If validated, eccDNA-based assays could eventually complement imaging and pathology by identifying patients who need closer surveillance or additional postoperative treatment after lung cancer surgery.

News Publication Date: 6 July 2026

Web References: Precision Clinical Medicine, https://academic.oup.com/pcm; DOI: https://doi.org/10.1093/pcmedi/pbag017

References: 10.1093/pcmedi/pbag017