A Study on the Diagnostic and Prognostic Value of Extrachromosomal Circular DNA in Breast Cancer

Abstract

Objectives: To investigate the clinical diagnostic and prognostic value of extrachromosomal circular DNA (eccDNA) in breast cancer, eccDNA profiles were constructed for 81 breast cancer tumor tissues and 33 adjacent non-tumor tissues.

Methods: The distribution characteristics of eccDNA across functional genomic elements and repetitive sequences were systematically analyzed. Furthermore, a diagnostic model for differentiating malignant and normal breast tissues, as well as a prognostic prediction model, was developed using a random forest algorithm.

Results: EccDNA in breast cancer tissues harbor a higher proportion of functional elements and repetitive sequences, with their annotated genes significantly enriched in tumor- and immune-related pathways. However, no significant differences in eccDNA features were observed across breast cancer subtypes or pathological stages. In the validation cohort, the eccDNA-based diagnostic model achieved an AUC of 0.83, with repetitive elements and enhancer-associated features contributing the most to diagnostic performance. The prognostic model achieved an AUC of 0.78, with repetitive element annotations also showing strong prognostic relevance.

Conclusions: These findings highlight the promising potential of eccDNA in the development of precision diagnostics and prognostic systems for breast cancer.

Keywords: breast cancer; diagnostic model; eccDNA; extrachromosomal circular DNA; prognostic evaluation.

Figure 1

Extrachromosomal circular DNA (eccDNA) profiles of breast cancer tissues and matched adjacent normal tissues. (A) Box plot of eccDNA numbers per million mapped reads (Wilcoxon test, p = 0.38). (B) The size distribution of eccDNA (Wilcoxon test, p < 2.22 × 10⁻¹⁶). (C) Genome-wide distribution of eccDNA numbers in 1 Mb genomic windows across breast cancer tissues. (D) The Kyoto encyclopedia of genes and genomes (KEGG) pathway enriched by eccDNA-annotated genes.

Figure 2

Comparative distribution of extrachromosomal circular DNA (eccDNA) across genomic elements in breast tumor versus adjacent normal tissues. p-values are determined using the Wilcoxon test. Orange point, tumor tissue sample; gray point, adjacent normal tissue sample. PM, per million mapped reads; DHS, DNase I hypersensitive site; SINE, short interspersed nuclear elements; LINE, long interspersed nuclear elements; LTR, long terminal repeat elements; MIR, Mammalian-wide interspersed repeat.

Figure 3

Performance of breast cancer diagnostic model based on extrachromosomal circular DNA (eccDNA) features. (A) Receiver operating characteristic (ROC) curves of all models on the training data. (B) ROC curves of the random forest model on the training and test data. (C) Feature importance based on Shapley Additive Explanation (SHAP) analysis of the random forest model. AUC, area under curve; SVM, support vector machine; KNN, K-nearest neighbors; DHS, DNase I hypersensitive site; LINE, long interspersed nuclear elements; LTR, long terminal repeat elements; MIR, Mammalian-wide interspersed repeat.

Figure 4

Survival analysis and prognostic prediction model for disease-free survival (DFS) based on extrachromosomal circular DNA (eccDNA) features in breast cancer. (A) Kaplan–Meier curves of intron-annotated eccDNA features. (B) Kaplan–Meier curves of Mammalian-wide interspersed repeat (MIR)-annotated eccDNA features. (C) Receiver operating characteristic (ROC) curves of all models on the training data. (D) ROC curves of random forest model on the training and test data. AUC, area under curve; SVM, support vector machine; KNN, K-nearest neighbors.