5-Hydroxymethylcytosine profiles in circulating cell-free DNA serve as potential biomarkers for diagnosis and classification of adenomyosis.

Abstract

Adenomyosis diagnosis and classification relies primarily on magnetic resonance imaging or pathological examination. Liquid biopsy has been widely applied in oncology but remains underexplored for adenomyosis. We aimed to identify liquid-biopsy-derived 5hmC biomarkers to aid adenomyosis diagnosis and molecular subtyping. Genome-wide 5hmC profiles were obtained by 5hmC-Seal from plasma cfDNA (adenomyosis: n = 51; controls: n = 46) and tissue DNA (adenomyosis: n = 26; controls: n = 21). A paired-tissue-driven splitting strategy was used: the training set comprised 26 adenomyosis plasmas with matched tissues plus 23 controls, and the validation set comprised remaining samples. A 5hmC-based logistic regression model was developed and validated. The identified marker amyloid precursor protein (APP) was functionally characterised in cellular and mouse models. Ten 5hmC markers were identified by machine learning. The model showed a sensitivity of 0.88 and a specificity of 0.87 (AUC = 0.91) in the validation set, with better performance than CA125 and CA199. 5hmC signatures could distinguish intrinsic from extrinsic adenomyosis subtypes. APP was related to epithelial cell migration and fibrosis in experimental models, which indicated its potential involvement in adenomyosis pathophysiology. Plasma cfDNA 5hmC markers show promise for adenomyosis diagnosis and molecular classification. The APP-HOXD9 axis represents a potential therapeutic target warranting further validation. These findings require prospective, multi-centre confirmation before clinical implementation. This work was supported by grant 82274034 from the National Natural Science Foundation of China, grant 2022YFB3604700 from the National Key Research and Development Program of China, and project BYESS2022031 from the Beijing Association for Science and Technology-Young Elite Scientist Sponsorship Program (BAST-YESS).