Introduction: Breast cancer remains a major health burden, causing considerable morbidity and mortality worldwide. Emodin is a natural anthraquinone compound with anti-tumor activity and is obtained mainly from the separation and purification of the Chinese medicine Rheum palmatum. This study aimed to investigate the potential targets and mechanisms of Rheum palmatum in the treatment of breast cancer using unsupervised machine learning techniques, thereby elucidating its pharmacological mechanisms and therapeutic potential and providing a theoretical basis for clinical applications. Methods: We conducted a literature search across four major databases, including PubMed, Embase, Web of Science, and the Cochrane Library. The duplicate studies were identified and removed from the EndNote library. All studies were first screened by title and abstract according to predefined inclusion and exclusion criteria. Study quality was assessed using the ARRIVE 2.0 guidelines, evaluating 23 sub-items and categorizing risk of bias as low, medium, or high. Data were analyzed using IBM SPSS. Categorical variables were summarized as frequencies and proportions. The study used several machine learning techniques, including unsupervised methods such as K-means clustering and Principal Component Analysis (PCA), as well as the Apriori algorithm for association analysis. Results: Of the 337 initially identified records, 61 experimental studies met the inclusion criteria. Quality assessment using ARRIVE 2.0 indicated a high risk of bias, primarily in the randomization and blinding domains. Frequency analysis revealed emodin and aloe-emodin as the most frequently studied compounds, predominantly exerting anti-proliferative and pro-apoptotic effects. Major therapeutic targets included MMP-9, Bcl-2, MMP-2, and NF-κB. The Apriori algorithm revealed strong associations between targets such as Caspase-3/Caspase-9 and MMP-9/MMP-2. PCA identified three principal components encompassing apoptosis regulation, angiogenesis/metastasis inhibition, and endoplasmic reticulum stress pathways. K-means clustering defined five molecular clusters, with Bcl-2, IL-1β, p38, ERK, TNF-α, Nrf2, and p53 emerging as central nodes. Protein-protein interaction analysis further underscored the significance of apoptosis and inflammatory signaling pathways in Rheum-mediated anticancer mechanisms. Conclusion: This study provides a comprehensive analysis of the molecular mechanisms and therapeutic targets of Rheum palmatum in breast cancer, offering insights into the development of targeted, plant-based therapeutic strategies.
Liu et al. (Tue,) studied this question.