Abstract

Colorectal cancer (CRC) is a prevalent malignant tumor with high incidence and mortality rates, with recurrence being the primary cause of death among patients. Traditional Chinese Medicine (TCM) utilizes a holistic cognitive approach to develop herbal treatment strategies, displaying significant efficacy in slowing tumor progression and improving patients' quality of life. However, discrepancies in herbal prescription strategies arise from variations in clinicians’ syndrome differentiation and experiential knowledge, leading to a lack of systematic understanding of compound prescription compatibility rules. This study aims to identify TCM prescriptions that are effective in preventing or treating CRC recurrence through evidence-supported and clinically applied prescriptions. By employing Apriori association rule mining and graph convolutional network analysis, a core herb spectrum associated with anti-recurrence prescriptions was identified, emphasizing a high-confidence herbal combination for further mechanistic exploration. Subsequent network pharmacology analysis revealed quercetin and kaempferol as representative active compounds with PTGS2 identified as a potential key target. Molecular docking, molecular dynamics simulation, public database analysis, and in vitro experiments provided initial evidence supporting their interaction and biological effects in CRC cells. Overall, this study reveals the compatibility characteristics of TCM prescriptions for preventing CRC recurrence and offers a scientific foundation for the continued use of TCM-based strategies in managing CRC recurrence.