Identification and risk prediction of cancer symptom clusters based on integrated Chinese–Western multimodal data: a prospective real-world study protocol

Authors:

Abstract
Objectives To identify stable cancer symptom clusters (CSCs) phenotypes and trajectories in patients with breast and ovarian cancer, examine associations with traditional Chinese medicine (TCM) syndromes and develop prediction models.Methods A multicentre, prospective, longitudinal real-world cohort will enrol 2000 patients with breast cancer and 1000 patients with ovarian cancer. Data include clinical characteristics, symptoms, TCM syndrome information and multiomics data from subgroups. CSCs will be identified using exploratory factor analysis with bootstrap resampling; associations and trajectories will be analysed using regression and mixed-effects models. Prediction models will use LASSO(least absolute shrinkage and selection operator), random forest and XGBoost(extreme gradient boosting) with internal validation.Results Expected outcomes include stable CSCs phenotypes and clinically useful prediction models.Discussion This study integrates TCM syndrome differentiation to explore CSC heterogeneity and evolution in real-world settings.Conclusion Findings may support personalised symptom management and improve quality of life among patients with cancer.Trial registration number ChiCTR2500111873 (Approval time: 6 November 2025); ITMCTR2025001939 (registration time: 17 July 2025).
Funding Information
  • Noncommunicable Chronic Diseases-National Science and Technology Major Project (2024ZD0521401)

This publication has 16 references indexed in Scilit: