Unraveling the Dynamic Trajectories and Predictive Determinants of Fear of Cancer Recurrence in Breast Cancer Survivors: A Prospective Cohort Study

BACKGROUND: Fear of cancer recurrence (FCR) is a dynamic phenomenon linked to negative outcomes among breast cancer survivors. Longitudinal studies on its causal links are limited. The aim of the study is to chart FCR trajectories in breast cancer patients, identify distinct patterns, and develop a predictive model. DESIGN: This multicenter prospective cohort study was conducted at two tertiary hospitals in Wuhan, China, from March 2022 to December 2023. 1095 breast cancer patients, with 790 patients' data used for analysis. Data collection occurred immediately after treatment planning, during treatment phases, and up to 3 months post-treatment. METHODS: Latent Class Growth Model (LCGM) identified FCR trajectories. LASSO regression analyzed predictors in the training cohort. Six machine learning models, including Shapley Additive Explanations (SHAP) and decision curve analysis (DCA), assessed predictive efficacy. RESULTS: Three FCR trajectories were found: "persistently low" (47.85%), "increasing" (32.28%), and "decreasing" (19.87%). The Random Forest (RF) model outperformed others, with an ROC AUC of 0.901 and PR AUC of 0.854, indicating high predictive accuracy. Key predictors included demographics, socioeconomics, clinical, and psychosocial factors. DCA validated the RF model's substantial net benefit. LIMITATION: We only measured FCR levels at six time points; assessing patients' psychological state as they reintegrate into society is our next focus. Future research should expand scope and duration. CONCLUSIONS: Diverse FCR trajectories exist among breast cancer patients, influencing prognosis. Early identification of high-risk factors is crucial for improving outcomes. Future research should focus on evidence-based interventions for FCR.