Asian Oncology Research Journal

Background: Breast cancer is one of the leading causes of mortality in women in Ghana; prognostic studies are still dominated by single endpoint Kaplan-Meier and Cox models, which do not consider recurrence, progression, and competing mortality. The study aimed to come up with a robust multistate competing risks survival model to define disease processes in Ghanaian breast cancer patients and provide dynamic prognostic profiles.

Objectives: To measure transition-specific hazards between diagnosis and progression, breast cancer event with competing risk death; to obtain the cumulative probability of breast cancer event in the presence of competing mortality; and to obtain state occupation probabilities over follow-up time.

Methods: A retrospective cohort of 558 patients was analysed using cause-specific Cox multistate models, Fine-Gray sub-distribution hazards, nonparametric cumulative incidence functions, and Aalen-Johansen estimators. The baseline predictors were age, stage, grade, lymph node involvement, tumour size, molecular subtype and hormone receptor status in determining the Hazard Ratio (HR) and the Sub-distribution Hazard Ratio (SHR).

Findings: After five years of follow-up with a median of about five years, 241 incidences of breast cancer and 40 competing risks deaths were verified. The advanced stage was a strong predictor of transition adverse outcomes (HR = 1.55, 95% CI: 1.192.01, p = 0.001), whereas lymph node involvement was a strong predictor of adverse outcomes (HR = 1.23, 95% CI: 1.051.44, p = 0.012). The cumulative incidence of breast cancer events was strongly increased by stage (SHR = 2.12, 95% CI: 1.48-3.04, p < 0.001) and lymph node status (SHR = 1.49, 95% CI: 1.25-1.78, p < 0.001), as observed in the Fine-Grey model.

Conclusion: This research integrates progression and competing mortality into a multistate survival model in Ghana to overcome the statistical limitations of single-endpoint models. The evidence supports early diagnosis, expanded molecular profiling, and the implementation of state-based prognostic instruments to strengthen precision oncology and cancer control policy.