Work

Is Breast Cancer Relapse Driven by Chance? Testing Dormancy-Escape Mechanisms in Three Patient Cohorts

Year: 2026
Type: preprint
Abstract: This study investigates whether breast cancer recurrence is primarily driven by chance, gradual changes in tumour dormancy, or transient shocks. Building on earlier theoretical work, we test the predicted recurrence-hazard signatures using three publicly available breast cancer cohorts: Rotterdam (2,982 patients), GBSG (686 patients), and METABRIC (2,376 patients), with METABRIC serving as a pre-specified independent replication cohort. Across all three cohorts, the recurrence hazard peaks in the second year and subsequently declines, a pattern consistent with chance-driven escape in a heterogeneous patient population. However, the Rotterdam cohort shows a possible late recurrence wave around 9–10 years. This finding was not robustly replicated in METABRIC, despite its larger number of patients at risk during the corresponding period. The study further compares chance-based, change-driven, and shock-based mechanisms using order-restricted statistical tests and parametric hazard models. The early recurrence peak is better explained by delayed chance in Rotterdam and GBSG, whereas a shock-type model fits better in METABRIC. Overall, the findings support chance-dominated escape as the best-supported explanation for the main recurrence pattern across the three cohorts. Evidence for a late change-driven recurrence wave remains weak and requires further investigation using additional long-term patient cohorts.
Language: English
Open Access status: green
OpenAlex ID: W7235008071
Cites:
Cited by:
Related to: 10