Time-varying effects of FOXA1 on breast cancer prognosis

Purpose Results of previous studies on the associations between Forkhead box A1 (FOXA1) expression in breast cancer tissues and the prognosis varied depending on the follow-up durations. The present study would investigate whether there is a time-varying effect of FOXA1 in breast cancer tissues on the prognosis. Methods FOXA1 expressions were evaluated in 1041 primary invasive breast tumors with tissue microarrays by immunohistochemistry. Cox models with restricted cubic splines and Kaplan–Meier survival analysis were used to examine the associations between FOXA1 and the prognosis. Flexible parametric models were applied to explore the time-varying effect of FOXA1. Results Overall, the association between FOXA1 expression and the prognosis was not significant but varied on the time of follow-up. Compared to FOXA1 ≤ 270 of H-score, the hazard ratios ( HR s) of death for those with 271–285 of FOXA1 expression increased from 0.35 (95% CI 0.14–0.86) at 6 months after diagnosis to 2.88 (95% CI 1.35–6.15) at 120 months with a crossover at around 36 months. Similar patterns were also observed for FOXA1 > 285 of H-score and for progression free survival (PFS). Moreover, when allowed both FOXA1 and estrogen receptor (ER) to change over time in the model (considering that ER had a similar time-varying effect), these time-varying effects remained for FOXA1 on both overall survival (OS) ( P