Time dependence of biomarkers: non-proportional effects of immunohistochemical panels predicting relapse risk in early breast cancer

Background: We investigated the impact of follow-up duration to determine whether two immunohistochemical prognostic panels, IHC4 and Mammostrat, provide information on the risk of early or late distant recurrence using the Edinburgh Breast Conservation Series and the Tamoxifen vs Exemestane Adjuvant Multinational (TEAM) trial. Methods: The multivariable fractional polynomial time (MFPT) algorithm was used to determine which variables had possible nonproportional effects. The performance of the scores was assessed at various lengths of follow-up and Cox regression modelling was performed over the intervals of 0- 5 years and 45 years. Results: We observed a strong time dependence of both the IHC4 and Mammostrat scores, with their effects decreasing over time. In the first 5 years of follow- up only, the addition of both scores to clinical factors provided statistically significant information (Po0.05), with increases in R-2 between 5 and 6% and increases in D-statistic between 0.16 and 0.21. Conclusions: Our analyses confirm that the IHC4 and Mammostrat scores are strong prognostic factors for time to distant recurrence but this is restricted to the first 5 years after diagnosis. This provides evidence for their combined use to predict early recurrence events in order to select those patients who may/will benefit from adjuvant chemotherapy.