Bias in Meta-Analytic Modeling of Surrogate Endpoints in Cancer Screening Trials

TL;DR

This paper investigates how uncertainty in trial-level estimates can bias meta-analytic models assessing surrogacy in cancer screening, highlighting the need to incorporate this uncertainty for accurate evaluations.

Contribution

It demonstrates through simulation and theory that ignoring trial estimate uncertainty biases surrogacy assessments, and recommends incorporating this uncertainty in meta-analytic models.

Findings

Uncertainty biases meta-analytic surrogacy evaluation.

Limited information from completed trials on late-stage incidence surrogate.

Recommends restricting meta-analytic regression to models accounting for estimate uncertainty.

Abstract

In meta-analytic modeling, the functional relationship between a primary and surrogate endpoint is estimated using summary data from a set of completed clinical trials. Parameters in the meta-analytic model are used to assess the quality of the proposed surrogate. Recently, meta-analytic models have been employed to evaluate whether late-stage cancer incidence can serve as a surrogate for cancer mortality in cancer screening trials. A major challenge in meta-analytic models is that uncertainty of trial-level estimates affects the evaluation of surrogacy, since each trial provides only estimates of the primary and surrogate endpoints rather than their true parameter values. In this work, we show via simulation and theory that trial-level estimate uncertainty may bias the results of meta-analytic models towards positive findings of the quality of the surrogate. We focus on cancer screening trials and the late stage incidence surrogate. We reassess correlations between primary and surrogate endpoints in Ovarian cancer screening trials. Our findings indicate that completed trials provide limited information regarding quality of the late-stage incidence surrogate. These results support restricting meta-analytic regression usage to settings where trial-level estimate uncertainty is incorporated into the model.