Prognostic and Predictive Values and Statistical Interactions in the Era of Personalized Medicine

• Activity Number: 153
• Type: Invited
• Date/Time: Monday, August 10, 2015 : 10:30 AM to 12:20 PM
• Sponsor: WNAR
• Author(s): Jaya M. Satagopan* and Alexia Iasonos and Qin Zhou
• Companies: Memorial Sloan Kettering Cancer Center and Memorial Sloan Kettering Cancer Center and Memorial Sloan Kettering Cancer Center
• Keywords: interaction ; likelihood ratio statistic ; AUC ; excess risk ; prognostic ; predictive

Abstract

The current era of targeted treatment has accelerated interest in studying gene-treatment, gene-gene and gene-environment interactions using statistical models. Interactions are incorporated in models as product terms of risk factors, and their statistical significance is examined using a likelihood ratio test (LRT). Epidemiological and clinical studies are interested in understanding whether interactions contribute to the prognostic and predictive values of genetic factors. Prognostic values can be examined via improvements in the area under the receiver operating characteristic curve with the inclusion of interaction terms (DeltaAUC). Predictive values provide insights into whether carriers of the genetic factors benifit from specific treatment or preventive interventions relative to non-carriers, and can be evaluated using the relative excess risk due to interaction (RERI). This work investigates the properties of epidemiological and clinical insights about interactions provided by LRT, DeltaAUC and RERI, and illustrates these approaches using published data on MC1R and sun exposure in melanoma.