Improving outcome prediction for breast cancer patients

Abstract:

Breast cancer is composed of distinct diseases with different outcomes. To determine the prognosis of patients, clinical and pathological factors are currently employed. There are several clinical decision making tools that use a combination of these prognostic factors to adapt the adjuvant treatment based on the prognosis prediction. However, due to insufficiently accurate prognosis predictions, a substantial proportion of breast cancer patients with inherently good outcome breast cancer receive adjuvant systemic therapy without gaining any benefit. Several gene expression signatures have been proposed recently which have been demonstrated to be predictive of outcome in breast cancer. Clinical trials that are currently assessing the use of some of these gene expression signatures while ignoring the traditional clinicopathological variables.

We investigate whether signature performance can be improved by including these variables.

About the speaker: Dr Nicola Armstrong received her PhD from the Department of Statistics at UC Berkeley where she worked with Terry Speed on incorporating crossover interference in the analysis of experimental crosses. After graduating, she moved to the Netherlands, spending two years as a post doc at Eurandom followed by the Department of Mathematics at the Vrije Universiteit in Amsterdam. Then she held a senior statistician position at the Netherlands Cancer Institute in Amsterdam. She has recently returned to Australia and is now working at the Garvan Institute in Sydney.