Meeting Banner
Abstract #0608

Neoadjuvant Chemotherapy Treatment Prediction: A Classification Model Based Approach Utilising Pre-treatment DCE-MRI

Martin D Pickles 1 , Peter Gibbs 1 , Martin Lowry 1 , and Lindsay W Turnbull 1

1 Centre for Magnetic Resonance Investigations, Hull York Medical School at University of Hull, Hull, East Yorkshire, United Kingdom

The aim of this work was to develop a classification model to predict pCR, in patients undergoing neoadjuvant chemotherapy. To generate empirical vascular parameters dynamic data was interrogated in a pixel-by-pixel manner. Following pathological analysis Synthetic Minority Over-sampling TEchnique (SMOTE) was utilised to balance the pCR and non-pCR classes and a classification model was developed. High predictive accuracy was obtained from only 4 DCE-MRI parameters. This study suggests that prediction of pathological complete response, secondary to NAC treatment, can be made even prior to the initiation of chemotherapy from DCE-MRI parameters with a 86% accuracy.

This abstract and the presentation materials are available to members only; a login is required.

Join Here