Quantitative Ultrasound and Texture Predictors of Breast Tumour Response to Chemotherapy

Previous studies have demonstrated that quantitative ultrasound (QUS)is an effective tool for monitoring breast cancer patients undergoing neo-adjuvant chemotherapy (NAC). Here, we demonstrate the clinical utility of pre-treatment and early stage treatment QUS texture features in predicting the response of breast cancer patients to NAC. Radiofrequency (RF)ultrasound data were acquired from 100 locally advanced breast cancer (LABC)patients prior to treatment, and during the first, fourth and eighth week of treatment. QUS Spectral and backscatter parameters were computed from regions of interest (ROI)in the tumour core and its margin. Subsequently, employing gray-level co-occurrence matrices (GLCM), four textural features and image quality features including core-to-margin ratio (CMR)and core-to-margin contrast ratio (CMCR)were extracted from the parametric images as potential predictive indicators. QUS results were compared with the clinical and pathological response of each patient determined at the end of their NAC. Results from the 100 patients indicate that a combined QUS and texture feature model demonstrated a favourable clinical and pathological based response prediction with area under the receiver operating characteristics curves (AUC)of 82%, 80%, 87%, and 92 % prior to treatment, during treatment at week 1, 4, and 8, respectively. Best results were obtained using a radial-basis-function support vector machine (RBF -SVM)machine learning algorithm. Only four features were selected in each binary response group classification. The findings of this study suggest that QUS features of a breast tumour are strongly linked to tumour responsiveness. The ability to identify patients that would not benefit from NAC would facilitate salvage therapy and clinical management that has minimum toxicity and maximum outcome in terms of patient survival.