Assessment of a breast cancer response to neoadjuvant chemotherapy using backscatter ultrasound statistics

Abstract

Neoadjuvant chemotherapy (NAC) is used in breast cancer treatment. A correct assessment of a cancer response to the therapy is necessary to decide whether to continue the treatment, change the drug or refer the patient to surgery to remove the tumor. The aim of this study was to assess the tumor response to the therapy, based on backscattered ultrasound statistics. Radio-frequency (RF) images were acquired from 30 tumors being treated using NAC. After the NAC therapy, the tumors were excised and examined histopathologically. As a part of the ultrasound data analysis, an experienced physician specified a tumor area on each image. Then, a probability density function (PDF) of the signal amplitude within the tumor was estimated, as well as tumor size (TS). Next, a difference between the PDF corresponding to the data acquired before the therapy and the PDFs corresponding to the successive NAC courses was measured for each tumor using the Kullback-Leibler symmetrized divergence (KLSD). A statistical analysis was performed to assess the ability of the KLSD and TS, obtained in the subsequent NAC rounds, to predict the final tumor response to the treatment. Results suggest that such parameters can be useful in the NAC monitoring.Neoadjuvant chemotherapy (NAC) is used in breast cancer treatment. A correct assessment of a cancer response to the therapy is necessary to decide whether to continue the treatment, change the drug or refer the patient to surgery to remove the tumor. The aim of this study was to assess the tumor response to the therapy, based on backscattered ultrasound statistics. Radio-frequency (RF) images were acquired from 30 tumors being treated using NAC. After the NAC therapy, the tumors were excised and examined histopathologically. As a part of the ultrasound data analysis, an experienced physician specified a tumor area on each image. Then, a probability density function (PDF) of the signal amplitude within the tumor was estimated, as well as tumor size (TS). Next, a difference between the PDF corresponding to the data acquired before the therapy and the PDFs corresponding to the successive NAC courses was measured for each tumor using the Kullback-Leibler symmetrized divergence (KLSD). A statistical analysis w...