Parametric Analysis of Ultrasound Backscatter Signals for Monitoring Cancer Cell Structural Changes During Cancer Treatment

High frequency ultrasound backscattered signals (20 - 60 MHz) from normal and apoptotic cell pellets differ in their backscatter intensity, and analyzing these signals could assist in the non-invasive monitoring of cancer therapy. In this work, the reflection coefficients of the lattice prediction error filter are used as feature set for parametric analysis and signal classification. The ultrasound (US) backscattered signal databases consisted of combinations of treated (apoptotic) and untreated (normal) cells mixed in different proportions. A 40 MHz commercial ultrasound imaging system was used. A classification accuracy of 97-100% for normal and apoptotic signals were obtained with a model order 15. The positive results ascertain that the reflection coefficient is a potential tool for analyzing biomedical signals such as US backscattered signals