Electrical Impedance Spectroscopy with Multi-modal Tissue Discrimination for Ultrasound-Guided Targeted Biopsy

Abstract

Prostate cancer (PCa) diagnosis has undergone a dramatic evolution in recent years.Men with clinical risk factors and important lesions observed via magnetic resonance imaging (MRI) typically undergo a targeted biopsy. In a typical prostate biopsy procedure, the operator uses ultrasound (US) imaging to place a biopsy needle in the lesion(s) identified pre-operation via MRI for sampling. Since the lesion is not always visible in US, there is no way to ascertain that the needle is correctly placed in the lesion before sampling. Thus, negative biopsy results occur often and lead to uncertainty in treatment decisions. This thesis proposes an instrumented needle probe that uses electrical impedance spectroscopy and algorithms to identify the tissue at the needle tip in real-time based on the tissue's bioelectrical properties. A new probe is designed, constructed, and validated on a collected dataset of ex-vivo tissue samples. Two data augmentation methods are proposed to improve the performance of tissue classification algorithms running on small datasets. Lastly, the thesis shows that US waves induce an electric potential in the tissue, which can be effectively measured with the sensorised needle and may provide additional information to classify the tissue. In the future, the use of the sensorised needle for tissue identification could improve confidence in biopsy results, and reduce procedure time and number of biopsies performed. i Components of Chapter 2, at time of writing, are under review as: C. McDermott, H. Asilian-Bidgoli, and C. Rossa -"Electrical impedance spectroscopy sensorized needle for prostate cancer localization during MRI-ultrasound-fused targeted biopsy",