Thesis Committee

Ryan J. Halter, Ph.D. (Chair)

Alexander Hartov, Ph.D.

Elias S. Hyams, M.D.

Richard Bayford, Ph.D.

Abstract

Currently there is no reliable and efficient clinical protocol to intraoperatively assess surgical margins during radical prostatectomy (RP). The objective of this study is to demonstrate feasibility of near real-time intraoperative surgical margin assessment during a Robot Assisted Laparoscopic Prostatectomy (RALP) procedure using a drop-in endoscopic electrical impedance sensing probe.

This thesis focuses primarily on design, evaluation, and construction of an optimized endoscopic probe that can be deployed through a 12 mm diameter laparoscopic port positioned during a RALP procedure. The technology is evaluated through two ex-vivo studies and one in-vivo study of patients undergoing a RALP procedure. Two main techniques analyzed during this study are: Electrical Impedance Spectroscopy (EIS) and Electrical Impedance Tomography (EIT). Significant benign to malignant contrasts were observed for ex-vivo study.  Both EIS and EIT show good predictive performance of 0.76 and 0.80 area-under-curve (AUC), respectively, when considering discrete frequencies only. A machine learning algorithm is implemented to combined features, which improves the AUCs of EIS and EIT to 0.83 and 0.85, respectively. Finally, in-vivo data collected from patients undergoing RALP procedures from the two anatomical locations on the prostate and two surrounding peri-prostatic regions demonstrate clinical feasibility of using bio-impedance within a clinical setting. In summary, the significant benign to malignant contrasts observed ex-vivo and the successful use of this probe intraoperatively to collect impedance data suggests the feasibility of deploying this technology to provide surgeons with real-time information regarding the pathological state of margins during RP. This work provides a path for a larger follow-up clinical study.