Real-Time Intraoperative Spectroscopic Margin Detection in Robotic-Assisted Radical Prostatectomy on Prostate Cancer Patients

Prostate cancer is often treated through a radical prostatectomy procedure consisting of the removal of the entirety of the prostate gland. The success of this procedure relies on the surgeon's ability to resect prostate tissue maximally and ensure no cancer is left behind. Surgical margins are currently assessed by pathologists only hours to days following surgery and positive surgical margins (cancer cells having spread to the boundary of resection) occur in as many as 1 in 5 cases. The ability to characterize tissue intraoperatively, in real time is therefore an important clinical need and is unmet by current technology. Raman spectroscopy (RS) is an optical technology based on the scattering of light by specific molecules/molecular bonds. Combining the molecular specificity of RS with machine learning techniques has shown promise for tissue characterization based on molecular composition. This sort of molecular fingerprinting could benefit prostatectomy procedures by ensuring maximal resection of prostate tissue and cancer. In this thesis, we present the design and development of an RS system integrated to a surgical robotics platform commonly used in RP. Design, engineering, and validation work on phantoms is presented. A proof of concept study was performed to demonstrate the ability of RS to characterize prostate tissue wherein the system was used to obtain data on 20 whole human prostates ex vivo. Preliminary results show successful distinction between prostatic and extra-prostatic tissue with accuracy, sensitivity and specificity over 90%. Also presented is the successful use of the system with a surgical robot to perform in vivo measurements in humans on 5 patients for the first time.