Three-dimensional automatic artificial intelligence driven augmented-reality selective biopsy during nerve-sparing...

Area of Science:

• Urology
• Surgical Oncology
• Medical Imaging

Background:

• Robot-assisted radical prostatectomy (RARP) aims for cancer removal while preserving neurovascular bundles (NVBs).
• Accurate intraoperative localization of tumors near NVBs is crucial for nerve sparing and oncological control.
• Existing methods may have limitations in precisely identifying tumor extent at the NVB level.

Purpose of the Study:

• To evaluate the accuracy of a novel 3D artificial intelligence (AI)-driven automatic augmented reality (AAR) system.
• To assess the system's ability to identify tumor location at the NVB during RARP.
• To determine the impact of this technology on surgical outcomes.

Main Methods:

• Prospective study of 34 prostate cancer patients (cT1c-3, cN0, cM0) undergoing RARP.
• Preoperative multiparametric MRI used to assess capsular contact or extracapsular extension.
• AI-driven 3D AAR system employed at the end of the extirpative phase to guide selective biopsies of the NVB.

Main Results:

• The 3D AAR system identified tumor location on the NVB in 87.5% of pT3 patients.
• Positive surgical margin (PSM) rates were 0% for pT2 and 7.1% for pT3 cases.
• The system facilitated tailored nerve-sparing surgery in locally advanced disease.

Conclusions:

• The AI-driven 3D AAR system accurately localizes tumors at the NVB in RARP.
• This technology allows for precise nerve-sparing surgery, even in locally advanced prostate cancer.
• Oncological safety, measured by PSM rates, is maintained with this advanced system.