Urological disease management relies extensively on diagnostic radiology with ultrasound, computerized tomography (CT) and magnetic resonance imaging to dictate treatment outside the operating room. Exciting advances with intraoperative imaging will similarly guide surgeons to improve their surgical care in the future. We briefly highlight key advances in intraoperative imaging with prostate, kidney and bladder cancer.

Prostate

The main goal of radical prostatectomy (RP) is to offer complete tumor resection while preserving urinary continence and erectile function. Gallium-68 (⁶⁸Ga) prostate specific membrane antigen (PSMA) Cerenkov luminescence imaging (CLI) is a recently developed advanced intraoperative imaging technique that facilitates assessment of surgical margins during RP. Darr et al performed a feasibility study using ⁶⁸Ga-PSMA CLI during robot-assisted radical prostatectomy (RARP).^{1 68}Ga-PSMA-11 was injected intravenously and ⁶⁸Ga-PSMA positron emission tomography (PET)/CT was performed within an hour of administration. On completion of RP, the extracted prostate specimen was imaged using the LightPath™ CLI system, which captured a Cerenkov image as well as a photographic image. Although this study only consisted of 10 patients, CLI accurately detected 2 of 3 patients with positive surgical margins (PSMs) when CLI was compared to histopathology. Thus, CLI is a promising imaging technique that could be used for intraoperative identification of PSMs during RP.

Kidney

At last year’s annual meeting of the American Urological Association, Samiei et al described an innovative imaging technique for intraoperative renal tumor detection with the capability of predicting pathological type.² They developed a noninvasive intraoperative molecular chemical imaging (MCI) device that utilizes molecular spectroscopy and digital imaging. This device uses machine learning and computer vision strategies, and is able to differentiate malignant from surrounding benign tissue. In this study of 22 patients, the MCI device had 93.5% accuracy in identifying tumor from nontumor tissue.

Additionally, Sentell et al studied the role of indocyanine green (ICG) dye with near-infrared fluorescence (NIRF) imaging in differentiating renal tumors from normal surrounding renal parenchyma during robot-assisted partial nephrectomy.³ The authors evaluated a total of 330 tumors, and the overall rate of successful differential fluorescence (fluorescence of normal parenchyma in the absence of tumor fluorescence) was 87.3%. However, the differential fluorescence varied significantly by tumor histology; it was 100% for cystic and benign lesions, 90% for renal cell carcinomas and 72% for oncocytomas.

Bladder

In the past few years, there has been growing interest in using deep learning for nonmuscle invasive bladder cancer (NMIBC). NMIBC accounts for 75% of newly diagnosed bladder cancer and is characterized by high recurrence rates that require frequent endoscopic procedures. The frequent recurrence of NMIBC has been associated with incomplete resection of the previously diagnosed lesion(s). Shkolyar et al utilized deep learning and constructed an image analysis platform that detects bladder tumor during white light cystoscopy.⁴ This technique of “augmented” cystoscopy demonstrated a 91% sensitivity for bladder tumor identification. While deep learning is not yet being used in daily clinical practice, similar techniques have the potential to improve bladder resection practices among providers and facilitate diagnostic decision-making. Thus, this could lead to better resections and decrease the rate of NMIBC recurrence.

A different approach to improving bladder tumor resection is endoscopic molecular imaging (EMI).⁵ The goal of EMI is to provide real-time dynamic imaging during bladder resections and help urologists identify the tumor boundaries accurately. A variety of molecular tracers, including antibodies, protein scaffolds, peptides and small molecules, have been assessed in preclinical studies with promising results. The sensitivity and specificity of bladder cancer detection were 85% to 90%, and clear surgical margins were provided in xenograft studies. The major limitation of molecular imaging is the need to administer the cancer-specific tracer and be equipped with a paired detection medical device. Thus, this new imaging technology could serve as a supporting tool that enhances the surgeon’s visualization of the tumor during white light cystoscopy, although clinical trials are still lacking.

The intraoperative imaging armamentarium of surgeons continues to evolve rapidly. We have briefly discussed the most recent and promising imaging advances that will likely infiltrate daily practice in the coming years.