Prostate cancer (PCa) continues to pose a significant health challenge, prompting around 90,000 radical prostatectomies annually in the United States alone as a primary mode of treatment.¹ A central objective during these surgeries is the attainment of negative surgical margins. However, positive surgical margins are found in 13%-25% of cases, increasing the risk of cancer recurrence.^2,3 Current methodology utilizing frozen sections from the ex vivo prostate, while effective, are often resource-intensive and time-consuming, hindering widespread adoption.⁴

Our study proposes the use of stimulated Raman histology (SRH), a cutting-edge imaging technique enabling swift pathological examination of fresh, unprocessed prostate biopsies.⁵ SRH presents high-resolution images, offering transformative benefits for intraoperative decision-making, especially in detecting positive surgical margins. By identifying such margins intraoperatively during a nerve-sparing procedure, SRH could direct surgeons toward a more balanced nerve-sparing approach, effectively preserving crucial nerves while ensuring comprehensive cancer removal.

Our study hypothesized that SRH can visualize periprostatic tissue, enabling surgeons to intraoperatively identify positive surgical margins. This innovative approach, integrating SRH into the surgical workflow, could revolutionize intraoperative decision-making by providing high-resolution images of prostate tissue and guiding surgeons toward a judicious nerve-sparing approach. We aimed to identify positive surgical margins via sampling surrounding periprostatic tissues from the resection bed. The primary end point was to introduce this novel technique for surgical margin assessment during radical prostatectomy and to evaluate the surgical team’s interpretative accuracy, offering an alternative to traditional methods that require extensive pathological tissue processing and resources.

In this prospective, Institutional Review Board–approved study, we recruited 22 participants scheduled for robotic-assisted laparoscopic radical prostatectomy (RALP) between February and June 2022. Recruitment was consecutive, contingent on the availability of the SRH microscope (NIO). All participants underwent a comprehensive evaluation, including a 3-T multiparametric prostate MRI using PI-RADS (Prostate Imaging Reporting & Data System) v2.^6,7 An MRI-ultrasound fusion targeted biopsy was performed using the ARTEMIS system and Profuse Bx software, guiding precise segmentation, coregistration, and 3D biopsy planning; all participants underwent MRI-ultrasound fusion targeted biopsy with systematic biopsies (4 targeted prostate biopsies from the region[s] of interest and 12 software-populated, spatially distributed systemic biopsies).⁶

We devised disease-specific treatment strategies for performing radical prostatectomies using the da Vinci Xi surgical system. These were based on numerous considerations, including MRI results, prostate biopsy findings, PSA levels, outcomes of digital rectal examinations, patient preference, and preoperative sexual function. The goal was to achieve negative surgical margins during the operation, and then surgeons also selectively sampled 3- to 5-mm areas from the bilateral neurovascular bundles or other areas of interest from within the participant for SRH assessment (see Figure).

Figure. Post-resection view of pelvis during a laparoscopic robotic-assisted radical prostatectomy. A, The overall area, with circles B and C marking regions of surgical margin sampling imaged with SRH shown in parts B and C of Figure. B, A magnified view of Circle B, illustrating a benign surgical margin which comprises stroma and neurovascular tissue from the left para-apical neurovascular bundle. C, A magnified view of Circle C reveals a positive surgical margin featuring International Society of Urological Pathology grade group 2 prostate cancer along with perineural invasion from the right para-apical neurovascular bundle.

A novel aspect of our approach was the intraoperative use of the NIO SRH microscope. This allowed us to scan freshly resected, unstained tissue samples—compressed to a thickness of 225 µm on an NIO slide—directly in the operating room. We acquired SRH digital images of 3 × 3 mm or 5 × 5 mm regions through strip-tiling at a constant imaging depth of 20 µm from the sample surface, utilizing 2 Raman shifts: 2,845 cm⁻¹ and 2,930 cm⁻¹. To enhance image interpretation, we applied pseudo-coloring to make these images resemble traditional hematoxylin and eosin (H&E) staining.^5,8

In this real-time setting, the surgical team could interpret the surgical margins, making immediate decisions about additional margin resection or necessity of confirmatory frozen section analysis based on the SRH findings. This contrasts with traditional postoperative assessment of removed prostates. Our innovative procedure even permitted routine pathological processing with H&E staining based on intraoperative SRH findings. The final interpretation of the specimens was completed by a genitourinary pathologist, who used H&E and/or frozen sections for SRH comparison.

Overall, in our study, we performed a RALP with SRH surgical margin assessment on 22 participants. These assessments targeted the neurovascular bundles and any other areas the surgeons found concerning. We conducted a total of 121 margin evaluations, averaging 5.5 per participant. Each assessment took approximately 5.5 minutes, facilitating a near real-time analysis of the surgical margins during the surgery itself.

Our participants represented a high-risk cohort, with half being diagnosed with stage pT3 PCa. The final histopathological evaluations concurred with the SRH findings on 5 margins identified as positive during surgery. However, there were 2 discordant margin assessments between SRH and H&E. One surgical margin intraoperatively concerning for PCa was deemed to contain benign prostate glands on H&E analysis; though discordant, the finding of benign glands at a surgical margin necessitated wider resection. Additionally, 1 case where a margin was deemed benign intraoperatively was found to contain PCa on the final H&E evaluation. This discrepancy occurred as only 85% of the sample was scanned intraoperatively by SRH, and the PCa focus was at the edge of the sample. Nonetheless, the surgical team’s interpretation of the SRH results showed an overall accuracy of 95.5%, with a sensitivity of 83.3% and a specificity of 98.3%.

Our study underscores the promise of SRH as an effective tool for real-time, intraoperative surgical margin assessment during RALP. We found that SRH provides surgeons with a practical method for immediate pathological evaluation of resected margins. This real-time assessment is crucial for confirming complete tumor removal and minimizing the risk of leaving a positive surgical margin. With high accuracy, sensitivity, and specificity shown in this study, SRH stands as a potentially significant addition to surgical procedures, aiming to reduce the burden of surgical margin assessments while improving both oncologic outcomes and patient quality of life. By adopting SRH, we could see decreased local recurrence rates, improved functional outcomes like urinary continence and sexual function, and reduced necessity for salvage therapies.^9,10