Abstract

Radical prostatectomy is a standard curative approach for high-risk prostate cancer, yet it entails significant risks of postsurgical complications like erectile dysfunction and urinary incontinence. Accurately defining tumor margins during surgery remains a major challenge. Intraoperative assessment of prostate tumor malignancy, particularly those with high aggressiveness catalogued in Gleason grade group (GG) ≥3, is crucial to prevent positive surgical margins and minimize postoperative complications. Here, we develop a surface-enhanced Raman scattering (SERS)-based navigation system for intraoperative localization of high-grade malignant regions by simultaneously accessing tissue acidity and prostate specific antigen (PSA) enzymatic activity. This system integrates a sampling pen for automatic biomarkers extraction from exposed tissue surfaces at the discretion of the operator, a nano-imprinted SERS array producing ratiometric Raman signal in response to acidity and PSA activity, and  a 2D-transformed deep-learning model determining pH by processing Raman spectra swiftly and robustly. We show that the system can intraoperatively identify GG≥3 malignancies in fresh prostate tissues from 144 Chinese patients with an area under the receiver operating characteristic curve of 0.89. This SERS-based navigation system holds strong potential to enhance surgical precision, minimize tumor residue, and ultimately improve patient outcomes.