Abstract

Prostate cancer (PCa) remains a major global health burden, yet current screening tools often lead to overdiagnosis due to low specificity, highlighting the urgent need for more precise diagnostic approaches. Prostatic fluid (PSF) represents a promising but underexplored biofluid with exceptional diagnostic potential due to its direct contact with the PCa microenvironment. Here, we employed molecule-level interpretable surface-enhanced Raman spectroscopy (SERS) to comprehensively investigate PCa-associated alterations in two PSF components including metabolites and small extracellular vesicles (sEVs) and explored their potential interrelations via correlation analysis. Through molecule-resolvable SERS spectral set (MORE SERSome) technique, we identified ergothioneine and deoxyguanosine as differential metabolites between PCa and benign prostatic hyperplasia patients. We further constructed a fusion diagnostic model by integrating metabolites and sEVs information. The fusion model significantly outperformed the diagnostic accuracy by applying any single component, suggesting diagnostic complementarity between PSF metabolites and sEVs. Integration with clinical variables such as age and plasma prostate-specific antigen concentration further enhanced performance with the area under the curve as high as 0.93 for PCa diagnosis, substantially surpassing existing screening methods. These findings strengthen the importance of in-depth analysis of specific PSF components and further promise the potential of SERS-based PSF profiling as a noninvasive strategy for PCa diagnosis and biopsy guidance.

Figure 1. Separating and SERSome measurement of PSF metabolites and PSF sEV. (a) Schematic workflow including metabolites extraction, sEVs separation, SERSome measurement, and following application. (b) NTA measurements of sEV isolating. The gray line represents the particle size and concentration distribution before EXODUS enrichment, while the red line indicates the distribution after enrichment. (c) TEM image of the isolated sEVs. (d) WB analysis of sEV protein markers. (e) Extinction spectrum and TEM image of Ag NPs. (f) Histogram of the hydrodynamic diameter of Ag NPs.

Figure 2. Spectral characteristics of PSF-metabolite SERSomes. (a, b) Representative PSF-metabolite SERSomes and their corresponding mean spectra from (a) a PCa patient and (b) a BPH patient. (c) Spectral fluctuations among single spectra within one SERSome. Bands with significant fluctuations are highlighted in red. (d) PCC versus spectral number per SERSome. The black curve indicates the mean value of all samples, and the gray shade exhibits the standard deviation (n = 77). (e) Heatmap of SERSomes from PCa and BPH groups, constructed using the mean spectrum of individual samples. (f) The p values of the SERS intensities at each wavenumber by ANOVA, converted to (−log10p) for clarity. Significant bands (p < 0.05) located at peak or shoulder positions are marked on the mean spectra. (g) Box plots of the significantly varied spectral bands between PCa and BPH groups by ANOVA, verified by outlier removal (*: p < 0.05).

https://doi.org/10.1021/acssensors.5c03331