Related Experiment Video
Updated: Aug 5, 2026
Detection and Isolation of Cancer in Prostate Biopsies Using Stimulated Raman Histology and Artificial Intelligence
Published on: June 10, 2025
Plasma infrared molecular fingerprinting for prostate cancer detection: Prospective evidence for a partially
Philip Jacob1, Michael Chaloupka2, Tarek Eissa1
1Ludwig-Maximilians-Universität München (LMU), Chair of Experimental Physics - Laser Physics, Am Coulombwall 1, Garching 85748, Germany; Max Planck Institute of Quantum Optics (MPQ), Laboratory for Attosecond Physics, Hans-Kopfermann-Straße 1, Garching 85748, Germany.
Background:
Prostate-specific antigen (PSA) testing has limited specificity, particularly at intermediate concentrations, contributing to diagnostic uncertainty in prostate cancer (PCa). Infrared molecular fingerprinting (IMF) provides a rapid plasma-based readout of systemic biochemical variation and may capture PCa-associated molecular information not fully reflected by PSA.
Methods:
In the prospective Lasers4Life study, plasma from 2024 men was analysed, including 1012 treatment-naïve PCa patients, 752 men with benign prostatic hyperplasia (BPH), and 260 asymptomatic controls. Classification models were evaluated using cross-validation and, for PCa-versus-BPH discrimination, temporally independent held-out testing. Performance was quantified by receiver operating characteristic area under the curve (ROC AUC) and examined across PSA strata, comorbidities, and medications.
Results:
IMF discriminated PCa from asymptomatic controls more strongly than from BPH, with cross-validated AUCs of 0.79 ± 0.08 and 0.67 ± 0.04, respectively. In an exploratory age- and PSA-matched subset of men with PSA below 12 ng/mL, comprising 200 PCa and 200 BPH cases, IMF showed higher discrimination than PSA alone, although performance remained modest (AUC 0.63 vs 0.56; p < 1 ×10⁻⁴). Most examined comorbidities and medications did not substantially alter IMF performance, although prostatitis reduced discrimination.
Conclusions:
Plasma IMF detected PCa-associated variation in a pre-treatment diagnostic setting, including information partly independent of PSA. However, discrimination was modest in the clinically relevant PCa-versus-BPH comparison and did not exceed PSA in the overall cohort. These findings support further evaluation of IMF as an adjunctive molecular-profiling approach, requiring independent multicentre validation, clinically significant PCa endpoints, and demonstration of additive value within established diagnostic pathways.
