Urinary volatile organic compounds in prostate cancer biopsy pathologic risk stratification using logistic regression...

Area of Science:

• Metabolomics
• Analytical Chemistry
• Computational Biology

Background:

• Prostate cancer (PCa) is a leading cause of cancer death in men.
• Current prostate-specific antigen (PSA) testing lacks specificity, leading to over-diagnosis and unnecessary biopsies.
• Novel diagnostic tools are needed for accurate PCa detection and grading.

Purpose of the Study:

• To explore the use of urinary volatile organic compounds (VOCs) for PCa diagnosis.
• To develop and evaluate machine learning models for PCa detection and grade assessment.
• To provide an alternative or adjunct to current PCa screening methods.

Main Methods:

• Urine samples from 386 men (247 with PCa, 139 controls) were analyzed.
• Volatile organic compounds (VOCs) were extracted using stir bar sorptive extraction (SBSE).
• Gas chromatography-mass spectrometry (GC-MS) and regularized logistic regression with machine learning were employed.

Main Results:

• A diagnostic model achieved an AUC of 0.99 (training) and 0.88 (testing).
• A model differentiating low-grade from intermediate/high-grade PCa achieved an average AUC of 0.78.
• Over 22,000 VOCs were identified in urine samples.

Conclusions:

• Urinary VOC analysis combined with machine learning shows high accuracy for PCa detection.
• This method can help differentiate PCa grades, aiding in biopsy targeting.
• This approach offers a promising non-invasive tool to improve PCa screening and management.