Related Experiment Video
Updated: Dec 23, 2025
08:05
Detection and Isolation of Cancer in Prostate Biopsies Using Stimulated Raman Histology and Artificial Intelligence
Published on: June 10, 2025
1.0K
Detecting Prostate Cancer Using Pattern Recognition Neural Networks With Flow Cytometry-Based Immunophenotyping in
George A Dominguez1, Alexander T Polo1, John Roop1
1Anixa Biosciences, Inc., San Jose, CA, USA.
Biomarker Insights
|April 29, 2020
Summary
This study introduces a novel liquid biopsy using flow cytometry and artificial neural networks (ANNs) for prostate cancer (PCa) detection. The method shows high accuracy in identifying PCa, potentially reducing unnecessary biopsies.
Area of Science:
- Urology
- Immunology
- Computational Biology
Background:
- Current prostate cancer (PCa) screening methods suffer from high false positive rates, complicating clinical assessment and necessitating improved early detection strategies.
- The need for accurate, non-invasive diagnostic tools is critical to differentiate between benign prostatic hyperplasia (BPH) and malignant PCa, and to stratify disease risk.
Purpose of the Study:
- To develop and evaluate a liquid biopsy test for prostate cancer (PCa) detection utilizing flow cytometry immunophenotyping combined with artificial neural network (ANN) analysis.
- To assess the efficacy of this novel approach in distinguishing PCa patients from healthy donors and patients with benign prostatic hyperplasia (BPH).
Main Methods:
- Measurement of numerous myeloid and lymphoid cell populations, including myeloid-derived suppressor cells, via flow cytometry.
- Analysis of collected immunophenotyping data using a pattern recognition neural network (PRNN), a type of ANN.
- Inclusion of data from 156 PCa patients, 123 BPH patients, and 99 healthy male donors (HD).
Main Results:
- The PRNN analysis demonstrated high performance in detecting PCa against healthy donors, achieving 96.6% sensitivity, 87.5% specificity, and an AUC of 0.97.
- The model also showed potential in differentiating higher-risk PCa (≥Gleason 7) from lower-risk disease (BPH/Gleason 6), with 92.0% sensitivity, 42.7% specificity, and an AUC of 0.72.
- Specific cell populations, including myeloid-derived suppressor cells, were identified as key indicators in the PCa detection model.
Conclusions:
- Flow cytometry-based immunophenotyping data analyzed with PRNNs offers a promising approach for improving prostate cancer (PCa) detection.
- This method has the potential to significantly reduce the number of unnecessary prostate biopsies, thereby improving patient management and healthcare efficiency.
- Further validation is warranted to establish this liquid biopsy as a standard clinical tool for PCa screening and diagnosis.
