• HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
Friday, November 20, 2026
BIOENGINEER.ORG
No Result
View All Result
  • Login
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
No Result
View All Result
Bioengineer.org
No Result
View All Result
Home NEWS Science News Cancer

Machine Learning Ensemble Accurately Predicts Prostate Cancer During Gray-Zone Biopsy Decisions

by
November 20, 2026
in Cancer, Health
Reading Time: 3 mins read
0
Machine Learning Ensemble Accurately Predicts Prostate Cancer During Gray-Zone Biopsy Decisions

Machine Learning Ensemble Accurately Predicts Prostate Cancer During Gray-Zone Biopsy Decisions

Share on FacebookShare on TwitterShare on LinkedinShare on RedditShare on Telegram

A newly developed machine-learning framework combining routine clinical markers demonstrates superior predictive performance in identifying prostate cancer among patients presenting with borderline prostate-specific antigen concentrations, according to findings from researchers at Dalian Medical University. The soft voting ensemble model, evaluated in a cohort of Chinese men with total serum prostate-specific antigen levels falling between 4 and 10 ng/mL, delivered an area under the receiver operating characteristic curve of 0.904 with a 95 percent confidence interval spanning 0.836 to 0.972. This diagnostic efficacy proved statistically significantly superior to that of any single standalone biomarker evaluated in the study.

Distinguishing between prostate cancer and benign prostatic hyperplasia has long represented a notable diagnostic challenge within this specific prostate-specific antigen concentration interval, frequently described in clinical urology as the diagnostic gray zone. Currently, clinicians lack reliable quantitative risk stratification tools capable of guiding individualized biopsy determinations for patients whose biomarker values sit within this intermediate spectrum.

All enrolled participants presented with baseline serum prostate-specific antigen measurements in the 4 to 10 ng/mL range.

These features comprised the free-to-total prostate-specific antigen ratio, the prostate health index, [-2]proPSA, prostate-specific antigen density, and total prostate volume.

The engineering of the machine-learning pipeline involved a soft voting ensemble algorithm configured through automated Optuna hyperparameter optimization. To counteract the pronounced numerical disparity between the 18 confirmed cancer cases and the 82 benign cases, the modeling framework incorporated specialized class imbalance correction techniques.

By maximizing Youden index, the investigators established an optimal decision cutoff value of 0.1726. At this diagnostic threshold, the ensemble model achieved an overall diagnostic accuracy of 81.00 percent, paired with a sensitivity of 94.44 percent and a specificity of 78.05 percent. Crucially for biopsy triage, the model demonstrated a negative predictive value of 98.46 percent, underscoring its ability to rule out malignancy with high reliability.

Subsequent evaluation using calibration curves and decision curve analysis further validated the practical utility of the machine-learning architecture. The calibration curve demonstrated close predictive consistency between estimated risk probabilities and observed pathological outcomes.

The investigators concluded that an ensemble model driven by five routine non-invasive clinical indicators offers dependable risk stratification for prostate cancer in Chinese men presenting with prostate-specific antigen levels in the 4 to 10 ng/mL range. Given its exceptionally high negative predictive value, the tool holds potential utility for accurately identifying low-risk individuals in whom invasive prostate biopsy could be safely deferred or avoided altogether.

Subject of Research: Machine learning ensemble risk prediction for prostate cancer biopsy decision-making in the PSA gray zone

Article Title: Clinical value of a soft voting ensemble model based on routine clinical indicators for risk prediction and biopsy decision-making of prostate cancer in Chinese men with PSA in the 4–10 ng/mL

Article References: Hua, Y., Li, Y., Xu, J., Wang, X., Li, X., & Wang, Y. (2026). Clinical value of a soft voting ensemble model based on routine clinical indicators for risk prediction and biopsy decision-making of prostate cancer in Chinese men with PSA in the 4–10 ng/mL. Frontiers in Oncology, 16, Article 1969579. https://doi.org/10.3389/fonc.2026.1969579

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1969579

Keywords: prostate cancer, prostate-specific antigen, PSA gray zone, machine learning, soft voting ensemble model, prostate biopsy, benign prostatic hyperplasia, prostate health index, free-to-total PSA ratio, p2PSA, PSA density, prostate volume, Optuna hyperparameter tuning, negative predictive value, decision curve analysis, repeated cross-validation, clinical decision support

News Source: Nathaniel Bowman. (November 20, 2026). Machine Learning Ensemble Accurately Predicts Prostate Cancer During Gray-Zone Biopsy Decisions. Scienmag.

Tags: Benign Prostatic HyperplasiaClinical decision supportdecision curve analysisfree-to-total PSA ratioMachine Learningnegative predictive valueOptuna hyperparameter tuningp2PSAprostate biopsyProstate CancerProstate Health Indexprostate volumeprostate-specific antigenPSA densityPSA gray zonerepeated cross-validationsoft voting ensemble model
Share12Tweet7Share2ShareShareShare1

Related Posts

Hospital Infection Control Training Shows Gaps Across Low- and Middle-Income Countries

Hospital Infection Control Training Shows Gaps Across Low- and Middle-Income Countries

November 20, 2026
Stroke Lesion Location and Unaffected Brain Inhibition Predict Motor Recovery Outcomes

Stroke Lesion Location and Unaffected Brain Inhibition Predict Motor Recovery Outcomes

November 20, 2026

Most South African Women with Genital Schistosomiasis Symptoms Prefer Biomedical Care

November 20, 2026

Dual First-Trimester Thyroid Antibodies Linked to Later Gestational Thyroid Dysfunction

November 20, 2026

POPULAR NEWS

  • Wild Custard Apple Leaf Size and Chemistry Vary Across Ghana Ecologies

    Wild Custard Apple Leaf Size and Chemistry Vary Across Ghana Ecologies

    29 shares
    Share 12 Tweet 7
  • Zanubrutinib Outperforms Older BTK Inhibitors in B Cell Lymphoma Meta-Analysis

    29 shares
    Share 12 Tweet 7
  • Child Care Pressures Spur Female Entrepreneurship Across Urban China via Reduced Compensation

    29 shares
    Share 12 Tweet 7
  • Pandemic Widened Physical Health Gap Between High and Low Earners in Germany

    29 shares
    Share 12 Tweet 7

About

We bring you the latest biotechnology news from best research centers and universities around the world. Check our website.

Follow us

Recent News

Urinary Proteomics Reveals Staged Heme Stress in Neonatal Hypoxic-Ischemic Encephalopathy

Hospital Infection Control Training Shows Gaps Across Low- and Middle-Income Countries

Stroke Lesion Location and Unaffected Brain Inhibition Predict Motor Recovery Outcomes

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 88 other subscribers
  • Contact Us

Bioengineer.org © Copyright 2023 All Rights Reserved.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In
No Result
View All Result
  • Homepages
    • Home Page 1
    • Home Page 2
  • News
  • National
  • Business
  • Health
  • Lifestyle
  • Science

Bioengineer.org © Copyright 2023 All Rights Reserved.