Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

A neurocomputational model for prostate carcinoma detection.

Pankaj Kalra1, Joanna Togami, Gaurav Bansal B S

  • 1Department of Urology, University of Illinois at Chicago, Chicago, Illinois 60612, USA.

Cancer
|October 30, 2003
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Journal of Urology Survey Abstracts: Craig Niederberger July.

The Journal of urology·2026
Same author

Evaluating Sperm DNA Damage: When and Why It Does Not Add to the Evaluation of Male Factor Infertility.

Advances in experimental medicine and biology·2026
Same author

Male Infertility.

The Journal of urology·2025
Same author

Association of Genomic Prostate Score at positive margin with recurrence after radical prostatectomy.

BJU international·2024
Same author

Women Urologists: Trends in Mentoring and Career Choices.

Urology practice·2023
Same author

Editorial Commentary.

Urology practice·2023

A new neural network model accurately predicts prostate cancer risk using seven clinical factors, outperforming traditional methods like PSA tests. This tool can aid clinicians in deciding on prostate biopsies.

Area of Science:

  • Urology
  • Oncology
  • Artificial Intelligence in Medicine

Background:

  • Current prostate carcinoma screening relies on digital rectal examination (DRE) and prostate-specific antigen (PSA).
  • Predicting individual patient risk is challenging due to nonlinear relationships between risk factors and prostate cancer.
  • Established risk factors include age, ethnicity, family history, and complexed PSA.

Purpose of the Study:

  • To investigate a neural network model for predicting prostate carcinoma risk.
  • To utilize seven readily available clinical features for risk assessment.
  • To improve the accuracy of prostate cancer risk prediction beyond linear univariate analysis.

Main Methods:

  • A dataset of 3268 men evaluated for prostate cancer detection was used.

Related Experiment Videos

  • Seven clinical features were analyzed: age, race, family history, International Prostate Symptom Score (IPSS), DRE, total PSA, and complexed PSA.
  • A neural computational system was developed and compared against linear and quadratic discriminant function analysis and logistic regression, with model accuracy evaluated using n1/n2 cross-validation.

    • Main Results:

    • The neural network model achieved a high receiving operating characteristic (ROC) area of 0.825 in the test set.
    • This significantly outperformed traditional methods, with PSA alone yielding ROC areas of 0.678 (total PSA) and 0.697 (complexed PSA).
    • All input features were found to be highly significant to the neural network model's performance.

    Conclusions:

    • A neural computational system effectively models prostate carcinoma risk using multiple clinical factors.
    • The model demonstrates critical significance of each of the seven input variables for accurate risk prediction.
    • The developed model is proposed for clinical use to assist in decisions regarding prostate biopsy.