Related Experiment Video
Updated: Jul 7, 2026
13:19
Microarray-based Identification of Individual HERV Loci Expression: Application to Biomarker Discovery in Prostate Cancer
Published on: November 2, 2013
Genetic adaptive neural network to predict biochemical failure after radical prostatectomy: a multi-institutional
A Tewari1, M Issa, R El-Galley
1Josephine Ford Cancer Center and Department of Urology, Henry Ford Medical Center, Detroit, Michigan 48202, USA. Atiwari1@hfhs.org
Molecular Urology
|January 16, 2002
Summary
A new neural network model accurately predicts prostate cancer recurrence after surgery. This tool helps physicians select treatments and plan follow-ups for patients with localized prostate cancer.
Area of Science:
- Urology
- Oncology
- Artificial Intelligence in Medicine
Background:
- Radical prostatectomy is a common treatment for localized prostate cancer.
- Predicting biochemical recurrence (rising prostate-specific antigen [PSA]) is crucial for treatment selection and follow-up.
- Current prediction methods are limited, with approximately 30% of patients experiencing recurrence.
Purpose of the Study:
- To develop and validate a computer-based genetic adaptive neural network model for predicting PSA recurrence after radical prostatectomy.
- To utilize preoperative parameters for accurate recurrence prediction.
- To aid primary care physicians and urologists in making management recommendations.
Main Methods:
- A neural network model was constructed using demographic data, preoperative PSA, clinical staging, and Gleason scores from 1400 patients.
- The model simulated human learning by training on a database and was then used to predict outcomes in new patients.
- Preoperative parameters were entered into the trained model to assess its predictive accuracy.
Main Results:
- The study included 1400 patients with a mean age of 68.4 years and a mean preoperative PSA of 11.6 ng/mL.
- Biochemical recurrence was observed in 30.6% of patients during a mean follow-up of 41.5 months.
- The neural network model achieved 76% accuracy in predicting recurrence, with an area under the curve of 0.831.
Conclusions:
- A neural network model can predict PSA recurrence after radical prostatectomy with 76% accuracy.
- This predictive tool offers objective guidance for treatment counseling, potentially leading to cost savings through appropriate selection and follow-up.
- The technology holds promise for application in other prostate cancer treatment modalities like watchful waiting, external-beam radiation, and brachytherapy.
