• Benign Prostatic Hyperplasia
  • Hormone Therapy
  • Genomic Testing
  • Next-Generation Imaging
  • UTUC
  • OAB and Incontinence
  • Genitourinary Cancers
  • Kidney Cancer
  • Men's Health
  • Pediatrics
  • Female Urology
  • Sexual Dysfunction
  • Kidney Stones
  • Urologic Surgery
  • Bladder Cancer
  • Benign Conditions
  • Prostate Cancer

Neural network model helps predict fertility success


Iowa City, IA-A neural network that takes into accountmaternal age, sperm retrieval technique, type of sperm used, andtype of male factor has been found to be clinically useful forpredicting the outcome of in vitro fertilization/intracytoplasmicsperm injection.

The network was developed by Moshe Wald, MD, and colleagues at the University of Iowa, Iowa City, in collaboration with Craig Niederberger, MD, associate professor of urology and chief of andrology at the University of Illinois at Chicago College of Medicine. It is available for clinicians to use free of charge on the Internet.

"This is a model that has been statistically shown to be very accurate," said Dr. Wald, assistant professor of male infertility and andrology at the University of Iowa. "We first made sure that the neural network was the best computational model available in terms of its ability to predict ICSI outcomes. Only then did we deploy it on the Web and make it available for handheld computers, too."

The model then supplies a prediction for success and the overall odds of achieving pregnancy.

"While the model is easy to use, we feel it is best designed for physicians treating infertile couples, specifically physicians doing IVF/ICSI," Dr. Wald said. "You need to insert certain inputs to generate a prediction of IVF/ICSI outcomes, and most of those are not available to laypeople."

Dr. Wald and his group retrospectively analyzed a data set of 113 exemplars derived from patients who underwent IVF/ICSI with surgically retrieved sperm. The data set, with appropriate input factors, was then randomized into a modeling/training set of 83 and a cross-validation/test set of 30. Then it was modeled with computer-aided linear and quadratic discriminant function analysis, logistic regression, and neural computation. Ultimately, researchers found that a four-hidden-node neural network had the most accuracy, and that maternal age was the most significant feature in predicting pregnancy (p=.025).

Sperm type was next in line in terms of predictive power (p=.076), while male factor (p=.47) and retrieval technique (p=.88) did not predict outcome.

The neural network can be accessed by going to http://www.urocomp.net/. The web site also features other computational models designed to predict the presence of hypo-gonadism, erectile dysfunction, and pr-ostate cancer, as well as stone size and outcomes of shockwave lithotripsy.

Dr. Wald and colleagues first described the computational model for IVF/ICSI outcomes in the September 2005 issue of Reproductive BioMedicine Online (2005; 11:325-31).

In the paper, the authors note that "in a non-linear mathematical model, such as neural computation, a simple statement about the superiority of frozen or fresh sperm in pregnancy outcomes cannot be derived. However, clinicians using the model may substitute either fresh or frozen spermatozoa into that variate to predict whether the outcome would be expected to be better or worse for that particular maternal age, retrieval technique, and male factor type.

Related Videos
doctor showing a patient some information on a digital tablet | Image Credit: © bongkarn - stock.adobe.com
Raveen Syan, MD, FPMRS, answers a question during a Zoom video interview
Blur image of hospital corridor | Image Credit: © zephyr_p - stock.adobe.com
Cassandra K. Kisby, MD, MS, answers a question during a Zoom video interview
Vikas Desai, MD, answers a question during a Zoom video interview
Bree Duncan, RN, BSN, answers a question during a Zoom video interview
3d rendered illustration of bladder cancer | Image Credit: © SciePro - stock.adobe.com
Roger R. Dmochowski, MD, MMHC, FACS, answers a question during a Zoom video interview
Doctor holding patient's hands | Image Credit: © bongkarn - stock.adobe.com
Related Content
© 2024 MJH Life Sciences

All rights reserved.