UroComp Home

Go to UroComp home

Abnormal Hormones in Male Infertility Neural Network

Portions of this document were adapted in part from our paper:

Powell CR, Desai1 RA, Makhlouf AA, Sigman M, Jarow JP, Ross LS and Niederberger CS: Computational models for detection of endocrinopathy in subfertile males: International Journal of Impotence Research, Volume 20, Issue 1 (January-February 2008), pages 79-84.

The reader is strongly encouraged to read this paper prior to using the network, as it contains a more specific description of the model.


About Hormonal Abnormalities in Infertile Males

We gratefully acknowledge Dr. Hussein Kandil's contribution to this section.

Approximately 15% of couples cannot conceive after 1 year of unprotected intercourse. About a third of couple's infertility is solely related to the male and in another third, the male contributes to the infertility. Hormonal abnormality is a relatively common cause of male infertility in up to half of infertile men depending on their sperm analysis. It is an important diagnosis to consider because it can be often easily treated.

The importance of identifying men with abnormal hormones is not only that they can be treated but also that other important related significant medical conditions may be detected. A good source of general information about male infertility including endocrine abnormalities can be found in its Google Knol.

Return to top of document

Neural Network Programming and Training

The data set included clinical information from 1035 men presenting for infertility evaluation. Endocrinopathy was defined as the presence of an abnormality in the serum hormonal panel without necessarily implying a primary endocrine cause of infertility. Testicular volume was determined using either a Seager or Prader orchidometer. All men had endocrine testing including FSH and testosterone. A minimum of two semen analyses were obtained from each man after a 2- to 3-day period of abstinence and evaluated by trained laboratory technicians. Seminal volume, sperm density, percent motility, forward progression and percentage of normal morphologic forms were recorded. Men were excluded who had undergone prior fertility workup or vasectomy.

We investigated four models including linear and quadratic discriminant function analysis (LDFA and QDFA), logistic regression (LR) and a neural network using "neUROn2++", a suite of C++ programs we designed to implement neural computational and statistical algorithms, which are cross compiled using Microsoft Visual C++ version 6 and GNU C++ (Cygwin port version 2.95).

The dataset was randomly divided into a training set of 777 subjects, and a test set of 258 subjects. The proportion of data from men with endocrinopathies was kept similar in both sets using a randomization algorithm, which preserved initial outcome frequencies in each. The test set was excluded from training and used only for cross validation (n1/n2 method). A single hidden layer with four hidden nodes was determined to represent an optimal topology which maintained acceptable goodness-of-fit without overlearning. The training method was canonical off-line backpropagation with weight decay, with the weight decay term lambda chosen to be 5e-05.1 The network to be trained to completion when the error was observed to be oscillating at a local error minimum and the error gradient less than 1e-06.

Wilk's Generalized Likelihood Ratio Test (GLRT) was employed to determine which input features were significant to the model's outcome in a reverse regression analysis.1 Three input variates were found to be significant to the model by this method of feature extraction, with testis volume and sperm count p < 1e-06 and sperm motility p = 0.25. The dataset was also modeled using logistic regression and linear and quadratic discriminant function analysis (LDFA and QDFA) to compare the nonlinear computational method of neural computation with traditional linear statistical modeling tools.

1A description of this method of training, including weight decay and feature extraction using Wilk's GLRT, may be found in Golden RM, Mathematical methods for neural network analysis and design, Cambridge, MA: MIT Press, 1996.

Return to top of document

Accuracy of the Neural Network Compared to Linear Methods

Method ROC Area
Neural Network
Logistic Regression
Quadratic Discriminant Function Analysis
Linear Discriminant Function Analysis

2Receiver Operating Characteristic Curve area. Numbers approach 1.0 as accuracy improves (a value of 1.0 would indicate sensitivity and specificity both 1.0). Areas were computed using the statistical method described by Wickens: Wickens TD, Elementary signal detection theory, New York: Oxford University Press, 2002.

Return to top of document