TY - JOUR JF - Int-J-Radiat-Res JO - Int J Radiat Res VL - 1 IS - 4 PY - 2004 Y1 - 2004/3/01 TI - Comparison of logistic regression and neural network models in predicting the outcome of biopsy in breast cancer from MRI findings TT - N2 - Background: We designed an algorithmic model based on the logistic regression analysis and a non-algorithmic model based on the Artificial Neural Network (ANN). Materials and methods: The ability of these models was compared together in clinical application to differentiate malignant from benign breast tumors in a study group of 161 patients' records. Each patient’s record consisted of 6 subjective features extracted from MRI appearance. These findings were encoded as features for an ANN as well as a logistic regression model (LRM) to predict biopsy outcome. After both models had been trained perfectly on samples (n=100), the validation samples (n=61) were presented to the trained network as well as the established LRMs. Finally, the diagnostic performance of models were compared to that of the radiologist in terms of sensitiv­ity, specificity and accuracy, using receiver operating characteristic curve (ROC) analysis. Results: The average output of the ANN yielded a perfect sensitivity (98%) and high accuracy (90%) similar to that one of an expert radiologist (96% and 92%) while specificity was smaller than that (67% verses 80%). The output of the LRM using significant features showed improvement in specificity from 60% for the LRM using all features to 93% for the reduced logistic regression model, keeping the accuracy around 90%. Conclusion: Results show that ANN and LRM prove the relationship between extracted morphological features and biopsy results. Using statistically significant variables reduced LRM outperformed of ANN with remarkable specificity while keeping high sensitivity is achieved. Iran . J. Radiat. Res., 2004 1(4): 217-228 SP - 217 EP - 228 AU - P. Abdolmaleki, AU - M. Yarmohammadi, AU - M. Gity, AD - KW - neural networks KW - logistic regression model KW - ROC curves UR - http://ijrr.com/article-1-33-en.html ER -