Using evolutionary computation to learn about detecting breast cancer

Disagreement or inconsistencies in mammographic interpretation motivates utilizing computerized pattern recognition algorithms to aid the assessment of radiographic features. This paper provides a review of recent efforts to evolve neural networks and linear classifiers to assist in the detection of breast cancer. Attention has been given to 216 cases (mammogram series) that presented suspicious characteristics. The domain expert (Wasson) quantified up to 12 radiographic features for each case based on guidelines from previous literature. Patient age was also included. The truth of malignancy or a benign case was available by examining the records of open surgical biopsy (111 malignant, 105 benign). Results indicate that both neural and linear models can yield suitable pattern classifiers and that fundamental relationships between input features and classification can be recognized