Diagnostic Accuracy of Different Machine Learning Algorithms for Breast Cancer Risk Calculation: a Meta-Analysis

Abstract

Objective: The aim of this study was to determine the diagnostic accuracy of different machine learning algorithms<br />for breast cancer risk calculation. Methods: A meta-analysis was conducted of published research articles on diagnostic<br />test accuracy of different machine learning algorithms for breast cancer risk calculation published between January 2000<br />and May 2018 in the online article databases of PubMed, ProQuest and EBSCO. Paired forest plots were employed for<br />the analysis. Numerical values for sensitivity and specificity were obtained from false negative (FN), false positive (FP),<br />true negative (TN) and true positive (TP) rates, presented alongside graphical representations with boxes marking the<br />values and horizontal lines showing the confidence intervals (CIs). Summary receiver operating characteristic (SROC)<br />curves were applied to assess the performance of diagnostic tests. Data were processed using Review Manager 5.3<br />(RevMan 5.3). Results: A total of 1,879 articles were reviewed, of which 11 were selected for systematic review and<br />meta-analysis. Fve algorithms for machine learning able to predict breast cancer risk were identified: Super Vector<br />Machine (SVM); Artificial Neural Networks (ANN); Decision Tree (DT); Naive Bayes (NB); and K-Nearest Neighbor<br />(KNN). With the SVM, the Area Under Curve (AUC) from the SROC was > 90%, therefore classified into the excellent<br />category. Conclusion: The meta-analysis confirmed that the SVM algorithm is able to calculate breast cancer risk with<br />better accuracy value than other machine learning algorithms.