Framework of Computer Aided Diagnosis Systems for Cancer Classification Based on Medical Images.

Area of Science:

• Medical Imaging
• Computational Biology
• Artificial Intelligence in Healthcare

Background:

• Early cancer detection significantly improves patient survival and treatment efficacy.
• Computer-aided diagnosis (CAD) systems enhance medical image analysis, reducing diagnostic errors and physician workload.
• Traditional cancer diagnosis relies on various medical imaging modalities, including microscopic images.

Purpose of the Study:

• To propose a novel computer-aided diagnosis (CAD) framework for enhanced cancer detection using medical images.
• To improve the accuracy of classifying suspicious regions as normal or abnormal.
• To apply the framework for diagnosing Acute Lymphoblastic Leukemia (ALL) from blood smear images.

Main Methods:

• A four-stage framework: preprocessing, segmentation, feature extraction/selection, and classification.
• Utilized Ant Colony Optimization (ACO) for optimal feature subset selection.
• Evaluated multiple classifiers: Decision Tree (DT), K-nearest neighbor (K-NN), Naïve Bayes (NB), and Support Vector Machine (SVM).

Main Results:

• The proposed CAD framework achieved high diagnostic performance on blood smear images for ALL detection.
• The Decision Tree classifier yielded the best results: 96.25% accuracy, 97.3% sensitivity, and 95.35% specificity.
• Feature selection using ACO maximized the classification performance.

Conclusions:

• The developed CAD framework demonstrates significant potential for accurate and efficient cancer diagnosis, specifically for ALL.
• The integration of ACO for feature selection enhances the robustness of the diagnostic system.
• This approach offers a valuable tool to assist physicians in medical image interpretation, leading to better patient outcomes.