IMAGE PROCESSING AND COMPUTER-AIDED DIAGNOSIS

An AI model of sonographer's evaluation+ S-Detect + elastography + clinical information improves the preoperative identification of benign and malignant breast masses

Purpose

The purpose of the study was to build an AI model with selected preoperative clinical features to further improve the accuracy of the assessment of benign and malignant breast nodules.

Methods

Patients who underwent ultrasound, strain elastography, and S-Detect before ultrasound-guided biopsy or surgical excision were enrolled. The diagnosis model was built using a logistic regression model. The diagnostic performances of different models were evaluated and compared.

Results

A total of 179 lesions (101 benign and 78 malignant) were included. The whole dataset consisted of a training set (145 patients) and an independent test set (34 patients). The AI models constructed based on clinical features, ultrasound features, and strain elastography to predict and classify benign and malignant breast nodules had ROC AUCs of 0.87, 0.81, and 0.79 in the test set. The AUCs of the sonographer and S-Detect were 0.75 and 0.82, respectively, in the test set. The AUC of the combined AI model with the best performance was 0.89 in the test set. The combined AI model showed a better specificity of 0.92 than the other models. The sonographer's assessment showed better sensitivity (0.97 in the test set).

Conclusion

The combined AI model could improve the preoperative identification of benign and malignant breast masses and may reduce unnecessary breast biopsies.

Visualization and Interpretation of Convolutional Neural Network Predictions in Detecting Pneumonia in Pediatric Chest Radiographs

Pneumonia affects 7% of the global population, resulting in 2 million pediatric deaths every year. Chest X-ray (CXR) analysis is routinely performed to diagnose the disease. Computer-aided diagnostic (CADx) tools aim to supplement decision-making. These tools process the handcrafted and/or convolutional neural network (CNN) extracted image features for visual recognition. However, CNNs are perceived as black boxes since their performance lack explanations. This is a serious bottleneck in applications involving medical screening/diagnosis since poorly interpreted model behavior could adversely affect the clinical decision. In this study, we evaluate, visualize, and explain the performance of customized CNNs to detect pneumonia and further differentiate between bacterial and viral types in pediatric CXRs. We present a novel visualization strategy to localize the region of interest (ROI) that is considered relevant for model predictions across all the inputs that belong to an expected class. We statistically validate the models' performance toward the underlying tasks. We observe that the customized VGG16 model achieves 96.2% and 93.6% accuracy in detecting the disease and distinguishing between bacterial and viral pneumonia respectively. The model outperforms the state-of-the-art in all performance metrics and demonstrates reduced bias and improved generalization.