No abstract

This study aimed at elucidating the relationship between the number of computed tomography (CT) images, including data concerning the accuracy of models and contrast enhancement for classifying the images. We enrolled 1539 patients who underwent contrast or noncontrast CT imaging, followed by dividing the CT imaging dataset for creating classification models into 10 classes for brain, neck, chest, abdomen, and pelvis with contrast-enhanced and plain imaging. The number of images prepared in each class were 100, 500, 1000, 2000, 3000, 4000, 5000, 6000, 7000, 8000, 9000, and 10,000. Accordingly, the names of datasets were defined as 0.1K, 0.5K, 1K, 2K, 3K, 4K, 5K, 6K, 7K, 8K, 9K, and 10K, respectively. We subsequently created and evaluated the models and compared the convolutional neural network (CNN) architecture between AlexNet and GoogLeNet. The time required for training models of AlexNet was lesser than that for GoogLeNet. The best overall accuracy for the classification of 10 classes was 0.721 with the 10K dataset of GoogLeNet. Furthermore, the best overall accuracy for the classification of the slice position without contrast media was 0.862 with the 2K dataset of AlexNet.

Objectives: To investigate whether subendocardial and transmural myocardial infarction can be identified and differentiated using the peak circumferential and longitudinal strains measured by fast strain-encoding (SENC).Methods: Nineteen patients with ischemic heart diseases underwent imaging with fast SENC and late gadolinium enhancement (LGE) MRI at 3T. Fast SENC measurements were performed in three short-axis slices (basal, mid-ventricular and apical levels) and one long-axis view (four-chamber) to assess peak longitudinal and circumferential systolic strains.Results: All patients showed myocardial infarction with an average of 7 positive LGE segments. A total of 304 segments for longitudinal strains (LS) and 114 segments for circumferential strains (CS) could be analysed. Positive LGE segments showed lower peak CS and LS compared with the no-LGE segments (P<0.0001 for both). Segments with subendocardial infarction showed reduced CS and LS compared with the no-LGE segments (P<0.0001 for both). There was a significant difference in CS between subendocardial and transmural infarct segments (P=0.03), but no significant difference in LS between them (P=0.64).Conclusions: Fast SENC can identify old myocardial infarction and differentiate subendocardial from transmural infarction.

The use of dynamic perfusion MRI at 3T is useful for estimating MBF and can be applied for patients with coronary artery disease.