Search

Detection of early-stage liver fibrosis has direct clinical implications on patient management and treatment. The aim of this paper is to develop a non-invasive, cost-effective method for classifying liver disease between "non-fibrosis" (F0) and "fibrosis" (F1-F4), and to evaluate the classification performance quantitatively.Image data from 75 patients who underwent a simultaneous liver biopsy and non-contrast CT examination were used for this study. Non-contrast CT image texture features such as wavelet-based features, standard deviation of variance filter, and mean CT number were calculated in volumes of interest (VOIs) positioned within the liver parenchyma. In addition, a combined feature was calculated using logistic regression with L2-norm regularization to further improve fibrosis detection. Based on the final pathology from the liver biopsy, the patients were labelled either as "non-fibrosis" or "fibrosis". Receiver-operating characteristic (ROC) curve, area under the ROC curve (AUROC), specificity, sensitivity, and accuracy were determined for the algorithm to differentiate between "non-fibrosis" and "fibrosis".The combined feature showed the highest classification performance with an AUROC of 0.86, compared to the wavelet-based feature (AUROC, 0.76), the standard deviation of variance filter (AUROC, 0.65), and mean CT number (AUROC, 0.84). The combined feature's specificity, sensitivity, and accuracy were 0.66, 0.88, and 0.76, respectively, showing the most promising results.A new non-invasive and cost-effective method was developed to classify liver diseases between "non-fibrosis" (F0) and "fibrosis" (F1-F4). The proposed method makes it possible to detect liver fibrosis in asymptomatic patients using non-contrast CT images for better patient management and treatment.

Purpose We evaluated the diagnostic performance of the texture features of dynamic contrast-enhanced (DCE) MRI for breast cancer diagnosis in which the discriminator was optimized, so that the specificity was maximized via the restriction of the negative predictive value (NPV) to greater than 98%. Methods Histologically proven benign and malignant mass lesions of DCE MRI were enrolled retrospectively. Training and testing sets consist of 166 masses (49 benign, 117 malignant) and 50 masses (15 benign, 35 malignant), respectively. Lesions were classified via MRI review by a radiologist into 4 shape types: smooth (S-type, 34 masses in training set and 8 masses in testing set), irregular without rim-enhancement (I-type, 60 in training and 14 in testing), irregular with rim-enhancement (R-type, 56 in training and 22 in testing), and spicula (16 in training and 6 in testing). Spicula were immediately classified as malignant. For the remaining masses, 298 texture features were calculated using a parametric map of DCE MRI in 3D mass regions. Masses were classified into malignant or benign using two thresholds on a feature pair. On the training set, several feature pairs and their thresholds were selected and optimized for each mass shape type to maximize specificity with the restriction of NPV > 98%. NPV and specificity were computed using the testing set by comparison with histopathologic results and averaged on the selected feature pairs. Results In the training set, 27, 12, and 15 texture feature pairs are selected for S-type, I-type, and R-type masses, respectively, and thresholds are determined. In the testing set, average NPV and specificity using the selected texture features were 99.0% and 45.2%, respectively, compared to the NPV (85.7%) and specificity (40.0%) in visually assessed MRI category-based diagnosis. Conclusion We, therefore, suggest that the NPV of our texture-based features method described performs similarly to or greater than the NPV of the MRI category-based diagnosis.