An Edge-Based Selection Method for Improving Regions-of-Interest Localizations Obtained Using Multiple Deep Learning Object-Detection Models in Breast Ultrasound Images.

Computer-aided diagnosis (CAD) systems can be used to process breast ultrasound (BUS) images with the goal of enhancing the capability of diagnosing breast cancer. Many CAD systems operate by analyzing the region-of-interest (ROI) that contains the tumor in the BUS image using conventional texture-based classification models and deep learning-based classification models. Hence, the development of these systems requires automatic methods to localize the ROI that contains the tumor in the BUS image. Deep learning object-detection models can be used to localize the ROI that contains the tumor, but the ROI generated by one model might be better than the ROIs generated by other models. In this study, a new method, called the edge-based selection method, is proposed to analyze the ROIs generated by different deep learning object-detection models with the goal of selecting the ROI that improves the localization of the tumor region. The proposed method employs edge maps computed for BUS images using the recently introduced Dense Extreme Inception Network (DexiNed) deep learning edge-detection model. To the best of our knowledge, our study is the first study that has employed a deep learning edge-detection model to detect the tumor edges in BUS images. The proposed edge-based selection method is applied to analyze the ROIs generated by four deep learning object-detection models. The performance of the proposed edge-based selection method and the four deep learning object-detection models is evaluated using two BUS image datasets. The first dataset, which is used to perform cross-validation evaluation analysis, is a private dataset that includes 380 BUS images. The second dataset, which is used to perform generalization evaluation analysis, is a public dataset that includes 630 BUS images. For both the cross-validation evaluation analysis and the generalization evaluation analysis, the proposed method obtained the overall ROI detection rate, mean precision, mean recall, and mean F1-score values of 98%, 0.91, 0.90, and 0.90, respectively. Moreover, the results show that the proposed edge-based selection method outperformed the four deep learning object-detection models as well as three baseline-combining methods that can be used to combine the ROIs generated by the four deep learning object-detection models. These findings suggest the potential of employing our proposed method to analyze the ROIs generated using different deep learning object-detection models to select the ROI that improves the localization of the tumor region.

Breast Tumor Classification in Ultrasound Images Using Combined Deep and Handcrafted Features.

This study aims to enable effective breast ultrasound image classification by combining deep features with conventional handcrafted features to classify the tumors. In particular, the deep features are extracted from a pre-trained convolutional neural network model, namely the VGG19 model, at six different extraction levels. The deep features extracted at each level are analyzed using a features selection algorithm to identify the deep feature combination that achieves the highest classification performance. Furthermore, the extracted deep features are combined with handcrafted texture and morphological features and processed using features selection to investigate the possibility of improving the classification performance. The cross-validation analysis, which is performed using 380 breast ultrasound images, shows that the best combination of deep features is obtained using a feature set, denoted by CONV features that include convolution features extracted from all convolution blocks of the VGG19 model. In particular, the CONV features achieved mean accuracy, sensitivity, and specificity values of 94.2%, 93.3%, and 94.9%, respectively. The analysis also shows that the performance of the CONV features degrades substantially when the features selection algorithm is not applied. The classification performance of the CONV features is improved by combining these features with handcrafted morphological features to achieve mean accuracy, sensitivity, and specificity values of 96.1%, 95.7%, and 96.3%, respectively. Furthermore, the cross-validation analysis demonstrates that the CONV features and the combined CONV and morphological features outperform the handcrafted texture and morphological features as well as the fine-tuned VGG19 model. The generalization performance of the CONV features and the combined CONV and morphological features is demonstrated by performing the training using the 380 breast ultrasound images and the testing using another dataset that includes 163 images. The results suggest that the combined CONV and morphological features can achieve effective breast ultrasound image classifications that increase the capability of detecting malignant tumors and reduce the potential of misclassifying benign tumors.

MRI Evaluation of Mullerian Duct Anomalies: Practical Classification by the New ASRM System.

Objective: To describe variable mullerian duct anomalies using magnetic resonance imaging (MRI) and to classify these anomalies according to the available classification systems, namely the American Fertility Society (AFS) system, the European Society of Human Reproduction and Embryology (ESHRE) system, and the new American Society for Reproductive Medicine (ASRM) system. Design: Retrospective chart review. Subjects: The pelvic MRI studies and the clinical records of 64 females with mullerian congenital anomalies were retrospectively reviewed between January 2010 and December 2021. The mean age was 22 years (age range 2-63 years). Main Outcome Measures: Detailed imaging findings were recorded, and the resulting mullerian anomalies were then classified according to the three classification systems of interest. Results: Variable mullerian anomalies were found among patients with multiple frequencies. Mullerian agenesis and hypoplasia were found in 12 patients (19%) and 16 patients (25%), respectively. Uterus didelphys was found in 5 patients (8%). Twelve (19%) patients had septate uterus, while 8 (12.5%) had a bicornuate anomaly. Unicornuate uterus was present in 7 patients (11%). Isolated vaginal anomaly was diagnosed in 4 patients (6%). Renal/urinary tract imaging was available for 27 (42%) patients, and accompanying urinary tract anomalies were noted in 10 of them (37%). Few ovarian and other extra-renal anomalies were observed. Conclusion: MRI could efficiently delineate the mullerian anomalies regardless of their complexity. Most of these anomalies were more efficaciously categorized by the ESHRE and the new ASRM systems, compared to the originally widely used AFS system. The new ASRM classification was found to be more practical as it is a modification of the original AFS system, using drawings with clear descriptions instead of symbols. This is particularly helpful in the radiological era, saving time and effort.

Pseudoaneurysm of the breast after blunt trauma: successful treatment with ultrasound-guided compression.

We present a case of breast pseudoaneurysm following a blunt trauma in a 58-year-old woman. Few cases of breast pseudoaneurysm have been reported in the literature, and most of these are related to previous interventional procedures. Pseudoaneurysm was suspected on real-time sonography and confirmed with color Doppler and spectral wave analysis, which revealed a characteristic to-and-fro pattern. Unlike previously reported cases, treatmentwith ultrasound-guided compression was successful.

The usefulness of MR defecography in the evaluation of pelvic floor dysfunction: our experience using 3T MRI.

PURPOSE: To assess the usefulness of MR defecography in evaluating pelvic floor dysfunction, and to correlate several pelvic organ abnormalities with each other and with patients' symptoms and characteristics. METHODS: MR defecographic examinations performed in 3T MRI machine of 95 patients (70 females, 25 males; mean age 48) were retrospectively reviewed. Pelvic organ abnormalities from all three compartments were recorded, including the anorectal junction descent, anterior rectocele, and cystocele. These were graded according to the known HMO system in relation to the pubococcygeal line. The correlation between these different abnormalities and their relation to patient symptoms and characteristics were evaluated. RESULTS: Anorectal junction descent and anterior rectocele were most commonly observed, predominantly manifesting in female patients. Both were associated with abnormalities from all compartments. The middle compartment was the least affected, and its abnormality of uterine/vaginal descent tended to occur in association with the anterior compartment abnormality (cystocele). Anismus was low in incidence, and was not associated with other compartments abnormalities. Both enterocele/peritoneocele and intussusception were uncommon. CONCLUSION: MR defecography is the modality of choice in assessing pelvic floor dysfunction, because it can neatly show various pelvic organ abnormalities from all compartments in a dynamic fashion, which are frequently coexistent. It can even show clinically silent or unsuspected abnormalities which can impact the management of patients.

Arachnoid granulations in the cerebral dural sinuses as demonstrated by contrast-enhanced 3D magnetic resonance venography.

INTRODUCTION: Identification of normal filling defects within the intracranial dural sinuses reduces the erroneous diagnosis of the presence of an intrasinus pathologic process. The aim of this prospective study was to assess the prevalence, distribution, and morphological characteristics of arachnoid granulations (AGs) in the dural sinuses. METHODS: This prospective study was carried out on 110 patients who had both normal conventional brain MRI and contrast-enhanced (CE) 3D turbo flash magnetic resonance venography (MRV). The dural sinuses were viewed on MRV images for the presence of filling defects. The prevalence, site, size, number, shape, outlines, internal structure, and presence of associated cortical vein were determined. RESULTS: One hundred and twenty-six AGs were observed among 71 patients. The superior sagittal sinus was the most common site of filling defects (58 AGs). The mean size of AGs was 6.45 +/- 3.55 mm. Eighty-three percent of AGs were round or oval, with sharp outlines and homogeneous internal structure; of these 81% were associated with cortical vein. CONCLUSIONS: In the majority of cases, the identification of AGs can be facilitated by their characteristic appearances: rounded or oval shaped, well-defined outlines and homogenous intensity. The presence of an adjacent cortical vein can be considered as an additional supportive element.