Search

Purpose Machine learning techniques, especially convolutional neural networks (CNN), have revolutionized the spectrum of computer vision tasks with a primary focus on supervised and labelled image datasets. We aimed to assess a novel method to segment the liver from the abdomen computed tomography (CT) image using the CNN network, and to train a unique method to locate and classify liver lesion pre-histological findings using multi-channel deep learning CNN (MDL-CNN). Material and methods The post-contrast CT images of the liver with a resolution of 0.625 mm were chosen for the study. In a random method, 50 examples of each hepatocellular carcinomas, metastases tumours, haemangiomas, hepatic cysts were chosen and evaluated. Results The dice score quantitatively analyses the similarity of segmentation results with the training dataset. In the first CNN model for segmenting the liver, the dice score was 96.18%. The MDL-CNN model yielded 98.78% accuracy in classification, and the dice score for locating liver lesions was 95.70%. Additionally, the performance of this model was compared to various other existing models. Conclusions According to our study, the machine learning approach can be successfully implemented to segment the liver and classify lesions, which will help radiologists impart better diagnosis.

Summary Background Infections of odontogenic origin are the most common cause of inflammatory disease of head and neck region. Computed tomography allows for defining localization and extent of inflammatory lesions, visualizes soft tissue involvement, presence of an abscess or an osteolytic lesion around causative tooth. The aim of this study was to assess pathways, by which odontogenic infections spread into respective deep head and neck structures in computed tomography examination, taking into account the following criteria: frequency of involvement of respective deep cervical spaces, possibility to determine a probable causative tooth and concordance with the results of clinical examination. Material/Methods Thirty-eight patients cervicofacial inflammatory disease had undergone CT examination of head and neck region with a 64-slice CT scanner after intravenous contrast administration. Results Abscess was reported in 30 (79%) cases, while inflammatory infiltration was diagnosed in remaining 8 (21%) patients. There was full concordance between radiological report and intraoperative report In 33 cases (87%). The most commonly involved cervical space was masticator space – 31 patients (82%), followed by submandibular space – 27 patients (71%). Dental examination was impossible in 29 patient because of trismus. During analysis of CT studies we evaluated maxillary and mandibular alveolar processes for presence of osteolytic bone lesions around causative teeth roots and we found them in 30 cases (79%). In 32 cases (84%) cervicofacial infection were of mandibular odontogenic origin. Conclusions In most cases CT study in patients suspected of odontogenic craniofacial infection revealed presence of an abscess, needing urgent surgery. Inflammatory infiltration of dental origin most frequently involves masticator space, followed by submandibular space. In most cases CT scanning allows for identification of causative teeth, especially when trismus makes detailed clinical examination impossible.

BACKGROUND Dolichoectasia of intracranial arteries is a rare arteriopathy characterized by elongation and widening of the arteries and disturbance of the laminar blood flow. It involves mostly vertebral and basilar arteries. In advanced cases, formation of a fusiform aneurysm is possible. CASE REPORT A sixty-four-year-old female with hypertension was admitted to the hospital with severe non-systemic vertigo and dysarthria, which had lasted for a couple of weeks. Imaging of the brain revealed dolichoectasia of arteries of the circle of Willis coexisting with a fusiform aneurysm of the basilar artery. CONCLUSIONS Intracranial arterial dolichoectasia may be asymptomatic for a long time. However, in many cases it leads to neurological symptoms associated with haemodynamic disturbance (due to unstable wall clots) and mass effect caused by the widened vessel.

Summary Background: Takayasu arteritis is an inflammatory disease of large-diameter arteries. Aorta and its branches are most frequently affected. Takayasu arteritis occurs mainly in young women and, if left untreated, leads to fatal complications. Digital subtraction angiography (DSA) is considered the gold standard in imaging of Takayasu arteritis. Case Report: A thirty-five-year-old woman was admitted to the hospital with transient loss of consciousness, effort-associated vertigo, upper limb weakness and temporary vision problems. On admission, there was no pulse on the left radial artery while there were bruits over subclavian arteries. Imaging of the aortic arch (computed tomography angiography, DSA) revealed stenoses of its main branches, indicating Takayasu arteritis. Conclusions: Computed tomography angiography (CTA) performed with a 64-slice unit revealed high effectiveness in localization of vascular wall and lumen pathologies resulting from Takayasu arteritis. Thanks to this fast diagnostic method, it is now possible to perform successful monitoring of patients with Takayasu arteritis and to plan possible interventional treatment.