The segmentation and characterization of the lung lobes are important tasks for Computer Aided Diagnosis (CAD) systems related to pulmonary disease. The detection of the fissures that divide the lung lobes is non-trivial when using classical methods that rely on anatomical information like the localization of the airways and vessels. This work presents a fully automatic and supervised approach to the problem of the segmentation of the five pulmonary lobes from a chest Computer Tomography (CT) scan using a Fully Regularized V-Net (FRV-Net), a 3D Fully Convolutional Neural Network trained end-toend. Our network was trained and tested in a custom dataset that we make publicly available. It can correctly separate the lobes even in cases when the fissure is not well delineated, achieving 0.93 in per-lobe Dice Coefficient and 0.85 in the inter-lobar Dice Coefficient in the test set. Both quantitative and qualitative results show that the proposed method can learn to produce correct lobe segmentations even when trained on a reduced dataset.

Pulmonary epithelioid hemangioendothelioma (PEH) is a rare neoplasm of vascular origin. There are three different major imaging patterns identified in thoracic manifestation of epithelioid hemangioendothelioma: (1) multiple pulmonary nodules; (2) multiple pulmonary reticulonodular opacities; and (3) diffuse infiltrative pleural thickening. Radiographically, presence of bilateral multiple nodules is the most common pattern of presentation. The diagnosis is made on the basis of histopathological findings and confirmed by positive immunohistochemistry staining. Although the prognostic factors for PEH have not yet been well established, a better prognosis is usually associated with the multinodular pattern. We report two different imagological presentations of this rare disease, based on two institutional experiences, along with a review of the relevant literature.

The Abernethy malformation consists of a congenital extrahepatic portosystemic shunt and is believed to be extremely rare in humans. The potential implications of abnormal portovenous shunting and decreased hepatic portal flow are numerous and potentially serious. Although congenital extrahepatic portosystemic shunts are increasingly suspected and diagnosed in specialized centres, much of their clinical presentation and natural history is not fully understood. Symptoms of portosystemic shunt are mainly caused by increased levels of ammonia, which lead to signs of encephalopathy. Therapeutic options depend on the type of shunt and its clinical course, so the classification of the congenital portosystemic shunt is a key finding in these patients.

Staphylococcus aureus is both a human commensal and a pathogen, that causes serious nosocomial and community-acquired infections. Despite nostrils being considered its preferred host habitat, the oral cavity has been demonstrated to be an ideal starting point for auto-infection and transmission. The antibiotic resistance assessment of S. aureus is a priority and is often reported in clinical settings. This study aimed to explore the prevalence and antimicrobial susceptibility of S. aureus in the oral and nasal cavities of healthy individuals. The participants (n = 101) were subjected to a demographic and clinical background survey, a caries evaluation, and to oral and nasal swabbing. Swabs were cultured in differential/selective media and S. aureus isolates were identified (MALDI-TOF MS) and tested for antibiotic susceptibility (EUCAST/CLSI). Similar S. aureus prevalence was found exclusively on nasal (13.9%) or oral (12.0%) habitats, whereas 9.9% of the population were simultaneous nasal and oral carriers. In oro-nasal cavities, similar antibiotic resistance rates (83.3–81.5%), including MDR (20.8–29.6%), were observed. Notably, 60% (6/10) of the simultaneous nasal and oral carriers exhibited different antibiotic resistance profiles between cavities. This study demonstrates the relevance of the oral cavity as an independent colonization site for S. aureus and as a potential source of antimicrobial resistance, a role which has been widely neglected so far.

A 33-year-old female smoker came to our emergency department complaining of dyspnea, pleuritic left chest pain, and an episode of mild hemoptysis. One week previously, she had experienced left calf pain. There was no history of recent trauma or immobilization. Her only medication was an oral contraceptive. On admission, her heart rate was 110 beats per minute in sinus rhythm, with a normal physical examination and arterial blood gases and without evidence of right ventricular strain on ECG (Figure 1). A chest x-ray showed a wedgeshaped peripheral opacity in the lower half of the left lung field (Hampton's hump; Figure 2, long arrow) and an enlarged right descending pulmonary artery (Palla's sign; Figure 2, short arrow). Plasma D-dimers were elevated (2.3 μg/mL; normal value, <0.5 μg/mL). The patient had a chest computed tomography contrast scan that showed luminal filling defects in the left pulmonary artery (Figure 3, asterisk) and right segmental pulmonary arteries, as well as a consolidation area in the left lower lobe probably associated with lung infarction.The patient had normal cardiac biomarkers and a normal transthoracic echocardiogram. Anticoagulation was initiated, and the patient was admitted to our intermediate care unit.Although contrast-enhanced thoracic computed tomography scan allows a rapid and straightforward diagnosis of pulmonary embolism, older radiological findings should not be forgotten. The Hampton's hump, described in 1940 by the radiologist Aubrey Hampton, consists of a wedge-shaped opacity of the peripheral lung field.1 The Palla's sign was named after the description by Antonio Palla of an association between pulmonary embolism and a chest x-ray sign of right descending pulmonary artery enlargement.2 The simultaneous presence of these findings in chest x-ray is infrequent 3 but is a valuable aid in the diagnosis of pulmonary embolism, especially in patients at risk of complications with contrastenhanced imaging techniques. DisclosuresNone.