Jeremy Burt

CardiacNET: Segmentation of left atrium and proximal pulmonary veins from MRI using multi-view CNN

Anatomical and biophysical modeling of left atrium (LA) and proximal pulmonary veins (PPVs) is important for clinical management of several cardiac diseases. Magnetic resonance imaging (MRI) allows qualitative assessment of LA and PPVs through visualization. However, there is a strong need for an advanced image segmentation method to be applied to cardiac MRI for quantitative analysis of LA and PPVs. In this study, we address this unmet clinical need by exploring a new deep learning-based segmentation strategy for quantification of LA and PPVs with high accuracy and heightened efficiency. Our approach is based on a multi-view convolutional neural network (CNN) with an adaptive fusion strategy and a new loss function that allows fast and more accurate convergence of the backpropagation based optimization. After training our network from scratch by using more than 60K 2D MRI images (slices), we have evaluated our segmentation strategy to the STACOM 2013 cardiac segmentation challenge benchmark. Qualitative and quantitative evaluations, obtained from the segmentation challenge, indicate that the proposed method achieved the state-of-the-art sensitivity (90%), specificity (99%), precision (94%), and efficiency levels (10s in GPU, and 7.5 min in CPU).

Multi-Planar Deep Segmentation Networks for Cardiac Substructures from MRI and CT

Non-invasive detection of cardiovascular disorders from radiology scans requires quantitative image analysis of the heart and its substructures. There are well-established measurements that radiologists use for diseases assessment such as ejection fraction, volume of four chambers, and myocardium mass. These measurements are derived as outcomes of precise segmentation of the heart and its substructures. The aim of this paper is to provide such measurements through an accurate image segmentation algorithm that automatically delineates seven substructures of the heart from MRI and/or CT scans. Our proposed method is based on multi-planar deep convolutional neural networks (CNN) with an adaptive fusion strategy where we automatically utilize complementary information from different planes of the 3D scans for improved delineations. For CT and MRI, we have separately designed three CNNs (the same architectural configuration) for three planes, and have trained the networks from scratch for voxel-wise labeling for the following cardiac structures: myocardium of left ventricle (Myo), left atrium (LA), left ventricle (LV), right atrium (RA), right ventricle (RV), ascending aorta (Ao), and main pulmonary artery (PA). We have evaluated the proposed method with 4-fold-cross-validation on the multi-modality whole heart segmentation challenge (MM-WHS 2017) dataset. A precision and dice index of 0.93 and 0.90, and 0.87 and 0.85 were achieved for CT and MR images, respectively. Cardiac CT volume was segmented in about 50 s, with cardiac MRI segmentation requiring around 17 s with multi-GPU/CUDA implementation.

Invasive Cardiac Lipoma: a case report and review of literature

BackgroundCardiac lipomas are rare benign tumors of the heart. They are usually asymptomatic and are thus most often diagnosed on autopsies. Symptoms, when present, depend upon the location within the heart. Typical locations are the endocardium of the right atrium and the left ventricle. Diagnostic modality of choice is cardiac MRI. Treatment guidelines have not yet been established due to the very low prevalence of these tumors and are thus guided by the patient’s symptomatology.Case presentationWe describe a case of an invasive cardiac lipoma, wherein the initial symptom of the patient was shortness of breath. Although the echocardiogram visualized the tumor in the right atrium, a cardiac MRI was performed for better tissue characterization. The MRI revealed a large, fat containing, septated mass in the right atrium with invasion into the inter-atrial septum and inferior cavoatrial junction. There was also invasion of the coronary sinus along the inferior and left lateral aspect of the posterior atrioventricular groove. Although the mass appeared to represent a lipoma by imaging characteristics, the unusual extension into the coronary sinus led to consideration of a low-grade liposarcoma in the differential. Thus a pre-operative biopsy was performed along with MDM2 gene amplification to rule out a liposarcoma preceding surgical excision.ConclusionCardiac lipomas are well-characterized on cardiac MRI, which is the diagnostic modality of choice. Typical locations are the right atrium and the left ventricle. However, in those with atypical features such as invasion of the coronary sinus, pre-operative biopsy for histopathologic confirmation is imperative to exclude well-differentiated liposarcoma. Our patient with a simple lipoma underwent partial resection to relieve symptoms. We discuss prognosis and treatment of symptomatic cardiac lipomas.

Concurrent Multiple Myeloma and Metastatic Osteosarcoma: A Case Report and Literature Review

Multiple myeloma (MM) and osteosarcoma (OS) are two common bone malignancies, however, the simultaneous occurrence of both primary bone tumors in the same patient has not been reported in the United States to date. We present a unique case in which both malignancies present concurrently in a 72-year-old man. Results of spinal magnetic resonance imaging (MRI), radiographic skeletal survey, and hematological workup established the initial diagnosis of MM. Approximately three months later, the patient was admitted with severe right hip pain and shortness of breath and was evaluated with computed tomography (CT) of the right hip, abdomen, pelvis, and chest, revealing an osseous mass with a “sunburst” pattern in the right hip, and several calcified nodules in the lungs. Subsequent wedge resection and histological evaluation of the lung nodules confirmed the diagnosis of metastatic OS to the lungs, with a presumptive diagnosis of primary OS of the right hip. The clinical findings and imaging characteristics in this case are presented. Two similar cases found in the literature are also briefly discussed. The findings of this case report suggest that, in rare instances, MM patients with sclerotic bone findings may have a concurrent diagnosis of OS.

CT-guided Native Medical Renal Biopsy: Cortical Tangential versus Non-Tangential Approaches—A Comparison of Efficacy and Safety

Purpose To review a single-center experience with the cortical tangential approach during computed tomography (CT)-guided native medical renal biopsy and to evaluate its efficacy and safety compared with those of a non-cortical tangential approach. Materials and Methods This retrospective study received institutional review board approval, with a waiver of the HIPAA requirement for informed consent. The number of cores, glomeruli, and complications were reviewed in 431 CT-guided medical renal biopsies performed between July 2007 and September 2015. A biopsy followed a cortical tangential approach if the needle path was parallel to the renal cortical surface, at a depth closer to the renal capsule than the renal pelvic fat. A sample was considered adequate if the biopsy yielded at least 10 glomeruli at light microscopy, one glomerulus at immunofluorescence microscopy, and one glomerulus at electron microscopy. The χ2 test, the t test, the Mann-Whitney test, and logistic regression modeling of sample adequacy were performed. Results One hundred fifty-six (36%) of 431 biopsies were performed with the cortical tangential approach. More cores were obtained for the cortical tangential group (2.6 vs 2.4, P = .001); biopsy needle gauge was not significantly different (P = .076). More adequate samples were obtained in the cortical tangential group (66.7% vs 49.8%, P = .001), with more glomeruli (23 vs 16, P = .014). Results were significant after controlling for needle gauge and number of cores (P = .008). The cortical tangential group had fewer complications (1.9% vs 7.3%, P = .018). Conclusion The cortical tangential approach, when applied to CT-guided native medical renal biopsies, results in higher rates of sample adequacy and lower rates of postprocedural complications.

A Novel Imaging Finding in Williams Syndrome: The Coral Sign

A 16-year-old female, with a history of Williams syndrome, presented to our institution with a 2-week history of intermittent dizziness. Holter monitoring demonstrated occasional premature ventricular contractions with rare couplets and triplets as well as one short run of nonsustained ventricular tachycardia. Echocardiography revealed an abnormal and irregular left ventricular septum with multiple mobile, pedunculated muscular projections extending into the left ventricular cavity. Cardiac MR confirmed abnormally thickened trabeculations consisting of multiple parallel ridges of myocardium crossing the left ventricle. The appearance of these findings closely resembled bands of coral lining the ocean floor. As such, this finding can henceforth be known as the “coral sign.” To our knowledge, no other reports of this finding in patients with Williams syndrome have been published.

Catastrophic Prosthetic Valve Endocarditis Caused by Rare Black Fungi

Fungal infection of prosthetic heart valves is rare and can lead to severe complications including death. Dematiaceous mold, also known as “black fungi,” are an extremely rare cause of endocarditis that usually affect immunocompromised hosts. The infection is usually chronic and can lead to heart failure and embolic complications. These fungi have limited antifungal treatment modalities. We present a rare case of prosthetic aortic valve, root, and graft infection in an immunocompetent host that revealed itself through renal, mesenteric, and cerebral embolic phenomenon. The patient underwent removal and replacement of the aortic graft followed by small bowel resection for mesenteric infarction. Patient had a successful postoperative course and underwent a long-term antifungal treatment with amphotericin B and voriconazole.

Unusual Vascular Manifestations of Noonan Syndrome

Noonan syndrome is a relatively common autosomal dominant disorder. We present a patient with Noonan syndrome and multiple cardiac and vascular manifestations, some of which are unusual and rarely reported. Further research is necessary to determine whether these defects are truly secondary to Noonan syndrome or possibly another underlying congenital abnormality.

Safety of CT-Guided Bone Marrow Biopsy in Thrombocytopenic Patients: A Retrospective Review

PURPOSE To test the hypothesis that computed tomography (CT)-guided bone marrow biopsy in patients with a platelet count between 20,000/uL and 50,000/uL is safe and that preprocedure platelet transfusion is unnecessary. MATERIALS AND METHODS This single-center retrospective study included bone marrow biopsies performed between May 2009 and May 2016. The study population included 981 patients-age range, 15-93 years; average age, 57 years; 505 (51.5%) men; and 476 (48.5%) women. One hundred eighty-seven biopsies were performed in patients with a platelet count of 20,000-50,000/μL; 33 were performed in patients with a platelet count of < 20,000/μL. The primary endpoint was hemorrhagic complications, Society of Interventional Radiology (SIR) complication class C or above. The complication rates in thrombocytopenic patients were compared to patients with a platelet count of ≥ 50,000/uL. Ninety-five percent confidence intervals (CIs) for the complication rate in each group were also calculated. RESULTS There were no SIR class C or above postprocedure bleeding-related complications, including interventions or transfusions. For patients with a platelet count of < 20,000/μL and of 20,000-50,000/μL, hemorrhagic complications rates were 0% (95% CI: 0-9.1%) and 0% (95% CI: 0-1.6%), respectively. CONCLUSIONS CT-guided bone marrow biopsy is safe in thrombocytopenic patients, with a hemorrhagic complication rate below 1.6% for patients with a platelet count of 20,000-50,000/μL. Routine preprocedure platelet transfusion may not be necessary for patients with a platelet count of 20,000-50,000/μL.

Abdominal Cocoon Syndrome: a Rare Cause for Recurrent Abdominal Pain

Abdominal cocoon syndrome, or sclerosing peritonitis, is a rare condition characterized by encasement of small bowel loops by a thick fibrous scar. It most commonly presents as nonspecific vague chronic abdominal pain and weight loss, and is difficult to recognize clinically until the patient develops symptoms of bowel obstruction. We present a case of abdominal cocoon syndrome in a 65-year-old female and describe its clinical, imaging, and pathologic features.