Deep learning to detect left ventricular structural abnormalities in chest X-rays.

BACKGROUND AND AIMS: Early identification of cardiac structural abnormalities indicative of heart failure is crucial to improving patient outcomes. Chest X-rays (CXRs) are routinely conducted on a broad population of patients, presenting an opportunity to build scalable screening tools for structural abnormalities indicative of Stage B or worse heart failure with deep learning methods. In this study, a model was developed to identify severe left ventricular hypertrophy (SLVH) and dilated left ventricle (DLV) using CXRs. METHODS: A total of 71 589 unique CXRs from 24 689 different patients completed within 1 year of echocardiograms were identified. Labels for SLVH, DLV, and a composite label indicating the presence of either were extracted from echocardiograms. A deep learning model was developed and evaluated using area under the receiver operating characteristic curve (AUROC). Performance was additionally validated on 8003 CXRs from an external site and compared against visual assessment by 15 board-certified radiologists. RESULTS: The model yielded an AUROC of 0.79 (0.76-0.81) for SLVH, 0.80 (0.77-0.84) for DLV, and 0.80 (0.78-0.83) for the composite label, with similar performance on an external data set. The model outperformed all 15 individual radiologists for predicting the composite label and achieved a sensitivity of 71% vs. 66% against the consensus vote across all radiologists at a fixed specificity of 73%. CONCLUSIONS: Deep learning analysis of CXRs can accurately detect the presence of certain structural abnormalities and may be useful in early identification of patients with LV hypertrophy and dilation. As a resource to promote further innovation, 71 589 CXRs with adjoining echocardiographic labels have been made publicly available.

Duplicated Prostate Artery Central Gland Blood Supply: A Retrospective Analysis and Classification System.

Angiograms and cone-beam computed tomography scans of 36 consecutive prostate artery embolization patients (72 hemipelves) between October 2014 and February 2018 were reviewed. The hemipelves were classified according to the presence of dual central gland (CG) blood supply and the pattern of vascularization: Type 1 with a single CG blood supply (83.3%; n = 60); Type 2 with 2 independent CG arteries with overlapping territories (9.7%; n = 7); and Type 3 with 2 independent CG arteries with isolated territories (7%; n = 5). Up to 20% of pelvic sides may have more than 1 independent CG prostate artery that should be searched for during prostate artery embolization.

Serial Uterine Artery Embolization for the Treatment of Placenta Percreta in the First Trimester: A Case Report.

Two patients with placenta percreta underwent uterine artery embolization (UAE) for abnormally invasive placenta (AIP) in the first trimester. Patient 1 had a 9-week cervical ectopic, while Patient 2 had a 9-week cesarean scar pregnancy. Elective termination of pregnancy was performed in both patients. UAE was performed with tris-acryl gelatin microspheres as well as gelfoam until stasis and was repeated in cases of revascularization. Both patients were followed with US/MRI/MRA scans and ß-hCG levels. Revascularization occurred in both patients following UAE, requiring multiple embolizations to achieve complete placental involution. Serial bland UAE may be an effective technique in the treatment of first-trimester AIP, with the distinct advantage of maintaining a patient's fertility. LEVEL OF EVIDENCE: Level IV.

Active hemorrhage into a postresection cavity detected by neuro-CT angiography.

We describe a case demonstrating active extravasation of contrast material into a hematoma resection cavity during CT angiography (CTA) that necessitated emergent reexploration, decompression, and hemostatic control. Our case highlights the value of neuro-CTA in the immediate postoperative setting and describes another scenario where CTA has added value. Prompt recognition of contrast extravasation is critical to the diagnosis and ultimately affects the quality of patient care.