Supervised Domain Adaptation for Automated Semantic Segmentation of the Atrial Cavity.

Atrial fibrillation (AF) is the most common cardiac arrhythmia. At present, cardiac ablation is the main treatment procedure for AF. To guide and plan this procedure, it is essential for clinicians to obtain patient-specific 3D geometrical models of the atria. For this, there is an interest in automatic image segmentation algorithms, such as deep learning (DL) methods, as opposed to manual segmentation, an error-prone and time-consuming method. However, to optimize DL algorithms, many annotated examples are required, increasing acquisition costs. The aim of this work is to develop automatic and high-performance computational models for left and right atrium (LA and RA) segmentation from a few labelled MRI volumetric images with a 3D Dual U-Net algorithm. For this, a supervised domain adaptation (SDA) method is introduced to infer knowledge from late gadolinium enhanced (LGE) MRI volumetric training samples (80 LA annotated samples) to a network trained with balanced steady-state free precession (bSSFP) MR images of limited number of annotations (19 RA and LA annotated samples). The resulting knowledge-transferred model SDA outperformed the same network trained from scratch in both RA (Dice equals 0.9160) and LA (Dice equals 0.8813) segmentation tasks.

Incidence, risk factors, and thrombotic load of pulmonary embolism in patients hospitalized for COVID-19 infection.

OBJECTIVE: To determine the incidence, characteristics, and risk factors of pulmonary embolism (PE) among patients hospitalized for COVID-19. PATIENTS AND METHODS: We performed a prospective observational study of a randomly selected cohort of consecutive patients hospitalized for COVID-19 infection between March 8, 2020 through April 25, 2020. All eligible patients underwent a computed tomography pulmonary angiography independently of their PE clinical suspicion and were pre-screened for a baseline elevated D-dimer level. RESULTS: 119 patients were randomly selected from the 372 admitted to one tertiary hospital in Valencia (Spain) for COVID-19 infection during the period of study. Seventy-three patients fulfilled both the inclusion criteria and none of the exclusion criteria and were finally included in the study. Despite a high level of pharmacological thromboprophylaxis (89%), the incidence of PE was 35.6% (95% confidence interval [CI], 29.6 to 41.6%), mostly with a peripheral location and low thrombotic load (Qanadli score 18.5%). Multivariate analysis showed that heart rate (Hazard Ratio [HR], 1.04), room-air oxygen saturation (spO2) (HR, 0.87), D-dimer (HR, 1.02), and C-reactive protein (CRP) levels (HR, 1.01) at the time of admission were independent predictors of incident PE during hospitalization. A risk score was constructed with these four variables showing a high predictive value of incident PE (AUC-ROC: 0.86; 95% CI: 0.80 to 0.93). CONCLUSIONS: Our findings confirmed a high incidence of PE in hospitalized COVID-19 patients. Heart rate, spO2, D-dimer, and CRP levels at admission were associated with higher rates of PE during hospitalization.

Giant Circumflex Aneurysm With Arteriovenous Fistula Surgery Guided by Three-Dimensional Printing.

A giant coronary artery aneurysm is an infrequent anomaly encountered as an incidental finding. An even more rare condition is its association with a coronary artery fistula. We report the case of a 61-year-old woman that combines two of the rarest coronary anomalies: a 70-mm giant circumflex aneurysm with an arteriovenous fistula tract draining to the superior vena cava. Considering the unique display of anatomy, the surgery was guided with 3-dimensional printing technology. Direct fistula occlusion near its outflow union, orifice source closure, aneurysmal sac exclusion, and two coronary artery bypass grafts were surgically performed. The postoperative course was satisfactory.