Classification of Echocardiogram View using A Convolutional Neural Network

The standard views in echocardiography capture distinct slices of the heart which can be used to assess cardiac function. Determining the view of a given echocardiogram is the first step for analysis. To automate this step, a deep network of the ResNet-18 architecture was used to classify between six standard views. The network parameters were pre-trained with the ImageNet database and prediction quality was assessed with a visualization tool known as gradient-weighted class activation mapping (Grad-CAM). The network was able to distinguish between three parasternal short axis views and three apical views to ~99\% accuracy. 10-fold cross validation showed a 97\%-98\% accuracy for the apical view subcategories (which included apical two-, three-, and four- chamber views). Grad-CAM images of these views highlighted features that were similar to those used by experts in manual classification. Parasternal short axis subcategories (which included apex level, mitral valve level, and papillary muscle level) had accuracies of 54\%-73\%. Grad-CAM images illustrate that the network classifies most parasternal short axis views as belonging to the papillary muscle level. Likely more images and incorporating time-dependent features would increase the parasternal short axis view accuracy. Overall, a convolutional neural network can be used to reliably classify echocardiogram views.

Single-View Echocardiography by Nonexpert Practitioners to Detect Rheumatic Heart Disease: A Prospective Study of Diagnostic Accuracy

Background: Echocardiographic screening can detect asymptomatic cases of rheumatic heart disease (RHD), facilitating access to treatment. Barriers to implementation of echocardiographic screening include the requirement for expensive equipment and expert practitioners. We aimed to evaluate the diagnostic accuracy of an abbreviated echocardiographic screening protocol (single parasternal-long-axis view with a sweep of the heart) performed by briefly trained, nonexpert practitioners using handheld ultrasound devices. Methods: Participants aged 5 to 20 years in Timor-Leste and the Northern Territory of Australia had 2 echocardiograms: one performed by an expert echocardiographer using a GE Vivid I or Vivid Q portable ultrasound device (reference test), and one performed by a nonexpert practitioner using a GE Vscan handheld ultrasound device (index test). The accuracy of the index test, compared with the reference test, for identifying cases with definite or borderline RHD was determined. Results: There were 3111 enrolled participants; 2573 had both an index test and reference test. Median age was 12 years (interquartile range, 10–15); 58.2% were female. Proportion with definite or borderline RHD was 5.52% (95% CI, 4.70–6.47); proportion with definite RHD was 3.23% (95% CI, 2.61–3.98). Compared with the reference test, sensitivity of the index test for definite or borderline RHD was 70.4% (95% CI, 62.2–77.8), specificity was 78.1% (95% CI, 76.4–79.8). Conclusions: Nonexpert practitioners can be trained to perform single parasternal-long-axis view with a sweep of the heart echocardiography. However, the specificity and sensitivity are inadequate for echocardiographic screening. Improved training for nonexpert practitioners should be investigated.

Complications during Ultrasound Guided Internal Jugular Vein Cannulation in ICU and OT Patients

This study was carried out in 80 patients admitted in the ICU and OT of IGMC Shimla. They were divided into 2 groups of 40 patients each. Internal jugular venous cannulations were done using ultrasound guided techniques (short axis and long axis view). The complications encountered during cannulation were noted. All the patients from both the groups were successfully cannulated by the operator under ultrasound guidance. Artery punctures occurred in 3 patients which was successfully managed and the cannulations were reattempted and were successful. None of the cannulation was abandoned. Keywords: ultrasound guided internal jugular vein cannulation, central venous pressure, internal jugular vein cannulation

Comparing Time Taken for Cannulating Internal Jugular Vein by USG Guided Approach: Short Axis versus Long Axis Technique

We have conducted our study in 80 patients admitted in the general ICU, requiring internal jugular vein cannulation. We formed 2 groups of 40 patients each that is:- Group 1 (short axis) and Group 2 (long axis). Under USG guidance, we cannulated the internal jugular vein with short axis view in group 1 and long axis view in group 2. Meanwhile the time taken to perform these cannulations was noted and then compared. After the study, we found that the internal jugular vein was cannulated much faster in short axis group as compared to the long axis group. Keywords: Central venous cannulation, internal jugular vein, USG guided approach, short axis versus long axis technique.

Distorted assessment of left atrial size by echocardiography in patients with increased aortic root diameter

Background Left atrial (LA) size is frequently assessed by posterior-anterior linear measurement of LA (LAD P-A) in the parasternal long axis to expedite examination. Aging, changes in body surface area, and several cardiovascular pathologies can affect aortic root (AoR) size, thereby affecting LA anatomical shape. We hypothesized that AoR dilatation influences LAD P-A and consequently correct assessment of LA size. Results We tested our hypothesis in a study of 70 patients with AoR diameter ranging from 2.7 to 4.8 cm. LA size assessed in parasternal long axis view as LAD P-A was compared to that with LA width and length acquired in the apical two and four chamber view. Simpson’s method of discs was used as standard measurement to assess LA volume. We observed that LAD P-A in the parasternal long axis decreases when AoR diameter increases. Thus, the increase in LA size assessed in parasternal long axis did not correlate with the increase of LA volume. Further analysis revealed that a significant positive correlation was observed when LAV was plotted as a function of LAD P-A only for those with a normal size AoR. In contrast, LA volume increase correlated with LA diameters assessed in the apical two and four chamber view regardless of AoR size. Conclusions Our study documents that increases in AoR impact on the linear measurement of LA, resulting in an underestimated LAD P-A. LA size ought to be calculated from the apical two and four chambers view parameters, especially in patients with AoR dilatation.

Prevalence of pericardial late gadolinium enhancement in patients after cardiac surgery: clinical and histological correlation

Funding Acknowledgements Type of funding sources: None. Background Opening of the pericardial sack during cardiac surgery induce usually, a mild inflammatory reaction. Late gadolinium enhancement of the pericardium (pLGE) still has been observed on cardiovascular magnetic resonance (CMR) in patients even long-time after cardiac surgery suggesting ongoing pericardial inflammation. Clinical relevance and histological correlation are unknown. We present a study evaluating the prevalence of pLGE and correlating it to clinical and histological findings. Materials and Methods 185 patients after cardiac surgery underwent CMR on a 1.5 or 3.0 T system. Presence (LGE+) or absence (LGE-) of pLGE was rated by 2 independent operators blinded to clinical characteristics. In case of discordance a third observer served as referee. Information on clinical symptoms were obtained at the time of CMR or from medical records. A pericardial biopsy was performed in 4 patients who underwent a second cardiac surgical intervention after CMR. Results Mean time between CMR and cardiac surgery was 158 ± 110 months. Pericardial LGE was observed in 83 patients (38%), two independent observers agreed in 73 (89%). The presence of LGE was not significant correlated to the type (p = 0.812) or duration of surgery (p = 0.734), nor the use of intrapericardial foreign material (p = 0.534). Two biopsies in LGE + patients showed mild inflammation and calcification, one biopsy in a LGE + patient showed the presence of fibrosis without inflammation while one biopsy in a LGE – patient was negative for inflammation. None of patients presented clinical signs for an active pericarditis. Discussion Presence of pericardial LGE is frequent in patient after cardiac surgery, however without clinical features of pericarditis. The CMR findings appear to be histologically correlated to the presence of fibrosis or mild chronic inflammation which remains to be confirmed in a larger patient population. Figure 1: Basal ventricular short axis view in LGE sequences showing the presence of LGE of the pericardium (Panel A, yellow arrows) compatible with mild inflammation present in histological findings, shown by the presence of T lymphocytes CD3+ (Panel B) and the presence of fibrin (Panel C) : (hematoxylin and eosin) and (Panel D) : (FAOG). Mid-ventricular short axis view in LGE sequences showing the presence of LGE of the pericardium (Panel E, yellow arrows), which is compatible with histological findings, showing the presence of granulomatous inflammation in a fibrinous pericardium (Panel F,H) : (hematoxylin and eosin) and (Panel G) : (fils polarized). Visible granuloma (Panel F).

Aortic valve regurgitation

Transthoracic echocardiography (TTE) is the first-line imaging tool to assess aortic valve (AV), aorta, and subsequent aortic regurgitation (AR). The parasternal long-axis view is classically used to measure the left outflow tract, the aortic annulus, and the aortic sinuses. Leaflet thickening and morphology can be visualized from this window as well as from the parasternal short-axis view and the apical five-chamber view. Nevertheless, 2D TTE may be limited and not enabling correct identification of the anatomy and causes of AR. In this situation, 3D echocardiography and cardiac magnetic resonance (CMR) could provide better delineation of the AV morphology. In some cases, transoesophageal echocardiography (TOE) could be required, more particularly for assessing the aortic root dimensions.

Ultrasound-guided arterial catheterization

Ultrasonography facilitates arterial catheterization compared to traditional palpation techniques, especially in small arteries. For successful catheterization without complications, practitioners should be familiar with the anatomic characteristics of the artery and ultrasound-guided techniques. There are two approaches for ultrasound-guided arterial catheterization: the short-axis view out-of-plane approach and the long-axis view in-plane approach. There are several modified techniques and tips to facilitate ultrasound-guided arterial catheterization. This review deals with the anatomy relevant to arterial catheterization, several methods to improve success rates, and decrease complications associated with arterial catheterization.