Student Ultrasound Education, Current Views and Controversies; Who Should be Teaching?

Acquiring diagnostic ultrasound competencies and skills is crucial in modern health care, and achieving the practical experience is vital in developing the necessary anatomy interpretation and scan acquisition skills. However, traditional teaching methods may not be sufficient to provide hands-on practice, which is essential for this skill acquisition. This paper explores various modalities and instructors involved in ultrasound education to identify the most effective approaches. The field of ultrasound instruction is enriched by the diverse roles of physicians, anatomists, peer tutors, and sonographers. All these healthcare professionals can inspire and empower the next generation of ultrasound practitioners with continuous training and support. Physicians bring their clinical expertise to the table, while anatomists enhance the understanding of anatomical knowledge through ultrasound integration. Peer tutors, often medical students, provide a layer of social congruence and motivation to the learning process. Sonographers provide intensive practical experience and structured learning plans to students. By combining different instructors and teaching methods, success can be achieved in ultrasound education. An ultrasound curriculum organized by experts in the field can lead to more efficient use of resources and better learning outcomes. Empowering students through peer-assisted learning can also ease the burden on faculty. Every instructor must receive comprehensive didactic training to ensure high-quality education in diagnostic ultrasound.

Student ultrasound education - current views and controversies.

As an extension of the clinical examination and as a diagnostic and problem-solving tool, ultrasound has become an established technique for clinicians. A prerequisite for high-quality clinical ultrasound practice is adequate student ultrasound training. In light of the considerable heterogeneity of ultrasound curricula in medical studies worldwide, this review presents basic principles of modern medical student ultrasound education and advocates for the establishment of an ultrasound core curriculum embedded both horizontally and vertically in medical studies.

Deep Learning Pitfall: Impact of Novel Ultrasound Equipment Introduction on Algorithm Performance and the Realities of Domain Adaptation.

OBJECTIVES: To test deep learning (DL) algorithm performance repercussions by introducing novel ultrasound equipment into a clinical setting. METHODS: Researchers introduced prospectively obtained inferior vena cava (IVC) videos from a similar patient population using novel ultrasound equipment to challenge a previously validated DL algorithm (trained on a common point of care ultrasound [POCUS] machine) to assess IVC collapse. Twenty-one new videos were obtained for each novel ultrasound machine. The videos were analyzed for complete collapse by the algorithm and by 2 blinded POCUS experts. Cohen's kappa was calculated for agreement between the 2 POCUS experts and DL algorithm. Previous testing showed substantial agreement between algorithm and experts with Cohen's kappa of 0.78 (95% CI 0.49-1.0) and 0.66 (95% CI 0.31-1.0) on new patient data using, the same ultrasound equipment. RESULTS: Challenged with higher image quality (IQ) POCUS cart ultrasound videos, algorithm performance declined with kappa values of 0.31 (95% CI 0.19-0.81) and 0.39 (95% CI 0.11-0.89), showing fair agreement. Algorithm performance plummeted on a lower IQ, smartphone device with a kappa value of -0.09 (95% CI -0.95 to 0.76) and 0.09 (95% CI -0.65 to 0.82), respectively, showing less agreement than would be expected by chance. Two POCUS experts had near perfect agreement with a kappa value of 0.88 (95% CI 0.64-1.0) regarding IVC collapse. CONCLUSIONS: Performance of this previously validated DL algorithm worsened when faced with ultrasound studies from 2 novel ultrasound machines. Performance was much worse on images from a lower IQ hand-held device than from a superior cart-based device.

Making Artificial Intelligence Lemonade Out of Data Lemons: Adaptation of a Public Apical Echo Database for Creation of a Subxiphoid Visual Estimation Automatic Ejection Fraction Machine Learning Algorithm.

OBJECTIVES: A paucity of point-of-care ultrasound (POCUS) databases limits machine learning (ML). Assess feasibility of training ML algorithms to visually estimate left ventricular ejection fraction (EF) from a subxiphoid (SX) window using only apical 4-chamber (A4C) images. METHODS: Researchers used a long-short-term-memory algorithm for image analysis. Using the Stanford EchoNet-Dynamic database of 10,036 A4C videos with calculated exact EF, researchers tested 3 ML training permeations. First, training on unaltered Stanford A4C videos, then unaltered and 90° clockwise (CW) rotated videos and finally unaltered, 90° rotated and horizontally flipped videos. As a real-world test, we obtained 615 SX videos from Harbor-UCLA (HUCLA) with EF calculations in 5% ranges. Researchers performed 1000 randomizations of EF point estimation within HUCLA EF ranges to compensate for ML and HUCLA EF mismatch, obtaining a mean value for absolute error (MAE) comparison and performed Bland-Altman analyses. RESULTS: The ML algorithm EF mean MAE was estimated at 23.0, with a range of 22.8-23.3 using unaltered A4C video, mean MAE was 16.7, with a range of 16.5-16.9 using unaltered and 90° CW rotated video, mean MAE was 16.6, with a range of 16.3-16.8 using unaltered, 90° CW rotated and horizontally flipped video training. Bland-Altman showed weakest agreement at 40-45% EF. CONCLUSIONS: Researchers successfully adapted unrelated ultrasound window data to train a POCUS ML algorithm with fair MAE using data manipulation to simulate a different ultrasound examination. This may be important for future POCUS algorithm design to help overcome a paucity of POCUS databases.

Optic Nerve Sheath Diameter Ultrasound for Raised Intracranial Pressure: A Literature Review and Meta-analysis of its Diagnostic Accuracy.

Optic nerve sheath diameter (ONSD) ultrasound is becoming increasingly more popular for estimating raised intracranial pressure (ICP). We performed a systematic review and analysis of the diagnostic accuracy of ONSD when compared to the standard invasive ICP measurement. METHOD: We performed a systematic search of PUBMED and EMBASE for studies including adult patients with suspected elevated ICP and comparing sonographic ONSD measurement to a standard invasive method. Quality of studies was assessed using the QUADAS-2 tool by two independent authors. We used a bivariate model of random effects to summarize pooled sensitivity, specificity, and diagnostic odds ratio (DOR). Heterogeneity was investigated by meta-regression and sub-group analyses. RESULTS: We included 18 prospective studies (16 studies including 619 patients for primary outcome). Only one study was of low quality, and there was no apparent publication bias. Pooled sensitivity was 0.9 [95% confidence intervals (CI): 0.85-0.94], specificity was 0.85 (95% CI: 0.8-0.89), and DOR was 46.7 (95% CI: 26.2-83.2) with partial evidence of heterogeneity. The Area-Under-the-Curve of the summary Receiver-Operator-Curve was 0.93 (95% CI: 0.91-0.95, P < .05). No covariates were significant in the meta-regression. Subgroup analysis of severe traumatic brain injury and parenchymal ICP found no heterogeneity. ICP and ONSD had a correlation coefficient of 0.7 (95% CI: 0.63-0.76, P < .05). CONCLUSION: ONSD is a useful adjunct in ICP evaluation but is currently not a replacement for invasive methods where they are feasible.

Early Lung Ultrasound Findings in Patients With COVID-19 Pneumonia: A Retrospective Multicenter Study of 479 Patients.

OBJECTIVES: Lung ultrasound (LUS) holds the promise of an accurate, radiation-free, and affordable diagnostic and monitoring tool in coronavirus disease 2019 (COVID-19) pneumonia. We sought to evaluate the usefulness of LUS in the diagnosis of patients with respiratory distress and suspicion of interstitial severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pneumonia, in comparison to other imaging modalities. METHODS: This was a multicenter, retrospective study. LUS was performed, on Emergency Department (ED) arrival of patients presenting for possible COVID-19 evaluation, by trained emergency physicians, before undergoing conventional radiologic examination or while waiting for the report. Scans were performed using longitudinal transducer orientation of the lung regions. CXR was interpreted by radiologists staffing ED radiology. Subjects were divided into two group based on molecular test results. LUS findings were compared to COVID test results, nonlaboratory data, and other imaging for each patient. Categorical variables were expressed as percentages and continuous variables as median ± standard error. RESULTS: A total of 479 patients were enrolled, 87% diagnosed with SARS-CoV-2 by molecular testing. COVID positive and COVID negative patients differed with respect to sex, presence of fever, and white blood cells count. Most common findings on lung point of care ultrasound (POCUS) for COVID-positive patients were B-lines, irregular pleural lines, and small consolidation. Normal chest X-ray was found in 17.89% of cases. CONCLUSIONS: This 479 patient cohort, with COVID-19, found LUS to be noninferior to chest X-ray (CXR) for diagnostic accuracy. In this study, COVID-positive patients are most likely to show B lines and sub-pleural consolidations on LUS examination.

Respiratory Variation in Carotid Artery Peak Systolic Velocity Is Unable to Predict Fluid Responsiveness in Spontaneously Breathing Critically Ill Patients When Assessed by Novice Physician Sonologists.

BACKGROUND: Respiratory variation in carotid artery peak systolic velocity (ΔVpeak) assessed by point-of-care ultrasound (POCUS) has been proposed as a noninvasive means to predict fluid responsiveness. We aimed to evaluate the ability of carotid ΔVpeak as assessed by novice physician sonologists to predict fluid responsiveness. METHODS: This study was conducted in 2 intensive care units. Spontaneously breathing, nonintubated patients with signs of volume depletion were included. Patients with atrial fibrillation/flutter, cardiogenic, obstructive or neurogenic shock, or those for whom further intravenous (IV) fluid administration would be harmful were excluded. Three novice physician sonologists were trained in POCUS assessment of carotid ΔVpeak. They assessed the carotid ΔVpeak in study participants prior to the administration of a 500 mL IV fluid bolus. Fluid responsiveness was defined as a ≥10% increase in cardiac index as measured using bioreactance. RESULTS: Eighty-six participants were enrolled, 50 (58.1%) were fluid responders. Carotid ΔVpeak performed poorly at predicting fluid responsiveness. Test characteristics for the optimum carotid ΔVpeak of 8.0% were: area under the receiver operating curve = 0.61 (95% CI: 0.48-0.73), sensitivity = 72.0% (95% CI: 58.3-82.56), specificity = 50.0% (95% CI: 34.5-65.5). CONCLUSIONS: Novice physician sonologists using POCUS are unable to predict fluid responsiveness using carotid ΔVpeak. Until further research identifies key limiting factors, clinicians should use caution directing IV fluid resuscitation using carotid ΔVpeak.

Regulating Critical Care Ultrasound, It Is All in the Interpretation.

Point-of-care ultrasound (POCUS) use is rapidly expanding as a practice in adult and pediatric critical care environments. In January 2020, the Joint Commission endorsed a statement from the Emergency Care Research Institute citing point-of-care ultrasound as a potential hazard to patients for reasons related to training and skill verification, oversight of use, and recordkeeping and accountability mechanisms for clinical use; however, no evidence was presented to support these concerns. Existing data on point-of-care ultrasound practices in pediatric critical care settings verify that point-of-care ultrasound use continues to increase, and contrary to the concerns raised, resources are becoming increasingly available for point-of-care ultrasound use. Many institutions have recognized a successful approach to addressing these concerns that can be achieved through multispecialty collaborations.

The COVID-19 Worsening Score (COWS)-a predictive bedside tool for critical illness.

OBJECTIVES: To evaluate the accuracy of a new COVID-19 prognostic score based on lung ultrasound (LUS) and previously validated variables in predicting critical illness. METHODS: We conducted a single-center retrospective cohort development and internal validation study of the COVID-19 Worsening Score (COWS), based on a combination of the previously validated COVID-GRAM score (GRAM) variables and LUS. Adult COVID-19 patients admitted to the emergency department (ED) were enrolled. Ten variables previously identified by GRAM, days from symptom onset, LUS findings, and peripheral oxygen saturation/fraction of inspired oxygen (P/F) ratio were analyzed. LUS score as a single predictor was assessed. We evaluated GRAM model's performance, the impact of adding LUS, and then developed a new model based on the most predictive variables. RESULTS: Among 274 COVID-19 patients enrolled, 174 developed critical illness. The GRAM score identified 51 patients at high risk of developing critical illness and 132 at low risk. LUS score over 15 (range 0 to 36) was associated with a higher risk ratio of critical illness (RR, 2.05; 95% confidence interval [CI], 1.52-2.77; area under the curve [AUC], 0.63; 95% CI 0.676-0.634). The newly developed COVID-19 Worsening Score relies on five variables to classify high- and low-risk patients with an overall accuracy of 80% and negative predictive value of 93% (95% CI, 87%-98%). Patients scoring more than 0.183 on COWS showed a RR of developing critical illness of 8.07 (95% CI, 4.97-11.1). CONCLUSIONS: COWS accurately identify patients who are unlikely to need intensive care unit (ICU) admission, preserving resources for the remaining high-risk patients.

Are Convolutional Neural Networks Trained on ImageNet Images Wearing Rose-Colored Glasses?: A Quantitative Comparison of ImageNet, Computed Tomographic, Magnetic Resonance, Chest X-Ray, and Point-of-Care Ultrasound Images for Quality.

OBJECTIVES: Deep learning for medical imaging analysis uses convolutional neural networks pretrained on ImageNet (Stanford Vision Lab, Stanford, CA). Little is known about how such color- and scene-rich standard training images compare quantitatively to medical images. We sought to quantitatively compare ImageNet images to point-of-care ultrasound (POCUS), computed tomographic (CT), magnetic resonance (MR), and chest x-ray (CXR) images. METHODS: Using a quantitative image quality assessment technique (Blind/Referenceless Image Spatial Quality Evaluator), we compared images based on pixel complexity, relationships, variation, and distinguishing features. We compared 5500 ImageNet images to 2700 CXR, 2300 CT, 1800 MR, and 18,000 POCUS images. Image quality results ranged from 0 to 100 (worst). A 1-way analysis of variance was performed, and the standardized mean-difference effect size value (d) was calculated. RESULTS: ImageNet images showed the best image quality rating of 21.7 (95% confidence interval [CI], 0.41) except for CXR at 13.2 (95% CI, 0.28), followed by CT at 35.1 (95% CI, 0.79), MR at 31.6 (95% CI, 0.75), and POCUS at 56.6 (95% CI, 0.21). The differences between ImageNet and all of the medical images were statistically significant (P ≤ .000001). The greatest difference in image quality was between ImageNet and POCUS (d = 2.38). CONCLUSIONS: Point-of-care ultrasound (US) quality is significantly different from that of ImageNet and other medical images. This brings considerable implications for convolutional neural network training with medical images for various applications, which may be even more significant in the case of US images. Ultrasound deep learning developers should consider pretraining networks from scratch on US images, as training techniques used for CT, CXR, and MR images may not apply to US.

Lung Ultrasound Point-of-View in Pediatric and Adult COVID-19 Infection.

From its start in China in December 2019, infection by the new SARS-CoV2 spread fast all over the world. It can present as severe respiratory distress in the elderly or a vasculitis in a child, most of whom are typically completely asymptomatic. This makes infection detection based on clinical grounds exceedingly difficult. Lung ultrasound has become an important tool in diagnosis and follow-up of patient with COVID-19 infection.Here we review available, up to date literature on ultrasound use for COVID-19 suspected pediatric patients and contrast it to published findings in adult patients.

Development of a Deep Learning Network to Classify Inferior Vena Cava Collapse to Predict Fluid Responsiveness.

OBJECTIVES: To create a deep learning algorithm capable of video classification, using a long short-term memory (LSTM) network, to analyze collapsibility of the inferior vena cava (IVC) to predict fluid responsiveness in critically ill patients. METHODS: We used a data set of IVC ultrasound (US) videos to train the LSTM network. The data set was created from IVC US videos of spontaneously breathing critically ill patients undergoing intravenous fluid resuscitation as part of 2 prior prospective studies. We randomly selected 90% of the IVC videos to train the LSTM network and 10% of the videos to test the LSTM network's ability to predict fluid responsiveness. Fluid responsiveness was defined as a greater than 10% increase in the cardiac index after a 500-mL fluid bolus, as measured by bioreactance. RESULTS: We analyzed 211 videos from 175 critically ill patients: 191 to train the LSTM network and 20 to test it. Using standard data augmentation techniques, we increased our sample size from 191 to 3820 videos. Of the 175 patients, 91 (52%) were fluid responders. The LSTM network was able to predict fluid responsiveness moderately well, with an area under the receiver operating characteristic curve of 0.70 (95% confidence interval [CI], 0.43-1.00), a positive likelihood ratio of infinity, and a negative likelihood ratio of 0.3 (95% CI, 0.12-0.77). In comparison, point-of-care US experts using video review offline and manual diameter measurement via software caliper tools achieved an area under the receiver operating characteristic curve of 0.94 (95% CI, 0.83-0.99). CONCLUSIONS: We demonstrated that an LSTM network can be trained by using videos of IVC US to classify IVC collapse to predict fluid responsiveness. Our LSTM network performed moderately well given the small training cohort but worse than point-of-care US experts. Further training and testing of the LSTM network with a larger data sets is warranted.

Prospective Longitudinal Evaluation of Point-of-Care Lung Ultrasound in Critically Ill Patients With Severe COVID-19 Pneumonia.

OBJECTIVES: To perform a prospective longitudinal analysis of lung ultrasound findings in critically ill patients with coronavirus disease 2019 (COVID-19). METHODS: Eighty-nine intensive care unit (ICU) patients with confirmed COVID-19 were prospectively enrolled and tracked. Point-of-care ultrasound (POCUS) examinations were performed with phased array, convex, and linear transducers using portable machines. The thorax was scanned in 12 lung areas: anterior, lateral, and posterior (superior/inferior) bilaterally. Lower limbs were scanned for deep venous thrombosis and chest computed tomographic angiography was performed to exclude suspected pulmonary embolism (PE). Follow-up POCUS was performed weekly and before hospital discharge. RESULTS: Patients were predominantly male (84.2%), with a median age of 43 years. The median duration of mechanical ventilation was 17 (interquartile range, 10-22) days; the ICU length of stay was 22 (interquartile range, 20.2-25.2) days; and the 28-day mortality rate was 28.1%. On ICU admission, POCUS detected bilateral irregular pleural lines (78.6%) with accompanying confluent and separate B-lines (100%), variable consolidations (61.7%), and pleural and cardiac effusions (22.4% and 13.4%, respectively). These findings appeared to signify a late stage of COVID-19 pneumonia. Deep venous thrombosis was identified in 16.8% of patients, whereas chest computed tomographic angiography confirmed PE in 24.7% of patients. Five to six weeks after ICU admission, follow-up POCUS examinations detected significantly lower rates (P < .05) of lung abnormalities in survivors. CONCLUSIONS: Point-of-care ultrasound depicted B-lines, pleural line irregularities, and variable consolidations. Lung ultrasound findings were significantly decreased by ICU discharge, suggesting persistent but slow resolution of at least some COVID-19 lung lesions. Although POCUS identified deep venous thrombosis in less than 20% of patients at the bedside, nearly one-fourth of all patients were found to have computed tomography-proven PE.

Residual Lung Injury in Patients Recovering From COVID-19 Critical Illness: A Prospective Longitudinal Point-of-Care Lung Ultrasound Study.

Scarce data exist regarding the natural history of lung lesions detected on ultrasound in those who survive severe COVID-19 pneumonia. OBJECTIVE: We performed a prospective analysis of point-of-care ultrasound (POCUS) findings in critically ill COVID-19 patients during and after hospitalization. METHODS: We enrolled 171 COVID-19 intensive care unit patients. POCUS of the lungs was performed with phased array (2-4 MHz), convex (2-6 MHz) and linear (10-15 MHz) transducers, scanning 12 lung areas. Chest computed tomography angiography was performed to exclude suspected pulmonary embolism. Survivors were clinically and sonographically evaluated during a 4 month period for evidence of residual lung injury. Chest computed tomography angiography and echocardiography were used to exclude pulmonary hypertension (PH) and chest high-resolution-computed-tomography to exclude interstitial lung disease (ILD) in symptomatic survivors. RESULTS: Cox regression analysis showed that lymphocytopenia (hazard ratio [HR]: 0.88, 95% confidence intervals [CI]: 0.68-0.96, p = .048), increased lactate (HR: 1.17, 95% CI: 0.94-1.46, p = 0.049), and D-dimers (HR: 1.21, 95% CI: 1.03-1.44, p = .03) were mortality predictors. Non-survivors had increased incidence of pulmonary abnormalities (B-lines, pleural line irregularities, and consolidations) compared to survivors (p < .05). During follow-up, POCUS with clinical and laboratory parameters integrated in the semi-quantitative Riyadh-Residual-Lung-Injury scale had sensitivity of 0.82 (95% CI: 0.76-0.89) and specificity of 0.91 (95% CI: 0.94-0.95) in predicting ILD. The prevalence of PH and ILD (non-specific-interstitial-pneumonia) was 7% and 11.8%, respectively. CONCLUSION: POCUS showed ability to monitor the evolution of severe COVID-19 pneumonia after hospital discharge, supporting its integration in clinical predictive models of residual lung injury.

Performance of a 25% Inferior Vena Cava Collapsibility in Detecting Fluid Responsiveness When Assessed by Novice Versus Expert Physician Sonologists.

OBJECTIVES: Inferior vena cava collapsibility (cIVC) measured by point-of-care ultrasound (POCUS) has been proposed as a noninvasive means of assessing fluid responsiveness. We aimed to prospectively evaluate the performance of a 25% cIVC cutoff value to detect fluid responsiveness among spontaneously breathing intensive care unit (ICU) patients when assessed with POCUS by novice versus expert physician sonologists. METHODS: Prospective observational study of spontaneously breathing ICU patients. Fluid responsiveness was defined as a >10% increase in cardiac index following a 500 mL fluid bolus, measured by bioreactance. Novice sonologist measured cIVC with POCUS. Their measurements were later compared to an expert physician sonologist who independently reviewed the POCUS images and assessed cIVCs. RESULTS: Of the 85 participants, 44 (52%) were fluid responders. A 25% cIVC cutoff value performed better when assessed by expert sonologists than novice physician sonologists (receiver-operator characteristic curve, ROC = 0.82 [0.74-0.88] vs ROC = 0.69 [0.60-0.77]). CONCLUSIONS: A 25% cIVC cutoff value measured by POCUS detects fluid responsiveness. However, the experience of the physician sonologist affects test performance and should be considered when interpreting and clinically using cIVC to direct intravenous fluid resuscitation.

Are All Deep Learning Architectures Alike for Point-of-Care Ultrasound?: Evidence From a Cardiac Image Classification Model Suggests Otherwise.

OBJECTIVES: Little is known about optimal deep learning (DL) approaches for point-of-care ultrasound (POCUS) applications. We compared 6 popular DL architectures for POCUS cardiac image classification to determine whether an optimal DL architecture exists for future DL algorithm development in POCUS. METHODS: We trained 6 convolutional neural networks (CNNs) with a range of complexities and ages (AlexNet, VGG-16, VGG-19, ResNet50, DenseNet201, and Inception-v4). Each CNN was trained by using images of 5 typical POCUS cardiac views. Images were extracted from 225 publicly available deidentified POCUS cardiac videos. A total of 750,018 individual images were extracted, with 90% used for model training and 10% for cross-validation. The training time and accuracy achieved were tracked. A real-world test of the algorithms was performed on a set of 125 completely new cardiac images. Descriptive statistics, Pearson R values, and κ values were calculated for each CNN. RESULTS: Accuracy ranged from 96% to 85.6% correct for the 6 CNNs. VGG-16, one of the oldest and simplest CNNs, performed best at 96% correct with 232 minutes to train (R = 0.97; κ = 0.95; P < .00001). The worst-performing CNN was the newer DenseNet201, with 85.6% accuracy and 429 minutes to train (R = 0.92; κ = 0.82; P < .00001). CONCLUSIONS: Six common image classification DL algorithms showed considerable variability in their accuracy and training time when trained and tested on identical data, suggesting that not all will perform optimally for POCUS DL applications. Contrary to well-established accuracies for CNNs, more modern and deeper algorithms yielded poorer results.

Creation and Testing of a Deep Learning Algorithm to Automatically Identify and Label Vessels, Nerves, Tendons, and Bones on Cross-sectional Point-of-Care Ultrasound Scans for Peripheral Intravenous Catheter Placement by Novices.

OBJECTIVES: We sought to create a deep learning (DL) algorithm to identify vessels, bones, nerves, and tendons on transverse upper extremity (UE) ultrasound (US) images to enable providers new to US-guided peripheral vascular access to identify anatomy. METHODS: We used publicly available DL architecture (YOLOv3) and deidentified transverse US videos of the UE for algorithm development. Vessels, bones, tendons, and nerves were labeled with bounding boxes. A total of 203,966 images were generated from videos, with corresponding label box coordinates in a YOLOv3 format. Training accuracy, losses, and learning curves were tracked. As a final real-world test, 50 randomly selected images from unrelated UE US videos were used to test the DL algorithm. Four different versions of the YOLOv3 algorithm were tested with varied amounts of training and sensitivity settings. The same 50 images were labeled by 2 blinded point-of-care ultrasound (POCUS) experts. The area under the curve (AUC) was calculated for the DL algorithm and POCUS expert performance. RESULTS: The algorithm outperformed POCUS experts in detection of all structures in the UE, with an AUC of 0.78 versus 0.69 and 0.71, respectively. When considering vessels, only one of the POCUS experts attained an AUC of 0.85, just ahead of the DL algorithm, with an AUC of 0.83. CONCLUSIONS: Our DL algorithm proved accurate at identifying 4 common structures on cross-sectional US imaging of the UE, which would allow novice POCUS providers to more confidently and accurately target vessels for cannulation, avoiding other structures. Overall, the algorithm outperformed 2 blinded POCUS experts.

Enhanced Point-of-Care Ultrasound Applications by Integrating Automated Feature-Learning Systems Using Deep Learning.

Recent applications of artificial intelligence (AI) and deep learning (DL) in health care include enhanced diagnostic imaging modalities to support clinical decisions and improve patients' outcomes. Focused on using automated DL-based systems to improve point-of-care ultrasound (POCUS), we look at DL-based automation as a key field in expanding and improving POCUS applications in various clinical settings. A promising additional value would be the ability to automate training model selections for teaching POCUS to medical trainees and novice sonologists. The diversity of POCUS applications and ultrasound equipment, each requiring specialized AI models and domain expertise, limits the use of DL as a generic solution. In this article, we highlight the most advanced potential applications of AI in POCUS tailored to high-yield models in automated image interpretations, with the premise of improving the accuracy and efficacy of POCUS scans.

Transesophageal Echocardiography: Guidelines for Point-of-Care Applications in Cardiac Arrest Resuscitation.

Cardiac arrest is one of the most challenging patient presentations managed by emergency care providers, and echocardiography can be instrumental in the diagnosis, prognosis, and treatment guidance in these critically ill patients. Transesophageal echocardiography has many advantages over transthoracic echocardiography in a cardiac arrest resuscitation. As transesophageal echocardiography is implemented more widely at the point of care during cardiac arrest resuscitations, guidelines are needed to assist emergency providers in acquiring the equipment and skills necessary to successfully incorporate it into the management of cardiac arrest victims.

Diagnostic value of cardiopulmonary ultrasound in elderly patients with acute respiratory distress syndrome.

BACKGROUND: Lung ultrasound and echocardiography are mainly applied in critical care and emergency medicine. However, the diagnostic value of cardiopulmonary ultrasound in elderly patients with acute respiratory distress syndrome (ARDS) is still unclear. METHODS: Consecutive patients admitted to ICU with the diagnosis of suspected ARDS based on clinical grounds were enrolled. Cardiopulmonary ultrasound was performed as part of monitoring on day 1, day 2 and day 3. On each day a bedside ultrasound was performed to examine the lungs and calculate the Left Ventricular Ejection Fraction (LVEF). On day 3, a thoracic CT was performed on each patient as gold standard for ARDS imaging diagnosis. According to the results from CT scan, patients were grouped into ARDS group or Non-ARDS group. The relation between the cardiopulmonary ultrasound results on each day and the results of CT scan was analyzed. RESULTS: Fifty one consecutive patients aged from 73 to 97 years old were enrolled. Based on CT criteria, 33 patients were classified into the ARDS group, while 18 patients were included in non-ARDS group. There was no significant difference between the two groups in baseline characteristics, including gender, age, underlying disease, comorbidities, APACHE II score, SOFA score, and PaO2/FiO2 ratio (P > 0.05). Lung ultrasound (LUS) examination results were consistent with the CT scan results in diagnosis of pulmonary lesions. The Kappa values were 0.55, 0.74 and 0.82 on day 1, day 2 and day 3, respectively. The ROC analysis showed that the sensitivity, specificity and area under curve of ROC (AUROC) for lung ultrasound in diagnose ARDS were 0.788,0.778,0.783;0.909,0.833,0.871;0.970,0.833,0.902 on day 1, day 2 and day 3, respectively. However, cardiopulmonary ultrasound performed better in diagnosing ARDS in elderly patients. The sensitivity, specificity and AUROC were 0.879,0.889,0.924;0.939,0.889,0.961;and 0.970,0.833,0.956 on day 1, day 2 and day 3, respectively. The combined performances of cardiopulmonary ultrasound, N-terminal pro-brain natriuretic peptide (NT-proBNP), and PaO2/FiO2 ratio improved the specificity of the diagnosis of ARDS in elderly patients. CONCLUSIONS: LUS examination results were consistent with the CT scan results in diagnosis of pulmonary lesions. Cardiopulmonary ultrasound has a greater diagnostic accuracy in elderly patients with ARDS, compared with LUS alone. The combined performances of cardiopulmonary ultrasound, NT-proBNP, and PaO2/FiO2 increased the specificity of the diagnosis of ARDS in elderly patients.