Quantitative lung ultrasonography: a putative new algorithm for automatic detection and quantification of B-lines

Combining quantitative and qualitative analysis for scoring pleural line in lung ultrasound

Objective.Accurate assessment of pleural line is crucial for the application of lung ultrasound (LUS) in monitoring lung diseases, thereby aim of this study is to develop a quantitative and qualitative analysis method for pleural line.Approach.The novel cascaded deep learning model based on convolution and multilayer perceptron was proposed to locate and segment the pleural line in LUS images, whose results were applied for quantitative analysis of textural and morphological features, respectively. By using gray-level co-occurrence matrix and self-designed statistical methods, eight textural and three morphological features were generated to characterize the pleural lines. Furthermore, the machine learning-based classifiers were employed to qualitatively evaluate the lesion degree of pleural line in LUS images.Main results.We prospectively evaluated 3770 LUS images acquired from 31 pneumonia patients. Experimental results demonstrated that the proposed pleural line extraction and evaluation methods all have good performance, with dice and accuracy of 0.87 and 94.47%, respectively, and the comparison with previous methods found statistical significance (P< 0.001 for all). Meanwhile, the generalization verification proved the feasibility of the proposed method in multiple data scenarios.Significance.The proposed method has great application potential for assessment of pleural line in LUS images and aiding lung disease diagnosis and treatment.

Lung Ultrasound Score on Postoperative Day 1 Is Predictive of the Occurrence of Pulmonary Complications after Major Abdominal Surgery: A Multicenter Prospective Observational Study

BACKGROUND

Postoperative pulmonary complications after major abdominal surgery are frequent and carry high morbidity and mortality. Early identification of patients at risk of pulmonary complications by lung ultrasound may allow the implementation of preemptive strategies. The authors hypothesized that lung ultrasound score would be associated with pulmonary postoperative complications. The main objective of the study was to evaluate the performance of lung ultrasound score on postoperative day 1 in predicting pulmonary complications after major abdominal surgery. Secondary objectives included the evaluation of other related measures for their potential prediction accuracy.

METHODS

A total of 149 patients scheduled for major abdominal surgery were enrolled in a bicenter observational study. Lung ultrasound score was performed before the surgery and on days 1, 4, and 7 after surgery. Pulmonary complications occurring before postoperative day 10 were recorded.

RESULTS

Lung ultrasound score on postoperative day 1 was higher in patients developing pulmonary complications before day 10 (median, 13; interquartile range, 8.25 to 18; vs. median, 10; interquartile range, 6.5 to 12; Mann-Whitney P = 0.002). The area under the curve for predicting postoperative pulmonary complications before day 10 was 0.65 (95% CI, 0.55 to 0.75; P = 0.003). Lung ultrasound score greater than 12 had a sensitivity of 0.54 (95% CI, 0.40 to 0.67), specificity of 0.77 (95% CI, 0.67 to 0.85), and negative predictive value of 0.74 (95% CI, 0.65 to 0.83). Lung ultrasound score greater than 17 had sensitivity of 0.33 (95% CI, 0.21 to 0.47), specificity of 0.95 (95% CI, 0.88 to 0.98), and positive predictive value of 0.78 (95% CI, 0.56 to 0.93). Anterolateral lung ultrasound score and composite scores using lung ultrasound score and other patient characteristics showed similar predictive accuracies.

CONCLUSIONS

An elevated lung ultrasound score on postoperative day 1 is associated with the occurrence of pulmonary complications within the first 10 days after major abdominal surgery.

EDITOR’S PERSPECTIVE

Lung ultrasound training: how short is too short? observational study on the effects of a focused theoretical training for novice learners

BACKGROUND

Lung ultrasound has been increasingly used in the last years for the assessment of patients with respiratory diseases; it is considered a simple technique, now spreading from physicians to other healthcare professionals as nurses and physiotherapists, as well as to medical students. These providers may require a different training to acquire lung ultrasound skills, since they are expected to have no previous experience with ultrasound. The aim of the study was to assess the impact of a short theoretical training focused on lung ultrasound pattern recognition in a population of novice nurse learners with no previous experience with ultrasound.

METHODS

We included the nurses attending a critical care advanced course for nurses performed at the University of Pavia. Images' interpretation skills were tested on two slide sets (a 25-clip set focused on B-pattern recognition and a 25-clip set focused on identification of pleural movement as lung sliding, lung pulse, lung point, no movement) before and after three 30-minute teaching modules dedicated to general ultrasound principles, B-lines assessment and lung sliding assessment. A cut off of 80% was considered acceptable for correctly interpreted images after this basic course.

RESULTS

22 nurses were enrolled (age 26.0 [24.0-28.0] years; men 4 (18%)); one nurse had previous experience with other ultrasound techniques, none of them had previous experience with lung ultrasound. After the training, the number of correctly interpreted clips improved from 3.5 [0.0-13.0] to 22.0 [19.0-23.0] (p < 0.0001) for B-pattern and from 0.5 [0.0-2.0] to 8.5 [6.0-12.0] (p < 0.0001) for lung sliding assessment. The number of correct answers for B-pattern recognition was significantly higher than for lung sliding assessment, both before (3.5 [0.0-13.0] vs. 0.5 [0.0-2.0]; p = 0.0036) and after (22.0 [19.0-23.0] vs. 8.5 [6.0-12.0]; p < 0.0001) the training. After the training, nurses were able to correctly recognize the presence or the absence of a B-pattern in 84.2 ± 10.3% of cases; lung sliding was correctly assessed in 37.1 ± 15.3% of cases.

CONCLUSIONS

Lung ultrasound is considered a simple technique; while a short, focused training significantly improves B-pattern recognition, lung sliding assessment may require a longer training for novice learners.

TRIAL REGISTRATION

Not applicable.

Lung ultrasound score in dogs and cats: A reliability study

BACKGROUND

Lung ultrasound (LUS) is a noninvasive tool for examining respiratory distress patients. The lung ultrasound score (LUSS) can be used to quantify and monitor lung aeration loss with good reliability.

HYPOTHESIS/OBJECTIVES

Assess the reliability of a new LUSS among raters with different levels of experience and determine how well the same raters agree on identifying patterns of LUS abnormalities.

ANIMALS

Forty LUS examinations of dogs and cats and 320 videos were reviewed from a digital database.

METHODS

Retrospective reliability study with post hoc analysis. Protocolized LUS were randomly selected; intrarater and interrater reliability of the LUSS and pattern recognition agreement among 4 raters with different levels of experience in LUS were tested.

RESULTS

The intrarater intraclass correlation coefficient (ICC) single measurement, absolute agreement, and 2-way mixed effects model was 0.967 for the high-experience rater (H-Exp), 0.963 and 0.952 for the medium-experience raters (M-Exp-1; M-Exp-2), and 0.950 for the low-experience rater (L-Exp). The interrater ICC average measurement, absolute agreement, and 2-way random effects model among the observers was 0.980. The Fleiss' kappa (k) values showed almost perfect agreement (k = 1) among raters in identifying pleural effusion and translobar tissue-like pattern, strong agreement for A-lines (k = 0.881) and B-lines (k = 0.806), moderate agreement (k = 0.693) for subpleural loss of aeration, and weak agreement (k = 0.474) for irregularities of the pleural line.

CONCLUSIONS AND CLINICAL IMPORTANCE

Our results indicate excellent intra- and interrater reliability for LUS scoring and pattern identification, providing a foundation for the use of the LUSS in emergency medicine and intensive care.

The value of a dynamic echocardiographic approach to diastolic dysfunction in intensive care medicine

Diastolic dysfunction is an underestimated feature in the context of the critically ill setting and perioperative medicine. Advances in echocardiography, its noninvasive, safe and easy use, have allowed Doppler echocardiography to become a cornerstone for diagnosing diastolic dysfunction in clinical practice. The diagnosis of diastolic dysfunction and increased filling pressures is nevertheless complex. Using an echocardiographic assessment and the routine application of preload stress maneuvers during echocardiographic examination can help identify early stages of diastolic dysfunction leading to better management of patients at risk of acute heart decompensation in the perioperative period or during ICU stay.

Noninvasive Imaging Methods for Quantification of Pulmonary Edema and Congestion: A Systematic Review

Quantification of pulmonary edema and congestion is important to guide diagnosis and risk stratification, and to objectively evaluate new therapies in heart failure. Herein, we review the validation, diagnostic performance, and clinical utility of noninvasive imaging modalities in this setting, including chest x-ray, lung ultrasound (LUS), computed tomography (CT), nuclear medicine imaging methods (positron emission tomography [PET], single photon emission CT), and magnetic resonance imaging (MRI). LUS is a clinically useful bedside modality, and fully quantitative methods (CT, MRI, PET) are likely to be important contributors to a more accurate and precise evaluation of new heart failure therapies and for clinical use in conjunction with cardiac imaging. There are only a limited number of studies evaluating pulmonary congestion during stress. Taken together, noninvasive imaging of pulmonary congestion provides utility for both clinical and research assessment, and continued refinement of methodologic accuracy, validation, and workflow has the potential to increase broader clinical adoption.

An Interpretable Neonatal Lung Ultrasound Feature Extraction and Lung Sliding Detection System Using Object Detectors

The objective of this study was to develop an interpretable system that could detect specific lung features in neonates. A challenging aspect of this work was that normal lungs showed the same visual features (as that of Pneumothorax (PTX)). M-mode is typically necessary to differentiate between the two cases, but its generation in clinics is time-consuming and requires expertise for interpretation, which remains limited. Therefore, our system automates M-mode generation by extracting Regions of Interest (ROIs) without human in the loop. Object detection models such as faster Region Based Convolutional Neural Network (fRCNN) and RetinaNet models were employed to detect seven common Lung Ultrasound (LUS) features. fRCNN predictions were then stored and further used to generate M-modes. Beyond static feature extraction, we used a Hough transform based statistical method to detect "lung sliding" in these M-modes. Results showed that fRCNN achieved a greater mean Average Precision (mAP) of 86.57% (Intersection-over-Union (IoU) = 0.2) than RetinaNet, which only displayed a mAP of 61.15%. The calculated accuracy for the generated RoIs was 97.59% for Normal videos and 96.37% for PTX videos. Using this system, we successfully classified 5 PTX and 6 Normal video cases with 100% accuracy. Automating the process of detecting seven prominent LUS features addresses the time-consuming manual evaluation of Lung ultrasound in a fast paced environment. Clinical impact: Our research work provides a significant clinical impact as it provides a more accurate and efficient method for diagnosing lung diseases in neonates.

Ultrasound imaging of lung disease and its relationship to histopathology: An experimentally validated simulation approach

Lung ultrasound (LUS) is a widely used technique in clinical lung assessment, yet the relationship between LUS images and the underlying disease remains poorly understood due in part to the complexity of the wave propagation physics in complex tissue/air structures. Establishing a clear link between visual patterns in ultrasound images and underlying lung anatomy could improve the diagnostic accuracy and clinical deployment of LUS. Reverberation that occurs at the lung interface is complex, resulting in images that require interpretation of the artifacts deep in the lungs. These images are not accurate spatial representations of the anatomy due to the almost total reflectivity and high impedance mismatch between aerated lung and chest wall. Here, we develop an approach based on the first principles of wave propagation physics in highly realistic maps of the human chest wall and lung to unveil a relationship between lung disease, tissue structure, and its resulting effects on ultrasound images. It is shown that Fullwave numerical simulations of ultrasound propagation and histology-derived acoustical maps model the multiple scattering physics at the lung interface and reproduce LUS B-mode images that are comparable to clinical images. However, unlike clinical imaging, the underlying tissue structure model is known and controllable. The amount of fluid and connective tissue components in the lung were gradually modified to model disease progression, and the resulting changes in B-mode images and non-imaging reverberation measures were analyzed to explain the relationship between pathological modifications of lung tissue and observed LUS.

Deep Learning for Detection and Localization of B-Lines in Lung Ultrasound

Lung ultrasound (LUS) is an important imaging modality used by emergency physicians to assess pulmonary congestion at the patient bedside. B-line artifacts in LUS videos are key findings associated with pulmonary congestion. Not only can the interpretation of LUS be challenging for novice operators, but visual quantification of B-lines remains subject to observer variability. In this work, we investigate the strengths and weaknesses of multiple deep learning approaches for automated B-line detection and localization in LUS videos. We curate and publish, BEDLUS, a new ultrasound dataset comprising 1,419 videos from 113 patients with a total of 15,755 expert-annotated B-lines. Based on this dataset, we present a benchmark of established deep learning methods applied to the task of B-line detection. To pave the way for interpretable quantification of B-lines, we propose a novel "single-point" approach to B-line localization using only the point of origin. Our results show that (a) the area under the receiver operating characteristic curve ranges from 0.864 to 0.955 for the benchmarked detection methods, (b) within this range, the best performance is achieved by models that leverage multiple successive frames as input, and (c) the proposed single-point approach for B-line localization reaches an F 1-score of 0.65, performing on par with the inter-observer agreement. The dataset and developed methods can facilitate further biomedical research on automated interpretation of lung ultrasound with the potential to expand the clinical utility.

Machine Learning Algorithm Detection of Confluent B-Lines

OBJECTIVE

B-lines are a ring-down artifact of lung ultrasound that arise with increased alveolar water in conditions such as pulmonary edema and infectious pneumonitis. Confluent B-line presence may signify a different level of pathology compared with single B-lines. Existing algorithms aimed at B-line counting do not distinguish between single and confluent B-lines. The objective of this study was to test a machine learning algorithm for confluent B-line identification.

METHODS

This study used a subset of 416 clips from 157 subjects, previously acquired in a prospective study enrolling adults with shortness of breath at two academic medical centers, using a hand-held tablet and a 14-zone protocol. After exclusions, random sampling generated a total of 416 clips (146 curvilinear, 150 sector and 120 linear) for review. A group of five experts in point-of-care ultrasound blindly evaluated the clips for presence/absence of confluent B-lines. Ground truth was defined as majority agreement among the experts and used for comparison with the algorithm.

RESULTS

Confluent B-lines were present in 206 of 416 clips (49.5%). Sensitivity and specificity of confluent B-line detection by algorithm compared with expert determination were 83% (95% confidence interval [CI]: 0.77-0.88) and 92% (95% CI: 0.88-0.96). Sensitivity and specificity did not statistically differ between transducers. Agreement between algorithm and expert for confluent B-lines measured by unweighted κ was 0.75 (95% CI: 0.69-0.81) for the overall set.

CONCLUSION

The confluent B-line detection algorithm had high sensitivity and specificity for detection of confluent B-lines in lung ultrasound point-of-care clips, compared with expert determination.

Can Dialysis Patients Identify and Diagnose Pulmonary Congestion Using Self-Lung Ultrasound?

BACKGROUND

With the recent developments in automated tools, smaller and cheaper machines for lung ultrasound (LUS) are leading us toward the potential to conduct POCUS tele-guidance for the early detection of pulmonary congestion. This study aims to evaluate the feasibility and accuracy of a self-lung ultrasound study conducted by hemodialysis (HD) patients to detect pulmonary congestion, with and without artificial intelligence (AI)-based automatic tools.

METHODS

This prospective pilot study was conducted between November 2020 and September 2021. Nineteen chronic HD patients were enrolled in the Soroka University Medical Center (SUMC) Dialysis Clinic. First, we examined the patient's ability to obtain a self-lung US. Then, we used interrater reliability (IRR) to compare the self-detection results reported by the patients to the observation of POCUS experts and an ultrasound (US) machine with an AI-based automatic B-line counting tool. All the videos were reviewed by a specialist blinded to the performer. We examined their agreement degree using the weighted Cohen's kappa (Kw) index.

RESULTS

A total of 19 patients were included in our analysis. We found moderate to substantial agreement between the POCUS expert review and the automatic counting both when the patient performed the LUS (Kw = 0.49 [95% CI: 0.05-0.93]) and when the researcher performed it (Kw = 0.67 [95% CI: 0.67-0.67]). Patients were able to place the probe in the correct position and present a lung image well even weeks from the teaching session, but did not show good abilities in correctly saving or counting B-lines compared to an expert or an automatic counting tool.

CONCLUSIONS

Our results suggest that LUS self-monitoring for pulmonary congestion can be a reliable option if the patient's count is combined with an AI application for the B-line count. This study provides insight into the possibility of utilizing home US devices to detect pulmonary congestion, enabling patients to have a more active role in their health care.

Monitoring of pulmonary involvement in critically ill COVID-19 patients - should lung ultrasound be preferred over CT?

BACKGROUND

It is unclear if relevant changes in pulmonary involvement in critically ill COVID-19 patients can be reliably detected by the CT severity score (CTSS) and lung ultrasound score (LUSS), or if these changes have prognostic implications. In addition, it has been argued that adding pleural abnormalities to the LUSS could improve its prognostic value. The objective of this study was to compare LUSS and CTSS for the monitoring of COVID-19 pulmonary involvement through: first, establishing the correlation of LUSS (± pleural abnormalities) and CTSS throughout admission; second, assessing agreement and measurement error between raters for LUSS, pleural abnormalities, and CTSS; third, evaluating the association of the LUSS (± pleural abnormalities) and CTSS with mortality at different timepoints.

METHODS

This is a prospective, observational study, conducted during the second COVID-19 wave at the AmsterdamUMC, location VUmc. Adult COVID-19 ICU patients were prospectively included when a CT or a 12-zone LUS was performed at admission or at weekly intervals according to local protocol. Patients were followed 90 days or until death. We calculated the: (1) Correlation of the LUSS (± pleural abnormalities) and CTSS throughout admission with mixed models; (2) Intra-class correlation coefficients (ICCs) and smallest detectable changes (SDCs) between raters; (3) Association between the LUSS (± pleural abnormalities) and CTSS with mixed models.

RESULTS

82 consecutive patients were included. Correlation between LUSS and CTSS was 0.45 (95% CI 0.31-0.59). ICCs for LUSS, pleural abnormalities, and CTSS were 0.88 (95% CI 0.73-0.95), 0.94 (95% CI 0.90-0.96), and 0.84 (95% CI 0.65-0.93), with SDCs of 4.8, 1.4, and 3.9. The LUSS was associated with mortality in week 2, with a score difference between patients who survived or died greater than its SDC. Addition of pleural abnormalities was not beneficial. The CTSS was associated with mortality only in week 1, but with a score difference less than its SDC.

CONCLUSIONS

LUSS correlated with CTSS throughout ICU admission but performed similar or better at agreement between raters and mortality prognostication. Given the benefits of LUS over CT, it should be preferred as initial monitoring tool.

Automatic detection of A-line in lung ultrasound images using deep learning and image processing

BACKGROUND

Auxiliary diagnosis and monitoring of lung diseases based on lung ultrasound (LUS) images is important clinical research. A-line is one of the most common indicators of LUS that can offer support for the assessment of lung diseases. A traditional A-line detection method mainly relies on experienced clinicians, which is inefficient and cannot meet the needs of these areas with backward medical level. Therefore, how to realize the automatic detection of A-line in LUS image is important.

PURPOSE

In order to solve the disadvantages of traditional A-line detection methods, realize automatic and accurate detection, and provide theoretical support for clinical application, we proposed a novel A-line detection method for LUS images with different probe types in this paper.

METHODS

First, the improved Faster R-CNN model with a selection strategy of localization box was designed to accurately locate the pleural line. Then, the LUS image below the pleural line was segmented for independent analysis excluding the influence of other similar structures. Next, image-processing methods based on total variation, matched filter, and gray difference were applied to achieve the automatic A-line detection. Finally, the "depth" index was designed to verify the accuracy by judging whether the automatic measurement results belong to corresponding manual results (±5%). In experiments, 3000 convex array LUS images were used for training and validating the improved pleural line localization model by five-fold cross validation. 850 convex array LUS images and 1080 linear array LUS images were used for testing the trained pleural line localization model and the proposed image-processing-based A-line detection method. The accuracy analysis, error statistics, and Harsdorff distance were employed to evaluate the experimental results.

RESULTS

After 100 epochs, the mean loss value of training and validation set of improved Faster R-CNN model reached 0.6540 and 0.7882, with the validation accuracy of 98.70%. The trained pleural line localization model was applied in the testing set of convex and linear probes and reached the accuracy of 97.88% and 97.11%, respectively, which were 3.83% and 8.70% higher than the original Faster R-CNN model. The accuracy, sensitivity, and specificity of A-line detection reached 95.41%, 0.9244%, 0.9875%, and 94.63%, 0.9230%, and 0.9766% for convex and linear probes, respectively. Compared to the experienced clinicians' results, the mean value and p value of depth error were 1.5342 ± 1.2097 and 0.9021, respectively, and the Harsdorff distance was 5.7305 ± 1.8311. In addition, the accumulated accuracy of the two-stage experiment (pleural line localization and A-line detection) was calculated as the final accuracy of the whole A-line detection system. They were 93.39% and 91.90% for convex and linear probes, respectively, which were higher than these previous methods.

CONCLUSIONS

The proposed method combining image processing and deep learning can automatically and accurately detect A-line in LUS images with different probe types, which has important application value for clinical diagnosis.

Current Advances in Computational Lung Ultrasound Imaging: A Review

In the field of biomedical imaging, ultrasonography has become common practice, and used as an important auxiliary diagnostic tool with unique advantages, such as being non-ionizing and often portable. This article reviews the state-of-the-art in medical ultrasound (US) image processing and in particular its applications in the examination of the lungs. First, we briefly introduce the basis of lung US (LUS) examination. We focus on (i) the characteristics of lung ultrasonography and (ii) its ability to detect a variety of diseases through the identification of various artifacts exhibiting on LUS images. We group medical US image computing methods into two categories: 1) model-based methods and 2) data-driven methods. We particularly discuss inverse problem-based methods exploited in US image despeckling, deconvolution, and line artifacts detection for the former, while we exemplify various works based on deep/machine learning (ML), which exploit various network architectures through supervised, weakly supervised, and unsupervised learning for the data-driven approaches.

Automated lung ultrasound image assessment using artificial intelligence to identify fluid overload in dialysis patients

BACKGROUND

Fluid assessment is challenging, and fluid overload poses a significant problem among dialysis patients, with pulmonary oedema being the most serious consequence. Our study aims to develop a simple objective fluid assessment strategy using lung ultrasound (LUS) and artificial intelligence (AI) to assess the fluid status of dialysis patients.

METHODS

This was a single-centre study of 76 hemodialysis and peritoneal dialysis patients carried out between July 2020 to May 2022. The fluid status of dialysis patients was assessed via a simplified 8-point LUS method using a portable handheld ultrasound device (HHUSD), clinical examination and bioimpedance analysis (BIA). The primary outcome was the performance of 8-point LUS using a portable HHUSD in diagnosing fluid overload compared to physical examination and BIA. The secondary outcome was to develop and validate a novel AI software program to quantify B-line count and assess the fluid status of dialysis patients.

RESULTS

Our study showed a moderate correlation between LUS B-line count and fluid overload assessed by clinical examination (r = 0.475, p < 0.001) and BIA (r = 0.356. p < 0.001). The use of AI to detect B-lines on LUS in our study for dialysis patients was shown to have good agreement with LUS B lines observed by physicians; (r = 0.825, p < 0.001) for the training dataset and (r = 0.844, p < 0.001) for the validation dataset.

CONCLUSION

Our study confirms that 8-point LUS using HHUSD, with AI-based detection of B lines, can provide clinically useful information on the assessment of hydration status and diagnosis of fluid overload for dialysis patients in a user-friendly and time-efficient way.

Lung Ultrasound Induction of Pulmonary Capillary Hemorrhage in Neonatal Swine

This study investigated induction of pulmonary capillary hemorrhage (PCH) in neonatal pigs (piglets) using three different machines: a GE Venue R1 point-of-care system with C1-5 and L4-12t probes, a GE Vivid 7 Dimension with a 7L probe and a SuperSonic Imagine machine with an SL15-4 probe and shear wave elastography (SWE). Female piglets were anesthetized, and each was mounted vertically in a warm bath for scanning at two or three intercostal spaces. After aiming at an innocuous output, the power was raised for a test exposure. Hydrophone measurements were used to calculate in situ values of mechanical index (MIIS). Inflated lungs were removed and scored for PCH area. For the C1-5 probe at 50% and 100% acoustical output (AO), a PCH threshold of 0.53 MIIS was obtained by linear regression (r2 = 0.42). The L4-12t probe did not induce PCH, but the 7L probe induced zones of PCH in the scan planes. The Venue R1 automated B-line tool applied with the C1-5 probe did not detect PCH induced by the C1-5 probe as B-line counts. However, when PCH induced by C1-5 and 7L exposures were subsequently scanned with the L4-12t probe using the automated tool, B-lines were counted in association with the PCH. The SWE induced PCH at push-pulse positions for 3, 30 and 300 s of SWE with PCH accumulating at 0.33 mm2/s and an exponential rise to a maximum of 18.4 mm2 (r2 = 0.61). This study demonstrated the induction of PCH by LUS of piglets, and supports the safety recommendation for use of MIs <0.4 in neonatal LUS.

Implementation and Assessment of Lung Ultrasound Training Curriculum for Physiotherapists With a Focus on Image Acquisition and Calculation of an Aeration Score

Described here is the implementation of a lung ultrasound course for physiotherapists focused on the acquisition and retention of knowledge and skills. Initially, we provided online lectures in a virtual learning environment (VLE), in which we taught the semiquantification of edema through a lung ultrasound score (LUS). Afterward, the physiotherapists participated in face-to-face lectures (which resumed the online lectures), followed by hands-on training and simulation with ultrasound. We assessed knowledge acquisition through a multiple-choice test with 30 questions (totaling 10 points). The test was applied before accessing the VLE (pre-VLE), before the face-to-face course and at its end (pre- and post-course). Physiotherapists collected actual patients' ultrasound scans, which were uploaded to the VLE and assessed by three supervisors, who performed a consensus LUS calculation and gave virtual written feedback. Thirteen physiotherapists collected 59 exams. The test results were 3.60 ± 1.58 (pre-VLE), 5.94 ± 1.45 (pre-course) and 8.50 ± 0.71 (post-course), with p &lt; 0.001 for all. The intraclass correlation coefficient for LUS between physiotherapists and supervisors was 0.814 (p &lt; 0.001), with moderate-to-weak agreement for LUS of the lung apical, median and basal zones, with κ = 0.455.334, and 0.417 (p &lt; 0.001 for all). Trainees were found to have increased short-term acquisition and retention of knowledge and skills, with a good intraclass correlation coefficient between them and the consensus of supervisors for the LUS of actual patients.

Interobserver Agreement and Correlation of an Automated Algorithm for B-Line Identification and Quantification With Expert Sonologist Review in a Handheld Ultrasound Device

OBJECTIVES

B-lines are ultrasound artifacts that can be used to detect a variety of pathologic lung conditions. Computer-aided methods to detect and quantify B-lines may standardize quantification and improve diagnosis by novice users. We sought to test the performance of an automated algorithm for the detection and quantification of B-lines in a handheld ultrasound device (HHUD).

METHODS

Ultrasound images were prospectively collected on adult emergency department patients with dyspnea. Images from the first 124 patients were used for algorithm development. Clips from 80 unique subjects for testing were randomly selected in a predefined proportion of B-lines (0 B-lines, 1-2 B-lines, 3 or more B-lines) and blindly reviewed by five experts using both a manual and reviewer-adjusted process. Intraclass correlation coefficient (ICC) and weighted kappa were used to measure agreement, while an a priori threshold of an ICC (3,k) of 0.75 and precision of 0.3 were used to define adequate performance.

RESULTS

ICC between the algorithm and manual count was 0.84 (95% confidence interval [CI] 0.75-0.90), with a precision of 0.15. ICC between the reviewer-adjusted count and the algorithm count was 0.94 (95% CI 0.90-0.96), and the ICC between the manual and reviewer-adjusted counts was 0.94 (95% CI 0.90-0.96). Weighted kappa was 0.72 (95% CI 0.49-0.95), 0.88 (95% CI 0.74-1), and 0.85 (95% CI 0.89-0.96), respectively.

CONCLUSIONS

This study demonstrates a high correlation between point-of-care ultrasound experts and an automated algorithm to identify and quantify B-lines using an HHUD. Future research may incorporate this HHUD in clinical studies in multiple settings and users of varying experience levels.

Bedside assessment of left atrial pressure in critical care: a multifaceted gem

Evaluating left atrial pressure (LAP) solely from the left ventricular preload perspective is a restrained approach. Accurate assessment of LAP is particularly relevant when pulmonary congestion and/or right heart dysfunction are present since it is the pressure most closely related to pulmonary venous pressure and thus pulmonary haemodynamic load. Amalgamation of LAP measurement into assessment of the ‘transpulmonary circuit’ may have a particular role in differentiating cardiac failure phenotypes in critical care. Most of the literature in this area involves cardiology patients, and gaps of knowledge in application to the bedside of the critically ill patient remain significant. Explored in this review is an overview of left atrial physiology, invasive and non-invasive methods of LAP measurement and their potential clinical application.

Quantitative Assessment of Lung Ultrasound Grayscale Images Based on Shannon Entropy for the Detection of Pulmonary Aeration: An Animal Study

OBJECTIVE

Lung ultrasound (LUS) is a radiation-free, affordable, and bedside monitoring method that can detect changes in pulmonary aeration before hypoxic damage. However, visual scoring methods of LUS only enable subjective diagnosis. Therefore, quantitative analysis of LUS is necessary for obtaining objective information on pulmonary aeration. Because raw data are not always available in conventional ultrasound systems, Shannon entropy (ShanEn) of information theory without the requirement of raw data is valuable. In this study, we explored the feasibility of ShanEn estimated through grayscale histogram (GSH) analysis of LUS images for the quantification of pulmonary aeration.

METHODS

Different degrees of pulmonary aeration caused by edema was induced in 32 male New Zealand rabbits intravenously injected with 0.1 mL/kg saline (the control group) and 0.025, 0.05, and 0.1 mL/kg oleic acid (mild, moderate, and severe groups, respectively). In vivo grayscale LUS images were acquired using a commercial point-of-care ultrasound system for estimation of GSH and corresponding ShanEn. Both lungs of each rabbit were dissected, weighed, and dried to determine the wet weight-to-dry weight ratio (W/D) through gravimetry.

RESULTS

The determination coefficients of linear correlations between ShanEn and W/D increased from 0.0487 to 0.7477 with gain and dynamic range (DR). In contrast to visual scoring methods of pulmonary aeration that use median gain and low DR, ShanEn for quantifying pulmonary aeration requires high gain and DR.

CONCLUSION

The current findings indicate that ShanEn estimated through GSH analysis of LUS images acquired using conventional ultrasonic imaging systems has great potential to provide objective information on pulmonary aeration.

Integrated lung ultrasound score for early clinical decision-making in patients with COVID-19: results and implications

Background and objectives

Lung Ultrasound Score (LUS) identifies and monitors pneumonia by assigning increasing scores. However, it does not include parameters, such as inferior vena cava (IVC) diameter and index of collapse, diaphragmatic excursions and search for pleural and pericardial effusions. Therefore, we propose a new improved scoring system, termed “integrated” lung ultrasound score (i-LUS) which incorporates previously mentioned parameters that can help in prediction of disease severity and survival, choice of oxygenation mode/ventilation and assignment to subsequent areas of care in patients with COVID-19 pneumonia.

Methods

Upon admission at the sub-intensive section of the emergency medical department (SEMD), 143 consecutively examined COVID-19 patients underwent i-LUS together with all other routine analysis. A database for anamnestic information, laboratory data, gas analysis and i-LUS parameters was created and analyzed.

Results

Of 143 enrolled patients, 59.4% were male (mean age 71 years) and 40.6% female. (mean age 79 years: p = 0.005). Patients that survived at 1 month had i-LUS score of 16, which was lower than that of non-survivors (median 20; p = 0.005). Survivors had a higher PaO2/FiO2 (median 321.5) compared to non-survivors (median 229, p < 0.001). There was a correlation between i-LUS and PaO2/FiO2 ratio (rho:-0.4452; p < 0.001), PaO2/FiO2 and survival status (rho:-0.3452; p < 0.001), as well as i-LUS score and disease outcome (rho:0.24; p = 0.005). In non-survivors, the serum values of different significant COVID indicators were severely expressed. The i-LUS score was higher (median 20) in patients who required non-invasive ventilation (NIV) than in those treated only by oxygen therapy (median 15.42; p = 0.003). The odds ratio for death outcome was 1.08 (confidence interval 1.02–1.15) for each point increased. At 1-month follow-up, 65 patients (45.5%) died and 78 (54.5%) survived. Patients admitted to the high critical ward had higher i-LUS score than those admitted to the low critical one (p < 0.003).

Conclusions

i-LUS could be used as a helpful clinical tool for early decision-making in patients with COVID-19 pneumonia.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13089-022-00264-8.

The Role of Lung Ultrasound Monitoring in Early Detection of Ventilator-Associated Pneumonia in COVID-19 Patients: A Retrospective Observational Study

Specific lung ultrasound signs combined with clinical parameters allow for early diagnosis of ventilator-associated pneumonia in the general ICU population. This retrospective cohort study aimed to determine the accuracy of lung ultrasound monitoring for ventilator-associated pneumonia diagnosis in COVID-19 patients. Clinical (i.e., clinical pulmonary infection score) and ultrasound (i.e., presence of consolidation and a dynamic linear−arborescent air bronchogram, lung ultrasound score, ventilator-associated lung ultrasound score) data were collected on the day of the microbiological sample (pneumonia-day) and 48 h before (baseline) on 55 bronchoalveolar lavages of 33 mechanically-ventilated COVID-19 patients who were monitored daily with lung ultrasounds. A total of 26 samples in 23 patients were positive for ventilator-associated pneumonia (pneumonia cases). The onset of a dynamic linear−arborescent air bronchogram was 100% specific for ventilator-associated pneumonia. The ventilator-associated lung ultrasound score was higher in pneumonia-cases (2.5 (IQR 1.0 to 4.0) vs. 1.0 (IQR 1.0 to 1.0); p < 0.001); the lung ultrasound score increased from baseline in pneumonia-cases only (3.5 (IQR 2.0 to 6.0) vs. −1.0 (IQR −2.0 to 1.0); p = 0.0001). The area under the curve for clinical parameters, ventilator-associated pneumonia lung ultrasound score, and lung ultrasound score variations were 0.472, 0.716, and 0.800, respectively. A newly appeared dynamic linear−arborescent air bronchogram is highly specific for ventilator-associated pneumonia in COVID-19 patients. A high ventilator-associated pneumonia lung ultrasound score (or an increase in the lung ultrasound score) orients to ventilator-associated pneumonia.

Visual Rounds Based on Multiorgan Point-of-Care Ultrasound in the ICU

Point-of-care ultrasonography (POCUS) is performed by a treating clinician at the patient's bedside, provides a acquisition, interpretation, and immediate clinical integration based on ultrasonographic imaging. The use of POCUS is not limited to one specialty, protocol, or organ system. POCUS provides the treating clinician with real-time diagnostic and monitoring information. Visual rounds based on multiorgan POCUS act as an initiative to improve clinical practice in the Intensive Care Unit and are urgently needed as part of routine clinical practice.

Evaluation of Pulmonary Edema Using Ultrasound Imaging in Patients With COVID-19 Pneumonia Based on a Non-local Channel Attention ResNet

Recent research has revealed that COVID-19 pneumonia is often accompanied by pulmonary edema. Pulmonary edema is a manifestation of acute lung injury (ALI), and may progress to hypoxemia and potentially acute respiratory distress syndrome (ARDS), which have higher mortality. Precise classification of the degree of pulmonary edema in patients is of great significance in choosing a treatment plan and improving the chance of survival. Here we propose a deep learning neural network named Non-local Channel Attention ResNet to analyze the lung ultrasound images and automatically score the degree of pulmonary edema of patients with COVID-19 pneumonia. The proposed method was designed by combining the ResNet with the non-local module and the channel attention mechanism. The non-local module was used to extract the information on characteristics of A-lines and B-lines, on the basis of which the degree of pulmonary edema could be defined. The channel attention mechanism was used to assign weights to decisive channels. The data set contains 2220 lung ultrasound images provided by Huoshenshan Hospital, Wuhan, China, of which 2062 effective images with accurate scores assigned by two experienced clinicians were used in the experiment. The experimental results indicated that our method achieved high accuracy in classifying the degree of pulmonary edema in patients with COVID-19 pneumonia by comparison with previous deep learning methods, indicating its potential to monitor patients with COVID-19 pneumonia.

Automated lung ultrasound scoring for evaluation of coronavirus disease 2019 pneumonia using two-stage cascaded deep learning model

Coronavirus disease 2019 (COVID-19) pneumonia has erupted worldwide, causing massive population deaths and huge economic losses. In clinic, lung ultrasound (LUS) plays an important role in the auxiliary diagnosis of COVID-19 pneumonia. However, the lack of medical resources leads to the low using efficiency of the LUS, to address this problem, a novel automated LUS scoring system for evaluating COVID-19 pneumonia based on the two-stage cascaded deep learning model was proposed in this paper. 18,330 LUS images collected from 26 COVID-19 pneumonia patients were successfully assigned scores by two experienced doctors according to the designed four-level scoring standard for training the model. At the first stage, we made a secondary selection of these scored images through five ResNet-50 models and five-fold cross validation to obtain the available 12,949 LUS images which were highly relevant to the initial scoring results. At the second stage, three deep learning models including ResNet-50, Vgg-19, and GoogLeNet were formed the cascaded scored model and trained using the new dataset, whose predictive result was obtained by the voting mechanism. In addition, 1000 LUS images collected another 5 COVID-19 pneumonia patients were employed to test the model. Experiments results showed that the automated LUS scoring model was evaluated in terms of accuracy, sensitivity, specificity, and F1-score, being 96.1%, 96.3%, 98.8%, and 96.1%, respectively. They proved the proposed two-stage cascaded deep learning model could automatically score an LUS image, which has great potential for application to the clinics on various occasions.

Ten conditions where lung ultrasonography may fail: limits, pitfalls and lessons learned from a computer-aided algorithmic approach

Lung ultrasonography provides relevant information on morphological and functional changes occurring in the lungs. However, it correlates weakly with pulmonary congestion and extra vascular lung water. Moreover, there is lack of consensus on scoring systems and acquisition protocols. The automation of this technique may provide promising easy-to-use clinical tools to reduce inter- and intra-observer variability and to standardize scores, allowing faster data collection without increased costs and patients risks.

Lung ultrasound to detect cardiopulmonary interactions in acutely ill children

OBJECTIVE AND DESIGN

Our prospective observational study is the first study that evaluates the lung ultrasound (LUS) findings of cardiopulmonary interactions in acutely ill children with elevated pro-brain natriuretic peptide (BNP) levels, with the aim of establishing the specific LUS pattern in this category of patients without primary lung diseases.

METHODOLOGY

We prospectively analyzed epidemiological, clinical, laboratory, instrumental, and lung ultrasound parameters in acutely ill children aged 1 month to 18 years admitted to the Department of Pediatrics between March 2020 to August 2020. Among the acutely ill patients evaluated, only patients with pro-BNP > 300 pg/ml and who underwent LUS before the start of any treatment were included. They were stratified into three subcategories based on the diagnosis (A) cardiac disease, (B) systemic inflammatory disease/sepsis without functional and/or organic alterations of the myocardium, and (C) systemic inflammatory disease/sepsis and cardiac disease, and were classified into two groups based on the level of pro-BNP. We also enrolled patients belonging to two other categories (patients with primary infectious lung disease and completely healthy patients) analyzing their epidemiological, clinical, laboratory, instrumental parameters, and lung ultrasound findings and comparing them with those of acutely ill children.

RESULTS AND CONCLUSION

We found that LUS findings in these acutely ill children are different from the ultrasound pattern of other categories of children and in particular (1) children with acute lower respiratory tract infections and (2) healthy infants. The finding in a child of a sonographic interstitial syndrome with multiple, bright, long, separate, and nonconfluent B-lines/long vertical artifacts deriving from a normal and regular pleural line, in the absence of subpleural consolidations, is strongly predictive of cardiogenic pulmonary edema or pulmonary congestion in the course of systemic inflammatory disease/sepsis.

Automated versus manual B-lines counting, left ventricular outflow tract velocity time integral and inferior vena cava collapsibility index in COVID-19 patients

Background and Aims:

The incorporation of artificial intelligence (AI) in point-of-care ultrasound (POCUS) has become a very useful tool to quickly assess cardiorespiratory function in coronavirus disease (COVID)-19 patients. The objective of this study was to test the agreement between manual and automated B-lines counting, left ventricular outflow tract velocity time integral (LVOT-VTI) and inferior vena cava collapsibility index (IVC-CI) in suspected or confirmed COVID-19 patients using AI integrated POCUS. In addition, we investigated the inter-observer, intra-observer variability and reliability of assessment of echocardiographic parameters using AI by a novice.

Methods:

Two experienced sonographers in POCUS and one novice learner independently and consecutively performed ultrasound assessment of B-lines counting, LVOT-VTI and IVC-CI in 83 suspected and confirmed COVID-19 cases which included both manual and AI methods.

Results:

Agreement between automated and manual assessment of LVOT-VTI, and IVC-CI were excellent [intraclass correlation coefficient (ICC) 0.98, P < 0.001]. Intra-observer reliability and inter-observer reliability of these parameters were excellent [ICC 0.96-0.99, P < 0.001]. Moreover, agreement between novice and experts using AI for LVOT-VTI and IVC-CI assessment was also excellent [ICC 0.95-0.97, P < 0.001]. However, correlation and intra-observer reliability between automated and manual B-lines counting was moderate [(ICC) 0.52-0.53, P < 0.001] and [ICC 0.56-0.69, P < 0.001], respectively. Inter-observer reliability was good [ICC 0.79-0.87, P < 0.001]. Agreement of B-lines counting between novice and experts using AI was weak [ICC 0.18, P < 0.001].

Conclusion:

AI-guided assessment of LVOT-VTI, IVC-CI and B-lines counting is reliable and consistent with manual assessment in COVID-19 patients. Novices can reliably estimate LVOT-VTI and IVC-CI using AI software in COVID-19 patients.

Lung Ultrasound in Pediatric Acute Respiratory Distress Syndrome Received Extracorporeal Membrane Oxygenation: A Prospective Cohort Study

OBJECTIVE

The aim of this study was to assess the prognostic value of the lung ultrasound (LUS) score in patients with pediatric acute respiratory distress syndrome (pARDS) who received extracorporeal membrane oxygenation (ECMO).

METHODS

A prospective cohort study was conducted in a pediatric intensive care unit (PICU) of a tertiary hospital from January 2016 to June 2021. The severe pARDS patients who received ECMO were enrolled in this study. LUS score was measured at initiation of ECMO (LUS-0 h), then at 24 h (LUS-24 h), 48 h (LUS-48 h), and 72 h (LUS-72 h) during ECMO, and when weaned from ECMO (LUS-wean). The value of LUS scores at the first 3 days of ECMO as a prognostic predictor was analyzed.

RESULTS

Twenty-nine children with severe pARDS who received ECMO were enrolled with a median age of 26 (IQR 9, 79) months. The median duration of ECMO support was 162 (IQR 86, 273) h and the PICU mortality was 31.0% (9/29). The values of LUS-72 h and LUS-wean were significantly lower in survivors than that in non-survivors (both P < 0.001). Daily fluid balance volume during the first 3 days of ECMO support were strongly correlated with LUS score [1st day: r = 0.460, P = 0.014; 2nd day: r = 0.540, P = 0.003; 3rd day: r = 0.589, P = 0.001]. The AUC of LUS-72 h for predicting PICU mortality in these patients was 1.000, and the cutoff value of LUS-72 h was 24 with a sensitivity of 100.0% and a specificity of 100.0%. Furthermore, patients were stratified in two groups of LUS-72 h ≥ 24 and LUS-72 h < 24. Consistently, PICU mortality, length of PICU stay, ratio of shock, vasoactive index score value, and the need for continuous renal replacement therapy were significantly higher in the group of LUS-72 h ≥ 24 than in the group of LUS-72 h < 24 (all P < 0.05).

CONCLUSION

Lung ultrasound score is a promising tool for predicting the prognosis in patients with ARDS under ECMO support. Moreover, LUS-72 h ≥ 24 is associated with high risk of PICU mortality in patients with pARDS who received ECMO.

Quantitative Analysis of Pleural Line and B-Lines in Lung Ultrasound Images for Severity Assessment of COVID-19 Pneumonia

Specific patterns of lung ultrasound (LUS) images are used to assess the severity of coronavirus disease 2019 (COVID-19) pneumonia, while such assessment is mainly based on clinicians' qualitative and subjective observations. In this study, we quantitatively analyze the LUS images to assess the severity of COVID-19 pneumonia by characterizing the patterns related to the pleural line (PL) and B-lines (BLs). Twenty-seven patients with COVID-19 pneumonia, including 13 moderate cases, seven severe cases, and seven critical cases, are enrolled. Features related to the PL, including the thickness (TPL) and roughness of the PL (RPL), and the mean (MPLI) and standard deviation (SDPLI) of the PL intensities are extracted from the LUS images. Features related to the BLs, including the number (NBL), accumulated width (AWBL), attenuation coefficient (ACBL), and accumulated intensity (AIBL) of BLs, are also extracted. The correlations of these features with the disease severity are evaluated. The performances of the binary severe/non-severe classification are assessed for each feature and support vector machine (SVM) classifiers with various combinations of features as input. Several features, including the RPL, NBL, AWBL, and AIBL, show significant correlations with disease severity (all ). The classification performance is optimal using the SVM classifier using all the features as input (area under the receiver operating characteristic (ROC) curve = 0.96, sensitivity = 0.93, and specificity = 1). These findings demonstrate that the proposed method may be a promising tool for automatic grading diagnosis and follow-up of patients with COVID-19 pneumonia.

ERS International Congress 2020 Virtual: highlights from the Respiratory Intensive Care Assembly

During the virtual European Respiratory Society Congress 2020, early career members summarised the sessions organised by the Respiratory Intensive Care Assembly. The topics covered included diagnostic strategies in patients admitted to the intensive care unit with acute respiratory failure, with a focus on patients with interstitial lung disease and for obvious reasons, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. These sessions are summarised in this article, with take-home messages highlighted.

Updates from #ERSCongress 2020 on diagnostic strategies in patients admitted to the ICU with acute respiratory failure and on the management of #SARSCoV2 infectionhttps://bit.ly/38cx0Pi

Interobserver Agreement of Lung Ultrasound Findings of COVID-19

BACKGROUND

Lung ultrasound (LUS) has received considerable interest in the clinical evaluation of patients with COVID-19. Previously described LUS manifestations for COVID-19 include B-lines, consolidations, and pleural thickening. The interrater reliability (IRR) of these findings for COVID-19 is unknown.

METHODS

This study was conducted between March and June 2020. Nine physicians (hospitalists: n = 4; emergency medicine: n = 5) from 3 medical centers independently evaluated n = 20 LUS scans (n = 180 independent observations) collected from patients with COVID-19, diagnosed via RT-PCR. These studies were randomly selected from an image database consisting of COVID-19 patients evaluated in the emergency department with portable ultrasound devices. Physicians were blinded to any patient information or previous LUS interpretation. Kappa values (κ) were used to calculate IRR.

RESULTS

There was substantial IRR on the following items: normal LUS scan (κ = 0.79 [95% CI: 0.72-0.87]), presence of B-lines (κ = 0.79 [95% CI: 0.72-0.87]), ≥3 B-lines observed (κ = 0.72 [95% CI: 0.64-0.79]). Moderate IRR was observed for the presence of any consolidation (κ = 0.57 [95% CI: 0.50-0.64]), subpleural consolidation (κ = 0.49 [95% CI: 0.42-0.56]), and presence of effusion (κ = 0.49 [95% CI: 0.41-0.56]). Fair IRR was observed for pleural thickening (κ = 0.23 [95% CI: 0.15-0.30]).

DISCUSSION

Many LUS manifestations for COVID-19 appear to have moderate to substantial IRR across providers from multiple specialties utilizing differing portable devices. The most reliable LUS findings with COVID-19 may include the presence/count of B-lines or determining if a scan is normal. Clinical protocols for LUS with COVID-19 may require additional observers for the confirmation of less reliable findings such as consolidations.

An Interpretable Object Detection-Based Model For The Diagnosis Of Neonatal Lung Diseases Using Ultrasound Images

Over the last few decades, Lung Ultrasound (LUS) has been increasingly used to diagnose and monitor different lung diseases in neonates. It is a noninvasive tool that allows a fast bedside examination while minimally handling the neonate. Acquiring a LUS scan is easy, but understanding the artifacts concerned with each respiratory disease is challenging. Mixed artifact patterns found in different respiratory diseases may limit LUS readability by the operator. While machine learning (ML), especially deep learning can assist in automated analysis, simply feeding the ultrasound images to an ML model for diagnosis is not enough to earn the trust of medical professionals. The algorithm should output LUS features that are familiar to the operator instead. Therefore, in this paper we present a unique approach for extracting seven meaningful LUS features that can be easily associated with a specific pathological lung condition: Normal pleura, irregular pleura, thick pleura, A- lines, Coalescent B-lines, Separate B-lines and Consolidations. These artifacts can lead to early prediction of infants developing later respiratory distress symptoms. A single multi-class region proposal-based object detection model faster-RCNN (fRCNN) was trained on lower posterior lung ultrasound videos to detect these LUS features which are further linked to four common neonatal diseases. Our results show that fRCNN surpasses single stage models such as RetinaNet and can successfully detect the aforementioned LUS features with a mean average precision of 86.4%. Instead of a fully automatic diagnosis from images without any interpretability, detection of such LUS features leave the ultimate control of diagnosis to the clinician, which can result in a more trustworthy intelligent system.

A new standard scoring for interstitial pneumonia based on quantitative analysis of ultrasonographic data: A study on COVID-19 patients

OBJECTIVE

To assess the effectiveness of 3 novel lung ultrasound (LUS)-based parameters: Pneumonia Score and Lung Staging for pneumonia staging and COVID Index, indicating the probability of SARS-CoV-2 infection.

METHODS

Adult patients admitted to the emergency department with symptoms potentially related to pneumonia, healthy volunteers and clinical cases from online accessible databases were evaluated. The patients underwent a clinical-epidemiological questionnaire and a LUS acquisition, following a 14-zone protocol. For each zone, a Pneumonia score from 0 to 4 was assigned by the algorithm and by an expert operator (kept blind with respect to the algorithm results) on the basis of the identified imaging signs and the patient Lung Staging was derived as the highest observed score. The output of the operator was considered as the ground truth. The algorithm calculated also the COVID Index by combining the automatically identified LUS markers with the questionnaire answers and compared with the nasopharyngeal swab results.

RESULTS

Overall, 556 patients were analysed. A high agreement between the algorithm assignments and the expert operator evaluations was observed, both for Pneumonia Score and Lung Staging, with the latter having sensitivity and specificity over 92% both in the discrimination between healthy/sick patients and between sick patients with mild/severe pneumonia. Regarding the COVID Index, an area under the curve of 0.826 was observed for the classification of patients with/without SARS-CoV-2.

CONCLUSION

The proposed methodology allowed the identification and staging of patients suffering from pneumonia with high accuracy. Moreover, it provided the probability of being infected by SARS-CoV-2.

Lung Ultrasound Assessment of Focal and Non-focal Lung Morphology in Patients With Acute Respiratory Distress Syndrome

Background: The identification of phenotypes based on lung morphology can be helpful to better target mechanical ventilation of individual patients with acute respiratory distress syndrome (ARDS). We aimed to assess the accuracy of lung ultrasound (LUS) methods for classification of lung morphology in critically ill ARDS patients under mechanical ventilation.

Methods: This was a post hoc analysis on two prospective studies that performed LUS and chest computed tomography (CT) scanning at the same time. Expert panels from the two participating centers separately developed two LUS methods for classifying lung morphology based on LUS aeration scores from a 12-region exam (Amsterdam and Lombardy method). Moreover, a previously developed LUS method based on anterior LUS scores was tested (Piedmont method). Sensitivity and specificity of all three LUS methods was assessed in the cohort of the other center(s) by using CT as the gold standard for classification of lung morphology.

Results: The Amsterdam and Lombardy cohorts consisted of 32 and 19 ARDS patients, respectively. From these patients, 23 (45%) had focal lung morphology while others had non-focal lung morphology. The Amsterdam method could classify focal lung morphology with a sensitivity of 77% and a specificity of 100%, while the Lombardy method had a sensitivity and specificity of 100 and 61%. The Piedmont method had a sensitivity and specificity of 91 and 75% when tested on both cohorts. With both the Amsterdam and Lombardy method, most patients could be classified based on the anterior regions alone.

Conclusion: LUS-based methods can accurately classify lung morphology in invasively ventilated ARDS patients compared to gold standard chest CT. The anterior LUS regions showed to be the most discriminant between focal and non-focal lung morphology, although accuracy increased moderately when lateral and posterior LUS regions were integrated in the method.

Lung Ultrasound Integration in Assessment of Patients with Noncritical COVID-19

OBJECTIVES

Performing lung ultrasound during the clinical assessment of patients with suspicion of noncritical COVID-19 may increase the diagnostic rate of pulmonary involvement over other diagnostic techniques used in routine clinical practice. This study aims to compare complications (readmissions, emergency department [ED] visits, and length of outpatient follow-up) in the first 30 days after ED discharge in patients with confirmed COVID-19 who were managed with versus without lung ultrasound.

MATERIALS AND METHODS

Prospective, observational, analytical study in noncritical patients with confirmed respiratory disease due to SARS-CoV-2, assessed in the ED of a tertiary Spanish hospital in March and April 2020. We compared 2 cohorts, differentiated by the use of lung ultrasound as a diagnostic tool. Complications were assessed (hospital admissions, ED revisits and days of outpatient follow-up) at 30 days postdischarge.

RESULTS

Of the 88 included patients, 31% (n = 27) underwent an initial lung ultrasound, while 61 (68%) did not. In 82.5% of the patients evaluated with ultrasound, the most predominant areas affected were the posterobasal regions, in the form of focalized and confluent B-lines; 70.4% showed pleural irregularity in these same areas. Use of the lung ultrasound was associated with a greater probability of hospital admission (odds ratio 5.63, 95% confidence interval 3.31 to 9.57; p < 0.001). However, it was not significantly associated with mortality or short-term complications.

CONCLUSIONS

Lung ultrasound could identify noncritical patients with lung impairment due to SARS-CoV-2, in whom other tests used routinely show no abnormalities. However, it has not shown a prognostic value in these patients and could generate a higher percentage of hospital admissions. More studies are still needed to demonstrate the clear benefit of this use.

Quantitative Lung Ultrasound: Technical Aspects and Clinical Applications

Lung ultrasound is increasingly used in emergency departments, medical wards, and critical care units-adult, pediatric, and neonatal. In vitro and in vivo studies show that the number and type of artifacts visualized change with lung density. This has led to the idea of a quantitative lung ultrasound approach, opening up new prospects for use not only as a diagnostic but also as a monitoring tool. Consequently, the multiple scoring systems proposed in the last few years have different technical approaches and specific clinical indications, adaptable for more or less time-dependent patients. However, multiple scoring systems may generate confusion among physicians aiming at introducing lung ultrasound in their clinical practice. This review describes the various lung ultrasound scoring systems and aims to clarify their use in different settings, focusing on technical aspects, validation with reference techniques, and clinical applications.

Quantitative Analysis and Automated Lung Ultrasound Scoring for Evaluating COVID-19 Pneumonia With Neural Networks

As being radiation-free, portable, and capable of repetitive use, ultrasonography is playing an important role in diagnosing and evaluating the COVID-19 Pneumonia (PN) in this epidemic. By virtue of lung ultrasound scores (LUSS), lung ultrasound (LUS) was used to estimate the excessive lung fluid that is an important clinical manifestation of COVID-19 PN, with high sensitivity and specificity. However, as a qualitative method, LUSS suffered from large interobserver variations and requirement for experienced clinicians. Considering this limitation, we developed a quantitative and automatic lung ultrasound scoring system for evaluating the COVID-19 PN. A total of 1527 ultrasound images prospectively collected from 31 COVID-19 PN patients with different clinical conditions were evaluated and scored with LUSS by experienced clinicians. All images were processed via a series of computer-aided analysis, including curve-to-linear conversion, pleural line detection, region-of-interest (ROI) selection, and feature extraction. A collection of 28 features extracted from the ROI was specifically defined for mimicking the LUSS. Multilayer fully connected neural networks, support vector machines, and decision trees were developed for scoring LUS images using the fivefold cross validation. The model with 128×256 two fully connected layers gave the best accuracy of 87%. It is concluded that the proposed method could assess the ultrasound images by assigning LUSS automatically with high accuracy, potentially applicable to the clinics.

B-line quantification: comparing learners novice to lung ultrasound assisted by machine artificial intelligence technology to expert review

BACKGROUND

The goal of this study was to assess the ability of machine artificial intelligence (AI) to quantitatively assess lung ultrasound (LUS) B-line presence using images obtained by learners novice to LUS in patients with acute heart failure (AHF), compared to expert interpretation.

METHODS

This was a prospective, multicenter observational study conducted at two urban academic institutions. Learners novice to LUS completed a 30-min training session on lung image acquisition which included lecture and hands-on patient scanning. Learners independently acquired images on patients with suspected AHF. Automatic B-line quantification was obtained offline after completion of the study. Machine AI counted the maximum number of B-lines visualized during a clip. The criterion standard for B-line counts was semi-quantitative analysis by a blinded point-of-care LUS expert reviewer. Image quality was blindly determined by an expert reviewer. A second expert reviewer blindly determined B-line counts and image quality. Intraclass correlation was used to determine agreement between machine AI and expert, and expert to expert.

RESULTS

Fifty-one novice learners completed 87 scans on 29 patients. We analyzed data from 611 lung zones. The overall intraclass correlation for agreement between novice learner images post-processed with AI technology and expert review was 0.56 (confidence interval [CI] 0.51-0.62), and 0.82 (CI 0.73-0.91) between experts. Median image quality was 4 (on a 5-point scale), and correlation between experts for quality assessment was 0.65 (CI 0.48-0.82).

CONCLUSION

After a short training session, novice learners were able to obtain high-quality images. When the AI deep learning algorithm was applied to those images, it quantified B-lines with moderate-to-fair correlation as compared to semi-quantitative analysis by expert review. This data shows promise, but further development is needed before widespread clinical use.

A-lines and B-lines in patients with acute heart failure

AIMS

Lung ultrasound (LUS) relies on detecting artefacts, including A-lines and B-lines, when assessing dyspnoeic patients. A-lines are horizontal artefacts and characterize normal lung, whereas multiple vertical B-lines are associated with increased lung density. We sought to assess the prevalence of A-lines and B-lines in patients with acute heart failure (AHF) and examine their clinical correlates and their relationship with outcomes.

METHODS AND RESULTS

In a prospective cohort study of adults with AHF, eight-zone LUS and echocardiography were performed early during the hospitalization and pre-discharge at an imaging depth of 18 cm. A- and B-lines were analysed separately off-line, blinded to clinical and outcome data. Of 164 patients [median age 71 years, 61% men, mean ejection fraction (EF) 40%], the sum of A-lines at baseline ranged from 0 to 19 and B-line number from 0 to 36. One hundred and fifty-six patients (95%) had co-existing A-lines and B-lines at baseline. Lower body mass index and lower chest wall thickness were associated with a higher number of A-lines (P trend < 0.001 for both). In contrast to B-lines, there was no significant change in the number of A-lines from baseline to discharge (median 6 vs. 5, P = 0.80). While B-lines were associated with 90-day HF readmission or death, A-lines were not [HR 1.67, 95% confidence interval (CI) 1.11-2.51 vs. HR 0.97, 95% CI 0.65-1.43].

CONCLUSIONS

A-lines and B-lines on LUS co-exist in the vast majority of hospitalized patients with AHF. In contrast to B-lines, A-lines were not associated with adverse outcomes.

Lung, Heart, Vascular, and Diaphragm Ultrasound Examination of COVID-19 Patients: A Comprehensive Approach

Lung ultrasound (LU) has a multitude of features and capacities that make it a useful medical tool to assist physicians contending with the pandemic spread of novel coronavirus disease-2019 (COVID-19) caused by coronavirus severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Thus, an LU approach to patients with suspected COVID-19 is being implemented worldwide. In noncritical COVID-19 patients, 2 new LU signs have been described and proposed, the "waterfall" and the "light beam" signs. Both signs have been hypothesized to increase the diagnostic accuracy of LU for COVID-19 interstitial pneumonia. In critically ill patients, a distinct pattern of LU changes seems to follow the disease's progression, and this information can be used to guide decisions about when a patient needs to be ventilated, as occurs in other disease states similar to COVID-19. Furthermore, a new algorithm has been published, which enables the automatic detection of B-lines as well as quantification of the percentage of the pleural line associated with lung disease. In COVID-19 patients, a direct involvement of cardiac function has been demonstrated, and ventilator-induced diaphragm dysfunction might be present due to the prolonged mechanical ventilation often involved, as reported for similar diseases. For this reason, cardiac and diaphragm ultrasound evaluation are highly important. Last but not least, due to the thrombotic tendency of COVID-19 patients, particular attention also should be paid to vascular ultrasound. This review is primarily devoted to the study of LU in COVID-19 patients. The authors explain the significance of its "light and shadows," bearing in mind the context in which LU is being used-the emergency department and the intensive care setting. The use of cardiac, vascular, and diaphragm ultrasound is also discussed, as a comprehensive approach to patient care.

Evaluation of lung involvement in COVID-19 pneumonia based on ultrasound images

BACKGROUND

Lung ultrasound (LUS) can be an important imaging tool for the diagnosis and assessment of lung involvement. Ultrasound sonograms have been confirmed to illustrate damage to a person's lungs, which means that the correct classification and scoring of a patient's sonogram can be used to assess lung involvement.

METHODS

The purpose of this study was to establish a lung involvement assessment model based on deep learning. A novel multimodal channel and receptive field attention network combined with ResNeXt (MCRFNet) was proposed to classify sonograms, and the network can automatically fuse shallow features and determine the importance of different channels and respective fields. Finally, sonogram classes were transformed into scores to evaluate lung involvement from the initial diagnosis to rehabilitation.

RESULTS AND CONCLUSION

Using multicenter and multimodal ultrasound data from 104 patients, the diagnostic model achieved 94.39% accuracy, 82.28% precision, 76.27% sensitivity, and 96.44% specificity. The lung involvement severity and the trend of COVID-19 pneumonia were evaluated quantitatively.

Sonographic Diagnosis of COVID-19: A Review of Image Processing for Lung Ultrasound

The sustained increase in new cases of COVID-19 across the world and potential for subsequent outbreaks call for new tools to assist health professionals with early diagnosis and patient monitoring. Growing evidence around the world is showing that lung ultrasound examination can detect manifestations of COVID-19 infection. Ultrasound imaging has several characteristics that make it ideally suited for routine use: small hand-held systems can be contained inside a protective sheath, making it easier to disinfect than X-ray or computed tomography equipment; lung ultrasound allows triage of patients in long term care homes, tents or other areas outside of the hospital where other imaging modalities are not available; and it can determine lung involvement during the early phases of the disease and monitor affected patients at bedside on a daily basis. However, some challenges still remain with routine use of lung ultrasound. Namely, current examination practices and image interpretation are quite challenging, especially for unspecialized personnel. This paper reviews how lung ultrasound (LUS) imaging can be used for COVID-19 diagnosis and explores different image processing methods that have the potential to detect manifestations of COVID-19 in LUS images. Then, the paper reviews how general lung ultrasound examinations are performed before addressing how COVID-19 manifests itself in the images. This will provide the basis to study contemporary methods for both segmentation and classification of lung ultrasound images. The paper concludes with a discussion regarding practical considerations of lung ultrasound image processing use and draws parallels between different methods to allow researchers to decide which particular method may be best considering their needs. With the deficit of trained sonographers who are working to diagnose the thousands of people afflicted by COVID-19, a partially or totally automated lung ultrasound detection and diagnosis tool would be a major asset to fight the pandemic at the front lines.

Development of a convolutional neural network to differentiate among the etiology of similar appearing pathological B lines on lung ultrasound: a deep learning study

OBJECTIVES

Lung ultrasound (LUS) is a portable, low-cost respiratory imaging tool but is challenged by user dependence and lack of diagnostic specificity. It is unknown whether the advantages of LUS implementation could be paired with deep learning (DL) techniques to match or exceed human-level, diagnostic specificity among similar appearing, pathological LUS images.

DESIGN

A convolutional neural network (CNN) was trained on LUS images with B lines of different aetiologies. CNN diagnostic performance, as validated using a 10% data holdback set, was compared with surveyed LUS-competent physicians.

SETTING

Two tertiary Canadian hospitals.

PARTICIPANTS

612 LUS videos (121 381 frames) of B lines from 243 distinct patients with either (1) COVID-19 (COVID), non-COVID acute respiratory distress syndrome (NCOVID) or (3) hydrostatic pulmonary edema (HPE).

RESULTS

The trained CNN performance on the independent dataset showed an ability to discriminate between COVID (area under the receiver operating characteristic curve (AUC) 1.0), NCOVID (AUC 0.934) and HPE (AUC 1.0) pathologies. This was significantly better than physician ability (AUCs of 0.697, 0.704, 0.967 for the COVID, NCOVID and HPE classes, respectively), p<0.01.

CONCLUSIONS

A DL model can distinguish similar appearing LUS pathology, including COVID-19, that cannot be distinguished by humans. The performance gap between humans and the model suggests that subvisible biomarkers within ultrasound images could exist and multicentre research is merited.

Lung ultrasound and computed tomography to monitor COVID-19 pneumonia in critically ill patients: a two-center prospective cohort study

BACKGROUND

Lung ultrasound can adequately monitor disease severity in pneumonia and acute respiratory distress syndrome. We hypothesize lung ultrasound can adequately monitor COVID-19 pneumonia in critically ill patients.

METHODS

Adult patients with COVID-19 pneumonia admitted to the intensive care unit of two academic hospitals who underwent a 12-zone lung ultrasound and a chest CT examination were included. Baseline characteristics, and outcomes including composite endpoint death or ICU stay > 30 days were recorded. Lung ultrasound and CT images were quantified as a lung ultrasound score involvement index (LUSI) and CT severity involvement index (CTSI). Primary outcome was the correlation, agreement, and concordance between LUSI and CTSI. Secondary outcome was the association of LUSI and CTSI with the composite endpoints.

RESULTS

We included 55 ultrasound examinations in 34 patients, which were 88% were male, with a mean age of 63 years and mean P/F ratio of 151. The correlation between LUSI and CTSI was strong (r = 0.795), with an overall 15% bias, and limits of agreement ranging - 40 to 9.7. Concordance between changes in sequentially measured LUSI and CTSI was 81%. In the univariate model, high involvement on LUSI and CTSI were associated with a composite endpoint. In the multivariate model, LUSI was the only remaining independent predictor.

CONCLUSIONS

Lung ultrasound can be used as an alternative for chest CT in monitoring COVID-19 pneumonia in critically ill patients as it can quantify pulmonary involvement, register changes over the course of the disease, and predict death or ICU stay > 30 days.

TRIAL REGISTRATION

NTR, NL8584. Registered 01 May 2020-retrospectively registered, https://www.trialregister.nl/trial/8584.

The role of ultrasonographic lung aeration score in the prediction of postoperative pulmonary complications: an observational study

Background

Postoperative pulmonary complications (PPCs) are important contributors to mortality and morbidity after surgery. The available predicting models are useful in preoperative risk assessment, but there is a need for validated tools for the early postoperative period as well. Lung ultrasound is becoming popular in intensive and perioperative care and there is a growing interest to evaluate its role in the detection of postoperative pulmonary pathologies.

Objectives

We aimed to identify characteristics with the potential of recognizing patients at risk by comparing the lung ultrasound scores (LUS) of patients with/without PPC in a 24-h postoperative timeframe.

Methods

Observational study at a university clinic. We recruited ASA 2–3 patients undergoing elective major abdominal surgery under general anaesthesia. LUS was assessed preoperatively, and also 1 and 24 h after surgery. Baseline and operative characteristics were also collected. A one-week follow up identified PPC+ and PPC- patients. Significantly differing LUS values underwent ROC analysis. A multi-variate logistic regression analysis with forward stepwise model building was performed to find independent predictors of PPCs.

Results

Out of the 77 recruited patients, 67 were included in the study. We evaluated 18 patients in the PPC+ and 49 in the PPC- group. Mean ages were 68.4 ± 10.2 and 66.4 ± 9.6 years, respectively (p = 0.4829). Patients conforming to ASA 3 class were significantly more represented in the PPC+ group (66.7 and 26.5%; p = 0.0026). LUS at baseline and in the postoperative hour were similar in both populations. The median LUS at 0 h was 1.5 (IQR 1–2) and 1 (IQR 0–2; p = 0.4625) in the PPC+ and PPC- groups, respectively. In the first postoperative hour, both groups had a marked increase, resulting in scores of 6.5 (IQR 3–9) and 5 (IQR 3–7; p = 0.1925). However, in the 24th hour, median LUS were significantly higher in the PPC+ group (6; IQR 6–10 vs 3; IQR 2–4; p < 0.0001) and it was an independent risk factor (OR = 2.6448 CI95% 1.5555–4.4971; p = 0.0003). ROC analysis identified the optimal cut-off at 5 points with high sensitivity (0.9444) and good specificity (0.7755).

Conclusion

Postoperative LUS at 24 h can identify patients at risk of or in an early phase of PPCs.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12871-021-01236-6.

Good clinical practice for the use of vasopressor and inotropic drugs in critically ill patients: state-of-the-art and expert consensus

Vasopressors and inotropic agents are widely used in critical care. However, strong evidence supporting their use in critically ill patients is lacking in many clinical scenarios. Thus, the Italian Society of Anesthesia and Intensive Care (SIAARTI) promoted a project aimed to provide indications for good clinical practice on the use of vasopressors and inotropes, and on the management of critically ill patients with shock. A panel of 16 experts in the field of intensive care medicine and hemodynamics has been established. Systematic review of the available literature was performed based on PICO questions. Basing on available evidence, the panel prepared a summary of evidence and then wrote the clinical questions. A modified semi-quantitative RAND/UCLA appropriateness method has been used to determine the appropriateness of specific clinical scenarios. The panel identified 29 clinical questions for the use of vasopressors and inotropes in patients with septic shock and cardiogenic shock. High level of agreement exists among the panel members about appropriateness of inotropes/vasopressors' use in patients with septic shock and cardiogenic shock.