Quantitative assessment of COVID-19 pneumonia in neonates using lung ultrasound score

Review on chest pathogies detection systems using deep learning techniques

Chest radiography is the standard and most affordable way to diagnose, analyze, and examine different thoracic and chest diseases. Typically, the radiograph is examined by an expert radiologist or physician to decide about a particular anomaly, if exists. Moreover, computer-aided methods are used to assist radiologists and make the analysis process accurate, fast, and more automated. A tremendous improvement in automatic chest pathologies detection and analysis can be observed with the emergence of deep learning. The survey aims to review, technically evaluate, and synthesize the different computer-aided chest pathologies detection systems. The state-of-the-art of single and multi-pathologies detection systems, which are published in the last five years, are thoroughly discussed. The taxonomy of image acquisition, dataset preprocessing, feature extraction, and deep learning models are presented. The mathematical concepts related to feature extraction model architectures are discussed. Moreover, the different articles are compared based on their contributions, datasets, methods used, and the results achieved. The article ends with the main findings, current trends, challenges, and future recommendations.

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.

The Assessment of COVID-19 Pneumonia in Neonates: Observed by Lung Ultrasound Technique and Correlated with Biomarkers and Symptoms

Newborns infected with SARS-CoV2 infection develop different symptoms in comparison with adults, but one thing is clear: some of the most common manifestations include cough and other respiratory symptoms that need to be evaluated. In these cases, lung ultrasound is a useful imaging technique that can evaluate the newborns’ lung damage caused by COVID-19 pneumonia and can be used for the surveillance of the patients as well, being non-irradiating and easy to use. Nineteen neonates who were confirmed as having SARS-CoV2 infection were investigated using this imaging tool, and the results were compared and correlated with their symptoms and biomarkers. The mean of LUSS was 12.21 ± 3.56 (S.D), while the 95% CI for the arithmetic mean was 10.49–13.93. The difference of an independent t-test between the LUSS for the patient who presented cough and the LUSS for the patient without cough was −4.48 with an associated p-value of p = 0.02. The Pearson’s correlation coefficient r = 0.89 (p = 0.03, 95% CI 0.0642 to 0.993) between the LUSS and IL-6 level showed a positive strong correlation. This reliable correlation between lung ultrasound score and inflammatory markers suggests that LUS could be used for monitoring inflammatory lung diseases in the future.

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.

The Value of Lung Ultrasound Score in Neonatology

Point-of-care lung ultrasound (LUS) is increasingly applied in the neonatal intensive care unit (NICU). Diagnostic applications for LUS in the NICU contain the diagnosis of many common neonatal pulmonary diseases (such as Respiratory distress syndrome, Transient tachypnea of the newborn, Meconium aspiration syndrome, Pneumonia, Pneumothorax, and Pleural effusion) which have been validated. In addition to being employed as a diagnostic tool in the classical sense of the term, recent studies have shown that the number and type of artifacts are associated with lung aeration. Based on this theory, over the last few years, LUS has also been used as a semi-quantitative method or as a "functional" tool. Scores have been proposed to monitor the progress of neonatal lung diseases and to decide whether or not to perform a specific treatment. The semi-quantitative LUS scores (LUSs) have been developed to predict the demand for surfactant therapy, the need of respiratory support and the progress of bronchopulmonary dysplasia. Given their ease of use, accuracy and lack of invasiveness, the use of LUSs is increasing in clinical practice. Therefore, this manuscript will review the application of LUSs in neonatal lung diseases.

Lung Ultrasound Characteristics in Neonates With Positive Real Time Polymerase Chain Reaction for SARS-CoV-2 on a Tertiary Level Referral Hospital in Mexico City

INTRODUCTION

Acute respiratory syndrome secondary to SARS-CoV-2 virus infection has been declared a pandemic since December 2019. On neonates, severe presentations are infrequent but possible. Lung ultrasound (LUS) has been shown to be useful in diagnosing lung involvement and following up patients, giving more information, and reducing exposure compared to traditional examination.

METHODS

LUS was performed after the diagnosis of SARS-CoV-2 infection with respiratory Real Time Polymerase Chain Reaction RT-PCR with portable equipment protected with a silicone sleeve. If hemodynamic or cardiology consultation was necessary, a prepared complete ultrasound machine was used. Ten regions were explored (anterior superior and inferior, lateral, and posterior superior and inferior, right and left), and a semiquantitative score (LUSS) was calculated. Disease severity was determined with a pediatric modified score.

RESULTS

Thirty-eight patients with positive RT-PCR were admitted, 32 (81%) of which underwent LUS. Included patients had heterogenous diagnosis and gestational ages as expected on a referral neonatal intensive care unit (NICU) (median, ICR: 36, 30-38). LUS abnormalities found were B-line interstitial pattern 90%, irregular/interrupted/thick pleural line 88%, compact B-lines 65%, small consolidations (≤5 mm) 34%, and extensive consolidations (≥5 mm) 37%. Consolidations showed posterior predominance (70%). LUSS showed a median difference between levels of disease severity and ventilatory support (Kruskal-Wallis, p = 0.001) and decreased with patient improvement (Wilcoxon signed-rank test p = 0.005). There was a positive correlation between LUSS and FiO₂ needed (Spearman r = 0.72, p = 0.01). The most common recommendation to the attending team was pronation (41%) and increase in positive end expiratory pressure (34%). Five patients with comorbidities died. A significant rank difference of LUSS and FiO₂ needed between survivors and non-survivors was found (Mann-Whitney U-test, p = 0.005).

CONCLUSION

LUS patterns found were like the ones described in other series (neonatal and pediatrics). Eighty-eight percent of the studies were performed with handheld affordable equipment. While there is no specific pattern, it varies according to gestational age and baseline diagnosis LUS, which were shown to be useful in assessing lung involvement that correlated with the degree of disease severity and respiratory support.

Is Lung Ultrasound Helpful in COVID-19 Neonates?-A Systematic Review

Background: The SARS-CoV-2 infection has occurred in neonates, but it is a fact that radiation exposure is not recommended given their age. The aim of this review is to assess the evidence on the utility of lung ultrasound (LUS) in neonates diagnosed with COVID-19. Methods: A systematic literature review was performed so as to find a number of published studies assessing the benefits of lung ultrasound for newborns diagnosed with COVID and, in the end, to make a comparison between LUS and the other two more conventional procedures of chest X-rays or CT exam. The key terms used in the search of several databases were: “lung ultrasound”, “sonography”, “newborn”, “neonate”, and “COVID-19′. Results: In total, 447 studies were eligible for this review, and after removing the duplicates, 123 studies referring to LU were further examined, but only 7 included cases of neonates. These studies were considered for the present research paper. Conclusions: As a non-invasive, easy-to-use, and reliable method for lung lesion detection in neonates with COVID-19, lung ultrasound can be used as a useful diagnosis tool for the evaluation of COVID-19-associated lung lesions. The benefits of this method in this pandemic period are likely to arouse interest in opening new research horizons, with immediate practical applicability.

Usefulness of Lung Ultrasound in Paediatric Respiratory Diseases

Respiratory infection diseases are among the major causes of morbidity and mortality in children. Diagnosis is focused on clinical presentation, yet signs and symptoms are not specific and there is a need for new non-radiating diagnostic tools. Among these, lung ultrasound (LUS) has recently been included in point-of-care protocols showing interesting results. In comparison to other imaging techniques, such as chest X-ray and computed tomography, ultrasonography does not use ionizing radiations. Therefore, it is particularly suitable for clinical follow-up of paediatric patients. LUS requires only 5-10 min and allows physicians to make quick decisions about the patient's management. Nowadays, LUS has become an early diagnostic tool to detect pneumonia during the COVID-19 pandemic. In this narrative review, we show the most recent scientific literature about advantages and limits of LUS performance in children. Furthermore, we discuss the major paediatric indications separately, with a paragraph fully dedicated to COVID-19. Finally, we mention potential future perspectives about LUS application in paediatric respiratory diseases.

Quantitative assessment of COVID-19 pneumonia in neonates using lung ultrasound score

BACKGROUND

Lung ultrasound (LUS) and lung ultrasound score (LUSS) have been successfully used to diagnose neonatal pneumonia, assess the lesion distribution, and quantify the aeration loss. The present study design determines the diagnostic value of LUSS in the semi-quantitative assessment of pneumonia in coronavirus disease 2019 (COVID-19) neonates.

METHODS

Eleven COVID-19 neonates born to mothers with COVID-19 infection and 11 age- and gender-matched controls were retrospectively studied. LUSS was acquired by assessing the lesions and aeration loss in 12 lung regions per subject.

RESULTS

Most of the COVID-19 newborns presented with mild and atypical symptoms, mainly involving respiratory and digestive systems. In the COVID-19 group, a total of 132 regions of the lung were examined, 83 regions (62.8%) of which were detected abnormalities by LUS. Compared with controls, COVID-19 neonates showed sparse or confluent B-lines (83 regions), disappearing A-lines (83 regions), abnormal pleural lines (29 regions), and subpleural consolidations (2 regions). The LUSS was significantly higher in the COVID-19 group. In total, 49 regions (37%) were normal, 73 regions (55%) scored 1, and 10 regions (8%) scored 2 by LUSS. All the lesions were bilateral, with multiple regions involved. The majority of the lesions were located in the bilateral inferior and posterior regions. LUS detected abnormalities in three COVID-19 neonates with normal radiological performance. The intra-observer and inter-observer reproducibility of LUSS was excellent.

CONCLUSIONS

LUS is a noninvasive, convenient, and sensitive method to assess neonatal COVID-19 pneumonia, and can be used as an alternative to the use of diagnostic radiography. LUSS provides valuable semi-quantitative information on the lesion distribution and severity.