Artificial Intelligence and Democratization of the Use of Lung Ultrasound in COVID-19: On the Feasibility of Automatic Calculation of Lung Ultrasound Score

During the COVID-19 pandemic, lung ultrasound has been revealed as a powerful technique for diagnosis and follow-up of pneumonia, the principal complication of SARS-CoV-2 infection. Nevertheless, being a relatively new and unknown technique, the lack of trained personnel has limited its application worldwide. Computer-aided diagnosis could possibly help to reduce the learning curve for less experienced physicians, and to extend such a new technique such as lung ultrasound more quickly. This work presents the preliminary results of the ULTRACOV (Ultrasound in Coronavirus disease) study, aimed to explore the feasibility of a real-time image processing algorithm for automatic calculation of the lung ultrasound score (LUS). A total of 28 patients positive on COVID-19 were recruited and scanned in 12 thorax zones following the lung score protocol, saving a 3 s video at each probe position. Those videos were evaluated by an experienced physician and by a custom developed automated detection algorithm, looking for A-Lines, B-Lines, consolidations, and pleural effusions. The agreement between the findings of the expert and the algorithm was 88.0% for B-Lines, 93.4% for consolidations and 99.7% for pleural effusion detection, and 72.8% for the individual video score. The standard deviation of the patient lung score difference between the expert and the algorithm was ±2.2 points over 36. The exam average time with the ULTRACOV prototype was 5.3 min, while with a conventional scanner was 12.6 min. Conclusion: A good agreement between the algorithm output and an experienced physician was observed, which is a first step on the feasibility of developing a real-time aided-diagnosis lung ultrasound equipment. Additionally, the examination time was reduced to less than half with regard to a conventional ultrasound exam. Acquiring a complete lung ultrasound exam within a few minutes is possible using fairly simple ultrasound machines that are enhanced with artificial intelligence, such as the one we propose. This step is critical to democratize the use of lung ultrasound in these difficult times.

The Value of Lung Ultrasound Score Predicts the Ventilator use in Preterm with Dyspnea within 24h after Hospitalization

This study aims to explore the predictive value of LUSsc(Lung Ultrasound Score) in the selection of respiratory support mode for premature infants with dyspnea.We prospectively included 857 preterm infants and performed LU in the first 2 hours of admission and scored LUSsc by two specialist sonographers. They were divided into two stratification according to gestational age (<32 +0 weeks and 32 +0 -36+6 weeks), and had two main outcomes: invasive and non-invasive respiratory support. In the training set, analysis the clinical factors finding the best cut-off value of lung ultrasound score then verified the consistency in the verification set. The choice of invasive respiratory support is based on neonatal mechanical ventilation rules. Preterm infants with invasive respiratory support had higher LUS scores and lower OI 、birth weight、than those with non-invasive support. For preterm <32 +0 weeks the cut-off point of LUSsc was 6.5 that the area under curve was 0.749 (95% CI: 0.689-0.809), which was statistically significant (P<0.05), and the sensitivity and specificity were 74.0% and 68.3%, for preterm 32 +0 -36 +6 weeks, cut-off point was 6.5 and the area under curve was 0.863 (95% CI: 0.811-0.911), sensitivity and specificity were 75.3% and 0.836%.In the validation set, use actual clinical respiratory support selection results to verify, for preterm <32 +0 weeks (Kappa value 0.660, P<0.05, McNemar test P ＞0.05),for preterm 32 +0 -36 +6 weeks (Kappa value 0.779, P<0.05, McNemar test P ＞0.05). Conclusion: The LUS score shows good reliability to predict respiratory support mode for preterm infants with dyspnea

Platelet-To-Lymphocyte Ratio Can Be Used As Marker To Predict Activity And Severity In Moderate And High Disease Activity of Rheumatoid Arthritis: A Retrospective Study

Objective: This study aimed to study and evaluate the value of platelet-to-lymphocyte ratio (PLR) and neutrophil-to-lymphocyte ratio (NLR) as markers to predict the disease activity and severity in patients with rheumatoid arthritis (RA).Methods: A total of 139 inpatients with RA were divided into two groups: moderate activity group (3.2<DAS28-CRP score<5.1) and high activity group (≥5.1). The correlation of routine hematological indices with DAS28 and joint ultrasound score (gray-scale ultrasound score, GS-US; power Doppler ultrasound score, PD-US; total ultrasound score, T-US) were analyzed by Pearson's correlation and logistic regression analyses. Receiver operating characteristics (ROC) analysis was performed to compare the efficacy of blood indices, ESR, or CRP in reflecting the disease activity and severity of RA.Results: The values of PLR, NLR, PD-US, and T-US were significantly different between moderate and high disease activity groups (p<0.001), and PLR was significantly correlated with PD-US and DAS28. Logistic regression analyses showed that PLR was an independent risk factor for disease activity by DAS 28 and joint damage severity by PD-US and T-US. ROC analysis showed that the efficacy of using PLR alone to evaluate the disease activity and joint severity of RA was similar to that of using combined CRP and ESR. The best cut-off value of PLR for predicting high disease activity and high joint severity was determined as 236.6.Conclusions: PLR can be used as a marker to predict activity and severity in patients with moderate and high RA disease activity.

Correlation of Serum Albumin Level to Lung Ultrasound Score and Its Role as Predictors of Outcome in Acute Respiratory Distress Syndrome Patients: A Prospective Observational Study

Background. There is ambiguity in the literature regarding hypoalbuminemia as a cause of extravascular lung water and acute respiratory distress syndrome (ARDS) outcomes. The aim of the study was to determine if low serum albumin on admission leads to lung deaeration and higher lung ultrasound score (LUSS) in ARDS patients. Patients and Methods. It was a prospective observational study in which 110 ARDS patients aged between 18 and 70 years were recruited. Serum albumin level and lung ultrasound score were assessed on the day of ICU admission. Length of ICU stay and hospital mortality were recorded. Results. The mean and standard deviation of serum albumin level in mild, moderate, and severe ARDS was 2.92 ± 0.65 g/dL, 2.91 ± 0.77 g/dL, and 3.21 ± 0.85 g/dL, respectively. Albumin level was not correlated to the global LUSS (Pearson correlation r −0.006, p = 0.949 ) and basal LUSS (r −0.066, p = 0.513 ). The cut-off value of albumin for predicting a prolonged length of ICU stay (≥10 days) in ARDS patients was <3.25 g/dL with AUC 0.623, p < 0.05 , sensitivity of 86.67%, specificity of 45.45%, and 95% confidence interval (CI) [0.513–0.732], and on multivariate analysis it increased the odds of prolonged ICU stay by 8.9 times (Hosmer and Lemeshow p value 0.810, 95% CI [2.760–28.72]). Serum albumin at admission was not a predictor of mortality. LUSS on the day of admission was not useful to predict either a prolonged length of ICU stay or mortality. Basal LUSS contributed about 56% of the global LUSS in mild and moderate ARDS, and 53% in severe ARDS. Conclusion. Serum albumin level was unrelated to LUSS on admission in ARDS patients. Albumin level <3.25 g/dL increased the chances of a prolonged length of ICU stay (≥10 days) but was not associated with an increase in mortality. LUSS on the day of admission could not predict either a prolonged length of ICU stay or mortality. This trial is registered with CTRI/2019/11/021857.

Roles of Lung Ultrasound Score in the Extubation Failure From Mechanical Ventilation Among Premature Infants With Neonatal Respiratory Distress Syndrome

Objective: To investigate the predictive value of lung ultrasound score (LUS) in the extubation failure from mechanical ventilation (MV) among premature infants with neonatal respiratory distress syndrome (RDS).Methods: The retrospective cohort study was conducted with a total of 314 RDS newborns who received MV support for over 24 h. After extubation from MV, infants were divided into extubation success and extubation failure groups. Extubation failure was defined as re-intubation within 48 h after extubation. Univariate and multivariate logistic regression analyses were used to identify the predictors of the extubation failure. The predictive effectiveness of the combined model and LUS in the extubation failure was assessed by receiver operating characteristic curve, area under curve (AUC), and internal validation.Results: 106 infants failed extubation from MV. The combined model for predicting the extubation failure was performed according to the predictors of gestational age, body length, birth weight, and LUS. The AUC of this combined model was 0.871 (sensitivity: 86.67%, specificity: 74.31%). The AUC of LUS was 0.858 (sensitivity: 84.00%, specificity: 80.69%), and the cutoff value was 18. There was no statistical difference in the predictive power between the combined model and LUS (Z = 0.880, P = 0.379). The internal validation result showed that the AUC of LUS was 0.855.Conclusions: LUS presented a good ability in predicting the extubation failure among RDS newborns after MV.

Elevated serum S14 levels are associated with more severe liver steatosis by ultrasonography

S14 has been identified as a potent stimulator of de novo hepatic lipogenesis (DNL) in rodents. However, it is unclear how S14 is regulated in humans with non-alcoholic fatty liver disease (NAFLD). The aim of this study was to investigate the relationship between serum S14 and liver steatosis in humans with NAFLD. A total of 614 participants were recruited from community. Liver steatosis were evaluated according to the Ultrasonographic Fatty Liver Indicator (US-FLI), which is a semi-quantitative liver ultrasound score. Anthropometric and biochemical indices were collected for further analysis. The risk of liver steatosis severity was estimated by a cumulative logistic regression model. NAFLD was found in 52.2% of the participants. The subjects with NAFLD showed higher levels of waist circumference, body mass index, insulin resistance, aspartate aminotransferase, dyslipidemia, visceral fat, serum S14 and risk of metabolic syndrome (MetS) than those of controls. Compared with the first tertile of serum S14, the odds ratios for the risk of more severe liver steatosis were 1.22 (95% confidence interval [CI]: 0.78–1.92) for those of the second tertile and 2.08 (95% CI: 1.28–3.39) for the third tertile (P for trend < 0.05) after adjusting for confounding factors. Higher serum S14 level was not only found in NAFLD subjects but also was positively correlated with the severity of liver steatosis. S14 may play an important role in the mechanism of DNL for NAFLD in humans.