MyoVision-US: an Artificial Intelligence-Powered Software for Automated Analysis of Skeletal Muscle Ultrasonography.

Introduction/Aims: Muscle ultrasound has high utility in clinical practice and research; however, the main challenges are the training and time required for manual analysis to achieve objective quantification of morphometry. This study aimed to develop and validate a software tool powered by artificial intelligence (AI) by measuring its consistency and predictability of expert manual analysis quantifying lower limb muscle ultrasound images across healthy, acute, and chronic illness subjects. Methods: Quadriceps complex (QC [rectus femoris and vastus intermedius]) and tibialis anterior (TA) muscle ultrasound images of healthy, intensive care unit, and/or lung cancer subjects were captured with portable devices. Automated analyses of muscle morphometry were performed using a custom-built deep-learning model (MyoVision-US), while manual analyses were performed by experts. Consistency between manual and automated analyses was determined using intraclass correlation coefficients (ICC), while predictability of MyoVision -US was calculated using adjusted linear regression (adj.R 2 ). Results: Manual analysis took approximately 24 hours to analyze all 180 images, while MyoVision - US took 247 seconds, saving roughly 99.8%. Consistency between the manual and automated analyses by ICC was good to excellent for all QC (ICC:0.85-0.99) and TA (ICC:0.93-0.99) measurements, even for critically ill (ICC:0.91-0.98) and lung cancer (ICC:0.85-0.99) images. The predictability of MyoVision-US was moderate to strong for QC (adj.R 2 :0.56-0.94) and TA parameters (adj.R 2 :0.81-0.97). Discussion: The application of AI automating lower limb muscle ultrasound analyses showed excellent consistency and strong predictability compared with human analysis. Future work needs to explore AI-powered models for the evaluation of other skeletal muscle groups.

Construct and criterion validity of muscle ultrasonography for assessment of skeletal muscle in patients recovering from COVID-19.

Background: The purpose was to investigate the content, construct, and criterion validity of muscle ultrasound in a mixed cohort of participants recovering from mild and critical COVID-19. Methods: A secondary analysis of a prospective cross-sectional study was conducted on data obtained from a battery of muscle and physical function assessments including a muscle biopsy and muscle ultrasonography (US). Rectus femoris (RF) muscle thickness (mT), quadricep complex (QC) mT, RF muscle cross-sectional area (CSA) using 2D freeform trace and estimated from Feret's diameter, and RF echo intensity (EI) were assessed with US. Muscle fiber CSA, fiber type, protein content in muscle fibers, extracellular matrix content (ECM; wheat-germ agglutin), and percent area of collagen in ECM (picrosirius red) were examined from vastus lateralis muscle biopsies. Spearman rho correlations (r) were performed to assess validity of ultrasound parameters. Results: Thirty-three individuals participated including 11 patients surviving critical COVID-19, 15 individuals recovering from mild-COVID, and 7 controls. There were several significant correlations between RF mT, QC mT, RF CSA, and RF EI with age, comorbid burden, body-mass index, and measures of muscle strength, muscle power, and physical function (range r = 0.35-0.83). RF Feret's CSA correlated to CSA of type II muscle fibers (r = 0.41, p = 0.022) and the average size of all muscle fibers (r = 0.39, p = 0.031). RF EI was correlated with collagen in muscle ECM (r = 0.53, p = 0.003) and protein content in muscle tissue (r = -0.52, p = 0.012). Conclusion: Muscle size and quality measured using US has moderate content and construct validity, and to lesser extent, fair to moderate criterion validity in a mixed cohort of individuals recovering from COVID. Muscle ultrasound quality (EI) appears to be sensitive at detecting muscle dysfunction as it is associated with strength, power, physical function, and collagen distribution in a mixed group of individuals recovering from COVID-19.

Transcranial ultrasonography to detect intracranial pathology: A systematic review and meta-analysis.

BACKGROUND AND PURPOSE: Transcranial ultrasonography (TCU) can be a useful diagnostic tool in evaluating intracranial pathology in patients with limited or delayed access to routine neuroimaging in critical care or austere settings. We reviewed available literature investigating the diagnostic utility of TCU for detecting pediatric and adult patient's intracranial pathology in patients with intact skulls and reported diagnostic accuracy measures. METHODS: We performed a systematic review of PubMed® , Cochrane Library, Embase® , Scopus® , Web of Science™, and Cumulative Index to Nursing and Allied Health Literature databases to identify articles evaluating ultrasound-based detection of intracranial pathology in comparison to routine imaging using broad Medical Subject Heading sets. Two independent reviewers reviewed the retrieved articles for bias using the Quality Assessment of Diagnostic Accuracy Studies tools and extracted measures of diagnostic accuracy and ultrasound parameters. Data were pooled using meta-analysis implementing a random-effects approach to examine the sensitivity, specificity, and accuracy of ultrasound-based diagnosis. RESULTS: A total of 44 studies out of the 3432 articles screened met the eligibility criteria, totaling 2426 patients (Mean age: 60.1 ± 14.52 years). We found tumors, intracranial hemorrhage (ICH), and neurodegenerative diseases in the eligible studies. Sensitivity, specificity, and accuracy of TCU and their 95% confidence intervals were 0.80 (0.72, 0.89), 0.71 (0.59, 0.82), and 0.76 (0.71, 0.82) for neurodegenerative diseases; 0.88 (0.74, 1.02), 0.81 (0.50, 1.12), and 0.94 (0.92, 0.96) for ICH; and 0.97 (0.92, 1.03), 0.99 (0.96, 1.01), and 0.99 (0.97, 1.01) for intracranial masses. No studies reported ultrasound presets. CONCLUSIONS: TCU has a reasonable sensitivity and specificity for detecting intracranial pathology involving ICH and tumors with clinical applications in remote locations or where standard imaging is unavailable. Future studies should investigate ultrasound parameters to enhance diagnostic accuracy in diagnosing intracranial pathology.

Early posthospitalization recovery after extracorporeal membrane oxygenation in survivors of COVID-19.

OBJECTIVE: We sought to determine the influence of venovenous extracorporeal membrane oxygenation (ECMO) on outcomes of mechanically ventilated patients with COVID-19 during the first 120 days after hospital discharge. METHODS: Five academic centers conducted a retrospective analysis of mechanically ventilated patients with COVID-19 admitted during March through May 2020. Survivors had access to a multidisciplinary postintensive care recovery clinic. Physical, psychological, and cognitive deficits were measured using validated instruments and compared based on ECMO status. RESULTS: Two hundred sixty two mechanically ventilated patients were compared with 46 patients cannulated for venovenous ECMO. Patients receiving ECMO were younger and traveled farther but there was no significant difference in gender, race, or body mass index. ECMO patients were mechanically ventilated for longer durations (median, 26 days [interquartile range, 19.5-41 days] vs 13 days [interquartile range, 7-20 days]) and were more likely to receive inhaled pulmonary vasodilators, neuromuscular blockade, investigational COVID-19 therapies, blood transfusions, and inotropes. Patients receiving ECMO experienced greater bleeding and clotting events (P < .01). However, survival at discharge was similar (69.6% vs 70.6%). Of the 217 survivors, 65.0% had documented follow-up within 120 days. Overall, 95.5% were residing at home, 25.7% had returned to work or usual activity, and 23.1% were still using supplemental oxygen; these rates did not differ significantly based on ECMO status. Rates of physical, psychological, and cognitive deficits were similar. CONCLUSIONS: Our data suggest that COVID-19 survivors experience significant physical, psychological, and cognitive deficits following intensive care unit admission. Despite a more complex critical illness course, longer average duration of mechanical ventilation, and longer average length of stay, patients treated with venovenous ECMO had similar survival at discharge and outcomes within 120 days of discharge.

Worse survival in patients with right ventricular dysfunction and COVID-19-associated acute respiratory distress requiring extracorporeal membrane oxygenation: A multicenter study from the ORACLE Group.

OBJECTIVE: We sought to determine the impact of right ventricular dysfunction on the outcomes of mechanically ventilated patients with COVID-19 requiring veno-venous extracorporeal membrane oxygenation. METHODS: Six academic centers conducted a retrospective analysis of mechanically ventilated patients with COVID-19 stratified by support with veno-venous extracorporeal membrane oxygenation during the first wave of the pandemic (March to August 2020). Echocardiograms performed for clinical indications were reviewed for right and left ventricular function. Baseline characteristics, hospitalization characteristics, and survival were compared. RESULTS: The cohort included 424 mechanically ventilated patients with COVID-19, 126 of whom were cannulated for veno-venous extracorporeal membrane oxygenation. Right ventricular dysfunction was observed in 38.1% of patients who received extracorporeal membrane oxygenation and 27.4% of patients who did not receive extracorporeal membrane oxygenation with an echocardiogram. Biventricular dysfunction was observed in 5.5% of patients who received extracorporeal membrane oxygenation. Baseline patient characteristics were similar in both the extracorporeal membrane oxygenation and non-extracorporeal membrane oxygenation cohorts stratified by the presence of right ventricular dysfunction. In the extracorporeal membrane oxygenation cohort, right ventricular dysfunction was associated with increased inotrope use (66.7% vs 24.4%, P < .001), bleeding complications (77.1% vs 53.8%, P = .015), and worse survival independent of left ventricular dysfunction (39.6% vs 64.1%, P = .012). There was no significant difference in days ventilated before extracorporeal membrane oxygenation, length of hospital stay, hours on extracorporeal membrane oxygenation, duration of mechanical ventilation, vasopressor use, inhaled pulmonary vasodilator use, infectious complications, clotting complications, or stroke. The cohort without extracorporeal membrane oxygenation cohort demonstrated no statistically significant differences in in-hospital outcomes. CONCLUSIONS: The presence of right ventricular dysfunction in patients with COVID-19-related acute respiratory distress syndrome supported with veno-venous extracorporeal membrane oxygenation was associated with increased in-hospital mortality. Additional studies are required to determine if mitigating right ventricular dysfunction in patients requiring veno-venous extracorporeal membrane oxygenation improves mortality.

Physical Function Measured Prior to Lung Transplantation Is Associated With Posttransplant Patient Outcomes.

INTRODUCTION: The primary objective of this study was to determine whether pretransplant physical function is correlated with posttransplantation outcomes. METHODS: We performed a retrospective study of patients that participated in pretransplantation screening and subsequently underwent lung transplantation. Pretransplant variables of interest included demographics, muscle mass, body composition, physical function, and physical frailty. Correlation tests were performed to assess relationships with significance set at 0.05. RESULTS: Twenty-five patients with a mean age of 57 ± 13 years (68% male) with pretransplant lung allocation score of 45 ± 14 were included. This cohort had a 3-year mortality rate of 32% (n = 8). Pretransplant 4-m gait speed was significantly related to performance on the Short Physical Performance Battery (r = 0.74, P = .02) and distance ambulated on the 6-minute walk test (r = 0.62, P = .07) at hospital discharge. Older age was associated with slower gait speed and worse performance on sit-to-stand testing at hospital discharge (r = -0.76, P = .01 and r = -0.75, P = .01, respectively). Statistically, only diagnosis of cystic fibrosis was associated with 3-year mortality. DISCUSSION: Our study demonstrates that demographic, clinical, and physical function assessed prior to lung transplantation may be indicators of functional recovery.

Development, implementation and outcomes of a quality assurance system for the provision of continuous renal replacement therapy in the intensive care unit.

Critically ill patients with requirement of continuous renal replacement therapy (CRRT) represent a growing intensive care unit (ICU) population. Optimal CRRT delivery demands continuous communication between stakeholders, iterative adjustment of therapy, and quality assurance systems. This Quality Improvement (QI) study reports the development, implementation and outcomes of a quality assurance system to support the provision of CRRT in the ICU. This study was carried out at the University of Kentucky Medical Center between September 2016 and June 2019. We implemented a quality assurance system using a step-wise approach based on the (a) assembly of a multidisciplinary team, (b) standardization of the CRRT protocol, (c) creation of electronic CRRT flowsheets, (d) selection, monitoring and reporting of quality metrics of CRRT deliverables, and (e) enhancement of education. We examined 34-month data comprising 1185 adult patients on CRRT (~ 7420 patient-days of CRRT) and tracked selected QI outcomes/metrics of CRRT delivery. As a result of the QI interventions, we increased the number of multidisciplinary experts in the CRRT team and ensured a continuum of education to health care professionals. We maximized to 100% the use of continuous veno-venous hemodiafiltration and doubled the percentage of patients using regional citrate anticoagulation. The delivered CRRT effluent dose (~ 30 ml/kg/h) and the delivered/prescribed effluent dose ratio (~ 0.89) remained stable within the study period. The average filter life increased from 26 to 31 h (p = 0.020), reducing the mean utilization of filters per patient from 3.56 to 2.67 (p = 0.054) despite similar CRRT duration and mortality rates. The number of CRRT access alarms per treatment day was reduced by 43%. The improvement in filter utilization translated into ~ 20,000 USD gross savings in filter cost per 100-patient receiving CRRT. We satisfactorily developed and implemented a quality assurance system for the provision of CRRT in the ICU that enabled sustainable tracking of CRRT deliverables and reduced filter resource utilization at our institution.