Accelerated MRI reconstructions via variational network and feature domain learning.

We introduce three architecture modifications to enhance the performance of the end-to-end (E2E) variational network (VarNet) for undersampled MRI reconstructions. We first implemented the Feature VarNet, which propagates information throughout the cascades of the network in an N-channel feature-space instead of a 2-channel feature-space. Then, we add an attention layer that utilizes the spatial locations of Cartesian undersampling artifacts to further improve performance. Lastly, we combined the Feature and E2E VarNets into the Feature-Image (FI) VarNet, to facilitate cross-domain learning and boost accuracy. Reconstructions were evaluated on the fastMRI dataset using standard metrics and clinical scoring by three neuroradiologists. Feature and FI VarNets outperformed the E2E VarNet for 4 × , 5 × and 8 × Cartesian undersampling in all studied metrics. FI VarNet secured second place in the public fastMRI leaderboard for 4 × Cartesian undersampling, outperforming all open-source models in the leaderboard. Radiologists rated FI VarNet brain reconstructions with higher quality and sharpness than the E2E VarNet reconstructions. FI VarNet excelled in preserving anatomical details, including blood vessels, whereas E2E VarNet discarded or blurred them in some cases. The proposed FI VarNet enhances the reconstruction quality of undersampled MRI and could enable clinically acceptable reconstructions at higher acceleration factors than currently possible.

EARLY TRAJECTORY OF VENOUS EXCESS ULTRASOUND SCORE IS ASSOCIATED WITH CLINICAL OUTCOMES OF GENERAL ICU PATIENTS.

ABSTRACT: Background: Systemic venous congestion, assessed by the venous excess ultrasound (VExUS) score, has been associated with adverse effects, including acute kidney injury (AKI), in patients with cardiac disease. In general intensive care unit (ICU) patients, the association between VExUS score and outcomes is understudied. We aimed to investigate the association between the trajectory of VExUS score within the first 3 days of ICU admission and the composite clinical outcome of major adverse kidney events within 30 days (MAKE30). Methods: In this prospective observational study, including patients consecutively admitted to the ICU, VExUS score was calculated within 24 h after ICU admission (day 1) and at 48 to 72 h (day 3). D-VExUS was calculated as the difference between the VExUS score on day 3 minus that on day 1. Development of AKI within 7 days and all-cause mortality within 30 days were recorded. Results: A total of 89 patients (62% men; median age, 62 years; median Acute Physiology and Chronic Health Evaluation II score, 24) were included. Sixty (67%) patients developed AKI within 7 days, and 17 (19%) patients died within 30 days after ICU admission. D-VExUS was associated with MAKE30, even after adjustment for confounders (hazard ratio, 2.07; 95% confidence interval, 1.17-3.66; P = 0.01). VExUS scores on days 1 or 3 were not associated with MAKE30. Also, VExUS scores on day 1 or on day 3 and D-VExUS were not associated with development of AKI or mortality. Conclusions: In a general ICU cohort, early trajectory of VExUS score, but not individual VExUS scores at different time points, was associated with the patient-centered MAKE30 outcome. Dynamic changes rather than snapshot measurements may unmask the adverse effects of systemic venous congestion on important clinical outcomes.

Immunomodulators for immunocompromised patients hospitalized for COVID-19: a meta-analysis of randomized controlled trials.

Background: Although immunomodulators have established benefit against the new coronavirus disease (COVID-19) in general, it is uncertain whether such agents improve outcomes without increasing the risk of secondary infections in the specific subgroup of previously immunocompromised patients. We assessed the effect of immunomodulators on outcomes of immunocompromised patients hospitalized for COVID-19. Methods: The protocol was prospectively registered with PROSPERO (CRD42022335397). MEDLINE, Cochrane Central Register of Controlled Trials and references of relevant articles were searched up to 01-06-2022. Authors of potentially eligible randomized controlled trials were contacted to provide data on immunocompromised patients randomized to immunomodulators vs control (i.e., placebo or standard-of-care). Findings: Eleven randomized controlled trials involving 397 immunocompromised patients hospitalized for COVID-19 were included. Ten trials had low risk of bias. There was no difference between immunocompromised patients randomized to immunomodulators vs control regarding mortality [30/182 (16.5%) vs 41/215 (19.1%); RR 0.93, 95% CI 0.61-1.41; p = 0.74], secondary infections (RR 1.00, 95% CI 0.64-1.58; p = 0.99) and change in World Health Organization ordinal scale from baseline to day 15 (weighed mean difference 0.27, 95% CI -0.09-0.63; p = 0.15). In subgroup analyses including only patients with hematologic malignancy, only trials with low risk of bias, only trials administering IL-6 inhibitors, or only trials administering immunosuppressants, there was no difference between comparators regarding mortality. Interpretation: Immunomodulators, compared to control, were not associated with harmful or beneficial outcomes, including mortality, secondary infections, and change in ordinal scale, when administered to immunocompromised patients hospitalized for COVID-19. Funding: Hellenic Foundation for Research and Innovation.

A combination of mild-moderate hypoxemia and low compliance is highly prevalent in persistent ARDS: a retrospective study.

BACKGROUND: The Acute Respiratory Distress Syndrome (ARDS) is characterized by lung inflammation and edema, impairing both oxygenation and lung compliance. Recent studies reported a dissociation between oxygenation and compliance (severe hypoxemia with preserved compliance) in early ARDS and COVID-19-related-ARDS (CARDS). During the pandemic, in patients requiring prolonged mechanical ventilation, we observed the opposite combination (mild-moderate hypoxemia but significantly impaired compliance). The purpose of our study was to investigate the prevalence of this combination of mild-moderate hypoxemia and impaired compliance in persistent ARDS and CARDS. METHODS: For this retrospective study, we used individual patient-level data from two independent cohorts of ARDS patients. The ARDSNet cohort included patients from four ARDS Network randomized controlled trials. The CARDS cohort included patients with ARDS due to COVID-19 hospitalized in two intensive care units in Greece. We used a threshold of 150 for PaO2/FiO2 and 30 ml/cmH2O for compliance, estimated the prevalence of each of the four combinations of oxygenation and compliance at baseline, and examined the change in its prevalence from baseline to day 21 in the ARDSNet and CARDS cohorts. RESULTS: The ARDSNet cohort included 2909 patients and the CARDS cohort included 349 patients. The prevalence of the combination of mild-moderate hypoxemia and low compliance increased from baseline to day 21 both in the ARDSNet cohort (from 22.2 to 42.7%) and in the CARDS cohort (from 3.1 to 33.3%). Among surviving patients with low compliance, oxygenation improved over time. The 60-day mortality rate was higher for patients who had mild-moderate hypoxemia and low compliance on day 21 (28% and 56% in ARDSNet and CARDS), compared to those who had mild-moderate hypoxemia and high compliance (20% and 50%, respectively). CONCLUSIONS: Among patients with ARDS who require prolonged controlled mechanical ventilation, regardless of ARDS etiology, a dissociation between oxygenation and compliance characterized by mild-moderate hypoxemia but low compliance becomes increasingly prevalent. The findings of this study highlight the importance of monitoring mechanics in patients with persistent ARDS.

PROLONGED MECHANICAL VENTILATION IN ACUTE RESPIRATORY DISTRESS SYNDROME.

ABSTRACT: Purpose: Trajectory of acute respiratory distress syndrome (ARDS) spans from rapidly improving cases to cases receiving prolonged mechanical ventilation (PMV). We attempted to estimate temporal trends of prevalence and mortality of PMV and to identify risk factors associated with mortality of patients with ARDS receiving PMV. Methods: We performed a secondary analysis of individual patient data from six randomized controlled clinical trials conducted by the ARDS Network. Prolonged mechanical ventilation was defined as the need for mechanical ventilation for >21 consecutive days. Results: Of 4,216 patients with ARDS, 646 (15.3%) received PMV. Prevalence of PMV gradually declined from 18.4% in the ARDS Network: Low-Tidal-Volume Trial (published in 2000) trial to 10.9% in the SAILS (2014) trial ( R2 = 0.728, P = 0.031). Ninety-day mortality of patients receiving PMV did not change over time ( R2 = 0.271, P = 0.290) and remained as high as 36.8%. Ιn the three most recent trials, risk factors associated with mortality among the 250 patients with ARDS receiving PMV included age, malignancy, pneumonia as the cause of ARDS, coagulation dysfunction, and hepatic dysfunction during the first 21 days after trial enrollment. Conclusion: Although prevalence of PMV among patients enrolled in ARDS Network trials gradually declined, mortality did not change. Risk factors associated with mortality were mostly nonmodifiable.

Racial and ethnic minority participants in clinical trials of acute respiratory distress syndrome.

PURPOSE: There is growing interest in improving the inclusiveness of racial and ethnic minority participants in trials of acute respiratory distress syndrome (ARDS). With our study we aimed to examine temporal trends of representation and mortality of racial and ethnic minority participants in randomized controlled trials of ARDS. METHODS: We performed a secondary analysis of eight ARDS Network and PETAL Network therapeutic clinical trials, published between 2000 and 2019. We classified race/ethnicity into "White", "Black", "Hispanic", or "Other" (including Asian, American Indian or Alaskan Native, Native Hawaiian, or other Pacific Islander participants). RESULTS: Of 5375 participants with ARDS, 1634 (30.4%) were Black, Hispanic, or Other race participants. Representation of racial and ethnic minority participants in trials did not change significantly over time (p = 0.257). However, among participants with moderate to severe ARDS (i.e., partial pressure of arterial oxygen to fraction of inspired oxygen ratio < 150), the difference in mortality between racial and ethnic minority participants and White participants decreased over time. In the five most recent trials, including 2923 participants with ARDS, there were no statistically significant differences in mortality between racial/ethnic groups, even after adjusting for potential confounders. In these five most recent trials, mortality was 31% for White, 31.9% for Black, 30.3% for Hispanic, and 37.1% for Other race participants (p = 0.633). CONCLUSION: Representation of racial and ethnic minority participants in ARDS trials from North America, published between 2000 and 2019, did not change over time. Black and Hispanic participants with ARDS may have similar mortality as White participants within trials.

Role of Clinical Insight at First Month in Predicting Relapse at the Year in First Episode of Psychosis (FEP) Patients.

INTRODUCTION: Clinical insight constitutes a useful marker of the progress and outcome of the First Episode of Psychosis (FEP), and lack of insight has been associated with more severe psychopathology, treatment non-adherence, and rehospitalization/relapse. In this study, we aimed to further investigate the possible role of insight as a predictor of relapse, its relation to diagnosis, and other parameters of positive psychotic symptomatology (delusions, hallucinations, and suspiciousness). METHODS: The Athens FEP study employed a prospective, longitudinal cohort design in which consecutive newly diagnosed patients with psychosis were interviewed and asked to voluntarily participate after completing informed consent. A total of 88/225 patients were examined at three different time points (baseline, month, and year). Their scores in the relevant items of the Positive and Negative Syndrome Scale (PANSS) were compared (G12 for insight, P1 for delusions, P3 for hallucinations, and P6 for suspiciousness), and they were further associated to diagnosis and the outcome at the end of the year (remission/relapse). RESULTS: In total, 22/88 patients with relapse at the year had greater scores in G12 for both the month and the year, and this finding was corroborated after adjusting the statistical analysis for demographics, diagnosis, social environment, and depression via multiple logistic regression analysis. Moreover, delusions and suspiciousness were significantly higher in patients diagnosed with non-affective psychosis compared to those diagnosed with affective psychosis (p < 0.001) at the first month. CONCLUSIONS: Lack of insight at the first month may serve as a predictor of relapse at the year.

Postoperative acute respiratory distress syndrome in randomized controlled trials.

BACKGROUND: Acute respiratory distress syndrome is a potentially fatal postoperative complication. We aimed to estimate temporal trends of the representation of patients with postoperative acute respiratory distress syndrome in clinical trials, determine their distinct clinical features, and identify predictors of mortality. METHODS: This is a secondary analysis of 7 randomized controlled clinical trials conducted by the Acute Respiratory Distress Syndrome Network and the Clinical Trials Network for the Prevention and Early Treatment of Acute Lung Injury. Patients with acute respiratory distress syndrome were classified into a postoperative acute respiratory distress syndrome group (ie, patients who had undergone elective surgery in the immediate period before trial enrollment) and a non-postoperative acute respiratory distress syndrome group. RESULTS: Out of 5,316 patients with acute respiratory distress syndrome, 256 (4.8%) had postoperative acute respiratory distress syndrome. Representation of postoperative acute respiratory distress syndrome in trials gradually declined from 2000 to 2011, but it remained stable afterward at 2.7%. Postoperative acute respiratory distress syndrome was associated with lower 90-day mortality (24.6% vs 30.9%, P = .032) than non-postoperative acute respiratory distress syndrome, even after adjusting for age, acute respiratory distress syndrome severity, usage of vasopressors at baseline, and the study publication year (hazard ratio 0.63, 95% confidence interval 0.49-0.82). Age (odds ratio 1.07, 95% confidence interval 1.04-1.09), immunosuppression (odds ratio 4.12, 95% confidence interval 1.43-11.86), and positive fluid balance (odds ratio 1.09, 95% confidence interval 1.04-1.14) were associated with 90-day mortality among patients with postoperative acute respiratory distress syndrome. CONCLUSION: Representation of postoperative acute respiratory distress syndrome in trials of the Acute Respiratory Distress Syndrome Network and the Clinical Trials Network for the Prevention and Early Treatment of Acute Lung Injury gradually declined from 2000 to 2011 but remained stable afterward. Postoperative acute respiratory distress syndrome was associated with lower mortality than non-postoperative acute respiratory distress syndrome. These findings may put both temporal trends and the prognosis of postoperative acute respiratory distress syndrome in perspective. Also, positive fluid balance was associated with the mortality of patients with postoperative acute respiratory distress syndrome.

Effect of Different Early Oxygenation Levels on Clinical Outcomes of Patients Presenting in the Emergency Department With Severe Traumatic Brain Injury.

STUDY OBJECTIVE: Despite the almost universal administration of supplemental oxygen in patients presenting in the emergency department (ED) with severe traumatic brain injury, optimal early oxygenation levels are unknown. Therefore, we aimed to examine the effect of different early oxygenation levels on the clinical outcomes of patients presenting in the emergency department with severe traumatic brain injury. METHODS: We performed a secondary analysis of the Resuscitation Outcomes Consortium Traumatic Brain Injury Hypertonic Saline randomized controlled trial by including patients with Glasgow Coma Scale ≤8. Early oxygenation levels were assessed by the worst value of arterial partial pressure of oxygen (PaO2) during the first 4 hours of presentation in the emergency department. The primary outcome was 6-month neurologic status, as assessed by the Extended Glasgow Outcome Scale. A binary logistic regression was utilized, and an odds ratio (OR) with 95% (95% confidence intervals) was calculated. RESULTS: A total of 910 patients were included. In unadjusted (crude) analysis, a PaO2 of 101 to 250 mmHg (OR, 0.59 [0.38 to 0.91]), or 251 to 400 mmHg (OR, 0.53 [0.34 to 0.83]) or ≥401 mmHg (OR, 0.31 [0.20 to 0.49]) was less likely to be associated with poor neurologic status when compared with a PaO2 of ≤100 mmHg. This was also the case for adjusted analyses (including age, pupillary reactivity, and Revised Trauma Score). CONCLUSION: High oxygenation levels as early as the first 4 hours of presentation in the emergency department may not be adversely associated with the long-term neurologic status of patients with severe traumatic brain injury. Therefore, during the early phase of trauma, clinicians may focus on stabilizing patients while giving low priority to the titration of oxygenation levels.

MARIE 2.0: A Perturbation Matrix Based Patient-Specific MRI Field Simulator.

High static field MR scanners can produce human tissue images of astounding clarity, but rely on high frequency electromagnetic radiation that generates complicated in-tissue field patterns that are patient-specific and potentially harmful. Many such scanners use parallel transmitters to better control field patterns, but then adjust the transmitters based on general guidelines rather than optimizing for the specific patient, mostly because computing patient-specific fields was presumed far too slow. It was recently demonstrated that the combination of fast low-resolution tissue mapping and fast voxel-based field simulation can be used to perform a patient-specific MR safety check in minutes. However, the field simulation required several of those minutes, making it too slow to perform the dozens of simulations that would be needed for patient-specific optimization. In this paper we describe a compressed-perturbation-matrix technique that nearly eliminates the computational cost of including complex coils (or coils and shields) in voxel-based field simulation of tissue, thereby reducing simulation time from minutes to seconds. The approach is demonstrated on a wide variety of head+coil and head+coil+shield configurations, using the implementation in MARIE 2.0, the latest version of the open-source MR field simulator MARIE.

A Hybrid Volume-Surface Integral Equation Method for Rapid Electromagnetic Simulations in MRI.

OBJECTIVE: We developed a hybrid volume surface integral equation (VSIE) method based on domain decomposition to perform fast and accurate magnetic resonance imaging (MRI) simulations that include both remote and local conductive elements. METHODS: We separated the conductive surfaces present in MRI setups into two domains and optimized electromagnetic (EM) modeling for each case. Specifically, interactions between the body and EM waves originating from local radiofrequency (RF) coils were modeled with the precorrected fast Fourier transform, whereas the interactions with remote conductive surfaces (RF shield, scanner bore) were modeled with a novel cross tensor train-based algorithm. We compared the hybrid-VSIE with other VSIE methods for realistic MRI simulation setups. RESULTS: The hybrid-VSIE was the only practical method for simulation using 1 mm voxel isotropic resolution (VIR). For 2 mm VIR, our method could be solved at least 23 times faster and required 760 times lower memory than traditional VSIE methods. CONCLUSION: The hybrid-VSIE demonstrated a marked improvement in terms of convergence times of the numerical EM simulation compared to traditional approaches in multiple realistic MRI scenarios. SIGNIFICANCE: The efficiency of the novel hybrid-VSIE method could enable rapid simulations of complex and comprehensive MRI setups.