Automatic deep learning-based pleural effusion segmentation in lung ultrasound images.

BACKGROUND: Point-of-care lung ultrasound (LUS) allows real-time patient scanning to help diagnose pleural effusion (PE) and plan further investigation and treatment. LUS typically requires training and experience from the clinician to accurately interpret the images. To address this limitation, we previously demonstrated a deep-learning model capable of detecting the presence of PE on LUS at an accuracy greater than 90%, when compared to an experienced LUS operator. METHODS: This follow-up study aimed to develop a deep-learning model to provide segmentations for PE in LUS. Three thousand and forty-one LUS images from twenty-four patients diagnosed with PE were selected for this study. Two LUS experts provided the ground truth for training by reviewing and segmenting the images. The algorithm was then trained using ten-fold cross-validation. Once training was completed, the algorithm segmented a separate subset of patients. RESULTS: Comparing the segmentations, we demonstrated an average Dice Similarity Coefficient (DSC) of 0.70 between the algorithm and experts. In contrast, an average DSC of 0.61 was observed between the experts. CONCLUSION: In summary, we showed that the trained algorithm achieved a comparable average DSC at PE segmentation. This represents a promising step toward developing a computational tool for accurately augmenting PE diagnosis and treatment.

Evaluation of the safety and efficacy of extracorporeal carbon dioxide removal in the critically ill using the PrismaLung+ device.

BACKGROUND: Several extracorporeal carbon dioxide removal (ECCO2R) devices are currently in use with variable efficacy and safety profiles. PrismaLung+ is an ECCO2R device that was recently introduced into clinical practice. It is a minimally invasive, low flow device that provides partial respiratory support with or without renal replacement therapy. Our aim was to describe the clinical characteristics, efficacy, and safety of PrismaLung+ in patients with acute hypercapnic respiratory failure. METHODS: All adult patients who required ECCO2R with PrismaLung+ for hypercapnic respiratory failure in our intensive care unit (ICU) during a 6-month period between March and September 2022 were included. RESULTS: Ten patients were included. The median age was 55.5 (IQR 41-68) years, with 8 (80%) male patients. Six patients had acute respiratory distress syndrome (ARDS), and two patients each had exacerbations of asthma and chronic obstructive pulmonary disease (COPD). All patients were receiving invasive mechanical ventilation at the time of initiation of ECCO2R. The median duration of ECCO2R was 71 h (IQR 57-219). A significant improvement in pH and PaCO2 was noted within 30 min of initiation of ECCO2R. Nine patients (90%) survived to weaning of ECCO2R, eight (80%) survived to ICU discharge and seven (70%) survived to hospital discharge. The median duration of ICU and hospital stays were 14.5 (IQR 8-30) and 17 (IQR 11-38) days, respectively. There were no patient-related complications with the use of ECCO2R. A total of 18 circuits were used in ten patients (median 2 per patient; IQR 1-2). Circuit thrombosis was noted in five circuits (28%) prior to reaching the expected circuit life with no adverse clinical consequences. CONCLUSION(S): PrismaLung+ rapidly improved PaCO2 and pH with a good clinical safety profile. Circuit thrombosis was the only complication. This data provides insight into the safety and efficacy of PrismaLung+ that could be useful for centres aspiring to introduce ECCO2R into their clinical practice.

Automatic deep learning-based consolidation/collapse classification in lung ultrasound images for COVID-19 induced pneumonia.

Our automated deep learning-based approach identifies consolidation/collapse in LUS images to aid in the identification of late stages of COVID-19 induced pneumonia, where consolidation/collapse is one of the possible associated pathologies. A common challenge in training such models is that annotating each frame of an ultrasound video requires high labelling effort. This effort in practice becomes prohibitive for large ultrasound datasets. To understand the impact of various degrees of labelling precision, we compare labelling strategies to train fully supervised models (frame-based method, higher labelling effort) and inaccurately supervised models (video-based methods, lower labelling effort), both of which yield binary predictions for LUS videos on a frame-by-frame level. We moreover introduce a novel sampled quaternary method which randomly samples only 10% of the LUS video frames and subsequently assigns (ordinal) categorical labels to all frames in the video based on the fraction of positively annotated samples. This method outperformed the inaccurately supervised video-based method and more surprisingly, the supervised frame-based approach with respect to metrics such as precision-recall area under curve (PR-AUC) and F1 score, despite being a form of inaccurate learning. We argue that our video-based method is more robust with respect to label noise and mitigates overfitting in a manner similar to label smoothing. The algorithm was trained using a ten-fold cross validation, which resulted in a PR-AUC score of 73% and an accuracy of 89%. While the efficacy of our classifier using the sampled quaternary method significantly lowers the labelling effort, it must be verified on a larger consolidation/collapse dataset, our proposed classifier using the sampled quaternary video-based method is clinically comparable with trained experts' performance.

Prolonged encephalopathy and associated nonconvulsive seizures from suspected pregabalin and tapentadol: Two case reports and review of literature.

A combination of pregabalin and tapentadol may be associated with prolonged encephalopathy.

Automatic deep learning-based pleural effusion classification in lung ultrasound images for respiratory pathology diagnosis.

Lung ultrasound (LUS) imaging as a point-of-care diagnostic tool for lung pathologies has been proven superior to X-ray and comparable to CT, enabling earlier and more accurate diagnosis in real-time at the patient's bedside. The main limitation to widespread use is its dependence on the operator training and experience. COVID-19 lung ultrasound findings predominantly reflect a pneumonitis pattern, with pleural effusion being infrequent. However, pleural effusion is easy to detect and to quantify, therefore it was selected as the subject of this study, which aims to develop an automated system for the interpretation of LUS of pleural effusion. A LUS dataset was collected at the Royal Melbourne Hospital which consisted of 623 videos containing 99,209 2D ultrasound images of 70 patients using a phased array transducer. A standardized protocol was followed that involved scanning six anatomical regions providing complete coverage of the lungs for diagnosis of respiratory pathology. This protocol combined with a deep learning algorithm using a Spatial Transformer Network provides a basis for automatic pathology classification on an image-based level. In this work, the deep learning model was trained using supervised and weakly supervised approaches which used frame- and video-based ground truth labels respectively. The reference was expert clinician image interpretation. Both approaches show comparable accuracy scores on the test set of 92.4% and 91.1%, respectively, not statistically significantly different. However, the video-based labelling approach requires significantly less effort from clinical experts for ground truth labelling.

Improving Sleep in Intensive Care Unit: An Overview of Diagnostic and Therapeutic Options.

Good quality sleep is considered to be essential for healthy living and recovering from illness. It would be logical to think that good quality sleep is most required when a patient is critically ill in an intensive care unit (ICU). Several studies have demonstrated poor quality of sleep while the patients are in ICU. Subjective tools such as questionnaires while simple are unreliable to accurately assess sleep quality. Relatively few studies have used standardized polysomnography. The use of novel biological markers of sleep such as serum brain-derived neurotrophic factor concentrations may help in conjunction with polysomnography to assess sleep quality in critically ill patients. Attempts to improve sleep included nonpharmacological interventions including the use of earplugs, eye sleep masks, and pharmacological agents including ketamine, propofol, dexmedetomidine, and benzodiazepines. The evidence for these interventions remains unclear. Further research is needed to assess quality of sleep and improve the sleep quality in intensive care settings.

Meropenem-induced vanishing bile duct syndrome: A case report.

Vanishing bile duct syndrome (VBDS) refers to a group of acquired disorders associated with progressive destruction and disappearance of the intrahepatic bile ducts. We report a case of meropenem-induced VBDS in a patient who had undergone surgical repair of a ruptured abdominal aortic aneurysm. Meropenem was used to treat Serratia marcescens isolated from blood, urine, sputum, and wound swab cultures. The patient developed severe mixed liver injury with no obstruction noted in radiological imaging. Because of the patient's increasing serum bilirubin level, VBDS was suspected and the meropenem was therefore changed to ciprofloxacin on postoperative day 18. Although the bilirubin level decreased, meropenem was restarted 3 days later because of clinical concerns regarding worsening fever and sepsis. Restarting meropenem was associated with an immediate increase in the serum bilirubin level. This further increase in bilirubin after reintroduction of meropenem strongly suggested meropenem-induced VBDS. The antibiotic therapy was changed from meropenem to ciprofloxacin and metronidazole, leading to a dramatic decrease in the bilirubin level to normal within a few weeks. In patients receiving meropenem, VBDS as a cause of deranged liver function and cholestasis should be considered after ruling out mechanical and other probable causes of liver injury.

Point-of-care lung ultrasound in the assessment of patients with COVID-19: A tutorial.

The adoption of point-of-care lung ultrasound for both suspected and confirmed COVID-19 patients highlights the issues of accessibility to ultrasound training and equipment. Lung ultrasound is more sensitive than chest radiography in detecting viral pneumonitis and preferred over computed tomography for reasons including its portability, reduced healthcare worker exposure and repeatability. The main lung ultrasound findings in COVID-19 patients are interstitial syndrome, irregular pleural line and subpleural consolidations. Consolidations are most likely found in critical patients in need of ventilatory support. Hence, lung ultrasound may be used to timely triage patients who may have evolving pneumonitis. Other respiratory pathology that may be detected by lung ultrasound includes pulmonary oedema, pneumothorax, consolidation and large effusion. A key barrier to incorporate lung ultrasound in the assessment of COVID-19 patients is adequate decontamination of ultrasound equipment to avoid viral spread. This tutorial provides a practical method to learn lung ultrasound and a cost-effective method of preventing contamination of ultrasound equipment and a practical method for performing and interpreting lung ultrasound.

Bleomycin-induced pulmonary toxicity and treatment with infliximab: A case report.

Given the current understanding of bleomycin-induced pneumonitis (BIP), the use of tumor necrosis factor alpha (TNF-α) inhibitors such as infliximab for late-stage disease appears to be of limited benefit. Further research regarding prevention and management of advanced BIP is required.

The impact of heart, lung and diaphragmatic ultrasound on prediction of failed extubation from mechanical ventilation in critically ill patients: a prospective observational pilot study.

BACKGROUND: Failed extubation from mechanical ventilation in critically ill patients is multifactorial, complex and not well understood. We aimed to identify whether combined transthoracic echocardiography, lung and diaphragmatic ultrasound can predict extubation failure in critically ill patients. RESULTS: Fifty-three participants who were intubated > 48 h and deemed by the treating intensivist ready for extubation underwent a 60-min pre-extubation weaning trial (pressure support ≤ 10 cmH2O and positive end expiratory pressure 5 cmH2O). Prior to extubation, data collected included ultrasound assessment of left ventricular ejection fraction, left atrial area, early diastolic trans-mitral flow velocity wave (E), early diastolic trans-mitral flow velocity wave/late diastolic trans-mitral flow velocity wave (E/A), early diastolic trans-mitral flow velocity wave/early diastolic mitral annulus velocity (E/E'), interatrial septal motion, lung loss of aeration score and diaphragm movement. At the end of the weaning trial, the rapid shallow breathing index and serum B-type natriuretic peptide concentration were measured. Success and failure of weaning was assessed by defined criteria. Decision to extubate was at the discretion of the treating intensivist. Failure of extubation was defined as re-intubation, non-invasive ventilation or death within 48 h after extubation. Of 53 extubated participants, 11 failed extubation. Failed extubation was associated with diabetes, ischaemic heart disease, higher E/E' (OR 1.27, 95% CI 1.05-1.54), left atrial area (OR 1.14, CI 1.02-1.28), fixed rightward curvature of the interatrial septum (OR 12.95, CI 2.73-61.41), and higher loss of aeration score of anterior and lateral regions of the lungs (OR 1.41, CI 1.01-1.82). CONCLUSIONS: Failed extubation in mechanically ventilated patients is more prevalent if markers of left ventricular diastolic dysfunction and loss of lung aeration are present.

Proportional assist ventilation versus pressure support ventilation in weaning ventilation: a pilot randomised controlled trial.

OBJECTIVE: Proportional assist ventilation with load-adjustable gain factors (PAV+) is a mode of ventilation that provides assistance in proportion to patient effort. This may have physiological and clinical advantages when compared with pressure support ventilation (PSV). Our objective was to compare these two modes in patients being weaned from mechanical ventilation. DESIGN: Prospective randomised controlled trial comparing PSV with PAV+. SETTING: University-affiliated, tertiary referral intensive care unit (ICU). PARTICIPANTS: Mechanically ventilated patients on a controlled mode of ventilation for at least 24 hours, who were anticipated to be spontaneously ventilated for at least 48 hours after randomisation. INTERVENTIONS: Nil. MAIN OUTCOME MEASURES: The primary outcome was time to successful liberation from the ventilator after the commencement of a spontaneous mode of ventilation. Secondary outcomes were requirement of rescue (mandatory) ventilation, requirement of sedative drugs, requirement for tracheostomy, re-intubation within 48 hours of extubation, ICU length of stay (LOS), hospital LOS, and ICU and hospital mortality. RESULTS: 50 patients were randomised to either PSV (n = 25) or PAV+ (n = 25). There was no significant difference between the PAV+ and PSV groups in time to successful weaning (84.3 v 135.9 hours, respectively; P = 0.536). Four patients randomised to PAV+ were crossed over to PSV during weaning. There was no significant difference between groups for rescue ventilation, reintubation within 48 hours, tracheostomy, sedatives and analgesics prescribed, and ICU and hospital LOS. ICU mortality was higher in the PSV group (25% v 4 %; P = 0.002). CONCLUSIONS: Both modes of ventilation were comparable in time to liberation from the ventilator.

The Feasibility and Impact of Routine Combined Limited Transthoracic Echocardiography and Lung Ultrasound on Diagnosis and Management of Patients Admitted to ICU: A Prospective Observational Study.

OBJECTIVES: Limited transthoracic echocardiography (TTE) and lung ultrasound increasingly is performed in the intensive care unit (ICU), though used in a goal-directed rather than routine manner. DESIGN: Prospective observational study. SETTINGS: Tertiary ICU. PARTICIPANTS: Ninety-three critically ill participants within 24 hours of admission to ICU. METHODS: A treating intensivist documented a clinical diagnosis and management plan before and after combined limited TTE and lung ultrasound. Ultrasound was performed by an independent intensivist and checked for accuracy offline by a second reviewer. RESULTS: Ultrasound images were interpretable in 99%, with good interobserver agreement. The hemodynamic diagnosis was altered in 66% of participants, including new (14%) or altered (25%) abnormal states or exclusion of clinically diagnosed abnormal state (27%). Valve pathology of at least moderate severity was diagnosed for mitral regurgitation (7%), aortic stenosis (1%), aortic stenosis and mitral regurgitation (1%), tricuspid regurgitation (3%), and 1 case of mitral regurgitation was excluded. Lung pathology diagnosis was changed in 58% of participants including consolidation (13%), interstitial syndrome (4%), and pleural effusion (23%), and exclusion of clinically diagnosed consolidation (6%), interstitial syndrome (3%), and pleural effusion (9%). Management changed in 65% of participants including increased (12%) or decreased (23%) fluid therapy, initiation (10%), changing (6%) or cessation (9%) of inotropic, vasoactive or diuretic drugs, non-invasive ventilation (3%), and pleural drainage (2%). CONCLUSION: Routine screening of patients with combined limited TTE and lung ultrasound on admission to ICU is feasible and frequently alters diagnosis and management.

Severe Rhabdomyolysis due to Presumed Drug Interactions between Atorvastatin with Amlodipine and Ticagrelor.

Atorvastatin and ticagrelor combination is a widely accepted therapy for secondary prevention of ischaemic heart disease. However, rhabdomyolysis is a well-known rare side effect of statins which should be considered when treatments are combined with cytochrome P450 3A4 enzyme inhibitors. We report a case of atorvastatin and ticagrelor associated severe rhabdomyolysis that progressed to multiorgan failure requiring renal replacement therapy, inotropes, intubation, and mechanical ventilation. Despite withdrawal of the precipitating cause and the supportive measures including renal replacement therapy, creatinine kinase increased due to ongoing rhabdomyolysis rapidly progressing to upper and lower limbs weakness. A muscle biopsy was performed to exclude myositis which confirmed extensive myonecrosis, consistent with statin associated rhabdomyolysis. After a prolonged ventilatory course in the intensive care unit, patient's condition improved with recovery from renal and liver dysfunction. The patient slowly regained her upper and lower limb function; she was successfully weaned off the ventilator and was discharged for rehabilitation. To our knowledge, this is a second case of statin associated rhabdomyolysis due to interaction between atorvastatin and ticagrelor. However, our case differed in that the patient was also on amlodipine, which is considered to be a weak cytochrome P450 3A4 inhibitor and may have further potentiated myotoxicity.

Low-flow veno-venous extracorporeal carbon dioxide removal in the management of severe status asthmatics: a case report.

Status asthmaticus is a life-threatening condition that requires intensive care management. Most of these patients have severe hypercapnic acidosis that requires lung protective mechanical ventilation. A small proportion of these patients do not respond to conventional lung protective mechanical ventilation or pharmacotherapy. Such patients have an increased mortality and morbidity. Successful use of extracorporeal membrane oxygenation (ECMO) is reported in such patients. However, the use of ECMO is invasive with its associated morbidity and is limited to specialised centres. In this report, we report the use of a novel, minimally invasive, low-flow extracorporeal carbon dioxide removal device in management of severe hypercapnic acidosis in a patient with life threatening status asthmaticus.

A case of chronic inflammatory demyelinating polyneuropathy with reversible alternating diaphragmatic paralysis: case study.

Respiratory failure requiring mechanical ventilation has been reported in patients with bilateral diaphragmatic paralysis due to CIDP. We report a case of CIDP that progressed to respiratory failure with normal chest radiography despite unilateral diaphragmatic paralysis. This manifestation would have been missed if ultrasound was not employed.

Diaphragmatic regional displacement assessed by ultrasound and correlated to subphrenic organ movement in the critically ill patients--an observational study.

INTRODUCTION: The objectives of the study are to identify the most reliably imaged regions of the diaphragm, to evaluate the correlation of movement between different parts of each hemidiaphragm, and to assess the agreement between liver or spleen displacement and movement of the ipsilateral hemidiaphragm. METHODS: Images of the diaphragm, liver, and spleen were obtained using 2-dimensional ultrasound. Acceptable agreement between regions of the diaphragm, liver, and spleen was defined as an absence of fixed or proportional bias using Deming regression analysis and limits of agreement of 2 SDs of the difference less than 30% of the mean value. RESULTS: We included 90 critically ill patients. The medial (87%) and middle (73%) regions of the right hemidiaphragm, liver (87.7%), and spleen (81%) and medial (71%) and middle regions (51%) of the left hemidiaphragm were most frequently imaged. In nonintubated patients, acceptable agreement was present for comparisons of the left middle and medial, right middle and medial, and left middle regions and spleen displacement. In intubated patients and in all patients when combined, acceptable agreement was only present for comparisons of the left middle and medial and right middle and medial regions of the diaphragm. Acceptable agreement was not present for intubated and all patients for diaphragmatic and solid organ movement. CONCLUSION: The diaphragm medial part is visualized in the majority of studied patients. The medial and middle thirds may be used interchangeably to assess hemidiaphragm movement. Acceptable agreement does not exist for diaphragm and solid organ movement, other than for the left middle region and the spleen.

Stress hyperglycemia may not be harmful in critically ill patients with sepsis.

BACKGROUND: Stress hyperglycemia (SH) is commonly seen in critically ill patients. It has been shown to be associated with adverse outcomes in some groups of patients. The effects of SH on critically ill patients with sepsis have not been well studied. We aimed to evaluate the effects of SH in critically ill patients with sepsis. METHODS: In this retrospective study, patients with sepsis admitted to intensive care unit (ICU) over a 5-year period were included. RESULTS: Of 297 patients, 204 (68.7%) had SH during the study period. The mean blood glucose level in patients with SH was 8.7 mmol/L compared with 5.9 mmol/L in those without SH (P < .05). There were no statistically significant differences in age; sex; sepsis severity; cardiovascular, respiratory, and renal comorbidities; requirement of mechanical ventilation; inotropes; and Acute Physiology, Age, and Chronic Health Evaluation III and Simplified Acute Physiology 2 scores on ICU admission. Intensive care unit mortality was significantly lower in patients who had SH. The median duration of ICU and hospital length of stay was longer in patients with SH. On logistic regression analysis, the presence of SH was associated with reduced ICU mortality. Subgroup analysis revealed SH to be protective in patients with septic shock. CONCLUSION: Stress hyperglycemia may not be harmful in critically ill patients with sepsis. Patients with SH had lower ICU mortality.