Pneumothorax in patients with respiratory failure in ICU

Deep Learning for Detection of Pneumothorax and Pleural Effusion on Chest Radiographs: Validation Against Computed Tomography, Impact on Resident Reading Time, and Interreader Concordance

PURPOSE

To study the performance of artificial intelligence (AI) for detecting pleural pathology on chest radiographs (CXRs) using computed tomography as ground truth.

PATIENTS AND METHODS

Retrospective study of subjects undergoing CXR in various clinical settings. Computed tomography obtained within 24 hours of the CXR was used to volumetrically quantify pleural effusions (PEfs) and pneumothoraxes (Ptxs). CXR was evaluated by AI software (INSIGHT CXR; Lunit) and by 3 second-year radiology residents, followed by AI-assisted reassessment after a 3-month washout period. We used the area under the receiver operating characteristics curve (AUROC) to assess AI versus residents' performance and mixed-model analyses to investigate differences in reading time and interreader concordance.

RESULTS

There were 96 control subjects, 165 with PEf, and 101 with Ptx. AI-AUROC was noninferior to aggregate resident-AUROC for PEf (0.82 vs 0.86, P < 0.001) and Ptx (0.80 vs 0.84, P = 0.001) detection. AI-assisted resident-AUROC was higher but not significantly different from the baseline. AI-assisted reading time was reduced by 49% (157 vs 80 s per case, P = 0.009), and Fleiss kappa for Ptx detection increased from 0.70 to 0.78 ( P = 0.003). AI decreased detection error for PEf (odds ratio = 0.74, P = 0.024) and Ptx (odds ratio = 0.39, P < 0.001).

CONCLUSION

Current AI technology for the detection of PEf and Ptx on CXR was noninferior to second-year resident performance and could help decrease reading time and detection error.

Occult tension pneumothorax discovered following imaging for adult trauma patients in the modern major trauma system: a multicentre observational study

BACKGROUND

Tension pneumothorax following trauma is a life-threatening emergency and radiological investigation is normally discouraged prior to treatment in traditional trauma doctrines such as ATLS. Some trauma patients may be physiologically stable enough for diagnostic imaging and occult tension pneumothorax is discovered radiologically. We assessed the outcomes of these patients and compared them with those with clinical diagnosis of tension pneumothorax prior to imaging.

METHODS

A multicentre civilian-military collaborative network of six major trauma centres in the UK collected observational data from adult patients who had a diagnosis of traumatic tension pneumothorax during a 33-month period. Patients were divided into 'radiological' (diagnosis following CT/CXR) or 'clinical' (no prior CT/CXR) groups. The effect of radiological diagnosis on survival was analysed using multivariable logistic regression that included the covariates of age, gender, comorbidities and Injury Severity Score.

RESULTS

There were 133 patients, with a median age of 41 (IQR 24-61); 108 (81%) were male. Survivors included 49 of 59 (83%) in the radiological group and 59 of 74 (80%) in the clinical group (p=0.487). Multivariable logistic regression showed no significant association between radiological diagnosis and survival (OR 2.40, 95% CI 0.80 to 7.95; p=0.130). There was no significant difference in mortality between the groups.

CONCLUSION

Radiological imaging may be appropriate for selected trauma patients at risk of tension pneumothorax if they are considered haemodynamically stable. Trauma patients may be physiologically stable enough for radiological imaging but have occult tension pneumothorax because they did not have the typical clinical presentation. The historical dogma of the 'forbidden scan' no longer applies to such patients.

Risk factors for long-term invasive mechanical ventilation: a longitudinal study using German health claims data

BACKGROUND

Long-term invasive mechanical ventilation (IMV) is a major burden for those affected and causes high costs for the health care system. Early risk assessment is a prerequisite for the best possible support of high-risk patients during the weaning process. We aimed to identify risk factors for long-term IMV within 96 h (h) after the onset of IMV.

METHODS

The analysis was based on data from one of Germany's largest statutory health insurance funds; patients who received IMV ≥ 96 h and were admitted in January 2015 at the earliest and discharged in December 2017 at the latest were analysed. OPS and ICD codes of IMV patients were considered, including the 365 days before intubation and 30 days after discharge. Long-term IMV was defined as evidence of invasive home mechanical ventilation (HMV), IMV ≥ 500 h, or readmission with (re)prolonged ventilation.

RESULTS

In the analysis of 7758 hospitalisations, criteria for long-term IMV were met in 38.3% of cases, of which 13.9% had evidence of HMV, 73.1% received IMV ≥ 500 h and/or 40.3% were re-hospitalised with IMV. Several independent risk factors were identified (p < 0.005 each), including pre-diagnoses such as pneumothorax (OR 2.10), acute pancreatitis (OR 2.64), eating disorders (OR 1.99) or rheumatic mitral valve disease (OR 1.89). Among ICU admissions, previous dependence on an aspirator or respirator (OR 5.13), and previous tracheostomy (OR 2.17) were particularly important, while neurosurgery (OR 2.61), early tracheostomy (OR 3.97) and treatment for severe respiratory failure such as positioning treatment (OR 2.31) and extracorporeal lung support (OR 1.80) were relevant procedures in the first 96 h after intubation.

CONCLUSION

This comprehensive analysis of health claims has identified several risk factors for the risk of long-term ventilation. In addition to the known clinical risks, the information obtained may help to identify patients at risk at an early stage. Trial registration The PRiVENT study was retrospectively registered at ClinicalTrials.gov (NCT05260853). Registered at March 2, 2022.

Finger Thoracostomy for Tension Pneumothorax

One of the injuries associated with chest trauma is pneumothorax, a condition where air accumulates between the parietal and visceral pleura in the chest leading to collapse of the lung due to pressure. Left untreated, a tension pneumothorax may develop leading to cardiovascular collapse. This article reviews the development of a tension pneumothorax, discusses the clinical recognition of the diagnosis, and outlines the procedure for performing a finger (or simple) thoracostomy. A simple mnemonic for the procedure is offered as a memory aid to reduce cognitive load for this procedure.

Position-Dependent Symptoms of Pneumothorax During Mechanical Ventilation: A Case Report

A 54-year-old male with severe hypoxia was transferred to our hospital after choking on a mochi. Chest computed tomography revealed negative pressure pulmonary edema without pneumothorax. Endotracheal intubation was performed, and pressure-controlled ventilation was initiated. Following admission to the intensive care unit, his respiratory condition was stable in both the supine and left decubitus positions. However, every time he was placed in the right decubitus position, the tidal volume decreased by half, and SpO₂ dropped rapidly to 80%, which recovered soon after returning to the supine position. Chest radiography was performed the following day, revealing grade II right pneumothorax, and a chest tube placement stabilized his respiratory status in the right decubitus position. Air leakage ceased within a few hours. Extubation was successful on the fifth hospital day, and the chest tube was removed on the eighth hospital day. To our knowledge, there are no previous reports on position-dependent symptoms of pneumothorax during mechanical ventilation. Clinicians should consider the possibility of pneumothorax on that same side when respiratory deterioration is observed only in one lateral decubitus position during mechanical ventilation.

Point-of-Care Lung Ultrasound in the Intensive Care Unit-The Dark Side of Radiology: Where Do We Stand?

Patients in intensive care units (ICUs) are critically ill and require constant monitoring of clinical conditions. Due to the severity of the underlying disease and the need to monitor devices, imaging plays a crucial role in critically ill patients' care. Given the clinical complexity of these patients, who typically need respiratory assistance as well as continuous monitoring of vital functions and equipment, computed tomography (CT) can be regarded as the diagnostic gold standard, although it is not a bedside diagnostic technique. Despite its limitations, portable chest X-ray (CXR) is still today an essential diagnostic tool used in the ICU. Being a widely accessible imaging technique, which can be performed at the patient's bedside and at a low healthcare cost, it provides additional diagnostic support to the patient's clinical management. In recent years, the use of point-of-care lung ultrasound (LUS) in ICUs for procedure guidance, diagnosis, and screening has proliferated, and it is usually performed at the patient's bedside. This review illustrates the role of point-of-care LUS in ICUs from a purely radiological point of view as an advanced method in ICU CXR reports to improve the interpretation and monitoring of lung CXR findings.

Automated Rib Fracture Detection on Chest X-Ray Using Contrastive Learning

To develop a deep learning-based model for detecting rib fractures on chest X-Ray and to evaluate its performance based on a multicenter study. Chest digital radiography (DR) images from 18,631 subjects were used for the training, testing, and validation of the deep learning fracture detection model. We first built a pretrained model, a simple framework for contrastive learning of visual representations (simCLR), using contrastive learning with the training set. Then, simCLR was used as the backbone for a fully convolutional one-stage (FCOS) objective detection network to identify rib fractures from chest X-ray images. The detection performance of the network for four different types of rib fractures was evaluated using the testing set. A total of 127 images from Data-CZ and 109 images from Data-CH with the annotations for four types of rib fractures were used for evaluation. The results showed that for Data-CZ, the sensitivities of the detection model with no pretraining, pretrained ImageNet, and pretrained DR were 0.465, 0.735, and 0.822, respectively, and the average number of false positives per scan was five in all cases. For the Data-CH test set, the sensitivities of three different pretraining methods were 0.403, 0.655, and 0.748. In the identification of four fracture types, the detection model achieved the highest performance for displaced fractures, with sensitivities of 0.873 and 0.774 for the Data-CZ and Data-CH test sets, respectively, with 5 false positives per scan, followed by nondisplaced fractures, buckle fractures, and old fractures. A pretrained model can significantly improve the performance of the deep learning-based rib fracture detection based on X-ray images, which can reduce missed diagnoses and improve the diagnostic efficacy.

Timely Management of Simultaneous Bilateral Spontaneous Pneumothorax: A Near-Death Experience

Simultaneous bilateral spontaneous pneumothorax is a situation that rarely occurs. The patient can have various presentations, from dyspnoea and chest pain to significant respiratory failure. Although it causes lung collapse in nearly two thirds of cases, early diagnosis and treatment are of the utmost importance. Here, we present the case of an 18-year-old boy who presented with complaints of sudden onset respiratory distress. He was immediately put on mechanical ventilation. He was diagnosed with simultaneous bilateral spontaneous pneumothorax on chest X-ray. He needed bilateral intra-thoracic drainage, following which a video-assisted thoracoscopy was done on the left side.

Risk factors for air leakage during invasive mechanical ventilation in pediatric intensive care units

Purpose

This study aimed to investigate air leakage during invasive mechanical ventilation (IMV) in a pediatric intensive care unit (PICU) and explore potential risk factors.

Methods

We conducted a retrospective cohort study of children who underwent IMV in a single-center PICU in a tertiary referral hospital. Air leakage risk factors and factors associated with an improved outcome were assessed.

Results

A total of 548 children who underwent IMV were enrolled in this study. Air leakage occurred in 7.5% (41/548) of the cases in the PICU. Air leakage increased the duration of IMV and hospitalization time. Multivariate logistic regression analysis showed a higher risk of air leakage during IMV for PICU patients with acute respiratory dyspnea syndrome (ARDS) (OR = 4.38), a higher pediatric critical illness score (PCIS) (OR = 1.08), or a higher peak inspiratory pressure (PIP) (OR = 1.08), whereas the risk was lower for patients with central respiratory failure (OR = 0.14). The logistic model had excellent predictive power for air leakage, with an area under the curve of 0.883 and tenfold cross-validation. Patients aged between 1 and 6 years who were diagnosed with measles or pneumonia and had a low positive end-expiratory pressure (PEEP) or high PaO2/FiO2 ratio were associated with improved outcomes. Patients diagnosed with central respiratory failure or congenital heart diseases were associated with less desirable outcomes.

Conclusions

Patients with ARDS, a higher PCIS at admission or a higher PIP were at higher risk of air leakage.

Effects of Non-intubated Video-Assisted Thoracic Surgery on Patients With Pulmonary Dysfunction

Background: Non-intubated video-assisted thoracic surgery (NIVATS) can be safely performed in lung volume reduction surgery for patients with severe pulmonary dysfunction. However, there is still no cohort observation on the effects of NIVATS on patients with pulmonary dysfunction undergoing different types of thoracic procedures. This retrospective study aimed to observe the effects of NIVATS for this kind of patients.

Methods: Three hundred and twenty-eight patients with moderate to severe obstructive pulmonary dysfunction, who underwent video-assisted thoracic surgery (VATS), were retrospectively collected from June 1st, 2017 to September 30th, 2019. Patients in NIVATS were case-matched with those in intubated video-assisted thoracic surgery (IVATS) by a propensity score-matched analysis. The primary outcome was the comparison of perioperative values, the secondary outcome was the risk factors for postoperative clinical complications (PCP) which were identified by binary logistic regression analysis.

Results: After being matched, there were no differences in demographics and preoperative values of pulmonary function between NIVATS and IVATS groups. The duration of surgery and anesthesia had no difference (P = 0.091 and P = 0.467). As for the postoperative recovery, except for the mean intensive care unit (ICU) stay was longer in the IVATS group than in the NIVATS group (P = 0.015), the chest tube removal time and the postoperative hospital stay had no difference (P = 0.394 and P = 0.453), and the incidence of PCP also had no difference (P = 0.121). The binary logistic regression analysis revealed that the history of pulmonary disease, anesthesia method, and surgical location were risk factors of PCP.

Conclusion: For patients with pulmonary dysfunction when undergoing different types of thoracic procedures, the NIVATS can be performed as effectively and safely as the IVATS, and can reduce the ICU stay.