Improving the Generalizability and Performance of an Ultrasound Deep Learning Model Using Limited Multicenter Data for Lung Sliding Artifact Identification.

Deep learning (DL) models for medical image classification frequently struggle to generalize to data from outside institutions. Additional clinical data are also rarely collected to comprehensively assess and understand model performance amongst subgroups. Following the development of a single-center model to identify the lung sliding artifact on lung ultrasound (LUS), we pursued a validation strategy using external LUS data. As annotated LUS data are relatively scarce-compared to other medical imaging data-we adopted a novel technique to optimize the use of limited external data to improve model generalizability. Externally acquired LUS data from three tertiary care centers, totaling 641 clips from 238 patients, were used to assess the baseline generalizability of our lung sliding model. We then employed our novel Threshold-Aware Accumulative Fine-Tuning (TAAFT) method to fine-tune the baseline model and determine the minimum amount of data required to achieve predefined performance goals. A subgroup analysis was also performed and Grad-CAM++ explanations were examined. The final model was fine-tuned on one-third of the external dataset to achieve 0.917 sensitivity, 0.817 specificity, and 0.920 area under the receiver operator characteristic curve (AUC) on the external validation dataset, exceeding our predefined performance goals. Subgroup analyses identified LUS characteristics that most greatly challenged the model's performance. Grad-CAM++ saliency maps highlighted clinically relevant regions on M-mode images. We report a multicenter study that exploits limited available external data to improve the generalizability and performance of our lung sliding model while identifying poorly performing subgroups to inform future iterative improvements. This approach may contribute to efficiencies for DL researchers working with smaller quantities of external validation data.

Prospective Real-Time Validation of a Lung Ultrasound Deep Learning Model in the ICU.

OBJECTIVES: To evaluate the accuracy of a bedside, real-time deployment of a deep learning (DL) model capable of distinguishing between normal (A line pattern) and abnormal (B line pattern) lung parenchyma on lung ultrasound (LUS) in critically ill patients. DESIGN: Prospective, observational study evaluating the performance of a previously trained LUS DL model. Enrolled patients received a LUS examination with simultaneous DL model predictions using a portable device. Clip-level model predictions were analyzed and compared with blinded expert review for A versus B line pattern. Four prediction thresholding approaches were applied to maximize model sensitivity and specificity at bedside. SETTING: Academic ICU. PATIENTS: One-hundred critically ill patients admitted to ICU, receiving oxygen therapy, and eligible for respiratory imaging were included. Patients who were unstable or could not undergo an LUS examination were excluded. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: A total of 100 unique ICU patients (400 clips) were enrolled from two tertiary-care sites. Fifty-six patients were mechanically ventilated. When compared with gold standard expert annotation, the real-time inference yielded an accuracy of 95%, sensitivity of 93%, and specificity of 96% for identification of the B line pattern. Varying prediction thresholds showed that real-time modification of sensitivity and specificity according to clinical priorities is possible. CONCLUSIONS: A previously validated DL classification model performs equally well in real-time at the bedside when platformed on a portable device. As the first study to test the feasibility and performance of a DL classification model for LUS in a dedicated ICU environment, our results justify further inquiry into the impact of employing real-time automation of medical imaging into the care of the critically ill.

Diagnostic sensitivity of pleural fluid cytology in malignant pleural effusions: systematic review and meta-analysis.

BACKGROUND: Pleural fluid cytology is an important diagnostic test used for the investigation of pleural effusions. There is considerable variability in the reported sensitivity for the diagnosis of malignant pleural effusions (MPE) in the literature. OBJECTIVE: The purpose of this review is to determine the diagnostic sensitivity of pleural fluid cytology for MPE, both overall and by tumour type, to better inform the decision-making process when investigating pleural effusions. DATA SOURCES: A literature search of EMBASE and MEDLINE was performed by four reviewers. Articles satisfying inclusion criteria were evaluated for bias using the QUADAS-2 tool. DATA EXTRACTION: For quantitative analysis, we performed a metaanalysis using a binary random-effects model to determine pooled sensitivity. Subgroup analysis was performed based on primary cancer site and meta-regression by year of publication. SYNTHESIS: Thirty-six studies with 6057 patients with MPE were included in the meta-analysis. The overall diagnostic sensitivity of pleural fluid cytology for MPE was 58.2% (95% CI 52.5% to 63.9%; range 20.5%-86.0%). There was substantial heterogeneity present among studies (I2 95.5%). For primary thoracic malignancies, sensitivity was highest in lung adenocarcinoma (83.6%; 95% CI 77.7% to 89.6%) and lowest in lung squamous cell carcinoma (24.2%; 95% CI 17.0% to 31.5%) and mesothelioma (28.9%; 95% CI 16.2% to 41.5%). For malignancies with extrathoracic origin, sensitivity was high for ovarian cancer (85.2%; 95% CI 74.2% to 96.1%) and modest for breast cancer (65.3%; 95% CI 49.8% to 80.8%). CONCLUSIONS: Pleural fluid cytology has an overall sensitivity of 58.2% for the diagnosis of MPE. Clinicians should be aware of the high variability in diagnostic sensitivity by primary tumour type as well as the potential reasons for false-negative cytology results.PROSPERO registration numberCRD42021231473.

Enhancing Annotation Efficiency with Machine Learning: Automated Partitioning of a Lung Ultrasound Dataset by View.

BACKGROUND: Annotating large medical imaging datasets is an arduous and expensive task, especially when the datasets in question are not organized according to deep learning goals. Here, we propose a method that exploits the hierarchical organization of annotating tasks to optimize efficiency. METHODS: We trained a machine learning model to accurately distinguish between one of two classes of lung ultrasound (LUS) views using 2908 clips from a larger dataset. Partitioning the remaining dataset by view would reduce downstream labelling efforts by enabling annotators to focus on annotating pathological features specific to each view. RESULTS: In a sample view-specific annotation task, we found that automatically partitioning a 780-clip dataset by view saved 42 min of manual annotation time and resulted in 55±6 additional relevant labels per hour. CONCLUSIONS: Automatic partitioning of a LUS dataset by view significantly increases annotator efficiency, resulting in higher throughput relevant to the annotating task at hand. The strategy described in this work can be applied to other hierarchical annotation schemes.

Accurate assessment of the lung sliding artefact on lung ultrasonography using a deep learning approach.

Pneumothorax is a potentially life-threatening condition that can be rapidly and accurately assessed via the lung sliding artefact generated using lung ultrasound (LUS). Access to LUS is challenged by user dependence and shortage of training. Image classification using deep learning methods can automate interpretation in LUS and has not been thoroughly studied for lung sliding. Using a labelled LUS dataset from 2 academic hospitals, clinical B-mode (also known as brightness or two-dimensional mode) videos featuring both presence and absence of lung sliding were transformed into motion (M) mode images. These images were subsequently used to train a deep neural network binary classifier that was evaluated using a holdout set comprising 15% of the total data. Grad-CAM explanations were examined. Our binary classifier using the EfficientNetB0 architecture was trained using 2535 LUS clips from 614 patients. When evaluated on a test set of data uninvolved in training (540 clips from 124 patients), the model performed with a sensitivity of 93.5%, specificity of 87.3% and an area under the receiver operating characteristic curve (AUC) of 0.973. Grad-CAM explanations confirmed the model's focus on relevant regions on M-mode images. Our solution accurately distinguishes between the presence and absence of lung sliding artefacts on LUS.

Automation of Lung Ultrasound Interpretation via Deep Learning for the Classification of Normal versus Abnormal Lung Parenchyma: A Multicenter Study.

Lung ultrasound (LUS) is an accurate thoracic imaging technique distinguished by its handheld size, low-cost, and lack of radiation. User dependence and poor access to training have limited the impact and dissemination of LUS outside of acute care hospital environments. Automated interpretation of LUS using deep learning can overcome these barriers by increasing accuracy while allowing point-of-care use by non-experts. In this multicenter study, we seek to automate the clinically vital distinction between A line (normal parenchyma) and B line (abnormal parenchyma) on LUS by training a customized neural network using 272,891 labelled LUS images. After external validation on 23,393 frames, pragmatic clinical application at the clip level was performed on 1162 videos. The trained classifier demonstrated an area under the receiver operating curve (AUC) of 0.96 (±0.02) through 10-fold cross-validation on local frames and an AUC of 0.93 on the external validation dataset. Clip-level inference yielded sensitivities and specificities of 90% and 92% (local) and 83% and 82% (external), respectively, for detecting the B line pattern. This study demonstrates accurate deep-learning-enabled LUS interpretation between normal and abnormal lung parenchyma on ultrasound frames while rendering diagnostically important sensitivity and specificity at the video clip level.

Development of a convolutional neural network to differentiate among the etiology of similar appearing pathological B lines on lung ultrasound: a deep learning study.

OBJECTIVES: Lung ultrasound (LUS) is a portable, low-cost respiratory imaging tool but is challenged by user dependence and lack of diagnostic specificity. It is unknown whether the advantages of LUS implementation could be paired with deep learning (DL) techniques to match or exceed human-level, diagnostic specificity among similar appearing, pathological LUS images. DESIGN: A convolutional neural network (CNN) was trained on LUS images with B lines of different aetiologies. CNN diagnostic performance, as validated using a 10% data holdback set, was compared with surveyed LUS-competent physicians. SETTING: Two tertiary Canadian hospitals. PARTICIPANTS: 612 LUS videos (121 381 frames) of B lines from 243 distinct patients with either (1) COVID-19 (COVID), non-COVID acute respiratory distress syndrome (NCOVID) or (3) hydrostatic pulmonary edema (HPE). RESULTS: The trained CNN performance on the independent dataset showed an ability to discriminate between COVID (area under the receiver operating characteristic curve (AUC) 1.0), NCOVID (AUC 0.934) and HPE (AUC 1.0) pathologies. This was significantly better than physician ability (AUCs of 0.697, 0.704, 0.967 for the COVID, NCOVID and HPE classes, respectively), p<0.01. CONCLUSIONS: A DL model can distinguish similar appearing LUS pathology, including COVID-19, that cannot be distinguished by humans. The performance gap between humans and the model suggests that subvisible biomarkers within ultrasound images could exist and multicentre research is merited.

Effectiveness of guideline dissemination and implementation strategies on health care professionals' behaviour and patient outcomes in the cancer care context: a systematic review.

BACKGROUND: Health care professionals (HCPs) use clinical practice guidelines (CPGs) to make evidence-informed decisions regarding patient care. Although a large number of cancer-related CPGs exist, it is unknown which CPG dissemination and implementation strategies are effective for improving HCP behaviour and patient outcomes in a cancer care context. This review aimed to determine the effectiveness of CPG dissemination and/or implementation strategies among HCPs in a cancer care context. METHODS: A comprehensive search of five electronic databases was conducted. Studies were limited to the dissemination and/or implementation of a CPG targeting both medical and/or allied HCPs in cancer care. Two reviewers independently coded strategies using the Mazza taxonomy, extracted study findings, and assessed study quality. RESULTS: The search strategy identified 33 studies targeting medical and/or allied HCPs. Across the 33 studies, 23 of a possible 49 strategies in the Mazza taxonomy were used, with a mean number of 3.25 (SD = 1.45) strategies per intervention. The number of strategies used per intervention was not associated with positive outcomes. Educational strategies (n = 24), feedback on guideline compliance (n = 11), and providing reminders (n = 10) were the most utilized strategies. When used independently, providing reminders and feedback on CPG compliance corresponded with positive significant changes in outcomes. Further, when used as part of multi-strategy interventions, group education and organizational strategies (e.g. creation of an implementation team) corresponded with positive significant changes in outcomes. CONCLUSIONS: Future CPG dissemination and implementation interventions for cancer care HCPs may benefit from utilizing the identified strategies. Research in this area should aim for better alignment between study objectives, intervention design, and evaluation measures, and should seek to incorporate theory in intervention design, so that behavioural antecedents are considered and measured; doing so would enhance the field's understanding of the causal mechanisms by which interventions lead, or do not lead, to changes in outcomes at all levels.

Legionella co-infection in a patient with COVID-19.

We describe a case of an 80-year-old man with COVID-19 and Legionella bacterial co-infection who initially presented to hospital with fever, respiratory symptoms, and diarrhea with radiographic evidence of atypical infection. His initial nasopharyngeal swab was negative; however, a subsequent swab was positive. A Legionella urinary antigen test was positive for Legionella pneumophilia serogroup 1 antigen. Despite a low prevalence of bacterial co-infection in patients with COVID-19, a large number of patients receive antimicrobial therapy. Based on clinical context, a high index of suspicion is warranted for both bacterial and viral infectious processes during the COVID-19 pandemic; this will help to ensure that appropriate antimicrobial therapy is used.

Les auteurs décrivent le cas d'un homme de 80 ans co-infecté par la COVID-19 et la légionellose bactérienne qui a consulté à l'hôpital à cause de fièvre, de symptômes respiratoires et de diarrhée et dont la radiographie démontrait une infection atypique. Le premier écouvillon nasopharyngé a donné un résultat négatif, mais un écouvillon subséquent s'est révélé positif. Un test d'antigène urinaire des légionelles était positif à l'antigène Legionella pneumophilia du sérogroupe 1. Malgré une faible prévalence de co-infection bactérienne chez les patients atteints de la COVID-19, de nombreux patients reçoivent des antimicrobiens. D'après le contexte clinique, il faut faire preuve de vigilance à l'égard des processus bactériens et viraux pendant la pandémie de COVID-19 afin de s'assurer d'utiliser des antimicrobiens appropriés.

Effectiveness of guideline dissemination and implementation strategies on health care professionals' behaviour and patient outcomes in the cancer care context: a systematic review protocol.

BACKGROUND: Health care professionals (HCPs) are able to make effective decisions regarding patient care through the use of systematically developed clinical practice guidelines (CPGs). These recommendations are especially important in a cancer health care context as patients are exposed to a multitude of interdisciplinary HCPs offering high-quality care throughout diagnosis, treatment, survivorship and palliative care. Although a large number of CPGs targeted towards cancer are widely disseminated, it is unknown whether implementation strategies targeting the use of these guidelines are effective in effecting HCP behaviour and patient outcomes in the cancer care context. The purpose of this systematic review will be to determine the effectiveness of different CPG dissemination and implementation interventions on HCPs' behaviour and patient outcomes in the cancer health care context. METHODS/DESIGN: Five electronic databases (CINAHL, the Cochrane Controlled Trials Register, MEDLINE via Ovid, EMBASE via Ovid and PsycINFO via Ovid) will be searched to include all studies examining the dissemination and/or implementation of CPGs in a cancer care setting targeting all HCPs. CPG implementation strategies will be included if the CPGs were systematically developed (e.g. literature review/evidence-informed, expert panel, evidence appraisal). The studies will be limited to randomized controlled trials, controlled clinical trials and quasi-experimental (interrupted time series, controlled before-and-after designs) studies. Two independent reviewers will assess articles for eligibility, data extraction and quality appraisal. DISCUSSION: The aim of this review is to inform cancer care health care professionals and policymakers about evidence-based implementation strategies that will allow for effective use of CPGs. SYSTEMATIC REVIEW REGISTRATION: PROSPERO CRD42015019331.