Search

Background: Spirometry quality assurance is a challenging task across levels of healthcare tiers, especially in primary care. Deep learning may serve as a support tool for enhancing spirometry quality. We aimed to develop a high accuracy and sensitive deep learning-based model aiming at assisting high-quality spirometry assurance., Methods: Spirometry PDF files retrieved from one hospital between October 2017 and October 2020 were labeled according to ATS/ERS 2019 criteria and divided into training and internal test sets. Additional files from three hospitals were used for external testing. A deep learning-based model was constructed and assessed to determine acceptability, usability, and quality rating for FEV 1 and FVC. System warning messages and patient instructions were also generated for general practitioners (GPs)., Results: A total of 16,502 files were labeled. Of these, 4592 curves were assigned to the internal test set, the remaining constituted the training set. In the internal test set, the model generated 95.1%, 92.4%, and 94.3% accuracy for FEV 1 acceptability, usability, and rating. The accuracy for FVC acceptability, usability, and rating were 93.6%, 94.3%, and 92.2%. With the assistance of the model, the performance of GPs in terms of monthly percentages of good quality (A, B, or C grades) tests for FEV 1 and FVC was higher by ~ 21% and ~ 36%, respectively., Conclusion: The proposed model assisted GPs in spirometry quality assurance, resulting in enhancing the performance of GPs in quality control of spirometry., (© 2022. The Author(s).)

Background: Due to the similar symptoms of upper airway obstruction to asthma, misdiagnosis is common. Spirometry is a cost-effective screening test for upper airway obstruction and its characteristic patterns involving fixed, variable intrathoracic and extrathoracic lesions. We aimed to develop a deep learning model to detect upper airway obstruction patterns and compared its performance with that of lung function clinicians., Methods: Spirometry records were reviewed to detect the possible condition of airway stenosis. Then they were confirmed by the gold standard (e.g., computed tomography, endoscopy, or clinic diagnosis of upper airway obstruction). Images and indices derived from flow-volume curves were used for training and testing the model. Clinicians determined cases using spirometry records from the test set. The deep learning model evaluated the same data., Results: Of 45,831 patients' spirometry records, 564 subjects with curves suggesting upper airway obstruction, after verified by the gold standard, 351 patients were confirmed. These cases and another 200 cases without airway stenosis were used as the training and testing sets. 432 clinicians evaluated 20 cases of each of the three patterns and 20 no airway stenosis cases (n = 80). They assigned an accuracy of 41.2% (±15.4) (interquartile range: 27.5-52.5%), with poor agreements (κ = 0.12). For the same cases, the model generated a correct detection of 81.3% (p < 0.0001)., Conclusions: Deep learning could detect upper airway obstruction patterns from other classic patterns of ventilatory defects with high accuracy, whereas clinicians presented marked errors and variabilities. The model may serve as a support tool to enhance clinicians' correct diagnosis of upper airway obstruction using spirometry., (© 2022 S. Karger AG, Basel.)

As COVID-19 is highly infectious, many patients can simultaneously flood into hospitals for diagnosis and treatment, which has greatly challenged public medical systems. Treatment priority is often determined by the symptom severity based on first assessment. However, clinical observation suggests that some patients with mild symptoms may quickly deteriorate. Hence, it is crucial to identify patient early deterioration to optimize treatment strategy. To this end, we develop an early-warning system with deep learning techniques to predict COVID-19 malignant progression. Our method leverages CT scans and the clinical data of outpatients and achieves an AUC of 0.920 in the single-center study. We also propose a domain adaptation approach to improve the generalization of our model and achieve an average AUC of 0.874 in the multicenter study. Moreover, our model automatically identifies crucial indicators that contribute to the malignant progression, including Troponin, Brain natriuretic peptide, White cell count, Aspartate aminotransferase, Creatinine, and Hypersensitive C-reactive protein., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2021 Elsevier B.V. All rights reserved.)

Background Cerebral aneurysm detection is a challenging task. Deep learning may become a supportive tool for more accurate interpretation. Purpose To develop a highly sensitive deep learning-based algorithm that assists in the detection of cerebral aneurysms on CT angiography images. Materials and Methods Head CT angiography images were retrospectively retrieved from two hospital databases acquired across four different scanners between January 2015 and June 2019. The data were divided into training and validation sets; 400 additional independent CT angiograms acquired between July and December 2019 were used for external validation. A deep learning-based algorithm was constructed and assessed. Both internal and external validation were performed. Jackknife alternative free-response receiver operating characteristic analysis was performed. Results A total of 1068 patients (mean age, 57 years ± 11 [standard deviation]; 660 women) were evaluated for a total of 1068 CT angiograms encompassing 1337 cerebral aneurysms. Of these, 534 CT angiograms (688 aneurysms) were assigned to the training set, and the remaining 534 CT angiograms (649 aneurysms) constituted the validation set. The sensitivity of the proposed algorithm for detecting cerebral aneurysms was 97.5% (633 of 649; 95% CI: 96.0, 98.6). Moreover, eight new aneurysms that had been overlooked in the initial reports were detected (1.2%, eight of 649). With the aid of the algorithm, the overall performance of radiologists in terms of area under the weighted alternative free-response receiver operating characteristic curve was higher by 0.01 (95% CI: 0.00, 0.03). Conclusion The proposed deep learning algorithm assisted radiologists in detecting cerebral aneurysms on CT angiography images, resulting in a higher detection rate. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Kallmes and Erickson in this issue.