Late Breaking Abstract - A Deep Learning Algorithm for Classifying Fibrotic Lung Disease on High Resolution Computed Tomography

Author(s):  
Simon Walsh ◽  
Lucio Calandriello ◽  
Mario Silva ◽  
Nicola Sverzellati
2020 ◽  
Vol 79 (Suppl 1) ◽  
pp. 1598.2-1599
Author(s):  
I. Rusu ◽  
L. Muntean ◽  
M. M. Tamas ◽  
I. Felea ◽  
L. Damian ◽  
...  

Background:Interstitial lung disease (ILD) is a common manifestation of connective tissue diseases (CTDs), and is associated with significant morbidity and mortality. Chest high-resolution computed tomography (HRCT) play an important role in the diagnosis of ILD and may provide prognostic information.Objectives:We aimed to characterize the clinical profile and chest HRCT abnormalities and patterns of patients diagnosed with CTDs and ILD.Methods:In this retrospective, observational study we included 80 consecutive patients with CTDs and ILD referred to a tertiary rheumatology center between 2015 and 2019. From hospital charts we collected clinical data, immunologic profile, chest HRCT findings. HRCT patterns were defined according to new international recommendations.Results:Out of 80 patients, 64 (80%) were women, with a mean age of 55 years old. The most common CTD associated with ILD was systemic sclerosis (38.8%), followed by polymyositis (22.5%) and rheumatoid arthritis (18.8%). The majority of patients had dyspnea on exertion (71.3%), bibasilar inspiratory crackles were present in 56.3% patients and 10% had clubbing fingers. Antinuclear antibodies (ANA) were present in 78.8% patients, and the most frequently detected autoantibodies against extractable nuclear antigen were anti-Scl 70 (28.8%), followed by anti-SSA (anti-Ro, 17.5%), anti-Ro52 (11.3%) and anti-Jo (7.5%). Intravenous cyclophosphamide therapy for 6-12 months was used in 35% of patients, while 5% of patients were treated with mycophenolate mofetil.The most frequent HRCT abnormalities were reticular abnormalities and ground glass opacity. Non-specific interstitial pneumonia (NSIP) was identified in 46.3% CTDs patients. A pattern suggestive of usual interstitial pneumonia (UIP) was present in 32.5% patients, mainly in patients with systemic sclerosis. In 21.3% patients the HRCT showed reticulo-nodular pattern, micronodules and other abnormalities, not diagnostic for UIP or NSIP pattern.Conclusion:Nonspecific interstitial pneumonia (NSIP) is the most common HRCT pattern associated with CTDs. Further prospective longitudinal studies are needed in order to determine the clinical and prognostic significance of various HRCT patterns encountered in CTD-associated ILD and for better patient management.References:[1]Ohno Y, Koyama H, Yoshikaua T, Seki S. State-of-the-Art Imaging of the Lung for Connective Tissue Disease (CTD). Curr Rheumatol Rep. 2015;17(12):69.[2]Walsh SLF, Devaraj A, Enghelmeyer JI, Kishi K, Silva RS, Patel N, et al. Role of imaging in progressive-fibrosing interstitial lung diseases. Eur Respir Rev. 2018;27(150)Disclosure of Interests:None declared


Cancers ◽  
2021 ◽  
Vol 13 (4) ◽  
pp. 652 ◽  
Author(s):  
Carlo Augusto Mallio ◽  
Andrea Napolitano ◽  
Gennaro Castiello ◽  
Francesco Maria Giordano ◽  
Pasquale D'Alessio ◽  
...  

Background: Coronavirus disease 2019 (COVID-19) pneumonia and immune checkpoint inhibitor (ICI) therapy-related pneumonitis share common features. The aim of this study was to determine on chest computed tomography (CT) images whether a deep convolutional neural network algorithm is able to solve the challenge of differential diagnosis between COVID-19 pneumonia and ICI therapy-related pneumonitis. Methods: We enrolled three groups: a pneumonia-free group (n = 30), a COVID-19 group (n = 34), and a group of patients with ICI therapy-related pneumonitis (n = 21). Computed tomography images were analyzed with an artificial intelligence (AI) algorithm based on a deep convolutional neural network structure. Statistical analysis included the Mann–Whitney U test (significance threshold at p < 0.05) and the receiver operating characteristic curve (ROC curve). Results: The algorithm showed low specificity in distinguishing COVID-19 from ICI therapy-related pneumonitis (sensitivity 97.1%, specificity 14.3%, area under the curve (AUC) = 0.62). ICI therapy-related pneumonitis was identified by the AI when compared to pneumonia-free controls (sensitivity = 85.7%, specificity 100%, AUC = 0.97). Conclusions: The deep learning algorithm is not able to distinguish between COVID-19 pneumonia and ICI therapy-related pneumonitis. Awareness must be increased among clinicians about imaging similarities between COVID-19 and ICI therapy-related pneumonitis. ICI therapy-related pneumonitis can be applied as a challenge population for cross-validation to test the robustness of AI models used to analyze interstitial pneumonias of variable etiology.


Sign in / Sign up

Export Citation Format

Share Document