Artificial neural network models to support the diagnosis of pleural tuberculosis in adult patients

• Published in: The international journal of tuberculosis and lung disease Vol. 17; no. 5; pp. 682 - 686
• Main Authors: Seixas, J. M., Faria, J., Souza Filho, J. B. O., Vieira, A. F. M., Kritski, A., Trajman, A.
• Format: Journal Article
• Language: English
• Published: Paris, France International Union Against Tuberculosis and Lung Disease 01-05-2013; International Union against Tuberculosis and Lung Disease
• Subjects: Accuracy; Adenosine deaminase; Adult; Artificial Intelligence; Bacterial diseases; Bacterial diseases of the respiratory system; Bacteriological Techniques; Biological and medical sciences; Biopsy; Coinfection; Diagnosis; Diagnosis, Computer-Assisted; Discriminant Analysis; Disease Progression; Early Diagnosis; HIV Infections - diagnosis; Human bacterial diseases; Human immunodeficiency virus; Humans; Infectious diseases; Linear Models; Medical sciences; Mycobacterium tuberculosis - isolation & purification; Neural Networks (Computer); Nonlinear Dynamics; Paracentesis; Pleural Effusion - microbiology; Pleurisy; Pneumology; Predictive Value of Tests; Prognosis; Prospective Studies; Respiratory system : syndromes and miscellaneous diseases; Tuberculosis; Tuberculosis and atypical mycobacterial infections; Tuberculosis, Pleural - diagnosis; Tuberculosis, Pleural - microbiology; Human; Support; Respiratory disease; Pleural tuberculosis; tuberculosis; Patient; Neural network; Accuracy; Pleurisy; Pleural disease; Adult; Models; Diagnosis; Artificial intelligence
• ISSN: 1027-3719; 1815-7920
• DOI: 10.5588/ijtld.12.0829

BACKGROUND: Clinicians in countries with high tuberculosis (TB) prevalence often treat pleural TB based on clinical grounds, as the availability and sensitivity of diagnostic tests are poor.OBJECTIVE: To evaluate the role of artificial neural networks (ANN) as an aid for the non-invasive
diagnosis of pleural TB. These tools can be used in simple computer devices (tablets) without remote internet connection.METHODS: The clinical history and human immunodeficiency virus (HIV) status of 137 patients were prospectively entered in a database. Both non-linear ANN and the linear
Fisher discriminant were used to calculate performance indexes based on clinical grounds. The same procedure was performed including pleural fluid test results (smear, culture, adenosine deaminase, serology and nucleic acid amplification test). The gold standard was any positive test for TB.RESULTS:
In pre-test modelling, the neural model reached >90% accuracy (Fisher discriminant 74.5%). Under pre-test conditions, ANN had better accuracy compared to each test considered separately.CONCLUSIONS: ANN are highly reliable for diagnosing pleural TB based on clinical grounds and HIV
status only, and are useful even in remote conditions lacking access to sophisticated medical or computer infrastructure. In other better-equipped scenarios, these tools should be evaluated as substitutes for thoracocentesis and pleural biopsy.