Using artificial intelligence to read chest radiographs for tuberculosis detection: A multi-site evaluation of the diagnostic accuracy of three deep learning systems

• Published in: Scientific reports 2019-10, Vol.9 (1), p.15000-10, Article 15000
• Main Authors: Qin, Zhi Zhen, Sander, Melissa S., Rai, Bishwa, Titahong, Collins N., Sudrungrot, Santat, Laah, Sylvain N., Adhikari, Lal Mani, Carter, E. Jane, Puri, Lekha, Codlin, Andrew J., Creswell, Jacob
• Format: Article
• Language: English
• Subjects: 692/699/255/1856; 692/700/1421/1770; 692/700/478/2772; Adult; Area Under Curve; Artificial intelligence; Cameroon - epidemiology; Chest; Data Accuracy; Deep Learning; DNA, Bacterial - genetics; Female; Humanities and Social Sciences; Humans; Male; Mass Screening - methods; Middle Aged; multidisciplinary; Mycobacterium tuberculosis - genetics; Nepal - epidemiology; Neural networks; Nucleic Acid Amplification Techniques; Radiography; Radiography, Thoracic - methods; Retrospective Studies; Science; Science (multidisciplinary); Sensitivity and Specificity; Triage; Tuberculosis; Tuberculosis, Pulmonary - diagnostic imaging; Tuberculosis, Pulmonary - epidemiology; Tuberculosis, Pulmonary - microbiology
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-019-51503-3

Deep learning (DL) neural networks have only recently been employed to interpret chest radiography (CXR) to screen and triage people for pulmonary tuberculosis (TB). No published studies have compared multiple DL systems and populations. We conducted a retrospective evaluation of three DL systems (CAD4TB, Lunit INSIGHT, and qXR) for detecting TB-associated abnormalities in chest radiographs from outpatients in Nepal and Cameroon. All 1196 individuals received a Xpert MTB/RIF assay and a CXR read by two groups of radiologists and the DL systems. Xpert was used as the reference standard. The area under the curve of the three systems was similar: Lunit (0.94, 95% CI: 0.93–0.96), qXR (0.94, 95% CI: 0.92–0.97) and CAD4TB (0.92, 95% CI: 0.90–0.95). When matching the sensitivity of the radiologists, the specificities of the DL systems were significantly higher except for one. Using DL systems to read CXRs could reduce the number of Xpert MTB/RIF tests needed by 66% while maintaining sensitivity at 95% or better. Using a universal cutoff score resulted different performance in each site, highlighting the need to select scores based on the population screened. These DL systems should be considered by TB programs where human resources are constrained, and automated technology is available.