Chest x-ray analysis with deep learning-based software as a triage test for pulmonary tuberculosis: a prospective study of diagnostic accuracy for culture-confirmed disease

Abstract

Background Deep learning-based radiological image analysis could facilitate use of chest x-rays as triage tests for pulmonary tuberculosis in resource-limited settings. We sought to determine whether commercially available chest x-ray analysis software meet WHO recommendations for minimal sensitivity and specificity as pulmonary tuberculosis triage tests. MethodsWe recruited symptomatic adults at the Indus Hospital, Karachi, Pakistan. We compared two software, qXR version 2.0 (qXRv2) and CAD4TB version 6.0 (CAD4TBv6), with a reference of mycobacterial culture of two sputa. We assessed qXRv2 using its manufacturer prespecified threshold score for chest x-ray classification as tuberculosis present versus not present. For CAD4TBv6, we used a data-derived threshold, because it does not have a prespecified one. We tested for non-inferiority to preset WHO recommendations (0•90 for sensitivity, 0•70 for specificity) using a non-inferiority limit of 0•05. We identified factors associated with accuracy by stratification and logistic regression. Findings We included 2198 (92•7%) of 2370 enrolled participants. 2187 (99•5%) of 2198 were HIV-negative, and 272 (12•4%) had culture-confirmed pulmonary tuberculosis. For both software, accuracy was non-inferior to WHO-recommended minimum values (qXRv2 sensitivity 0•93 [95% CI 0•89-0•95], non-inferiority p=0•0002; CAD4TBv6 sensitivity 0•93 [0•90-0•96], p<0•0001; qXRv2 specificity 0•75 [0•73-0•77], p<0•0001; CAD4TBv6 specificity 0•69 [0•67-0•71], p=0•0003). Sensitivity was lower in smear-negative pulmonary tuberculosis for both software, and in women for CAD4TBv6. Specificity was lower in men and in those with previous tuberculosis, and reduced with increasing age and decreasing body mass index. Smoking and diabetes did not affect accuracy.Interpretation In an HIV-negative population, these software met WHO-recommended minimal accuracy for pulmonary tuberculosis triage tests. Sensitivity will be lower when smear-negative pulmonary tuberculosis is more prevalent.

Application of artificial intelligence in digital chest radiography reading for pulmonary tuberculosis screening

“…A prospective study was conducted in a tertiary hospital from Pakistan to evaluate the performance of qXR(v2) and CAD4TB(v6) with mycobacterial culture being a standard reference, which collected a dataset of 2187 CXRs and reported that both software achieved non-inferior accuracy (qXR: sensitivity = 0.93; specificity = 0.75; CAD4TB: sensitivity = 0.93; specificity = 0.69) to WHO-recommended minimum values, while sensitivity might be lower when smear-negative pulmonary TB was more prevalent. 27 A retrospective case–control study was conducted in a tertiary hospital from India using microbiologically-confirmed TB as the reference standard, which reported that qXR (v2) could detect TB with an AUC of 0.81, a sensitivity of 71% and a specificity of 80%. 28 As many patients who present at tertiary hospitals have symptoms suggestive of TB, accurate and rapid triage tests that can rule out the disease are needed.…”

Accuracy of computer-aided chest X-ray screening in the Kenya National Tuberculosis Prevalence Survey

“…(6,7,15,16) Previous evaluations of CAD software have been mostly conducted in triage testing use situations, with very little data available to evaluate accuracy in community-based TB screening interventions. (6,8,16,17) Our aim was to evaluate the accuracy of the Computer-Aided Detection for Tuberculosis version 6 (CAD4TBv6) system for TB screening using a large data set (n=61,848) from the 2016 Kenya National TB prevalence survey. (18,19) To do this we used a Bayesian modelling approach to evaluate the accuracy of CAD4TBv6 and Clinical Officer CXR interpretation against the bacteriological reference standard used within the prevalence survey.…”

How does artificial intelligence in radiology improve efficiency and health outcomes?

“…This can be used as an autonomous pre-screening tool to reduce the use of microbiological tests, Study demonstrates the impact of the software on society by performing an economic analysis Fig. 1 Six objectives that can be pursued with artificial intelligence in radiology to improve efficiency and health outcomes which are more time-consuming and costly (levels 2, 3, 6) [10][11][12][13]. This is one of the first AI applications in radiology where the software functions autonomously and has taken over the task of the radiologist.…”