Radiological significance of high-resolution computed tomography for elderly pulmonary tuberculosis patients - an...

Pulmonary tuberculosis (PTB) is one of the major health problems in the elderly population, causing significant morbidity and mortality. The aim of this study is to evaluate the significance of the high-resolution computed tomography (HRCT) modality for the diagnosis of PTB, in comparison to culture test.

Thoracic HRCT images of the study population, comprising 124 patients clinically suspected for PTB with smear and culture reports, were analysed for sensitivity and specificity of the HRCT test. Features of active PTB were centrilobular nodules, 'tree-in-bud' pattern densities, macro-nodules, consolidations, cavitary lesions, ground-glass opacities, and miliary nodules.

Among the study population, 108 cases presented HRCT features of active PTB and the remaining cases were negative but had presented a few features mimicking PTB. As inferred from positive culture test results, 106 cases had active PTB, the remaining cases were culture negative for PTB. False-positive (FP) or 'type I error' cases, and false-negative (FN) or 'type II error' cases were ascertained by Bayes' theorem. Sensitivity (true positive rate) and specificity (true negative rate) of HRCT test were 0.8125 and 0.8571, respectively.

For proper diagnosis the predictive capability, as two values of 'a posteriori probability', was computed; the mean value of 'a posteriori probability' for HRCT was 0.6358. When its culture test was positive, the HRCT test was 69.56-92.85% efficient in ascertaining positive results with a sample; on the other hand, when its culture test was negative it was 66.66-100% efficient for a negative result. Thus, the HRCT test is considerably dependable.