Table 1. Summary of published studies evaluating cavitation on chest radiograph or computerized tomography (CT) scan after anti-tuberculous treatment.

Reference	Design	Population	Relevant measure(s)	Frequencies
Sonnenberg (2000)20	Prognostic cohort for TB outcomes (exclusion of MDR cases and those not cured of TB), stratified by HIV status	Gold miners in South Africa	Chest X-ray at 3 & 6 months after cure	Residual cavitation at cure - 69/326 (21%)
				[HIV strong first factor for TB recurrence among those without cavitation]
Bombarda (2003)21	Repeat measures (unclear if losses to follow-up) during active disease and then after treatment conclusion	Referral hospital PTB patients	Repeated conventional CT scans	Thick-walled cavities post treatment 1/20 (5%) and thin-walled cavities post treatment 5/20 (25%); together, 30%
		? selected population
De Vallière (2004)19	Post-treatment cross-sectional survey (reporting on 33/42 patients who completed treatment – no comment on selection)	Patients registered in MDR-TB programme in Limpopo province, South Africa	Chest X-ray based on two observers	% with cavities – observer 2, 17/33 (52%) to observer 1, 23/33 (70%)
Hamilton (2008)17	Prognostic cohort of post-treatment (6 months), 170 exclusions from 1 004 subjects primarily for missing test results	Multi-centre North American TB trials Consortium RCT	Chest X-ray, consensus criteria, reading kappa 0.54 for cavity (80% raw agreement)	EOT cavity, 23.3% (n = 834) in EOT chest X-ray analysis vs 19.1% among those excluded from main analysis (n = 170)
Lee (2008)18	Pre- and post-treatment repeat scans (excluded 31/83, primarily loss to follow-up, 23)	Taiwan, China, general hospital n = 52	High resolution chest CT scan	Post-treatment, 18/52 (35%) vs pre-treatment, 38/52 (73%)

EOT, end-of-treatment; HIV, human immunodeficiency virus; MDR, multi-drug resistant; PTB, pulmonary tuberculosis; RCT, randomized controlled trial; TB, tuberculosis.