EDITORIAL
INT J TUBERC LUNG DIS 18(7):757 Q 2014 The Union http://dx.doi.org/10.5588/ijtld.14.0354
Chest radiography for diagnosis of tuberculosis in children: a problem of interpretation CHEST RADIOGRAPHY is frequently used to complement clinical evaluation in a child with suspected tuberculosis (TB), as sputum smear microscopy has a low sensitivity and a good quality sputum specimen can be difficult to obtain, particularly in young children. Chest radiography is also usually more readily available than other diagnostic modalities in resource-limited settings.1 A wide range of radiological abnormalities might be observed in a child with TB. The specificity of most of these abnormalities is low, and dependent on the epidemiological setting. Specificity is lower in high TB endemic settings where severe pneumonia, malnutrition and human immunodeficiency virus (HIV) infection in children are often also prevalent, with coinfections of respiratory pathogens also common in those settings. The sensitivity of chest radiography is also low, particularly for detecting enlargement of the regional lymph nodes, which is the commonest pathology following primary infection with Mycobacterium tuberculosis.2,3 The challenge is further exacerbated when the quality of the radiograph is poor or there is no lateral view available, both common occurrences. Therefore, the reliance on chest radiography can lead to a potentially large number of misdiagnoses, including overdiagnosis of TB in children. Nonetheless, given that improved diagnostic imaging alternatives such as computerised tomography are not available or even appropriate for wider use, can we do better with the imperfect tool that we have? In this issue of the Journal, Seddon and colleagues describe the outcomes of a training course designed to teach clinicians who regularly manage children with TB in HIV endemic settings how to interpret chest radiographs.4 The assessment tool and final classification groups were designed to be clear and pragmatic. The chest radiographs for those correctly classified as ‘abnormal - likely TB’ were from symptomatic children with microbiologically confirmed TB. Those correctly classified as ‘abnormal unlikely TB’ were from children with a proven alternative diagnosis, although evidence of prior infection with M. tuberculosis in this latter group was not stated. The training was provided by an experienced and expert group. The modest improvements recorded highlight the challenges. While the participants’ ability to correctly identify normal and unreadable chest radiographs from children improved slightly, the training did not improve their ability to correctly identify abnormal-
ities due to TB. A less experienced group of participants might perhaps have shown greater improvement by presumably starting at a lower baseline of knowledge, but the study found the opposite. The greatest improvements were in those with previous training, suggesting that ongoing revision of the skill is required. The study group were blinded to clinical information and were more likely to interpret abnormal chest radiographs as due to TB post-training. The participants may have been sensitised to TB, as it was the focus of the training itself. This has relevance for clinical practice in that readers of chest radiographs are not blinded (and nor should they be) to the clinical or epidemiological information which may influence interpretation. Diagnosis of TB in children is the most critical and central challenge to improving case detection and treatment outcome, as well as accuracy of surveillance data, research and advocacy. An imaging tool that is safe, low-cost, user-friendly and accurate, particularly for the identification (or exclusion) of perihilar adenopathy, could be a major advance. STEPHEN M. GRAHAM, FRACP, PHD Centre for International Child Health The University of Melbourne Department of Paediatrics and Murdoch Childrens Research Institute Royal Children’s Hospital Melbourne, VIC, Australia International Union Against Tuberculosis and Lung Disease Paris, France e-mail: [email protected]
References 1 Weismuller M M, Graham S M, Claessens N J M, Meijnen S, Salaniponi F, Harries A D. Diagnosis of childhood tuberculosis in Malawi: an audit of hospital practice. Int J Tuberc Lung Dis 2002; 6: 432–438. 2 Marais B J, Gie R P, Schaaf H S, Hesseling A C, Enarson D A, Beyers N. The spectrum of disease in children treated for tuberculosis in a highly endemic area. Int J Tuberc Lung Dis 2006; 10: 732–738. 3 Swingler G H, du Toit G, Andronikou S, van der Merwe L, Zar H J. Diagnostic accuracy of chest radiography in detecting mediastinal lymphadenopathy in suspected pulmonary tuberculosis. Arch Dis Child 2005; 90: 1153–1156. 4 Seddon J A, Padayachee T, Du Plessis A-M, et al. Teaching chest radiograph reading for child tuberculosis suspects. Int J Tuberc Lung Dis 2014; 18: 763–769.
INT J TUBERC LUNG DIS 18(7):757 Q 2014 The Union http://dx.doi.org/10.5588/ijtld.14.0354
Chest radiography for diagnosis of tuberculosis in children: a problem of interpretation CHEST RADIOGRAPHY is frequently used to complement clinical evaluation in a child with suspected tuberculosis (TB), as sputum smear microscopy has a low sensitivity and a good quality sputum specimen can be difficult to obtain, particularly in young children. Chest radiography is also usually more readily available than other diagnostic modalities in resource-limited settings.1 A wide range of radiological abnormalities might be observed in a child with TB. The specificity of most of these abnormalities is low, and dependent on the epidemiological setting. Specificity is lower in high TB endemic settings where severe pneumonia, malnutrition and human immunodeficiency virus (HIV) infection in children are often also prevalent, with coinfections of respiratory pathogens also common in those settings. The sensitivity of chest radiography is also low, particularly for detecting enlargement of the regional lymph nodes, which is the commonest pathology following primary infection with Mycobacterium tuberculosis.2,3 The challenge is further exacerbated when the quality of the radiograph is poor or there is no lateral view available, both common occurrences. Therefore, the reliance on chest radiography can lead to a potentially large number of misdiagnoses, including overdiagnosis of TB in children. Nonetheless, given that improved diagnostic imaging alternatives such as computerised tomography are not available or even appropriate for wider use, can we do better with the imperfect tool that we have? In this issue of the Journal, Seddon and colleagues describe the outcomes of a training course designed to teach clinicians who regularly manage children with TB in HIV endemic settings how to interpret chest radiographs.4 The assessment tool and final classification groups were designed to be clear and pragmatic. The chest radiographs for those correctly classified as ‘abnormal - likely TB’ were from symptomatic children with microbiologically confirmed TB. Those correctly classified as ‘abnormal unlikely TB’ were from children with a proven alternative diagnosis, although evidence of prior infection with M. tuberculosis in this latter group was not stated. The training was provided by an experienced and expert group. The modest improvements recorded highlight the challenges. While the participants’ ability to correctly identify normal and unreadable chest radiographs from children improved slightly, the training did not improve their ability to correctly identify abnormal-
ities due to TB. A less experienced group of participants might perhaps have shown greater improvement by presumably starting at a lower baseline of knowledge, but the study found the opposite. The greatest improvements were in those with previous training, suggesting that ongoing revision of the skill is required. The study group were blinded to clinical information and were more likely to interpret abnormal chest radiographs as due to TB post-training. The participants may have been sensitised to TB, as it was the focus of the training itself. This has relevance for clinical practice in that readers of chest radiographs are not blinded (and nor should they be) to the clinical or epidemiological information which may influence interpretation. Diagnosis of TB in children is the most critical and central challenge to improving case detection and treatment outcome, as well as accuracy of surveillance data, research and advocacy. An imaging tool that is safe, low-cost, user-friendly and accurate, particularly for the identification (or exclusion) of perihilar adenopathy, could be a major advance. STEPHEN M. GRAHAM, FRACP, PHD Centre for International Child Health The University of Melbourne Department of Paediatrics and Murdoch Childrens Research Institute Royal Children’s Hospital Melbourne, VIC, Australia International Union Against Tuberculosis and Lung Disease Paris, France e-mail: [email protected]
References 1 Weismuller M M, Graham S M, Claessens N J M, Meijnen S, Salaniponi F, Harries A D. Diagnosis of childhood tuberculosis in Malawi: an audit of hospital practice. Int J Tuberc Lung Dis 2002; 6: 432–438. 2 Marais B J, Gie R P, Schaaf H S, Hesseling A C, Enarson D A, Beyers N. The spectrum of disease in children treated for tuberculosis in a highly endemic area. Int J Tuberc Lung Dis 2006; 10: 732–738. 3 Swingler G H, du Toit G, Andronikou S, van der Merwe L, Zar H J. Diagnostic accuracy of chest radiography in detecting mediastinal lymphadenopathy in suspected pulmonary tuberculosis. Arch Dis Child 2005; 90: 1153–1156. 4 Seddon J A, Padayachee T, Du Plessis A-M, et al. Teaching chest radiograph reading for child tuberculosis suspects. Int J Tuberc Lung Dis 2014; 18: 763–769.
