Purpose or Case Report: Diagnosing paediatric TB can be challenging due to non-specific signs and symptoms and difficulties in getting a microbiological confirmation. Diagnosis relies mainly on symptoms and chest X-ray (CXR), with hilar or mediastinal lymphadenopathy as the most common finding. However, CXR has poor inter- and intra-reader agreement for mediastinal lymphadenopathy. Ultrasound (US) does not expose children to ionizing radiation; is repeatable and due to the recent development of portable, low-cost US machines it can be used at the point of care and is cost-effective.
We investigated clinician-performed mediastinal US, comparing with CXR, for the detection of lymphadenopathy in children with suspected pulmonary TB.
Methods & Materials: Consecutive children younger than 13 years with suspected pulmonary TB were enrolled between October 2014 and September 2015. Comprehensive microbiologic investigations were done, and children were categorised into one of the three groups: definite TB (microbiological confirmed), possible TB (clinical diagnosis with TB treatment), unlikely TB (improvement on follow-up without TB treatment).
Mediastinal US scan was performed on a low-cost, black and white ultrasound machine (Mindray DP10, China), with the patient in a supine position examining the mediastinum with a micro-convex array transducer 7.5 MHz (range 5.0 - 8.5 MHz), through the suprasternal notch. A transverse and the coronal oblique view were obtained, figure 1. The sonographer, blinded for the final diagnosis, recorded the side and size of lymph nodes larger than 1 cm. A radiologist blinded for the sonographer’s findings and final diagnosis, recorded lymphadenopathy seen on the CXR.
Results: Ninety-four children (55.3% male) with a median age of 33.9 months (range: 5.2 – 145.6 months) were enrolled, 10.6% were HIV infected. Twelve children (12.8%) were categorized as definite TB, 64 (68.0%) as clinical TB and 18 (19.2%) as unlikely TB.
Sixteen children (17.0%) had lymphadenopathy on mediastinal US, in eight of these children lymphadenopathy was not seen on CXR, table 1. Two of the eight children with lymphadenopathy only on mediastinal US were categorized as definite TB, five as clinical TB and one as unlikely TB. In 26 children lymphadenopathy was noted on CXR but not on mediastinal US. See table 1 for a summary of the results.
Conclusions: These preliminary results show that mediastinal US is a useful tool, in addition to CXR, for the detection of mediastinal lymphadenopathy in children with pulmonary TB.