Purpose or Case Report: Diagnosing pediatric pulmonary tuberculosis (PTB) may be challenging as getting microbiological confirmation is difficult. Chest X-Ray (CXR) involves radiation exposure and has poor inter-reader agreement. Mediastinal lymphadenopathy can also be detected by mediastinal ultrasound (US). We describe technical aspects of performing mediastinal US and the findings in children with suspected PTB.
Methods & Materials: Consecutive children <13 years with suspected PTB were enrolled between June 2014 and May 2016. Children were categorized as definite TB (microbiological confirmed), possible TB (clinical diagnosis with TB treatment) or unlikely TB (improvement on follow-up without TB treatment).

Technique
A black-and-white US machine with a micro-convex array transducer 7.5 MHz (range 5.0 - 8.5 MHz) was used. The child was placed in the supine position with the neck slightly extended. A transverse and coronal oblique view were obtained. The probe was placed in the suprasternal notch, for the transverse view the probe was placed transverse, a tilting motion pointed the probe caudally into the chest to visualize the mediastinum. For the coronal oblique view the probe was rotated anti-clockwise, oriented diagonally with the inferior point on the patient’s right; a tilting motion pointed the US beam into the thorax in line with the aortic arch.
Figure 1 shows the identifiable anatomic for these two views.

Lymphadenopathy was recognised as round or oval hypoechoic structures between the anatomic landmarks, easily differentiated from blood vessels (echo-free structures that elongate, branch and can present with a hyper-echoic rim representing the vessel's wall). Lymphadenopathy >1.0 cm can cause mass effect on adjacent structures. Figure 1 shows the predefined zones for searching and recording mediastinal lymphadenopathy.
Results: 186 children (58.6% male) with a mean age of 31.7 months (range: 5.2 - 153.5 month) were enrolled in this study. 31 children had mediastinal lymphadenopathy (16.6%), most commonly seen in zone A (45.2%) and D (32.3%), with a mean size of 1.48 cm (95% CI 1.30 – 1.66 cm). Nine children (20.5%) with definite TB, 14 children (13.9%) with possible TB and 7 children (17.1%) with unlikely TB, had lymphadenopathy on mediastinal US, table 1.
Conclusions: One in five children with definite PTB had mediastinal lymphadenopathy on US, this was not significantly different from the unlikely TB category. Further prospective evaluation of the accuracy of mediastinal US, compared to CXR and MRI are needed.