The utility of abdominal ultrasound in the diagnosis of paediatric abdominal tuberculosis: a single centre review

Master Thesis

2020

Permanent link to this Item
Authors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
License
Series
Abstract
Background: Childhood tuberculosis (TB) is a common disease worldwide, with an increased propensity for severe, disseminated disease in settings with a high burden of concomitant HIV infection. Ultrasound is commonly used in diagnosing abdominal TB, however the indications for its use are unclear and often vary amongst clinicians. Objective: In this study, we describe the findings of ultrasound examinations performed for suspected abdominal TB at a tertiary children's hospital and examine the variability in reporting patterns amongst radiologists performing these imaging investigations. Materials and methods: Ultrasound studies performed for “suspected abdominal TB” between 01 January 2013 – 31 December 2018 were reviewed. In studies reported as suggestive of abdominal TB, evidence of microbiologically confirmed disease was sought. Subsequently, a selection of images from these studies were independently reviewed by three paediatric radiologists to determine their level of agreement when interpreting imaging findings. Results: During the study period 1093 studies were performed for suspected abdominal TB, of which 166 (15%) had abnormal features suggestive of TB. Forty-seven percent of these patients (78/166) had microbiologically confirmed disease. The commonest reported features were lymphadenopathy, 77% (128/166) and splenic microabscesses, 55% (92/166) for which substantial inter-reader agreement was documented, Fleiss' kappa = 0.64 and 0.66 respectively. There was moderate inter-reader agreement in the diagnosis of abdominal TB among radiologists (Fleiss' kappa=0.47). Conclusion: Caution is advised when basing clinical decisions on ultrasound studies performed for suspected abdominal TB, as imaging features are non-specific and there is considerable variability in interpretation of studies among reporting radiologists.
Description

Reference:

Collections