Maria Babar, Usman Zafar, Irum Shahzad, Arsalan Jamil Raja, Mutahira Mannan, Hannan Zafar

Previous Published Issues



574



ABSTRACT

 

Aim: To determine the diagnostic accuracy of diffusion weighted magnetic resonance imaging for differentiation of Brain abscesses from brain lesions taking surgical findings as gold standard.

Methods: This cross sectional study was done at department of Radiology, Lahore General Hospital, Lahore. The study was completed in Six Months (October 1, 2014 till March 31, 2105). After taking informed consent, demographic information (name, age, gender and contact) was obtained. Then patients underwent through Diffusion weighted imaging (DWI). Diffusion-sensitizing gradients were applied with a diffusion sensitivity of b = 0, 500 and 1000s/mm2. Apparent diffusion coefficient (ADC) values were also calculated on computer software. ADC values were taken from the center and wall of the lesion as well as from the surrounding oedema. DWI were analyzed on the basis of signal intensity on diffusion images and corresponding ADC maps. Then patients were underwent biopsy for histopathology. All biopsy samples were sent to the histopathology laboratory of hospital and results were compared with results of DWI.

Results: The mean age of the patients was 43.74±14.64 years with minimum and maximum ages of 20 & 70 years respectively. Out of 145(65.91%) patients were male and 75(34.09%) patients were females. The male to female ratio was 1.933:1. The sensitivity of DWI in our study was 91.67%, specificity was 93.18%, positive predictive value was 95.28%, negative predictive value was 88.17% and the diagnostic accuracy was 92.27% taking histopathology as gold standard.

Conclusion: Diffusion weighted imaging is an effective method with high sensitivity and specificity which can help in differentiation of Brain abscesses from brain lesions.

Keywords: Diffusion weighted imaging, DWI, Histopathology, Brain Lesions, Restriction, Abscess



Copyright © Pakistan Journal of Medical & Health Sciences 2020. All rights reserved!