Diagnostic accuracy of computed tomography in differentiating peritoneal tuberculosis from peritoneal carcinomatosis

Document Type

Article

Department

Radiology

Abstract

Introduction: Peritoneal tuberculosis is difficult to diagnose as it may mimic peritoneal carcinomatosis, which has similar symptomatology. We sought to determine the diagnostic accuracy of computed tomography in differentiating peritoneal tuberculosis versus peritoneal carcinomatosis.
Materials and methods: The associations of radiological findings in 124 patients with peritoneal carcinomatosis (n = 55) or tuberculosis (n = 69) were determined using Chi-square test. Sensitivity, specificity, positive and negative predictive value, and total diagnostic accuracy of CT imaging, with histopathology as gold standard, was determined. Subgroup analyses to determine these parameters by age (>40 years and ≤40 years) and gender (male and female) were performed.
Results: Mean age of study population was 44.1 ± 13.2 years with 61 males (49.2%) and 63 females (50.8%). The most common radiological abnormality in both peritoneal carcinomatosis (90.9%) and peritoneal tuberculosis (89.9%) was omental smudging, followed by presence of extraperitoneal mass (81.8%) in carcinomatosis and presence of micro-nodules in tuberculosis (88.4%). The findings significantly different in both the carcinomatosis and tuberculosis groups were high-density ascites, splenic calcification, splenomegaly, lymph node calcifications, micro-nodules, and macro-nodules. The diagnostic accuracy of CT in differentiating peritoneal tuberculosis from peritoneal carcinomatosis was 83.8%; sensitivity and specificity for peritoneal tuberculosis were 88.4% and 78.2%, respectively.
Conclusion: The diagnostic accuracy of CT in differentiating peritoneal tuberculosis from peritoneal carcinomatosis revealed an overall diagnostic accuracy of 83.8%. Subgroup analysis revealed that CT may be a more specific diagnostic tool to predict peritoneal tuberculosis in female patients and in those over 40 years old.

Comments

Issue no. are not provided by the author/publisher

Publication (Name of Journal)

Clinical Imaging

Share

COinS