Multidetector computed tomography versus platelet/spleen diameter ratio as methods for the detection of gastroesophageal varices

Authors Andreas Karatzas, Christos Triantos, Maria Kalafateli, Misiel Marzigie, Chryssoula Labropoulou-Karatza, Konstantinos Thomopoulos, Theodoros Petsas, Christina Kalogeropoulou.

Abstract

Background All patients with liver cirrhosis should undergo screening endoscopy, but there are limitations and this approach places a heavy burden upon endoscopy units. The aim of this study was to compare multidetector computed tomography (MDCT) and the platelet/spleen diameter ratio as non-invasive methods for the detection of gastroesophageal varices.

Methods The study included 38 cirrhotics who underwent upper gastrointestinal (GI) endoscopy and MDCT within one month. Two radiologists reviewed the scans, in order to determine the presence and the size of varices. Blood tests and measurement of the spleen maximum diameter were also carried out and the platelet/spleen diameter ratio was calculated. Endoscopy was considered the gold standard and the results of the two methods were compared to it.

Results Varices were detected by upper GI endoscopy in 24 of 38 patients. The mean sensitivity and specificity of MDCT for the two observers was 86.1% and 57.1% respectively. In patients with large varices (>5 mm), the sensitivity was 100% (4/4). Using 909 as a cut-off value of the platelet/spleen diameter ratio this method yielded a sensitivity of 56.5% and a specifi city of 35.7%. The difference in sensitivity and specifi city between the two methods was statistically signifi cant P<0.05.

Conclusion MDCT was accurate for the detection of gastroesophageal varices, especially those with clinically significant size (>5 mm), and superior to platelet/spleen diameter ratio. MDCT could replace, in selected patients, upper GI endoscopy as a method for detecting gastroesophageal varices in cirrhotic patients.

Keywords Varices, multidetector computed tomography, platelet/spleen ratio, endoscopy

Ann Gastroenterol 2016; 29 (1): 71-78

Published
2016-01-07
Section
Original Articles