Computed tomography poorly detects gastroesophageal varices compared with endoscopy: a real-world single-center...

Screening esophagogastroduodenoscopy (EGD) is recommended for all cirrhotic patients at the time of diagnosis, however, non-invasive screening practices for varices have been emerging. CT scans have demonstrated high sensitivity in the detection of esophageal varices. The purpose of this study was to evaluate the sensitivity of CT scans in detecting esophageal and gastric varices when compared to the gold standard of EGD.

After institutional board review was obtained, patients from any University of Florida- Jacksonville location with esophageal varices, gastric varices, or both seen on CT abdomen and pelvis, CT chest, CT chest and abdomen pelvis, or EGD were included (n = 470). Exclusion criteria comprised cases where medical record numbers were invalid (n = 239), if esophageal or gastric varices were not reported (n = 25), varices other than esophageal or gastric were present (n = 5) and if CT scans were performed after an EGD. 201 total patients were included in the analysis.

Compared to EGD-verified varices, CT scans demonstrated a sensitivity of 38.3% (95% CI [0.312-0.457]), with scans performed within 7 days of each other improving sensitivity to 39.51% (95% CI [0.294-0.501]), and 40.1% (95% CI [0.327-0.478]), in contrasted studies. CT scans correctly detected varices in 45.90% of thrombocytopenic patients (95% CI [36.85%- 55.16%]).

Although this study's sensitivity differs markedly from that reported in prior literature, the findings can be understood in the context of methodological limitations. More than the absolute percentages, the observed trends highlight which patient groups may derive the greatest benefit from CT scan surveillance of varices. Overall, routine CT scans may not be surrogates for dedicated radiographic variceal screening. By presenting real-world data on variceal detection rates with CT, this study underscores the potential role of routine imaging not only in risk stratification but also in guiding resource allocation, such as identifying patients who may safely defer endoscopic evaluation versus those who warrant earlier interventions.

Retrospectively registered.