Computed tomography angiography compared to digital subtraction angiography for
finding and grading mesenteric artery stenoses in patients suspected of having chronic
mesenteric ischemia: a retrospective comparative study
M.P. de Wringer1, L.M.G. Moons1, J.N.L. Schouten1, A. Moelker2, E.J. Kuipers1, M.
Ouhlous2
1
Dept. of Gastroenterology and Hepatology, Erasmus MC, Rotterdam, The Netherlands
2
Dept. of Radiology, Erasmus MC, Rotterdam, The Netherlands
Computed tomography angiography (CTA) is often used in the diagnostic work-up of patients
suspected of having chronic mesenteric ischemia (CMI). It is presumed that CTA has a good
sensitivity and specificity for detecting mesenteric artery stenoses. However, to our
knowledge, CTA has never been properly compared to the gold standard, digital subtraction
angiography (DSA). Since the indication for treatment depends on the presence of a stenosis
detected by CTA, the diagnostic value of CTA was studied.
In this retrospective study, all patients analyzed for CMI with a CTA in the period of 2006-
2011 and a DSA within 6 months after CTA were included from a hospital database. Stenoses
of both the celiac artery (CA) and the superior mesenteric artery (SMA) were graded as:
70%. Stenoses on CTA were graded by the radiologist, and agreed upon
in a consensus meeting with radiologists, vascular surgeons and gastroenterologists. Stenoses
on DSA were assessed by the interventional radiologist. Stenosis etiology was determined by
the radiologist. CTA was thought to best show atherosclerotic lesions, and DSA to show
respiratory dependent stenoses: CA compression syndrome (CACS).
In total, 165 patients (mean age: 64, std. dev.: 15 yrs; 72% female) were included. Data was
available on CA stenosis of 109 pts, and of 88 pts for the SMA. Etiology of the stenosis was
atherosclerosis in 82%, CACS in 11%, and of other origin in 7%. CTA categorized 75 out of
109 CA stenoses correctly (agreement 69%; kappa 0,445; p=0,067). Underestimation
occurred in 19/109 (17%) cases, and overestimation in 15/109 (14%) cases. Using a threshold
of ≥50% for a significant CA stenosis, agreement was 82% (kappa 0,499; p= 0,096) with a
sens. of 92%, spec. 53%, positive predictive value (PPV) 84%, and NPV of 73%. Using a
threshold of >70%, agreement was 80% (kappa 0,581; p=0,078) with a sens. of 78%, spec.
84%, PPV 90% and NPV of 67%.
For SMA stenoses, CTA categorized 64 out of 88 SMA stenoses correctly (agreement 73%;
kappa 0,547; p=0,071) Underestimation occurred in 10/88 (11%) cases, overestimation
occurred in 14/88 (16%) cases. Using a ≥50% threshold for a significant SMA stenosis,
agreement was 83% (kappa 0,657; p = 0,079) with a sens. of 93%, spec. 72%, PPV 78% and
NPV of 91%. Using the 70% threshold, agreement was 81% (kappa 0,610; p=0.085) with a
sens. of 76%, spec. 85%, PPV 82%, and NPV of 80%.
Conclusion: CTA is a fairly accurate technique to identify mesenteric artery stenosis, with a
better test performance for correctly grading SMA than CA stenoses.