Stenoses of the cervical vessels - computed tomographic angiography (CTA) versus digital subtraction angiography (DSA)

The stroke ranks among the most common lethal diseases in Germany. Approximately 20% of the insults are caused by occlusion or embolic effective material of plaque ruptures of constricted cervical vessels. An early diagnosis can prevent serious consequences. Therefore procedures without complications and a high diagnostic reliability are required. For many years the digital subtraction angiography (DSA) has been established as gold standard, but due to invasivity and risks like vessel dissection, post-operative hemorrhage or thrombo-embolism it is unsuitable as a routine examination. The computed tomographed angiography (CTA) is a non-invasive alternative. The aim of this study was to determining the validity of CTA-based stenoses measurements. To that end the CTA and corresponding DSA- datasets of 154 cervical vessels have been compared. The data of 66 patients of the UMG, who underwent CTA as well as DSA for further diagnosis or treatment, have been examined by two blinded independent investigators retrospectively. The arithmetic average of those two measurements was used for further statistic interpretation. The CTA diagnostic was performed using a 128-row CTA biplanar flat panel detector to gain the DSA data. Multi-planar reconstruction projections (MPR) were used to measure the stenosis of the Aa. carotides and Aa. vertebrales based on the method of the North American Symptomatic Carotid Endarterectomy Trial (NASCET), European Carotid Surgery Trial (ECST), and in the case of the Aa. carotides the Common Carotid (CC), and correlated to the gold standard NASCET-DSA. Additionally semiautomatic software based on maximum intensity projections (MIP) was used. Furthermore the influence of hard (B46) and soft (B26) kernels to stenoses measurements at different arteriosclerosis was examined. Through four fold table sensitivity and specificity were calculated, for this purpose the stenoses were categorized in occlusion (100%), high-grade (>70%50<71%) and low-grade (<50%). For inter-rater reliability testing Cohens κ was determined. Manual measuring of the lumina revealed high (r=0,76-0,98) correlation between CTA and DSA. The best correlation was shown for NASCET using a hard kernel (rA.vertebralis B46NASCET= 0,9775; rA.carotis B46NASCET=0,9510). The semiautomatic measuring presented correlations between r=0,61 for A.vertebralis B26 and r = 0,88 for A. carotis B46. For all modalities there had been higher correlations if hard kernels were used. The inter-rater variability resulted in a κ of >0,6. The calculated sensitivity and specificity were 100% for occlusions, for high grade stenoses 84% respectively 96%, for moderate stenoses 81% respectively 86% and for low-grade stenoses 83% respectively 95%. Under consideration of the regression-line corresponds a 70% NASCET stenosis with a 78,9& CC-stenosis, respectively a 81,3% ECST stenosis, which is in accord with the literature. The determined high correlation between CTA and DSA, and also the high sensitivity and specificity underline the results of previous studies on this subject. The results of the current study verify that the CTA is an appropriate device to diagnose vertebralis stenoses. Hard kernels show higher correlations with the DSA results. The differences between hard and soft kernels are obvious especially for semiautomatic stenosis measuring. Showing a κ >0,6 inter-rater variability is to be classified as low. According to this the 128-row CTA is an appropriate non-invasive device for reliable diagnoses of cervical vessel stenosis with high correlation to the DSA and excellent sensitivity and specificity for high grade stenoses.2013-09-1