Comparing the Accuracy of Digital Subtraction Angiography, CT Angiography and MR Angiography at Estimating the Volume of Cerebral Aneurysms

While there are many studies that compare imaging modalities in the detection of cerebral aneurysms there are no existing studies that compare two dimensional digital subtraction angiography (DSA), CT angiography (CTA) and MR angiography (MRA) in calculating the volume of cerebral aneurysms. This study will compare these imaging modalities on seven in-vitro models of known volume. Seven silicone models of cerebral aneurysms were chosen representing slight variations in geometric shape and size. The volume of each model was measured by weighing the amount of water required to fill the aneurysm to the parent artery. Contrast enhanced images of the models were taken with DSA, CTA and MRA. The images were interpreted by four independent readers and the volumes were calculated. The measured volumes from the water weight analysis were compared to the volumes calculated from the interpreter's measurements. The accuracy of DSA, CTA and MRA were compared using the Percent of absolute and true variance from the measured volume. The average percent absolute variance for DSA was 14.3%, CTA was 16.8% and MRA was 18.6%. While these differences were minimal, comparing the percent of true variance demonstrated an average variance of -1.9% for DSA, 16.1% for CTA and -15.9% for MRA. Calculating the volume of cerebral aneurysms, while increasingly important, is difficult and error prone. It is important to understand the limitations and inherent errors before relying on calculated volumes in clinical decision-making. Regardless of imaging modality, one should consider error rates of 14-19% for calculating volume while keeping in mind the tendency for CTA to overestimate volume, MRA to underestimate volume and DSA to both under and overestimate equally.