Accuracy assessment of computerized tomography-angiography for 3D measurement of intracranial aneurysm geometry

The purpose of this study was to develop a valid, reliable, and accurate system of measurement of intracranial aneurysm geometry using volumetric data obtained by computerized tomography-angiography (CTA). A simple model of lateral saccular aneurysms was created using a rubber balloon inserted in a plastic straw with a small lateral hole. Diluted contrast material was injected allowing the balloon to protrude through the lateral hole simulating an aneurysm. Volumetric CT data of models with side holes of various size and configuration were acquired using a 1-mm beam width and 1-mm/second couch movement at settings similar to clinical CTA. Axial source images of 1 mm thickness were generated every 0.5 mm for subsequent two- and three-dimensional (3-D) reconstruction. The aneurysm sac and its neck were evaluated using a volume-rendered image based method and multiplanar reconstruction (MPR)-based methods. Evaluation included measurements of longitudinal and transverse diameter of an aneurysm sac and its neck in relation to the parent vessel as well as their morphology. These measurements were compared with actual caliper measurements of the model. Clinical cases of lateral aneurysms arising from the supraclinoid internal carotid artery (three patients), and the middle cerebral artery (two patients) were also studied. Both the volume-rendered image-based and MPR-based methods provided accurate assessment of size and shape of an aneurysm sac and its neck. Variability of measurements by different methods was within 10%. Complex neck morphology of the clinical cases was also clearly delineated in both methods. The volume-rendering technique provided superior 3-D images but required editing and thresholding by an operator and thus was subjective. The MPR-based method was more objective and better suited for quantitative analysis. Using these complementary tools, morphological and geometrical information of an intracranial aneurysm, which is critical for optimum neurosurgical and endovascular treatment, can be derived from the volumetric data generated bv CTA.