gms | German Medical Science

Objective: Follow-up imaging of treated and of untreated cerebral aneurysms is currently being performed using MRA, standard 2D DSA, rarely 3D rotation DSA, or CTA. It remains difficult to compare these different imaging modalities. Therefore, we attempted to adopt a semi-automated image fusion algorithm for various volumetric data sets, to investigate 1) feasibility of image reconstruction and co-registration, 2) current 2D based analysis with 3D based analysis, and 3) typical growth / revascularization patterns of treated aneurysms (Ax).

Method: Informed consent was obtained from all patients. Prospective acquisition of image data sets was instituted. Out of 225 Ax that had repeated vascular imaging using volume data, 30 were available that had multiple consecutive imaging data sets (median 4, range 3-9) with a median observation time of 35 months (range 12-120 months) Distribution of aneurysms were ACA/Acomm:18, MCA:7, Pcomm,PCA/basilar:11. The ratio of clipping to coiling was 5:25. All image data sets were imported into Amira visualization software. Semi-automated segmentation, rendering, and rigid co-registration were performed using in-house developed algorithms.

Results: Initial pre-treatment aneurysm volume was in the reperfusion group 489 mm³ vs. in the completely obliterated aneurysm group 243mm³. Neither carrier vessel (2.9 vs 2.8mm) nor aneurysm neck diameter (4.67 vs 4.60mm) differed significantly between the reperfusion / obliterated groups. Several reperfusion patterns were observed: 11 reperfusions at the base or neck, 4 diffuse reperfusions in the aneurysm due to coil compaction, and 2 reperfusions detected on the dome of the aneurysms. Five Ax had a spontaneous regression of the reperfused volume over time. Nine reperfused Ax were stable during the observation time, whereas two Ax showed marked growth.

Conclusions: We were able to co-register and analyze different vascular imaging modalities with good overlap. Future analysis would benefit from elastic co-registration. 3D based image comparison and analysis provided a new quality, in particular with respect to the shape of revascularization. Revascularization was most often seen at the distal base/neck entrance of the carrying vessel. A correlation between initial Ax volume and rate of reperfusion was seen. However, interpretation of our data is limited due to low case number and a predominantly imaging follow-up for endovascular treated rather than clipped patients.