Object
Cerebrovascular lesions can have complicated abnormal anatomy that is not completely characterized by CT or MR angiography. Although 3D rotational angiography provides superior spatial and temporal resolution, catheter angiograms are not easily registered to the patient, limiting the use of these images as a source for neuronavigation. However, 3D digital subtraction angiography (DSA) contains not only vascular anatomy but also facial surface anatomy data. The authors report a novel technique to register 3D DSA images by using only the surface anatomy contained within the data set without having to fuse the DSA image set to other imaging modalities or use fiducial markers.
Methods
A cadaver model was first created to assess the accuracy of neuronavigation based on 3D DSA images registered by facial surface anatomy. A 3D DSA scan was obtained of a formalin-fixed cadaver head, with acquisitions of mask and contrast runs. The right common carotid artery was injected prior to the contrast run with a 45% contrast solution diluted with water-soluble red liquid latex. One week later, the head was registered to a neuronavigation system loaded with the 3D DSA images acquired earlier using facial surface anatomy. A right pterional craniotomy was performed and 10 different vascular landmarks were identified and measured for accuracy using the neuronavigation system. Neuronavigation based only on 3D DSA was then used to guide an open clipping procedure for a patient who presented with a ruptured distal lenticulostriate aneurysm.
Results
The accuracy of the measurements for the cadaver model was 0.71 ± 0.25 mm (mean ± SE), which is superior to the 1.8–5 mm reported for neuronavigation. The 3D DSA–based navigation-assisted surgery for the distal lenticulostriate aneurysm aided in localization, resulting in a small craniotomy and minimal brain dissection.
Conclusions
This is the first example of frameless neuronavigation based on 3D catheter angiography registered by only the surface anatomy data contained within the 3D DSA image set. This is an easily applied technique that is beneficial for accurately locating vascular pathological entities and reducing the dissection burden of vascular lesions.
Hepatocellular Carcinoma Segmentation from Digital Subtraction Angiography Videos Using Learnable Temporal Difference