Stereotactic radiosurgery (SRS) is an indispensable, definitive treatment option for patients with symptomatic arteriovenous malformation (AVM) and/or dural arteriovenous fistula (DAVF), which are not amenable to direct surgery or endovascular embolization alone [1,2]. Traditionally, the stereotactic co-registration of one or a few representative frames of orthogonally paired 2D digital subtraction angiography (2D-DSA) into a 3D setup, as well as planning images via a dedicated image localizer under rigid frame fixation to the skull, has been a standard prerequisite procedure for target delineation in SRS for intracranial AVM and DVAF. However, the use of frame fixation for a prolonged period and performing follow-up angiography on the day of SRS impose substantial burdens on patients .
Among the 3D imaging techniques for visualizing affected angioarchitectures, cone-beam computed tomography (CT) angiography (CBCTA) acquired by C-arm-based selective 3D-rotational angiography (3DRA) is recognized as the most suitable foundation for target definition in terms of spatial resolution, can visualize tiny lesions and/or fine caliber vessels, and distinguish nidus segmentation from different feeding arteries [1,3-17]. With the advent of image-guidance systems and the application of CBCTA, a frameless SRS workflow has been adopted in an increasing number of institutions implementing CyberKnife, conventional linac, and Gamma Knife, where treatment planning is usually completed by the day before SRS, thus allowing for the omission of angiography on the day of SRS [3-5,11,13-16,18]. However, in this frameless workflow, 2D-DSA cannot be directly integrated into a 3D planning image such as CBCTA and/or non-contrast planning CT and is only available as a reference image. Although target outlining based on CBCTA alone may be adequate for cases with a relatively simple lesion configuration and relevant angioarchitecture [3,5,7,8,11], the relatively low temporal resolution of CBCTA likely becomes problematic in cases with a complex vasculature, such as an irregularly shaped and/or diffuse nidus and rare localization. Biplane 2D-DSA, including oblique angulation with the adequate temporal resolution, provides dynamic, comprehensive, and panoramic views of the affected angioarchitecture, whereas paired 2D-DSA alone is inadequate for 3D target outlining owing to its low spatial resolution [5-7].
To redeem the low temporal resolution of CBCTA and to expand the applicability of frameless radiosurgery, the frameless integration of 2D-DSA (whole frames, if possible) into CBCTA is required [5,10]. Thus far, the frameless integration of 2D-DSA into CT and/or 3D vascular images has been attempted using various approaches, such as the implantation of fiducial markers or originally developed software prototypes [13,14]. Recently, an upgraded version of dedicated planning-subsidiary software Brainlab® Elements (Brainlab AG, Munich, Germany) has provided the frameless co-registration functionality of 3D vascular images and orthogonally paired 2D-DSA . To the best of our knowledge, Brainlab® Elements has so far been the only commercially available software enabling frameless 6 degree-of-freedom (6DoF) co-registration of 2D-DSA whole frames into any stereotactic 3D vasculature-visible image. Thus, we examined the accuracy and practicality of this software by using actual clinical images in a preclinical feasibility study to verify whether the frameless co-registration of 2D-DSA and CBCTA is feasible in a clinical setting.
The synopsis of this study was presented at the 13th annual meeting of the Japan Radiosurgery Society held online on February 5, 2022.