DESCRIPTION (provided by applicant): Intracranial arteriovenous lesions (AVLs) - i.e. dural arteriovenous fistulas (dAVFs) and arteriovenous malformations (AVMs) - are an important, treatable cause of long-term neurological disability and death. They are characterized by direct passage of blood from arteries to veins (shunting). dAVFs (Fig. A) are acquired arteriovenous shunts within the wall of a dural venous sinus, supplied by dural arteries, whereas AVMs (Fig. B) are congenital lesions, characterized by a "nidus" (latin: 'nest') of malformed vessels in which
arteries are connected directly to veins, without an intervening capillary bed. Once an AVL is diagnosed, characterization is necessary for treatment planning and risk stratification; delineation of venous drainage in particular is important for prognostication. Due to the inaccuracy and lack of spatio-temporal resolution of non-invasive methods, the current standardof- care test for the diagnosis and characterization of AVLs is catheter-based digital subtraction angiography (DSA). Due to the non-specific clinical presentation of AVLs, which includes intracranial hemorrhage (ICH - the most dreaded complication), seizures and objective pulsatile tinnitus, many patients without an AVL end up undergoing DSA. This invasive procedure, which is usually performed as a stand-alone diagnostic procedure prior to any endovascular therapy, is costly (approx. $75,000), requires highly specialized physicians and infrastructure, and carries a risk to the patient of complications. A non-invasive method that can sensitively detect AVLs would allow DSA to be avoided in the large number of patients with a suspicious clinical presentation in whom the diagnosis can be confidently excluded. Accurate noninvasive delineation of venous drainage would also allow treatment planning to be performed without the need for a preliminary, diagnostic DSA. In addition to avoiding the risks of an invasive procedure and ionizing radiation, an MR-based method would also be more accessible, decrease demand on neuro-interventional services, and translate to cost reduction as it is much cheaper (approx. $5,000) than a DSA. Preliminary results indicate that arterial spin label (ASL) MRI demonstrates AVLs with high conspicuity, due to passage of labeled blood into the draining veins in the setting of arteriovenous shunting. However, current ASL methods do not allow dynamic imaging of the passage of labeled blood through the cerebral vasculature, which is necessary for AVL characterization. The goal of this project is therefore: to develop robust time-resolved pseudo-continuous ASL (pCASL)-based angiography and perfusion MR methods; and validate the diagnostic accuracy and sensitivity of these methods, when added to a routine MR protocol, for the detection and venous drainage characterization of AVLs. In Aim 1 we will develop a new pCASL-based MR angio sequence with a segmented cine readout using 3D cones trajectories and compressed sensing; the sequence will be expanded to also perform perfusion imaging. Aim 2 is to assess the improvement in diagnostic sensitivity and specificity for the detection of AVLs when pCASL is added to standard MR protocols. Aim 3 is to assess the accuracy of pCASL methods for delineation of the venous drainage pattern. The study end-point is a new MR protocol augmented by pCASL methods together with scan parameter recommendations based on comparative assessment of the diagnostic yield of pCASL methods. In summary, we aim to create a new, clinically feasible, time-resolved pCASL MR angiographic and perfusion imaging method with high spatio-temporal resolution. We seek to establish whether this technique is comparable to DSA in terms of: sensitivity for the detection, and reliability for the delineation of venous drainage. If so, this would obviate the need for purely diagnostic DSA prior to treatment, a radical and disruptive change in the workup of AVL patients, with high impact on cost savings, dose reduction and quicker access to diagnosis and (if needed) intervention/surgery.