Clinical Radiology and Imaging Journal (CRIJ)

ISSN: 2640-2343

Research Article

A MIP-based 3D Contour Technique for Radiosurgery of Arteriovenous Malformation

Authors: Zhou Y, Zheng Y, Stuart B, Zhao W* and Wu X*

DOI: 10.23880/crij-16000189

Abstract

Purpose: Traditional, orthogonal digital subtraction angiography (DSA) has been the gold standard in the diagnosis of Arteriovenous Malformation (AVM) and in nidus delineation for its treatment using stereotactic radiosurgery (SRS), notably with the GammaKnife. The progressively advanced Time-evolved, CTA, MRA and rotational angiography that are capable of accounting for haemodynamic features of AVM have made them attractive alternatives. However, the maximum intensity projection (MIP) that is a frequent tool in AVM analysis has not been easily integrated with most treatment planning systems. Herein, a novel technique is presented that allows conversion of the MIP images of the nidus to axial images. As such, contours of the AVM nidus can be established more accurately for SRS treatment when DSA is not available. Methods: In the proposed method, the AVM nidus delineated on the MIPs from multiple projection angels are back-projected to obtain a “reconstructed” 3D AVM nidus volume. Adjusting the threshold accordingly, the nidus border can then be reconstructed and transcribed onto the axial images. Results: The method was tested on a digital phantom with an artificial AVM volume, as well as on a patient MRA data set containing a real AVM. The reconstructed AVM contours, when compared to that of direct delineation from axial images, show satisfactory agreement in both phantom and in a clinical example. Conclusions: Using MIP from 4D MRA or CTA to facilitate AVM nidus delineations demonstrates both technical feasibility and potential clinical advantages for SRS of AVM.

Keywords: Delineation; Radiosurgery; AVM; Back-Projection; MRA; MIP

View PDF

Google_Scholar_logo Academic Research index asi ISI_logo logo_wcmasthead_en scilitLogo_white F1 search-result-logo-horizontal-TEST cas_color europub infobase logo_world_of_journals_no_margin