THE ROLE OF MRA AND CTA IN THE PREOPERATIVE IDENTIFICATION OF THE ARTERY OF ADAMKIEWICZ IN PATIENTS UNDERGOING THORACO-ABDOMINAL AORTIC ANEURYSM REPAIR

Hinojosa CA¹, Carvalho-Perron LB¹, Garrido-Olivares L¹, Franchetto AA2, Ellins ML², Grynspan J², Jichici D³, Cinà CS1,^4
1 Division of Vascular Surgery, ² Department of Radiology,³ Department of Critical Care Medicine,
⁴ Department of Epidemiology and Biostatistics McMaster University

Background: Traditional open and endovascular repair of thoracic and thoracoabdominal aortic aneurysms are associated with a risk of lower limb neurologic deficits.  Surgical and radiological reports underscore the importance of preserving the intercostal artery from which the artery of Adamkiewicz (AKA) supplying the spinal cord originates. Ideally this artery should be identified before surgery.

Purpose:  To report the results of using Magnetic Resonance Angiography (MRA) and Computed Tomography Angiography (CTA) to localize the AKA, in patients with thoraco-abdominal aortic aneurysm (TAAA) candidates for surgical treatment.

Methods: This is a prospective cohort of elective patients operated upon for TAAA who underwent CTA and MRA preoperatively to identify the AKA. The studies were performed using a 16 or 64 slice multidetector CT, and a 1.5T MRI, respectively. Expert radiologists interpreted the results of the studies. Exclusion criteria for MRA were: presence of foreign bodies non MRA-compatible, claustrophobia or logistical reasons.  CTA was not done for serum Cr level greater than 250mmoL/L.  Outcomes of the imaging studies were defined as follows: localization of the AKA, defined as identification of the vessel moving up to the anterior midsagittal surface of the spinal cord from the radicular-medullary artery originating from the dorsal branch of the intercostal or lumbar artery;  identification of the AKA, defined as identification of the artery, but not of the level of the intercostal or lumbar artery from which it originated; and non identification of the AKA, when the artery was not identified or localized. Inter-test reliability (i.e., the extent to which the tests revealed the same findings) was assessed using generalizability theory.

Results: From 2002 to 2006, 16 patients underwent studies to identify the AKA; age 57 (SD 13.5) and 11 69% males.  A CTA was done in 11 (68%), an MRA in 13 (81%) and both tests in 8 (50%) patients, respectively.  CTA alone was able to localize the artery in 9 patients (82%) and in 2 (18%) the artery was not identified. MRA alone localized the artery in 9 patients (69%), identify it in 1 (8%), and was not diagnostic in 3 (23%).  By CTA and MRA, the AKA was always identified from T10 to L1. In the CTA the frequency at which each artery was localized is one artery at T10, three at T11, two at T12 and three at L1, for the MRA the frequency is equally distributed in three arteries by each level from T11 to L1.
When the group of patients undergoing CTA and MRA were analyzed, agreement between the two tests was moderate (G = 0.55).

Conclusions: CTA and MRA are noninvasive procedures capable of localizing the AKA.  In our study the CTA appear to be superior to MRA, but this may be different if a 3D time-of-flight MRA is used. The test levels assigned by each tool in identifying the AKA were in moderate agreement.