With great interest, we read the article by Uotani and colleagues 1 in the February issue of the American Journal of Neuroradiology. The authors compared intravenous and intra-arterial contrast agent application for the localization of the Adamkiewicz artery (AKA) in different patients scheduled for thoracic or thoracoabdominal aortic aneurysm repair with use of CT angiography (CTA). Intra-arterial injection seemed to perform much better than intravenous injection regarding AKA localization (94% vs 60%, respectively). The focus of this study was visualization of the AKA and its direct segmental supplier. The continuity between the AKA and the segmental artery could not be visualized in all cases, possibly because of the proximity of vertebral bony structures with similar signal intensity on CT. However, occlusion of the segmental artery might also be a possible explanation. Two thirds of the patient population had an atherosclerotic aortic aneurysm, but an indirect supply to the aorta was not visualized in any of the patients. This finding is interesting because, in other studies, it has been found that the segmental artery directly connecting to the AKA was occluded in up to 40% of patients.2 In these cases, the AKA supply was provided through an open segmental artery originating 1 or more levels above or below the occluded segmental artery. Apart from the segmental supply to the AKA, there are alternative routes by which the anterior spinal axis might be supplied. From our own experience 2 and that of others, 3 we know that spinal cord function is not always crucially dependent on segmental supply directly connecting to the AKA. Collaterals can substitute the spinal cord supply to maintain spinal cord function. 3 Recently, these collateral pathways have been visualized with MR angiography (MRA) 2 (Fig 1) and correlated with intraoperative spinal cord motor-evoked potentials. It is interesting to note in this study that when the segmental supply to the AKA was inside the cross-clamped aortic region and connecting collaterals originating from outside the cross-clamped aortic area were noticed, this had a predictive value of 97% for stable spinal cord function. This finding can be of crucial importance for the surgeon during the preoperative and intraoperative strategic planning. Therefore, we believe that preoperative imaging in thoracoabdominal aortic aneurysm surgery should not only locate the AKA with its segmental supply but should also locate the possible collateral supply. An important aspect that is mentioned by Uotani et al 1 is that a multiphase approach is necessary to convincingly separate the AKA from the similar spatially shaped outlet vein. With the use of CTA, this would mean that the patient is exposed twice to a high effective dose (we calculated 20 mSv for total thoracoabdominal aortic coverage). A smaller vertebral coverage as used in the protocol of Uotani 1 (coverage T7-L3) can be used to lower the effective dose. However, in our opinion, this is not an option because of the possibility of m...