Arjen Lindenholz

190 CHAPTER 9 Interpreting vessel wall thickening Another challenge in vessel wall assessment is the used cutoff value for vessel wall thickening. Previous studies have shown that healthy participants also show vessel wall thickening, and that vessel wall diameter increases somewhat with age. 29- 31 It is currently not clear whether or up to which stage vessel wall thickening still represents normal age-related variation of the intracranial vessel wall, and what can be seen as early indicators of (non-stenotic) atherosclerotic plaques. In general, the vessel wall is considered radiologically thickened when thickness has increased with more than 50% compared with the adjacent vessel wall ( Chapter 3 ). 4,32 However, this is an arbitrary cutoff value chosen for practical purposes and not validated. Most vessel wall pathologies that eventually result in wall thickening, especially intracranial atherosclerosis, have a continuous spectrum of vessel wall thickening that varies depending on disease severity; therefore, this cutoff point should be validated with histopathology and with clinical outcome. An additional question that arises in this regard is how much the vessel wall should be thickened to be reliably seen on MRI. Histopathological studies have shown that intracranial vessel walls of the circle of Willis range in thickness from 0.2 - 0.4 mm; in contrast, radiological measurements have reported vessel wall thicknesses of approximately 1.0 mm for the large intracranial arteries of the circle of Willis. 33,34 This difference may be caused by an underestimation in histopathological studies due to e.g., cell shrinkage, or an overestimation in radiological studies due to partial volume effects, motion artefacts and pulsatility effects, or both. 4,35 The latter is supported by a recent study which showed that with currently used spatial resolutions for vessel wall MRI – with varying voxel sizes between 0.4 - 0.8 mm – vessel wall thicknesses cannot be reliably measured. 35 To perform accurate vessel wall thickness measurements, one needs at least two voxels to prevent partial volume effects, necessitating voxel sizes of at least 0.18 mm isotropic (or smaller) which are currently not feasible within clinically reasonable acquisition times, not even at 7T. 4,36 Notwithstanding these drawbacks, the fact that thickness variations of the vessel wall are visible using in vivo vessel wall sequences means that qualitative assessment of the intracranial vessel wall is possible, by inferring signal intensity and vessel wall thickness throughout the whole vascular tree and comparing these with the adjacent parts of the artery or with the contralateral arteries. Indirect evidence to the nature of vessel wall lesions Since the current relative lack of histopathological validation of intracranial vessel wall lesions precludes strong statements regarding their exact nature, other more indirect ways have been used to further elucidate their possible etiology and clinical relevance, several of which have been described in Part II of this thesis. From these studies, two important findings regarding the nature and possible consequence of vessel wall lesions for clinical practice can be distilled.