There has been increased attention paid to intracranial vascular calcifications, with evidence that there is associated increased association with prior and future cerebrovascular ischemic events and development of dementia(1-4). Thin slice CT imaging is the reference standard for detection of calcification, however considering increasing concerns over radiation exposure, the value of MR imaging for this evaluation can be seen. Typically when MR is employed for the evaluation of calcifications, time-of-flight (TOF) MRA is employed. Simultaneous Non-contrast Angiography with intraPlaque hemorrhage (SNAP)(5) is a slab selective phase sensitive inversion recovery technique that produces a bright blood MRA and heavily T1 weighted inverted image for detection of intraplaque hemorrhage. Notably, SNAP sequence contains a proton density (PD) weighted reference acquisition (Ref) without tissue suppression for phase sensitive reconstruction, which is usually discarded. This Ref image generated from SNAP provides an opportunity for the evaluation of atherosclerotic plaque outer wall boundary and wall calcifications. This study sought to evaluate the ability of SNAP Ref images to detect intracranial artery wall calcifications as compared 3D time of flight (TOF) MRA relative to the reference standard, thin slice CTA.Imaging Protocol and Analysis After IRB approval, the radiology database was reviewed for consecutive subjects with thin-slice (.625 mm) CT or CTA, TOF MRA and SNAP imaging of the brain scanned on 3 Tesla Philips Ingenia MRI system (Philips Healthcare, Best, the Netherlands). A new SNAP Ref image (Ref2) was generated by weighted addition of SNAP MRA and SNAP Ref. The parameters of each MR sequence are shown in Table 1. A blinded review was performed by a board certified neuroradiologist, with consecutive Ref2 sequences reviewed in random order, followed by randomized review of 3D TOF MRA and finally thin slice CTA images in consecutive days. The intracranial arterial segments that were evaluated individually include: cavernous, ophthalmic, supraclinoid and terminal carotid artery segments, M1 middle cerebral, A1 anterior cerebral, P1 posterior cerebral arterial segments on the right and left and the basilar artery. Statistical methods Sensitivity and specificity for detecting calcification per vessel was computed for SNAP and TOF-MRA using CT as the reference standard. Overall agreement with CT was assessed using unweighted Cohen’s kappa and linearly weighted Cohen’s kappa for both SNAP and TOF-MRA. Agreement was assessed for presence/absence of calcification per vessel and calcification size category per vessel (none, <50% circumferential involvement and >50% circumferential involvement) based on previously established evaluation(6). Diagnostic performance and agreement metrics were compared between SNAP and TOF-MRA using the non-parametric bootstrap, to account for potential dependence between vessels from the same subject.11 subjects were imaged. 143 segments were reviewed on each modality (basilar artery and bilateral cavernous carotid, supraclinoid carotid, carotid terminus, A1 and M1 segments). 14 segments were not evaluable on all modalities (5 basilar arteries, 4 cavernous segments, 4 opthalmic segments and 1 supraclinoid segment), leaving 129 segments available for analysis. Of the 11 subjects, 7 had calcification identified in at least one intracranial segment on CT. Of the 129 vessels evaluated, 19% had calcification by CT, 22% by SNAP and 13% by TOF-MRA (Table 2). Using CT as the reference standard, SNAP had higher sensitivity (75.0% vs. 29.2%, p=0.01) and similar specificity (89.5% vs. 90.5%, p=0.8) compared to TOF-MRA (Table 3). SNAP also had higher overall agreement with CT for calcification presence/absence (kappa: 0.60 vs. 0.22, p=0.01) and calcification size categories (weighted kappa: 0.61 vs. 0.20, p=0.008) than TOF-MRA.In this study, the feasibility of utilization of Ref2 images for the evaluation of intracranial calcifications is shown. In comparison to the typically used 3D TOF MRA, Ref2 more accurately depicted intracranial arterial calcifications. When combined with the MRA and intraplaque hemorrhage images generated by SNAP, this technique can potentially provide first line luminal and vessel wall imaging information.