Paradoxical parasellar high signals resembling shunt diseases on routine 3D time-of-flight MR angiography of the brain: mechanism for the signals and differential diagnosis from shunt diseases

BACKGROUND AND PURPOSE: We occasionally encounter phenomena in which venous flow signals of the cavernous sinus (CS) and/or inferior petrosal sinus (IPS) are visualized paradoxically in patients without arteriovenous shunt in 3D time-of-flight magnetic resonance angiography (3D-TOF MRA) of the brain. The aims of this study are to examine the frequency and cause of this phenomenon ("pseudo-shunt" image) and to determine points of differentiation from definite arteriovenous shunt images ("real shunt"). METHODS: We retrospectively examined 85 maximum intensity projection images obtained by MRA in the absence of arteriovenous shunts to detect pseudo-shunt images, and evaluated source images of pseudo-shunt studies for venous structures. Four real-shunt MRA studies were compared with pseudo-shunt studies on three points: (1) extension of sinuses, (2) extension of cortical veins, and (3) signal intensity of sinuses as assessed by the scoring method (1 point when these findings exist, 0 when they do not). RESULTS: We detected five CS (3%) and six IPS (4%) signals in 9 (11%) of the 85 cases. In the source images of four pseudo-shunt images in the CS, we detected signals from the sphenoparietal sinus (SPS). The average score was significantly lower in the pseudo-shunt (0.22) than the real-shunt (2.75) images (P < .0001). CONCLUSION: In cerebral 3D-TOF MRA, pseudo-shunt images were seen in 11% (9/85) of the study population, with antegrade upward blood flow of the SPS considered as one of the causes. Real-shunt signals can be distinguished from pseudo-shunt signals by evaluation of source images.