Does Size, Shape, or Location Limit the Central Halo and the Polar Phase Signals of Susceptibility-Weighted Imaging in Differentiating Intracranial Hemorrhages from Calcifications?

Abstract

Background and purpose: Evaluation of polar phase signals on susceptibility-weighted imaging (SWI) has shown success in differentiating hemorrhage from calcification, particularly in subcentimeter spherical foci located in the brain. However, aliasing of phase signals near the center of larger susceptibility lesions presents a challenge in accurately classifying lesions with signal dropout and blooming on SWI. We have investigated the use of central halo, in addition to the polar signals, to broaden the use of SWI phase images in classifying lesions with a wider range of locations, sizes, and shapes.

Materials and methods: This retrospective study included 50 consecutive patients who underwent MRI with SWI of the brain. Phase signals from the two polar regions and the central halo were evaluated. Susceptibility foci of all sizes, shapes, and locations were included, except for the basal ganglia calcifications. CT images were used as the gold standard for differentiating hemorrhages from calcifications. Appropriate statistical analyses were performed.

Results: The study cohort included 22 males and 28 females aged 2-90 years (mean age: 61.19 ± 21.13 years). SWI identified 406 hemorrhages: 305 intraparenchymal, 45 subdural, 22 subarachnoid, 15 intraventricular, and 19 cortical vein thromboses. There were 202 calcifications observed on SWI: 24 intraparenchymal, 41 pineal, 83 choroid plexus, 18 dural, and 36 arachnoid granulations. Hemorrhage sizes ranged from 1.5 mm to 145.2 mm (mean: 11.5 ± 15.81 mm), while calcifications ranged from 1.5 mm to 71.9 mm (mean: 8.16 ± 7.13 mm). Hemorrhagic lesions were round (300), linear (75), or irregular (31), while calcifications were round (139), linear (95), or irregular (1). Sensitivity and specificity for hemorrhages were 99.5% (95% CI: 98.23-99.4) and 100% (95% CI: 98.06- 100), respectively. For calcifications, sensitivity was 84.26% (95% CI: 78.96-88.67) and specificity was 95.42% (95% CI: 90.30-98.30). The area under the curve (AUC) was ≥0.97 for all three phase sectors in hemorrhages and ≥0.93 for the caudal and halo regions in calcifications.

Conclusions: Phase signals of SWI, analyzed across both poles and the central halo, can successfully distinguish most intracranial hemorrhages and calcifications, regardless of their size, shape, or location.

Abbreviations: AUC = Area Under the Curve; COVT = Cortical Vein Thrombosis; SSS = Superior Sagittal Sinus; AG = Arachnoid Granulation; SDH = Subdural Hemorrhage; IVH = Intraventricular Hemorrhage QSM = Quantitative Susceptibility Mapping; PPV = Positive Predictive Value; NPV = Negative Predictive Value.