Program Information
Classification of Intracranial Calcific and Hemorrhagic Lesions Using Quantitative Susceptibility Mapping: Preliminary Results of a Human Trial
MC Jacobsen1*, K Hwang1 , LG Le Roux2 , C Kale2 , D Cody1 , D Schellingerhout2,3 , (1) UT MD Anderson Cancer Center, Department of Imaging Physics, Houston, Texas (2) The University of Texas MD Anderson Cancer Center, Department of Cancer Systems Imaging, Houston, Texas, (3) University of Texas MD Anderson Cancer Center, Department of Diagnostic Imaging, Neuroradiology Section, Houston, Texas
Presentations
WE-G-201-5 (Wednesday, August 2, 2017) 4:30 PM - 6:00 PM Room: 201
Purpose: Intracranial lesions presenting with susceptibility effects on T₂*-weighted Gradient Recalled Echo (T₂* GRE) may be either calcific (diamagnetic, negative susceptibility) or hemorrhagic (paramagnetic, positive susceptibility). The purpose of this trial was to classify intracranial susceptibility lesions as either calcific or hemorrhagic.
Methods: Ten patients with susceptibility lesions on T₂* GRE were recruited into this IRB-approved trial. Patients underwent both CT and MRI with 3D T₁-weighted Fast Gradient Echo and 3D Multi-echo Fast GRE (MFGRE) with 12 echoes to perform Quantitative Susceptibility Mapping (QSM). All images were registered to the T₁-weighted MRI and susceptibility lesions were manually segmented on the first echo of the MFGRE (TE = 4.5 ms). The regions-of-interest were applied to the QSM and CT images, and mean susceptibility (ppb) and HU were recorded for each lesion. Lesions were classified as calcific on CT with HU > 100, hemorrhagic otherwise. Lesions were classified as calcific on MR when mean susceptibility was negative, hemorrhagic otherwise.
Results: A total of 63 susceptibility lesions were identified. CT classified 37 lesions as calcifications (178 +/- 79 HU), with 26 hemorrhage (40 +/- 15 HU). By QSM, 39 lesions were classified as calcifications (-109 +/- 85 ppb) and 28 as hemorrhage (882 +/- 574 ppb). For 37 lesions (59%) the classification into calcific was in agreement between CT and MR, for 24 lesions (38%) the classification for hemorrhage was in agreement between modalities. 2 cases (3%) gave discordant results where CT pointed to hemorrhage and MR to calcification (-45 and -129 ppb; CT numbers 87 and 78 HU). The QSM-measured mean susceptibility of CT-proven calcifications had a negative correlation with CT number (R² = 0.496, p<0.001).
Conclusion: QSM and CT are in good agreement on the classification of known calcific lesions, and may allow for imaging-based classification of susceptibility lesions.
Funding Support, Disclosures, and Conflict of Interest: This research was supported by GE Healthcare and funds from Dr. William Murphy, Jr., the John S. Dunn, Sr. Distinguished Chair in Diagnostic Imaging at MD Anderson Cancer Center.
Contact Email: