Program Information
Sub-MSv Cerebral CT Perfusion Imaging Using Prior Image Constrained Compressed Sensing: A Prospective Clinical Trial
K Li1*, P Yang1,2 , M Cathey1 , J Tang1,3 , H Rowley1 , P Turski1 , V Prabhakaran1 , C Strother1 , G.-H. Chen1 , (1) University of Wisconsin-Madison, Madison, WI (2) Changhai Hospital, Second Military Medical University, Shanghai, China, (3) GE Healthcare, Waukesha, WI
Presentations
TH-AB-207A-10 (Thursday, August 4, 2016) 7:30 AM - 9:30 AM Room: 207A
Purpose: A prospective clinical study was conducted to assess whether sub-mSv cerebral CT perfusion (CTP) imaging can be achieved using the Prior Image Constrained Compressed Sensing (PICSS) reconstruction algorithm.
Methods: Following a standard dose (SD) CTP acquisition, 31 patients (median age, 62 years; 15 M/16 F; median BMI, 27) underwent an additional reduced dose (RD) CTP acquisition at 20% of the standard dose level. Several algorithms (PICCS, ASIR, and FBP) were used to reconstruct CTP source images of the RD series. SD-FBP served as the reference standard. Two clinical observers (one neuroradiologist and one neurosurgeon) scored CTP maps using qualitative scoring system and quantitative noise measurements were performed. Statistical analyses were performed to interpret the qualitative and quantitative assessments.
Results: The CTDIvol and effective dose for the RD series was 44.5 mGy and 0.75 mSv respectively, compared with 222.6 mGy and 3.74 mSv respectively for the SD series. The average image quality scores of the RD-PICCS series were 3.8±0.7 for CBF, 3.5±0.9 for CBV, 3.9±0.9 for MTT. In comparison, subjective image quality scores of SD-FBP were 3.5±0.8 for CBF, 3.5±0.9 for CBV, 3.3±0.8 for MTT. The average noise of RD-PICCS (CBF, 5.8±1.9 mL/100 g/min; CBV, 0.74±0.26 mL/100 g; MTT, 5.5±1.2 s) were lower than those of SD-FBP (CBF, 8.1±2.5 mL/100 g/min; CBV, 0.86±0.28 mL/100 g; MTT, 6.4±1.2 s) with statistical significance (p < 0.001). Both subjective scores and objective noise of RD-ASIR and RD-FBP were significantly inferior to those of SD-FBP (p < 0.001).
Conclusions: PICCS enables sub-mSv cerebral CTP imaging without compromising diagnostic image quality.
Funding Support, Disclosures, and Conflict of Interest: Funding Support: 2015 AAPM Research Seed Grant. Disclosures: J. Tang: Employee, GE Healthcare; H. Rowley: Research Consultant, Bracco Group, Guerbet SA, GE Healthcare, F. Hoffmann-La Roche Ltd, W.L. Gore & Associates, Lundbeck Group; C. Strother: Research Consultant, Siemens AX; G.-H. Chen: Research funded, GE Healthcare, Siemens AX.
Contact Email: