Photon-Counting CT Versus Energy-Integrating CT Angiography for Craniocervical Fibromuscular Dysplasia: A Preliminary Retrospective Paired Analysis
Pranjal Rai, Vincent E. Y. Chan, Neetu Soni, Ahmed O. El Sadaney, Shravani Kharat, John C. Benson, Amit Agarwal, Victoria M. Silvera, Lifeng Yu, Felix E. Diehn, Girish BathlaBackground:
Assessing craniocervical fibromuscular dysplasia (cc-FMD) can be challenging using cross-sectional imaging. We hypothesize that photon-counting detector CT (PCD-CT), owing to its superior spatial resolution over conventional energy-integrating detector CT (EID-CT), may better help appreciate changes in cc-FMD.
Materials and Methods:
In this retrospective cohort, 24 patients with radiologically reported cc-FMD (mean age: 54 y; 75% women) underwent both PCD-CT and EID-CT within 180 days (median: 52 d). Bilateral internal carotid (ICA) and vertebral artery (VA) segments without dissection/occlusion were graded on a 3-point scale (absent/likely present/present). Intermodality agreement was evaluated with weighted Cohen κ, and radiation dose metrics (CTDIvol) were compared using the paired Wilcoxon signed-rank tests.
Results:
Eighty-seven paired segments were analyzed. PCD-CT detected FMD in 48 segments (55.1%) versus 30 (34.5%) on EID-CT. The mean reader confidence scores were consistently higher with PCD-CT compared with EID-CT, for both ICAs (2.16±0.95 vs. 1.93±0.99) and VAs (1.86±0.95 vs. 1.36±0.75). Intermodality agreement was almost perfect for ICAs (κ=0.807) and substantial for VAs (κ=0.662). Under our institutional protocols, median CTDIvol was lower with PCD-CT (16.6 mGy) than EID-CT (21.5 mGy;
Conclusion:
PCD-CT demonstrated more frequent reader categorization of cc-FMD features than EID-CT, particularly in the VAs, and had a lower median CTDIvol under our institutional protocols. These preliminary findings suggest improved cc-FMD detection with PCD-CT as compared with EID-CT and require prospective, reference-standard validation to establish accuracy and clinical impact.