Purpose: We hypothesized that two-phase (the arterial phase and the phase between the peak and late venous phases) CTA may work comparably with 3-phase CTA, and sought to compare a simulated 3-phase CTA with a simulated 2-phase CTA for assessment of collaterals.
Methods: We enrolled 155 patients (76 male; median, 72 years) who underwent 9.6-cm, 70-kVp perfusion CT (PCT) for 60 seconds within 8 hours after symptom onset and had occlusion in the ICA (ICA with/without M1; n=63) or M1 segment (with/without M2 segment; n=92). PCT data were reconstructed with a section width of 1 mm, and were used to create the simulated 3-phase CTA (3-CTA; peak arterial, peak venous, and late venous phase; 3 seconds each) with an increment of 4.5 seconds. Another 3-second CTA immediately after peak venous phase was additionally generated, which consisted of the second CTA of the simulated 2-phase CTA (2-CTA) with the peak arterial CTA. Each CTA was reconstructed into 25-mm axial maximum intensity projections, and was assessed by 2 independent reviewers at a separate session. For the 2-CTA, good collaterals were determined when the ipsilateral side showed both similar vascular filling and enhancement (ASPECT>7) compared to the contralateral brain on the peak arterial CTA. For the 3-CTA, we followed the published criteria to determine good/poor collaterals.
Results: Interobserver agreements were excellent (k=0.931 for 3-CTA; k=0.929 for 2-CTA). Consensus review determined good and poor collaterals in 97 and 58 patients, respectively, on 3-CTA. All 97 patients with good collaterals on 3-CTA showed good collaterals on 2-CTA; 2 out of 58 patients with poor collaterals on 3-CTA were determined to have good collaterals on 2-CTA, showing no significant difference (P=0.5, McNemar test).
Conclusion: Compared to the simulated 3-phase CTA, the simulated 2-phase CTA has comparable diagnostic performance for determining the status of collateral circulation.