A timing invariant CT angiography (CTA) derived from CT perfusion (CTP) data avoids additional contrast, radiation exposure and time associated with traditional CTA. Previous studies have validated this technique for LVO confined to M1 MCA. With advances in endovascular techniques, the previously considered “distal” branches of MCA like M2 segments are now considered amenable to thrombectomy.

To test reliability of time invariant CTA derived from CTP for identifying M2 MCA occlusions. 

We retrospectively review consecutive patients with CTP. CT perfusion protocol used 40 cc of iodinated contrast and 40 seconds acquisition time. Timing invariant CTA images were generated from the CT perfusion data. Volume perfusion CT data were routinely reconstructed with a slice width of 1.5 mm every 1 mm for CTA analysis .We scored  those scans for subjective CTA image quality of the M1/M2 segments, presence, absence and site of occlusion recorded by three blinded independent reviewers and compared to diagnostic cerebral angiogram (DSA) read by an independent neurointerventionalist in cases 24/45 cases that underwent endovascular intervention. Observed probability of agreement was computed to measure agreement. 

Out of 45 patients analyzed 1 had no occlusion,20 had M1 and 24 had M2 occlusion. The image quality observed agreement ranged from 96% (95% CI: 92-99%) for M1 occlusion to 77% (95% CI:  72-81%) for distal M2 occlusion. The observed agreements compared 3 image reviewers vs DSA in M2 patients (n=24) was 98% for identifying occlusion (95% CI 95-100%), 94% for identifying proximal M2 occlusion (95% CI 88-98%). Beyond the bifurcation there was 91% (95% CI 84-97%) and 90% ( 95% CI 83-95%) agreement respectively for correctly identifying inferior and superior branch of M2 occlusion. 

Time invariant CTA derived from CTP data preserves high diagnostic yield for correctly identifying M2 MCA occlusion without addition of dedicated CTA