We know that non-contrast head cts fail us in the evaluation of posterior strokes, but can posterior circulation stroke (without posterior large vessel occlusion) be diagnosed with CT perfusion (CTP) in lieu of MRI?
Based on current literature and discussions with some experts in the field, the answer seems to be no. MRI remains the gold standard for the diagnosis of stroke in general. But if you pose the question a different way the answer becomes more interesting: is CTP useful in early diagnosis of posterior stroke? Classically we are taught that the posterior fossa is difficult to appreciate on CT due to artifact from the skull base, but with improved resolution and adjunctive studies like CT angiogram (CTA) and CTP is this still true?
I sat down with Dr. Eugene Gu, neurointensivist at Stony Brook, to discuss this topic. The benefit of CTP comes in its wide availability, especially at institutions where emergent or urgent MRI is not available. CTP is important specifically when considering neurointervention as it allows one to detect a perfusion mismatch suggesting viable, at risk tissue. The diagnostic accuracy of CTP in general decreases as the size of the stroke decreases, and posterior strokes are often quite small leading to difficulty in diagnosis by this modality. Much of this can be attributed to the difference between the vasculature of the anterior and posterior circulations: the main arteries of the anterior circulation have multiple large branching vessels that if affected lead to relatively large volume stroke while the main arteries of the posterior circulation have small perforating branches which affect small volume areas. Therefore a more distal posterior circulation stroke is basically impossible to see on CTP. This sort of stroke would also not be amenable to neuro-intervention.
CTP is known to have high sensitivity and specificity for identification of anterior circulation ischemic stroke1. Two recent studies have shown the additive benefit of CTP when performed in conjunction with non-contrast CT and CTA.
A retrospective study performed by Sporns et al. in 2016 evaluated the additive benefit of whole brain volume perfusion scan upon initial presentation in conjunction with non-contrast CT and CTA in patients admitted for suspected ischemic stroke at a tertiary care center, focusing on those diagnosed with posterior stroke. Interestingly only 267 of the 3,011 patients diagnosed with acute stroke during the study period were included (about 1 in 5 ischemic strokes are posterior so this would work out to about 600 patients; it is unclear which exclusion criteria caused such significant attrition). 76.6% of posterior circulation infarcts detected on followup MRI were previously detected by CT+CTA+CTP (vs 21.3% on non-con CT only and 43.6% CT+CTA). Sensitivity was lowest in pons and midbrain stroke1. A similar study by Van der Hoeven et al. in 2015 found comparable results, citing a sensitivity of 80% for CTP in the detection of posterior circulation stroke2. My take away message from this is not that using this modality will cause me to miss 20-25% of patients with posterior stroke, but rather I may be able to identify a large number of posterior strokes in a timely manner with CTP in patients who may now be candidates for intervention.
Interestingly patients in whom not all regions of the posterior circulation were included in the CT images were excluded in both of these studies. Some studies which have evaluated the diagnostic impact of CTP exclude patients with posterior circulation in entirety due to technical issues as the standard axial plains do not routinely cover the infra-tentorial structures sufficiently3. Therefore it’s important to know what your radiology department is capturing.
In conclusion, if your patient awoke with the symptoms or is within the window for intervention or tPA (saving all arguments about tPA for another day), CTP has a limited but potential role for the detection of posterior stroke in borderline patients, but should not be used to exclude this diagnosis.