International Stroke Conference
February 19–21, 2020

Parneet Grewal, MD                           

Moderators: Dr. Ashutosh Jadhav and Dr. Robin Novakovic

This session at the International Stroke Conference 2020 in Los Angeles included extensive discussion about various imaging modalities that are being used to select patients as candidates for endovascular treatment (EVT) in the real world, along with their pitfalls. The discussion was led by Dr. Albert J. Yoo, Dr. Achala Vagal, and Dr. Bernard Yan. There were also case scenarios presented by Dr. Richard Aviv of challenging CT angiography (CTA) cases.

Early time, late time, large core, small core: Non-con CT is all you need (Dr. Albert J. Yoo)

In his presentation, Dr. Yoo discussed the importance of non-contrast CT head in selecting patients for EVT and urged the clinicians to consider CT head and CTA as the only imaging modalities that are needed prior to making decisions for EVT. The EVT candidate patients are broadly divided into early window (0-6 hours) and late window (> 6-24 hours) based on their time of presentation, as well as into small core and large core based on the size of ischemic infarct. All the major landmark trials in the early window, such as MR CLEAN, ESCAPE, EXTEND IA, SWIFT PRIME and REVASCAT, utilized CT head/CTA for patient selection with EXTEND IA and SWIFT PRIME also using CT perfusion (CTP) imaging. All the trials demonstrated improved perfusion and functional outcome for patients with large vessel occlusion who underwent EVT but had different criteria for patient selection. A study by Tawil et al. on comparing eligibility for different trial protocols to estimate the number of patients eligible for treatment showed that 53% of the patient population met criteria for MR CLEAN, which decreased to only 17% for EXTEND IA. Secondary analysis of the MR CLEAN trial has also shown that the patients who were ineligible per EXTEND IA criterion also benefitted from the EVT, which means that excluding patients in the early window using CTP might lead us to miss a subset of the population which can still benefit from thrombectomy.

Dr. Yoo also discussed pitfalls of CT perfusion imaging, such as ghost infarct core in the early window, which is associated with earlier imaging and recanalization. He suggested use of high-contrast setting (35 HU/35 HU) on non-con CT head, which increases sensitivity of detection of ischemia to 71% while retaining 100% specificity. Hence, in the early window, CTP should not be used, per Dr. Yoo, both in small core and large core patients. In the late window (>6-24 hours), the only randomized controlled trials that have shown benefit of EVT are DAWN and DEFUSE 3, and per AHA/ASA 2019 guidelines, perfusion imaging is recommended to aid in patient selection for thrombectomy. However, since the number needed to treat (NNT) is 3 for late window patients, most likely the patients who do not have mismatch on perfusion imaging will also benefit from EVT. This is supported by single-center studies which showed benefit of EVT in patients who did not meet DAWN and DEFUSE 3 mismatch criteria in the 6-24 hours window. The ongoing MR CLEAN Late trial will help in providing randomized evidence regarding the same, and we might forgo the need for perfusion imaging and only select patients based on non-con CT head ASPECTS score even in the late window. In the end, Dr. Yoo urged to not undertake plurality in imaging without necessity.

Availability of advanced imaging techniques at endovascular-capable centers (Dr. Bernard Yan)

Dr. Yan from Royal Melbourne Hospital (RMH) in Australia discussed advanced imaging availability in the state of Victoria, Australia, along with the need for advanced imaging for endovascular centers. At RMH, all the code stroke patients, including direct admissions, mobile stroke unit patients, as well as transfers from spoke sites, undergo CT/CTA and CTP as part of their initial workup either at RMH or at the spoke sites. All the imaging modalities are uploaded to a single central server, and the imaging is easily available for review to the clinicians making treatment decisions. Per Dr. Yan, for patients in the late window (>6-24 hours), use of advanced perfusion imaging techniques is mandated by the AHA/ASA guidelines based on the DAWN and DEFUSE 3 trials and should be undertaken unless further evidence is available to the contrary. Regarding the patients in the early window (0-6 hours), Dr. Yan discussed different case scenarios in which getting perfusion imaging might prove to be useful and provide additional information to guide in decision making such as borderline low ASPECTS score on non-contrast CT head, unclear assessment of pre-morbid function or lack of critical information, distal occlusions beyond proximal M2 and in patients older than 80+ years. He also stressed the variability in calculation of ASPECTS score, which might necessitate getting advanced imaging with automated software in many patients, and having a system that is already established in getting perfusion imaging might save time for such cases. Imaging repository MOSES (MOnitoring of Stroke Endovascular Services), which is actively recruiting comprehensive stroke centers in Australia, was also discussed by Dr. Yan.

Can you trust the perfusion maps? Pitfalls of automated perfusion imaging (Dr. Achala Vagal)

In CT perfusion (CTP),  we observe a dynamic flow as the iodinated contrast washes in and out of circulation through the capillary bed. The images are repeated (cine mode) after contrast injection, time density curves (TDC) are generated and highly complicated mathematical models are used to calculate perfusion parameters in each voxel. Using such automated perfusion software has several advantages such as the ease of use, reproducibility, efficiency in workflow along with validation in multiple trials. However, there are several technical and diagnostic pitfalls that are associated with this imaging modality, and Dr. Vagal talked about these in her presentation. To get an adequate CTP study, one needs to have optimal contrast bolus (can be seen by reviewing the TDC, ideal CTP acquisition time is 60-70 seconds), properly selected arterial (usually A1 segment of anterior cerebral artery) and venous (usually superior sagittal sinus) phase along with no patient motion. There are also several artifacts which can alter the final results of perfusion mismatch ratio, such as skull base/orbit artifacts. Several other diagnostic pitfalls also necessitate the need to review all the imaging modalities, including non-con CT head and CTA along with CTP for an accurate ischemic core and penumbral estimation. One such example is over-estimation of the core infarct due to a chronic infarct and ghost core or under-estimation of the ischemic core due to partial reperfusion. Another pitfall is misclassification of penumbral mismatch, which can happen due to chronic infarcts, carotid stenosis, chronic occlusion, cardiac arrythmias or aortic dissection leading to over-estimation of penumbra. CTP is also not a very reliable imaging modality for ischemic infarcts in posterior fossa, lacunar infarcts or in patients with severe white matter disease. In conclusion, Dr. Vagal recommends knowing the perfusion software, always checking for technical adequacy and interpreting CTP with clinical history and other imaging such as non-contrast CT head and CTA.

Did I nearly miss that? Challenging CTA cases (Dr. Richard Aviv)

In this talk, Dr. Aviv discussed several interesting cases that gave him and his colleagues a pause for thought and were the causes of interpretive errors. The most common causes for such errors are high volume of images that need to be interpreted in a limited time, fatigue, lack of knowledge, radiologist bias, poor communication, complacency, faulty reasoning, under-reading (missed finding), poor technique and inaccurate history. Studies show that error rate drops if the attending radiologist has a resident or a fellow reading the imaging along with them. The cases discussed included:

  • C6 fracture in a patient with vertebral artery dissection as seen on CTA.
  • Vocal cord paralysis in a patient undergoing CT venogram for venous sinus thrombosis.
  • Hyper-density on non-contrast CT head attributed to peri mesencephalic subarachnoid hemorrhage, which was later diagnosed as tuberculous meningitis.
  • Carotid web in a patient with acute ischemic stroke and intracranial anterior circulation large vessel occlusion.
  • Pseudo-aneurysm formation due to vertebral artery dissection in a patient with whiplash injury.
  • Abnormal CTA due to intra-arterial iodinated contrast injection.