Richard I. Aviv

Since its inception, I have been convinced of the extraordinary potential of CT perfusion imaging in acute stroke to improve upon the landmark NINDS and ECASS stroke studies that used noncontrast CT alone to define a target population for stroke thrombolysis. It is incumbent on all clinicians using CTP to understand the potential and limitations of the technique. Responsible research ensures the appropriate application of CTP and balances concerns for radiation exposure with patient benefit.

The two featured papers form a small part of my ongoing interest in exploring the limits of CTP to diagnose infarct presence and predict patient outcome. Our research shows that CTP can rapidly and efficiently define infarct extent with heightened confidence while excluding important mimics, thus ensuring that the desired population receives thrombolysis. Sometimes this basic clinical message is obfuscated by much noise arguing the evils of CTP in favor of MR perfusion. After years of CTP research I remain convinced of the central role CTP will play in patient selection in acute ischemic stroke.

Reinforced by our research practice of prospectively enrolling every patient into our stroke database, CTP remains an indispensable part of our stroke workup at Sunnybrook Health Sciences Center. Clinically, CTP increases the confidence of our stroke and general neurologists to diagnose and manage acute stroke and stroke-like presentations by confirming the presence and distribution of infarct and ischemia and the collateral status. Increasingly, my research has focused on multiparametric imaging, combining both baseline multimodal CT data and patient-specific clinical variables to create a predictive tool for outcome determination. Current models are too simplistic, ignoring important variables that modify response to therapy and outcome. Factors such as age, gender, baseline presentation severity, glycemic status, and time to presentation are critical determinants that require consideration and can be integrated into baseline

perfusion assessment to give a better snapshot of infarct and tissue-at-risk status. This approach is not unique to stroke, as evidenced by a surge in "genomic imaging," which emphasizes a more holistic role of the radiologist in outcome/response prediction. It is my hope that the approach will be adopted by manufacturers of existing automated packages that currently define core and penumbral volumes for patient selection in acute stroke intervention based simply on nonspecific thresholds. We are also independently developing a GUI that will achieve this goal through a vendor neutral user-friendly, open source interactive interface. Through collaborative efforts by a North American CTP interest group we expect to test the performance of our method against other products on the market.

I anticipate the next generation of acute stroke studies using multiparametric physiological selection parameters to identify target patients, and I remain confident that superior outcomes will be achieved through these more refined selection strategies.

Read these articles at AJNR.org …

Toward patient-tailored perfusion thresholds for prediction of stroke outcome

Multimodal CT provides improved performance for lacunar infarct detection