Shaun Patel MD, Presenter: Nothing to Disclose

Yasser Naveed Mir MD, Abstract Co-Author: Nothing to Disclose

Eytan Raz MD, Abstract Co-Author: Nothing to Disclose

Keith Siller MD, Abstract Co-Author: Nothing to Disclose

Bidyut Kumar Pramanik MD, Abstract Co-Author: Nothing to Disclose

Ke Lin MD, Abstract Co-Author: Speakers Bureau, Siemens AG

The operational paradigm for defining the acute infarct core on CT perfusion (CTP) has been the low cerebral blood volume (CBV) lesion.  However, this definition arose from previous software packages that were not equipped with delay-correction which can overestimate the size of the core on both the CBV and cerebral blood flow (CBF) maps.  Our aim was to determine whether CBV or CBF was optimal for delineating the infarct core on delay-insensitive CTP software.

18 patients were retrospectively selected from our prospective stroke registry who (1) arrived to the ED with acute MCA infarction <3 hrs, (2) were evaluated by CTP, (3) were given IV tPA or had thrombectomy, (4) achieved acute clinical recovery with recanalization, and (5) had diffusion-weighted imaging (DWI) follow-up for defining infarct size. CTP was performed centered at the mid-basal ganglia and acquisition of 60s. CTP software employed an integration technique for the impulse residue function that includes arrival time of the arterial input function as one of the fitting parameters and is delay insensitive by design. A range of thresholds were used for CBF (20 to 2 mL/100g/min in 2 unit intervals) and CBV (2.0 to 0.2 mL/100g in 0.2 unit intervals) for which the software delineated the area of the perfusion lesion. The area of the lesion on coregistered DWI was calculated on ImageJ. Pearson correlation coefficients (R) were calculated for each parameter at each threshold. Wilcoxon test was used to compare the best correlated threshold on CBF vs. CBV vs. DWI. 

Total 32 perfusion lesions were evaluated. Final DWI lesion areas ranged 0 to 28.2 cm2, median 5.8 cm2. For CBF, R ranged 0.382 (CBF≤2) to 0.500 (CBF≤14) with all p-values <0.031. For CBV, R ranged 0.407 (CBV≤0.2) to 0.649 (CBV≤1.6) with all p-values <0.021. The median area for core was 8.9 cm2 for optimal CBF threshold (CBF≤14) and differed signficantly from DWI (p=0.017) and the median area for core of 5.9 cm2 from the optimal CBV threshold (CBV≤1.6) (p<0.0001). CBV and DWI did not differ (p=0.985).

Using more advanced delay-insensitive CTP software, CBV outperformed CBF for defining the acute infarct core with an optimal threshold at 1.6 mL/100g.

The infarct core size is a critical triage variable for predicting outcome and symptomatic hemorrhagic transformation with aggressive treatment.  It is best predicted by the CBV parameter on CTP.

Patel, S, Mir, Y, Raz, E, Siller, K, Pramanik, B, Lin, K, Comparison of Cerebral Blood Volume and Cerebral Blood Flow as the Optimal Parameter for Delineation of the Acute Infarct Core on Delay-Insensitive CT Perfusion Postprocessing.  Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL. http://archive.rsna.org/2010/9015817.html