Correlation of T max volumes with clinical outcome in anterior circulation stroke

Abstract Background and Purpose The recent thrombectomy trials have shown that perfusion imaging is helpful in proper patient selection in thromboembolic stroke. In this study, we analyzed the correlation of pretreatment T max volumes in MR and CT perfusion with clinical outcome after thrombectomy. Methods Forty‐one consecutive patients with middle cerebral artery occlusion (MCA) or carotid T occlusion treated with thrombectomy were included. T max volumes at delays of >4, 6, 8, and 10 s as well as infarct core and mismatch ratio were automatically estimated in preinterventional MRI or CT perfusion using RAPID software. These perfusion parameters were correlated with clinical outcome. Outcome was assessed using modified Rankin scale at 90 days. Results In patients with successful recanalization of MCA occlusion, T max > 8 and 10 s showed the best linear correlation with clinical outcome (r = 0.75; p = .0139 and r = 0.73; p = .0139), better than infarct core (r = 0.43; p = .2592). In terminal internal carotid artery occlusions, none of the perfusion parameters showed a significant correlation with clinical outcome. Conclusions T max at delays of >8 and 10 s is a good predictor for clinical outcome in MCA occlusions. In carotid T occlusion, however, T max volumes do not correlate with outcome.


| INTRODUCTION
MR and CT perfusion are successfully used in clinical routine for the estimation of salvageable tissue in acute ischemic stroke, which can be helpful in identifying patients eligible for mechanical thrombectomy (Butcher et al., 2005;Campbell et al., 2012. Often time to peak (TTP) or T max are used for the estimation of hypoperfused volume (Lansberg et al., 2012). Compared to TTP parameter, T max has the advantage of a reduced dependence on bolus shape and cardiac output (Olivot et al., 2009;Straka, Albers, & Bammer, 2010). Also, T max values in MR and CT perfusion have been shown to be comparable (Campbell et al., 2012;Lin, Bivard, Levi, & Parsons, 2014).
T max volumes with different delay thresholds, infarct core, and mismatch ratio can be estimated automatically using analysis software such as RAPID (Austein et al., 2016;Dehkharghani et al., 2015;Lansberg et al., 2011). These estimations are based on diffusionweighted images (DWI) and deconvolution of dynamic susceptibility contrast perfusion (DSC-PWI) in MRI or volume perfusion CT.
Previous studies have shown that infarct core correlates well with collaterals and final infarct volume in anterior circulation stroke (Lee et al., 2015;Mokin et al., 2017;Seker, Potreck, Möhlenbruch, Bendszus, & Pham, 2016). The correlation of various T max volumes with clinical outcome, however, is not clear yet. Therefore, we analyzed the correlation of T max volumes, infarct core, and mismatch ratio with clinical outcome after thrombectomy in anterior circulation stroke.

| Patients
Between March 2014 and April 2015 all consecutive patients undergoing neurothrombectomy for the treatment of acute ischemic stroke were collected in a prospective database and analyzed retrospectively. Inclusion criteria for this study were an acute occlusion of the terminal carotid internal artery (carotid T) or the M1 segment of the middle cerebral artery (MCA), premorbid modified Rankin scale score of ≤2, and treatment with thrombectomy. In total, 41 patients met the inclusion criteria for this retrospective study. Baseline demographic data, initial National Institutes of Health Stroke Scale (NIHSS) score, and modified Rankin scale score at 90 days after stroke onset (90d-mRS) were regularly documented by a neurologist. Good neurological outcome was defined as 90d-mRS score of ≤2 and poor outcome was defined as 90d-mRS >2. Our institutional review board approved the study.

| Imaging protocol
All patients underwent either multimodal CT imaging or MRI prior to thrombectomy as described recently Seker et al., 2016).
Briefly, multimodal CT imaging was performed using a 64-multislice CT (SOMATOM Definition AS, Siemens, Erlangen, Germany). Scanning order was nonenhanced CT, volume perfusion CT, and single-phase CT angiography. Acquisition parameters for CT perfusion were 180 kV and 80 mAs, acquisition duration was 60 s. A single contrast bolus of 36 ml Xenetix 350 (Guerbet, Sulzbach, Germany) was applied at a flow rate of 6.0 ml/s, followed by a saline flush of 20 ml and a flow rate of 6.0 ml/s. CT perfusion was reconstructed with a slice thickness of 5 mm every 3 mm. Acquisition parameters for CTA were 120 kV and 20 mAs. A single contrast bolus of 65 ml Xenetix 350 was given with a flow rate of 4.0 ml/s, followed by a saline flush of 20 ml and a flow rate 2D-T2-TSE-weighted and 2D-FLAIR-weighted images were acquired with TR=5,000 ms/8,500 ms, TE=85 ms/133 ms, ST = 5 mm, and slice gap = 0.5 mm. DSC-PWI was performed with a T2*-weighted gradient-echo echo-planar imaging sequence and was started with bolus injection of a 0.1 mmol/kg of intravenous gadoterate meglumine. Bolus and prebolus were injected at an injection rate of 5 ml/s. Sections (25-27) were imaged with fat suppression (TR/ TE = 2220 ms/36 ms, flip angle: 90°, field of view: 240 × 240 mm, image matrix: 128 × 128 mm, ST=5 mm). In total, 50-75 dynamic measurements were performed (including at least eight prebolus measurements) .

| Image analysis
Recanalization success was assessed on final angiograms. Successful recanalization was defined as Thrombolysis In Cerebral Infarction (TICI) score of 2b-3.

| Postprocessing
Preinterventional DWI and DSC-PWI or volume perfusion CT data were processed by RAPID software (research version, Stanford University). RAPID automatically identified infarct core, mismatch ratio, and volumes at four T max thresholds (delay of >4 s, >6 s, >8 s, and >10 s).
Estimation of ischemic core was based on either relative CBF (threshold of <31% of mean contralateral CBF) or apparent diffusion coefficient maps (Campbell et al., 2011;Straka et al., 2010). Mismatch was defined as perfusion lesion volume (T max delay of >6 s) to infarct core (diffusion lesion or CT-relative hemispheric cerebral blood flow decrease <30%) volume ratio >1.2 with >10 ml absolute mismatch and ischemic core volume <70 ml .

| Statistical analysis
Statistical analyses were performed using R version 3.2.2 and RStudio version 0.99. T max volumes at thresholds of 4, 6, 8, and 10 s, infarct core volume and mismatch ratio were correlated with 90d-mRS using Pearson's correlation. T test was used for nonparametric univariate analyses. Statistical significance was set at p < .05. Given multiple comparisons, Benjamini-Hochberg correction was used to minimize the false-discovery rate.

| Patient characteristics
The demographic characteristics of the patients are shown in Table 1.

| MCA occlusion
In MCA occlusion, T max volumes at a delay of >8 and 10 s showed the best linear correlation with clinical outcome (r = 0.68; p = .0139 and r = 0.67; p = .0139), better than infarct core (r = 0.51; p = .5299). T max at >8 and 10 s did not show a significant correlation with infarct core (r = 0.10; p = .6581 and r = 0.19; p = .3678). Linear correlation was even better in patients with successful recanalization (T max at >8 s, r = 0.75; p = .0139 and T max at >10 s, r = 0.73; p = .0139) ( Table 2).
Furthermore, T max volumes at >6, 8, and 10 s were significantly higher in patients with poor outcome compared to good outcome (Figure 1a).

| Carotid T occlusion
In terminal internal carotid artery occlusion, none of the perfusion parameters including infarct core and mismatch ratio showed a significant correlation with clinical outcome (Table 2). Also, no significant differences in T max volumes were determined between patients with good and poor clinical outcomes (Figure 1b).

| DISCUSSION
The past thrombectomy trials have shown that proper patient selection is important in mechanical thrombectomy (Meckel & Herweh, 2015;Sheth et al., 2016). In EXTEND-IA and SWIFT PRIME, MR and CT perfusion were successfully used for patient selection Saver et al., 2015). Both MR and CT perfusion help distinguishing between infarct core and salvageable tissue. In EXTEND-IA, for instance, T max at a delay of >6 s was defined as hypoperfused volume and either cerebral blood flow (CBF) in CT or DWI in MRI were used to determine infarct core volume.
Unfortunately, there are differences in perfusion maps gener-   perfusion delays as well as infarct core volume and mismatch ratio in a user-friendly way which is currently not offered by manufacturers (Dehkharghani et al., 2015).
Various perfusion parameters in both CT and MRI have been studied (Lansberg et al., 2011;Mundiyanapurath et al., 2017;Seker et al., 2016;Straka et al., 2010). Mokin et al. (2017) and Albers et al. (2016) showed that CBF and cerebral blood volume (CBV) provide accurate prediction of final infarct volume after successful recanalization. Past studies have also demonstrated that T max volume at a delay of >6 and/or 8 s accurately differentiate penumbra from salvageable tissue (Shih et al., 2003;Zhang et al., 2015). The correlation of T max volumes with clinical outcome is not clear yet, though. We, therefore, performed a retrospective analysis on the correlation of T max values at perfusion delays of >4, 6, 8, and 10 s with clinical outcome in anterior circulation stroke using RAPID.
In patients with successful recanalization of MCA occlusion, T max volumes at thresholds of 8 and 10 s showed good linear correlation with clinical outcome with significant results (r = 0.75; p = .0139 and r = 0.73; p = .0139), even better than infarct core (r = 0.43; p = .2592) which is based on CBF or DWI in RAPID algorithm (  (Austein et al., 2016;Campbell et al., 2012;Dehkharghani et al., 2015;Straka et al., 2010). Finally, patients who were not treated with endovascular therapy were not included in this analysis which is a potential source of selection bias.

| CONCLUSIONS
T max volumes at a delay of >8 and 10 s have a strong correlation with clinical outcome in middle cerebral artery occlusions as measured by RAPID. In terminal internal carotid artery occlusion, however, T max volumes do not correlate with outcome. Further studies with larger study populations are necessary to confirm our findings, though.
F I G U R E 1 (a) In middle cerebral artery occlusion, T max volumes at delays of 6, 8, and 10 s in patients with poor outcome were significantly higher compared to patients with good outcome (p = .1350, p = .0157, p = .0139, and p = .0146, respectively). (b) In carotid T occlusion, there were no significant differences (p = .9743, p = .7805, p = .2593, and p = .4216). The asterisks indicate statistically significant differences (p < .05)