Clinical Radiology 70 (2015) e41ee50
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Comparing perfusion CT evaluation algorithms for predicting outcome after endovascular treatment in anterior circulation ischaemic stroke A.V. Khaw a, b, A. Angermaier a, M. Kirsch c, C. Kessler a, N. Hosten c, S. Langner c, * a
Dept. of Neurology, University Medicine Greifswald, Greifswald, Germany Dept. of Clinical Neurosciences, University of Western Ontario, London Health Sciences Centre, London, Ontario, Canada c Dept. of Diagnostic Radiology and Neuroradiology, University Medicine Greifswald, Greifswald, Germany b
art icl e i nformat ion Article history: Received 17 August 2014 Received in revised form 27 November 2014 Accepted 2 February 2015
AIM: To analyse perfusion CT (PCT) evaluation algorithms for their predictive value for outcome after endovascular therapy (ET) in acute ischaemic stroke. MATERIALS AND METHODS: Twenty-six patients were prospectively enrolled to undergo endovascular therapy for moderate to severe [National Institute of Health Stroke Scale (NIHSS) score of 5] anterior circulation stroke 6 h of onset. PCT datasets were evaluated according to three algorithms: visual mismatch estimate (VME), Alberta Stroke Programme Early CT Score (ASPECTS) perfusion, and quantitative perfusion ratios (QPRs: RCBF, RCBV) of cerebral blood flow (CBF) and volume (CBV). Results were correlated with outcome measures [NIHSS score at discharge, NIHSS score change until discharge (DNIHSSA/D), mRS at 90 days (mRS90d)] and compared with a matched control group. RESULTS: Recanalization was achieved in 73%, median NIHSS score decreased from 14 to 5 at discharge. The treatment and control group did not differ by VME and ASPECTS perfusion, nor did VME correlate with any of the three outcome measures. ASPECTS perfusion was not predictive of any outcome measure in the ET group. RCBF and RCBV were associated with DNIHSSA/D in controls and, inversely, the ET group, but not with mRS90d. Receiver operating characteristic (ROC) analysis of RCBF (and RCBV) showed a positive predictive and negative predictive value of 87% (78%) and 74% (73%), respectively, for discriminating major neurological improvement (DNIHSSA/D 2, NIHSS at discharge (NIHSSD) 1 versus >1, and DNIHSSA/D 7 versus >7.
Statistical analysis Spearman’s and Pearson’s correlation coefficients were calculated to analyse the relationship between PCT parameters and mRS score and NIHSS score, respectively. In the case of a significant correlation with an outcome measure, differences in imaging parameters were tested for by dichotomized outcome measures. Finally, receiver operating characteristic (ROC) area under the curves (AUC) were computed. All statistical analyses were performed using SPSS version 14 (Chicago, IL, USA). A p-value 3e4, >4e5, >5e6 h or dichotomized to 3 versus >3 h. At discharge, controls had a median NIHSS score of 14 (p ¼ 0.001 versus ET group), a median DNIHSSA/D of 1 (p < 0.001), and a median mRS score of 4 (p ¼ 0.001), with a proportion of 9% (n ¼ 2) achieving favourable outcomes (mRS2, p ¼ 0.004).
Imaging The PCT scans of four patients undergoing ET and two controls were excluded from analysis due to technical Table 3 Distribution of occlusion sites and TICI grades of recanalization. Occlusion site
Carotid T Intracranial ICA M1 M1/M2 M2 Subtotals by TIMI (n,%)
Subtotals by site, n (%) 2 1 6 5 12
(8) (4) (23) (19) (46)
TICI (n,%) 0 0 1 1 1 2 5
(4) (4) (4) (8) (19)
1
2a
2b
3
1 (4) 0 1 (4) 0 0 2 (8)
0 0 0 0 0 0
0 0 2 2 5 9
1 (4) 0 2 (8) 2 (8) 5 (19) 10 (38)
(8) (8) (19) (35)
TICI, thrombolysis in cerebral infarction; ICA, internal carotid artery; M1 and M2, middle cerebral artery segment.
e46
A.V. Khaw et al. / Clinical Radiology 70 (2015) e41ee50
Figure 3 Early neurological improvement, measured as difference (D) in NIHSS score between admission (pre-treatment) and discharge and functional outcome at 90 days (mRS score) according to initial stroke severity of the patient group undergoing intra-arterial thrombolysis. NIHSS values 0.22).
VME
ASPECTS perfusion
In the ET group, presence of a mismatch between CBF and CBV lesion size suggesting tissue at risk was found in 20 of 22 patients (91%), which did not differ from controls (17 of 20 patients, 85%, p ¼ 0.9). Semiquantitative VMEs ranged from 0e60% (median 20%) in the ET group and from 0e40% (median 10%) in controls, with no difference between groups (p ¼ 0.23). Neither the dichotomous nor the
Median ASPECT scores for CBF, CBV, and TTP were not different between the ET and the control group (2.5, 3, 2, respectively; p > 0.25). In the control group, the ASPECT scores of all three perfusion parameters were inversely associated with the NIHSS score at discharge and mRS90d. No association was found with early neurological improvement (Table 4).
Figure 4 Box plots of early neurological improvement (DNIHSS score discharge-admission) and long-term functional outcome (mRS at 90 days) according to recanalization status after intra-arterial thrombolysis. NIHSS values
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Comparing perfusion CT evaluation algorithms for predicting outcome after endovascular treatment in anterior circulation ischaemic stroke A.V. Khaw a, b, A. Angermaier a, M. Kirsch c, C. Kessler a, N. Hosten c, S. Langner c, * a
Dept. of Neurology, University Medicine Greifswald, Greifswald, Germany Dept. of Clinical Neurosciences, University of Western Ontario, London Health Sciences Centre, London, Ontario, Canada c Dept. of Diagnostic Radiology and Neuroradiology, University Medicine Greifswald, Greifswald, Germany b
art icl e i nformat ion Article history: Received 17 August 2014 Received in revised form 27 November 2014 Accepted 2 February 2015
AIM: To analyse perfusion CT (PCT) evaluation algorithms for their predictive value for outcome after endovascular therapy (ET) in acute ischaemic stroke. MATERIALS AND METHODS: Twenty-six patients were prospectively enrolled to undergo endovascular therapy for moderate to severe [National Institute of Health Stroke Scale (NIHSS) score of 5] anterior circulation stroke 6 h of onset. PCT datasets were evaluated according to three algorithms: visual mismatch estimate (VME), Alberta Stroke Programme Early CT Score (ASPECTS) perfusion, and quantitative perfusion ratios (QPRs: RCBF, RCBV) of cerebral blood flow (CBF) and volume (CBV). Results were correlated with outcome measures [NIHSS score at discharge, NIHSS score change until discharge (DNIHSSA/D), mRS at 90 days (mRS90d)] and compared with a matched control group. RESULTS: Recanalization was achieved in 73%, median NIHSS score decreased from 14 to 5 at discharge. The treatment and control group did not differ by VME and ASPECTS perfusion, nor did VME correlate with any of the three outcome measures. ASPECTS perfusion was not predictive of any outcome measure in the ET group. RCBF and RCBV were associated with DNIHSSA/D in controls and, inversely, the ET group, but not with mRS90d. Receiver operating characteristic (ROC) analysis of RCBF (and RCBV) showed a positive predictive and negative predictive value of 87% (78%) and 74% (73%), respectively, for discriminating major neurological improvement (DNIHSSA/D 2, NIHSS at discharge (NIHSSD) 1 versus >1, and DNIHSSA/D 7 versus >7.
Statistical analysis Spearman’s and Pearson’s correlation coefficients were calculated to analyse the relationship between PCT parameters and mRS score and NIHSS score, respectively. In the case of a significant correlation with an outcome measure, differences in imaging parameters were tested for by dichotomized outcome measures. Finally, receiver operating characteristic (ROC) area under the curves (AUC) were computed. All statistical analyses were performed using SPSS version 14 (Chicago, IL, USA). A p-value 3e4, >4e5, >5e6 h or dichotomized to 3 versus >3 h. At discharge, controls had a median NIHSS score of 14 (p ¼ 0.001 versus ET group), a median DNIHSSA/D of 1 (p < 0.001), and a median mRS score of 4 (p ¼ 0.001), with a proportion of 9% (n ¼ 2) achieving favourable outcomes (mRS2, p ¼ 0.004).
Imaging The PCT scans of four patients undergoing ET and two controls were excluded from analysis due to technical Table 3 Distribution of occlusion sites and TICI grades of recanalization. Occlusion site
Carotid T Intracranial ICA M1 M1/M2 M2 Subtotals by TIMI (n,%)
Subtotals by site, n (%) 2 1 6 5 12
(8) (4) (23) (19) (46)
TICI (n,%) 0 0 1 1 1 2 5
(4) (4) (4) (8) (19)
1
2a
2b
3
1 (4) 0 1 (4) 0 0 2 (8)
0 0 0 0 0 0
0 0 2 2 5 9
1 (4) 0 2 (8) 2 (8) 5 (19) 10 (38)
(8) (8) (19) (35)
TICI, thrombolysis in cerebral infarction; ICA, internal carotid artery; M1 and M2, middle cerebral artery segment.
e46
A.V. Khaw et al. / Clinical Radiology 70 (2015) e41ee50
Figure 3 Early neurological improvement, measured as difference (D) in NIHSS score between admission (pre-treatment) and discharge and functional outcome at 90 days (mRS score) according to initial stroke severity of the patient group undergoing intra-arterial thrombolysis. NIHSS values 0.22).
VME
ASPECTS perfusion
In the ET group, presence of a mismatch between CBF and CBV lesion size suggesting tissue at risk was found in 20 of 22 patients (91%), which did not differ from controls (17 of 20 patients, 85%, p ¼ 0.9). Semiquantitative VMEs ranged from 0e60% (median 20%) in the ET group and from 0e40% (median 10%) in controls, with no difference between groups (p ¼ 0.23). Neither the dichotomous nor the
Median ASPECT scores for CBF, CBV, and TTP were not different between the ET and the control group (2.5, 3, 2, respectively; p > 0.25). In the control group, the ASPECT scores of all three perfusion parameters were inversely associated with the NIHSS score at discharge and mRS90d. No association was found with early neurological improvement (Table 4).
Figure 4 Box plots of early neurological improvement (DNIHSS score discharge-admission) and long-term functional outcome (mRS at 90 days) according to recanalization status after intra-arterial thrombolysis. NIHSS values
