Acute Ischemic Stroke : Infarct Core Estimation On Ct Angiography Source Images Depends On Ct Angiography Protocol (Page 5 of 12) in pdf

NEURORADIOLOGY:

Infarct Core Estimation on CT Angiography Source Images Depends on Protocol

Pulli et al

Table 3

and clinical data were compared between

groups by using the Mann-Whitney U test

Comparison of Clinical and Demographic Characteristics between Patients Imaged

and were presented as medians and IQRs.

with Protocol 1 and Those Imaged with Protocol 2

Categoric data were compared by using

Patients Imaged with

the Fisher exact test and were presented

Characteristic

Protocol 1 ( n = 35)

Protocol 2 ( n = 65)

P Value

as percentages.

Interrater agreement for CT angi-

†

Age (y) *

71 (60.3–78.0)

73 (56.8–81.0)

.621

ography source images and DW images

‡

Male sex

15 (42.9)

29 (44.6)

. .99

was examined by using the intraclass

†

NIHSSS *

15 (9–21)

16 (13.75–20.25)

.230

correlation coefﬁ cient, as well as Bland-

‡

Right-side involvement

15 (42.9)

28 (43.1)

. .99

Altman analysis ( 25 ). We sought to con-

‡

Atrial ﬁ brillation

9 (25.7)

12 (18.5)

.445

‡

Level of occlusion

.733

ﬁ rm previous reports of high interrater

Proximal ICA

1 (2.9)

2 (3.1)

agreement for DW imaging in the setting

Terminal ICA (with or without

10 (28.6)

19 (29.2)

of acute ischemic stroke ( 26,27 ) and to

extension into MCA M1)

compare the interrater agreement for

MCA M1

17 (48.6)

36 (55.4)

DW imaging to that for CT angiography

MCA M2

7 (20.0)

7 (10.8)

source images.

MCA M3

1 (1.5)

Within-group comparisons of lesion

†

Time to imaging (h:min) *

4:07 (2:47–5:39)

4:00 (2:31–4:59)

.370

volumes on CT angiography source im-

†

Time from CT angiography to MR

0:40 (0:35–1:04)

0:27 (0:19–0:37)

, .0001

ages and lesion volumes at DW imaging

imaging (h:min) *

were performed by using the Wilcoxon

‡

Received intravenous tPA before

10 (28.6)

23 (35.4)

.514

signed rank test (paired analysis). In-

imaging

termethod agreement between CT angi-

ography source images and DW images

Note.—Unless otherwise indicated, data are numbers of patients, with percentages in parentheses. NIHSSS = National Institutes

of Health Stroke Scale Score.

was further examined by using Bland-

* Data are medians, with IQRs in parentheses.

Altman analysis ( 25 ), as well as Spear-

†

Calculated with the Mann-Whitney U test.

man rank correlation. For relative com-

‡

Calculated with the Fisher exact test.

parisons, only instances in which the

infarct volume at DW imaging was 5

mL or greater were considered because

small volumes are prone to measure-

(1%) had an M3 segment occlusion.

40 minutes in group 1, versus 27 minutes

ment errors ( 28 ).

Thirty-three patients (33%) received

in group 2 ( P , .0001). However, the

Finally, the effects of the aforemen-

intravenous tPA before imaging was per-

time to MR imaging and DW imaging

tioned protocol-related parameters were

formed. These patients were admitted to

infarct volume did not correlate signif-

analyzed in a stepwise multivariate

the hospital through a “telestroke” service,

icantly ( r = 0.13 [ P = .17] for all pa-

tients, r = 0.008 [ P = .96] for group 1,

logistic regression analysis to determine

and intracranial hemorrhage was excluded

predictors of marked overestimation of

at the referring hospital by means of un-

and r = 0.28 [ P = .82] for group 2).

lesions on CT angiography source images

enhanced CT. Median time to CT imaging

Table 4 shows the results of com-

( 20% of the DW imaging lesion) and in

was 4 hours 1 minute (IQR, 2:36–5:04).

parison of regions of hypoattenuation

a multivariate linear regression analysis

Median time between CT angiography and

on CT angiography source images and

to evaluate the correlation of these pa-

DW imaging was 31 minutes (IQR, 22–

hyperintensity volumes at DW imaging

rameters with absolute (or relative) over-

41 minutes). Twenty-one patients (21%)

according to protocol. Volume on CT

estimation on CT angiography source

were found to have atrial ﬁ brillation just

angiography source images was slightly

images. Only variables with a univariate

before imaging.

smaller than the volume at DW imaging

P value of less than .1 were included.

Clinical, demographic, and imaging

in group 1 (33.0 vs 41.6 mL, P = .01)

data are given in Table 3 . Between pa-

but was signiﬁ cantly larger in group 2

tients imaged with protocol 1 (group 1;

(94.8 vs 17.8 mL, P 5 .0001 [ Fig 2 ]).

Results

n = 35) and those imaged with protocol

The median ratio of volume on CT an-

Among 100 patients who satisﬁ ed the in-

2 (group 2; n = 65), we observed no

giography source images to volume at

clusion criteria, three (3%) had an iso-

difference in age, male sex, right hemi-

DW imaging was 0.83 in group 1, ver-

lated extracranial internal carotid artery

sphere involvement, median NIHSSS,

sus 3.5 in group 2 ( P , .0001). Patients

(ICA) occlusion, 29 (29%) had a termi-

level of occlusion, time from symptom

with an infarct volume at DW imaging

nal ICA occlusion with or without MCA

onset to CT imaging, atrial ﬁ brillation,

of less than 100 mL (or , 70 mL) were

M1 segment occlusion, 53 (53%) had an

and administration of intravenous tPA

found to have a volume on CT angiogra-

MCA M1 segment occlusion, 14 (14%)

before imaging. The time from CT angi-

phy source images of greater than 100

had an M2 segment occlusion, and one

ography to MR imaging acquisition was

mL (or . 70 mL) 44.4% (or 59.6%)

597

Radiology: Volume 262: Number 2—February 2012

Acute Ischemic Stroke : Infarct Core Estimation On Ct Angiography Source Images Depends On Ct Angiography Protocol Page 5

Related Articles

Related forms

Related Categories