Acta Radiol OnlineFirst, published on February 6, 2016 as doi:10.1177/0284185116628338
Original Article
The applied research of MRI with ASSET-EPI-FLAIR combined with 3D TOF MRA sequences in the assessment of patients with acute cerebral infarction
Acta Radiologica 0(0) 1–9 ! The Foundation Acta Radiologica 2016 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0284185116628338 acr.sagepub.com
Zhichao Lin1, Zexiong Guo2, Lin Qiu3, Wanyoug Yang1 and Mingxia Lin1
Abstract Background: To extend the time window for thrombolysis, reducing the time for diagnosis and detection of acute cerebral infarction seems to be warranted. Purpose: To evaluate the feasibility of implementing an array spatial sensitivity technique (ASSET)-echo-planar imaging (EPI)-fluid attenuated inversion recovery (FLAIR) (AE-FLAIR) sequence into an acute cerebral infarction magnetic resonance (MR) evaluation protocol, and to assess the diagnostic value of AE-FLAIR combined with three-dimensional time-of-flight MR angiography (3D TOF MRA). Material and Methods: A total of 100 patients (68 men, 32 women; age range, 44–82 years) with acute cerebral infarction, including 50 consecutive uncooperative and 50 cooperative patients, were evaluated with T1-weighted (T1W) imaging, T2-weighted (T2W) imaging, FLAIR, diffusion-weighted imaging (DWI), 3D TOF, EPI-FLAIR, and AE-FLAIR. Conventional FLAIR, EPI-FLAIR, and AE-FLAIR were assessed by two observers independently for image quality. The optimized group (AE-FLAIR and 3D TOF) and the control group (T1W imaging, T2W imaging, conventional FLAIR, DWI, and 3D TOF) were compared for evaluation time and diagnostic accuracy. Results: One hundred and twenty-five lesions were detected and images having adequate diagnostic image quality were in 73% of conventional FLAIR, 62% of EPI-FLAIR, and 89% of AE-FLAIR. The detection time was 12 1 min with 76% accuracy and 4 0.5 min with 100% accuracy in the control and the optimized groups, respectively. Inter-observer agreements of k ¼ 0.78 and k ¼ 0.81 were for the optimized group and control group, respectively. Conclusion: With reduced acquisition time and better image quality, AE-FLAIR combined with 3D TOF may be used as a rapid diagnosis tool in patients with acute cerebral infarction, especially in uncooperative patients.
Keywords Acute cerebral infarction, magnetic resonance imaging (MRI), three-dimensional time-of-flight magnetic resonance angiography (3D TOF MRA), image quality, fluid attenuated inversion recovery (FLAIR) Date received: 3 November 2015; accepted: 30 December 2015
Introduction Cerebral stroke is one of the most common causes of long-term disability or death worldwide and results in a marked burden in healthcare system (1). Typically, a stroke is caused by the blood supply reduction to the brain due to cerebral artery occlusion by a clot (2). This so-called ischemic stroke (cerebral infarction), which may result from cardiogenic, lacunar, arterosclerotic, hemodynamic, or cryptogenic sources, accounts for
1 Department of Radiology, the First Affiliated Hospital of Jinan University, Guangzhou, PR China 2 Department of Urology, the First Affiliated Hospital of Jinan University, Guangzhou, PR China 3 Department of Neurology, the First Affiliated Hospital of Jinan University, Guangzhou, PR China
Corresponding author: Zexiong Guo, Department of Urology, the First Affiliated Hospital of Jinan University, 613 West Huangpu Dadao, Guangzhou City, Guangdong Province, PR China. Email: [email protected]
Downloaded from acr.sagepub.com at Flinders University on February 8, 2016
2
Acta Radiologica 0(0)
about 80% of strokes (3). The minority of strokes is related to disruption of a cerebral artery resulting in intracerebral hemorrhage (2). Currently, early recanalization of an occluded cerebral artery with use of a neurothrombectomy device or thrombolytic drug to salvage ischemic tissue at risk of infarction and decrease the size of the infarct may lead to a better functional outcome of patients with acute ischemic stroke (4). Thus, to extend the time window for thrombolysis, reducing the time for diagnosis and detection of acute cerebral infarction seems to be warranted. Neuroimaging has been demonstrated to play a significant role in the evaluation of patients with acute cerebral infarction (5). In order to better investigate anatomical details and pathologic lesions in the brain, routine magnetic resonance imaging (MRI) protocols include various types of sequences such as T1-weighted (T1W) and T2-weighted (T2W) imaging (5). However, adequate image quality is often impaired by motion artifacts which may sometimes obscure or mimic pathology in uncooperative patients with unconsciousness or respiratory failure symptoms (5,6). Because of longer acquisition time and limited availability, MRI has been mainly used in comprehensive stroke centers (7). With recent advances in MR technology, a comprehensive MR protocol including parenchymal imaging (diffusion-weighted imaging [DWI], gradient recalled-echo [GRE], fluid attenuated inversion recovery [FLAIR]), MR angiography (MRA), and MR perfusion can be obtained in 20 min as demonstrated in several clinical trials (5,8). The tissue contrast of DWI aids in the assessment of alterations of water diffusion related to the cytotoxic edema in the acute phase of brain ischemia (9). Besides, three-dimensional time-offlight MRA (3D TOF MRA), an available noninvasive technique, is a commonly used pulse sequence in the MR evaluation of intracranial arteries (10). FLAIR imaging is a T2W imaging sequence and is an indispensable part of common multi-parametric stroke MRI protocols (9). Recently, introduction of fast imaging techniques, such as parallel acquisition and echo-planar imaging (EPI), has obviously enhanced the performance of MRI with regard to acquisition speed (5,7). For instance, a study has demonstrated that the application of EPI to the FLAIR sequence (EPI-FLAIR) was feasible with comparable quantitative and qualitative results to the conventional FLAIR and resulted in reduced acquisition time (5). The reduction of acquisition time may prove useful for examining the uncooperative, medically unstable, or claustrophobic patients. However, few studies have investigated the ability of the application of both EPI and parallel acquisition techniques to the FLAIR sequence in evaluation of patients with acute cerebral infarction.
Parallel acquisition imaging spatial encoding from coil elements of the receiver array to compress the gradient encoding steps and hence accelerate acquisition (11). In this study, array spatial sensitivity technique (ASSET) parallel acquisition, a sensitivity encoding (SENSE)-based algorithm, was employed. The purpose of this study was to evaluate the feasibility of implementing an ASSET-EPI-FLAIR (AE-FLAIR) sequence into an acute cerebral infarction MR evaluation protocol, to assess the diagnostic value of AEFLAIR combined with 3D TOF MRA, and to compare with the results with conventional MR protocol.
Material and Methods Patients From March 2010 and March 2015, 100 consecutive patients with acute ischemic stroke who had undergone computed tomography (CT) to rule out intracranial hemorrhage were recruited for this study. Consecutive selection of patients was performed in this study, and these 100 patients included 50 cooperative and 50 uncooperative patients. All these patients met the following inclusion criteria: age >40 years; within 72 h after onset of ictus; in accordance with the Chinese guidelines of diagnosis and treatment for acute ischemic stroke (2010) (12) and having ruled out intracranial hemorrhage by CT evaluation; and a National Institutes of Health Stroke Scale (NIHSS) score at admission (>20). The exclusion criteria were as follows: intracranial hemorrhage; hypoxic-ischemic encephalopathy; toxication; hypoglycemia; hypotension; hepatic encephalopathy; pulmonary encephalopathy; other intracephalic diseases such as encephalitis; tumor; and parasitosis. Besides, the clinical stages of these 100 patients with acute ischemic stroke were determined as previously described (13). Additionally, this study was approved by the local institutional review committee, and written informed consent from the patients or the closest family member of the patients was obtained before the start of the study.
MRI protocol All MRI scans were performed on a 1.5-T MR scanner (Optima, MR360; GE Healthcare, Beijing, China) retrofitted for EPI and ASSET capabilities with a standard circularly polarized head coil. The following sequences were performed in all participants (the cooperative group and the uncooperative group of patients) in an axial plane: DWI, 3D TOF MRA, conventional FLAIR, Prol-T2W imaging, fast spin echo (FSE)-T1W imaging, EPI-FLAIR, and AE-FLAIR. The parameters used for these sequences were presented
Downloaded from acr.sagepub.com at Flinders University on February 8, 2016
Lin et al.
3
Table 1. Summary of imaging parameters for the sequences that were applied for MRI scans
Parameter
DWI
3D TOF MRA
Field of view (cm) TR/TE (ms) Slice thickness/inter-slice gap (mm) Matrix size NEX Asset R value b (s/mm2) BW TI (ms) Echo train length
24 24 5000/minimum 6/2
22 18 28/6.9 1.2/0
24 18 8600/140 6/2
24 18 5200/126 6/2
24 18 450/minimum 6/2
24 24 9000/minimum 6/2
24 24 9000/minimum 6/2
128 128 1 2 1000/0
320 192 1
320 192 1
320 328 1.5
320 224 2
192 64 1 1000/0
192 64 1 2 1000/0
2000–2200
2000–2200
15.63
FLAIR
Prol-T2W imaging
FSE-T1W imaging
EPI-FLAIR
AE-FLAIR
27.78 2000–2200 24
31.25
31.25
24
20
3D TOF MRA, three-dimensional time-of-flight magnetic resonance angiography; AE-FLAIR, array spatial sensitivity technique-EPI-FLAIR; BW, bandwidth; DWI, diffusion-weighted imaging; EPI-FLAIR, echo-planar imaging-FLAIR; FLAIR, fluid attenuated inversion recovery; FSE-T1W, fast spin echoT1-weighted; NEX, number of excitations; asset R value, scan time reduction factor; TI, inversion time; TR/TE, repetition time/echo time.
in Table 1. For each sequence, 20 contiguous sections that covered the entire brain from the foramen magnum to the vertex were acquired.
Image analysis All datasets were transferred to a workstation. Image analysis was done by a fully trained neuroradiologist and a neurologist who were experienced in reading of brain images from patients with stroke. The two observers were masked to clinical information including the final extent of infarction. The readers were asked to grade FLAIR, EPI-FLAIR, and AE-FLAIR maps using a scale (range, 0–3) scoring system in terms of the presence of image motion artifacts, visualization of infarction, delineation of major structures, susceptibility mediated distortion at tissue interfaces: 0, poor image quality for diagnosis, not interpretable; 1, evident distortion, lesions with visible, limited detail of delineation of lesions with surrounding brain tissue, acceptable image quality for diagnosis; 2, minimal distortion with detailed delineation of all structures, moderate diagnostic image quality; and 3, excellent image quality for confident diagnosis and sharply defined borders, no distortion. A score of 2 or 3 was considered as overall adequate diagnostic image quality. The readers were also asked to assess the clinical value of the MRI results in the optimized group (AE-FLAIR and 3D TOF MRA) and the control group (T1W imaging, T2W imaging, conventional FLAIR, DWI, and 3D TOF). Individual observer’s judgments were used only for the calculation of inter-observer agreement.
Statistical evaluation All statistical analyses were performed with the SPSS 13.0 software (SPSS Inc., Chicago, IL, USA). A kappa coefficient (k) test was used to assess inter-observer agreement, and the median score of these two observers was used in the follow-up analysis. Good inter-observer agreement was noted between the two readers when k 0.75. A 2 test was carried out to determine differences in the ratio of good image quality (score of 2 or 3) in the optimized group and the control group. The differences with a value of P < 0.05 were regarded as statistically significant.
Results Baseline characteristics The age of these 100 patients (68 men, 32 women) was in the range of 44–82 years (mean age, 63 0.5 years). The NIHSS score was 25 3. The 100 patients were also categorized into hyperacute phase (n ¼ 18) and acute phase (n ¼ 82).
Clinical imaging findings and image quality As shown in Table 2, a total of 125 brain lesions (acute infarction) were identified by both observers in these 100 patients with conventional FLAIR, EPI-FLAIR, or AE-FLAIR. Among these 125 brain lesions, 20 lesions were in the fronto-parietal region, 22 in the temporo-parietal region, 18 in the parieto-occipital
Downloaded from acr.sagepub.com at Flinders University on February 8, 2016
4
Acta Radiologica 0(0)
Table 2. Summary of the results of the quantitative comparison between the cooperative group and the uncooperative group of patients.
Cooperative group (n ¼ 50)
Uncooperative group (n ¼ 50)
Total patients (n ¼ 100)
Sequences
Score of 3
Score of 2
Score of 1
Score of 0
Ratio of good image quality
Group comparison
2
df
P value
1. 2. 3. 1. 2. 3. 1. 2. 3.
62 8 2 5 10 4 67 18 6
5 55 40 20 38 32 25 93 72
3 7 26 18 6 12 21 13 38
0 0 2 12 1 7 1 1 9
96% 90% 60% 45% 87% 65% 73% 89% 62%
1 1 2 1 1 2 1 1 2
1.723 25.895 16.800 21.544 4.453 7.253 9.459 3.603 23.614
1 1 1 1 1 1 1 1 1
0.326 0.0005 0.0005 0.0005 0.055 0.013 0.003 0.078 0.0005
FLAIR AE-FLAIR EPI-FLAIR FLAIR AE-FLAIR EPI-FLAIR FLAIR AE-FLAIR EPI-FLAIR
and and and and and and and and and
2 3 3 2 3 3 2 3 3
AE-FLAIR, array spatial sensitivity technique-EPI-FLAIR; df, degrees of freedom; EPI-FLAIR, echoplanar imaging-FLAIR; FLAIR, fluid attenuated inversion recovery.
region, 30 in the corona radiata region, 14 in the pons, 11 in the cerebellum, and 10 in the temporal lobe. Images were rated as having adequate diagnostic image quality (score of 2 or 3) in 73% (92/125) of conventional FLAIR, 62% (78/125) of EPI-FLAIR, and 89% (101/125) of AE-FLAIR for both the group of cooperative patients and the group of uncooperative patients. In the uncooperative group of patients, the ratio of good image quality of the AE-FLAIR sequences (48/55, 87%) was higher than those of the conventional FLAIR (25/55, 45%) and EPI-FLAIR (36/55, 65%) sequences in all the cases (P < 0.05). There were no discrepancies between FLAIR, EPIFLAIR, and AE-FLAIR in identification of the anatomic location of infarctions (Figs. 1–3). Nevertheless, 28 lesions detected by conventional FLAIR sequences, 47 lesions by EPI-FLAIR sequences, and 14 lesions by AE-FLAIR sequences were deemed non-diagnostic (image quality
Original Article
The applied research of MRI with ASSET-EPI-FLAIR combined with 3D TOF MRA sequences in the assessment of patients with acute cerebral infarction
Acta Radiologica 0(0) 1–9 ! The Foundation Acta Radiologica 2016 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0284185116628338 acr.sagepub.com
Zhichao Lin1, Zexiong Guo2, Lin Qiu3, Wanyoug Yang1 and Mingxia Lin1
Abstract Background: To extend the time window for thrombolysis, reducing the time for diagnosis and detection of acute cerebral infarction seems to be warranted. Purpose: To evaluate the feasibility of implementing an array spatial sensitivity technique (ASSET)-echo-planar imaging (EPI)-fluid attenuated inversion recovery (FLAIR) (AE-FLAIR) sequence into an acute cerebral infarction magnetic resonance (MR) evaluation protocol, and to assess the diagnostic value of AE-FLAIR combined with three-dimensional time-of-flight MR angiography (3D TOF MRA). Material and Methods: A total of 100 patients (68 men, 32 women; age range, 44–82 years) with acute cerebral infarction, including 50 consecutive uncooperative and 50 cooperative patients, were evaluated with T1-weighted (T1W) imaging, T2-weighted (T2W) imaging, FLAIR, diffusion-weighted imaging (DWI), 3D TOF, EPI-FLAIR, and AE-FLAIR. Conventional FLAIR, EPI-FLAIR, and AE-FLAIR were assessed by two observers independently for image quality. The optimized group (AE-FLAIR and 3D TOF) and the control group (T1W imaging, T2W imaging, conventional FLAIR, DWI, and 3D TOF) were compared for evaluation time and diagnostic accuracy. Results: One hundred and twenty-five lesions were detected and images having adequate diagnostic image quality were in 73% of conventional FLAIR, 62% of EPI-FLAIR, and 89% of AE-FLAIR. The detection time was 12 1 min with 76% accuracy and 4 0.5 min with 100% accuracy in the control and the optimized groups, respectively. Inter-observer agreements of k ¼ 0.78 and k ¼ 0.81 were for the optimized group and control group, respectively. Conclusion: With reduced acquisition time and better image quality, AE-FLAIR combined with 3D TOF may be used as a rapid diagnosis tool in patients with acute cerebral infarction, especially in uncooperative patients.
Keywords Acute cerebral infarction, magnetic resonance imaging (MRI), three-dimensional time-of-flight magnetic resonance angiography (3D TOF MRA), image quality, fluid attenuated inversion recovery (FLAIR) Date received: 3 November 2015; accepted: 30 December 2015
Introduction Cerebral stroke is one of the most common causes of long-term disability or death worldwide and results in a marked burden in healthcare system (1). Typically, a stroke is caused by the blood supply reduction to the brain due to cerebral artery occlusion by a clot (2). This so-called ischemic stroke (cerebral infarction), which may result from cardiogenic, lacunar, arterosclerotic, hemodynamic, or cryptogenic sources, accounts for
1 Department of Radiology, the First Affiliated Hospital of Jinan University, Guangzhou, PR China 2 Department of Urology, the First Affiliated Hospital of Jinan University, Guangzhou, PR China 3 Department of Neurology, the First Affiliated Hospital of Jinan University, Guangzhou, PR China
Corresponding author: Zexiong Guo, Department of Urology, the First Affiliated Hospital of Jinan University, 613 West Huangpu Dadao, Guangzhou City, Guangdong Province, PR China. Email: [email protected]
Downloaded from acr.sagepub.com at Flinders University on February 8, 2016
2
Acta Radiologica 0(0)
about 80% of strokes (3). The minority of strokes is related to disruption of a cerebral artery resulting in intracerebral hemorrhage (2). Currently, early recanalization of an occluded cerebral artery with use of a neurothrombectomy device or thrombolytic drug to salvage ischemic tissue at risk of infarction and decrease the size of the infarct may lead to a better functional outcome of patients with acute ischemic stroke (4). Thus, to extend the time window for thrombolysis, reducing the time for diagnosis and detection of acute cerebral infarction seems to be warranted. Neuroimaging has been demonstrated to play a significant role in the evaluation of patients with acute cerebral infarction (5). In order to better investigate anatomical details and pathologic lesions in the brain, routine magnetic resonance imaging (MRI) protocols include various types of sequences such as T1-weighted (T1W) and T2-weighted (T2W) imaging (5). However, adequate image quality is often impaired by motion artifacts which may sometimes obscure or mimic pathology in uncooperative patients with unconsciousness or respiratory failure symptoms (5,6). Because of longer acquisition time and limited availability, MRI has been mainly used in comprehensive stroke centers (7). With recent advances in MR technology, a comprehensive MR protocol including parenchymal imaging (diffusion-weighted imaging [DWI], gradient recalled-echo [GRE], fluid attenuated inversion recovery [FLAIR]), MR angiography (MRA), and MR perfusion can be obtained in 20 min as demonstrated in several clinical trials (5,8). The tissue contrast of DWI aids in the assessment of alterations of water diffusion related to the cytotoxic edema in the acute phase of brain ischemia (9). Besides, three-dimensional time-offlight MRA (3D TOF MRA), an available noninvasive technique, is a commonly used pulse sequence in the MR evaluation of intracranial arteries (10). FLAIR imaging is a T2W imaging sequence and is an indispensable part of common multi-parametric stroke MRI protocols (9). Recently, introduction of fast imaging techniques, such as parallel acquisition and echo-planar imaging (EPI), has obviously enhanced the performance of MRI with regard to acquisition speed (5,7). For instance, a study has demonstrated that the application of EPI to the FLAIR sequence (EPI-FLAIR) was feasible with comparable quantitative and qualitative results to the conventional FLAIR and resulted in reduced acquisition time (5). The reduction of acquisition time may prove useful for examining the uncooperative, medically unstable, or claustrophobic patients. However, few studies have investigated the ability of the application of both EPI and parallel acquisition techniques to the FLAIR sequence in evaluation of patients with acute cerebral infarction.
Parallel acquisition imaging spatial encoding from coil elements of the receiver array to compress the gradient encoding steps and hence accelerate acquisition (11). In this study, array spatial sensitivity technique (ASSET) parallel acquisition, a sensitivity encoding (SENSE)-based algorithm, was employed. The purpose of this study was to evaluate the feasibility of implementing an ASSET-EPI-FLAIR (AE-FLAIR) sequence into an acute cerebral infarction MR evaluation protocol, to assess the diagnostic value of AEFLAIR combined with 3D TOF MRA, and to compare with the results with conventional MR protocol.
Material and Methods Patients From March 2010 and March 2015, 100 consecutive patients with acute ischemic stroke who had undergone computed tomography (CT) to rule out intracranial hemorrhage were recruited for this study. Consecutive selection of patients was performed in this study, and these 100 patients included 50 cooperative and 50 uncooperative patients. All these patients met the following inclusion criteria: age >40 years; within 72 h after onset of ictus; in accordance with the Chinese guidelines of diagnosis and treatment for acute ischemic stroke (2010) (12) and having ruled out intracranial hemorrhage by CT evaluation; and a National Institutes of Health Stroke Scale (NIHSS) score at admission (>20). The exclusion criteria were as follows: intracranial hemorrhage; hypoxic-ischemic encephalopathy; toxication; hypoglycemia; hypotension; hepatic encephalopathy; pulmonary encephalopathy; other intracephalic diseases such as encephalitis; tumor; and parasitosis. Besides, the clinical stages of these 100 patients with acute ischemic stroke were determined as previously described (13). Additionally, this study was approved by the local institutional review committee, and written informed consent from the patients or the closest family member of the patients was obtained before the start of the study.
MRI protocol All MRI scans were performed on a 1.5-T MR scanner (Optima, MR360; GE Healthcare, Beijing, China) retrofitted for EPI and ASSET capabilities with a standard circularly polarized head coil. The following sequences were performed in all participants (the cooperative group and the uncooperative group of patients) in an axial plane: DWI, 3D TOF MRA, conventional FLAIR, Prol-T2W imaging, fast spin echo (FSE)-T1W imaging, EPI-FLAIR, and AE-FLAIR. The parameters used for these sequences were presented
Downloaded from acr.sagepub.com at Flinders University on February 8, 2016
Lin et al.
3
Table 1. Summary of imaging parameters for the sequences that were applied for MRI scans
Parameter
DWI
3D TOF MRA
Field of view (cm) TR/TE (ms) Slice thickness/inter-slice gap (mm) Matrix size NEX Asset R value b (s/mm2) BW TI (ms) Echo train length
24 24 5000/minimum 6/2
22 18 28/6.9 1.2/0
24 18 8600/140 6/2
24 18 5200/126 6/2
24 18 450/minimum 6/2
24 24 9000/minimum 6/2
24 24 9000/minimum 6/2
128 128 1 2 1000/0
320 192 1
320 192 1
320 328 1.5
320 224 2
192 64 1 1000/0
192 64 1 2 1000/0
2000–2200
2000–2200
15.63
FLAIR
Prol-T2W imaging
FSE-T1W imaging
EPI-FLAIR
AE-FLAIR
27.78 2000–2200 24
31.25
31.25
24
20
3D TOF MRA, three-dimensional time-of-flight magnetic resonance angiography; AE-FLAIR, array spatial sensitivity technique-EPI-FLAIR; BW, bandwidth; DWI, diffusion-weighted imaging; EPI-FLAIR, echo-planar imaging-FLAIR; FLAIR, fluid attenuated inversion recovery; FSE-T1W, fast spin echoT1-weighted; NEX, number of excitations; asset R value, scan time reduction factor; TI, inversion time; TR/TE, repetition time/echo time.
in Table 1. For each sequence, 20 contiguous sections that covered the entire brain from the foramen magnum to the vertex were acquired.
Image analysis All datasets were transferred to a workstation. Image analysis was done by a fully trained neuroradiologist and a neurologist who were experienced in reading of brain images from patients with stroke. The two observers were masked to clinical information including the final extent of infarction. The readers were asked to grade FLAIR, EPI-FLAIR, and AE-FLAIR maps using a scale (range, 0–3) scoring system in terms of the presence of image motion artifacts, visualization of infarction, delineation of major structures, susceptibility mediated distortion at tissue interfaces: 0, poor image quality for diagnosis, not interpretable; 1, evident distortion, lesions with visible, limited detail of delineation of lesions with surrounding brain tissue, acceptable image quality for diagnosis; 2, minimal distortion with detailed delineation of all structures, moderate diagnostic image quality; and 3, excellent image quality for confident diagnosis and sharply defined borders, no distortion. A score of 2 or 3 was considered as overall adequate diagnostic image quality. The readers were also asked to assess the clinical value of the MRI results in the optimized group (AE-FLAIR and 3D TOF MRA) and the control group (T1W imaging, T2W imaging, conventional FLAIR, DWI, and 3D TOF). Individual observer’s judgments were used only for the calculation of inter-observer agreement.
Statistical evaluation All statistical analyses were performed with the SPSS 13.0 software (SPSS Inc., Chicago, IL, USA). A kappa coefficient (k) test was used to assess inter-observer agreement, and the median score of these two observers was used in the follow-up analysis. Good inter-observer agreement was noted between the two readers when k 0.75. A 2 test was carried out to determine differences in the ratio of good image quality (score of 2 or 3) in the optimized group and the control group. The differences with a value of P < 0.05 were regarded as statistically significant.
Results Baseline characteristics The age of these 100 patients (68 men, 32 women) was in the range of 44–82 years (mean age, 63 0.5 years). The NIHSS score was 25 3. The 100 patients were also categorized into hyperacute phase (n ¼ 18) and acute phase (n ¼ 82).
Clinical imaging findings and image quality As shown in Table 2, a total of 125 brain lesions (acute infarction) were identified by both observers in these 100 patients with conventional FLAIR, EPI-FLAIR, or AE-FLAIR. Among these 125 brain lesions, 20 lesions were in the fronto-parietal region, 22 in the temporo-parietal region, 18 in the parieto-occipital
Downloaded from acr.sagepub.com at Flinders University on February 8, 2016
4
Acta Radiologica 0(0)
Table 2. Summary of the results of the quantitative comparison between the cooperative group and the uncooperative group of patients.
Cooperative group (n ¼ 50)
Uncooperative group (n ¼ 50)
Total patients (n ¼ 100)
Sequences
Score of 3
Score of 2
Score of 1
Score of 0
Ratio of good image quality
Group comparison
2
df
P value
1. 2. 3. 1. 2. 3. 1. 2. 3.
62 8 2 5 10 4 67 18 6
5 55 40 20 38 32 25 93 72
3 7 26 18 6 12 21 13 38
0 0 2 12 1 7 1 1 9
96% 90% 60% 45% 87% 65% 73% 89% 62%
1 1 2 1 1 2 1 1 2
1.723 25.895 16.800 21.544 4.453 7.253 9.459 3.603 23.614
1 1 1 1 1 1 1 1 1
0.326 0.0005 0.0005 0.0005 0.055 0.013 0.003 0.078 0.0005
FLAIR AE-FLAIR EPI-FLAIR FLAIR AE-FLAIR EPI-FLAIR FLAIR AE-FLAIR EPI-FLAIR
and and and and and and and and and
2 3 3 2 3 3 2 3 3
AE-FLAIR, array spatial sensitivity technique-EPI-FLAIR; df, degrees of freedom; EPI-FLAIR, echoplanar imaging-FLAIR; FLAIR, fluid attenuated inversion recovery.
region, 30 in the corona radiata region, 14 in the pons, 11 in the cerebellum, and 10 in the temporal lobe. Images were rated as having adequate diagnostic image quality (score of 2 or 3) in 73% (92/125) of conventional FLAIR, 62% (78/125) of EPI-FLAIR, and 89% (101/125) of AE-FLAIR for both the group of cooperative patients and the group of uncooperative patients. In the uncooperative group of patients, the ratio of good image quality of the AE-FLAIR sequences (48/55, 87%) was higher than those of the conventional FLAIR (25/55, 45%) and EPI-FLAIR (36/55, 65%) sequences in all the cases (P < 0.05). There were no discrepancies between FLAIR, EPIFLAIR, and AE-FLAIR in identification of the anatomic location of infarctions (Figs. 1–3). Nevertheless, 28 lesions detected by conventional FLAIR sequences, 47 lesions by EPI-FLAIR sequences, and 14 lesions by AE-FLAIR sequences were deemed non-diagnostic (image quality
