- Email: [email protected]
MRI scout images can detect the acute intracerebral hemorrhage on CT
Journal of the Neurological Sciences 387 (2018) 147–149
Contents lists available at ScienceDirect
Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns
MRI scout images can detect the acute intracerebral hemorrhage on CT ⁎
T
Toshiyuki Hayashi , Junya Aoki, Kentaro Suzuki, Yuki Sakamoto, Satoshi Suda, Seiji Okubo, Masahiro Mishina, Kazumi Kimura Department of Neurological Science, Graduate School of Medicine, Nippon Medical School, Japan
A R T I C L E I N F O
A B S T R A C T
Keywords: MRI Scout images Intracerebral hemorrhage
Introduction: Magnetic resonance imaging (MRI) has recently emerged as a first-line tool for investigating acute stroke. However, MRI requires long scan times, which could be detrimental for severe stroke patients with a large intracerebral hemorrhage (ICH). MRI scout images, which are taken prior to a study to determine the range of subsequent images, can be used to rapidly screen the whole brain. We examined whether MRI scout imaging can detect ICHs observed by computed tomography (CT). Methods: Between September 2014 and March 2016, consecutive acute ICH patients who underwent both MRI scout and CT imaging in the acute setting were studied. ICHs on MRI scout images were defined as spaceoccupying lesions. Two neurologists independently assessed the scout images. We investigated whether ICHs on CT scans can be detected on MRI scout images and the characteristics of ICHs not detected by MRI scout images. Results: One hundred and forty-eight ICH patients (median age, 68 [interquartile range, 59–77] years; 99 [67%] males; median National Institutes of Health Stroke Scale score, 11 [4–17]) were enrolled. Among these, 138 (93%) patients were diagnosed as having ICH by MRI scout imaging (positive group), and 10 (7%) patients were not (negative group). The bleeding volume was 9.3 [4.5–22.4] ml in the positive group and 1.0 [0.4–2.0] ml in the negative group (p < .001). The cut-off value of bleeding volume calculated from the receiver operating characteristic curve was 2.0 ml. Regarding ICH lesions, 4 (44%) of the 9 pontine hemorrhages were detected on MRI scout images, whereas 134 (96%) of the 139 other hemorrhages were diagnosed (p < .001). Conclusions: We diagnosed > 90% of ICHs using MRI scout images. Low levels of ICH and pontine hemorrhaging might be difficult to detect using MRI scout imaging.
1. Introduction
MRI scout imaging.
Clinical use of computed tomography (CT) and magnetic resonance imaging (MRI) began in the 1970s and 1980s, respectively. Le Bihan [1] first reported diffusion-weighted imaging (DWI), and the diagnostic accuracy of acute stroke imaging has since improved dramatically. However, the common MRI test for patients with stroke includes sequences such as DWI, fluid-attenuated inversion recovery, and magnetic resonance angiography, which require long scan times that could be detrimental for patients with severe intracerebral hemorrhage (ICH). MRI scout images, which are acquired preliminarily to determine the range of subsequent images, can be used to rapidly screen the whole brain. It takes less than one minute and shorter than DWI. Although no studies have reported the use of MRI scout imaging for diagnosis, if MRI scout images have a high diagnostic accuracy for ICHs detected by CT, we can stop to obtain additional MRI sequences for patients with severe ICH, thus shortening the overall scan time and increasing the safety of MRI. We therefore investigated whether ICH can be diagnosed using
2. Methods
⁎
Data on consecutive patients with acute ICH who were admitted to our stroke center between September 2014 and March 2016 were retrospectively analyzed. This study is approved by the institutional review board at Nippon Medical School. Only patients with both MRI and CT images acquired in our hospital were studied. ICH on an MRI scout image was defined as a space-occupying lesion (Fig. 1). Two neurologists (T.H and J.A), who were blinded to the bleeding lesions of all patients and with knowledge of their diagnoses, retrospectively and independently reviewed all MRI scout images and judged whether and where MRI scout images demonstrated space-occupying lesions. Only when the judgment regarding MRI scout images matched the CT findings was it said to positively diagnose the ICH. When the two evaluators returned different assessments, the final judgment was made after discussion.
Corresponding author at: Department of Neurological Science, Nippon Medical School, Graduate School of Medicine, 1-1-5, Sendagi, Bunkyo-ku, Tokyo 113-8602, Japan. E-mail address: [email protected] (T. Hayashi).
https://doi.org/10.1016/j.jns.2018.01.041 Received 19 October 2017; Received in revised form 24 January 2018; Accepted 31 January 2018 Available online 02 February 2018 0022-510X/ © 2018 Elsevier B.V. All rights reserved.
Journal of the Neurological Sciences 387 (2018) 147–149
T. Hayashi et al.
Table 1 Clinical characteristics compared between patients in MRI scout images positive and negative group.
Age, y, (IQR) Men, n (%) NIHSS score (IQR) Time from onset to images, h (IQR) Bleeding volume, ml (IQR) Bleeding regions, n (%) Thalamus Putamen Pons Subcortical Cerebellum
Group with MRI scout positive
Group with MRI scout negative
p
n = 138
n = 10
68(58–77) 94(64%) 11(4–18) 14.0(1.9–86.0)
72(61–88) 5(50%) 2(1−12) 72.0(39.1–101.1)
0.160 0.300 0.003 0.031
9.3(4.5–22.4)
1.0(0.4–2.0)
< 0.001
52(38%) 54(39%) 4(3%) 24(17%) 4(3%)
2(20%) 1(10%) 5(50%) 1(10%) 1(10%)
0.320 0.091 < 0.001 1.000 0.290
IQR, indicates interquartile range; NIHSS, National Institutes of Health Stroke Scale.
4. Results Between September 2014 and March 2016, 169 consecutive acute ICH patients were admitted to our stroke center. Of these, 148 (median age, 68 [IQR, 59–77] years; 99 [67%] males; median NIHSS score, 11 [4–17]) patients who had undergone both MRI and CT were enrolled to the present study. The etiology of ICH includes 138 patients with hypertension, 7 patients with amyloid angiopathy, two patients with cavernous hemangioma, and one patients with sinus thrombosis. One neurologist (T.H.) assessed 133 patients as positive, and another neurologist (J.A,) assessed 138 patients as positive. After discussion, 138 patients were classified in the positive group, and 10 patients were classified in the negative group. Interrater agreement was 97%. The interrater reliability for the evaluators was found to be Kappa = 0.78 (p < .001). Clinical characteristics and examination findings based on MRI scout images are summarized in Table 1. The NIHSS score (11 [4–18] vs. 2 [1−12]; p = .003) and bleeding volume (9.3 [4.5–22.4] ml vs. 1.0 [0.4–2.0] ml; p < .001) were higher in the positive group than the negative group. Time from onset to imaging was 14.0 (1.9–86.0) hours in the positive group and 72.0 (39.1–101.0) hours in the negative group (p = .031). Regarding bleeding regions, pontine hemorrhages were less frequent in the positive group than the negative group (4 [3%] vs. 5 [50%]; p < .001). Regarding continuous parameters, ROC analysis showed that NIHSS score on admission (area under the curve [AUC] = 0.757; p = .010) and bleeding volume (AUC = 0.942; p < .001) were significantly associated with MRI scout positive ICH finding (Fig. 2). Cut-off values were calculated as bleeding volume of 2 ml, with a sensitivity of 0.99, specificity of 0.50, positive predictive value of 0.93, and negative predictive value of 0.90; and NIIHSS score of 3 with sensitivity of 0.97, specificity of 0.30, positive predictive value of 0.88, and negative predictive value of 0.70 (Table 2).
Fig. 1. Representative case with intracerebral hemorrhage. Left putaminal hemorrhage in CT (A) can be recognized as a space-occupying lesion in MRI scout images (B). The time from onset to the CT and MRI scans are 2.3 h and 1.9 h. By contrast, a pontine hemorrhage observed in a CT scan (C) could not be detected in an MRI scout image (D). The time from onset to the CT and MRI scans are 73.0 h and 74.9 h.
The following clinical information was obtained: age; sex; time from onset to MRI scout and CT imaging; National Institutes of Health Stroke Scale (NIHSS) on admission; bleeding volume; bleeding regions. Bleeding regions were classified as thalamus, putamen, pons, subcortical, and cerebellum. MRI scout images were acquired on a 1.5-Tesla scanner (Echelon Oval; Hitachi, Tokyo, Japan). Scout images were obtained using the following parameters: repetition time, 30 ms; echo time, 1.7 ms; field of view, 100 cm; acquisition matrix and slice thickness, 10 mm. The number of slice is 5 in each three orientations (horizontal, coronal, and sagittal). CT scans were carried out with a 5-mm slice thickness, without contrast enhancement, at 120 kV and 100 mA, filmed at an appropriate window width and level setting of 100/40 (LightSpeed VCT; GE Healthcare, Little Chalfont, UK). 3. Statistical analysis All patients were initially classified into positive and negative groups based on MRI scout and CT images. Clinical information and imaging findings were compared between the two groups. Next, regarding parameters with p < .05, receiver operating characteristic (ROC) curves were drawn to calculate optimal cut-off values to differentiate patients in the positive and negative groups. The sensitivity, specificity, positive predictive value, and negative predictive value of each cut-off value for differentiating the positive group were then calculated. Interrater agreement (percentage) and reliability analysis (Kappa statistic) of scout findings were also performed to determine consistency between two evaluators. The Mann-Whitney U test was used to analyze differences in continuous variables, and the χ2 test was used to analyze differences in categorical variables. Data are presented as median values (interquartile range [IQR]) or frequencies (%). All statistical analyses were performed using IBM SPSS Statistics 22. Results were considered significant at p < .05.
5. Discussion We found that 93% of acute ICH patients could be diagnosed as having ICH based on MRI scout images. By contrast, ICHs with a low NIHSS score and low volume, as well as those located in the pons, were difficult to diagnose using MRI scout imaging. We conclude that the volume of the ICH is the factor that determines detectability of ICHs on MRI scout images. Generally, low NIHSS scores are indicative of small ICHs. The volume of pontine hemorrhages is also low, though the NIHSS score is high. A bleeding volume of 2 ml was determined as the 148
Journal of the Neurological Sciences 387 (2018) 147–149
T. Hayashi et al.
patients suspected of having temporomandibular joint arthrosis based on MRI scout images and indicated that scout images should be carefully examined for not only positioning the slice but also for diagnosis. In that study, the MRI examination area was limited to the temporomandibular joint area. However, scout imaging can screen the whole brain, thus improving detectability. MRI scout imaging enabled the detection of a 20-mm-diameter neoplastic lesion of approximately 4.2 ml. This space volume corresponded approximately to our cut-off bleeding volume of 2 ml, which can be used for the diagnosis of ICH based on the high-sensitivity of MRI scout imaging. We believe that the characteristic features of MRI scout images that show the whole brain in 3 orientations can be helpful for the diagnosis of lesions > 2 ml. There are several limitations to the present study; namely, this was a single-center observational study, the sample size was small, and the mass lesions seen on the MRI scout images were not specific for ICH. Second, in this study, we focused only on acute ICH patients, and the imaging characteristics of other intracerebral diseases on scout images are not fully understood. The repetition and the echo times of scout imaging resemble those of T1-weighted imaging. Thus, subacute or chronic ischemic lesions, which also show as hypointense on T1 could potentially be difficult to distinguish from ICH on scout images. However, hypointense signal change due to ICH on scout images is clearly visible and marked by sharp margins, which may be due to the edema, and this enables the differentiation of ICH from subacute or chronic infarcts, especially when they are large in size. In this study we evaluated the detection of ICH from scout images as these are a prerequisite in any MRI scan. The detection of ICH from scout images should be compared to other sequences with short acquisition times, such as a DWI, which can be acquired in under 2 min. This could provide further information on early diagnosis on MRI. Further larger studies are needed to confirm our findings. In conclusion, > 90% of acute ICH patients can be diagnosed using MRI scout imaging. Attention to symptoms originating in the brainstem can increase the safety of MRI as a first-line diagnostic test for stroke.
Fig. 2. Receiver operating characteristic curve. NIHSS score on admission (AUC = 0.757; p = .010) and bleeding volume (AUC = 0.942; p < .001) were significantly associated with diagnostic ability of MRI scout imaging for ICH, whereas age (AUC = 0.351; p = .134) and time from onset to imaging (AUC = 0.291; p = .360) were not.
Table 2 Patients redistributed using cut-off value of bleeding volume and NIHSS score.
Bleeding volume ≥2 ml < 2 ml NIHSS score ≥3 <3
Group with MRI scout positive n = 138
Group with MRI scout negative n = 10
Total n = 148
129 (93%) 9 (7%)
1 (10%) 9 (90%)
130 (88%) 18 (12%)
122 (88%) 16 (12%)
3 (30%) 7 (70%)
125 (84%) 23 (16%)
cut-off value for detecting an ICH. Although MRI scout images are acquired in 3 different orientations (horizontal, coronal, and sagittal), they are acquired for subsequent positioning so that the slice thickness will not be thinner than in CT imaging. Therefore, low-volume ICHs could be difficult to detect given the limited range of MRI scout images, although they can be easily detected by general MRI and CT imaging. With an understanding of its advantages and disadvantages, we think MRI scout imaging can be a suitable diagnostic tool. A number of specialized stroke centers currently utilize MRI as the first-line imaging tool for patients suspected of having a stroke [2–4]. The advantage of MRI in patients with acute ischemic stroke is its high diagnostic accuracy. Within 3 h after stroke onset, over 80% of patients can be diagnosed based on diffusion-weighted images [5]. However, MRI has the disadvantage that it prevents physicians from detecting worsening conditions in patients, including respiratory status and vomiting, due to its long time requirement and the closed space in which patients are placed. If it is possible to diagnose severe ICH patients, who tend to get worse over time, we can stop to obtain additional MRI sequences, providing shortening the scan time and increasing the safety of MRI. Further study is expected to confirm our study's findings. To date, no previous reports have described the use of MRI scout imaging for diagnosis, although it is discussed in some case reports. Yanagi [6] reported the incidental discovery of tumors in 2 of 2776
Disclosures None References [1] D. Le Bihan, E. Breton, Imagerie de diffusion in vivo par resonance magnétique nucléaire, C R Acad. Sci. Paris T. 301 (1995) 1109–1112. [2] B.T. Buckley, A. Wainwright, T. Meagher, D. Briley, Audit of a policy of magnetic resonance imaging with diffusion-weighted imaging as first-line neuroimaging for inpatients with clinically suspected acute stroke, Clin. Radiol. 58 (2003) 234–237. [3] P.L. Tan, D. King, C.J. Durkin, T.M. Meagher, D. Briley, Diffusion weighted magnetic resonance imaging for acute stroke: practical and popular, Postgrad. Med. J. 82 (2006) 289–292. [4] R.R. Leker, G. Keigler, R. Eichel, T. Ben Hur, J.M. Gomori, J.E. Cohen, Should DWI MRI be the primary screening test for stroke? Int. J. Stroke 9 (2014) 696–697. [5] Junya Aoki, Kazumi Kimura, Yasuyuki Iguchi, Kensaku Shibazaki, Kenichiro Sakai, Takeshi Iwanaga, FLAIR can estimate the onset time in acute ischemic stroke patients, J. Neurol. Sci. 293 (2010) 39–44. [6] Yoshinobu Yanagi, Junichi Asaumi, Yuu Maki, Jun Murakami, Miki Hisatomi, Hidenobu Matsuzaki, et al., Incidentally found and unexpected tumors discovered by MRI examination for temporomandibular joint arthrosis, Eur. J. Radiol. 47 (2003) 6–9.
149
Contents lists available at ScienceDirect
Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns
MRI scout images can detect the acute intracerebral hemorrhage on CT ⁎
T
Toshiyuki Hayashi , Junya Aoki, Kentaro Suzuki, Yuki Sakamoto, Satoshi Suda, Seiji Okubo, Masahiro Mishina, Kazumi Kimura Department of Neurological Science, Graduate School of Medicine, Nippon Medical School, Japan
A R T I C L E I N F O
A B S T R A C T
Keywords: MRI Scout images Intracerebral hemorrhage
Introduction: Magnetic resonance imaging (MRI) has recently emerged as a first-line tool for investigating acute stroke. However, MRI requires long scan times, which could be detrimental for severe stroke patients with a large intracerebral hemorrhage (ICH). MRI scout images, which are taken prior to a study to determine the range of subsequent images, can be used to rapidly screen the whole brain. We examined whether MRI scout imaging can detect ICHs observed by computed tomography (CT). Methods: Between September 2014 and March 2016, consecutive acute ICH patients who underwent both MRI scout and CT imaging in the acute setting were studied. ICHs on MRI scout images were defined as spaceoccupying lesions. Two neurologists independently assessed the scout images. We investigated whether ICHs on CT scans can be detected on MRI scout images and the characteristics of ICHs not detected by MRI scout images. Results: One hundred and forty-eight ICH patients (median age, 68 [interquartile range, 59–77] years; 99 [67%] males; median National Institutes of Health Stroke Scale score, 11 [4–17]) were enrolled. Among these, 138 (93%) patients were diagnosed as having ICH by MRI scout imaging (positive group), and 10 (7%) patients were not (negative group). The bleeding volume was 9.3 [4.5–22.4] ml in the positive group and 1.0 [0.4–2.0] ml in the negative group (p < .001). The cut-off value of bleeding volume calculated from the receiver operating characteristic curve was 2.0 ml. Regarding ICH lesions, 4 (44%) of the 9 pontine hemorrhages were detected on MRI scout images, whereas 134 (96%) of the 139 other hemorrhages were diagnosed (p < .001). Conclusions: We diagnosed > 90% of ICHs using MRI scout images. Low levels of ICH and pontine hemorrhaging might be difficult to detect using MRI scout imaging.
1. Introduction
MRI scout imaging.
Clinical use of computed tomography (CT) and magnetic resonance imaging (MRI) began in the 1970s and 1980s, respectively. Le Bihan [1] first reported diffusion-weighted imaging (DWI), and the diagnostic accuracy of acute stroke imaging has since improved dramatically. However, the common MRI test for patients with stroke includes sequences such as DWI, fluid-attenuated inversion recovery, and magnetic resonance angiography, which require long scan times that could be detrimental for patients with severe intracerebral hemorrhage (ICH). MRI scout images, which are acquired preliminarily to determine the range of subsequent images, can be used to rapidly screen the whole brain. It takes less than one minute and shorter than DWI. Although no studies have reported the use of MRI scout imaging for diagnosis, if MRI scout images have a high diagnostic accuracy for ICHs detected by CT, we can stop to obtain additional MRI sequences for patients with severe ICH, thus shortening the overall scan time and increasing the safety of MRI. We therefore investigated whether ICH can be diagnosed using
2. Methods
⁎
Data on consecutive patients with acute ICH who were admitted to our stroke center between September 2014 and March 2016 were retrospectively analyzed. This study is approved by the institutional review board at Nippon Medical School. Only patients with both MRI and CT images acquired in our hospital were studied. ICH on an MRI scout image was defined as a space-occupying lesion (Fig. 1). Two neurologists (T.H and J.A), who were blinded to the bleeding lesions of all patients and with knowledge of their diagnoses, retrospectively and independently reviewed all MRI scout images and judged whether and where MRI scout images demonstrated space-occupying lesions. Only when the judgment regarding MRI scout images matched the CT findings was it said to positively diagnose the ICH. When the two evaluators returned different assessments, the final judgment was made after discussion.
Corresponding author at: Department of Neurological Science, Nippon Medical School, Graduate School of Medicine, 1-1-5, Sendagi, Bunkyo-ku, Tokyo 113-8602, Japan. E-mail address: [email protected] (T. Hayashi).
https://doi.org/10.1016/j.jns.2018.01.041 Received 19 October 2017; Received in revised form 24 January 2018; Accepted 31 January 2018 Available online 02 February 2018 0022-510X/ © 2018 Elsevier B.V. All rights reserved.
Journal of the Neurological Sciences 387 (2018) 147–149
T. Hayashi et al.
Table 1 Clinical characteristics compared between patients in MRI scout images positive and negative group.
Age, y, (IQR) Men, n (%) NIHSS score (IQR) Time from onset to images, h (IQR) Bleeding volume, ml (IQR) Bleeding regions, n (%) Thalamus Putamen Pons Subcortical Cerebellum
Group with MRI scout positive
Group with MRI scout negative
p
n = 138
n = 10
68(58–77) 94(64%) 11(4–18) 14.0(1.9–86.0)
72(61–88) 5(50%) 2(1−12) 72.0(39.1–101.1)
0.160 0.300 0.003 0.031
9.3(4.5–22.4)
1.0(0.4–2.0)
< 0.001
52(38%) 54(39%) 4(3%) 24(17%) 4(3%)
2(20%) 1(10%) 5(50%) 1(10%) 1(10%)
0.320 0.091 < 0.001 1.000 0.290
IQR, indicates interquartile range; NIHSS, National Institutes of Health Stroke Scale.
4. Results Between September 2014 and March 2016, 169 consecutive acute ICH patients were admitted to our stroke center. Of these, 148 (median age, 68 [IQR, 59–77] years; 99 [67%] males; median NIHSS score, 11 [4–17]) patients who had undergone both MRI and CT were enrolled to the present study. The etiology of ICH includes 138 patients with hypertension, 7 patients with amyloid angiopathy, two patients with cavernous hemangioma, and one patients with sinus thrombosis. One neurologist (T.H.) assessed 133 patients as positive, and another neurologist (J.A,) assessed 138 patients as positive. After discussion, 138 patients were classified in the positive group, and 10 patients were classified in the negative group. Interrater agreement was 97%. The interrater reliability for the evaluators was found to be Kappa = 0.78 (p < .001). Clinical characteristics and examination findings based on MRI scout images are summarized in Table 1. The NIHSS score (11 [4–18] vs. 2 [1−12]; p = .003) and bleeding volume (9.3 [4.5–22.4] ml vs. 1.0 [0.4–2.0] ml; p < .001) were higher in the positive group than the negative group. Time from onset to imaging was 14.0 (1.9–86.0) hours in the positive group and 72.0 (39.1–101.0) hours in the negative group (p = .031). Regarding bleeding regions, pontine hemorrhages were less frequent in the positive group than the negative group (4 [3%] vs. 5 [50%]; p < .001). Regarding continuous parameters, ROC analysis showed that NIHSS score on admission (area under the curve [AUC] = 0.757; p = .010) and bleeding volume (AUC = 0.942; p < .001) were significantly associated with MRI scout positive ICH finding (Fig. 2). Cut-off values were calculated as bleeding volume of 2 ml, with a sensitivity of 0.99, specificity of 0.50, positive predictive value of 0.93, and negative predictive value of 0.90; and NIIHSS score of 3 with sensitivity of 0.97, specificity of 0.30, positive predictive value of 0.88, and negative predictive value of 0.70 (Table 2).
Fig. 1. Representative case with intracerebral hemorrhage. Left putaminal hemorrhage in CT (A) can be recognized as a space-occupying lesion in MRI scout images (B). The time from onset to the CT and MRI scans are 2.3 h and 1.9 h. By contrast, a pontine hemorrhage observed in a CT scan (C) could not be detected in an MRI scout image (D). The time from onset to the CT and MRI scans are 73.0 h and 74.9 h.
The following clinical information was obtained: age; sex; time from onset to MRI scout and CT imaging; National Institutes of Health Stroke Scale (NIHSS) on admission; bleeding volume; bleeding regions. Bleeding regions were classified as thalamus, putamen, pons, subcortical, and cerebellum. MRI scout images were acquired on a 1.5-Tesla scanner (Echelon Oval; Hitachi, Tokyo, Japan). Scout images were obtained using the following parameters: repetition time, 30 ms; echo time, 1.7 ms; field of view, 100 cm; acquisition matrix and slice thickness, 10 mm. The number of slice is 5 in each three orientations (horizontal, coronal, and sagittal). CT scans were carried out with a 5-mm slice thickness, without contrast enhancement, at 120 kV and 100 mA, filmed at an appropriate window width and level setting of 100/40 (LightSpeed VCT; GE Healthcare, Little Chalfont, UK). 3. Statistical analysis All patients were initially classified into positive and negative groups based on MRI scout and CT images. Clinical information and imaging findings were compared between the two groups. Next, regarding parameters with p < .05, receiver operating characteristic (ROC) curves were drawn to calculate optimal cut-off values to differentiate patients in the positive and negative groups. The sensitivity, specificity, positive predictive value, and negative predictive value of each cut-off value for differentiating the positive group were then calculated. Interrater agreement (percentage) and reliability analysis (Kappa statistic) of scout findings were also performed to determine consistency between two evaluators. The Mann-Whitney U test was used to analyze differences in continuous variables, and the χ2 test was used to analyze differences in categorical variables. Data are presented as median values (interquartile range [IQR]) or frequencies (%). All statistical analyses were performed using IBM SPSS Statistics 22. Results were considered significant at p < .05.
5. Discussion We found that 93% of acute ICH patients could be diagnosed as having ICH based on MRI scout images. By contrast, ICHs with a low NIHSS score and low volume, as well as those located in the pons, were difficult to diagnose using MRI scout imaging. We conclude that the volume of the ICH is the factor that determines detectability of ICHs on MRI scout images. Generally, low NIHSS scores are indicative of small ICHs. The volume of pontine hemorrhages is also low, though the NIHSS score is high. A bleeding volume of 2 ml was determined as the 148
Journal of the Neurological Sciences 387 (2018) 147–149
T. Hayashi et al.
patients suspected of having temporomandibular joint arthrosis based on MRI scout images and indicated that scout images should be carefully examined for not only positioning the slice but also for diagnosis. In that study, the MRI examination area was limited to the temporomandibular joint area. However, scout imaging can screen the whole brain, thus improving detectability. MRI scout imaging enabled the detection of a 20-mm-diameter neoplastic lesion of approximately 4.2 ml. This space volume corresponded approximately to our cut-off bleeding volume of 2 ml, which can be used for the diagnosis of ICH based on the high-sensitivity of MRI scout imaging. We believe that the characteristic features of MRI scout images that show the whole brain in 3 orientations can be helpful for the diagnosis of lesions > 2 ml. There are several limitations to the present study; namely, this was a single-center observational study, the sample size was small, and the mass lesions seen on the MRI scout images were not specific for ICH. Second, in this study, we focused only on acute ICH patients, and the imaging characteristics of other intracerebral diseases on scout images are not fully understood. The repetition and the echo times of scout imaging resemble those of T1-weighted imaging. Thus, subacute or chronic ischemic lesions, which also show as hypointense on T1 could potentially be difficult to distinguish from ICH on scout images. However, hypointense signal change due to ICH on scout images is clearly visible and marked by sharp margins, which may be due to the edema, and this enables the differentiation of ICH from subacute or chronic infarcts, especially when they are large in size. In this study we evaluated the detection of ICH from scout images as these are a prerequisite in any MRI scan. The detection of ICH from scout images should be compared to other sequences with short acquisition times, such as a DWI, which can be acquired in under 2 min. This could provide further information on early diagnosis on MRI. Further larger studies are needed to confirm our findings. In conclusion, > 90% of acute ICH patients can be diagnosed using MRI scout imaging. Attention to symptoms originating in the brainstem can increase the safety of MRI as a first-line diagnostic test for stroke.
Fig. 2. Receiver operating characteristic curve. NIHSS score on admission (AUC = 0.757; p = .010) and bleeding volume (AUC = 0.942; p < .001) were significantly associated with diagnostic ability of MRI scout imaging for ICH, whereas age (AUC = 0.351; p = .134) and time from onset to imaging (AUC = 0.291; p = .360) were not.
Table 2 Patients redistributed using cut-off value of bleeding volume and NIHSS score.
Bleeding volume ≥2 ml < 2 ml NIHSS score ≥3 <3
Group with MRI scout positive n = 138
Group with MRI scout negative n = 10
Total n = 148
129 (93%) 9 (7%)
1 (10%) 9 (90%)
130 (88%) 18 (12%)
122 (88%) 16 (12%)
3 (30%) 7 (70%)
125 (84%) 23 (16%)
cut-off value for detecting an ICH. Although MRI scout images are acquired in 3 different orientations (horizontal, coronal, and sagittal), they are acquired for subsequent positioning so that the slice thickness will not be thinner than in CT imaging. Therefore, low-volume ICHs could be difficult to detect given the limited range of MRI scout images, although they can be easily detected by general MRI and CT imaging. With an understanding of its advantages and disadvantages, we think MRI scout imaging can be a suitable diagnostic tool. A number of specialized stroke centers currently utilize MRI as the first-line imaging tool for patients suspected of having a stroke [2–4]. The advantage of MRI in patients with acute ischemic stroke is its high diagnostic accuracy. Within 3 h after stroke onset, over 80% of patients can be diagnosed based on diffusion-weighted images [5]. However, MRI has the disadvantage that it prevents physicians from detecting worsening conditions in patients, including respiratory status and vomiting, due to its long time requirement and the closed space in which patients are placed. If it is possible to diagnose severe ICH patients, who tend to get worse over time, we can stop to obtain additional MRI sequences, providing shortening the scan time and increasing the safety of MRI. Further study is expected to confirm our study's findings. To date, no previous reports have described the use of MRI scout imaging for diagnosis, although it is discussed in some case reports. Yanagi [6] reported the incidental discovery of tumors in 2 of 2776
Disclosures None References [1] D. Le Bihan, E. Breton, Imagerie de diffusion in vivo par resonance magnétique nucléaire, C R Acad. Sci. Paris T. 301 (1995) 1109–1112. [2] B.T. Buckley, A. Wainwright, T. Meagher, D. Briley, Audit of a policy of magnetic resonance imaging with diffusion-weighted imaging as first-line neuroimaging for inpatients with clinically suspected acute stroke, Clin. Radiol. 58 (2003) 234–237. [3] P.L. Tan, D. King, C.J. Durkin, T.M. Meagher, D. Briley, Diffusion weighted magnetic resonance imaging for acute stroke: practical and popular, Postgrad. Med. J. 82 (2006) 289–292. [4] R.R. Leker, G. Keigler, R. Eichel, T. Ben Hur, J.M. Gomori, J.E. Cohen, Should DWI MRI be the primary screening test for stroke? Int. J. Stroke 9 (2014) 696–697. [5] Junya Aoki, Kazumi Kimura, Yasuyuki Iguchi, Kensaku Shibazaki, Kenichiro Sakai, Takeshi Iwanaga, FLAIR can estimate the onset time in acute ischemic stroke patients, J. Neurol. Sci. 293 (2010) 39–44. [6] Yoshinobu Yanagi, Junichi Asaumi, Yuu Maki, Jun Murakami, Miki Hisatomi, Hidenobu Matsuzaki, et al., Incidentally found and unexpected tumors discovered by MRI examination for temporomandibular joint arthrosis, Eur. J. Radiol. 47 (2003) 6–9.
149