Neuroradiology DOI 10.1007/s00234-014-1479-z
DIAGNOSTIC NEURORADIOLOGY
Role of thalamic diffusion for disease differentiation between multiple sclerosis and ischemic cerebral small vessel disease Bilge Öztoprak & İbrahim Öztoprak & Kamil Topalkara & Mustafa F. Erkoç & İsmail Şalk
Received: 20 August 2014 / Accepted: 5 December 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Introduction Cerebral small vessel disease (CSVD) and multiple sclerosis (MS) both harbor multiple, T2-hyperintense white matter lesions on conventional magnetic resonance imaging (MRI).We aimed to determine the microstructural changes via diffusion-weighted imaging (DWI) in normal appearing thalami. We hypothesized that the apparent diffusion coefficient (ADC) values would be different in CSVD and MS, since the extent of arterial involvement is different in these two diseases. Methods DWI was performed for 50 patients with CSVD and 35 patients with MS along with gender- and age-matched controls whose conventional MRI revealed normal findings. DWI was done with 1.5 Tesla MR devices using echo planar imaging (EPI) for b=0, 1000 s/mm2. ADC values were obtained from the thalami which appeared normal on T2weighted and FLAIR images. Standard oval regions of interest (ROIs) of 0.5 cm2 which were oriented parallel to the long axis of the thalamus were used for this purpose. Results The mean ADC value of the thalamus was (0.99± 0.16)×10−3 mm2/s in patients with CSVD, whereas the mean ADC value was (0.78±0.06)×10−3 mm2/s in the control group. The mean ADC value was significantly higher in patients with CSVD compared to the controls (p0.05). Conclusion Our study revealed a difference in the diffusion of the thalami between CSVD and MS. DWI may aid in the radiological disease differentiation. Keywords Thalamus . Diffusion-weighted imaging . Cerebral small vessel disease . Multiple sclerosis Abbreviations MS Multiple sclerosis CSVD Cerebral small vessel disease MRI Magnetic resonance imaging DWI Diffusion-weighted imaging ADC Apparent diffusion coefficient ROI Region of interest
Introduction Multiple sclerosis (MS) focal demyelinating lesions and ischemic gliotic foci of cerebral small vessel disease (CSVD) located in the cerebral white matter may cause similar appearance on MRI [1]. CSVD and MS, which both show T2hyperintense lesions on conventional MRI, are the most common pathological entities that require differential diagnosis. Although clinical and laboratory findings provide important data in differential diagnosis, they may not always be sufficient in equivocal cases. Age is an important parameter, but it may be misleading. Although MS is primarily a disease of young adults, it may affect all age groups from infancy to eighth decade [2]. On the other hand, CSVD is more frequently seen in advanced age, but up to 20 % of patients are below 50 years of age; therefore, occurrence of lesions of these two pathologies in the same age group is not rare [3]. Diffusion
Neuroradiology
studies showed increased, decreased, or normal diffusion in T2-hyperintense areas in both CSVD and MS according to the age of the lesions [4–7], which are also nonspecific. Although lesion location (juxtacortical, infratentorial, spinal cord) and distribution (perivascular) are important in differentiating MS from vascular lesions, most of these types of lesions can also be observed in CSVD [8]. Small penetrating artery infarctions account for approximately 20–25 % of all ischemic strokes [9]. Fibrohyalinosis affecting the small penetrating arteries causes ischemic lesions in the deep gray matter together with the ischemic gliotic foci in the white matter, since penetrating arteries are located in the cerebral white matter and deep gray matter [10, 11]. On the other hand, MS pathology is primarily due to the deterioration of myelin sheath and oligodendrocytes as a result of perivascular inflammation. However, the pathologic events causing degeneration of axons in MS may affect the small arteries and may lead to infarction in advanced stages of the disease [12]. In recent years, with new methods of MRI, it was understood that the pathologic process is not limited to the areas identified by conventional MRI sequences but more widespread, such as normal-appearing white matter [13, 14]. From this perspective, normal-appearing deep gray matter areas may be expected to be affected in MS, either directly from demyelination or secondarily from vascular impairment. However, this influence is likely to be significantly less than in the case of CSVD, since CSVD is a primary vascular disease unlike MS in which vascular structures are secondarily involved. The aims of this study were to identify the microstructural changes in normal-appearing thalami on conventional MRI using DWI and to determine the difference between MS and CSVD quantitatively with the apparent diffusion coefficient (ADC) values. According to our hypothesis, the ADC values of the thalamus might be different since the degree of arterial involvement in these two pathologies is different. We expected to obtain data to distinguish these pathologies even though white matter lesions may mimic both pathologies.
groups constituted of 50 subjects for CSVD group and 35 subjects for MS group who were sent to our department for different reasons, such as headache, vertigo, tinnitus, hearing loss, seizures, and who had normal findings on conventional MRI. There was no history of any cardiac, metabolic, or hematologic disorder in the control groups. Patients with comorbidities which may be predisposing factors for CSVD, such as hypertension, diabetes mellitus, coronary heart disease, history of myocardial infarction, or any known autoimmune disorder, were not included in the control groups. Revised McDonald criteria valid for the date of admission were used for the diagnosis of MS [15, 16]. Thirty-one (89 %) of MS patients were previously diagnosed and were undergone MRI before at least once. Mean disease duration for these patients was 8 years. Four (11 %) patients were admitted with the suspicion of MS and diagnosed with MS based on clinical, laboratory, and imaging findings. MS patients were included in the study regardless of the clinical subtypes and stages of the disease. Patients who did not have the definitive diagnosis of MS throughout the course of the study were excluded. There was no evidence of vascular comorbidity in the MS patient group. Since standard diagnostic criteria did not exist for CSVD, the patients were selected among those who had multiple hyperintense foci in the cerebral white matter on conventional T2-weighted and FLAIR MR images. The presence of some risk factors such as hypertension, diabetes mellitus, dyslipidemia, and history of smoking was searched in the selection of patients, and among these patients whose clinical, laboratory, and imaging findings could not rule out such pathologies as thromboembolic disease, vasculitis, infection, demyelinating disease, trauma, and intracranial space-occupying lesions were excluded from the study. Fourteen male and 21 female subjects whose ages ranged between 21 and 53 composed the control group of MS patients. On the other hand, the control group of CSVD patients consisted of 26 male and 24 female individuals whose ages ranged from 29 to 76. Patients and controls were informed about the study. No special preparation was done before the examination.
Materials and methods Image acquisition Patient selection This prospective case-control study was approved by the Ethics Committee of Cumhuriyet University School of Medicine (No. 2007-4/8; decided May 1, 2007). Fifty patients with CSVD and 35 patients with MS were examined with MRI in our radiology department between May 2007 and April 2013. Since there are significant differences between CSVD and MS patients in terms of age and gender, two different control groups which were statistically comparable with patient groups were conducted. The control
MR imaging was done with two 1.5 Tesla MRI units (Magnetom Aera, Siemens, Erlangen, Germany and Exelart, Toshiba, Tokyo, Japan), using 16-channel phased array head coil and standard head coil, respectively. The conventional MRI protocol included axial and coronal T1-weighted spin echo (SE) (parameters for two vendors, Siemens and Toshiba, respectively, were as follows: TR 520 and 550 ms; TE 5.6 and 15 ms; FA 150 and 70/180; NEX 3 and 1.2; FOV 220×84 and 220×180 mm; matrix 256×100 and 256×160; slice thickness 5 and 5 mm; interslice gap 1.7 and 1 mm); axial, coronal, and
Neuroradiology
sagittal T2-weighted fast SE (TR 4400 and 5000 ms; TE 102 and 94 ms; FA 150 and 90/180; NEX 2 and 2; FOV 220×97 and 220×180 mm; matrix 320×90 and 320×224; slice thickness 5 and 5 mm; interslice gap 1.7 and 1 mm) and axial and sagittal FLAIR (TR 8000 and 7500 ms, TE 86 and 94 ms, TI 2384 and 2200 ms; FA 150 and 90/160; NEX 1 and 1; FOV 220×97 and 220×180 mm; matrix 256×100 and 256×160; slice thickness 5 and 5 mm; interslice gap 1.7 and 1 mm) images. Intravenous contrast agent was administered for all of the MS patients. For the other patient group and the controls, contrast media were used if there was an indication, such as suspicion of a neoplasm, meningitis, etc. Contrast-enhanced imaging was done with a 0.1 mmol/kg intravenous paramagnetic agent (gadolinium-DTPA or gadodiamide) in T1weighted SE axial and coronal images. Diffusion-weighted images were acquired using echo planar imaging (EPI) sequences (TR 5800 and 5000 ms; TE 98 and 130 ms; FA 150 and 90/180; NEX 1 and 1; FOV 225× 100 and 320×270 mm; matrix 192×100 and 128×128; slice thickness 5 and 5 mm; interslice gap 1.7 and 2 mm; b value 0 and 1000 s/mm2). Diffusion gradients were applied in three orthogonal planes in order to measure diffusion in three planes (x, y, and z).
Post-processing and image analysis Thalamus was selected for the measurement of ADC values of deep gray matter containing penetrating arteries, since the area of the thalamus was larger than the basal ganglia, and measurement without contamination was more easily provided. Besides, Virchow-Robin spaces which might change ADC values are common in the basal ganglia. The presence of these spaces is rare in the thalamus. Thus, the thalamus was considered to be the most suitable site to measure ADCs.
Fig. 1 Diffusion-weighted image (a) and corresponding ADC map (b) showing placement of ROIs for quantitative measurement
Having measured a set of at least two different b value images, the system calculated pixel by pixel the ADC by linear regression using Syngo VD13A software, in Siemens MRI unit. The ADC pixel values together formed the ADC map, and on a half logarithmic scale, the signal decay delivered a straight tilted line whose slope provided the ADC. With the Toshiba MRI unit, DW images were transferred to a separate workstation after the observation of absence of any signal change in conventional MR images, especially on T2weighted and FLAIR images, of the thalami of both patient and control groups. Direction-independent isotropic diffusion images were obtained by post-processing of diffusion images acquired from x-, y-, and z-axes. ADC maps were constructed via subtraction of SE EP b=0 images from these isotropic images. The regions of interest (ROIs), the areas from where the ADC values were quantitatively measured on ADC maps, were selected from the central portions of the thalami far away enough from the ventricles medially and internal capsules laterally to avoid contamination of these tissues. Standard oval ROIs, compatible with the anatomic shape of the thalamus, with an area of 0.5 cm2 which were oriented parallel to the long axis of the thalamus were used (Fig. 1). ADC values of the right and left thalami were calculated by the MR device automatically for each case. The difference between the thalamic diffusions of patient and control groups was determined.
Statistical analysis The data obtained were analyzed using SPSS software. The difference between the ages, ADC values of the right and left thalami, and mean ADC values of patient and control groups was analyzed with the t test for the significance of the difference of the means. The chi-square test
Neuroradiology
was used for the comparison of the gender of patient and control groups. p values less than 0.05 were considered to be statistically significant.
Results Fifty-two percent (n=26) of patients with CSVD were women, 48 % (n=24) were men, and their mean age was 58.7± 12.6 (29–80). Forty-eight percent (n=24) of patients in the control group of CSVD were women, and 52 % (n=26) were men with a mean age of 58.1±11.7 (29–80). The difference between the patient and control groups in terms of age (p= 0.842) and gender (p=0.806) was not statistically significant. On the other hand, 71 % (n=25), of patients with MS were women and 29 % (n=10) were men, with a mean age of 36.1± 8.5 (21–52). Twenty-one (60 %) women and 14 (40 %) men with a mean age of 35.5±8.3 (21–53) consisted the control group of MS. No statistically significant difference was found between these two groups in terms of age (p=0.755) and gender (p=0.450). According to the Fazekas scale [17], 27 (54 %) patients with CSVD were rated as Fazekas 1, 19 (38 %) patients were rated as Fazekas 2, and 4 (8 %) patients were rated as Fazekas 3 on the basis of FLAIR and T2 hyperintensities (Fig. 2). Most of the patients with CSVD had chronic hypertension (n=32, 64 %). Seventeen (34 %) patients had type 2 diabetes mellitus, 11 (22 %) had hyperlipidemia, and 5 (10 %) had coronary heart disease in the CSVD group. Other accompanying risk factors included smoking, obesity, history of previous ischemic attack, myocardial infarction, and atrial fibrillation. Clinical information about comorbidities and risk factors of patients are summarized in Table 1. Thirty-one (88.6 %) of patients had relapsing remitting (RR) subtype of MS, and four (11.4 %) patients had secondary progressive (SP) type. Six (17 %) MS patients had one to four
Fig. 2 Distribution of patients with CSVD according to Fazekas scale for white matter lesions (from 0 to 3)
enhancing lesions. Expanded disability status scale (EDSS) of MS patients ranged between 1 and 6, with a mean of 2.1±1.2 (Table 1). Most of the patients (n=29, 83 %) were under disease modifying treatment at the time of MR imaging. Six patients received no treatment at the time of admission to radiology department for various reasons, such as pregnancy or disinclination for medication due to side effects. All patients with CSVD and MS involved in the study showed multiple T2-hyperintense lesions in bilateral cerebral deep white matter, which are more prominent in the periventricular regions (Figs. 3 and 4). The mean ADC value of the thalamus was (0.99±0.16)× 10−3 mm2/s in patients with CSVD, whereas the mean ADC value was (0.78±0.06)×10−3 mm2/s in the control group. The mean ADC value was significantly higher in patients with CSVD compared to the control group (p0.05) (Table 2). There was no statistically significant difference between the right and left thalami in patient and control groups with respect to ADC values (p>0.05) (Table 2). There was statistically significant difference in mean thalamic ADC values (ADCmean) between CSVD patients and their controls (p30) Previous ischemic attack History of myocardial infarction History of atrial fibrillation EDSS (mean)
results show that there are in fact some microstructural changes in normal-appearing thalami on conventional MRI. This is probably due to tissue damage resulting from penetrating artery involvement, gliosis, and increased extracellular fluid motion as a result. In contrast to our study, Chun et al. [21] found normal thalamic diffusion in nine individuals who had less than five small focal ischemic white matter lesions in their study to evaluate age-related diffusion changes in 38 subjects. However, healthy subjects and patients were both included in their study, and the number of patients and lesions was too limited to exemplify CSVD with white matter T2 hyperintensities. Mean ADC values that we obtained from MS patients and matched controls were within normal limits similar to the results of Helenius et al. [6] in normal subjects in different age groups. We can interpret this result as normal-appearing thalamus is not affected from demyelination to a degree to change ADC values, or no small vessel involvement is present in MS, or if present, it is not to a degree to affect ADC values. Some reports, such as Griffin et al. [22] and Ciccarelli et al. [23], support our findings, whereas some others such as a
CSVD (n=50)
MS (n=35)
No. and ratio of patients (%)
No. and ratio of patients (%)
58.7±12.6
36.1±8.5
24 (48 %) 26 (52 %)
10 (29 %) 25 (71 %)
32 (64 %) 7.5 (min 1, max 23)
3 (9 %) 5 (min 4, max 7)
17 (34 %) 9 (min 1, max 23)
1 (3 %) 3
11 (22 %) 10 (min 3, max 18)
2 (6 %) 5 (min 3, max 7)
5 (10 %) 9 (min 2, max 15)
0 (0 %) N/A
13 (26 %) 15 (min 3, max 34) 9 (18 %) 3 (6 %) 1 (2 %) 0 (0 %) N/A
8 (23 %) 10 (min 3, max 16) 4 (11 %) 0 (0 %) 0 (0 %) 0 (0 %) 2.1±1.2
diffusion tensor imaging study by Tovar-Moll et al. revealed thalamic involvement, which may be attributable to the lesion load or subtype of MS patients as well as the higher resolution of high-field MR scanner [24–26]. Attempts are still being made to aid in differential diagnosis of MS and CSVD. Susceptibility-weighted imaging or FLAIR* images, which are a combination of FLAIR and T2*-weighted images, can be used for this purpose on 7T, 3T, and even 1.5T MRI devices to show intralesional veins and perilesional hypointense rim. A very recent study by Kilsdonk et al. [27] analyzed the lesions themselves and concluded that MS lesions can be differentiated from CSVD by displaying a vessel passing through the center of the lesion or a hypointense rim around the lesion. This promising study was carried out with a 7 Tesla MRI unit which is not available in all clinical settings with a limitation of evaluation of only supratentorial lesions. Similar studies were also carried out in 7 Tesla units to differentiate MS from other diseases such as Susac’s disease and neuromyelitis optica [28, 29]. There are some limitations of our study: Patients were included in our study irrespective of the subtype of MS, and
Neuroradiology Fig. 3 MR images of a 38-yearold female patient with CSVD obtained from the level of lateral ventricles (a, c) and centrum semiovale (b, d). T2-weighted (a, b) and FLAIR (c, d) images reveal multiple hyperintense foci in cerebral white matter, which are more prominent in the periventricular area (arrows). Thalami appear normal on T2weighted image (e) and ADC map (f)
the phase of the disease which might affect the results was also ignored. Patients with CSVD were not categorized in terms of severity, either, although most of them were grade 1 (mild)
and 2 (moderate) according to Fazekas scale [17]. In other words, the disease burden of both patient groups was limited, which might have also limited our study. The control groups
Neuroradiology Fig. 4 MR images of a 44-yearold female patient with MS. T2weighted (a, b) and corresponding FLAIR (c, and d) images from two different levels demonstrate multiple white matter hyperintensities, predominantly in the periventricular regions (arrows). Thalami show normal signal on T2-weighted image (e) and ADC map (f)
were conducted among subjects with normal cranial MRI findings instead of normal healthy controls, which is another limitation of our study. In addition, two different MRI systems were used in our study; even we know that ADC values are a
constant. Further studies with increased specificity and reproducibility are needed. Younger age of the patient, female predilection, oval shape and periventricular, pericallosal, infratentorial and
Neuroradiology Table 2 Mean ADC numbers in patients with CSVD, MS, and controls CSVD CSVD control p value MS MS control CSVD cerebral small vessel disease, MS multiple sclerosis, ADC-R apparent diffusion coefficient of the right thalamus, ADC-L apparent diffusion coefficient of the left thalamus, ADC-M mean apparent diffusion coefficient
p value CSVD MS p value CSVD control MS control p value
Number
Mean ADC-R
Mean ADC-L
Mean ADC-M
50 50
0.98±0.17×10−3 0.78±0.07×10−3
DIAGNOSTIC NEURORADIOLOGY
Role of thalamic diffusion for disease differentiation between multiple sclerosis and ischemic cerebral small vessel disease Bilge Öztoprak & İbrahim Öztoprak & Kamil Topalkara & Mustafa F. Erkoç & İsmail Şalk
Received: 20 August 2014 / Accepted: 5 December 2014 # Springer-Verlag Berlin Heidelberg 2014
Abstract Introduction Cerebral small vessel disease (CSVD) and multiple sclerosis (MS) both harbor multiple, T2-hyperintense white matter lesions on conventional magnetic resonance imaging (MRI).We aimed to determine the microstructural changes via diffusion-weighted imaging (DWI) in normal appearing thalami. We hypothesized that the apparent diffusion coefficient (ADC) values would be different in CSVD and MS, since the extent of arterial involvement is different in these two diseases. Methods DWI was performed for 50 patients with CSVD and 35 patients with MS along with gender- and age-matched controls whose conventional MRI revealed normal findings. DWI was done with 1.5 Tesla MR devices using echo planar imaging (EPI) for b=0, 1000 s/mm2. ADC values were obtained from the thalami which appeared normal on T2weighted and FLAIR images. Standard oval regions of interest (ROIs) of 0.5 cm2 which were oriented parallel to the long axis of the thalamus were used for this purpose. Results The mean ADC value of the thalamus was (0.99± 0.16)×10−3 mm2/s in patients with CSVD, whereas the mean ADC value was (0.78±0.06)×10−3 mm2/s in the control group. The mean ADC value was significantly higher in patients with CSVD compared to the controls (p0.05). Conclusion Our study revealed a difference in the diffusion of the thalami between CSVD and MS. DWI may aid in the radiological disease differentiation. Keywords Thalamus . Diffusion-weighted imaging . Cerebral small vessel disease . Multiple sclerosis Abbreviations MS Multiple sclerosis CSVD Cerebral small vessel disease MRI Magnetic resonance imaging DWI Diffusion-weighted imaging ADC Apparent diffusion coefficient ROI Region of interest
Introduction Multiple sclerosis (MS) focal demyelinating lesions and ischemic gliotic foci of cerebral small vessel disease (CSVD) located in the cerebral white matter may cause similar appearance on MRI [1]. CSVD and MS, which both show T2hyperintense lesions on conventional MRI, are the most common pathological entities that require differential diagnosis. Although clinical and laboratory findings provide important data in differential diagnosis, they may not always be sufficient in equivocal cases. Age is an important parameter, but it may be misleading. Although MS is primarily a disease of young adults, it may affect all age groups from infancy to eighth decade [2]. On the other hand, CSVD is more frequently seen in advanced age, but up to 20 % of patients are below 50 years of age; therefore, occurrence of lesions of these two pathologies in the same age group is not rare [3]. Diffusion
Neuroradiology
studies showed increased, decreased, or normal diffusion in T2-hyperintense areas in both CSVD and MS according to the age of the lesions [4–7], which are also nonspecific. Although lesion location (juxtacortical, infratentorial, spinal cord) and distribution (perivascular) are important in differentiating MS from vascular lesions, most of these types of lesions can also be observed in CSVD [8]. Small penetrating artery infarctions account for approximately 20–25 % of all ischemic strokes [9]. Fibrohyalinosis affecting the small penetrating arteries causes ischemic lesions in the deep gray matter together with the ischemic gliotic foci in the white matter, since penetrating arteries are located in the cerebral white matter and deep gray matter [10, 11]. On the other hand, MS pathology is primarily due to the deterioration of myelin sheath and oligodendrocytes as a result of perivascular inflammation. However, the pathologic events causing degeneration of axons in MS may affect the small arteries and may lead to infarction in advanced stages of the disease [12]. In recent years, with new methods of MRI, it was understood that the pathologic process is not limited to the areas identified by conventional MRI sequences but more widespread, such as normal-appearing white matter [13, 14]. From this perspective, normal-appearing deep gray matter areas may be expected to be affected in MS, either directly from demyelination or secondarily from vascular impairment. However, this influence is likely to be significantly less than in the case of CSVD, since CSVD is a primary vascular disease unlike MS in which vascular structures are secondarily involved. The aims of this study were to identify the microstructural changes in normal-appearing thalami on conventional MRI using DWI and to determine the difference between MS and CSVD quantitatively with the apparent diffusion coefficient (ADC) values. According to our hypothesis, the ADC values of the thalamus might be different since the degree of arterial involvement in these two pathologies is different. We expected to obtain data to distinguish these pathologies even though white matter lesions may mimic both pathologies.
groups constituted of 50 subjects for CSVD group and 35 subjects for MS group who were sent to our department for different reasons, such as headache, vertigo, tinnitus, hearing loss, seizures, and who had normal findings on conventional MRI. There was no history of any cardiac, metabolic, or hematologic disorder in the control groups. Patients with comorbidities which may be predisposing factors for CSVD, such as hypertension, diabetes mellitus, coronary heart disease, history of myocardial infarction, or any known autoimmune disorder, were not included in the control groups. Revised McDonald criteria valid for the date of admission were used for the diagnosis of MS [15, 16]. Thirty-one (89 %) of MS patients were previously diagnosed and were undergone MRI before at least once. Mean disease duration for these patients was 8 years. Four (11 %) patients were admitted with the suspicion of MS and diagnosed with MS based on clinical, laboratory, and imaging findings. MS patients were included in the study regardless of the clinical subtypes and stages of the disease. Patients who did not have the definitive diagnosis of MS throughout the course of the study were excluded. There was no evidence of vascular comorbidity in the MS patient group. Since standard diagnostic criteria did not exist for CSVD, the patients were selected among those who had multiple hyperintense foci in the cerebral white matter on conventional T2-weighted and FLAIR MR images. The presence of some risk factors such as hypertension, diabetes mellitus, dyslipidemia, and history of smoking was searched in the selection of patients, and among these patients whose clinical, laboratory, and imaging findings could not rule out such pathologies as thromboembolic disease, vasculitis, infection, demyelinating disease, trauma, and intracranial space-occupying lesions were excluded from the study. Fourteen male and 21 female subjects whose ages ranged between 21 and 53 composed the control group of MS patients. On the other hand, the control group of CSVD patients consisted of 26 male and 24 female individuals whose ages ranged from 29 to 76. Patients and controls were informed about the study. No special preparation was done before the examination.
Materials and methods Image acquisition Patient selection This prospective case-control study was approved by the Ethics Committee of Cumhuriyet University School of Medicine (No. 2007-4/8; decided May 1, 2007). Fifty patients with CSVD and 35 patients with MS were examined with MRI in our radiology department between May 2007 and April 2013. Since there are significant differences between CSVD and MS patients in terms of age and gender, two different control groups which were statistically comparable with patient groups were conducted. The control
MR imaging was done with two 1.5 Tesla MRI units (Magnetom Aera, Siemens, Erlangen, Germany and Exelart, Toshiba, Tokyo, Japan), using 16-channel phased array head coil and standard head coil, respectively. The conventional MRI protocol included axial and coronal T1-weighted spin echo (SE) (parameters for two vendors, Siemens and Toshiba, respectively, were as follows: TR 520 and 550 ms; TE 5.6 and 15 ms; FA 150 and 70/180; NEX 3 and 1.2; FOV 220×84 and 220×180 mm; matrix 256×100 and 256×160; slice thickness 5 and 5 mm; interslice gap 1.7 and 1 mm); axial, coronal, and
Neuroradiology
sagittal T2-weighted fast SE (TR 4400 and 5000 ms; TE 102 and 94 ms; FA 150 and 90/180; NEX 2 and 2; FOV 220×97 and 220×180 mm; matrix 320×90 and 320×224; slice thickness 5 and 5 mm; interslice gap 1.7 and 1 mm) and axial and sagittal FLAIR (TR 8000 and 7500 ms, TE 86 and 94 ms, TI 2384 and 2200 ms; FA 150 and 90/160; NEX 1 and 1; FOV 220×97 and 220×180 mm; matrix 256×100 and 256×160; slice thickness 5 and 5 mm; interslice gap 1.7 and 1 mm) images. Intravenous contrast agent was administered for all of the MS patients. For the other patient group and the controls, contrast media were used if there was an indication, such as suspicion of a neoplasm, meningitis, etc. Contrast-enhanced imaging was done with a 0.1 mmol/kg intravenous paramagnetic agent (gadolinium-DTPA or gadodiamide) in T1weighted SE axial and coronal images. Diffusion-weighted images were acquired using echo planar imaging (EPI) sequences (TR 5800 and 5000 ms; TE 98 and 130 ms; FA 150 and 90/180; NEX 1 and 1; FOV 225× 100 and 320×270 mm; matrix 192×100 and 128×128; slice thickness 5 and 5 mm; interslice gap 1.7 and 2 mm; b value 0 and 1000 s/mm2). Diffusion gradients were applied in three orthogonal planes in order to measure diffusion in three planes (x, y, and z).
Post-processing and image analysis Thalamus was selected for the measurement of ADC values of deep gray matter containing penetrating arteries, since the area of the thalamus was larger than the basal ganglia, and measurement without contamination was more easily provided. Besides, Virchow-Robin spaces which might change ADC values are common in the basal ganglia. The presence of these spaces is rare in the thalamus. Thus, the thalamus was considered to be the most suitable site to measure ADCs.
Fig. 1 Diffusion-weighted image (a) and corresponding ADC map (b) showing placement of ROIs for quantitative measurement
Having measured a set of at least two different b value images, the system calculated pixel by pixel the ADC by linear regression using Syngo VD13A software, in Siemens MRI unit. The ADC pixel values together formed the ADC map, and on a half logarithmic scale, the signal decay delivered a straight tilted line whose slope provided the ADC. With the Toshiba MRI unit, DW images were transferred to a separate workstation after the observation of absence of any signal change in conventional MR images, especially on T2weighted and FLAIR images, of the thalami of both patient and control groups. Direction-independent isotropic diffusion images were obtained by post-processing of diffusion images acquired from x-, y-, and z-axes. ADC maps were constructed via subtraction of SE EP b=0 images from these isotropic images. The regions of interest (ROIs), the areas from where the ADC values were quantitatively measured on ADC maps, were selected from the central portions of the thalami far away enough from the ventricles medially and internal capsules laterally to avoid contamination of these tissues. Standard oval ROIs, compatible with the anatomic shape of the thalamus, with an area of 0.5 cm2 which were oriented parallel to the long axis of the thalamus were used (Fig. 1). ADC values of the right and left thalami were calculated by the MR device automatically for each case. The difference between the thalamic diffusions of patient and control groups was determined.
Statistical analysis The data obtained were analyzed using SPSS software. The difference between the ages, ADC values of the right and left thalami, and mean ADC values of patient and control groups was analyzed with the t test for the significance of the difference of the means. The chi-square test
Neuroradiology
was used for the comparison of the gender of patient and control groups. p values less than 0.05 were considered to be statistically significant.
Results Fifty-two percent (n=26) of patients with CSVD were women, 48 % (n=24) were men, and their mean age was 58.7± 12.6 (29–80). Forty-eight percent (n=24) of patients in the control group of CSVD were women, and 52 % (n=26) were men with a mean age of 58.1±11.7 (29–80). The difference between the patient and control groups in terms of age (p= 0.842) and gender (p=0.806) was not statistically significant. On the other hand, 71 % (n=25), of patients with MS were women and 29 % (n=10) were men, with a mean age of 36.1± 8.5 (21–52). Twenty-one (60 %) women and 14 (40 %) men with a mean age of 35.5±8.3 (21–53) consisted the control group of MS. No statistically significant difference was found between these two groups in terms of age (p=0.755) and gender (p=0.450). According to the Fazekas scale [17], 27 (54 %) patients with CSVD were rated as Fazekas 1, 19 (38 %) patients were rated as Fazekas 2, and 4 (8 %) patients were rated as Fazekas 3 on the basis of FLAIR and T2 hyperintensities (Fig. 2). Most of the patients with CSVD had chronic hypertension (n=32, 64 %). Seventeen (34 %) patients had type 2 diabetes mellitus, 11 (22 %) had hyperlipidemia, and 5 (10 %) had coronary heart disease in the CSVD group. Other accompanying risk factors included smoking, obesity, history of previous ischemic attack, myocardial infarction, and atrial fibrillation. Clinical information about comorbidities and risk factors of patients are summarized in Table 1. Thirty-one (88.6 %) of patients had relapsing remitting (RR) subtype of MS, and four (11.4 %) patients had secondary progressive (SP) type. Six (17 %) MS patients had one to four
Fig. 2 Distribution of patients with CSVD according to Fazekas scale for white matter lesions (from 0 to 3)
enhancing lesions. Expanded disability status scale (EDSS) of MS patients ranged between 1 and 6, with a mean of 2.1±1.2 (Table 1). Most of the patients (n=29, 83 %) were under disease modifying treatment at the time of MR imaging. Six patients received no treatment at the time of admission to radiology department for various reasons, such as pregnancy or disinclination for medication due to side effects. All patients with CSVD and MS involved in the study showed multiple T2-hyperintense lesions in bilateral cerebral deep white matter, which are more prominent in the periventricular regions (Figs. 3 and 4). The mean ADC value of the thalamus was (0.99±0.16)× 10−3 mm2/s in patients with CSVD, whereas the mean ADC value was (0.78±0.06)×10−3 mm2/s in the control group. The mean ADC value was significantly higher in patients with CSVD compared to the control group (p0.05) (Table 2). There was no statistically significant difference between the right and left thalami in patient and control groups with respect to ADC values (p>0.05) (Table 2). There was statistically significant difference in mean thalamic ADC values (ADCmean) between CSVD patients and their controls (p30) Previous ischemic attack History of myocardial infarction History of atrial fibrillation EDSS (mean)
results show that there are in fact some microstructural changes in normal-appearing thalami on conventional MRI. This is probably due to tissue damage resulting from penetrating artery involvement, gliosis, and increased extracellular fluid motion as a result. In contrast to our study, Chun et al. [21] found normal thalamic diffusion in nine individuals who had less than five small focal ischemic white matter lesions in their study to evaluate age-related diffusion changes in 38 subjects. However, healthy subjects and patients were both included in their study, and the number of patients and lesions was too limited to exemplify CSVD with white matter T2 hyperintensities. Mean ADC values that we obtained from MS patients and matched controls were within normal limits similar to the results of Helenius et al. [6] in normal subjects in different age groups. We can interpret this result as normal-appearing thalamus is not affected from demyelination to a degree to change ADC values, or no small vessel involvement is present in MS, or if present, it is not to a degree to affect ADC values. Some reports, such as Griffin et al. [22] and Ciccarelli et al. [23], support our findings, whereas some others such as a
CSVD (n=50)
MS (n=35)
No. and ratio of patients (%)
No. and ratio of patients (%)
58.7±12.6
36.1±8.5
24 (48 %) 26 (52 %)
10 (29 %) 25 (71 %)
32 (64 %) 7.5 (min 1, max 23)
3 (9 %) 5 (min 4, max 7)
17 (34 %) 9 (min 1, max 23)
1 (3 %) 3
11 (22 %) 10 (min 3, max 18)
2 (6 %) 5 (min 3, max 7)
5 (10 %) 9 (min 2, max 15)
0 (0 %) N/A
13 (26 %) 15 (min 3, max 34) 9 (18 %) 3 (6 %) 1 (2 %) 0 (0 %) N/A
8 (23 %) 10 (min 3, max 16) 4 (11 %) 0 (0 %) 0 (0 %) 0 (0 %) 2.1±1.2
diffusion tensor imaging study by Tovar-Moll et al. revealed thalamic involvement, which may be attributable to the lesion load or subtype of MS patients as well as the higher resolution of high-field MR scanner [24–26]. Attempts are still being made to aid in differential diagnosis of MS and CSVD. Susceptibility-weighted imaging or FLAIR* images, which are a combination of FLAIR and T2*-weighted images, can be used for this purpose on 7T, 3T, and even 1.5T MRI devices to show intralesional veins and perilesional hypointense rim. A very recent study by Kilsdonk et al. [27] analyzed the lesions themselves and concluded that MS lesions can be differentiated from CSVD by displaying a vessel passing through the center of the lesion or a hypointense rim around the lesion. This promising study was carried out with a 7 Tesla MRI unit which is not available in all clinical settings with a limitation of evaluation of only supratentorial lesions. Similar studies were also carried out in 7 Tesla units to differentiate MS from other diseases such as Susac’s disease and neuromyelitis optica [28, 29]. There are some limitations of our study: Patients were included in our study irrespective of the subtype of MS, and
Neuroradiology Fig. 3 MR images of a 38-yearold female patient with CSVD obtained from the level of lateral ventricles (a, c) and centrum semiovale (b, d). T2-weighted (a, b) and FLAIR (c, d) images reveal multiple hyperintense foci in cerebral white matter, which are more prominent in the periventricular area (arrows). Thalami appear normal on T2weighted image (e) and ADC map (f)
the phase of the disease which might affect the results was also ignored. Patients with CSVD were not categorized in terms of severity, either, although most of them were grade 1 (mild)
and 2 (moderate) according to Fazekas scale [17]. In other words, the disease burden of both patient groups was limited, which might have also limited our study. The control groups
Neuroradiology Fig. 4 MR images of a 44-yearold female patient with MS. T2weighted (a, b) and corresponding FLAIR (c, and d) images from two different levels demonstrate multiple white matter hyperintensities, predominantly in the periventricular regions (arrows). Thalami show normal signal on T2-weighted image (e) and ADC map (f)
were conducted among subjects with normal cranial MRI findings instead of normal healthy controls, which is another limitation of our study. In addition, two different MRI systems were used in our study; even we know that ADC values are a
constant. Further studies with increased specificity and reproducibility are needed. Younger age of the patient, female predilection, oval shape and periventricular, pericallosal, infratentorial and
Neuroradiology Table 2 Mean ADC numbers in patients with CSVD, MS, and controls CSVD CSVD control p value MS MS control CSVD cerebral small vessel disease, MS multiple sclerosis, ADC-R apparent diffusion coefficient of the right thalamus, ADC-L apparent diffusion coefficient of the left thalamus, ADC-M mean apparent diffusion coefficient
p value CSVD MS p value CSVD control MS control p value
Number
Mean ADC-R
Mean ADC-L
Mean ADC-M
50 50
0.98±0.17×10−3 0.78±0.07×10−3
