- Email: [email protected]
Imaging Brain Iron and Diffusion Patterns
Imaging Brain Iron and Diffusion Patterns: A Follow-up Study of Parkinson’s Disease in the Initial Stages €ki, MD, PhD, Maija Elina Rossi, PhD, Hanna Ruottinen, MD, PhD, Tiia Saunama Irina Elovaara, MD, PhD, Prasun Dastidar, MD, PhD Rationale and Objectives: The aim of this study was to examine changes of brain iron content and diffusion patterns longitudinally in early-stage Parkinson’s disease (PD) patients using T2- and T2*-based magnetic resonance imaging (MRI) over 2-year follow-up. Materials and Methods: We imaged 32 PD patients with tremor and 19 healthy controls. A follow-up study (median 25 months, range 22–31 months) was accomplished for 25 patients (men:women = 11:14; age range 44–87 years, median 73 years). All patients and healthy volunteers underwent clinical, neuropsychological, and MRI examinations on the same day. Three different MRI sequences were used and their results were compared: T2-weighted imaging, susceptibility-weighted imaging, and T2* mapping. Additionally, we evaluated diffusion tensor data between groups using tract-based spatial statistics. Results: Over the 2-year follow-up, the iron-related relaxation increased in the globus pallidus anterior and the caudate nucleus and slightly in the substantia nigra pars compacta (SNc). In the globus pallidus anterior and medial SNc, the change was associated with mild cognitive impairment. In the caudate nucleus, the increase was pronounced in patients with disease onset at 67 years or older. In the SNc, medial transverse relaxation was increased, and in the thalamus, it was decreased, in patients with PD compared with healthy volunteers at 2-year follow-up. Tract-based spatial statistical data did not differ between groups based on gender or Unified Parkinson’s Disease Rating Scale, but a slight tendency to decreasing fractional anisotropy (P < .10) in the genu of corpus callosum and bilaterally in corona radiata was seen over 2 years. Conclusions: PD-related changes were found in putative iron content over 2 years. Although mild in the initial stages, these changes were consistent over MRI sequences. Rather than correlating with disease duration, the rate of changes was associated with individual characters, such as cognitive decline and age. Key Words: Brain iron; diffusion tensor imaging; magnetic resonance imaging; Parkinson’s disease. ªAUR, 2014
P
arkinson’s disease (PD) is a neurodegenerative disease that is characterized by tremor, rigidity, bradykinesia, and postural instability (1). These motor symptoms are the basis of the diagnosis. Although the disease is incurable, various treatment options are available to enhance the quality of life, and new neuroprotective agents are constantly being developed (1–3). Depending on treatment, early initiation of therapy may provide benefit for the patient (4). Therefore, early diagnosis and initiation of therapy including follow-up
Acad Radiol 2014; 21:64–71 From the Department of Radiology, Medical Imaging Centre (M.E.R., P.D.) and Department of Neurology and Rehabilitation (H.R., T.S., I.E.), Tampere University Hospital, Teiskontie 35, Post Box 2000, 33520 Tampere, Finland; and Tampere Medical School, University of Tampere, Tampere, Finland (I.E., P.D.). Received May 30, 2013; accepted September 18, 2013. This study was supported by grants for research funding from the Alfred Kordelin Foundation, the Instrumentarium Science Foundation, and Competitive Research Funding of the Tampere University Hospital, Pirkanmaa Hospital District. Address correspondence to: M.E.R. e-mail: maija.rossi@hotmail. com ªAUR, 2014 http://dx.doi.org/10.1016/j.acra.2013.09.018
64
are urgently needed. An accurate diagnosis of PD is challenging and increasingly complemented by imaging (5,6). Imaging findings in PD patients are limited, but slight changes may be found with improving imaging techniques. Imaging of PD patients has made the most progress in the area of imaging dopamine transporters using positron emission tomography and various techniques of magnetic resonance imaging (MRI) (5–7). In addition to spectroscopy, diffusion tensor imaging (DTI), and functional imaging, one of the aspects investigated with MRI is the brain iron content, which is increased in the substantia nigra pars compacta (SNc) of PD patients (5–11). The brain iron content is associated with the loss of dopamine, and their concentrations seem to correlate as earlier shown in putamen (12). Therefore, because diagnostically promising results on the loss of dopamine have been reported with positron emission tomography (5), the results may be indirectly repeated by imaging iron with nonionizing MRI. Earlier cross-sectional studies have investigated the association between brain iron content increase and disease duration. Although the iron content in SNc is increased in comparison
Academic Radiology, Vol 21, No 1, January 2014
with healthy volunteers, this has not correlated with the duration of PD symptoms (8,13). However, in substantia nigra pars reticulata (SNr), the hypointense part of substantia nigra, the iron content may increase over time during disease progression (8), while with later-onset PD, the iron content may, in contrast to early onset, decrease (10). In the putamen, reduced iron content with disease duration has been found (14,15). Although these studies agree on the majority of findings, they might be prone to individual variations and hemispheric differences (16). To our knowledge, longitudinal studies are not available. In DTI, fractional anisotropy (FA) indicates the degree of anisotropic orientation of water diffusion that is primarily induced by fiber tracts. FA can be processed by comparing two groups using tract-based spatial statistics (TBSS) (17–19). In healthy volunteers, FA usually decreases with age in white matter (WM) but may increase in gray matter (20). The relation between iron and DTI is unclear. The presence of iron in diffusion gradients may increase the measured signal decay; however, there may also be actual alterations in the tissue structure (20). In PD, the substantia nigra and frontal lobes show reduced FA in regions of interest (ROI) analysis (21,22). However, TBSS studies in this patient group are not found in literature. The aim of this study was to investigate the brain iron accumulation longitudinally over 2-year period in patients with Parkinson’s disease. The rate of such accumulation was compared with the patients’ clinical and neuropsychological data. To potentially support these tests, similar comparisons with the same group comparisons were made for DTI data using TBSS. MATERIALS AND METHODS This study included 52 patients who were referred from local health centers to the university hospital with symptoms indicative of PD. The inclusion criterion was to present two or more of the common symptoms of PD: resting tremor, bradykinesia or hypokinesia, rigidity, or postural instability. Exclusion criteria included the presence of Alzheimer’s disease or other dementia; other severe disease such as heart, lung, or gastrointestinal disease; hypofunction of liver or kidney; active cancer; neurological or psychiatric disease; a history of cerebrovascular attack; contraindications for MRI; alcohol or drug addiction; and pregnancy. All patients underwent clinical investigation and neuropsychological testing. Based on such thorough clinical investigation, PD was diagnosed in 37 patients. For this study, we only selected PD patients with tremor, leaving 33 patients. The patients underwent MRI, during which one patient withdrew because of claustrophobia, leaving 32 patients for this study (age range 42–86 years, median 71 years; men: women = 15:17). Twenty-five of 32 patients (age range 44–87 years, median 73 years, men:women = 11:14) could be included in a 2-year follow-up including MRI (median 25 months, range 22–31 months). Clinical investigation included the evaluation of Unified Parkinson’s Disease Rating
TIME-DEPENDENT IRON STORES IN PD
Scale (UPDRS) and Mini-Mental State Examination (MMSE). Based on the neuropsychological assessment, patients were classified as having mild cognitive impairment (MCI) or as being cognitively intact (23). The study also included 20 healthy volunteers. The inclusion criterion was a clinical neurological examination with no abnormal findings. The exclusion criteria were similar to those of the PD patients. One of the volunteers showed severe symptoms of dementia and was excluded, leaving 19 volunteers for this study (age range 58–80 years, median 65 years, men:women = 4:15). All subjects gave their informed consent for the study, and the study was approved by the hospital ethical committee. MRI Protocol and Analysis
All MRI examinations were performed with a 3-T Siemens TrioTim (Siemens, Erlangen, Germany). The MRI protocol included 3-dimensional T2-weighted imaging (T2WI) using the sampling perfection with application optimized contrasts using different flip-angle evolution (single-slab, 3-dimensional, T2-weighted turbo-spin-echo sequence with high sampling efficiency) acquisition method (24), susceptibility-weighted imaging (SWI) (25,26), and T2* mapping (27,28). Additionally, we performed DTI with 20 gradient encoding directions, b value of 1000, and 0 s/mm2. Imaging parameters are presented in Table 1. For the image analysis of the nonquantitative T2WI and SWI, we calculated the contrast, c, against the genu of the corpus callosum (gCC), Sa SgCC c¼ Sa þ SgCC where Sa and SgCC are the signal intensities of the concerned structure and gCC, respectively, as previously described in Rossi et al (29,30). In T2* mapping, the quantitative T2* (ms) was used. Representative images are shown in Figure 1. Analysis for Iron Content
Using ImageJ 1.42q (National Institutes of Health, Bethesda, MD, USA), ROIs were drawn in the lateral and medial SNr and SNc, red nuclei, nucleus dentatus, caudate nucleus, anterior and posterior putamen, anterior and posterior globus pallidus (GP), thalamus, and basilar pons, and, for contrast measurements, the gCC. The ROIs were drawn in a single slice where they were best seen (Figure 2). Their size was adapted to the size of the structure, but the borders of the structures were excluded to avoid partial volume effects. All images were analyzed within 3 weeks by the same person (M.E.R.) to ensure there were similar ROI selections through all patients. Although the first-year images had been analyzed earlier, they were reanalyzed to ensure similar ROI selections for the two time points because intraobserver variability may increase within a 2-year interval. 65
ROSSI ET AL
Academic Radiology, Vol 21, No 1, January 2014
TABLE 1. Imaging Parameters Parameter Slice thickness (mm) Slice gap (mm) Field of view Acquisition matrix Interpolated matrix Repetition time (ms) Echo time (ms)
T2WI
SWI
MapIt
DTI
3.0 0 172.5 230.0 384 290
1.5 0 172.5 230.0 256 182
4.0 0.8 220.0 220.0 384 384 768 768 422 4.18: 11.32: 18.46: 25.60: 32.74
3.0 3.9 230.0 230.0 128 128
3200 354
27 20
5043 94
DTI, diffusion tensor imaging; MapIt, multiple echo sequence for T2* measurement; SWI, susceptibility-weighted imaging; T2WI, T2-weighted imaging.
Figure 1. T2* map (a), T2-weighted imaging (b), and susceptibility-weighted imaging (c) of a 59-year-old male patient with 3-month duration of symptoms of Parkinson’s disease at the baseline.
Figure 2. Regions of interest in a T2-weighted imaging of a 54-year-old female patient with 3.5-year duration of symptoms at the 2-year follow-up. (a) Red nucleus (1), substantia nigra pars compacta medial (2) and lateral (3), and pars reticulata medial (4) and lateral (5), and cerebral peduncle (6). (b) caudate nucleus (7), putamen anterior (8) and posterior (9), globus pallidus anterior (10) and posterior (11), and thalamus (12).
66
Academic Radiology, Vol 21, No 1, January 2014
TIME-DEPENDENT IRON STORES IN PD
Figure 3. Tissue contrast against the genu of corpus callosum in T2-weighted imaging (T2WI) (a), susceptibility-weighted imaging (SWI) (b), and R2* (c) in healthy volunteers (n = 19) and patients (n = 25) at the baseline and at the 2-year follow-up. The bars represent 95% confidence interval. CN, caudate nucleus; CP, cerebral peduncle; GPa, globus pallidus anterior; SNc, substantia nigra pars compacta. *P < .10, **P < .05 between baseline and follow-up.
TBSS Analysis
The TBSS method refers to group comparisons of FA maps that can be derived from DTI. All of the following analyses were computed with the Software Library of Functional Magnetic Resonance Imaging of the Brain (FMRIB) Analysis Group (17), specifically TBSS (18). The FA images were brain extracted with use of a brain extraction tool (19). The extracted images were nonlinearly registered into a common space, on the basis of which the software calculated a mean FA map of all patients and a skeletonized FA image, representing the centers of the main tracts. Statistical Analysis
Statistical analysis was performed using the PASW Statistics 20.0.0 (SPSS Inc, Chicago, IL, USA). The normal distribution of data was confirmed using the one-sample Kolmogorov–Smirnov test. Differences between patients and healthy volunteers were calculated with the Student’s t-test for unrelated samples. The baseline and follow-up MRIs were compared using the Student’s t-test for paired samples. Correlation between MRI data and disease duration was assessed with the Pearson’s correlation test. The statistical testing between MRI results and clinical data was computed using the Student’s t-test for unrelated samples. Groups were formed based on gender, age at disease onset, UPDRS III at the baseline MRI, and MCI. The age limit was set at 67 years and UPDRS III at 20, both based on the patient cohort, to make approximately similar-sized groups. Bonferroni post-hoc correction was used to account for multiple comparisons, and the significance level was set at P < .05. To further decrease the possibility of false-positive results, we focused on results that were similar between two or three sequences.
TABLE 2. P Values for Comparison between the Baseline and 2-Year Follow-up of PD Patients Region
T2WI
SWI
R2*
GP anterior Caudate nucleus SNc lateral SNc medial Cerebral peduncle
.003* .489* .749* .031* .073y
.127* .599* .847* .322* 1.000y
.056* .081* 1.000y .943y .095y
GP, globus pallidus; PD, Parkinson’s disease; SNc, substantia nigra compacta; SWI, susceptibility-weighted imaging; T2WI, T2-weighted imaging. Student’s t-test for paired samples. The values have been corrected for multiple comparisons. *Increasing relaxation. y Decreasing relaxation. Statistically significant results (p > .05) in bold.
In TBSS, voxelwise comparisons were computed with use of an unpaired t-test to conduct comparisons between groups based on patient gender, age older than 67 at disease onset, UPDRS III lower than 20 at the baseline MRI session, and the presence of MCI. We also compared patients with healthy volunteers at both time points and, longitudinally, the MRI data of the patients at baseline and follow-up session.
RESULTS Longitudinal MRI Analysis of Iron Content
In specific structures, the signal average of the two hemispheres dropped over the 2-year follow-up, corresponding to decreasing T2 and T2*, that is, increasing R2 and R2*. These data are presented in Figure 3 and the corresponding 67
ROSSI ET AL
Academic Radiology, Vol 21, No 1, January 2014
Figure 4. Longitudinal changes in groups based on (a) age as younger (Y) or older (O) than 67 years at disease onset, (b) gender as male (M) or female (F), (c) cognitive status as cognitively intact (CI) or with mild cognitive impairment (MCI), and (d) Unified Parkinson’s Disease Rating Scale score as higher (H) or lower (L) than 20 at the baseline. CN, caudate nucleus; CP, cerebral peduncle; GPa, globus pallidus anterior; SNcl, substantia nigra pars compacta lateral; SNcm, SNc medial; SWI, susceptibility-weighted imaging; T2WI, T2-weighted imaging. *P < .05, **P < .10.
P values in Table 2. Increasing transverse relaxation over time was consistent in all three sequences in the GP anterior and caudate nucleus, whereby the increase in the GP anterior was more significant (Figure 3, Table 2). In the SNc, slightly increasing relaxation over 2 years was found in both T2WI and SWI (Figure 3a and 3b); however, only T2WI-based data of the medial part showed statistical significance (Table 2). Tendency to decreasing relaxation in the cerebral peduncles was shown by two sequences (Figure 3, Table 2). Increasing hypointensity of the GP posterior in T2WI (P = .047) during 2 years was not supported by other sequences. Other ROIs did not reveal significant iron-related signal changes over the 2-year interval.
Differences between groups are presented in Figure 4. In the GP anterior, where all three sequences showed increasing iron-related relaxation with disease progression as a whole group, this was associated with the presence of MCI, and possibly the male gender. The increase in caudate nucleus was stronger in patients with disease onset at 67 years or older. In SNc medial, where for the whole group the increasing relaxation over time was very modest, this increase was significant only in patients with MCI. In the cerebral peduncle, the same sequences showing decreasing relaxation for the whole group showed that this decrease was associated with disease onset at 67 years or younger, and possibly MCI. Comparison with Healthy Volunteers
Comparison with Clinical Data
Disease duration was not correlated with MRI parameters. Only in the caudate nucleus was weak correlation (r = 0.450, P = .042) found in R2* at the follow-up. In most of the structures, age was not significantly correlated with the MRI parameters. Only in the putamen posterior did two sequences show inverse correlation: in T2WI (r = 0.472, P = .027) and SWI (r = 0.451, P = .037). This correlation was present at the follow-up session but not at the baseline. We were unable to find correlation between MRI parameters and neither UPDRS III nor its change. MMSE did not significantly correlate with MRI parameters. 68
SWI showed hypointensity in PD patients compared with the healthy volunteers in both anterior GP (P = .040) and medial SNc (P = .007) at the 2-year follow-up. In the caudate nuclei, SWI showed hypointensity at both baseline (P = .047) and follow-up (P = .027). Also, R2* and T2WI showed increased relaxation but not at a statistically significant level. The differences are shown in Figure 3. In contrast to increasing relaxation, in the thalamus, 2-year follow-up (P = .006) showed decreased R2*. This was, however, not supported by the two other imaging sequences. Also in the SNr of patients, R2* was lower in patients than in healthy volunteers in both medial (P = .040) and lateral (P = .033) parts at the
Academic Radiology, Vol 21, No 1, January 2014
TIME-DEPENDENT IRON STORES IN PD
Figure 5. Mean fractional anisotropy image of all patients (grayscale) at the 2-year follow-up MRI, overlapped with the mean skeleton (green), on base of which the statistical group comparisons are calculated. Significant differences between groups are presented in red-yellow. The image showing differences based on patient age at disease onset (P < .05) is focused on the right thalamus (a) and that based on MCI (P < .1) is focused on the left capsula externa (b), both structures marked with an arrow. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
2-year follow-up, but this was not supported by T2WI or SWI. Of all these results, none of the results were significant after Bonferroni correction for multiple comparisons. TBSS Data
We found tendency to decreasing FA values during the 2-year follow-up in the gCC and bilaterally in the anterior corona radiata (P < .10). No P values lower than 0.05 were recorded. There were no differences between healthy controls and patients at either imaging time point. Age at disease onset was cut at older than 67 years, revealing that younger age at disease onset was associated (P < .05) with higher FA values in large areas in peripherical WM tracts: bilaterally in the frontal and parietal lobes and on the right side in the temporal and occipital lobes (Figure 5a). Additionally, similar difference was found in the right thalamus (Figure 5a). In patients with MCI, the FA values were slightly higher (P < .10) than those in cognitively intact patients in the left
corona radiata and capsula externa (Figure 5b). The figures present the state at the 2-year follow-up MRI and these were closely similar to the first year. Group comparisons between men and women or high and low UPDRS III score revealed no differences. DISCUSSION Several authors have studied the relation between disease duration and MRI-based putative iron content in PD. However, to date, longitudinal studies on brain iron deposition over time have not been reported (7,8). Therefore, we decided to perform a 2-year follow-up study on patients with symptoms of PD. Of the 36 patients initially diagnosed with PD, 25 agreed to attend a 2-year follow-up including MRI. Patients and healthy volunteers were imaged with clinical sequences that have previously been shown to correlate with nonhemin iron concentration of the landmark study by Hallgren and Sourander (29,31). 69
ROSSI ET AL
Increased PD-related putative iron content has been previously shown in SNc (8,32). In several structures, there are contradictions between possible increase and decrease. Such debated structures include GP (15,33), putamen (15,33,34), and caudate nucleus (15,35). Our data indicated that in the SNc, the putative iron content increase seems to be limited in the initial stages of the disease. However, in the medial part the content seemed to slightly grow over the 2-year follow-up in T2WI and SWI. The contrasting result of R2* at the follow-up (Fig 3c) might be explained by image distortion, combined with small ROIs. Similar behavior of the iron content was shown by GP anterior and in the caudate nucleus. Here, in the caudate nucleus, putative iron content may be slightly increased already from the early stages and may continue to slowly increase during the 2-year period. In contrary, slight indications of decreased iron content in cerebral peduncle was measured in all sequences over time, but even at the 2-year follow-up the difference to healthy volunteers was not significant. The eventual decrease may be affected by several factors as discussed in the following. Instead, in the SNc and GP anterior, corresponding putative iron increase is probably less affected by outer factors as these, especially in SNc, are continuously found throughout different studies. Additionally, in the anterior GP, the decrease is consistent and shown by all sequences in our study. In SNr and thalamus, similar decreases were found in PD patients compared with healthy volunteers, but as these decreases were not consistent over the sequences, we did not consider these changes significant. In the debated putamen, the patient cohort did not show any changes in either increasing or decreasing brain iron reserves. A previous study found no correlation between SNc hypointensity and disease duration (8). The studies of Wallis et al (34) and Zhang et al (36) do not state whether their ROI is located in pars compacta or reticulata, but their figures, with correlation between MRI findings and the iron content earlier measured by Hallgren and Sourander (31), indicate that they have probably measured pars reticulata and found no correlation with disease duration (34,36). In our patient cohort, neither pars compacta nor pars reticulata was correlated with disease duration. The limited signal changes in SNc are probably due to our study design with early PD. As the disease progresses, the iron stores are significantly higher than those of age-matched controls (5–11,32). In putamen, longer disease duration has been previously associated with higher T2 values (15), but not all studies agree on this issue (8,33), including ours. Although the literature gives MRI correlations with disease duration, there is a lack of longitudinal studies. In our data, decreasing relaxation over time was shown in several structures (Table 2, Fig 3), but only in the caudate nucleus was weak correlation found between R2* and disease duration. In several structures, there were associations between clinical data and putative changes in the brain iron content. Mild cognitive impairment predicted faster changes in each anterior GP, medial SNc, and cerebral peduncle. Late disease onset was associated with an increase 70
Academic Radiology, Vol 21, No 1, January 2014
in the putative iron content in the caudate nucleus and with a decrease in the cerebral peduncle. The presence of longitudinal changes in iron-related MRI parameters, however, combined with the lack of correlation between disease duration and putative iron content indicates that the rate of iron-related signal changes probably would not be predicted by disease duration. Instead, the rate of these changes might be affected by individual characteristics, such as age at disease onset or cognitive decline. TBSS has shown the well-known decline of various WM tracts, that is, lower FA, in older individuals (37). In PD patients compared with healthy controls, decreasing FA has been shown bilaterally in the frontal lobes and the supplementary motor area (22). Conversely, PDrelated increasing FA in the right frontal WM has been suggested (38). This is partly supported by an increase in the 25th percentile of the FA histogram, but the result was not localized in the study (39). Our TBSS results, based on DTI, revealed that patients with earlier disease onset had higher FA in various WM areas and thalamus. As our patient cohort consisted of patients in the initial stages of the disease, the age at disease onset is strongly correlated with actual patient age and resulting declination of WM tracts. However, comparing our results with those of healthy volunteers (37), thalamic FA decrease might be related to PD as thalamic age-related FA changes are not reported in healthy individuals. However, changes that we found in FA were not spatially associated with the changes in iron content, and further discussion in this particular study is therefore omitted. Four issues may affect the results of this study. First, outer factors that may deteriorate the measurement of iron content include several factors, such as the possibly increasing water content relating to neuronal loss and decreasing R2 and R2* (40), or calcifications and blood oxygenation changes that may appear similar to iron-induced signal changes but supposedly are primarily related to cerebral lesions (41,42), although microbleeds may be present in the deep gray matter structures. However, iron content has been shown to strongly correlate with various MRI sequences in deep gray matter (8,23,43). Additionally, MRI–iron correlations have been proved in multiple sclerosis lesions postmortem (44) and in liver biopsies (45). Second, ROI localization in the T2* maps is challenging because of unavoidable image distortion. This is reflected by the large error bars in Figure 3c compared with those of T2WI and SWI. In par, the area of substantia nigra is suffering, and unfortunately, this indeed is an area of great interest in PD. Third, it is well known that nonhemin iron deposits increase with age from birth until the third to sixth decades of age, depending on the structure (31). However, we did not correct for these changes in this study. As only in the posterior putamen weak correlation with age was found—and even this only at the second MRI session—it may be assumed that the age dependence of our results is minimal in these elderly patients. Fourth, the TBSS analysis did not account for the effect of crossing fibers. As the main focus of our study was to analyze iron content
Academic Radiology, Vol 21, No 1, January 2014
changes, the FA should be analyzed in more detail in future studies. CONCLUSION Comparing several sequences, slight consistency in putative iron content changes were found. Instead of correlating with disease duration, the rate of the changes was mildly affected by various individual characters, such as age at disease onset and eventual cognitive decline. REFERENCES 1. Davie CA. A review of Parkinson’s disease. Br Med Bull 2008; 86:109–127. 2. Hart RG, Pearce LA, Ravina BM, et al. Neuroprotection trials in Parkinson’s disease: systematic review. Mov Disord 2009; 24:647–654. 3. Massano J, Garrett C. Deep brain stimulation and cognitive decline in Parkinson’s disease: a clinical review. Front Neurol 2012; 3:66. 4. Olanow CW, Rascol O, Hauser R, et al., ADAGIO Study Investigators. A double-blind, delayed-start trial of rasagiline in Parkinson’s disease. N Engl J Med 2009; 361:1268–1278. 5. Stoessl AJ, Martin WW, McKeown MJ, et al. Advances in imaging in Parkinson’s disease. Lancet Neurol 2011; 10:987–1001. 6. Mascalchi M, Vella A, Ceravolo R. Movement disorders: role of imaging in diagnosis. J Magn Reson Imaging 2012; 35:239–256. 7. Martin WR. Quantitative estimation of regional brain iron with magnetic resonance imaging. Parkinsonism Relat Disord 2009; 15:215–218. 8. Martin WR, Wieler M, Gee M. Midbrain iron content in early Parkinson disease: a potential biomarker of disease status. Neurology 2008; 70: 1411–1417. 9. Dexter DT, Carayon A, Javoy-Agid F, et al. Alterations in the levels of iron, ferritin and other trace metals in Parkinson’s disease and other neurodegenerative diseases affecting the basal ganglia. Brain 1991; 114: 1953–1975. 10. Bartzokis G, Cummings JL, Markham CH, et al. MRI evaluation of brain iron in earlier- and later-onset Parkinson’s disease and normal subjects. Magn Reson Imaging 1999; 17:213–222. 11. Du G, Lewis MM, Styner M, et al. Combined R2* and diffusion tensor imaging changes in the substantia nigra in Parkinson’s disease. Mov Disord 2011; 26:1627–1632. € tz ME, Double K, Gerlach M, et al. The relevance of iron in the pathogen12. Go esis of Parkinson’s disease. Ann N Y Acad Sci 2004; 1012:193–208. 13. Antonini A, Leenders KL, Meier D, et al. T2 relaxation time in patients with Parkinson’s disease. Neurology 1993; 43:697–700. €newald RA, et al. Brain iron deposition in Par14. Graham JM, Paley MN, Gru kinson’s disease imaged using the PRIME magnetic resonance sequence. Brain 2000; 123:2423–2431. 15. Kosta P, Argyropoulou MI, Markoula S, et al. MRI evaluation of the basal ganglia size and iron content in patients with Parkinson’s disease. J Neurol 2006; 253:26–32. 16. Xu X, Wang Q, Zhang M. Age, gender, and hemispheric differences in iron deposition in the human brain: an in vivo MRI study. Neuroimage 2008; 40: 35–42. 17. Smith SM, Jenkinson M, Woolrich MW, et al. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 2004; 23:208–219. 18. Smith SM, Jenkinson M, Johansen-Berg H, et al. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 2006; 31:1487–1505. 19. Smith SM. Fast robust automated brain extraction. Hum Brain Mapp 2002; 17:143–155. 20. Pfefferbaum A, Adalsteinsson E, Rohlfing T, et al. Diffusion tensor imaging of deep gray matter brain structures: effects of age and iron concentration. Neurobiol Aging 2010; 31:482–493. 21. Vaillancourt DE, Spraker MB, Prodoehl J, et al. High-resolution diffusion tensor imaging in the substantia nigra of de novo Parkinson disease. Neurology 2009; 72:1378–1384.
TIME-DEPENDENT IRON STORES IN PD
22. Karagulle Kendi AT, Lehericy S, Luciana M, et al. Altered diffusion in the frontal lobe in Parkinson disease. AJNR Am J Neuroradiol 2008; 29:501–505. € ster AI, et al. Diagnostic criteria for mild cognitive 23. Litvan I, Goldman JG, Tro impairment in Parkinson’s disease: Movement Disorder Society Task Force guidelines. Mov Disord 2012; 27:349–356. 24. Lichy MP, Wietek BM, Mugler JP, 3rd, et al. Magnetic resonance imaging of the body trunk using a single-slab, 3-dimensional, T2-weighted turbo-spinecho sequence with high sampling efficiency (SPACE) for high spatial resolution imaging: initial clinical experiences. Invest Radiol 2005; 40:754–760. 25. Haacke EM, Mittal S, Wu Z, et al. Susceptibility-weighted imaging: technical aspects and clinical applications, part 1. AJNR Am J Neuroradiol 2009; 30:19–30. 26. Mittal S, Wu Z, Neelavalli J, et al. Susceptibility-weighted imaging: technical aspects and clinical applications, part 2. AJNR Am J Neuroradiol 2009; 30:232–252. 27. Welsch GH, Trattnig S, Hughes T, et al. T2 and T2* mapping in patients after matrix-associated autologous chondrocyte transplantation: initial results on clinical use with 3.0-Tesla MRI. Eur Radiol 2010; 20:1515–1523. 28. Welsch GH, Mamisch TC, Hughes T, et al. In vivo biochemical 7.0 Tesla magnetic resonance: preliminary results of dGEMRIC, zonal T2, and T2* mapping of articular cartilage. Invest Radiol 2008; 43:619–626. 29. Rossi M, Ruottinen H, Elovaara I, et al. Brain iron deposition and sequence characteristics in parkinsonism: comparison of SWI, T2* Maps, T2-weighted- and FLAIR-SPACE. Invest Radiol 2010; 45:795–802. 30. Rossi M, Ruottinen H, Soimakallio S, et al. Clinical MRI for iron detection in Parkinson’s disease. Clin Imaging 2013; 37:631–636. 31. Hallgren B, Sourander P. The effect of age on the non-haemin iron in the human brain. J Neurochem 1958; 3:41–51. € ger A, Berg D. Does structural neuroimaging reveal a disturbance of 32. Gro iron metabolism in Parkinson’s disease? Implications from MRI and TCS studies. J Neural Transm 2012; 119:1523–1528. 33. Ryvlin P, Broussolle E, Piollet H, et al. Magnetic resonance imaging evidence of decreased putamenal iron content in idiopathic Parkinson’s disease. Arch Neurol 1995; 52:583–588. 34. Wallis LI, Paley MN, Graham JM, et al. MRI assessment of basal ganglia iron deposition in Parkinson’s disease. J Magn Reson Imaging 2008; 28: 1061–1067. 35. Zhang W, Sun SG, Jiang YH, et al. Determination of brain iron content in patients with Parkinson’s disease using magnetic susceptibility imaging. Neurosci Bull 2009; 25:353–360. 36. Zhang J, Zhang Y, Wang J, et al. Characterizing iron deposition in Parkinson’s disease using susceptibility-weighted imaging: an in vivo MR study. Brain Res 2010; 1330:124–130. €ckman L, et al. Age-related differences in 37. Burzynska AZ, Preuschhof C, Ba white matter microstructure: region-specific patterns of diffusivity. Neuroimage 2010; 49:2104–2112. 38. Haller S, Badoud S, Nguyen D, et al. Individual detection of patients with Parkinson disease using support vector machine analysis of diffusion tensor imaging data: initial results. AJNR Am J Neuroradiol 2012; 33: 2123–2128. 39. Tessa C, Giannelli M, Della Nave R, et al. A whole-brain analysis in de novo Parkinson disease. AJNR Am J Neuroradiol 2008; 29:674–680. 40. Haacke EM, Cheng NY, House MJ, et al. Imaging iron stores in the brain using magnetic resonance imaging. Magn Reson Imaging 2005; 23:1–25. 41. Deistung A, Schweser F, Wiestler B, et al. Quantitative susceptibility mapping differentiates between blood depositions and calcifications in patients with glioblastoma. PLoS One 2013; 8:e57924. 42. Li M, Hu J, Miao Y, et al. In vivo measurement of oxygenation changes after stroke using susceptibility weighted imaging filtered phase data. PLoS One 2013; 8:e63013. 43. Haller S, Badoud S, Nguyen D, et al. Differentiation between Parkinson disease and other forms of Parkinsonism using support vector machine analysis of susceptibility-weighted imaging (SWI): initial results. Eur Radiol 2013; 23:12–19. 44. Walsh AJ, Lebel RM, Eissa A, et al. Multiple sclerosis: validation of MR imaging for quantification and detection of iron. Radiology 2013; 267: 531–542. 45. Wood JC, Enriquez C, Ghugre N, et al. MRI R2 and R2* mapping accurately estimates hepatic iron concentration in transfusion-dependent thalassemia and sickle cell disease patients. Blood 2005; 106:1460–1465.
71
P
arkinson’s disease (PD) is a neurodegenerative disease that is characterized by tremor, rigidity, bradykinesia, and postural instability (1). These motor symptoms are the basis of the diagnosis. Although the disease is incurable, various treatment options are available to enhance the quality of life, and new neuroprotective agents are constantly being developed (1–3). Depending on treatment, early initiation of therapy may provide benefit for the patient (4). Therefore, early diagnosis and initiation of therapy including follow-up
Acad Radiol 2014; 21:64–71 From the Department of Radiology, Medical Imaging Centre (M.E.R., P.D.) and Department of Neurology and Rehabilitation (H.R., T.S., I.E.), Tampere University Hospital, Teiskontie 35, Post Box 2000, 33520 Tampere, Finland; and Tampere Medical School, University of Tampere, Tampere, Finland (I.E., P.D.). Received May 30, 2013; accepted September 18, 2013. This study was supported by grants for research funding from the Alfred Kordelin Foundation, the Instrumentarium Science Foundation, and Competitive Research Funding of the Tampere University Hospital, Pirkanmaa Hospital District. Address correspondence to: M.E.R. e-mail: maija.rossi@hotmail. com ªAUR, 2014 http://dx.doi.org/10.1016/j.acra.2013.09.018
64
are urgently needed. An accurate diagnosis of PD is challenging and increasingly complemented by imaging (5,6). Imaging findings in PD patients are limited, but slight changes may be found with improving imaging techniques. Imaging of PD patients has made the most progress in the area of imaging dopamine transporters using positron emission tomography and various techniques of magnetic resonance imaging (MRI) (5–7). In addition to spectroscopy, diffusion tensor imaging (DTI), and functional imaging, one of the aspects investigated with MRI is the brain iron content, which is increased in the substantia nigra pars compacta (SNc) of PD patients (5–11). The brain iron content is associated with the loss of dopamine, and their concentrations seem to correlate as earlier shown in putamen (12). Therefore, because diagnostically promising results on the loss of dopamine have been reported with positron emission tomography (5), the results may be indirectly repeated by imaging iron with nonionizing MRI. Earlier cross-sectional studies have investigated the association between brain iron content increase and disease duration. Although the iron content in SNc is increased in comparison
Academic Radiology, Vol 21, No 1, January 2014
with healthy volunteers, this has not correlated with the duration of PD symptoms (8,13). However, in substantia nigra pars reticulata (SNr), the hypointense part of substantia nigra, the iron content may increase over time during disease progression (8), while with later-onset PD, the iron content may, in contrast to early onset, decrease (10). In the putamen, reduced iron content with disease duration has been found (14,15). Although these studies agree on the majority of findings, they might be prone to individual variations and hemispheric differences (16). To our knowledge, longitudinal studies are not available. In DTI, fractional anisotropy (FA) indicates the degree of anisotropic orientation of water diffusion that is primarily induced by fiber tracts. FA can be processed by comparing two groups using tract-based spatial statistics (TBSS) (17–19). In healthy volunteers, FA usually decreases with age in white matter (WM) but may increase in gray matter (20). The relation between iron and DTI is unclear. The presence of iron in diffusion gradients may increase the measured signal decay; however, there may also be actual alterations in the tissue structure (20). In PD, the substantia nigra and frontal lobes show reduced FA in regions of interest (ROI) analysis (21,22). However, TBSS studies in this patient group are not found in literature. The aim of this study was to investigate the brain iron accumulation longitudinally over 2-year period in patients with Parkinson’s disease. The rate of such accumulation was compared with the patients’ clinical and neuropsychological data. To potentially support these tests, similar comparisons with the same group comparisons were made for DTI data using TBSS. MATERIALS AND METHODS This study included 52 patients who were referred from local health centers to the university hospital with symptoms indicative of PD. The inclusion criterion was to present two or more of the common symptoms of PD: resting tremor, bradykinesia or hypokinesia, rigidity, or postural instability. Exclusion criteria included the presence of Alzheimer’s disease or other dementia; other severe disease such as heart, lung, or gastrointestinal disease; hypofunction of liver or kidney; active cancer; neurological or psychiatric disease; a history of cerebrovascular attack; contraindications for MRI; alcohol or drug addiction; and pregnancy. All patients underwent clinical investigation and neuropsychological testing. Based on such thorough clinical investigation, PD was diagnosed in 37 patients. For this study, we only selected PD patients with tremor, leaving 33 patients. The patients underwent MRI, during which one patient withdrew because of claustrophobia, leaving 32 patients for this study (age range 42–86 years, median 71 years; men: women = 15:17). Twenty-five of 32 patients (age range 44–87 years, median 73 years, men:women = 11:14) could be included in a 2-year follow-up including MRI (median 25 months, range 22–31 months). Clinical investigation included the evaluation of Unified Parkinson’s Disease Rating
TIME-DEPENDENT IRON STORES IN PD
Scale (UPDRS) and Mini-Mental State Examination (MMSE). Based on the neuropsychological assessment, patients were classified as having mild cognitive impairment (MCI) or as being cognitively intact (23). The study also included 20 healthy volunteers. The inclusion criterion was a clinical neurological examination with no abnormal findings. The exclusion criteria were similar to those of the PD patients. One of the volunteers showed severe symptoms of dementia and was excluded, leaving 19 volunteers for this study (age range 58–80 years, median 65 years, men:women = 4:15). All subjects gave their informed consent for the study, and the study was approved by the hospital ethical committee. MRI Protocol and Analysis
All MRI examinations were performed with a 3-T Siemens TrioTim (Siemens, Erlangen, Germany). The MRI protocol included 3-dimensional T2-weighted imaging (T2WI) using the sampling perfection with application optimized contrasts using different flip-angle evolution (single-slab, 3-dimensional, T2-weighted turbo-spin-echo sequence with high sampling efficiency) acquisition method (24), susceptibility-weighted imaging (SWI) (25,26), and T2* mapping (27,28). Additionally, we performed DTI with 20 gradient encoding directions, b value of 1000, and 0 s/mm2. Imaging parameters are presented in Table 1. For the image analysis of the nonquantitative T2WI and SWI, we calculated the contrast, c, against the genu of the corpus callosum (gCC), Sa SgCC c¼ Sa þ SgCC where Sa and SgCC are the signal intensities of the concerned structure and gCC, respectively, as previously described in Rossi et al (29,30). In T2* mapping, the quantitative T2* (ms) was used. Representative images are shown in Figure 1. Analysis for Iron Content
Using ImageJ 1.42q (National Institutes of Health, Bethesda, MD, USA), ROIs were drawn in the lateral and medial SNr and SNc, red nuclei, nucleus dentatus, caudate nucleus, anterior and posterior putamen, anterior and posterior globus pallidus (GP), thalamus, and basilar pons, and, for contrast measurements, the gCC. The ROIs were drawn in a single slice where they were best seen (Figure 2). Their size was adapted to the size of the structure, but the borders of the structures were excluded to avoid partial volume effects. All images were analyzed within 3 weeks by the same person (M.E.R.) to ensure there were similar ROI selections through all patients. Although the first-year images had been analyzed earlier, they were reanalyzed to ensure similar ROI selections for the two time points because intraobserver variability may increase within a 2-year interval. 65
ROSSI ET AL
Academic Radiology, Vol 21, No 1, January 2014
TABLE 1. Imaging Parameters Parameter Slice thickness (mm) Slice gap (mm) Field of view Acquisition matrix Interpolated matrix Repetition time (ms) Echo time (ms)
T2WI
SWI
MapIt
DTI
3.0 0 172.5 230.0 384 290
1.5 0 172.5 230.0 256 182
4.0 0.8 220.0 220.0 384 384 768 768 422 4.18: 11.32: 18.46: 25.60: 32.74
3.0 3.9 230.0 230.0 128 128
3200 354
27 20
5043 94
DTI, diffusion tensor imaging; MapIt, multiple echo sequence for T2* measurement; SWI, susceptibility-weighted imaging; T2WI, T2-weighted imaging.
Figure 1. T2* map (a), T2-weighted imaging (b), and susceptibility-weighted imaging (c) of a 59-year-old male patient with 3-month duration of symptoms of Parkinson’s disease at the baseline.
Figure 2. Regions of interest in a T2-weighted imaging of a 54-year-old female patient with 3.5-year duration of symptoms at the 2-year follow-up. (a) Red nucleus (1), substantia nigra pars compacta medial (2) and lateral (3), and pars reticulata medial (4) and lateral (5), and cerebral peduncle (6). (b) caudate nucleus (7), putamen anterior (8) and posterior (9), globus pallidus anterior (10) and posterior (11), and thalamus (12).
66
Academic Radiology, Vol 21, No 1, January 2014
TIME-DEPENDENT IRON STORES IN PD
Figure 3. Tissue contrast against the genu of corpus callosum in T2-weighted imaging (T2WI) (a), susceptibility-weighted imaging (SWI) (b), and R2* (c) in healthy volunteers (n = 19) and patients (n = 25) at the baseline and at the 2-year follow-up. The bars represent 95% confidence interval. CN, caudate nucleus; CP, cerebral peduncle; GPa, globus pallidus anterior; SNc, substantia nigra pars compacta. *P < .10, **P < .05 between baseline and follow-up.
TBSS Analysis
The TBSS method refers to group comparisons of FA maps that can be derived from DTI. All of the following analyses were computed with the Software Library of Functional Magnetic Resonance Imaging of the Brain (FMRIB) Analysis Group (17), specifically TBSS (18). The FA images were brain extracted with use of a brain extraction tool (19). The extracted images were nonlinearly registered into a common space, on the basis of which the software calculated a mean FA map of all patients and a skeletonized FA image, representing the centers of the main tracts. Statistical Analysis
Statistical analysis was performed using the PASW Statistics 20.0.0 (SPSS Inc, Chicago, IL, USA). The normal distribution of data was confirmed using the one-sample Kolmogorov–Smirnov test. Differences between patients and healthy volunteers were calculated with the Student’s t-test for unrelated samples. The baseline and follow-up MRIs were compared using the Student’s t-test for paired samples. Correlation between MRI data and disease duration was assessed with the Pearson’s correlation test. The statistical testing between MRI results and clinical data was computed using the Student’s t-test for unrelated samples. Groups were formed based on gender, age at disease onset, UPDRS III at the baseline MRI, and MCI. The age limit was set at 67 years and UPDRS III at 20, both based on the patient cohort, to make approximately similar-sized groups. Bonferroni post-hoc correction was used to account for multiple comparisons, and the significance level was set at P < .05. To further decrease the possibility of false-positive results, we focused on results that were similar between two or three sequences.
TABLE 2. P Values for Comparison between the Baseline and 2-Year Follow-up of PD Patients Region
T2WI
SWI
R2*
GP anterior Caudate nucleus SNc lateral SNc medial Cerebral peduncle
.003* .489* .749* .031* .073y
.127* .599* .847* .322* 1.000y
.056* .081* 1.000y .943y .095y
GP, globus pallidus; PD, Parkinson’s disease; SNc, substantia nigra compacta; SWI, susceptibility-weighted imaging; T2WI, T2-weighted imaging. Student’s t-test for paired samples. The values have been corrected for multiple comparisons. *Increasing relaxation. y Decreasing relaxation. Statistically significant results (p > .05) in bold.
In TBSS, voxelwise comparisons were computed with use of an unpaired t-test to conduct comparisons between groups based on patient gender, age older than 67 at disease onset, UPDRS III lower than 20 at the baseline MRI session, and the presence of MCI. We also compared patients with healthy volunteers at both time points and, longitudinally, the MRI data of the patients at baseline and follow-up session.
RESULTS Longitudinal MRI Analysis of Iron Content
In specific structures, the signal average of the two hemispheres dropped over the 2-year follow-up, corresponding to decreasing T2 and T2*, that is, increasing R2 and R2*. These data are presented in Figure 3 and the corresponding 67
ROSSI ET AL
Academic Radiology, Vol 21, No 1, January 2014
Figure 4. Longitudinal changes in groups based on (a) age as younger (Y) or older (O) than 67 years at disease onset, (b) gender as male (M) or female (F), (c) cognitive status as cognitively intact (CI) or with mild cognitive impairment (MCI), and (d) Unified Parkinson’s Disease Rating Scale score as higher (H) or lower (L) than 20 at the baseline. CN, caudate nucleus; CP, cerebral peduncle; GPa, globus pallidus anterior; SNcl, substantia nigra pars compacta lateral; SNcm, SNc medial; SWI, susceptibility-weighted imaging; T2WI, T2-weighted imaging. *P < .05, **P < .10.
P values in Table 2. Increasing transverse relaxation over time was consistent in all three sequences in the GP anterior and caudate nucleus, whereby the increase in the GP anterior was more significant (Figure 3, Table 2). In the SNc, slightly increasing relaxation over 2 years was found in both T2WI and SWI (Figure 3a and 3b); however, only T2WI-based data of the medial part showed statistical significance (Table 2). Tendency to decreasing relaxation in the cerebral peduncles was shown by two sequences (Figure 3, Table 2). Increasing hypointensity of the GP posterior in T2WI (P = .047) during 2 years was not supported by other sequences. Other ROIs did not reveal significant iron-related signal changes over the 2-year interval.
Differences between groups are presented in Figure 4. In the GP anterior, where all three sequences showed increasing iron-related relaxation with disease progression as a whole group, this was associated with the presence of MCI, and possibly the male gender. The increase in caudate nucleus was stronger in patients with disease onset at 67 years or older. In SNc medial, where for the whole group the increasing relaxation over time was very modest, this increase was significant only in patients with MCI. In the cerebral peduncle, the same sequences showing decreasing relaxation for the whole group showed that this decrease was associated with disease onset at 67 years or younger, and possibly MCI. Comparison with Healthy Volunteers
Comparison with Clinical Data
Disease duration was not correlated with MRI parameters. Only in the caudate nucleus was weak correlation (r = 0.450, P = .042) found in R2* at the follow-up. In most of the structures, age was not significantly correlated with the MRI parameters. Only in the putamen posterior did two sequences show inverse correlation: in T2WI (r = 0.472, P = .027) and SWI (r = 0.451, P = .037). This correlation was present at the follow-up session but not at the baseline. We were unable to find correlation between MRI parameters and neither UPDRS III nor its change. MMSE did not significantly correlate with MRI parameters. 68
SWI showed hypointensity in PD patients compared with the healthy volunteers in both anterior GP (P = .040) and medial SNc (P = .007) at the 2-year follow-up. In the caudate nuclei, SWI showed hypointensity at both baseline (P = .047) and follow-up (P = .027). Also, R2* and T2WI showed increased relaxation but not at a statistically significant level. The differences are shown in Figure 3. In contrast to increasing relaxation, in the thalamus, 2-year follow-up (P = .006) showed decreased R2*. This was, however, not supported by the two other imaging sequences. Also in the SNr of patients, R2* was lower in patients than in healthy volunteers in both medial (P = .040) and lateral (P = .033) parts at the
Academic Radiology, Vol 21, No 1, January 2014
TIME-DEPENDENT IRON STORES IN PD
Figure 5. Mean fractional anisotropy image of all patients (grayscale) at the 2-year follow-up MRI, overlapped with the mean skeleton (green), on base of which the statistical group comparisons are calculated. Significant differences between groups are presented in red-yellow. The image showing differences based on patient age at disease onset (P < .05) is focused on the right thalamus (a) and that based on MCI (P < .1) is focused on the left capsula externa (b), both structures marked with an arrow. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
2-year follow-up, but this was not supported by T2WI or SWI. Of all these results, none of the results were significant after Bonferroni correction for multiple comparisons. TBSS Data
We found tendency to decreasing FA values during the 2-year follow-up in the gCC and bilaterally in the anterior corona radiata (P < .10). No P values lower than 0.05 were recorded. There were no differences between healthy controls and patients at either imaging time point. Age at disease onset was cut at older than 67 years, revealing that younger age at disease onset was associated (P < .05) with higher FA values in large areas in peripherical WM tracts: bilaterally in the frontal and parietal lobes and on the right side in the temporal and occipital lobes (Figure 5a). Additionally, similar difference was found in the right thalamus (Figure 5a). In patients with MCI, the FA values were slightly higher (P < .10) than those in cognitively intact patients in the left
corona radiata and capsula externa (Figure 5b). The figures present the state at the 2-year follow-up MRI and these were closely similar to the first year. Group comparisons between men and women or high and low UPDRS III score revealed no differences. DISCUSSION Several authors have studied the relation between disease duration and MRI-based putative iron content in PD. However, to date, longitudinal studies on brain iron deposition over time have not been reported (7,8). Therefore, we decided to perform a 2-year follow-up study on patients with symptoms of PD. Of the 36 patients initially diagnosed with PD, 25 agreed to attend a 2-year follow-up including MRI. Patients and healthy volunteers were imaged with clinical sequences that have previously been shown to correlate with nonhemin iron concentration of the landmark study by Hallgren and Sourander (29,31). 69
ROSSI ET AL
Increased PD-related putative iron content has been previously shown in SNc (8,32). In several structures, there are contradictions between possible increase and decrease. Such debated structures include GP (15,33), putamen (15,33,34), and caudate nucleus (15,35). Our data indicated that in the SNc, the putative iron content increase seems to be limited in the initial stages of the disease. However, in the medial part the content seemed to slightly grow over the 2-year follow-up in T2WI and SWI. The contrasting result of R2* at the follow-up (Fig 3c) might be explained by image distortion, combined with small ROIs. Similar behavior of the iron content was shown by GP anterior and in the caudate nucleus. Here, in the caudate nucleus, putative iron content may be slightly increased already from the early stages and may continue to slowly increase during the 2-year period. In contrary, slight indications of decreased iron content in cerebral peduncle was measured in all sequences over time, but even at the 2-year follow-up the difference to healthy volunteers was not significant. The eventual decrease may be affected by several factors as discussed in the following. Instead, in the SNc and GP anterior, corresponding putative iron increase is probably less affected by outer factors as these, especially in SNc, are continuously found throughout different studies. Additionally, in the anterior GP, the decrease is consistent and shown by all sequences in our study. In SNr and thalamus, similar decreases were found in PD patients compared with healthy volunteers, but as these decreases were not consistent over the sequences, we did not consider these changes significant. In the debated putamen, the patient cohort did not show any changes in either increasing or decreasing brain iron reserves. A previous study found no correlation between SNc hypointensity and disease duration (8). The studies of Wallis et al (34) and Zhang et al (36) do not state whether their ROI is located in pars compacta or reticulata, but their figures, with correlation between MRI findings and the iron content earlier measured by Hallgren and Sourander (31), indicate that they have probably measured pars reticulata and found no correlation with disease duration (34,36). In our patient cohort, neither pars compacta nor pars reticulata was correlated with disease duration. The limited signal changes in SNc are probably due to our study design with early PD. As the disease progresses, the iron stores are significantly higher than those of age-matched controls (5–11,32). In putamen, longer disease duration has been previously associated with higher T2 values (15), but not all studies agree on this issue (8,33), including ours. Although the literature gives MRI correlations with disease duration, there is a lack of longitudinal studies. In our data, decreasing relaxation over time was shown in several structures (Table 2, Fig 3), but only in the caudate nucleus was weak correlation found between R2* and disease duration. In several structures, there were associations between clinical data and putative changes in the brain iron content. Mild cognitive impairment predicted faster changes in each anterior GP, medial SNc, and cerebral peduncle. Late disease onset was associated with an increase 70
Academic Radiology, Vol 21, No 1, January 2014
in the putative iron content in the caudate nucleus and with a decrease in the cerebral peduncle. The presence of longitudinal changes in iron-related MRI parameters, however, combined with the lack of correlation between disease duration and putative iron content indicates that the rate of iron-related signal changes probably would not be predicted by disease duration. Instead, the rate of these changes might be affected by individual characteristics, such as age at disease onset or cognitive decline. TBSS has shown the well-known decline of various WM tracts, that is, lower FA, in older individuals (37). In PD patients compared with healthy controls, decreasing FA has been shown bilaterally in the frontal lobes and the supplementary motor area (22). Conversely, PDrelated increasing FA in the right frontal WM has been suggested (38). This is partly supported by an increase in the 25th percentile of the FA histogram, but the result was not localized in the study (39). Our TBSS results, based on DTI, revealed that patients with earlier disease onset had higher FA in various WM areas and thalamus. As our patient cohort consisted of patients in the initial stages of the disease, the age at disease onset is strongly correlated with actual patient age and resulting declination of WM tracts. However, comparing our results with those of healthy volunteers (37), thalamic FA decrease might be related to PD as thalamic age-related FA changes are not reported in healthy individuals. However, changes that we found in FA were not spatially associated with the changes in iron content, and further discussion in this particular study is therefore omitted. Four issues may affect the results of this study. First, outer factors that may deteriorate the measurement of iron content include several factors, such as the possibly increasing water content relating to neuronal loss and decreasing R2 and R2* (40), or calcifications and blood oxygenation changes that may appear similar to iron-induced signal changes but supposedly are primarily related to cerebral lesions (41,42), although microbleeds may be present in the deep gray matter structures. However, iron content has been shown to strongly correlate with various MRI sequences in deep gray matter (8,23,43). Additionally, MRI–iron correlations have been proved in multiple sclerosis lesions postmortem (44) and in liver biopsies (45). Second, ROI localization in the T2* maps is challenging because of unavoidable image distortion. This is reflected by the large error bars in Figure 3c compared with those of T2WI and SWI. In par, the area of substantia nigra is suffering, and unfortunately, this indeed is an area of great interest in PD. Third, it is well known that nonhemin iron deposits increase with age from birth until the third to sixth decades of age, depending on the structure (31). However, we did not correct for these changes in this study. As only in the posterior putamen weak correlation with age was found—and even this only at the second MRI session—it may be assumed that the age dependence of our results is minimal in these elderly patients. Fourth, the TBSS analysis did not account for the effect of crossing fibers. As the main focus of our study was to analyze iron content
Academic Radiology, Vol 21, No 1, January 2014
changes, the FA should be analyzed in more detail in future studies. CONCLUSION Comparing several sequences, slight consistency in putative iron content changes were found. Instead of correlating with disease duration, the rate of the changes was mildly affected by various individual characters, such as age at disease onset and eventual cognitive decline. REFERENCES 1. Davie CA. A review of Parkinson’s disease. Br Med Bull 2008; 86:109–127. 2. Hart RG, Pearce LA, Ravina BM, et al. Neuroprotection trials in Parkinson’s disease: systematic review. Mov Disord 2009; 24:647–654. 3. Massano J, Garrett C. Deep brain stimulation and cognitive decline in Parkinson’s disease: a clinical review. Front Neurol 2012; 3:66. 4. Olanow CW, Rascol O, Hauser R, et al., ADAGIO Study Investigators. A double-blind, delayed-start trial of rasagiline in Parkinson’s disease. N Engl J Med 2009; 361:1268–1278. 5. Stoessl AJ, Martin WW, McKeown MJ, et al. Advances in imaging in Parkinson’s disease. Lancet Neurol 2011; 10:987–1001. 6. Mascalchi M, Vella A, Ceravolo R. Movement disorders: role of imaging in diagnosis. J Magn Reson Imaging 2012; 35:239–256. 7. Martin WR. Quantitative estimation of regional brain iron with magnetic resonance imaging. Parkinsonism Relat Disord 2009; 15:215–218. 8. Martin WR, Wieler M, Gee M. Midbrain iron content in early Parkinson disease: a potential biomarker of disease status. Neurology 2008; 70: 1411–1417. 9. Dexter DT, Carayon A, Javoy-Agid F, et al. Alterations in the levels of iron, ferritin and other trace metals in Parkinson’s disease and other neurodegenerative diseases affecting the basal ganglia. Brain 1991; 114: 1953–1975. 10. Bartzokis G, Cummings JL, Markham CH, et al. MRI evaluation of brain iron in earlier- and later-onset Parkinson’s disease and normal subjects. Magn Reson Imaging 1999; 17:213–222. 11. Du G, Lewis MM, Styner M, et al. Combined R2* and diffusion tensor imaging changes in the substantia nigra in Parkinson’s disease. Mov Disord 2011; 26:1627–1632. € tz ME, Double K, Gerlach M, et al. The relevance of iron in the pathogen12. Go esis of Parkinson’s disease. Ann N Y Acad Sci 2004; 1012:193–208. 13. Antonini A, Leenders KL, Meier D, et al. T2 relaxation time in patients with Parkinson’s disease. Neurology 1993; 43:697–700. €newald RA, et al. Brain iron deposition in Par14. Graham JM, Paley MN, Gru kinson’s disease imaged using the PRIME magnetic resonance sequence. Brain 2000; 123:2423–2431. 15. Kosta P, Argyropoulou MI, Markoula S, et al. MRI evaluation of the basal ganglia size and iron content in patients with Parkinson’s disease. J Neurol 2006; 253:26–32. 16. Xu X, Wang Q, Zhang M. Age, gender, and hemispheric differences in iron deposition in the human brain: an in vivo MRI study. Neuroimage 2008; 40: 35–42. 17. Smith SM, Jenkinson M, Woolrich MW, et al. Advances in functional and structural MR image analysis and implementation as FSL. Neuroimage 2004; 23:208–219. 18. Smith SM, Jenkinson M, Johansen-Berg H, et al. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 2006; 31:1487–1505. 19. Smith SM. Fast robust automated brain extraction. Hum Brain Mapp 2002; 17:143–155. 20. Pfefferbaum A, Adalsteinsson E, Rohlfing T, et al. Diffusion tensor imaging of deep gray matter brain structures: effects of age and iron concentration. Neurobiol Aging 2010; 31:482–493. 21. Vaillancourt DE, Spraker MB, Prodoehl J, et al. High-resolution diffusion tensor imaging in the substantia nigra of de novo Parkinson disease. Neurology 2009; 72:1378–1384.
TIME-DEPENDENT IRON STORES IN PD
22. Karagulle Kendi AT, Lehericy S, Luciana M, et al. Altered diffusion in the frontal lobe in Parkinson disease. AJNR Am J Neuroradiol 2008; 29:501–505. € ster AI, et al. Diagnostic criteria for mild cognitive 23. Litvan I, Goldman JG, Tro impairment in Parkinson’s disease: Movement Disorder Society Task Force guidelines. Mov Disord 2012; 27:349–356. 24. Lichy MP, Wietek BM, Mugler JP, 3rd, et al. Magnetic resonance imaging of the body trunk using a single-slab, 3-dimensional, T2-weighted turbo-spinecho sequence with high sampling efficiency (SPACE) for high spatial resolution imaging: initial clinical experiences. Invest Radiol 2005; 40:754–760. 25. Haacke EM, Mittal S, Wu Z, et al. Susceptibility-weighted imaging: technical aspects and clinical applications, part 1. AJNR Am J Neuroradiol 2009; 30:19–30. 26. Mittal S, Wu Z, Neelavalli J, et al. Susceptibility-weighted imaging: technical aspects and clinical applications, part 2. AJNR Am J Neuroradiol 2009; 30:232–252. 27. Welsch GH, Trattnig S, Hughes T, et al. T2 and T2* mapping in patients after matrix-associated autologous chondrocyte transplantation: initial results on clinical use with 3.0-Tesla MRI. Eur Radiol 2010; 20:1515–1523. 28. Welsch GH, Mamisch TC, Hughes T, et al. In vivo biochemical 7.0 Tesla magnetic resonance: preliminary results of dGEMRIC, zonal T2, and T2* mapping of articular cartilage. Invest Radiol 2008; 43:619–626. 29. Rossi M, Ruottinen H, Elovaara I, et al. Brain iron deposition and sequence characteristics in parkinsonism: comparison of SWI, T2* Maps, T2-weighted- and FLAIR-SPACE. Invest Radiol 2010; 45:795–802. 30. Rossi M, Ruottinen H, Soimakallio S, et al. Clinical MRI for iron detection in Parkinson’s disease. Clin Imaging 2013; 37:631–636. 31. Hallgren B, Sourander P. The effect of age on the non-haemin iron in the human brain. J Neurochem 1958; 3:41–51. € ger A, Berg D. Does structural neuroimaging reveal a disturbance of 32. Gro iron metabolism in Parkinson’s disease? Implications from MRI and TCS studies. J Neural Transm 2012; 119:1523–1528. 33. Ryvlin P, Broussolle E, Piollet H, et al. Magnetic resonance imaging evidence of decreased putamenal iron content in idiopathic Parkinson’s disease. Arch Neurol 1995; 52:583–588. 34. Wallis LI, Paley MN, Graham JM, et al. MRI assessment of basal ganglia iron deposition in Parkinson’s disease. J Magn Reson Imaging 2008; 28: 1061–1067. 35. Zhang W, Sun SG, Jiang YH, et al. Determination of brain iron content in patients with Parkinson’s disease using magnetic susceptibility imaging. Neurosci Bull 2009; 25:353–360. 36. Zhang J, Zhang Y, Wang J, et al. Characterizing iron deposition in Parkinson’s disease using susceptibility-weighted imaging: an in vivo MR study. Brain Res 2010; 1330:124–130. €ckman L, et al. Age-related differences in 37. Burzynska AZ, Preuschhof C, Ba white matter microstructure: region-specific patterns of diffusivity. Neuroimage 2010; 49:2104–2112. 38. Haller S, Badoud S, Nguyen D, et al. Individual detection of patients with Parkinson disease using support vector machine analysis of diffusion tensor imaging data: initial results. AJNR Am J Neuroradiol 2012; 33: 2123–2128. 39. Tessa C, Giannelli M, Della Nave R, et al. A whole-brain analysis in de novo Parkinson disease. AJNR Am J Neuroradiol 2008; 29:674–680. 40. Haacke EM, Cheng NY, House MJ, et al. Imaging iron stores in the brain using magnetic resonance imaging. Magn Reson Imaging 2005; 23:1–25. 41. Deistung A, Schweser F, Wiestler B, et al. Quantitative susceptibility mapping differentiates between blood depositions and calcifications in patients with glioblastoma. PLoS One 2013; 8:e57924. 42. Li M, Hu J, Miao Y, et al. In vivo measurement of oxygenation changes after stroke using susceptibility weighted imaging filtered phase data. PLoS One 2013; 8:e63013. 43. Haller S, Badoud S, Nguyen D, et al. Differentiation between Parkinson disease and other forms of Parkinsonism using support vector machine analysis of susceptibility-weighted imaging (SWI): initial results. Eur Radiol 2013; 23:12–19. 44. Walsh AJ, Lebel RM, Eissa A, et al. Multiple sclerosis: validation of MR imaging for quantification and detection of iron. Radiology 2013; 267: 531–542. 45. Wood JC, Enriquez C, Ghugre N, et al. MRI R2 and R2* mapping accurately estimates hepatic iron concentration in transfusion-dependent thalassemia and sickle cell disease patients. Blood 2005; 106:1460–1465.
71