Correlation of DTI-Derived Measures to Therapy-Mediated Recovery after Stroke: Preliminary Findings
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.329584
Source of Support: None, Conflict of Interest: None
Keywords: DTI, recovery, rehabilitation, stroke
Stroke is the second leading cause of chronic disability and mortality with 102 million disability-adjusted life years lost annually. While the incidence of stroke is decreasing in the developed world, it has peaked in India due to demographic transition and rapid shift in the socioeconomic milieu. The estimated adjusted prevalence rate of stroke reported a range of 84 to 262/100,000 in rural and 334–424/100,000 in urban India
Motor impairments in stroke often present in the form of hemiparesis and/or plegia of the contralateral side, coordination dysfunction, and gait abnormalities., The neurophysiology of motor control in humans is mediated by the pyramidal system i.e., corticospinal, rubrospinal, and reticulospinal tracts, of which corticospinal tract (CST) is the main descending fiber bundle that has been studied widely.
Evidence from clinical studies suggests that the integrity of the descending connections from the ipsilesional motor system is critical for restitution of motor function after stroke. Conventional magnetic resonance imaging (MRI) cannot provide reliable information about the integrity of white matter tracts, thereby limiting the ability to predict clinical outcomes. Diffusion tensor imaging (DTI) in stroke has opened up new possibilities of imaging fiber tracts in the brain based on the preferred direction and degree of water diffusion explaining microstructural properties of brain tissue. The degree of anisotropy of diffusion reflects the integrity and the degree of organization of the fiber tracts within the brain. Wallerian degeneration More Details (WD) in the chronic stage of stroke has been extensively studied and has shown that fractional anisotropy (FA) was reduced along the pyramidal tract of the affected side months to years after stroke., In addition to this, it is also evidenced that the degree of corticospinal involvement in the infarcts is related directly to stroke severity and inversely to functional recovery.
Longitudinal studies after stroke recovery suggest that approximately 50% of patients with significant arm paresis recover useful arm function within the first 3 years of stroke. The initial severity of paresis in the first month after stroke remains the best indicator of recovery of hand function.,
In the present study, we attempted to correlate the DTI-derived measures of CST to the clinical and functional outcomes after stroke and also to observe any therapy-induced changes on the DTI imaging parameters.
The data were analyzed from 20 chronic strokes (n = 20) patients which were within 3 months to 2 years; 18 to 70 years; MRC (Medical Research Council) grade of wrist and hand muscles at least 2. The exclusion criteria were progressive neurological disorders, pregnancy, and contraindications to MRI. The study was approved by the institutional ethics committee (IEC) and written informed consent was taken from all the subjects before participation in the study.
The subjects were recruited from the OPD and were examined and assessed by a neurologist and neurophysiotherapist, respectively. All of them underwent clinical evaluation and functional MR imaging at baseline (0 weeks), 8 weeks (post physiotherapy), and at 24 weeks (follow-up).
National Institute of Health Stroke Scale (NIHSS), Fugl-Meyer scale, motricity index (MI), modified Barthel Index (mBI), and Edinburgh Handedness Inventory was used to assess motor function and hand dominance activities. This data presented here was retrieved from the patients who were administered with a structured physiotherapy regime and formed a control group for our previous paper.
Blood oxygenation level-dependent (BOLD) and DTI were obtained with a 1.5 T MR scanner (Avanto, M/s. Siemens Medical Solutions, Germany) with a standard head coil. The encoding scheme consisted of 12 directions with three b values i.e., 0, 400, and 1000 s/mm2 with imaging parameters as follows: TE = 76 ms, TR = 10,726 ms, SENSE factor = 2, EPI factor = 127, NEX = 1, resolution = 128 × 128 matrix, field of view = 230 mm, and a slice thickness of 2.3 mm. The termination criteria used were FA <0.2, an angle change >30° samples per voxel length was 2, and step length 0.9 mm. The FA indices and ADC (apparent diffusion coefficient) values were calculated in the region of interest.
We calculated the FA ratio of the affected (FAah) to the unaffected hemisphere (FAuh). A threshold value of FA asymmetry of 0.2 or FA ratio of 0.6 and a tracking angle <30° were used for fiber tracking. The region of interest (ROI) is kept on the lesion with a fixed radius and the same radius was used for the non-lesioned hemisphere. Three-dimensional DTI-based color maps were later generated with the standard color-coding system (red, green, and blue colors indicating left-right, anterior-posterior, and superior-inferior directions respectively). The fiber tracking was done using tractography and DTI processing software (M/s. Siemens Medical Solutions, Erlangen Germany) which estimated the number of fibers, fiber length (mm), and several voxels in a given ROI. Seed points as a ROI were drawn in the infarcted area and CST portion i.e., precentral sulcus, posterior limb of the internal capsule, midbrain on a 2-D FA color maps of the affected and unaffected hemisphere. The selection of ROIs was repeated thrice by one rater and the average value was regarded as the unit of measurement. To have reproducible and concomitant results, we used two raters who drew the region of interest in the brain segments. The intraclass correlation coefficient (ICC) for intra-rater reliability was 0.88 and the inter-rater reliability was 0.89. Fiber tracts passing through these ROIs were designated as the final tracts of interest. We also calculated the fiber number (FN) and fiber length (FL) ratio in the given ROI of the affected hemisphere for the unaffected hemisphere (FNah/FAuh and FLah/FLuh). We also recorded the CST involvement criterion to the extent it passed the lesion as “whole,” “partial,” and “intact” as mentioned in the literature.,
The goals of the treatment were strength training, motor control and coordination, tone reduction, and functional training., It was administered for 5 days/week for 8 weeks for 60 to 90 min. All the patients received physiotherapy by the same therapist for the paretic upper and lower limbs.
The data of the study were analyzed using SPSS software (version 11.5). The within-group comparison was done using paired t-test (parametric) and Mann Whitney U-test (non-parametric) to compare the difference between two means i.e., baseline and 8 weeks; between 8 and 24 weeks. Repeated measures of ANOVA were used for repeated testing within a group.
We analyzed 20 (n = 20) chronic stroke patients data with mean age 45.5 ± 6.7 years and 18 male and 2 females. Among them, 13 (n = 13) were cortical lesions which included parietal, frontal, and temporoparietal areas and 7(n = 7) were subcortical with the capsule, caudate, anterior horn, and PLIC lesions. Twelve patients (n = 12) had right hemisphere lesion and eight patients (n = 8) were left hemispheric. Only one patient was having a hemorrhagic stroke while the remaining were ischemic in origin. We classified all 19 patients according to ischemic stroke subtypes and observed that 5 patients each formed large artery (LA) and small-vessel occlusion (SVO), 4 patients were cardioembolic (CE), 3 were undetermined (UD), and 2 were with determined etiology (OD). The mean age (±SD) of patients was 45.45 (±6.6) years. The mean time after stroke onset was 8.5 months. The male to female ratio was 18:2. The demographics and clinical scores at baseline, 8 weeks, and 24 weeks are given in [Table 1]. The mean FM score at baseline was 18.9 ± 7.6 and at 8 weeks was 29.4 ± 9.1 (t = -14.36, P = 0.001) and at follow-up, was 35.6 ± 8.5 with statistically significant improvement between 8 weeks and 24 weeks and between baseline and 24 weeks (P < 0.05). The mean mBI at baseline, 8 weeks, and 24 weeks was 46.95 ± 10.04, 58 ± 9.3, and 68.4 ± 9.2 showing significant improvement (P < 0.05).
In the DTI imaging analysis, we calculated the FA values, eigen vector values, ADC, and fiber number (FN), and the fiber length (FL) were calculated in the given region of interest (ROI) from affected and unaffected hemisphere. Anisotropy or FA which is the square root of the sum of squares (SRSS) of the diffusivity differences, divided by the SRSS of the diffusivities. The mean radial diffusivity (λ) in the affected hemisphere was 0.40 and in the unaffected hemisphere was 0.18. The axial diffusivity (λ) in the affected hemisphere 0.30 and the unaffected hemisphere was 0.14. The mean D of the affected hemisphere was 3.73 and in the unaffected hemisphere was 1.57 [Table 1].
The mean FA ratio (± SD) were 0.47 (±0.01), 0.53 (±0.08), and 0.57 (±0.06) at baseline, 8, and 24 weeks, respectively. There was an increased FA ratio at 8 weeks which correlated well with the FM, MI, and mBI scores. The mean fiber number (FN) ratio (± SD) was 0.27 (±0.01) at baseline, 0.33 (±0.04) at 8 weeks, and 0.41 (±0.03) at 24 weeks. Similarly, the fiber length (FL) ratios (± SD) were 0.22 (±0.01), 0.24 (±0.04) and 0.27 (±0.03) at baseline, 8 weeks and 24 weeks, respectively. From the percent change between the time intervals, we observed an increase of 12.7% in the FA ratio between 0 and 8 weeks and 9.11% between 8 and 24 weeks [Table 2]. There was a change of 22.6% from baseline till follow-up (24 weeks) in the FA ratio.
Fiber tracking of the CST was successful in all 20 patients [Table 2]. Group 1 lesion (intact type), in which the CST was close to the lesion but did not pass through it, included 4 patients [Figure 1], Group 2 (partial involvement type), in which the CST partly passed through the lesion, included 11 patients [Figure 2], and Group 3 (whole involvement type), in which lesion was centered in the pyramidal tract or involved a whole part of the CST, included 5 patients [Figure 3]. There was a change of 17.1% in FL ratio between baseline and 8 weeks and 10.01% between 8 and 24 weeks. We observed a change of 28% in the FL ratio between baseline and at follow-up (24 weeks). There was a change of 22.2% in the mean FN ratio from baseline to 8 weeks, 24% from 8 weeks to 24 weeks, and 51% from baseline to 24 weeks.
It has been validated that the FA index is a sensitive measure of structural change after stroke i.e., reduction in FA ratio reflects the deterioration of axonal integrity leading to axonal loss and WD., This research with a long-term follow-up of 24 weeks studying the correlation between the DTI parameters and clinical recovery after stroke. The results further reveal an association between the temporal evolution of FA indices and motor function. We observed a decline in the FA in patients with the poor motor outcome (MI) and good recovery patients had an FA index ratio of above 0.6 [Figure 4]. FM and mBI significantly improved at 8 weeks, suggesting an improvement in the motor recovery from mirror therapy with an increased BOLD activation shown in premotor and SMA. The percentage change in the mean FM score and mBI from baseline to 8 weeks was 55.08% and 23%, respectively whereas the change from 8 weeks to follow-up (24 weeks) was 21% and 17.9%, respectively. This suggested that a supervised exercise regime by a qualified therapist for 8 weeks led to greater improvement as compared when the supervised exercise regime is withdrawn (withdrawal effect of exercise).
A low FA ratio was observed in 5 patients (n = 5) at baseline and 7 patients (n = 7) at 8 weeks [Figure 5] with 12.7% change which may indicate the formation of motor tracts and neurons at microstructural levels. Repetitive exercise training administered to all patients leads to neurogenesis and axonal remodeling transforming the white matter integrity. The slightly increased FA values represent increased fiber tract “representations” after stroke. We also observed a change of 9.11% in the FA ratio between 8 and 24 weeks. It is also to be noted that 9 (n = 9) patients out of 20 recovered well at 24 weeks with FA ratio >0.6 with 22.6% change from baseline to 24 weeks.
White matter tracts of the cerebral hemispheres may be classified into three distinct types: (1) association - those that connect two different regions of the cerebral cortex within the same hemisphere; (2) projection - those that connect the cerebral cortex to subcortical structures, such as the thalamus and spinal cord; and (3) commissural - those that connect cortical regions of the left hemisphere with those of the right hemisphere.
It has been observed that decreased FA value and increased fiber number in the CST of the unaffected hemisphere appear to indicate a regenerative capacity of the CST following ICH, but the same was not found in our study owing to the chronicity of the disease in which the recovery mechanisms change over time. We observed a little change in the number of fibers from 8 weeks to 24 weeks as compared from baseline to 8 weeks indicating that a focussed exercise regime might have enhanced recovery with new growth of axons and nerve fibers leading to a better performance suggesting therapy-induced stroke recovery.
The stroke literature review also suggests that patients with pure cortical stroke have better motor outcomes than patients with purely subcortical stroke. Besides, patients with mixed (cortical and subcortical) stroke tended to do better than subcortical stroke despite the expected larger size of mixed lesions i.e., the involvement of corona radiata and posterior limb of the internal capsule (PLIC) with an understanding of redundant cortical motor representation and convergence of corticofugal motor efferents as they pass through the corona radiata to the PLIC. Our results also supported the above findings as it was evidenced by two patients with a PLIC lesion (id 14, 17 with a volume of lesion 45 and 12.2 Ml, respectively) did not improve much as compared to cortical stroke (patients id 16, 18 with an intact and partial involvement of corticospinal tracts). The LA strokes correlated with a large lesion volume (mL) than other ischemic subtypes, both LA and SVO showed more than 50% gains in FM score at 8 weeks as compared to another report which stated symptomatic ischemic strokes of the lacunar-type have peculiar clinical characteristics with a favorable short-term prognosis., Corticofugal outflow from MI, Primary motor area (PMA), and Supplementary motor area (SMA) is thought to be through PLIC, genu, and Anterior limb of internal capsule, respectively. Purely cortical strokes, therefore, do not interrupt cortical input to more primitive rubrospinal, reticulospinal, and vestibulospinal motor control systems.
Our results propel a long-standing debate regarding the relationship between infarct size and clinical outcomes of stroke. Although several studies have reported significant but moderate correlations between motor impairment and lesion size, others have argued that no statistically significant relationship exists. A recent study correlated the CST damage with impairments by calculating the CST lesion load for all patients by overlaying the lesion map with a probabalistic tract derived from DTI images of age-matched healthy subjects, but we did not use the same methodology in our study. However, our results were well-correlated with the FA ratios and the clinical status of the patients.
It is a well-accepted fact that intensive exercise regime leads to better performance of stroke patients both clinically and radiologically enhancing neural plasticity. Motor imagery or observational learning activates the primary/premotor areas which have been termed as “mirror neuron network” and our treatment was based on the basic principles of learning i.e., practice and active participation.
The ROI-based analysis of CST is a reliable and sensitive marker that has been reported. This study reported an intraclass correlation coefficient between two raters as 0.80 which was very similar to our results. A recent study reported a correlation that a larger infarct volume is associated with more pronounced tissue modifications in the chronic stage as observed with the mean diffusivity (MD) and FA alterations. Since our study included stroke from 3 months to 2 years we did not found a statistically significant correlation between the infarct volume and FA alterations (r = 0.012).
The time of symptom onset and site of the cortical lesion plays a major role in the degree of motor impairment, but may not be a potential marker of recovery., In the present study, the volume of the lesion varied from a small lacuna (5.8 mL or cc) to a large cortical lesion (45.5 mL) in patients. In our study, the location and volume of the lesion were not included the inclusion criteria to avoid any ethical issues in the small sample of this research. We recruited patients according to their functional potential (MRC grade at least 2, Brunnstrom stage 2 to 4) contrary to the site and side of the lesion, premorbid status, clinical status, the time of symptom onset, frequency of attack, acute stroke interventions, etc. It would be interesting to know whether DTI and BOLD parameters have gender differences in stroke patients. To remark a recovery pattern with only two females vis-a-vis 18 males did not let us conclude the gender basis of risk factors, stroke subtype, stroke severity, and outcome which has been proven otherwise.
It is concluded that clinical and functional recovery after stroke is well-correlated with the DTI parameters i.e., FA ratios, CST involvement, and fiber numbers at long-term follow-up. Physical therapy-induced changes correlate well with DTI imaging parameters.,
AB analyzed and wrote the manuscript; SSK helped in DTI imaging analysis, MVP, critically read the manuscript.
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms. In the form, the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.
Financial support and sponsorship
The study was funded by the Department of Science and Technology, DST, New Delhi, India.
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2]