Feasibility Study Of Corticoreticular Tract DTI For Prediction Of Post-Stroke Spasticity

Part I DTI Fiber Tracing of the Corticoreticular TractObjectiveThis study aimed to construct the trajectory of corticoreticular tract（CRT）using the diffusion tensor imaging（DTI）method,focusing on the origin of corticoreticular tracts,terminating pattern in the brainstem,and the intermediate morphology.MethodsThirty-one healthy volunteers from Shanghai Fourth People’s Hospital affiliated to Tongji University were recruited between September 2020 and March 2021.A new cortical partitioning method of HCPMMP（Human Connectome Project Multi-Modal Parcellation）was used to determine the origin of the cortex.The brainstem nucleus that receiving CRT terminals were manually delineated according to the literatures,and then the CRT fiber projection trajectories of healthy volunteers and the corresponding DTI indicators were analyzed using the professional fiber tract tracing software DSi-studio.Results 1.CRTs were reconstructed using DTI and the corresponding DTI indicators were obtained.2.Supplementary motor area（SMA）（6ma,6mp）and premotor cortex（PM）（6d）were identified as the origins of CRTs.The giantocellular reticular nucleus（Gi）in the medulla oblongata was the main nucleus receiving CRT fiber terminals.3.The majority of CRT fibers originated from the SMA（64.9%）,and to a lesser extent from the 6d area（35.0%）.CRTs bilaterally projected to the brainstem with an ipsilateral dominance.4.QA（Quantitative anisotropy）reflected the characteristic changes of the overall trajectory of fiber bundles.FA（Fractional anisotropy）reflected changes of fiber bundles in the specific region.5.CRTs and corticospinal tracts could be distinguished using DTI.Both were in close proximity at the center of the semioval,the corona radiata and the hind limb of the internal capsule.Conclusion 1.CRT originates primarily from the SMA,and to a lesser extent from the 6d.2.CRT fibers bilaterally project to the brainstem with an ipsilateral dominance,and particularly to the Gi of the medulla oblongata.3.The QA value reflects the characteristic changes of the overall trajectory of fiber bundles.Fractional anisotropy（FA）reflects changes of fiber bundles in specific regions.4.DTI imaging can distinguish CRTs from corticospinal tracts.Part II Study on the Evolution Features of DTI of Corticoreticular Tract after StrokeObjective The purpose of this study was to use DTI to track the trajectory of the CRTs in stroke patients and to elucidate their characteristic changes of DTI indicators.Dynamic changes of CRTs at different stages of stroke were analyzed through a cross-section study and characterized in the process of stroke rehabilitation.Finally,the correlation between dynamic changes of CRT and spasticity as well as stroke prognosis was explored.Methods Patients with ischemic stroke and age-,sex-matched healthy volunteers who presented to Shanghai Fourth People’s Hospital affiliated to Tongji University between September 2020 and March 2021 were recruited.MRI images（including DWI,DTI and T1）of patients in the acute phase（duration<1month）,recovery phase（duration=3 months）,and chronic phase（duration>6months）,as well as demographic and clinical information of all subjects were collected.Spasticity was assessed using the Modified Ashworth Scale（MAS）and categorized into spasticity（MAS>=1）and non-spasticity（MAS=0）groups.The modified Rankin Scale（mRS）was used to assess patient’s prognosis and patients were divided into two groups: good prognosis mRS=<1 and poor prognosis mRS>1.The DSI-studio software was used to reconstruct CRTs,and DTI indices,such as the number of fibers and QA,were analyzed to characterize their evolutionary changes.Finally,the correlation between CRT-DTI indices and clinical variables（spasticity and stroke prognosis）was analyzed.Results 1.A total of 19 stroke patients in the acute phase,19 in the recovery phase,18 in the chronic phase,and 31 healthy volunteers were recruited.2.CRT fibers projected bilaterally with an ipsilateral dominance.No significant difference in the number of CRT fibers on both sides was found between stroke patients in the acute phase,recovery phase,and the chronic phase and healthy volunteers（P>0.05）.3.Significant difference was found in the QA value of the contralateral side to the lesion between patients at different stages of cerebral infarction（p<0.001）.With the QA value steadily increased as the course of cerebral infarction prolonged.The QA ratio（rQA,the lesion side / the contralateral side）decreased gradually（p<0.001）.The results suggest that rQA was a more sensitive index than the number of fibers,and there was a dynamic change process,which was mainly due to the gradual increase of the QA value on the contralateral side of the lesion.4.In the spasticity group,the rQA value decreased significantly（p<0.05）.5.Significant difference was found in the rQA value of the ipsilateral project CRT fibers between the good and the poor prognosis groups.Further analysis found that the lower the rQA value was,the worse the prognosis was（p<0.05）.Conclusion 1.CRTs projected bilateral with an ipsilateral predominance.This projecting pattern and the number of fibers do not change at different stages of stroke.2.rQA is a more sensitive index than the number of fibers,and there is a dynamic change process.The main reason is that the QA value of the contralateral side of the lesion gradually increases.3.The rQA of the ipsilateral projection fibers is correlated with post-stroke spasticity and stroke prognosis.The smaller the value is,the more likely the patients are to have spasticity and worse prognosis.Part III Corticoreticular Tracts Predict Post-Stroke Spasticity:a Derivation and Validation StudyObjective The purpose of this study was to analyze DTI indicators of CRTs reconstructed using DTI on stroke patients,aiming to discover predictive factors in a derived cohort,and to verify the predictive value of these factors in a validation cohort.Finally,our aim was to determine whether CRT-based DTI indicators can predict the occurrence of post-stroke spasticity（PSS）.Methods The participants were divided into a derivation cohort and a validation cohort.The derivation cohort recruited patients with acute cerebral infarction admitted to Tongji Hospital Affiliated to Tongji University between April 2013 and July 2018.The validation cohort included patients with acute cerebral infarction and age-,sex-matched healthy volunteers presented to Shanghai Fourth People’s Hospital affiliated to Tongji University between September 2020 and June 2021.All patients completed a 3-month follow-up.Demographic,clinical information and baseline MRI scans,including DWI,DTI,and T1 images,were collected.The DSI-studio software was used to track CRT fibers,and to calculate the FA,MD,AD,and RD values in the meantime.Univariate analysis was performed for each of these two cohorts to identify possible predictors,followed by correlation analysis,and subsequently by multivariate logistic regression analysis to validate these independent predictors.The ROC curve was produced to determine the best cut-off value as well as its sensitivity and specificity.Finally,examples were demonstrated to show the applicability of these predictive factors in clinical practice.Results 1.The derivation cohort included 47 patients with acute stroke,the validation cohort included 31 patients with acute stroke and 49 healthy volunteers.The incidence of PSS in the two cohorts was 57.4% and 51.6%,respectively.2.Results of the derivation cohort suggested that rFA₆ma（ratio of FA on the side of the lesion to FA on the contralateral side of the lesion in the 6ma area）can independently predict PSS（OR=0.001,95%CI,0.001-0.168,p=0.009）.The validation cohort validated the predictive effect of rFA₆ma（RR=0.001,95%CI,<0.001-0.258,p=0.02）.3.The smaller the rFA₆ma value was,the more likely the patients were to develop spasticity,and this trend was statistically significant（p<0.05）4.Results of both derivation and validation cohorts suggested that a baseline value of rFA₆ma = 1 was the best threshold for predicting PSS（derivative cohort AUC,0.82;95% CI,0.68-0.92）（validation cohort AUC,0.87;95% CI,0.69-0.96）.The sensitivity was 97.5%（95% CI,64-97）and the specificity was 70.3%（95% CI,48-89）.Conclusion 1.rFA₆ma is an independent factor in predicting spasticity after acute stroke.2.The smaller the rFA₆ma value is,the more likely the patients are to develop spasticity.3.rFA₆ma value = 1 is the best threshold for predicting PSS with a high sensitivity and specificity.