RU2801034C1 - Method of predicting cognitive disorders in patients with age-dependent cerebral microangiopathy - Google Patents

Abstract

FIELD: medicine; neurology.

SUBSTANCE: invention can be used to predict cognitive disorders in patients with age-dependent cerebral microangiopathy. An MRI of the brain is carried out in the diffusion-tensor MRI mode, followed by the construction of maps of axial diffusion and the determination of its value in the posterior middle part of the corpus callosum. If the value of individual axial diffusion is more than 0.002, a high probability of cognitive disorders is predicted. If the value of individual axial diffusion is less than 0.002, a low probability of cognitive disorders is predicted.

EFFECT: method provides the possibility of predicting cognitive disorders in patients with age-dependent cerebral microangiopathy by determining axial diffusion.

1 cl, 2 tbl, 3 dwg, 5 ex

Description

The invention relates to medicine, in particular to neurology, and can be used to diagnose cognitive disorders (CD) in patients with age-dependent cerebral microangiopathy (CMA).

Age-dependent cerebral microangiopathy (syn. - small vessel disease) is a progressive disease characterized by a wide range of pathological processes in small vessels and the development of ischemic, hemorrhagic and inflammatory lesions in the brain. Differentiation in wide clinical practice of MRI changes taking into account the diagnostic criteria for age-dependent CMA (STRIVE) showed that age-dependent CMA is one of the most common diseases in older people. The prevalence of age-dependent CMA MRI signs increases with aging and in people aged 60 to 90 years is 50-98% for unchanged white matter (WWM), with an average increase rate of 0.25 cm ³ per year, lacunae - 8-28 %, microbleeds 5–23% (Hilal S., Mok V., Youn YC, et al. Prevalence, risk factors and consequences of cerebral small vessel diseases: data from three Asian countries. J Neurol Neurosurg Psychiatry. 2017; 88(8 ): 669-674).

Age-dependent CMA makes an extremely high contribution to disability and mortality and is the cause of up to 45% of cognitive disorders (CD) - vascular and mixed with neurodegeneration, a quarter of strokes and more than doubles the risk of a recurrent cerebrovascular event (Wardlaw J.M., Smith E.E., Biessels G.J., et al. Neuroimaging standards for research into small vessel disease and its contribution to aging and neurodegeneration. Lancet Neurol 2013; 12(8): 822-838. Assessment of the rate of progression of age-dependent CMA in a large cohort (1650 patients, mean age 65 years) in the pre-dementia stage with no stroke, in the interval of 7 years, showed the onset of death in 18.5%, dementia in 5%, stroke in 4%. MRI comparisons have established an increase in risks in the initial presence of a combination of MRI signs in patients (Yilmaz P., Ikram M.K., Niessen W.J., et al. Practical Small Vessel Disease Score Relates to Stroke, Dementia, and Death. Stroke. 2018; 49(12 ): 2857-2865).

At the same time, age-dependent CMA is a disease with no pathogenetic treatment. Prevention of age-dependent CMA based on the control of arterial hypertension (AH), the main risk factor, led to a decrease in the incidence of strokes, but not the prevalence of CR. Comparison in the SPRINT (Systolic Blood Pressure Intervention Trial) randomized study, intensive versus standard BP reduction, also found no effect of hypertension treatment on the risk of dementia and a dubious effect on delayed brain damage - less white matter damage with an increase in general atrophy. Antithrombotic therapy also did not show its effectiveness in age-dependent CMA and at the same time increased the risk of hemorrhagic complications.

In this regard, an extremely urgent problem of age-dependent CMA is the availability of sensitive surrogate markers for monitoring the progression of the disease and evaluating the effectiveness of new treatments.

There is a method for determining cognitive impairment in cerebral small vessel disease (SVD) by means of MRI with an assessment of a number of signs of the disease, including lacunar infarcts, T2 lesions, brain atrophy and cerebral microhemorrhages. In addition, white matter ultrastructure disturbance was assessed by diffusion tensor imaging (DTI), and in particular, a putative marker of axonal damage was obtained by axial diffusion and a demyelination index by radial diffusion. We then determined the contribution of these whole-brain MRI markers to cognitive impairment in SVD. In a multi-predictor analysis, the number of lacunar infarctions and the diffuseness of normal-appearing white matter on DTI were independent predictors of executive function in SVD. Number of lacunar infarcts and brain atrophy were independent predictors of processing speed. Radial diffusion was a stronger predictor of DTI than axial diffusion, suggesting that ischemic demyelination seen neuropathologically in SVD may be an important predictor of cognitive impairment in SVD (Lawrence AJ, Patel B, Morris RG, et al. Mechanisms of cognitive impairment in cerebral small vessel disease: multimodal MRI results from the St George's cognition and neuroimaging in stroke (SCANS) study PLoS One 2013;8(4): e61014 https://doi.org/10.1371/journal.pone.0061014). However, the authors of this work did not set the task of predicting cognitive impairment in age-dependent cerebral microangiopathy.

At this point in time, the problem of predicting cognitive impairment in age-dependent CMA in the short and medium term has been solved only using the neuropsychological assessment of age-dependent CMA, namely, screening cognitive scales and tests. At the same time, working memory, long-term (episodic) memory, processing speed and executive function were evaluated. Annualized change in cognition was calculated for 98 patients who completed at least two time intervals, which showed their low sensitivity in the short and medium term (Lawrence AJ, Brookes RL, Zeesstraten EA, et al. Pattern and Rate of Cognitive Decline in Cerebral Small Vessel Disease: A Prospective Study PLoS One 2015;10(8):1/15-15/15).

The technical result is to increase the reliability of the prognosis of cognitive disorders in patients with age-dependent cerebral microangiopathy.

The technical result is achieved by the fact that the prediction of cognitive disorders in patients with age-dependent cerebral microangiopathy is carried out by MRI examination of the brain, while the MRI examination of the brain is carried out in the diffusion-tensor MRI mode, followed by mapping of axial diffusion and determining its value in corpus callosum, then the values of axial diffusion indicators are compared with the established threshold value of 0.002, and if the value of individual axial diffusion is more than 0.002, a high probability is predicted, and if less than 0.002, a low probability of cognitive impairment in patients with age-dependent cerebral microangiopathy.

The method is carried out as follows.

A patient with complaints of decreased concentration and memory undergoes an MRI examination of the brain in the mode of diffusion tensor magnetic resonance imaging ((Diffusion Tensor Imaging / DT-MRI). Diffusion data is obtained using a spin-echo echo-planar sequence with 3 diffusion -weighted values (b=0, 1000 and 2500 s/mm ² ) for 64 directions of encoding diffusion gradients, TE/TR 115/12600 ms, matrix 100 x 100, resolution 2 x 2 x 2 mm ³ with subsequent data processing and construction maps of diffusion metrics in the ExploreDTI program.Processing of diffusion data included the following steps: evaluation and correction of noise for all diffusion images using a non-central chi-square distribution.Next, the inhomogeneities of the external magnetic field are corrected using topup and eddy utilities (based on the FSL program The correction included correction of geometric distortions induced by "stray currents" during magnetic gradient switching and spatial correction of shifts of different diffusion images relative to each other. In accordance with the general alignment of the individual diffusion images, the corresponding directions of the encoding gradients are also corrected using the found affine transformation. In order to remove artifacts resulting from finite k-space discretization, the so-called Gibbs effect, and to reduce the number of possible uncontrolled erroneous diffusion data, also known as random outliers, a Gaussian smoothing function with a kernel size of 1.5x1.5x1.5 mm ³ was applied. . Then, automatic segmentation of the brain is carried out with the exclusion of areas of the scalp. Segmentation results are evaluated visually and, if necessary, corrected in the ITK-SNAP program (http://itksnap.org). All maps of diffusion metrics are built using the Explore DTI program using the least squares method with weights. Next, the processed b0-images are loaded into the ITK-SNAP program (http://itksnap.org), where for each subject a 2x2 mm region of interest (Region-of-interest-based analysis, ROI) is manually allocated in the posterior middle corpus callosum . The resulting zone is saved as a binary mask and subsequently used to evaluate the value of the diffusion index on the map of axial diffusion (axial diffusivity, AD). Thus, when opening the axial diffusion map in ITK-SNAP (http://itksnap.org) with the obtained mask superimposed on it, in the "Volumes and Static Data" section, a numerical value for this area is obtained. Therefore, the diagnosis of cognitive disorders in patients with age-dependent CMA is carried out by mapping AD and determining its value in the posterior middle corpus callosum. The obtained AD values are compared with the established threshold value equal to 0.002, while the value of individual axial diffusion >0.002 corresponds to a high probability of cognitive disorders in the patient, and <0.002 to a low one.

We conducted a study to compare microstructural changes in DT-MRI in brain regions with neuropsychological profile data of patients with diagnostic MRI signs of age-dependent CMA.

On a group of 74 patients with age-dependent CMA (mean age 60.7±6.9 years) and 18 healthy volunteers (mean age 57.8±5.9), a binary logistic regression model of microstructural predictors of CR was obtained. We used the values of 29 regions of interest (ROI) (see Fig. 1) on diffusion tensor MRI metric maps (Siemens Verio, 3 Tesla, Erlangen, Germany) – mean diffusivity (MD), fractional anisotropy (FA), radial diffusion (RD) and axial diffusion (AD). In FIG. 1 shows selection of regions of interest (ROIs) in periventricular (circle), deep (square), and juxtacortical (triangle) NIV (a); anterior, anteromedian, posterior median, and posterior corpus callosum (b), anterior, middle, and posterior cingulate gyrus (c), hippocampus (d).

Criteria for inclusion in the study were: 1) age of patients 46-69 years; 2) justification for MRI - the presence of cognitive complaints (decrease in memory, attention, slow thinking, and others); 3) MRI changes corresponding to age-dependent CMA according to the STRIVE criteria, 2013 [Wardlaw J.M., Smith E.E., Biessels G.J., et al. Neuroimaging standards for research into small vessel disease and its contribution to aging and neurodegeneration. Lancet Neurol. 2013; 12(8): 822-838.]; 4) exclusion of CMA due to other independent causes: genetic, inflammatory, thrombophilic, systemic, toxic, history of severe migraine; 5) exclusion of probable Alzheimer's disease according to the NIA-AA criteria for moderate CR and dementia.

The study was approved by the local ethics committee of the Federal State Budget Scientific Institution of Science. All subjects signed an informed consent to the study and the processing of personal data.

The severity of CI was determined according to the Montreal Cognitive Function (CF) Scale (MoCA) [23] and independence in everyday life [24]: with a score of <26 and independence - moderate CI (MCR), with dependence - dementia, with a score of ≥ 26 - subjective CR.

Diffusion data were obtained using a spin echo planar echo sequence with 3 diffusion weighted values (b=0, 1000 and 2500 s/mm2 ) for 64 directions of encoding diffusion gradients; TE/TR 115/12600 ms; matrix 100 x 100, resolution 2 x 2 x 2 mm3. DT-MRI was used to generate maps of FA, MD, RD and AD followed by analysis of regions of interest (ROI analysis). ROI with a volume of 2 ^mm3 was isolated manually independently by two neuroradiologists (10 years of work experience) on b0-images, followed by projection onto 4 diffusion metric maps using the ITK-Snap program (http://itksnap.org). The reliability of the selected DT ROI values was assessed using intraclass coefficients, which for all cases was >0.7. The following analysis used the means of two ROI measurements.

The selected ROIs are shown in FIG. 1. They included the centers of the anterior, anterior medial, posterior median, and posterior sections of the corpus callosum; anterior, middle and posterior sections of the cingulate gyrus; hippocampal heads; NIBV of the hemispheres. To select the ROI in the NTI of the hemispheres, a formalized section was made on sagittal images through the bodies of the lateral ventricles above the subcortical structures, after which the ROI was identified on the axial images on the conditional axes of the anterior (anterofrontal region) and posterior (temporoparietal region) horns of the lateral ventricles, the body of the lateral ventricle (posterior frontal region), separately in the periventricular (3–13 mm from the wall of the lateral ventricles), juxtacortical (4 mm from the corticomedullary junction) and deep (between the two zones described) white matter. The control of ROI isolation was achieved by assessing the location of the marker in all three projections (axial, sagittal, frontal) using a 3D cursor; In the absence of NIBV hemispheres in the ROI projection, the ROI was distinguished vertically by 1–2 slices above and below, horizontally - with an indent to the sides up to 5 mm from the main axis. The resulting zones were stored in the form of binary masks and were subsequently used to estimate the values of diffusion metrics in each area.

All ROIs of the corpus callosum, cingulate gyrus, and NIBV of the left hemisphere were included in the predictive CR model.

Statistical analysis was performed using IBMSPSS 23.0 and R 3.4.3 software. Binary logistic regression was used to assess the predictive ability of individual indicators in the development of cognitive impairment. The adequacy of the fitted logistic model was additionally assessed by ROC-analysis (Receiver Operator Characteristic) according to the probabilities predicted by the model (binary outcome) with the determination of the area under the curve and optimal threshold values with their sensitivity and specificity.

It was shown that 3 indicators of AD have the greatest predictive ability in relation to CR - in the periventricular unchanged white matter (pNIW) of the posterior sections of the left frontal lobe, the middle section of the right cingulate gyrus and the posterior middle section of the corpus callosum. A further 5-year prospective follow-up of 22 patients showed that WMH foci formed in the areas of NIBV in some patients, and therefore this predictor lost its value, while AD in the posterior-middle corpus callosum showed significant changes in the 5-year interval as a criterion assessment of disease progression and CR.

At this stage, in order to search for diagnostic predictors (equivalents) of CR, which are easy to determine, optimal sensitivity and specificity, the group of studied patients with age-dependent CMA was expanded from 74 to 188 patients, healthy volunteers - from 18 to 53 people. The examination was carried out on two tomographs - Siemens Verio, 3 Tesla (Erlangen, Germany) (137 patients and 31 volunteers) and Siemens Prisma, 3 Tesla (Erlangen, Germany) (51 patients and 22 volunteers) and the same 29 ROI 4 were included in the analysis metrics (MD, FA, RD and AD) of diffusion tensor MRI. The results for each of the groups were evaluated separately.

In accordance with the new binary logistic regression model, axial diffusion in the posterior-middle corpus callosum (p=0.014), as well as pNIBP of the posterior parts of the left frontal lobe, which was not further analyzed due to the fact that in dynamics in some patients becomes WMH (see table 1).

Table 1. Predictors of CR severity (binary logistic regression, p<0.001).

Predictors B
(predictor coefficient) R OSH 95% CI, borders lower upper AD in pNIBV of the posterior sections of the left frontal lobe ( χ ₁ ) 2846.3 0.0001 9.09 0.957 86.487 AD in the posterior middle part of the corpus callosum ( χ ₂ ) 2097.6 0.014 13.7 1.012 186.503 AD in the middle sections of the right lumbar gyrus ( χ ₃ ) 1121.8 0.240 0.009 0.001 0.469 Constant -10.4 0.0001 Abbreviations: OR, odds ratios; CI - confidence interval; AD - Axial Diffusivity / axial diffusion; pNIBP, periventricular white matter unchanged.

In accordance with the ROC analysis, the threshold value for the predictive model of the development of age-dependent CMA based on axial diffusion in the posterior middle corpus callosum was 0.0024 (sensitivity - 77%, specificity - 77%) (see Table 2 and Fig. 2). ). That is, the choice of ROI in the posterior middle corpus callosum on the AD-maps of the patient and its comparison with the threshold value at a value of> 0.002 allows us to conclude that the probability of CR is high, and <0.002 is low.

In FIG. 2 shows the ROC curve of a predictive model of CR in patients with age-dependent CMA. At the same time, the area under the curve was 0.842, the confidence interval was 0.773-0.911 (p<0.001), which indicates a good predictive ability of the predictive model for the development of CMA.

Thus, the established predictor (equivalent) of CR can be used as a tool for diagnosing CR in age-dependent CMA based on individual AD in the posterior middle corpus callosum (Figure 3). In accordance with the available ideas obtained on model animals, the change in AD most likely corresponds to axonal damage. Our data indicate the obligatory nature of these processes in the development of CR in CMA. Probably, this indicator indicates the comorbidity of vascular and degenerative pathology. However, its use as a measure of treatment efficacy is currently limited by insufficient knowledge of its pathophysiological significance and histological equivalents.

In FIG. 3 shows the selection of a region of interest (ROI) in the corpus callosum for individual assessment of axial diffusion. Currently, there are no other methods of instrumental diagnosis of CR in patients with MRI signs of age-dependent CMA.

Examples of the implementation of the method.

Example 1

Patient K., a 68-year-old man, applied to the NCI with complaints of decreased concentration and memory and MRI signs of age-dependent cerebral microangiopathy (Fazekas stage 1). An additional examination was carried out - neuropsychological testing and MRI examination of the brain in the mode of diffusion-tensor MRI.

Neuropsychological testing showed no signs of cognitive impairment in the patient (according to the MoCA scale - 27 points, tests for executive functions of the brain (Stroop test, tracking test) - within the age limits).

DT-MRI was performed on a 3 Tesla device. Based on the obtained images, axial diffusion (AD) maps were constructed using the Explore DTI program. Next, b0-images were loaded into the ITK-SNAP program, where a 2x2 mm zone of interest in the posterior middle corpus callosum was manually selected and saved as a binary mask. This mask with the preserved area of interest was superimposed on the AD map, and in the "Volumes and static data" section of the ITK-SNAP program, the absolute values of the AD index for the posterior middle corpus callosum were obtained, which amounted to 0.0025 in this patient.

The obtained value is above the threshold of 0.002 and indicates a high probability of CR in the patient. This conclusion is confirmed by the results of neuropsychological testing of the patient in dynamics after 5 years, which showed the presence of moderate cognitive impairment in the patient (MoCA - 25 points), which indicates a greater sensitivity of the AD indicator in the posterior middle corpus callosum compared to classical neuropsychological testing for the diagnosis of CR in patients with age-dependent CMA.

Example 2

Patient L., a 64-year-old woman, applied to the Federal State Budget Scientific Institution of Science with complaints of forgetfulness and memory loss. MRI of the brain revealed signs of age-dependent cerebral microangiopathy (stage Fazekas 1). An additional examination was carried out - neuropsychological testing and MRI examination of the brain in the mode of diffusion-tensor MRI.

Neuropsychological testing showed no signs of cognitive disorders in the patient according to the MoCA scale - 26 points, self-care was not impaired. According to the results of testing for the control functions of the brain (Stroop test, tracking test), mild impairments were revealed.

DT-MRI was performed on a 3 Tesla device. Based on the obtained images, axial diffusion (AD) maps were constructed using the Explore DTI program. Next, b0-images were loaded into the ITK-SNAP program, where a 2x2 mm zone of interest in the posterior middle corpus callosum was manually selected and saved as a binary mask. This mask with the preserved area of interest was superimposed on the AD map, and in the "Volumes and static data" section of the ITK-SNAP program, the absolute values of the AD index for the posterior middle corpus callosum were obtained, which amounted to 0.0026 in this patient.

The obtained value is above the threshold of 0.002 and indicates a high probability of CR in the patient. This conclusion is confirmed by the results of neuropsychological testing of the patient over 5 years, which showed that the patient has moderate cognitive impairment (MoCA - 24 points), which indicates a greater sensitivity of AD in the posterior middle corpus callosum compared to classical neuropsychological testing for the diagnosis of CR in patients with age-dependent CMA.

Example 3

Patient Sh., a 65-year-old woman, applied to the Federal State Budget Scientific Institution of Science with complaints of memory loss, slowing of the pace of thinking. Brain MRI revealed foci of age-dependent cerebral microangiopathy (Fazekas stage 2). Conducted neuropsychological testing and MRI examination of the brain in the mode of diffusion-tensor MRI.

Neuropsychological testing revealed no objective cognitive impairments in the patient according to the results of the MoCA test and activity in everyday life (MoCA - 26 points, self-care is not impaired), according to the results of tests for executive functions of the brain (Stroop test, tracking test) - mild to moderate impairments were detected. degree.

DT-MRI was performed on a 3 Tesla device. Based on the obtained images, axial diffusion (AD) maps were constructed using the Explore DTI program. Next, b0-images were loaded into the ITK-SNAP program, where a 2x2 mm zone of interest in the posterior middle corpus callosum was manually selected and saved as a binary mask. This mask with the preserved area of interest was superimposed on the AD map, and in the "Volumes and static data" section of the ITK-SNAP program, the absolute values of the AD index for the posterior middle corpus callosum were obtained, which amounted to 0.0023 in this patient.

The obtained value is above the threshold of 0.002 and indicates a high probability of CR in the patient. This conclusion is confirmed by the results of neuropsychological testing of the patient in dynamics after 5 years, which showed the presence of moderate cognitive impairment (MoCA - 24 points), which indicates a greater sensitivity of the AD indicator in the posterior middle corpus callosum compared to classical neuropsychological testing for the diagnosis of CR in patients with age-dependent CMA.

Patient 4.

Patient L., a 58-year-old woman, applied to the Federal State Budget Scientific Institution of Science with complaints of forgetfulness and decreased concentration. MRI of the brain revealed signs of age-dependent cerebral microangiopathy (stage Fazekas 1). An additional examination was carried out - neuropsychological testing and MRI examination of the brain in the mode of diffusion-tensor MRI.

Neuropsychological testing showed no signs of cognitive disorders in the patient according to the MoCA scale - 28 points, self-care was not impaired. According to the results of testing for the control functions of the brain (Stroop test, tracking test) - the results correspond to the age norm.

DT-MRI was performed on a 3 Tesla machine. Based on the obtained images, axial diffusion (AD) maps were constructed using the Explore DTI program. Next, b0-images were loaded into the ITK-SNAP program, where a 2x2 mm zone of interest in the posterior middle corpus callosum was manually selected and saved as a binary mask. This mask with the preserved area of interest was superimposed on the AD map, and in the "Volumes and static data" section of the ITK-SNAP program, the absolute values of the AD index for the posterior middle corpus callosum were obtained, which amounted to 0.0019 in this patient.

The obtained value is below the threshold of 0.002 and indicates a low probability of CR in the patient. This conclusion is confirmed by the results of neuropsychological testing of the patient in dynamics after 5 years, which showed the absence of cognitive disorders (MoCA - 29 points).

Example 5

Patient A, a 50-year-old woman, applied to the FSBSI NCN with complaints of forgetfulness. MRI of the brain revealed signs of age-dependent cerebral microangiopathy (stage Fazekas 1). An additional examination was carried out - neuropsychological testing and MRI examination of the brain in the mode of diffusion-tensor MRI.

Neuropsychological testing showed no signs of cognitive disorders in the patient according to the MoCA scale - 30 points, self-care was not impaired. According to the results of testing for the control functions of the brain (Stroop test, tracking test) - the results correspond to the age norm.

DT-MRI was performed on a 3 Tesla device. Based on the obtained images, axial diffusion (AD) maps were constructed using the Explore DTI program. Next, b0-images were loaded into the ITK-SNAP program, where a 2x2 mm zone of interest in the posterior middle corpus callosum was manually selected and saved as a binary mask. This mask with the preserved area of interest was superimposed on the AD map, and in the "Volumes and static data" section of the ITK-SNAP program, the absolute values of the AD index for the posterior middle corpus callosum were obtained, which amounted to 0.00183 in this patient.

The obtained value is below the threshold of 0.002 and indicates a low probability of CR in the patient. This conclusion is confirmed by the results of neuropsychological testing of the patient in dynamics after 5 years, which showed the absence of cognitive disorders (MoCA - 30 points).

Claims (1)

A method for predicting cognitive disorders in patients with age-dependent cerebral microangiopathy, including an MRI examination of the brain, characterized in that the MRI examination of the brain is carried out in the diffusion-tensor MRI mode, followed by mapping of axial diffusion and determining its value in the posterior middle region corpus callosum, then the values of axial diffusion indicators are compared with the established threshold value of 0.002, and with an individual axial diffusion value of more than 0.002, a high probability is predicted, and at less than 0.002, a low probability of cognitive impairment in patients with age-dependent cerebral microangiopathy.