Correlation between the imaging manifestations of cerebellar infarction and the clinical characteristics of patients: a retrospective analysis

analysis of the correlation between the risk factors, etiology, distribution of blood vessels involved, infarct lesion size and scope, and clinical manifestations of cerebellar infarction.


Page 3/22
Cerebellar infarction is mainly caused by thrombosis or thromboembolism that leads to the blockage of the corresponding blood supply arteries of the cerebellum, which leads to a reduction or interruption of cerebellar blood ow and causes a series of symptoms and signs. Common risk factors for cerebellar infarction are hypertension, heart disease, diabetes, and atrial brillation [1]. Once the cerebellar infarction is formed, it causes a series of cell biochemical process disturbances caused by cerebral ischemia, leading to cytotoxic edema with increased intracellular water and opening or destruction of the bloodbrain barrier. Cerebral edema is the pathological basis of local brain swelling signs and imaging lowdensity signs caused by cerebral infarction in the super-acute phase [2,3]. Early diagnosis of cerebellar infarction is critical to the prognosis of patients. A physical examination is usually performed on the patient to get a preliminary understanding of the disease. However, its diagnosis often relies on head imaging examinations. In addition, examinations such as electrocardiogram, vascular ultrasound, blood lipids and coagulation function can further help clarify the cause of cerebellar infarction [4].
Computed tomography (CT) scan of the head is one of the most commonly used examination methods [5]. In the early stage of cerebral infarction (within three hours of onset), CT examination of the head can reveal some slight changes such as middle cerebral artery high density sign, unclear boundary between cortical edge and lenticular nucleus, and disappearance of brain sulci [6]. Magnetic resonance imaging (MRI) can nd small lesions that cannot be detected by CT. It is also of great signi cance when reviewing the effect of treatment. One hour after the onset of cerebral infarction, MRI showed swelling of the local gyrus and narrowing of the brain sulcus, followed by abnormal long T1 and long T2 signals. In addition, MRI diffusion and perfusion imaging can detect early cerebral infarction more timely [4,7,8]. Based on the imaging examination data and epidemiological statistics of patients with cerebellar infarction, this study conducted an in-depth analysis of the correlation between the risk factors, etiology, distribution of blood vessels involved, infarct lesion size and scope, and clinical manifestations of cerebellar infarction.

Ethics statement
The study was approved by the Research Ethics Committee of Tongji Hospital, Tongji University, This was an observational study and did not involve any form of active intervention in the diagnosis and treatment process of patients, the rights of patients were fully guaranteed. All patients were informed of the detailed study process and signed a written informed consent form.
All research procedures and methods were strictly implemented in accordance with relevant guidelines and regulations.

Patients
From January 2017 to June 2020, a total of 83 patients with acute cerebellar infarction (Tongji Hospital, Tongji University, Shanghai, China) were included in this analysis. The clinical data of the patients was extracted, including gender, age, personal history, past medical history, clinical symptoms, laboratory tests, National institute of Health Stroke Scale (NIHSS), Trial of org10172 in Acute Stroke Treatment (TOAST), and imaging manifestations. The demographic baseline data of the included patients are detailed in Table 1.

Diagnosis of cerebellar infarction
The diagnosis of cerebellar infarction was based on the criteria of the American Heart Association/American Stroke Association Deletes Sections from 2018 Stroke Guidelines [9] and was considered with clinical data and auxiliary examination. Clinical history and symptoms included: (1) suffered from diseases prone to thrombosis, such as heart disease, atrial brillation, cardiomyopathy, myocardial infarction, etc.; (2) sudden onset, reaching the peak of disease in a short time; (3) had signs or symptoms of the nervous system; and (4) severe patients experienced coma, gastrointestinal bleeding, brain herniation, and even died soon. CT and MRI examinations: (1) showed changes in ischemic infarction or hemorrhagic infarction; (2) CT and MRI of the brain showed infarcts at the embolization site, the boundary was not clear, and there was a certain space-occupying effect. Auxiliary examinations: (1) increased brain pressure and corresponding changes in cerebrospinal uid cytology; (2) the electrocardiogram showed abnormalities such as myocardial infarction or arrhythmia; and (3) carotid artery ultrasound examination evaluated the degree of lumen stenosis and atherosclerotic plaque and helped con rm carotid artery embolism.

NIHSS score
According to the National Institute of Health stroke scale (NIHSS) [10], the patients were scored for clinical prognosis assessment of patients. The evaluation includes consciousness, language, motor function, sensory loss, visual eld defect, eye movement, coordinated movement, neglect and articulation.
The scoring area is 0-42 points. In this study, the highest score was 15 points. We divided the patients according to NIHSS score into three groups (< 1, > 1 ≤ 5, and > 5 points).

TOAST classi cation
Based on the clinical data, the patients were classi ed by the standard of Trial of org10172 in Acute Stroke Treatment (TOAST) [11]. The concepts used in this classi cation are mainly derived from the Harvard stroke registration classi cation and the National Institute of Neurological Diseases and Stroke Database of the United States [11][12][13]. According to the criteria, the patients included in this study were mainly divided into the following types: small-artery occlusion (SAO), large-artery atherosclerosis (LAA) and cardioembolism (CE).

CT and MRI examination
All patients underwent head CT scan. The cross-sectional scan took the canthus ear line as the baseline, and scanned 8 to 10 slices upward in sequence, with a thickness of 5 mm, and a coronal scanning was also performed. All patients underwent cranial magnetic resonance (MRI) examination, including horizontal FSE, T1 FLAIR, T2 FLAIR and DWI sequence scans and sagittal FSE sequence. In order to further clarify the cause and scope of the disease, all patients received carotid/vertebral artery computed tomography angiography (CTA)/magnetic resonance angiography (MRA) and whole cerebral angiography [1,4,7,9].

Identi cation of the involved artery
With reference to the methods reported in the previous literature [14], the location of the infarct was con rmed by imaging examination. The blood supply arteries of the cerebellum were divided into three branches. Posterior inferior cerebellar artery (PICA): the blood supply area for the inferior vermis and the lower back of the cerebellar hemisphere. Anterior inferior cerebellar artery (AICA): the blood supply area for the midfoot, the cerebellar pompons, and the anterior lower part of the cerebellum. Superior cerebellar artery (SCA): the blood supply area for the upper half of the cerebellar hemisphere, the oral vermis and the dentate nucleus.

Blood laboratory tests
In the early morning on an empty stomach, 5 mL of venous blood was collected from the patient's elbow vein. The blood was kept in a centrifuge tube without anticoagulant. After the blood had clotted at room temperature, it was centrifuged for 8 minutes (at 3,000 rpm). The supernatant in the centrifuge tube was carefully aspirated, placed in a new EP tube, and stored in a refrigerator at -20°C for testing. The following biochemical indicators were detected on the automatic biochemical analyzer, which included D-dimer, blood sugar, high density lipoprotein (HDL), low-density lipoprotein (LDL), triglycerides and cholesterol.

Statistical processing of data
The data collected from this study included counting data and measurement data. The counting data was analyzed by the fourfold table and the R×C table, including chi-square test for comparing two rates or two constituent ratios, chi-square test for comparing multiple rates or multiple constituent ratios, correlation analysis for classi cation data and continuous adjusted chi-square test. The measurement data was analyzed by T test and analysis of variance (ANOVN). The T test was used for comparing of means between two groups and the analysis of variance was used for comparing the means of more than two groups. For the analysis of variance, each sample was required to be from a normal population and the overall variance of each group was homogeneous. The SPSS 22.0 software was hired to complete the data analysis of this study, a P value less than 0.05 was considered statistically signi cant.

Demographic characteristics and underlying diseases of patients
Eighty-three patients with cerebellar infarction were included in this study (

Correlations between the changes of D-dimer and blood lipids and TOAST classi cation in patients with cerebellar infarction
The concentrations of serum D-dimer in patients classi ed as CE (1.12 ± 0.56 mg/L) were higher than that in patients classi ed as LAA (0.58 ± 0.25 mg/L) and SAO (0.56 ± 0.22 mg/L) (P = 0.032) (Fig. 3E). The patients classi ed as LAA (3.58 ± 1.33 mmol/L) and SAO (3.76 ± 1.48 mmol/L) had higher serum concentrations of LDL than those classi ed as CE (2.85 ± 1.22 mmol/L) (P = 0.042) (Fig. 3F). However, the concentrations of blood glucose, HDL, triglycerides, and cholesterol did not show statistical signi cance among the three different types of patients (P > 0.05). The statistical data are shown in Table 6.

Correlations between the changes of D-dimer and blood lipids and the infarct sizes of cerebellar infarction (CT and MRI)
From the perspective of CT measurement, the serum concentrations of HDL in patients with the infarct diameter of < 20mm (1.27 ± 0.26 mmol/L) were higher than that in patients with ≥ 20mm (1.30 ± 0.28 mmol/L) (P = 0.019) (Fig. 4A). In terms of MRI, the serum concentrations of HDL in patients with the infarct diameter of < 20mm (1.05 ± 0.25 mmol/L) were higher than that in patients with ≥ 20mm (1.02 ± 0.36 mmol/L) (P = 0.013) (Fig. 4B). However, the concentrations of D-dimer, blood glucose, LDL, triglycerides, and cholesterol did not show statistical signi cance on the different infarct size in patients with cerebellar infarction (P > 0.05). The statistical data are shown in detail in Table 7.

Correlations between the NIHSS scores and TOAST typing and the diameter of infarct sizes in patients with cerebellar infarction
From the perspective of CT measurement, the infarct sizes of cerebellar infarction in patients with NIHSS scores of > 1 (31.86 ± 16.08 mm) were larger than that in those with NIHSS scores of ≤ 1 (18.98 ± 10.81 mm) (P = 0.028) (Fig. 4C). The patients classi ed as LAA (33.26 ± 13.68 mm) had larger infarct sizes of cerebellar infarction than those classi ed as SAO (9.31 ± 4.23 mm) and CE (7.16 ± 3.88 mm) (P < 0.0001) (Fig. 4D). From the perspective of MRI measurement, the infarct sizes of cerebellar infarction in patients with NIHSS scores of > 1 (33.98 ± 16.63 mm) were larger than that in those with NIHSS scores of ≤ 1 (21.87 ± 15.56 mm) (P = 0.035) (Fig. 4E). The infarct sizes in patients classi ed as LAA (40.05 ± 17.23 mm) were larger than that in patients classi ed as SAO (13.69 ± 5.68 mm) and CE (15.96 ± 8.65 mm) (P < 0.0001) (Fig. 4F). The statistical data are shown in Table 8.

Discussion
Cerebellar ischemia and hypoxia cause ischemic necrosis of local tissues, called cerebellar infarction. It is mainly caused by atherosclerosis, thrombosis or embolism in the posterior circulation (also known as the vertebrobasilar system) that supplies blood to the brain [15]. High-risk factors mainly include high blood pressure, diabetes, dyslipidemia, smoking, lack of exercise, abdominal obesity, heart disease, and poor eating habits [2,11,12,14,15]. Because of the extensive anastomosis between the cerebellar blood supply arteries, the incidence of cerebellar infarction is low, accounting for about 1.5-3% of ischemic stroke [16]. However, larger cerebellar infarcts are likely to cause hydrocephalus and lower brainstem compression, with severe clinical symptoms and higher mortality. So improving the understanding of the clinical characteristics of cerebellar infarction is essential for early diagnosis, treatment, prevention and assessment of prognosis.
Our research showed that the incidence of cerebellar infarction in men and women was close to 2:1 and there were more patients older than 65 years old. However, smoking and drinking did not seem to be associated with cerebellar infarction. We found that the included patients with cerebellar infarction had multiple underlying diseases, including diabetes, coronary heart disease, hyperlipidemia, carotid atherosclerosis, atrial brillation and hypertension. It is showed that hyperlipidemia, hypertension and carotid atherosclerosis were the main underlying causes. Especially, there was the most obvious correlation between carotid atherosclerosis and cerebellar infarction. It is generally believed that the risk factors of cerebellar infarction are consistent with the risk factors of cerebrovascular diseases, and the latter signi cantly increases the incidence of cerebellar infarction [11,12,17]. In terms of clinical manifestations, the incidence of ataxia and brainstem syndrome was relatively high, while the incidence of vestibular syndrome was relatively low. Common clinical manifestations of cerebellar infarction include dizziness/vertigo, nystagmus, limb and gait ataxia, cerebellar dysarthria, etc. Previous studies have found that infarcts in the blood supply area of PICA and AICA are more likely to cause dizziness in patients than infarcts in the SCA area. This is because the structure of the vestibular cerebellar nodules and occules are mainly supplied by PICA and AICA [18]. The prominent clinical manifestations of SCA are limb and gait ataxia and dysarthria [15]. Due to the limited sample size in this study, it is di cult to clarify the differences in clinical manifestations caused by infarctions in different vascular regions, which needs to be further explored in future. The NIHSS scores of the patients included in this study were mostly less than one point, and the TOAST classi cation indicated that the SAO and LAA were mostly.
Studies have shown that most patients with cerebellar infarction have lower NIHSS scores, and the TOAST etiological classi cation of cerebellar infarction has a certain relationship with the involved arteries [1,2,7,15,18].
We analyzed the vascular imaging results of patients with cerebellar infarction and found that the incidence of cerebellar infarction caused by SCA involvement was the highest regardless of whether it was performed by CTA or MRA. In addition to SCA, the probability of PICA involvement is also signi cantly greater than that of AICA. The involvement of different cerebellar blood vessels will produce different clinical symptoms, which is helpful for the diagnosis and prognosis of the disease. Obstruction of the PICA can lead to a headache and AICA territory infarction often leads to dysmetria, unilateral hearing loss and ipsilateral facial paralysis, and obstruction of the SCA tends to produce more ataxia, dysarthria and nystagmus [15]. By calculating the detection rate, we found that there was no statistical difference in the detection rate of different blood vessels between CTA and MRA. This indicates that if MRA cannot be performed for the patient, enhanced CT angiography and perfusion imaging will provide a good reference for the display of cerebellar infarcted blood vessels [7,15,16]. In this study, both CT and MRI examinations showed that the incidence of unilateral cerebellar infarction accounted for more than 80%; and the left hemisphere cerebellar infarction was the main one. This was a sure nding that MRI had an absolute advantage in detecting cerebellar infarction compared with CT. Brain CT scan is the most commonly used examination for cerebral infarction, but it is not sensitive to ultra-early ischemic lesions and small infarcts, especially brain stem. However, MRI can detect cerebral infarction early and has a high sensitivity [15,16]. One hour after the onset, local gyrus swelling, narrowing of the brain sulcus and abnormal long T1 and long T2 signals can be seen. In particular, MRI is sensitive to lacunar infarcts in the thalamus, cerebellum, and brainstem [7].
Our study found that patients with high NIHSS scores had higher serum concentrations of D-dimer, LDL, and triglycerides than those with low scores, while HDL levels were just the opposite. The increase of Ddimer suggests that it is related to thrombotic diseases caused, and it illustrates the enhancement of brinolytic activity. Existing studies generally support that the increase of D-dimer is a risk factor for cerebral infarction, and it has predictive signi cance for the course of the disease [19][20][21]. Studies have shown that cholesterol carried by LDL is prone to accumulate on the walls of blood vessels and arteries, forming unstable soft plaques, which blocks the cerebral artery and causes a cerebral infarction [22][23][24].
The main harm of triglycerides to the human body is mainly to cause atherosclerosis, block blood vessels and form thrombus. It is closely related to the occurrence of cerebral infarction [23,24]. Our research suggests that the D-dimer, LDL, HDL, and triglycerides can be used as indicators to predict the severity of the NIHSS scores in patients with cerebellar infarction [25]. The TOAST subtype classi cation standard focuses on the etiological classi cation, which has good credibility in clinical application [26]. We found that the serum D-dimer concentrations in patients classi ed as CE was higher than that in patients classi ed as LAA and SAO. The cardiogenic embolus is detached and embolized in the corresponding cerebral artery with the blood ow, causing ischemic cerebral infarction in the corresponding blood supply area. Among them, atrial brillation-related infarctions account for more than 79% of all cardiogenic infarctions [11,26]. Previous studies have also shown that D-dimer is intrinsically related to cardiogenic cerebral infarction [27,28]. Our research suggests that patients with cerebellar infarction with elevated D dimer concentrations should focus on the investigation of the possibility of cardiogenic embolism. Our study showed that patients classi ed as LAA and SAO had higher serum concentrations of LDL than those classi ed as CE, suggesting that LDL plays a prominent role in vascular embolism caused by atherosclerosis. LDL receptors are distributed on the cell membrane surface of various tissues throughout the body, such as arterial wall cells, and abnormally increased levels can cause atherosclerosis [29].
In the study, all patients with cerebellar infarction underwent cranial CT and MRI. We found that both CT and MRI showed that the serum HDL concentrations of patients with a cerebellar infarct area of less than 20mm in diameter was signi cantly higher than that of patients with an infarct area greater than or equal to 20mm in diameter. It is known that the high concentrations of HDL has a negative correlation with the occurrence of cardio-cerebral vascular embolism and infarction [30,31]. Our study suggests that HDL is a protective factor against cerebellar infarction and its high levels can reduce the occurrence of cerebellar infarction, which can also be understood that patients with higher HDL indicate a lower degree of cerebellar infarction and a better prognosis. Studies showed that HDL can bind to lipids attached to the walls of blood vessels to dissolve them. The dissolved products are transformed into excretable metabolites through the liver's biochemical reaction, thereby reducing the occurrence of vascular sclerosis and embolism [32,33]. Regardless of the result of CT or MRI, we found that the diameter of the cerebellar infarct area of patients with NIHSS scores more than one point was signi cantly larger than that of patients with NIHSS scores less or equal to 1 point. The tool of NIHSS scores is mainly used to evaluate the neurological status of stroke patients, and has important signi cance in individualized treatment, prognosis prediction and correct health education. The higher the scores are, the more severe the nerves damage [10,25]. Our research suggests that the NIHSS scores are positively correlated with the sizes of the cerebellar infarction area, which can indirectly re ect the severity of the infarction. In addition, we found that the cerebellar infarct area of patients classi ed as LAA by TOAST classi cation was signi cantly larger than that of patients classi ed as SAO and CE, and the measurement results from CT and MRI showed the same trend. The LAA patients often are found with signi cant stenosis in the carotid artery, anterior cerebral artery, middle cerebral artery, posterior cerebral artery, and vertebrobasilar artery during retrospective examination. Cerebral infarction caused by LAA often manifests as large area, rapid onset and high mortality rate [34,35].

Conclusions
The ndings of this study are summarized as follows. Hyperlipidemia, hypertension and carotid atherosclerosis are the main causes of cerebellar infarction. Ataxia and brainstem syndrome are the main clinical manifestations. SAO and LAA patients account for the majority. Cerebellar infarction caused by obstruction of SCA accounts for the majority, but there is no statistical difference in the detection ability of CTA and MRA. Unilateral infarction is the most and the incidence of left hemisphere cerebellar infarction is the highest. However, compared with CT, MRI has an absolute advantage in showing the area of cerebellar infarction. The serum concentrations of D-dimer, LDL, and triglycerides are higher in patients with high NIHSS scores than in those with low scores, while the serum HDL levels are just the opposite.