Bilateral middle cerebellar peduncle lesions: Neuroimaging features and differential diagnoses

Abstract Objectives Lesions limited to the bilateral middle cerebellar peduncles (MCPs) are uncommon. This retrospective study investigated diseases with a proclivity for the bilateral MCPs and explored the associations between their neuroimaging features and clinical findings for the differential diagnosis of such lesions. Methods We enrolled 26 patients who were admitted to our department between January 2016 and March 2019 with bilateral MCP abnormalities on magnetic resonance imaging (MRI). The demographic, clinical, and neuroimaging characteristics, and the biomarkers and diagnoses were evaluated. Results Although all patients exhibited symmetrical bilateral MCP hypointensities on T1‐weighted imaging and hyperintensities on T2‐weighted and fluid‐attenuated inversion recovery imaging, they were diagnosed with different conditions. Diagnoses included acute cerebral infarction (ACI) (n = 9, 34.62%), Wallerian degeneration (WD) (n = 8, 30.77%), multiple system atrophy (MSA) (n = 6, 23.08%), neuromyelitis optica (NMO) (n = 1, 3.85%), heroin‐induced leukoencephalopathy (n = 1, 3.85%), and primary central nervous system lymphoma (PCNSL) (n = 1, 3.85%). Patients with ACI exhibited bilateral MCP‐restricted diffusion hyperintensities on diffusion‐weighted imaging and corresponding stenosis or occlusion of the vertebrobasilar system. The initial MRI of patients with WD depicted pontine infarctions, while symmetrical MCP lesions were observed on follow‐up MRI. Symmetrical MCP lesions, cruciform hyperintensity, and marked atrophy in the posterior fossa were characteristic manifestations of MSA. Longitudinally extensive myelitis affecting more than three vertebral segments on cervical MRI and positive serum AQP4‐IgG may be indicative of NMO. Heroin‐induced leukoencephalopathy was characterized by extra‐symmetrical lesions in the posterior limbs of the internal capsules, while the anterior limbs were spared. PCNSL was indicated by a significant and characteristic “fist” sign on contrast‐enhanced MRI. Conclusions Bilateral MCP lesions were most frequently observed in cerebrovascular diseases, followed by neurodegenerative diseases, inflammatory diseases, toxic encephalopathies, and lymphomas. Our findings demonstrate that bilateral MCP signal abnormalities are more common in patients with ACI and WD, with fewer degenerative processes than previously believed. The high frequency of WD may be attributed to the specific awareness of this pathology. WD can also present with stage‐related restricted diffusion and should not be mistaken for a new infarction. The symmetrical bilateral MCP hypointensities on T1‐weighted imaging and hyperintensities on T2‐weighted imaging often raise concern regarding a demyelinating process. Our findings emphasize that neurologists should consider the aforementioned conditions and correlate the specific neuroimaging characteristics and medical history before arriving at the final diagnosis.


| INTRODUC TI ON
The middle cerebellar peduncle (MCP), also called the brachium pontis, is the largest afferent system of the cerebellum. It consists of pontocerebellar tract (PCT) fibers arising from the contralateral pontine nuclei (Perrini, Tiezzi, Castagna, & Vannozzi, 2013). The detection rate of posterior fossa lesions has improved with rapid advancements in neuroimaging, and abnormal signals involving the MCPs on magnetic resonance imaging (MRI) have been recently described (Hall, Fraint, & Dafer, 2018;Polat, 2019). However, symmetrical lesions limited to the MCPs are rare. Although some studies have reported bilateral hyperintensities of the MCPs on T2-weighted imaging in different diseases, their findings were not comprehensive (Okamoto, Tokiguchi, & Furusawa, 2003;Uchino, Sawada, Takase, & Kudo, 2004). Therefore, we retrospectively analyzed the data from 26 patients with bilateral MCP lesions on MRI and compared the clinical and neuroimaging features of different diseases, and correlated the imaging findings with clinical clues to aid differential diagnosis.

| Patient selection
This retrospective study enrolled 34 consecutive patients with bilateral MCP lesions treated in the Department of Neurology at the First Affiliated Hospital of China Medical University, a comprehensive academic hospital in a large urban area, between January 2016 and March 2019. We used the terms "bilateral MCP" and "bilateral brachium pontis" to retrieve data from the hospital's electronic information database system, which includes the information of a large number of patients who were hospitalized at our center. Patients were excluded from the study if they did not meet the respective diagnostic criteria (outlined below) or did not undergo radiological examination with MRI. Eight patients were excluded from the study based on these criteria. Twenty-six patients were included in the final review. This study was approved by the appropriate ethics review board.

| Clinical information and diagnostic criteria
The clinical information obtained included demographics (age of onset and sex), neurological signs and symptoms, disease-specific biomarkers (e.g., anti-aquaporin-4 [AQP4]-IgG), neuroimaging findings, and diagnoses. One patient underwent brain biopsy.
The diagnosis of each condition was based on its clinical diagnostic criteria, as follows. (a) Patients with acute cerebral infarction (ACI) met the Baltimore-Washington Cooperative Young Study Criteria, including neurological deficit lasting longer than 24 hr and MRI showing infarctions corresponding to their clinical findings, while those with transient ischemic attack or cerebral hemorrhage were excluded (Johnson, Kittner, & McCarter, 1995) with Wallerian degeneration (WD) were identified by the presence of initial pontine damage with subsequent abnormal bilateral MCP findings on MRI; moreover, they presented with clear fiber tract connections between the pontine and MCP lesions (De Simone, Regna-Gladin, Carriero, Farina, & Savoiardo, 2005;Shen et al., 2018). (c) Patients with multiple system atrophy (MSA) met the criteria for probable MSA detailed in the second consensus statement on the diagnosis of MSA: a sporadic, progressive, adult-onset disease characterized by autonomic failure and poorly levodopa-responsive parkinsonism or a cerebellar syndrome (Gilman, Wenning, & Low, 2008) (d) Patients with neuromyelitis optica (NMO) met the international consensus diagnostic criteria for NMO spectrum disorder (NMOSD) (Wingerchuk, Banwell, & Bennett, 2015): the core clinical characteristics, including optic neuritis and acute myelitis, a positive test for AQP4-IgG, and the exclusion of alternative diagnoses. (d) Heroininduced leukoencephalopathy was diagnosed by a positive heroin test, acute intoxication typically presenting as classic toxic leukoencephalopathy, supporting radiological abnormalities in the bilateral abnormalities are more common in patients with ACI and WD, with fewer degenerative processes than previously believed. The high frequency of WD may be attributed to the specific awareness of this pathology. WD can also present with stage-related restricted diffusion and should not be mistaken for a new infarction. The symmetrical bilateral MCP hypointensities on T1-weighted imaging and hyperintensities on T2weighted imaging often raise concern regarding a demyelinating process. Our findings emphasize that neurologists should consider the aforementioned conditions and correlate the specific neuroimaging characteristics and medical history before arriving at the final diagnosis.

K E Y W O R D S
magnetic resonance imaging, middle cerebellar peduncle, multiple system atrophy, neuromyelitis optica, stroke, Wallerian degeneration periventricular white matter or the cerebellum, and the exclusion of alternative diagnoses (Alambyan, Pace, & Miller, 2018), (e) Patients with primary central nervous system lymphoma (PCNSL) met the guidelines for the diagnosis and management of primary central nervous system diffuse large B-cell lymphoma: stereotactic biopsy and contrast-enhanced brain MRI (Fox, Phillips, & Smith, 2019).

| Magnetic resonance imaging findings
All patients had bilateral symmetrical MCP hypointensities on T1weighted imaging and hyperintensities on T2-weighted and FLAIR imaging. However, there were major differences in other neuroimaging findings for the different disorders (Table 1). Figure 1 describes a patient with isolated bilateral MCP infarctions and basilar artery (BA) occlusion. Eight of nine patients with ACI underwent angiography (five CTA, two MRA, and one digital subtraction angiography) of the involved vessels (Table 1). WD presented as an initial acute pontine infarction on initial MRI, with abnormal signals of the bilateral MCPs on follow-up MRI ( Figure 2). However, the characteristics of the initial infarctions and signals on DWI and ADC showed changes in the subsequent stages of WD in different patients (Table 1)

| D ISCUSS I ON
Isolated bilateral abnormalities of the MCPs on brain MRI are rare.
Previous studies have indicated that neurodegenerative diseases are most likely to affect both MCPs (Morales & Tomsick, 2015;Okamoto et al., 2003). However, bilateral MCP lesions were signs of other diseases in this study, the most common being ACI, followed by WD, MSA, NMO, heroin-induced leukoencephalopathy, and PCNSL.
Despite reports on all these pathologies mentioned previously, to the best of our knowledge, NMO and PCNSL lesions were not restricted to the bilateral MCPs but were often accompanied by other cerebral lesions in previous studies. We think that the differences in the frequency of disease distribution observed in our study, especially the increase in incidence of ACI and the decrease in that of degenerative processes, are related to the inclusion criteria or selection bias. Moreover, we attribute the high frequency of WD, which was not reported in the Okamoto review, to an increase in the specific awareness of this pathology (Okamoto et al., 2003).
Although these patients had similar abnormal signals in the bilateral MCPs on conventional MRI, resulting in similar clinical symptoms/ signs, there were major differences in their medical histories, other iconic imaging findings, and biomarkers. This study aimed to discuss the possible mechanisms underlying bilateral MCP lesions and explore the associations between neuroimaging features and clinical clues to facilitate differential diagnosis. Of them, one patient exhibited hypointensity (16 weeks after infarction) and three had slight hyperintensities (11, 16, and 17 weeks after F I G U R E 4 A 31-year-old woman with AQP4-IgG-seropositive neuromyelitis optica. Axial brain T1-weighted imaging (a) shows hypointensity in both MCPs, with hyperintensity on T2-weighted (b) and fluid-attenuated inversion recovery (c) imaging. Sagittal T1-weighted imaging (d) of the cervical spinal cord shows central linear hypointensity extending from C3 through C6, with high signal intensity on T2-weighted imaging (e)

F I G U R E 5
A 49-year-old woman with heroin-induced leukoencephalopathy. Axial brain T1-weighted imaging shows low signal intensity affecting the bilateral MCPs (a) and optic radiations, as well as the posterior limbs of the internal capsules, with sparing of the anterior limbs (f) and corresponding hyperintensity on T2-weighted (b,g), fluid-attenuated inversion recovery (c,h), and diffusion-weighted imaging (d,i). Enhanced scanning does not demonstrate significant enhancement in both the MCPs and internal capsules (e,j) infarction) on ADC, while information regarding the ADC sequence was unavailable for one patient (3 weeks after infarction). Castillo & Mukherji (1999) and Kang et al. (2000) have described several hyperintensities in CST during the early stage of WD (3-12 days). Shen et al. (2018) and Fitzek, Fitzek, & Stoeter (2004)  showed the complete "hot cross bun" sign within 5 years (Horimoto, Aiba, & Yasuda, 2002). Moreover, the sensitivity of cruciform hyperintensity for MSA-C is 50%, whereas that of the presence of MCP hyperintensity is 85% (Rizzo, Martinelli, & Manners, 2008;Schrag, Good, & Miszkiel, 2000). Consistent with these findings, our study suggests that a combination of these signs can contribute to the identification of MSA at different stages, and bilateral MCP abnormalities may be more sensitive to MSA identification compared to other infratentorial signs. As mentioned above, WD is a progressive process of degeneration and demyelination, which easily involves the fibers and MCPs (a massive bundle of fibers connecting the pons with the cerebellum). Moreover, neurodegenerative processes such as MSA (which also entails degeneration and demyelination) is well-known for its predilection for MCPs. Thus, we emphasized the importance of the predilection of the two pathologic processes (degeneration and demyelination) for MCPs. This information may help clinicians evaluate such cases and formulate a differential diagnosis.
Recently, numerous researchers have found that brain abnormalities can be observed in several patients with NMOSD with an increase in disease duration, and that the detection of cerebral white matter lesions by MRI, indicating multiple sclerosis (MS), no longer excludes a diagnosis of NMOSD (Cabrera-Gomez & Kister, 2012; Matthews & Palace, 2014). In 2015, the international consensus on the diagnostic criteria for NMOSD described relatively specific lesions in patients with NMOSD and brain abnormalities, including lesions in the dorsal medulla/area postrema, periependymal areas in the brainstem or diencephalic structures, and cerebral hemispheres, all of which are associated with a high AQP4 expression (Pittock et al., 2006;Wingerchuk et al., 2015). However, brain abnormalities were also observed in areas where the AQP4 expression was not particularly high, including extensive lesions spanning a considerable length of the corpus callosum and CST (Kim, Park, & Lee, 2010). Despite increasing knowledge of brain MRI characteristics in NMOSD, bilateral MCP lesions in patients with NMOSDrelated demyelination and inflammatory diseases are extremely F I G U R E 6 A 54-year-old woman with primary central nervous system lymphoma. Axial brain T2-weighted (b) and fluid-attenuated inversion recovery (c) imaging show ill-defined hyperintensity with a minimal associated mass effect involving the bilateral MCPs, with hypointensity on T1-weighted imaging (a). Contrast-enhanced T1-weighted MRI demonstrates abnormal patchy enhancement at the same location (d). Immunohistochemistry findings are CK − , CD3 + , CD20 + (f), Pax-5 + (h), Bcl2 + , CD21 − , CD23 − , CD10 − , CD5 + (e), CyclinD1 − , CD30 − , and Ki-67(80%+) (g), consistent with diffuse large B-cell lymphoma rare, which makes them difficult to distinguish from atypical MS lesions, especially in AQP4-IgG-seronegative patients (Okamoto et al., 2003;Uchino et al., 2004). In such cases, longitudinally extensive transverse myelitis lesions spanning three or more vertebral segments, which is very rare in adult MS, could be another important and specific neuroimaging characteristic of NMOSD (Kim, Paul, & Lana-Peixoto, 2015). Our patient experienced four consecutive attacks of bilateral optic neuritis and acute myelitis, was AQP4-IgGseropositive, and had a contiguous intramedullary lesion that met the international consensus diagnostic criteria for NMOSD (Wingerchuk et al., 2015). It is interesting that, unlike the circumventricular areas, the PCTs do not have high AQP4 expression; nevertheless, we believe the unique PCT involvement extending from the brainstem to the MCPs may be similar to longitudinal brain abnormalities commonly involving the CST from the internal capsule to the cerebral peduncle, providing new evidence for symmetrical MCP T2 hyperintensities in patients with AQP4-IgG-seropositive NMO.
Toxic leukoencephalopathy is an uncommon but critical neurological disorder seen in heroin abusers. There are various routes of heroin misuse, including inhalation, IV injection, intranasal (snorting), and subcutaneous injection (Alambyan et al., 2018). "Chasing the dragon," an inhalation method that involves heating heroin over aluminum foil and inhaling the resultant fumes, has recently become the most popular route of heroin intake due to its availability, greater ease of administration, and safer infectious profile compared to the other routes (Cicero, Ellis, Surratt, & Kurtz, 2014). A recent study showed that the brain MRIs of patients with heroin-induced leukoencephalopathy differed greatly according to the intake routes (Cheng, Chin, & Chang, 2019): The MRI findings of patients who inhaled heroin were characterized by posterior to anterior involvement of the cerebral white matter and lesions in the posterior limbs of the internal capsules, cerebellum, and brainstem, whereas those of patients who injected heroin IV were characterized by lesions in the subcortical U-fibers and genu of the internal capsule. The definitive mechanism underlying heroin-induced leukoencephalopathy remains unclear. One neuropathological study suggested that it was characterized by spongiform vacuolar degeneration of the white matter due to apoptosis of oligodendrocytes and subsequent demyelination. (Yin, Lu, Chen, Fan, & Lu, 2013). Oligodendrocyte apoptosis-induced demyelination in the white matter is more sensitive to ischemia/hypoxia because the axons and myelin sheaths in this area are thin (Yin et al., 2013). Therefore, in agreement with the pathological findings, brain MRI abnormalities caused by heroin inhalation often result in extensive, symmetrical lesions of the cerebral and cerebellar white matter, posterior limbs of the internal capsules, and splenium of the corpus callosum, which helps distinguish it from other causes of leukoencephalopathy such as toluene toxicity or reversible posterior leukoencephalopathy (Alambyan et al., 2018;Offiah & Hall, 2008). Our patient, who had an extensive medical history of heroin inhalation with presentations ranging from confusion to coma, exhibited these specific bilateral symmetrical lesions in the posterior limbs of the internal capsules, while the anterior limbs were spared, consistent with the aforementioned observations. Moreover, symmetrical lesions involving the bilateral MCPs might be another novel neuroimaging feature in those who "chase the dragon." Contrast-enhanced MRI is considered to be the best technique to detect and differentiate PCNSL, owing to the complexity of brain biopsy or cerebrospinal fluid analysis (Batchelor, 2019). A typical brain MRI in a patient with PCNSL shows a solitary, homogeneously enhancing mass or multiple lesions, frequently involving the frontal lobe, basal ganglia, corpus callosum, or periventricular white matter (Cheng & Zhang, 2019;Küker, Nägele, & Korfel, 2005). A significant and characteristic "incision," "angular," or "fist" sign, moderate edema, and the absence of necrosis is pathognomonic neuroimaging features (Patrick & Mohile, 2015). The presence of contrast enhancement in areas with abnormal signals in the MCP should evoke concern for a demyelinating process such as MS or progressive mul- In conclusion, symmetrical bilateral MCP lesions occur in various pathological conditions, and our findings indicated they were most frequent in cerebrovascular diseases, followed by neurodegenerative diseases, inflammatory diseases, toxic encephalopathies, and lymphoma. Patients with these diseases may have similar clinical symptoms/signs and conventional brain MRI presentations. Hence, clinicians should consider additional specific neuroimaging characteristics while formulating a differential diagnosis. The definitive mechanism underlying the pathogenesis of bilateral MCP lesions remains unclear; however, the degeneration of transverse pontocerebellar fibers, as a part of diffuse white matter lesions or the continuous spread of pontic lesions, has been considered as the most probable cause. Prospective research, including basic studies and large clinical trials, is needed to elucidate the mechanisms underlying MCP-related disorders and establish better guidelines for their differential diagnosis in clinical practice.

ACK N OWLED G M ENTS
The authors would like to thank Editage (www.edita ge.cn) for English language editing.

CO N FLI C T O F I NTE R E S T
The authors have no conflict of interest to declare.

AUTH O R CO NTR I B UTI O N
JJ: conceptualization and design of the study, interpretation of data, drafting and revising the manuscript. JW: critical revision of the manuscript for important intellectual content, ethics submission, and data analyses. ML: data collection and analyses. XW: ethics submission and data analyses. JZ: data collection and analyses. XS: critical revision of the manuscript for important intellectual content, study supervision, and fund support. All authors have read and approved the manuscript.

E TH I C A L S TATEM ENT
This study was approved by the appropriate ethics review board.

PE E R R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1002/brb3.1778.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data supporting the findings of this study are available from the corresponding author (Xiuli Shang) upon reasonable request.