Findings on Magnetic Resonance Imaging in Neuroborreliosis—A Nationwide Cohort Study

ABSTRACT

Objective

To explore pathological findings on magnetic resonance imaging (MRI) and their diagnostic implications in early‐stage neuroborreliosis (NB).

Method

Adult patients from the Danish neuroinfections cohort (DASGIB, 2015–2019) with confirmed NB, symptom duration < 6 months, and MRI performed within 14 days of diagnosis were included. MRIs were retrospectively reinterpreted by an unblinded neuroradiologist.

Results

In 116 patients, 123 MRIs were performed (99 brain, 44 spine, 46 with contrast). White matter lesions (WML) were common, but non‐specific and associated with increasing age (p < 0.001). Six patients showed WML not typical for small vessel disease. Acute infarctions occurred in four patients. Encephalitis was clinically diagnosed in five patients; one showed brainstem FLAIR hyperintensities. In 45 contrast‐enhanced brain scans, leptomeningeal enhancement was identified in 6 (13%) and cranial nerve enhancement in 29 (64%). There was poor correlation between facial palsy and enhancement. Spinal cord lesions (1.6–14 cm) were identified in 10 of 44 scans (23%) without symptoms of transverse myelitis. Among 13 contrast‐enhanced spine scans, 8 showed leptomeningeal enhancement (61%), and 8 showed nerve root enhancement (61%). Most enhancements did not match symptoms.

Conclusion

Pathological findings were found in 35 of 46 patients with contrast‐enhanced MRIs. Key findings included cranial nerve, spinal nerve root, and leptomeningeal enhancement—often without clinical correlation. Spinal cord lesions were relatively frequent; cerebral infarction was rare. While key findings can support the diagnosis, their absence does not exclude NB.

Key MRI findings in neuroborreliosis include leptomeningeal, cranial nerve, and nerve root enhancement, often without corresponding symptoms. Spinal cord lesions were relatively frequent and white matter lesions were non‐specific and associated with increasing age. Acute infarctions were rare but important differential diagnoses, particularly in younger stroke patients without an apparent cause. The absence of MRI abnormalities does not exclude neuroborreliosis, and imaging should be interpreted in clinical context.

1. Introduction

Neuroborreliosis (NB) is one of the most prevalent bacterial neuroinfections in Europe, with an estimated incidence of 7.3 per 100,000 person‐years [1, 2, 3, 4]. Symptoms typically emerge within two months following a tick bite infected with Borrelia burgdorferi sensu lato ( B. burgdorferi ) [5, 6, 7]. In adults, radicular pain is the most frequent manifestation, and nearly half develop cranial nerve (CN) palsy most often affecting the facial nerve [1, 5, 7, 8]. Additional symptoms include headache, sensory disturbances, and fatigue [9]. Despite a typically recognizable clinical presentation, the average diagnostic delay from the onset of neurological symptoms to diagnosis remains approximately 2–3 weeks and has shown little improvement over the last 30 years [1, 10, 11]. Increased healthcare‐seeking behavior and diagnostic procedures including imaging have been observed 12 weeks preceding NB diagnosis [12].

The role of magnetic resonance imaging (MRI) in NB diagnosis remains unsettled [13, 14, 15]. Reported findings range from normal scans to leptomeningeal enhancement, CN enhancement, spinal nerve root enhancement, and myelitis [8, 13, 15, 16, 17]. Recently, fluid‐attenuated inversion recovery (FLAIR) imaging has shown specific patterns in NB‐related encephalitis [16]. The significance of white matter lesions (WML) remains debated, as many may reflect age‐related findings [8, 18, 19].

Our earlier study of 368 well‐defined NB patients revealed that about 25% had findings suggestive of NB on the initial contrast‐enhanced MRI reports [15]. The scans were presumably conducted to rule out alternative diagnoses like stroke, malignancy, or disc herniation. The present study expands on this by reinterpreting MRI scans retrospectively with knowledge of the NB diagnosis, aiming to refine understanding of key MRI findings in NB patients.

2. Methods

2.1. Study Design

A nationwide observational retrospective cohort study.

2.2. Study Population and Setting

As of December 2023, Denmark had a total population of 5.9 million people. Healthcare in Denmark is universally accessible, tax‐financed and provided free of charge to all residents. The study population was identified through the Danish Study Group of Infections of the Brain (DASGIB) database from January 1, 2015, to November 1, 2019. DASGIB is a nationwide collaboration involving eight Danish Departments of Infectious Diseases, and it includes patients aged 18 years and older (≥ 15 years from 2015 to 2017) diagnosed with central nervous system (CNS) infections [20]. Two studies have previously been published on NB patients from the DASGIB cohort [10, 15]. Patients with NB were diagnosed based on European Federation of Neurological Societies (EFNS) guidelines [21]. To ensure focus on early‐stage NB, patients with symptoms exceeding 6 months duration were excluded. Patients who underwent MRI within 14 days of diagnosis were included, regardless of protocol.

2.3. Imaging Data and Protocols

MRI examinations were performed across multiple institutions using diverse scanner models and local protocols. Despite this variability, a core level of standardization was achieved.

All non‐contrast brain MRI protocols included diffusion‐weighted imaging with apparent diffusion coefficient (DWI/ADC) mapping, FLAIR, and T1‐ and T2‐weighted sequences. Contrast‐enhanced brain MRIs consistently included a post‐contrast T1‐weighted sequence, most acquired as a 3D magnetization‐prepared rapid gradient echo [MPRAGE], although some sites used alternative 3D sequences (e.g., SPACE, VIBE) or 2D spin‐echo sequences. Fat suppression techniques were often applied but varied in type and implementation across scanners.

For spinal MRI, non‐contrast imaging typically included sagittal T1, sagittal T2, and axial T2 sequences, while contrast‐enhanced scans included both sagittal and axial post‐contrast T1‐weighted sequences.

2.4. Clinical Data

Study data were systematically collected and managed using the secure, web‐based software tool Research Electronic Data Capture (REDCap), hosted at Region Nordjylland [22, 23]. As previously described [15], all infections were prospectively registered by the principal investigator using a standardized web‐based case report form, including medical records, laboratory test results, and radiological findings.

2.5. Outcomes

The primary outcomes of this study were the reinterpretation and identification of findings suggestive of NB. One experienced neuroradiologist, fully informed of the diagnosis, conducted the reinterpretation of all MRIs. The interpretation focused on identifying: (1) abnormalities in brain parenchymal with and without gadolinium (Gd) enhancement; (2) spinal cord abnormalities with and without Gd enhancement; (3) leptomeningeal enhancement; (4) CN and spinal nerve enhancement; (5) acute infarctions; and (6) other significant radiological findings.

MRI findings suggestive of NB were defined as leptomeningeal enhancement in brain and spine, CN enhancement, spinal nerve enhancement, and myelitis.

CN enhancement was considered abnormal when signal intensity was visibly and consistently higher than expected—either asymmetrical (compared to the contralateral nerve) or increased relative to adjacent muscle or parenchyma (in bilateral cases). For the facial nerve, enhancement of the cisternal, intracanalicular, or labyrinthine segments, or marked asymmetry in normally enhancing segments (geniculate to mastoid) was considered pathological.

Ventriculomegaly was assessed visually while normal pressure hydrocephalus (NPH) was evaluated using the established imaging criteria: an Evans Index > 0.3, presence of Disproportionately Enlarged Subarachnoid‐space Hydrocephalus (DESH), and a callosal angle ≤ 90° [24].

WMLs were counted and categorized into three groups: < 10, 10–20, and > 20 lesions. This semi‐quantitative approach, based on guidelines from the Swedish Neuroradiological Society [25], was chosen for its reproducibility and consistency compared to subjective severity ratings.

2.6. Statistical Analysis

Categorical variables were reported as counts and percentages and compared using the chi‐square test or Fisher's exact test, as appropriate. Continuous variables were summarized as medians with interquartile range (IQR) and compared using non‐parametric test including the Wilcoxon rank sum test and Kruskal–Wallis test if more than two groups are compared. SAS statistical software Enterprise guide version 8.1 (SAS Institute Inc., Cary, NC, USA) was used for data analysis, and p‐values below 0.05 (two‐sided) were considered statistically significant.

2.7. Ethics

The study and the DASGIB cohort were approved by the Danish Data Protection Agency (record number 2012‐58‐0018) and the Danish Patient Safety Authorization (journal number 3‐3013‐2579/1) permitting the use and registration of data without obtaining informed consent. Since no biomedical interventions were involved, ethical approval from the National Committee on Health Research Ethics was not required under Danish law.

3. Results

The DASGIB cohort identified 368 NB patients; 150 underwent MRI. After excluding 34 patients (25 scanned > 14 days after diagnosis, 8 with symptoms > 6 months, and 1 child), 116 patients with 123 scans were included: 99 brain MRIs, 44 spinal MRIs, including whole neuroaxis MRI in 20 patients, and brain with partial spine in 7. Contrast was administered in 46 patients: 45 brain, 13 spine, including 9 whole neuroaxis MRI, and 3 brain with partial spine. All patients had neurological symptoms and CSF pleocytosis; 91/116 (78%) had intrathecal B. burgdorferi antibody production. Baseline characteristics are shown in Table 1 (median age: 59 years [IQR 47–71], 60 males). Median symptom duration was 14 days (IQR 5–35), and most MRIs were conducted within 3 weeks of symptom onset. Brain and spine findings are detailed in Tables 2 and 3.

TABLE 1.

Baseline characteristics of 116 adult patients with neuroborreliosis diagnosed between 2015 and 2019 in Denmark.

Characteristics	Patients with NB (n = 116)
Age, years, median (IQR)	59 (47–71)
Sex, female, n (%)	56 (48)
No physical/cognitive deficits prior to admission, n (%)	90 (85)
Missing	10
Functioning level prior to admission, n (%)
Full‐time working	49 (49)
Retired but independent	36 (35)
Retired and dependent on some help	1 (1)
Parttime work	4 (4)
Disabled, on a welfare income, or in a nursing home	11 (11)
Missing	15
Days from the debut of symptoms to diagnosis, median (IQR)	14 (5–35)
Days from the debut of symptoms to MRI, median (IQR)	17 (7–43)
Hours from lumbar puncture to MRI, median (IQR)	2 (−27–46)
Symptoms
Peripheral neurological symptoms, n (%)
Paresis in the extremities	29 (28)
Sensory deficiencies or radicular pain	53 (51)
Missing	12
Gait disturbances, n (%)	17 (15)
Cranial nerve palsy, n (%)
None	59 (53)
Facial nerve	42 (37)
Trigeminal nerve	8 (7)
Abducens	4 (4)
Other cranial nerve a	16 (14)
Missing	4
Diagnosis b , n (%)
Definite NB	91 (78)
Possible NB	25 (22)

Note: Categorical variables are presented as n (%) and continuous variables as medians with interquartile rates (IQRs).

Abbreviations: B. burgdorferi , B. burgdorferi sensu lato complex; CSF, Cerebrospinal fluid; Ig, Immunoglobulin; IQR, Interquartile range; NB, neuroborreliosis..

^a Hearing loss, taste loss, reduced or affected/double vision, tinnitus, bulbar symptoms.

^b EFNS GUIDELINES: Three criteria should be fulfilled for definite NB and two of them for possible NB: (i) neurological symptoms; (ii) cerebrospinal fluid (CSF) pleocytosis; (iii) Bb‐specific antibodies produced intrathecally.

TABLE 2.

Findings on brain MRI in 99 patients with neuroborreliosis ^a .

	n = 99
White matter lesions b
Counts, n (%)
0	17 (17)
1–10	44 (45)
10–20	10 (10)
> 20	28 (28)
Primary location, n (%)
Subcortical	43 (53)
Deep white	1 (1)
Periventricular c	38 (46)
Character, n (%)
Punctate	40 (49)
Fluffy	29 (36)
Confluent	12 (15)
White matter lesions not typical for small vessel disease, n (%)
Inflammatory lesions resembling multiple sclerosis	3 (3)
Located in the cerebellar peduncle, medulla oblongata, or mesencephalon	3 (3)
Vascular changes
Acute infarct, n (%)	4
Cortical infarct	1
Lacunar infarct	3
Chronic stage infarct, n (%)	11 (11)
Other findings
Possible posterior reversible encephalopathy syndrome	1 (1)
Findings of normal pressure hydrocephalus d	1 (1)
Early signs of normal pressure hydrocephalus e	3 (3)
Ventriculomegaly	1 (1)
Contrast enhancement in 45 brain MRI with contrast
	n = 45
Findings suggestive of neuroborreliosis, n (%)	33 (73)
Leptomeningeal enhancement, n (%)	6 (13)
Pachymeningeal enhancement g	3 (7)
Lepto‐ and pachymeningeal enhancement g , n (%)	1 (2)
Cranial nerve enhancement, n (%)	29 (64)
Cranial nerve III, Oculomotor nerve, n (%)
Unilateral enhancement	3 (7)
Bilateral enhancement	12 (27)
Cranial nerve V, Trigeminal nerve, n (%)
Unilateral enhancement	1 (2)
Bilateral enhancement	14 (31)
Cranial nerve VI, Abducens nerve, n (%)
Unilateral	0 (0)
Bilateral enhancement	2 (4)
Cranial nerve VII, Facial nerve, n (%)
Unilateral enhancement	3 (7)
Bilateral enhancement	5 (11)
Cranial nerve VII and VIII, Facial and vestibulocochlear complex, n (%)
Unilateral enhancement	1 (2)
Bilateral enhancement	14 (31)
Cranial nerve IX, Glossopharyngeal nerve n (%)
Unilateral enhancement	0 (0)
Bilateral enhancement	1 (2)
Cranial nerve XII, Hypoglossal nerve, n (%)
Unilateral enhancement	0 (0)
Bilateral enhancement	1 (1)
Enhancement of more than 1 cranial nerve, n (%)	18 (40)

^a Including brain MRI from 20 patients having a combined spine and brain MRI.

^b White matter lesions of chronic small vessel disease character.

^c Not like multiple sclerosis.

^d Marked dilation of fissure sylvii, acute callosal angle, and crowding at the vertex.

^e Dilation of Sylvian fissure, callosal angle > 90°, and crowding at the vertex within normal range.

^f MRI findings suggestive of neuroborreliosis were defined as leptomeningeal enhancement in brain and spine, cranial nerve enhancement, spinal nerve enhancement, and myelitis.

^g The time from lumbar puncture to MRI was 3–6 days in three of the patients with pachymeningeal enhancement, one patient had the MRI performed 2 months prior to lumbar puncture.

TABLE 3.

MRI findings in the spinal canal in 44 patients with neuroborreliosis ^a .

Non contrast	n = 44
Protocol, n (%)
Full spine	29 (66)
Cervical spine	10 (23)
Lumbar spine	5 (11)
Spinal cord lesions
0 lesions (%)	34 (76)
1 lesion (%)	9 (22)
2 lesions (%)	1 (2)
Size, cm, mean (min‐max)	7.5 (1.6–14.0)
Swollen cord, n (%)	0 (0)
Affected tissue, n (%)
Peripheral	9 (82)
Central	2 (18)
Segment, n (%)
Cervical	6 (55)
Thoracic	1 (9)
Lumbar/conus	3 (27)
Entire spinal cord	1 (9)
Appearance, n (%)
Long segment lesion (≥ 1.5 vertebral body height)	8 (73)
Short segment lesion (< 1.5 vertebral body height)	2 (18)
Diffuse	1 (9)
Spinal lesions not attributed to other pathology, n (%)	8 (80)
Spinal lesions possibly caused by degenerative disc disease, n (%)	2 (20)
Contrast enhancement in 13 spine MRI with contrast
Protocol, n (%)
Full spine	10 (77)
Cervical spine	3 (23)
Findings suggestive of neuroborreliosis b , n (%)	10 (85)
Intramedullary Gadolinium enhancing lesion, n (%)	0 (0)
Leptomeningeal enhancement, n (%)	8 (62)
Enhancement of cervical nerve roots, n (%)	2 (23)
Enhancement of cauda equina, n (%)	6 (46)

^a Including MRI spine from 20 patients with a combined brain and spine MRI.

^b MRI findings suggestive of neuroborreliosis were defined as leptomeningeal enhancement in brain and spine, cranial nerve enhancement, spinal nerve enhancement, and myelitis.

3.1. White Matter Lesions

A total of 82/99 (83%) brain MRIs were with WMLs. WML count correlated significantly with age (Kruskal–Wallis test, p < 0.001), increasing across WMLs groups from a median of 41 years in those without WMLs to 72 years in those with > 20 lesions. No association was found between WML count and symptoms duration (Kruskal–Wallis test, p = 0.5).

WMLs not typical for small vessel disease (SVD) were identified in six patients, two of whom also had acute infarctions. Three lesions displayed inflammatory patterns resembling MS, with confluent periventricular involvement of the corpus callosum, pons, and cerebellar peduncles. The remaining three patients had lesions in less typical regions, including the mesencephalon, medulla, and cerebellar peduncles. Two of these also showed CN enhancement and additional FLAIR hyperintensities in the brainstem, cerebellar folia, posterior lentiform nucleus, and perivascular or sulcal spaces.

3.2. Acute Infarctions

Four patients (median age 65 years, 2 males) presented with acute cerebral infarctions on MRI.

Patient 1 showed left frontoparietal watershed infarcts, small foci of leptomeningeal enhancement, CN enhancement, sulcal FLAIR hyperintensity in a few left frontal sulci and along perivascular spaces, and WMLs not typical for SVD in the cerebellar peduncle. Symptoms included 4 months of fatigue and gait issues followed by acute onset of unilateral weakness and dysarthria. Fatigue persisted at 3‐month follow‐up.

Patient 2 had a left occipital cortical infarct (Figure 1a), frontal watershed infarcts, and inflammatory WMLs. Symptoms lasted 1 week, with progressing dyscoordination and sensory deficits; mild headache remained at follow‐up.

FIGURE 1.

Acute infarctions in the left occipital lobe (a) and the medulla oblongata (b) in patients with neuroborreliosis. (a) MRI brain of a male in his fifties with neuroborreliosis. Axial Diffusion‐weighted imaging demonstrated an acute infarction in the left occipital lobe. (b) MRI brain of a male in his seventies with neuroborreliosis. Axial Diffusion‐weighted imaging demonstrated an acute infarction in the midline of the medulla oblongata.

Patient 3 had a paramedian medullary infarct (Figure 1b) and chronic stage infarct in the basal ganglia. Balance disturbances had progressed over 7 weeks. Follow‐up was not available.

Patient 4 had a cerebellar infarct, extensive non‐specific WMLs including in the pons, leptomeningeal enhancement around the brainstem, diffuse FLAIR hyperintensities, and early signs of NPH. Symptoms progressed over 3 months; gait disturbances resolved post‐treatment, but new‐onset hearing loss persisted.

3.3. Encephalitis

Clinical encephalitis, defined as definite NB with altered mental status lasting > 24 h and at least three minor criteria [26], was demonstrated in five patients. MRI findings varied: Patient 1 demonstrated extensive FLAIR hyperintensities in the brainstem, corticospinal tracts, posterior lentiform nucleus, and cerebellar folia, along with early signs of NPH, subtle CN enhancement (CN5 and CN7/8 complex), and cervical leptomeningeal enhancement. Patient 2 had FLAIR hyperintensities lining the brainstem surface and the cranial portion of the central canal but not in the brainstem parenchyma, non‐specific WMLs in the pons, and early signs of NPH. Patient 3 had isolated CN7 enhancement. Patient 4 had non‐specific WMLs in the pons, and Patient 5 exhibited both WMLs and CN enhancements (CN3, CN5 and CN7/8 complex).

3.4. Brain, Meningeal and Cranial Nerve Enhancement

Leptomeningeal enhancement was observed in 6/45 (13%) brain MRIs, primarily around the brainstem (4 cases) and cerebellar folia (2 cases).

CN enhancement was present in 29/45 (64%) scans, mostly affecting CN3 (15), CN5 (15), CN7 (8), or the CN7/8 complex (15), often bilateral, and with multiple nerve involvement in 18/29 (62%) (Figure 2a–c). No consistent association with corresponding clinical symptoms was identified. Facial nerve enhancement occurred in 22 patients; 10 had no facial palsy. Among 20 patients with both facial palsy and contrast MRI, 8 (40%) showed no enhancement (Figure 3a). Trigeminal nerve enhancement was seen in 15 patients, 14 of whom had no trigeminal neuropathy; conversely, 2 patients with trigeminal neuropathy had no visible enhancement (Figure 3b).

FIGURE 2.

Bilateral enhancement of the oculomotor nerves (A), trigeminal nerves (B), and vestibulocochlear nerves (C) in a young female (< 18 years) with neuroborreliosis. (a) Axial T1‐weighted post‐contrast imaging using the Magnetization‐Prepared Rapid Gradient Echo (MPRAGE) sequence with gadolinium‐based contrast agent demonstrating avid enhancement of bilateral cisternal segments of the oculomotor nerves (White arrows). (b) Axial T1‐weighted post‐contrast imaging using the MPRAGE sequence with gadolinium‐based contrast agent demonstrating avid enhancement of bilateral cisternal segments of the trigeminal nerves (White arrows). (c) Axial T1‐weighted post‐contrast imaging using the MPRAGE sequence with gadolinium‐based contrast agent demonstrating avid enhancement of intracanalicular segments of bilateral 7th and 8th cranial nerve complexes (White arrows).

FIGURE 3.

Number of patients with palsy or enhancement of the facial nerve (A) and the trigeminal nerve (B). (a) Facial palsy and enhancement of the facial (CN7) and/or the facial/vestibulocochlear nerve complex (CN7/8) on contrast enhanced MRI (n = 45). A total of 22 patients exhibited enhancement of facial nerve or facial‐vestibulocochlear nerve complex (blue, black, purple, and green); 10 without facial palsy, 9 with unilateral facial palsy, and 3 with bilateral facial palsy. Among the 23 patients without enhancement of facial nerve or facial‐vestibulocochlear nerve complex (gray), 7 had a unilateral facial palsy, and 1 had a bilateral facial palsy. (b) Trigeminal palsy and enhancement of the trigeminal nerve (CN5) on contrast enhanced MRI (n = 45). A total of 15 patients exhibited enhancement of the trigeminal nerve (black and green); 14 without trigeminal palsy and 1 with unilateral trigeminal palsy. Among the 30 patients without enhancement of the trigeminal nerve (gray), 2 had unilateral trigeminal palsy.

3.5. Spinal Cord Lesions

Among 44 spine MRIs, 10 patients had 11 spinal cord lesions ranging between 1.6–14 cm, with 8/11 (73%) classified as long segment (≥ 1.5 vertebral body height). Two lesions were likely related to degenerative disc disease. No spinal cord swelling was observed. Leptomeningeal enhancement along the spinal cord or enhancement of nerve roots at the lesion level appeared in three cases and four cases also had CN enhancement. Another patient exhibited inflammatory WMLs. Seven patients had symptoms correlating with lesion location—typically radiating pain, sensory changes, or motor deficits—none met criteria for acute transverse myelitis. Three patients had unrelated symptoms.

3.6. Spinal Enhancement

Of the 13 contrast‐enhanced spine MRIs, 11 (85%) showed enhancement: 8 demonstrated leptomeningeal enhancement along the spinal cord, 6 showed cauda equina enhancement, and 2 had cervical nerve root enhancement (Figure 4). One brain MRI additionally revealed leptomeningeal and cervical nerve root enhancement in the partially scanned upper spine. Despite frequent radicular symptoms, anatomical correlation was poor—only 2 of 7 patients with cervical involvement reported upper limb symptoms, and 2 of 6 with cauda equina enhancement had leg or back pain. No cases showed intramedullary enhancement.

FIGURE 4.

Enhancement in spinal MRI of a neuroborreliosis patient. Sagittal T1 post‐contrast (gadolinium) of a young female (< 20 years) with neuroborreliosis showing leptomeningeal enhancement along the medulla and spinal cord, as well as enhancement of the cauda equina (white arrows).

3.7. Other Findings

Five patients had ventriculomegaly; one patient demonstrated radiological signs of NPH, with marked dilation of sylvian fissure and crowding at the vertex. Three other patients exhibited early radiological signs of NPH, but within the normal range. The remaining patient showed visual ventriculomegaly without other features. None were clinically diagnosed with NPH.

3.8. The Reinterpretation Compared With the Initial Findings

The reinterpretation identified findings suggestive of NB in 33 of 45 (73%) contrast‐enhanced brain MRIs and 10 of 13 (77%) spine MRIs, compared to only 11 of 45 (24%) and 3 of 13 (23%), respectively, in the original reports. This represents a significant increase in detection rates (Fisher's exact test: brain MRI, p < 0.001; spine MRI, p = 0.02). New findings included WMLs not typical for SVD, spinal cord lesions, and FLAIR hyperintensities not previously reported.

4. Discussion

Findings suggestive of NB increased significantly upon reevaluation; findings were observed in 35 of 46 patients with contrast‐enhanced MRIs (brain and spine scans). Key findings included spinal nerve root, CN and leptomeningeal enhancement. No consistent association was found between the anatomical location of contrast enhancement and corresponding clinical symptoms. Additionally, spinal cord lesions were identified in 10 of 44 patients. Non‐specific WMLs were common and associated with increasing age. Acute cerebral infarcts were observed in four patients. In five patients with clinical encephalitis, one showed brainstem FLAIR hyperintensities. These findings underscore the diagnostic value of MRI in patients with complex neurological presentations and highlight the importance of including NB in the differential diagnosis.

WML are common incidental findings, present in more than half of the elderly population, often progressing with time [27, 28]. While earlier studies reported a high prevalence of WML in NB patients [29, 30, 31], later research suggests that these lesions are primarily age‐related [13, 18, 19]. A Norwegian study found no significant differences in WML burden between NB patients and controls [32]. In our cohort, WMLs appeared in 82% of MRIs, mostly non‐specific and more frequent in older patients, supporting the likelihood that most lesions are unrelated to NB. However, the absence of a control group, standardized grading, and follow‐up imaging limits interpretation.

Six patients had WMLs not typical for SVD, often accompanied by findings of cerebral infarcts, spinal cord lesions, and CN enhancement. Inflammatory lesions resembling those seen in MS have been described, and the presence of symptoms such as radicular pain or CN involvement—which are uncommon in MS—warrant further assessment for NB [30, 33, 34]. The imaging patterns suggest a possible mixed meningeal and parenchymal inflammatory process, and the lesions may represent part of a broader inflammatory response rather than isolated vascular pathology.

Cerebrovascular events in NB are rare, accounting for < 1% of cases and < 0.1% of all strokes [35, 36, 37]. The pathophysiology may involve CSF inflammation with perivascular lymphocytic infiltration and vasoconstriction or thrombosis from vessel wall inflammation [38, 39]. Typically, small penetrating arteries are affected, leading to subcortical or lacunar infarcts (Figure 1a,b). NB‐associated strokes often occur in younger patients after a prodromal phase with one‐third exhibiting radiculitis or cranial neuritis [40, 41, 42]. However, NB may be unrecognized in older stroke patients, and mild cases may be missed due to symptom resolution after antibiotics [40, 41]. In our cohort, all four stroke patients had prodromal neurological symptoms, and two had MRI findings suggestive of NB. These findings underline the need to consider NB in patients with unexplained strokes, especially those with neurological prodromes or living in endemic regions. Unrecognized NB may lead to repeated infarctions or long‐term sequelae [41, 43].

Recent studies have increasingly focused on encephalitis and potential radiological features [16, 26]. Pfefferkorn et al. described the tarsier sign, characterized by symmetrical FLAIR hyperintensities in the midbrain and pons [16]. Although we did not specifically assess for the tarsier sign, among the five patients who fulfilled the clinical criteria for encephalitis, only one showed FLAIR hyperintensities in the brainstem which extended beyond the brainstem, and another showed hyperintensities along the brainstem surface but not in the parenchyma. Additionally, pontine FLAIR hyperintensities were observed in a patient without encephalitis. These findings suggest that while brainstem FLAIR hyperintensities may be associated with NB‐related encephalitis, they are variable and not consistently present [37].

Facial palsy affects about one‐third of NB patients but represents only 5% of all facial palsy cases [1, 7, 44]. CN enhancement is a recognized MRI finding in NB, though reported frequencies vary [8, 18, 44]. In a German study, contrast enhancement on the day of admission was observed in only 5 of the 29 patients with facial palsy caused by NB (17%). Lindland et al., however, reported contrast enhancement in 57% in NB patients scanned within 1 month—comparable to our 64% [8, 44]. By contrast, Lindland et al. found CN7 enhancement in 95% (35/37) of facial palsy cases, mostly in the distal internal auditory canal [8]. In our cohort, only 12 of 20 patients (60%) with facial palsy and contrast‐enhanced brain MRI showed CN7 or CN7/CN8 complex enhancement [8]. This discrepancy may reflect differences in imaging timing and methodology. Our MRIs were performed within 14 days of diagnosis, whereas Lindland allowed imaging up to 1 month, possibly detecting later‐stage inflammation. Lindland also used a standardized imaging protocol, while our multicenter study involved variations in scanners, contrast timing, and sequences. Both studies observed CN enhancement in patients without facial palsy [8]. While facial palsy often prompts lumbar puncture and NB testing, these findings suggest that CN enhancement can occur without clinical facial palsy, and when present alongside symptoms such as radicular pain, meningism, or sensory deficits, should prompt CSF analysis.

Spinal cord lesions were seen in 10 of 44 spine MRIs, with one patient presenting two separate lesions—representing a higher frequency than previously reported [37, 45]. Prior studies have mainly described transverse myelitis and corresponding spinal lesions; however, the lesions found in this study lacked clinical features of transverse myelitis. Instead, patients predominantly experienced radiating pains, which were not necessarily confined to the affected segment. Although fat‐suppression techniques are useful for detecting subtle pathological changes, they may introduce artifacts complicating interpretation. In this study, T2 sequences were predominantly used to assess high signal in the medulla, and fat suppression was not applied. Importantly, six patients had additional MRI findings which, similar to the WMLs not typical for SVD, suggest that the spinal cord lesions may reflect a broader inflammatory process in NB.

NPH secondary to NB is rare, with less than ten published cases [46, 47, 48]. The presumed mechanism involves inflammation impairing CSF resorption. Most cases respond to antibiotic therapy without requiring shunt placement [46]. Radiological signs of NPH have also been described in NB‐related encephalitis [26] consistent with our findings where two encephalitis patients had early signs of NPH. However, ventricular enlargement can also occur incidentally. A Japanese study reported incidental NPH‐like imaging in approximately 1% of asymptomatic elderly individuals [49], while the Austrian ASPS‐Fam cohort reported 8 cases (7.1%) in 113 community‐dwelling adults aged 50–75 without neuropsychiatric disease [50]. The frequency in our cohort does not exceed these population rates. As we lacked a control group and none of our patients were clinically diagnosed with NPH, a causal link between NB and NPH cannot be established. Further studies with appropriate controls are needed to explore this potential association.

Although we included only patients with early NB (symptom duration < 6 months), several exhibited CNS involvements like cerebral infarctions and clinical encephalitis. While such features are traditionally considered late manifestations, clinical staging does not always correlate with symptom duration, and CNS involvement may occur early in the disease course [21, 37, 51]. This is supported by our findings, where CNS pathology was observed even in patients with relatively short symptom duration.

4.1. Limitations and Clinical Implications

This study reflects real‐world diagnostic practices across Denmark and includes a large, well‐defined cohort of patients with NB. Since inclusion in the DASGIB cohort is based on confirmed diagnosis rather than presenting symptoms—and healthy controls are not part of the cohort—the inclusion of a true control group was not feasible.

Several limitations should be noted. First, imaging protocols varied across sites, including differences in magnet strength, sequence types, contrast use, and the extent of spinal imaging. Contrast‐enhanced MRI was not uniformly performed, and spinal imaging was sometimes limited. This heterogeneity may affect the consistency and generalizability of specific imaging findings. Second, MRI was likely more often performed in patients with more severe or atypical symptoms, introducing potential selection bias influenced by local imaging practices and referral pathways. Third, all imaging was re‐evaluated by a single neuroradiologist, ensuring consistency but limiting reproducibility. Reliability was not assessed. The unblinded re‐evaluation may also have increased sensitivity to subtle findings. Fourth, CN enhancement was not consistently assessed for laterality, as left/right side information was not systematically recorded in the database and could not be retrospectively retrieved with sufficient accuracy. Consequently, analyses related to facial palsy and enhancement were not stratified by side.

Despite these limitations, the diversity of imaging practices provides valuable clinical insight into how MRI is used in routine evaluation of complex neurological symptoms and highlights findings relevant to NB diagnosis in everyday practice. MRI should be considered a valuable tool in the diagnostic workup of neurological symptoms; while the absence of enhancement does not exclude NB, the presence of findings should prompt further investigation.

5. Conclusion

Key findings on contrast‐enhanced MRI were seen in 35 of 46 NB patients, including CN and spinal nerve root enhancement and leptomeningeal enhancement—often without corresponding clinical symptoms. Additionally, spinal cord lesions were identified in 10 of 44 patients. Cerebral infarctions were rare. While these findings can support the diagnosis, their absence does not exclude NB.

Author Contributions

All authors contributed to the acquisition of data. N.S. reevaluated all MRIs. M.Ø. and A.‐M.L. designed the study and drafted the manuscript. The statistics were performed by M.Ø. All authors revised, commented on, and approved the final manuscript.

Funding

M.Ø. has received a research grant from Rigshospitalets Forskningspuljer. A.‐M.L. received a research grant from Lundbeck Foundation (grant number: R366‐2021‐127) and from Rigshospitalets forskningspuljer.

Conflicts of Interest

Outside of the present work: A.‐M.L. reports speakers' honorarium/travel grants/advisory board activities from Gilead and GSK, and honorarium/advisory board activity from Pfizer. The other authors declare that they have no competing interests.

Ørbæk M., Shekhrajka N., Gynthersen R. M. M., et al., “Findings on Magnetic Resonance Imaging in Neuroborreliosis—A Nationwide Cohort Study,” European Journal of Neurology 33, no. 2 (2026): e70500, 10.1111/ene.70500.

Data Availability Statement

The descriptive data that support the findings of this study are available from the corresponding author upon reasonable request. Raw data including imaging is protected by regulations on distributions of personal data and is not available for external researchers.