MRI for diagnosing dementia – update 2025

Abstract

Background

Magnetic resonance imaging (MRI) plays a crucial role alongside clinical and neuropsychological assessments in diagnosing dementia. The recent and ongoing advancements in MRI technology have significantly enhanced the detection and characterization of the specific neurostructural changes seen in various neurodegenerative diseases, thereby significantly increasing the precision of diagnosis. Within this context of perpetual evolution, this review article explores the recent advances in MRI with regard to diagnosing dementia.

Methods

A retrospective literature review was conducted by searching the PubMed and ScienceDirect databases for the keywords “dementia”, “imaging”, and “MRI”. The inclusion criteria were scientific papers in English that revolved around the role of MRI as a diagnostic tool in the field of dementia. A specific time frame was not determined but the focus was on current articles, with an overall of 20 articles dating from the last 6 years (after 2018), corresponding to 55% of the total number of articles.

Results

This review provides a comprehensive overview of the latest advances in the radiologic diagnosis of dementia using MRI, with a particular focus on the last 6 years. Technical aspects of image acquisition for clinical and research purposes are discussed. MRI findings typical of dementia are described. The findings are divided into non-specific findings of dementia and characteristic findings for certain dementia subtypes. This provides information about possible causes of dementia. In addition, developed scoring systems that support MRI findings are presented, including the MTA score for Alzheimer’s disease with corresponding illustrative figures.

Conclusion

The symbiosis of clinical evaluation with high-field MRI methodologies enhances dementia diagnosis and offers a holistic and nuanced understanding of structural brain changes associated with dementia and its various subtypes. The latest advances, mainly involving the emergence of ultra-high-field (7T) MRI, despite having limited use in clinical practice, mark a pragmatic shift in the field of research.

Key Points

• High-field MRI (3T) and specialized sequences allow for the detection of early structural changes indicative of dementia.

• Characteristic neuroanatomical MRI patterns enable the differentiation between various subtypes of dementia.

• Established scales provide added value to the quantification and categorization of MRI findings in dementia.

Citation Format

• Akl E, Dyrba M, Görß D et al. MRI for diagnosing dementia – update 2024. Rofo 2025; DOI 10.1055/a-2563-0725

Zusammenfassung

Hintergrund

Die Magnetresonanztomografie (MRT) spielt neben klinischen und neuropsychologischen Untersuchungen eine entscheidende Rolle bei der Diagnose von Demenzerkrankungen. Die laufenden Fortschritte in der MRT-Technologie haben die Erkennung und Charakterisierung der spezifischen neurostrukturellen Veränderungen bei verschiedenen neurodegenerativen Erkrankungen erheblich verbessert, was die Genauigkeit der Demenzdiagnose deutlich erhöht. Dieser Übersichtsartikel untersucht die jüngsten Fortschritte der MRT bei der Diagnose von Demenzerkrankungen.

Methoden

Im Rahmen einer retrospektiven Literaturrecherche wurden die Datenbanken PubMed und ScienceDirect nach den Stichworten „dementia“, „imaging“ und „MRI“ durchsucht. Die Aufnahmekriterien waren wissenschaftliche Arbeiten in englischer Sprache, die sich mit MRT-Aspekten der Demenz befassen. Ein bestimmter Zeitrahmen wurde nicht festgelegt, der Schwerpunkt lag aber auf aktuelle Artikeln, wobei insgesamt 20 Artikel aus den letzten 6 Jahren stammen (nach 2018), was 55% der Gesamtzahl der Artikel entspricht.

Ergebnisse

Diese Übersicht gibt einen umfassenden Überblick über die neuesten Fortschritte bei der radiologischen Diagnose von Demenz mittels MRT, mit besonderem Fokus auf die letzten 6 Jahre. Es werden technische Aspekte der Bilderfassung für klinische und Forschungszwecke erörtert. Es werden MRT-Befunde beschrieben, welche typisch für Demenzerkrankungen sind. Die Befunde werden unterteilt in unspezifische Befunde einer Demenz und in charakteristische Befunde für bestimmte Demenz-Subtypen. Dies gibt Aufschluss über mögliche Ursachen der Demenz. Darüber hinaus werden entwickelte Scoring-Systeme vorgestellt, die die MRT-Befundung unterstützen, darunter der MTA-Score für Alzheimer-Krankheit mit entsprechenden, illustrierenden Abbildungen.

Schlussfolgerung

Die Symbiose zwischen der klinischen Bewertung mit Hochfeld-MRT-Nutzung erweitert die Demenzdiagnose und bietet ein ganzheitliches und nuanciertes Verständnis der strukturellen Veränderungen des Gehirns im Zusammenhang mit Demenz und ihren verschiedenen Subtypen. Die jüngsten Fortschritte, die vor allem das Aufkommen der Ultrahochfeld-MRT (7 T) enthalten, markieren trotz ihres begrenzten Einsatzes in der klinischen Praxis ein Paradigmenwechsel im Bereich der Forschung.

Kernaussagen

• Hochfeld-MRT (3T) und spezialisierte Sequenzen ermöglichen die Erkennung früher struktureller Veränderungen, die auf eine Demenz hindeuten.

• Charakteristische neuroanatomische MRT-Muster erlauben die Unterscheidung verschiedener Demenzsubtypen.

• Etablierte Skalen bieten einen zusätzlichen Mehrwert bei der Quantifizierung und Kategorisierung von MRT-Befunden bei Demenz.

Keywords

dementia - atrophy - neurodegenerative disorders

Introduction

Dementia is a syndrome characterized by an acquired disturbance of cortical functions, including memory, thinking, orientation, comprehension, calculation, learning capacity, language, and judgment; consciousness remains intact [1]. These disturbances lead to significant difficulties in daily living, and the increasing incidence of the disease is resulting in an unprecedented burden globally. As we are facing a major demographic shift, with a drastic increase in the elderly population, the prevalence of age-associated diseases, particularly dementia, is set to rise significantly [2].

There are numerous underlying causes of dementia. The meticulous diagnosis of dementia necessitates a synergistic approach, combining detailed clinical evaluation and paraclinical findings with neuroimaging to attain a nuanced understanding of the various underlying cognitive disorders. After a thorough history, physical examination, and neuropsychological tests to determine the deficits, it is crucial to exclude secondary causes of dementia associated with the cognitive deficits and to detect neurostructural changes associated with different forms of dementia. Neuroimaging plays a vital role in ruling out these potential secondary causes, helps to distinguish between the various types of dementia, helps with understanding progression, and aids in developing treatment strategies [3].

This review aims to collate and present the comprehensive body of scientific evidence regarding the use of magnetic resonance imaging (MRI) in diagnosing various forms of dementia in routine care and will focus on recent study results. It will also evaluate the nuanced capabilities of MRI in distinguishing between the most common causes of dementia, such as Alzheimer’s disease (AD) and vascular dementia (VD), among others.

Diagnostic Approach

Although alternative neuroimaging modalities are used as adjuncts in the workup of dementia, MRI is the most utilized modality for clarifying the causes of a dementia syndrome. The most recent German guidelines for dementia (S3-Leitlinie Demenzen, 11/2023) recommend the use of structural imaging with MRI due to its superior diagnostic capabilities and expert consensus based on current studies. While CT can be used to exclude secondary causes of dementia, MRI is particularly recommended to differentiate between various etiologies of dementia due to its detailed imaging of brain structures [4].

Structural MRI, in particular 3T MRI, is notable for its readily available comprehensive and noninvasive methodology while providing detailed images of brain structures and detecting changes like atrophy, which are crucial for identifying characteristic patterns for various types of dementia [2].

Step 1: Definition of the imaging protocol

The first step in the clinical evaluation of MRI is to choose the MRI sequences to be performed, each designed to highlight different aspects of brain structure and function. The protocol may vary depending on the available equipment, time resources, and specific patient needs, but a general model often includes the following sequences [5] [6] [7]:

• T1-weighted imaging (T1W): is good for its anatomical details and allows for quantitative analysis of the various brain atrophy patterns.

• T1-weighted Magnetization Prepared Rapid Gradient Echo (MPRAGE): provides high-resolution details and helps to detect focal atrophy patterns. It is typically acquired in the sagittal orientation and reformatted in the coronal orientation.

• T2-weighted imaging (T2W): helps in identifying areas of high water content, which can indicate edema or inflammation. In dementia, the presence of abnormal signal intensities in the white matter may suggest vascular contributions to cognitive impairment.

• Fluid-attenuated inversion recovery imaging (FLAIR) in a coronal or sagittal view: detects lesions near the cerebrospinal fluid, such as periventricular white matter changes. It is useful for detecting small vascular changes or demyelination, which can be associated with vascular dementia (VaD) or multiple sclerosis.

• Diffusion-weighted imaging (DWI): is sensitive to the movement of water molecules and can help detect acute stroke or chronic microinfarcts, which might mimic dementia symptoms.

• Susceptibility-weighted imaging (SWI): is sensitive to blood products and iron deposits. It is useful for identifying microbleeds, which are associated with cerebral amyloid angiopathy and VaD.

A proposed protocol, ideally using 3T-MRI, in the workup of dementia according to the most recent recommendations would include T1W MPRAGE, T2W, FLAIR, DWI, and SWI sequences [6]. The coronal plane of a T1W sequence is the method of choice to evaluate atrophy patterns as well as the morphological evaluation of the hippocampus, whereas the sagittal plane of a T1W sequence is best to evaluate the precuneus. Furthermore, to better evaluate the medial temporal lobe, a special reconstruction of a T2W sequence could be performed perpendicular to the long axis of the hippocampus to assess the medial temporal lobe [8].

Any other additional sequence or modality would be determined individually and in a case-based manner. For example, contrast-enhanced MRI could be considered in atypical cases to rule out neoplastic, inflammatory, or infectious processes.

In a prospective study that evaluated the clinical impact of MRI on the diagnosis of dementia subtypes using this acquisition protocol, an increase of 23.7% in the accuracy of diagnosis was reported [9].

Step 2: Excluding secondary causes of dementia

Once the case-specific protocol is defined, the next step is to exclude the presence of structural secondary causes of dementia. These conditions include, but are not limited to, brain tumors; normal pressure hydrocephalus (NPH), which involves an accumulation of cerebrospinal fluid in the brain’s ventricles; subdural hematomas; cerebral infections such as encephalitis or HIV-associated neurocognitive disorders; and demyelinating diseases like multiple sclerosis. All of these can contribute to cognitive impairment [9].

Step 3: Identifying common neuroanatomical changes on MRI

After excluding the secondary causes of dementia, a systematic approach to MRI interpretation is essential, encompassing a spectrum of neuroanatomical changes, from general to specific, aligning with various dementia etiologies. The most common MRI findings as well as corresponding examples of conditions causing dementia will be provided below.

1. Cerebral Atrophy

Brain atrophy initiates at approximately 30 years of age and slowly progresses with aging. There is a reduction in brain volume across the lifespan, with a 14% decrease in frontal lobe volume and a 13% decrease in hippocampal volume from ages 30 to 80. Additionally, significant shrinkage of white matter volume, particularly a 24% loss in the frontal lobe, is predominantly observed after 70 years of age. Considering the aforementioned information, when atrophy occurs at a rate faster than the normal rate of aging, there is a possibility of concomitant dementia. It should be noted that there is a large interindividual range and there is no linear correlation between age and brain volume loss [10]. Therefore, it is essential to compare the brain volume with age-matched healthy controls since brain volume decline slowly accelerates with advancing age [11].

Some factors also contribute to an acceleration in the rate of brain atrophy [4]. For example, alcohol consumption can cause significant structural and volumetric changes in the brain. Therefore, an appropriate medical history should be investigated, to guide clinicians through the diagnosis of dementia in conjunction with MRI evaluation. Alcoholism has been linked to significant gray matter volume loss, particularly in the cerebellum, which is the most affected region. Other affected areas include the prefrontal cortex, cingulate cortex, insula, and striatum, with the degree of loss correlating with the duration and intensity of alcohol consumption [12] [13].

Global Cerebral Atrophy: A fundamental aspect of dementia diagnosis using MRI is the evaluation of global cerebral atrophy. This encompasses a reduction in overall brain volume, discernible by widened cerebral sulci and ventricular enlargement. These changes are a common denominator in dementia with a very high sensitivity but low specificity [14].

Global cortical atrophy (GCA) scale:

A methodical tool designed to quantify atrophy across 13 distinct brain regions. The GCA score assesses brain atrophy by examining 13 specific areas, including the sulcal and ventricular dilations in the frontal, parieto-occipital, and temporal lobes on both sides of the brain and the third ventricle. Each area is scored from 0 to 3, indicating atrophy severity from “none” to “severe” [15] [16] ([Fig. 1]).

Fig. 1 Axial and coronal reformations of MPRAGE sequences. a, b Normal, age-appropriate brain volume for a 53-year-old woman, corresponding to GCA 0. c, d accentuated sulci (blue arrow) without relevant widening of the ventricles in a 65-year-old man, observed in GCA 1. e, f 84-year-old man with volume loss of gyri (white arrow) and moderate ventricular enlargement, as seen in GCA 2. g, h severe and global thinning of the gyri of the cerebral cortex (white arrow) with marked ventricular enlargement in an 81-year-old man with GCA 3.

Specific Patterns of Atrophy:

Lobar Atrophy: In early stages, the distribution of atrophy based on the affected lobes and the uni-/bilateral distribution pattern gives insights into the underlying etiology:

For example, medial temporal lobe atrophy is a hallmark of AD but can also be observed in other types of dementia like FTD [14].

Focal Atrophy in Specific Dementia Types: For the semantic variant primary progressive aphasia (svPPA), focal atrophy in the anterior temporal lobe (indicative of svPPA) is a defining feature [17].

Cortical Atrophy Patterns: Cortical atrophy in dementia is typified by regional thinning in specific lobes – the frontal, parietal, and temporal lobes being most commonly affected. The pattern of this cortical thinning, discerned through detailed MRI analysis, is instrumental in differentiating between dementia subtypes, as the frontal dominance observed in FTD [18].

In clinical practice, cerebral atrophy is still a predominantly qualitative measurement and is subject to interobserver variability. The systemic use of scoring systems can help quantify the visual assessments of the reporting radiologist and mitigate qualitative bias, thus facilitating a more objective and standardized assessment of neurodegenerative changes [19]. The use of brain volumetry systems is another recent advancement that helps overcome this potential variability. Automated image analysis tools, typically relying on volumetric T1-weighted (3-D T1W) brain MRI scans, are employed to quantify atrophy. Using automated volumetric analyses, structural MRI atrophy patterns demonstrated a sensitivity of 90% and a specificity of 84% in identifying AD [3].

2. FLAIR/ T2- Hyperintensities

White Matter Hyperintensities (WMH) on T2-weighted and FLAIR MR images together with a medical history of high blood pressure are indicative of small vessel disease and are particularly prevalent in VaD. The extent and topography of these hyperintensities is important to further classify the subtypes of small vessel disease, as well as to differentiate it from other demyelinating processes: The rarity of hyperintensity extension into the temporal poles usually suggests a non-vascular origin of the pathology, potentially indicating conditions like cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) or demyelinating diseases such as multiple sclerosis [7].

Subcortical Hyperintensities [7] :

Lacunar infarcts, typically 2 to 20 mm in size, primarily affect subcortical white matter and deep gray matter, identified as a cerebrospinal fluid signal intensity with a surrounding hyperintense rim on T2-weighted images. Lacunes are also indicative of cerebral small vessel disease.

Hyperintensities in striatum and/or neocortex: Hyperintensities in the cortex and/or basal ganglia, particularly in the putamen, seen on both FLAIR and DWI sequences, are commonly seen in Creutzfeldt-Jakob disease, a rare etiology of dementia.

3. Cerebral Microbleeds (CMB)

Seen as small, rounded, homogeneous hypointense lesions on SWI, these are relevant findings in cases often seen in various vascular-related dementia syndromes.

Hypertension-related CMBs typically manifest in deeper brain areas such as the basal ganglia, thalamus, and brainstem. In contrast, CMBs linked with cerebral amyloid angiopathy, often seen in AD, predominantly appear in cortical-subcortical (lobar) regions ([Fig. 2]). Notably, these two types can coexist in older individuals with plausible additive effects on cognitive decline [18].

Fig. 2 Axial view of susceptibility-weighted MR images showing multiple punctate bilateral hypointensities, corresponding to cerebral microbleeds in two different patients (a 56-year-old male patient, b 84-year-old male patient). In a, the microbleeds involve the basal ganglia (arrow), which is likely due to hypertension. The patient had a long-term history of hypertension which was also in agreement with the MRI findings. In contrast to a, in b the microbleeds are located more peripherally and subcortically (arrow) with relative sparing of the basal ganglia. This pattern of microbleeds is commonly seen in cerebral amyloid angiopathy. This patient had no known history of hypertension and had a history of dementia, which could have been associated with this finding.

4. Subcortical structural changes

Subcortical atrophy and hyperintensities observed on MRI are key features in Parkinson’s disease dementia and DLB, with notable changes in the basal ganglia and brainstem regions [19].

Recent updates regarding the detection of neurostructural changes of dementia:

The latest advances in MRI technologies, though not yet standard in the clinical routine, have transformative potential in the field of dementia research. Ultra-high-field 7T-MRI, with its enhanced signal-to-noise ratio, contrast, and spatial resolution, allows for unprecedented visualization of brain structures pertinent to neurodegenerative diseases [20] [21]. Complementing this, experimental protocols such as functional MRI (fMRI) offer insight into the broad-scale neural activity affected by dementia, whereas arterial spin labeling (ASL) perfusion MRI, magnetic resonance spectroscopy (MRS), voxel-based imaging, and diffusion tensor imaging (DTI) contribute by measuring cerebral blood flow, chemical composition, and white matter integrity, respectively. Collectively, these advancements promise a significant leap forward in the early detection, differentiation, and understanding of the underlying pathophysiology of dementia using MRI [3] [22] [23]. Further prospective clinical studies, which account for factors such as cost and time efficiency versus clinical applicability and therapeutic consequences, are needed and could bring us closer to establishing such modalities as part of routine examinations in the future. Ongoing research and technological advances are already addressing many of these barriers to enable these methods to become more validated, accessible, and standardized FTD [21].

Step 4: Identifying disease-specific patterns on MRI

After detection of the neuroanatomical pathologies, the correlation of the observed findings with the known specific patterns attributed to the various etiologies of dementia as follows.

The predominant forms of dementia include AD, LBD, VaD, and FTD [3]. These are classified and summarized with their corresponding common MRI findings in [Table 1].

Table 1 Common imaging findings and specific diagnosis of the most common dementia etiologies.

Entity	Classic Imaging Finding	Visual Rating Scale
Alzheimer’s disease:
Typical AD
Prominent hippocampal and mediotemporal atrophy. Early structural changes start in the entorhinal cortex and then involve the hippocampus and eventually the neocortex; a pattern consistent with the Braak staging system. Other specific changes include cortical atrophy, particularly in the parietal and temporal lobes.		MTA or ERICA scale
Atypical AD:
PCA	Posterior cortical atrophy. Prominent occipital and parietal lobe atrophy with relative sparing of the two hippocampi.	Koedam Visual Rating scale
Logopenic variant	Atrophy of the left posterior perisylvian and/or parietal region. Can be observed in the absence of other typical AD findings.	Non-specific
Frontal variant	Predominant temporoparietal atrophy, with additional involvement of frontal regions.	Non-specific
Frontotemporal dementia:
bvFTD	Pronounced frontal lobe atrophy, particularly in the orbitofrontal and medial prefrontal cortices, alongside atrophy in the anterior temporal lobes, often more pronounced on the right side.	Non-specific
PNFA	Left frontal atrophy, particularly in the inferior and dorsolateral prefrontal cortex, extending to the superior temporal region, displays asymmetric degeneration in the dominant hemisphere, particularly in the perisylvian region, visible on MRI as left frontal atrophy extending to the inferior and dorsolateral prefrontal cortex and the superior temporal region.	Non-specific
svPPA	Bilateral asymmetric anterior temporal lobe atrophy, often more pronounced on the left side.	Non-specific
Vascular dementia	Extensive white matter hyperintensities on T2-weighted images, lacunar infarcts, and microbleeds, predominantly in periventricular areas, subcortical zones, or within the deep white matter. Infarcts in critical brain area.	Fazekas scale
Lewy Body Dementia (LBD): (DLB and PDD types)	Lack of a typical high-signal-intensity swallow tail sign, mild generalized brain atrophy, atrophy in the dorsal mesopontine region and relative preservation of the temporal lobes. PDD type typically presents with more pronounced atrophic changes in the substantia nigra along with cortical atrophy that is less severe than in AD of the medial temporal lobes.	Non-specific

1. Alzheimer’s Disease

AD can manifest in different forms, with “typical” and “atypical” presentations representing distinct patterns of brain atrophy.

Typical AD is characterized by memory impairment, primarily resulting from prominent hippocampal and mediotemporal atrophy. Early structural changes start in the entorhinal cortex and progress to involve the hippocampus later in the disease, and eventually the neocortex – a pattern consistent with the Braak staging system [3]. Other specific changes include cortical atrophy, particularly in the parietal and temporal lobes [18]. In clinical practice, the classification of atrophy patterns in AD often employs semiquantitative visual rating scales, such as the medial temporal lobe atrophy scale (MTA) and the entorhinal cortical atrophy (ERICA) score, best seen on coronal views of T1-weighted scans [24]:

MTA score: This score assesses hippocampal atrophy on a scale from 0 to 4, with higher scores indicating more severe atrophy. Key evaluated features include the width of the choroid fissure, the lateral ventricle’s temporal horn, and the hippocampus’s height ([Fig. 3]). The score’s interpretation varies with age: a score of 2 or more is abnormal for those under 75, while 3 or more is abnormal for those 75 and older. This scoring system is particularly relevant for diagnosing conditions like AD and FTD and is associated with the progression from mild cognitive impairment to dementia. In addition, factors like gender, age, and education can influence the score.

Fig. 3 The different stages of medial temporal lobe atrophy as observed in coronal views of reformations of MPRAGE sequences. Scheltens’ original description assessed hippocampal height, the width of the choroid fissure, and the width of the inferior horn, while more recent protocols also include measurements of the entorhinal and occasionally the perirhinal cortex. Both sides are evaluated separately, and the scores are averaged, resulting in a 0–4 scale where 0 indicates no atrophy and 4 represents severe temporal mesial atrophy. For people less than 75 years of age, a score of ≥ 2 is defined as abnormal; for people over 75 years, a score ≥ 3 is abnormal [41] [42]. a No CSF is observed around the hippocampus of a 70-year-old woman, indicating MTA 0. b Slight enlargement of the temporal horn (white arrow) is compatible with MTA 1 in a 53-year-old man. c Widening of the choroid fissure (green arrow) and slightly enlarged temporal horn (white arrow) of a 67-year-old woman, as seen in MTA 2. d 80-year-old man with marked widening of the choroid fissure, moderate enlargement of the temporal horn, and moderate loss of hippocampal height (orange arrow), consistent with MTA 3. e Advanced widening of the choroid fissure and enlargement of the temporal horn of a 75-year-old man, the hippocampus is markedly atrophied. These findings indicate MTA 4. f Overview of the sagittal 3D T1-weighted gradient-echo sequence and the correct angulation of the reformatted view. As illustrated, the slices should be reformatted perpendicular to the anterior commissure-posterior commissure (AC-PC) line. The slices are then selected posterior to the amygdalae and posterior to or at the level of the mammillary bodies, to ensure homogeneous visualization of the hippocampi, pons, and cerebral peduncles.

Hippocampal atrophy and MTA have been shown to be among the most accurate imaging markers of AD [25]. Many studies have demonstrated that the MTA score is significantly linked to MMSE scores, where a higher MTA score is associated with lower MMSE scores [16].

ERICA score: This score has been developed as a supplement to the MTA score and aids in diagnosing AD by assessing atrophy in the entorhinal cortex, a brain region critical to memory processes located in the medial temporal lobe. This score categorizes atrophy into four stages: 0 indicating normal volume, 1 for mild atrophy, 2 for moderate atrophy with a notable “tentorial cleft sign”, which indicates the presence of CSF between the entorhinal cortex and the tentorium, and 3 for severe atrophy with a wide cleft between the entorhinal cortex and the cerebellar tentorium and pronounced atrophy of the parahippocampal gyrus [26].

Atypical AD:

The atypical AD subtypes mentioned below show distinctive patterns on MRI at an early onset. It is important to note that at a late stage of the disease, the patterns of atrophy overlap and become similar.

The posterior cortical atrophy variant shows prominent occipital and parietal lobe atrophy with relative sparing of the two hippocampi. The assessment of this atrophy pattern is typically conducted using the Koedam Visual Rating Scale [27].

Koedam Visual Rating Scale: Designed for the visual evaluation of parietal atrophy on MRI. It incorporates assessments in sagittal, axial, and coronal views, examining both the left and right hemispheres separately. This scale focuses on three key anatomical landmarks across these orientations: In the sagittal view, it examines the widening of the posterior cingulate and parieto-occipital sulcus, as well as the atrophy of the precuneus. In the axial orientation, it assesses the widening of the posterior cingulate sulcus and the sulci in the parietal lobes. The coronal view evaluates the widening of the posterior cingulate sulcus and changes in the parietal lobes.

The grading system ranges from 0 to 3: Grade 0 indicates no discernible changes in the sulci or atrophy in the assessed regions. Grade 1 shows mild widening of the sulci and mild atrophy in the relevant areas. Grade 2 is characterized by significant sulcal widening and atrophy, while grade 3 denotes severe, end-stage atrophy with pronounced sulcal widening and “knife-blade” atrophy in the parietal lobes and precuneus. In instances where different grades have been assigned in different orientations, the highest grade observed is considered definitive [28] ([Fig. 4]).

Fig. 4 The different grades of posterior parietal atrophy as seen on T1-weighted MPRAGE. a, b, c 57-year-old woman with closed sulci and no gyral atrophy of the precuneus and parietal gyrus. d, e, f 73-year-old man with grade 1 posterior atrophy, showing mild widening of the posterior cingulate and the parieto-occipital sulci (white arrows) as well mild gyral atrophy, particularly along the posterior cingulate sulcus and parietal lobes. g, h, i 79-year-old man with grade 2 posterior atrophy, showing substantial sulcal widening and substantial gyral atrophy in the posterior parietal regions. j, k, l Images of an 81-year-old woman with grade 3 posterior atrophy, showing marked sulcal widening and severe gyral atrophy with characteristic “knife-blade” appearance, particularly evident along the parieto-occipital and posterior cingulate sulci.

Logopenic variant: Atrophy of the left posterior perisylvian and/or parietal region. This atrophy pattern is highly suggestive of lvPPA and can be observed in the absence of other typical AD findings on MRI [17].

Frontal variant : MRI findings reveal predominant temporoparietal atrophy, with additional involvement of frontal regions [29].

2. Frontotemporal Dementia

Frontotemporal dementia (FTD) encompasses diverse phenotypes based on different etiologies, such as behavioral variant FTD (bvFTD), semantic dementia (SD), and progressive nonfluent aphasia (PNFA). The pathognomonic MRI pattern is asymmetric, predominantly frontal frontotemporal atrophy with relative sparing of parietal and occipital lobes. Subcortical atrophy is an early and significant feature in FTD, often preceding cortical atrophy. This subcortical degeneration, particularly in structures like the amygdala and striatum, is linked to early cognitive and behavioral changes. The subtypes are mainly characterized by the clinical picture of the patient but also show neurostructural features specific to each subtype, helpful for differential diagnosis [17] [30] [31]:

bvFTD: MRI typically reveals pronounced frontal lobe atrophy, particularly in the orbitofrontal and medial prefrontal cortices, alongside atrophy in the anterior temporal lobes, often more pronounced on the right side, correlating with the behavioral and personality changes seen in this variant.

PNFA is associated with left frontal atrophy, particularly in the inferior and dorsolateral prefrontal cortex, extending to the superior temporal region, and displays asymmetric degeneration in the dominant hemisphere, particularly in the perisylvian region, evidenced on MRI as left frontal atrophy extending to the inferior and dorsolateral prefrontal cortex and the superior temporal region.

svPPA consistently shows bilateral asymmetric anterior temporal lobe atrophy, which is often more pronounced on the left side, with the left posterior inferior temporal gyrus implicated in semantic processing deficits.

3. Vascular Dementia

VaD, the second most frequent cause of dementia after AD, is a heterogeneous group of cerebrovascular diseases leading to cognitive decline. It manifests through both small- and large-vessel diseases, primarily resulting in confluent alterations in white matter, lacunar infarctions, and/or cortical/subcortical cerebrovascular lesions subsequent to ischemic events [32].

In small vessel disease, MRI reveals extensive white matter hyperintensities on T2-weighted images, lacunar infarcts, and microbleeds, predominantly in periventricular areas, subcortical zones, or within the deep white matter. The severity of these changes is evaluated using the Fazekas scale, which offers a straightforward visual grading system to classify white matter lesions as mild, moderate, or severe [32].

The Fazekas scale is used to quantify the number of white matter lesions that are hyperintense on T2-weighted and FLAIR MR images. The scale categorizes white matter into periventricular white matter (PVWM) and deep white matter (DWM), with each region receiving a grade based on lesion size and aggregation. In the context of dementia assessment, it is the DWM score that holds significance and is commonly reported in clinical settings. DWM scoring is as follows: 0 for no lesions; 1 for punctate foci; 2 for beginning confluence of lesions; and 3 for large confluent areas [33] ([Fig. 5]).

Fig. 5 The different stages of perivascular changes demonstrated on axial FLAIR sequences. a No hyperintensities are seen, consistent with Fazekas grade 0. b punctate foci of hyperintense deep white matter lesions are observed (arrows), compatible with Fazekas grade I. c beginning confluence of periventrical hyperintensities detected (arrows), as seen in Fazekas grade 2. d Extensive confluent areas of periventricular hyperintensities of the deep white matter (arrows) consistent with Fazekas grade 3.

Strategic infarct dementia, another subtype, is characterized by infarcts in critical brain areas controlling cognition. These include, for example, the angular gyrus, the paramedian thalamic area, the inferior medial temporal lobe, and parietal watershed areas in the dominant hemisphere [2] [7]. These infarcts can also be indicative of VaD due to large vessel disease [34].

4. Lewy Body Dementia (LBD)

LBD, which includes dementia with Lewy Bodies (DLB) and Parkinson’s disease dementia (PDD), presents subtle yet distinguishable features on MRI.

The lack of a typical high-signal-intensity “swallow tail” sign in the substantia nigra, has significant diagnostic value on axial high-spatial-resolution SWI for identifying LBD, with a sensitivity of 64% and specificity of 86% for diagnosing LBD on 3T-MRI. Therefore, its presence can be a significant complement to the diagnosis of LBD, but its absence cannot rule out the disease [35].

Furthermore, DLB and PDD are characterized neuropathologically by the abnormal aggregation of the α-synuclein protein, which shows no specific correlating structural change on MRI. MRI, therefore, generally demonstrates nonspecific findings such as mild generalized brain atrophy and atrophy in the dorsal mesopontine region [31]. Moreover, on MR images the medial temporal lobes are relatively preserved in contrast to patients with AD, as mentioned in the consensus diagnostic criteria for DLB as a supportive criterion for diagnosis [29] [36] [37].

DLB: Cortical loss typically begins in the occipital lobes while sparing the posterior cingulate gyrus (cingulate island sign) [6].

PDD: Does not show characteristic imaging patterns that can be reliably used as definitive diagnostic markers. However, it could present with more pronounced atrophic changes in the substantia nigra along with cortical atrophy that is less severe than in AD [32].

Recent developments of disease-specific patterns on MRI:

Through enhanced resolution, 7T-MRI facilitates detailed imaging of cortical and subcortical degeneration, vascular pathology, functional brain disruptions, and iron dysregulation, thereby improving our understanding of AD pathology and paving the way for targeted therapeutic interventions. Furthermore, ultra-high-field MRI has improved the understanding of PD by allowing detailed visualization of the substantia nigra’s structure and the changes associated with the disease, such as iron accumulation and the loss of the “swallow tail” sign in nigrosome-1 [20]. Though mainly used for research purposes, understanding the neuroanatomical changes observed in the various etiologies will help put the focus on some specific areas during diagnosis.

Moreover, there is a new class of therapeutics expected to be approved by the European Medicines Agency (EMA) in 2024, such as donanemab, which allows for anti-amyloid treatment. Notably, monoclonal antibody immunotherapy has been shown to cause so-called “amyloid-related imaging abnormalities” (ARIA) that are further classified into two categories, ARIA-E and ARIA-H, representing edema and/or effusion and hemorrhage, respectively. Patients treated with these substances require regular MRI screening to monitor such complications. Cerebral amyloid angiopathy has an identical pathogenesis to that of ARIA and is its closest differential diagnosis, with imaging findings being the same for both entities and only a history of monoclonal antibody administration allowing differentiation [38] [39].

Conclusion

MRI plays an elementary role in the nuanced diagnosis of various forms of dementia. A core step in the diagnosis is to provide an optimal MRI protocol, which allows the analysis of various neuroanatomical changes including brain atrophy patterns, white matter changes, and other characteristic pathologies which are the key to distinguishing between different types of dementia. Furthermore, the future of dementia diagnosis appears promising with the integration of Artificial Intelligence (AI) workflows, which are capable of analyzing multifactorial imaging data quickly and robustly, thereby complementing traditional neuroimaging methods in offering new dimensions in the monitoring and management of the complex neurodegenerative diseases. For instance, many products have been integrated into routine diagnostic workflows in the European clinical landscape, exemplifying the transformative potential of AI in enhancing the precision of neurodegenerative disease management [40]. The major available tools until now include but are not limited to mdbrain and mdlesion from Mediaire in Germany, airascore from Airamed in Germany, and cNeuro and cDSI from combinostics in Finland.

Conflict of Interest

The authors declare that they have no conflict of interest.

• References

• 1 Margallo-Lana ML, Tyrer SP, Moore PB. Overview of the Diagnostic Instruments for Dementia in People with Intellectual Disability. In: Prasher V. , ed. Neuropsychological Assessments of Dementia in Down Syndrome and Intellectual Disabilities. Cham: Springer; 2018: 1-17
• 2 Ahmad DJ, Alghamdixx AA, Alshahrani NA. et al. The role of MRI in assessing cognitive impairment changes. Int J Community Med Public Health 2023; 10 (12)
• 3 Chouliaras L, O’Brien JT. The use of neuroimaging techniques in the early and differential diagnosis of dementia. Mol Psychiatry 2023; 28: 4084-4097
• 4 Deutsche Gesellschaft für Psychiatrie und Psychotherapie, Psychosomatik und Nervenheilkunde (DGPPN) und Deutsche Gesellschaft für Neurologie (DGN) (2023). S3-Leitlinie Demenz (Version 4.0). AWMF-Register Nr. 038–013.
• 5 Raji CA, Benzinger TLS. The Value of Neuroimaging in Dementia Diagnosis. Continuum (Minneap Minn) 2022; 28 (03) 800-821
• 6 Patel KP, Wymer DT, Bhatia VK. et al. Multimodality Imaging of Dementia: Clinical Importance and Role of Integrated Anatomic and Molecular Imaging. RadioGraphics 2020; 40: 200-222
• 7 Harper L, Barkhof F, Scheltens P. et al. An algorithmic approach to structural imaging in dementia. J Neurol Neurosurg Psychiatry 2014; 85: 692-698
• 8 Živanović M, Aracki Trenkić A, Milošević V. et al. The role of magnetic resonance imaging in the diagnosis and prognosis of dementia. Biomolecules and Biomedicine 2023; 23 (02) 209-224
• 9 Verhagen MV, Guit GL, Hafkamp GJ. et al. The impact of MRI combined with visual rating scales on the clinical diagnosis of dementia: a prospective study. Eur Radiol 2016; 26: 1716-1722
• 10 Bittner N, Jockwitz C, Franke K. et al. When your brain looks older than expected: combined lifestyle risk and BrainAGE. Brain Struct Funct 2021; 226 (03) 621-645
• 11 Ruiyang G, Yu Y, Qi YX. et al. Normative modelling of brain morphometry across the lifespan with CentileBrain: algorithm benchmarking and model optimisation. Lancet Digit Health 2024; 6 (03)
• 12 Nutt D, Hayes A, Fonville L. et al. Alcohol and the Brain. Nutrients 2021; 13 (11) 3938
• 13 Rossi DJ. Alcohol and the Cerebellum. In: Gruol DL, Koibuchi N, Manto M. , ed. Essentials of Cerebellum and Cerebellar Disorders. Cham: Springer; 2023: 431-439
• 14 Dhikav V, Duraiswamy S, Anand KS. Correlation between hippocampal volumes and medial temporal lobe atrophy in patients with Alzheimer’s disease. Ann Indian Acad Neurol 2017; 20 (01) 29-35
• 15 Pedraza MI, de Lera M, Bos D. et al. Brain atrophy and the risk of futile endovascular reperfusion in acute ischemic stroke. Stroke 2020; 51 (05) 1514-1521
• 16 Wei M, Shi J, Ni J. et al. A new age-related cutoff of medial temporal atrophy scale on MRI improving the diagnostic accuracy of neurodegeneration due to Alzheimer’s disease in a Chinese population. BMC Geriatrics 2019; 19: 59
• 17 Sajjadi SA, Sheikh-Bahaei N, Cross J. et al. Can MRI visual assessment differentiate the variants of primary-progressive aphasia?. AJNR Am J Neuroradiol 2017; 38 (05) 954-960
• 18 Du A-T, Schuff N, Kramer JH. et al. Different regional patterns of cortical thinning in Alzheimer’s disease and frontotemporal dementia. Brain: A J Neurol 2007; 130 (04) 1159-1166
• 19 Harper L, Fumagalli GG, Barkhof F. et al. MRI visual rating scales in the diagnosis of dementia: evaluation in 184 post-mortem confirmed cases. Brain 2016; 139 (04) 1211-1225
• 20 Düzel E, Costagli M, Donatelli G. et al. Studying Alzheimer disease, Parkinson disease, and amyotrophic lateral sclerosis with 7-T magnetic resonance. European Radiology Experimental 2021; 5: 36
• 21 Perera Molligoda Arachchige AS, Garner AK. Seven Tesla MRI in Alzheimer’s disease research: State of the art and future directions: A narrative review. AIMS Neurosci 2023; 10 (04) 401-422
• 22 Chen Y, Wang Y, Song Z. et al. Abnormal white matter changes in Alzheimer’s disease based on diffusion tensor imaging: A systematic review. Ageing Res Rev 2023; 87: 101911-101911
• 23 Alsop DC, Detre JA, Golay X. et al. Recommended implementation of arterial spin-labeled perfusion MRI for clinical applications: A consensus of the ISMRM perfusion study group and the European consortium for ASL in dementia. Magn Reson Med 2015; 73 (01) 102-116
• 24 Molinder A, Ziegelitz D, Maier SE. et al. Validity and reliability of the medial temporal lobe atrophy scale in a memory clinic population. BMC Neurol 2021; 21 (01) 289
• 25 Kaushik S, Vani K, Chumber S. et al. Evaluation of MR Visual Rating Scales in Major Forms of Dementia. J Neurosci Rural Pract 2021; 12 (01) 16-23
• 26 Enkirch SJ, Traschütz A, Müller A. et al. The ERICA Score: An MR Imaging–based Visual Scoring System for the Assessment of Entorhinal Cortex Atrophy in Alzheimer Disease. Radiology 2018; 288 (01) 226-233
• 27 Haller S, Jäger HR, Vernooij MW. et al. Neuroimaging in Dementia: More than Typical Alzheimer Disease. Radiology 2023; 308 (03) e230173
• 28 Koedam ELGE, Lehmann M, van der Flier WM. et al. Visual assessment of posterior atrophy development of a MRI rating scale. Eur Radiol 2011; 21 (12) 2618-2625
• 29 Ossenkoppele R, Pijnenburg YAL, Perry DC. et al. The behavioural/dysexecutive variant of Alzheimer’s disease: Clinical, neuroimaging and pathological features Brain. 2015; 138: 2732-2749
• 30 Snowden JS. Changing perspectives on frontotemporal dementia: A review. J Neuropsychol 2023; 17: 211-234
• 31 Grossman M. Frontotemporal dementia: A review. J Int Neuropsychol Soc 2002; 8: 566-583
• 32 Furtner J, Prayer D. Neuroimaging in dementia. Wiener Medizinische Wochenschrift 2021; 171: 274-281
• 33 Kim KW, MacFall JR, Payne ME. Classification of white matter lesions on magnetic resonance imaging in elderly persons. Biol Psychiatry 2008; 64 (04) 273-280
• 34 Kantarci K, Ferman TJ, Boeve BF. et al. Focal atrophy on MRI and neuropathologic classification of dementia with Lewy bodies. Neurology 2012; 79 (06) 553-560
• 35 Shams S, Fällmar D, Schwarz S. et al. MRI of the Swallow Tail Sign: A Useful Marker in the Diagnosis of Lewy Body Dementia?. AJNR Am J Neuroradiol 2017; 38 (09) 1737-1741
• 36 McKeith IG, Boeve BF, Dickson DW. et al. Diagnosis and management of dementia with Lewy bodies: Fourth consensus report of the DLB Consortium. Neurology 2017; 89 (01) 88-100
• 37 Broski SM, Hunt CH, Johnson GB. et al. Structural and Functional Imaging in Parkinsonian Syndromes. RadioGraphics 2014; 34: 1273-1292
• 38 Hampel H, Elhage A, Cho M. et al. Amyloid-related imaging abnormalities (ARIA): radiological, biological and clinical characteristics. BRAIN 2023; 146: 4414-4424
• 39 Roytman M, Mashriqi F, Al-Tawil K. et al. Amyloid-Related Imaging Abnormalities: An Update. AJR Am J Roentgenol 2023; 220 (04) 562-574
• 40 Battineni G, Chintalapudi N, Hossain MA. et al. Artificial Intelligence Models in the Diagnosis of Adult-Onset Dementia Disorders A Review. Bioengineering 2022; 9 (08) 370
• 41 Scheltens P, Leys D, Barkhof F. et al. Atrophy of medial temporal lobes on MRI in “probable” Alzheimer’s disease and normal ageing: diagnostic value and neuropsychological correlates. J Neurol Neurosurg Psychiatry 1992; 55: 967-972
• 42 Westman E, Cavallin L, Muehlboeck JS. et al. Sensitivity and specificity of medial temporal lobe visual ratings and multivariate regional MRI Classification in Alzheimer’s Disease. PLoS One 2011; 6 (07) e22506

Publication History

Received: 11 May 2024

Accepted after revision: 09 March 2025

Article published online:
10 April 2025