Corneal nerve status as a biomarker for neurodegenerative and neurological diseases: a narrative review

Abstract

Corneal nerve status has gained prominence as a potential adjunctive biomarker for neurodegenerative and neurological diseases, as corneal small-fibre abnormalities can be detected in peripheral neuropathy and may precede conventional manifestations in certain conditions. Advances in in vivo confocal microscopy (IVCM) allow non-invasive, cellular-level assessments of corneal nerve alterations in real-time. Corneal nerve metrics have been associated with disease progression and severity. This review explores the role of corneal nerve status in the evaluation and monitoring of neurodegenerative and neurological diseases or conditions, including diabetic peripheral neuropathy (DPN), Parkinson’s disease (PD), mild cognitive impairment (MCI) and dementia, multiple sclerosis (MS), schizophrenia, amyotrophic lateral sclerosis (ALS), migraine, tauopathies, and chronic inflammatory demyelinating polyradiculoneuropathy (CIDP). The strength and maturity of evidence vary substantially between conditions. DPN has the most established evidence base, with replicated cross-sectional and longitudinal studies showing reductions in corneal nerve fibre density, branch density and length, and evidence that corneal nerve loss may precede conventional manifestations of neuropathy. In MS, longitudinal studies demonstrate progressive corneal nerve loss associated with worsening disability. In PD, the extent of corneal nerve fibre loss correlates with severity of motor deterioration, cognitive impairment and development of autonomic dysfunction, although proposed applications in prodromal detection, disease phenotyping, and prognostication remain largely dependent on small, single-centre studies. Studies in MCI and dementia suggest potential diagnostic and prognostic utility, including comparisons with neuroimaging, but are largely derived from overlapping single-centre cohorts. Evidence in CIDP, migraine, ALS, and schizophrenia remains limited or conflicting, while evidence in tauopathies is predominantly pre-clinical. Overall, current literature supports corneal nerve parameters primarily as promising adjunctive rather than established diagnostic, prognostic, treatment-response biomarkers. Direct evidence that IVCM provides incremental diagnostic or prognostic value over established biomarkers remains limited. Furthermore, heterogeneity in confocal platforms, image acquisition, image sampling, and analytical methods must be considered in the interpretation across studies. Standardised acquisition and analysis protocols, prospective multicentre longitudinal studies, appropriate adjustment for confounders, and direct comparison with established biomarkers are necessary before broader clinical translation.

Introduction

The cornea is a transparent, avascular tissue in the eye that plays a crucial role in vision and has dense innervation and the ability to reflect peripheral nerve changes [1]. Recent research has suggested the potential value of corneal nerve architecture as a biomarker for various neurodegenerative and neurological diseases. The nerves observed in in vivo confocal microscopy (IVCM) typically consist of peripheral C and Aδ sensory neurons originating from the trigeminal ganglion [2]. Corneal neurons have shorter axons and exist in a slightly different tissue environment at their peripheral terminals. However, they seem similarly susceptible to metabolic and toxic insults, making them a promising non-invasive biomarker for detecting sensory peripheral neuropathy [3].

Advancements in automated analysis techniques have improved the accuracy and efficiency of IVCM assessments. Fully automated IVCM analysis, such as the ACCMetrics software developed by Dabbah et al., has been instrumental in these findings. Moreover, automated analysis enables rapid image analysis, facilitating scalability for large cohort studies and longitudinal assessments [4].

Emerging evidence has supported the use of corneal nerve fibre parameters as diagnostic and prognostic biomarkers for neurodegenerative and neurological diseases, including but not limited to diabetic peripheral neuropathy (DPN), Parkinson’s disease (PD), and multiple system atrophy (MSA). Quantitative changes in corneal nerve metrics, evaluated by IVCM, have been associated with disease progression and severity. This review synthesises current research findings and evaluates corneal nerve status as a potential biomarker for various neurodegenerative and neurological diseases. Through a comprehensive analysis of the literature, we aim to underscore the role of corneal nerve fibre assessments in advancing our understanding of neurodegenerative and neurological diseases and potentially improving clinical practice.

Search strategy and selection criteria

This article is a narrative review. To provide an overview of the role of corneal nerve status as a diagnostic and monitoring marker for neurodegenerative and neurological diseases, an electronic literature search of the PubMed database was performed. The PubMed search initially focused on articles published from 2015 to June 2025. Relevant earlier publications and newer papers encountered during revision were also included.

We employed a combination of keywords and medical subject headings (MeSH) to capture a broad yet focused range of relevant studies. The search terms included but were not limited to “corneal nerves” AND “diabetic peripheral neuropathy”, “peripheral neuropathy”, “corneal nerves” AND “neurodegenerative diseases”, “corneal confocal microscopy” AND “Parkinson’s disease”, “corneal nerve fibre” AND “dementia” OR “Alzheimer’s disease”, “corneal neuropathy” AND “multiple sclerosis”, “in vivo confocal microscopy” AND “schizophrenia” OR “amyotrophic lateral sclerosis”, “in vivo confocal microscopy” AND “motor neuron disease”, “corneal sensitivity” AND “migraine”, “tauopathies” AND “corneal nerves”, “chronic inflammatory demyelinating polyradiculoneuropathy” AND “corneal nerves”.

The authors included only articles published in English with full texts available, and restricted the date range to the most recent 10 years (2015–2025) to ensure the inclusion of current and relevant research. The authors also manually reviewed the bibliographies of the selected articles to identify additional pertinent studies that might not have appeared in the initial database search. Both observational and interventional studies that provided qualitative or quantitative data on corneal nerve changes in these conditions were identified. Large language models were used for language editing only; all scientific content was reviewed and verified by the authors.

Consistent with the narrative design of this review, articles were selected by the authors on the basis of relevance to each disease section rather than through a formal screening protocol. No PRISMA framework, duplicate independent screening or standardised risk-of-bias instrument was applied. Studies were instead appraised qualitatively, with attention to study design, sample size, the availability of longitudinal follow-up and whether findings had been replicated by independent groups. These appraisals are presented at the end of each disease section, are compared across conditions in the section comparing the strength of evidence across diseases, and the limitations of this approach are acknowledged under Limitations and Future Directions.

Corneal nerve status in relation to diabetic peripheral neuropathy (DPN)

Introduction

DPN is the most frequent long-term complication of diabetes [5], characterised by damage to the peripheral nerves, leading to neuropathic pain, loss of sensation, and in severe cases, limb amputation. The diagnosis and monitoring of DPN pose significant challenges due to the insidious onset and slow progression of the disease [6]. Recently, IVCM has emerged as a promising non-invasive technique to detect early signs of DPN by examining corneal nerve status. This technique offers unique advantages in the early detection and monitoring of DPN progression and severity [7].

Mechanisms underlying corneal nerve changes and DPN

The mechanisms underlying corneal nerve changes in DPN are largely driven by hyperglycaemia-induced metabolic and vascular disturbances. Chronic hyperglycaemia leads to the accumulation of advanced glycation end-products, oxidative stress, and inflammation, causing neuronal damage and loss [8]. Metabolically, excess glucose shunted through the polyol pathway leads to the accumulation of harmful by-products such as sorbitol and fructose, increasing osmotic and oxidative stress. Additionally, microvascular dysfunction reduces blood supply to nerves, exacerbating ischaemia and nerve degeneration [9]. Inflammatory mediators further worsen neurodegeneration [10]. While both diabetic corneal neuropathy (DCN) and DPN share these mechanisms, IVCM provides an imaging modality to directly visualise small-fibre nerve changes in the cornea, which are often the earliest detectable signs of neuropathy, making DCN a valuable adjunctive biomarker for DPN.

In-vivo confocal microscopy findings in DPN

On IVCM evaluation, corneal sub-basal nerves appear as well-defined, densely anastomosed, thin and linear structures with homogeneous reflectivity and Y-shaped or H-shaped branches [11]. IVCM presentations in DPN are characterised by decreased Corneal Nerve Fibre Density (CNFD), Corneal Nerve Fibre Length (CNFL), tortuous nerves, and nerve beading (Figure 1).

FIGURE 1

Early detection of peripheral nerve damage in DPN

In the early stages of DPN, IVCM can detect subclinical corneal nerve alterations before traditional diagnostics become abnormal: patients with diabetes show reduced CNFD, corneal nerve branch density (CNBD), and CNFL compared to healthy controls, with these deficits preceding both clinical symptoms and nerve conduction changes in DPN [7, 12–14]. In a longitudinal rat model of prediabetes and type 2 diabetes, animals fed a high-fat diet exhibited significant early reductions in CNFL and corneal sensitivity, even before alterations in intraepidermal nerve fibre density or thermal nociception, thereby corroborating IVCM’s superior sensitivity for detecting subclinical DPN [15]. Among IVCM parameters, reduced CNFL, particularly within the inferior whorl region, has emerged as the most sensitive early marker of diabetic sensorimotor polyneuropathy, often preceding overt symptoms by years and outperforming conventional nerve conduction studies in subclinical detection [16].

DCN in relation to diagnosis of DPN

Corneal nerve parameters are reliable and reproducible biomarkers for detecting and staging DPN. Quantitative corneal nerve parameters, including CNFL, CNFD, and CNBD, are significantly reduced in patients with DPN compared to those without [17, 18]. A meta-analysis that included 13 studies and 1680 participants concluded that corneal nerve parameters were significantly reduced in patients with diabetes mellitus (DM), especially when the DM status was complicated with DPN, implying that corneal nerve morphology assessment is useful for the detection and evaluation of early nerve damage in patients with DM [19]. CNFL has demonstrated good reproducibility as a potential biomarker of early DPN [20]. This was corroborated by a longitudinal assessment study examining 396 participants across three groups, which found CNFL was significantly lower in the DPN group versus diabetics without neuropathy (14.0 ± 6.4 mm/mm² vs. 19.1 ± 5.8 mm/mm², P < 0.001) [21].

Conventional DPN diagnostics, such as nerve conduction studies, demonstrate limited sensitivity for detecting small-fibre pathology. Skin biopsy, the current gold standard for diagnosing small fibre neuropathy, is invasive and hence less appealing as a diagnostic and screening tool [22]. In contrast, IVCM provides direct visualisation of small unmyelinated C-fibres and Aδ-fibres, which are the first neural structures affected by hyperglycaemia-induced metabolic toxicity. Corneal nerve parameters demonstrate substantial diagnostic utility, with Ahmed et al. reporting area under the curve (AUC) values of 0.88 for CNFL and 0.84 for CNFD for distinguishing participants with diabetes with versus without DPN [23]. Ferdousi et al. also reported a CNFD AUC of 0.81 [24]. Moreover, CNFD had a higher diagnostic AUC than intraepidermal nerve fibre density in the same cohort (0.81 versus 0.73) [25].

Corneal nerve metrics also provide predictive value for DPN development. A CNFL threshold of <14.9 mm/mm² predicted diabetic peripheral neuropathy development with a sensitivity of 0.82 and a specificity of 0.69 [26]. Furthermore, CNFL provided prognostic information up to 6 years before DPN diagnosis, with an optimal cut-off of 14.1 mm/mm², yielding a sensitivity of 67% and specificity of 71% [13]. A sustained abnormal CNFL for at least 50% of a 2-year observation period demonstrated superior predictive capacity for incident DPN (AUC: 0.763, p < 0.0001) compared to single measurements, emphasising the value of serial monitoring [27].

Disease progression and severity in DPN

Corneal nerve parameters correlate closely with DPN severity, providing both functional and structural assessments of disease progression. Functionally, corneal sensitivity decreases significantly with advancing DPN, showing strong correlations with clinical neuropathy scores (r = 0.441, P < 0.001) and quantitative sensory thresholds, including cold, vibration, and warm sensation [28, 29]. Structurally, CNFD, CNBD, and CNFL progressively decline with increasing DPN severity, as assessed by neuropathy scores and nerve conduction studies (P < 0.001), enabling stratification of patients across mild, moderate, and severe neuropathy stages [30, 31]. CNFD correlated inversely with the Total Neuropathy Score (ρ = −0.783, P < 0.01) [32]. Separately, cross-sectional evidence suggests that CCM can detect small-fibre abnormalities at earlier clinical stages of diabetic neuropathy than skin-biopsy-derived intraepidermal nerve fibre assessment, while providing a rapid, non-invasive examination [30].

These corneal metrics show significant negative correlations with neuropathic symptom severity scores and pain intensity, with CNFD, CNBD, and CNFL all declining with increasing pain severity on visual analogue scale ratings [33].

The progressive corneal nerve changes enable identification of patients at risk for serious DPN complications. CNFD demonstrates high specificity for early-stage small fibre neuropathy (84%), DPN diagnosis (75%), and individuals at risk for foot ulceration (72%), serving as an objective marker for complication stratification [34]. Furthermore, patients developing diabetic neuroosteoarthropathy show significantly reduced corneal nerve metrics compared to non-diabetic controls, and corneal nerve changes correlate with neuropathic pain severity [35, 36]. Thus, IVCM assessment of corneal nerve status enables risk stratification across the spectrum of DPN severity and complications, guiding the optimal timing of therapeutic interventions.

AI-based diagnostic approaches

Recent advances have harnessed artificial intelligence (AI) to automate and enhance IVCM analysis for DPN detection, addressing limitations of manual quantification, including inter-observer variability and time-intensive processing.

Deep learning algorithms have significantly improved diagnostic performance. A convolutional neural network (CNN)-based algorithm employing data augmentation enabled rapid quantification of multiple nerve parameters, including fibre length, branch points, tail points, nerve segments, and fractal dimensions, without manual intervention. This approach achieved an AUC of 0.83, specificity of 0.87, and sensitivity of 0.68 for DPN diagnosis, outperforming the ACCMetrics tool [37]. Another CNN trained on IVCM images achieved 96% accuracy in separating healthy controls from diabetics with neuropathy, without the need for manual nerve tracing or parameter extraction [38]. Advanced ResNet-50 models incorporating Grad-CAM (gradient-weighted class activation mapping) visualisation achieved sensitivity of 0.91, specificity of 0.93, and AUC of 0.95 in patients with Type 1 DM (T1DM), Type 2 DM (T2DM), and prediabetes using IVCM images of the sub-basal nerve plexus, without requiring prior manual quantification of corneal nerve parameters [39]. In a cross-sectional study of 369 participants, including healthy controls and people with T1DM, T2DM or prediabetes, a CNN classifier was developed to process single corneal images for DPN severity stratification, achieving F1-scores of 0.91 for healthy controls, 0.88 for participants without peripheral neuropathy, and 0.91 for participants with peripheral neuropathy [40].

U2Net-based segmentation tools outperformed ACCMetrics in consistency with manual annotation and achieved an AUC of 0.75 for DPN diagnosis, further improving the reliability of automated nerve quantification [41].

Transformer-based models represent recent advances. A Swin transformer network outperformed traditional CNNs with an AUC of 0.94 for DPN diagnosis, highlighting the potential of transformer architectures for rapid, accurate screening [42], while Vision Transformer models achieved an AUC of 0.99, sensitivity of 98%, and specificity of 92% for DPN diagnosis from IVCM images without manual segmentation [43]. Hybrid approaches combining deep learning with an adaptive neuro-fuzzy inference system for DPN diagnosis achieved an AUC of 0.95 with 92% sensitivity and 80% specificity, demonstrating the potential for rapid, automated, and accurate diagnosis [44].

A recent deep learning model extracted features directly from corneal images and, when combined with HbA1c, achieved an AUC of 0.94 for predicting diabetic foot risk [45]. Integration with serum creatinine produced an AUC of 0.93 for identifying high-risk patients with diabetic foot (Figure 2). Notably, quantitative nerve parameters did not enhance performance compared with raw images, indicating that deep learning can capture diagnostically relevant features without manual extraction. This approach provides a novel screening methodology for diabetic foot risk categories.

FIGURE 2

Table 1 provides a comprehensive overview of studies evaluating corneal nerve status in relation to DPN.

Counter-evidence: limitations of corneal nerve metrics

Despite the preponderance of data supporting the diagnostic value of using corneal nerve status, not all studies corroborate its reliability as a DPN biomarker. A recent meta-analysis found that standard corneal nerve parameters (CNFL, CNFD and CNBD) do not significantly differ between patients with painful and painless DPN, indicating that these metrics cannot distinguish between pain subtypes and are therefore limited as pain-specific biomarkers in DPN [48]. Furthermore, a lack of association was found between corneal nerve density and clinical measures of DPN, including neuropathy symptoms and disability scores, nerve conduction studies, and quantitative sensory tests [49]. These results underscore the need for cautious interpretation of IVCM findings and highlight the importance of developing more accurate quantitative tools and multimodal diagnostic approaches for DPN assessment.

Overall appraisal of the evidence in DPN

The evidence base in DPN is the strongest of any condition considered in this review. It comprises large cross-sectional cohorts, prospective longitudinal studies with follow-up of two to 6 years, and meta-analytic synthesis. The core findings - reduction of CNFD, CNBD and CNFL in diabetes, and the predictive value of CNFL for incident DPN, have been reproduced by independent groups in different countries. Several caveats nonetheless apply. Most diagnostic accuracy estimates derive from single-centre cohorts analysed with a common software pipeline, few studies report blinded assessment or pre-specified diagnostic thresholds, and the artificial intelligence studies in particular are susceptible to optimistic performance estimates where external validation datasets were not used. Findings relating to the detection and progression of DPN may therefore be regarded as robust, whereas the pain-specific and artificial intelligence-based applications remain exploratory.

Corneal nerve status in relation to Parkinson’s disease

Introduction

Parkinson’s disease (PD) is a progressive neurodegenerative disorder primarily characterised by motor symptoms such as tremors, bradykinesia, and rigidity. However, non-motor symptoms, including hyposmia, sleep disorders, and autonomic dysfunction, are increasingly recognised as significant contributors to the disease burden and often precede other motor manifestations [50]. Emerging data indicate that both dopaminergic and non-dopaminergic pathology underlie these non-motor features, with the alpha-synuclein pathology starting in the gastrointestinal autonomic plexus and spreading along the neuronal pathways [51]. There is also accumulating evidence demonstrating early α-synuclein deposition and small-fibre degeneration within the skin, autonomic nervous system, and enteric plexus [52, 53]. Given that corneal nerves are predominantly small, unmyelinated sensory fibres, analogous to those affected in cutaneous and autonomic neuropathy, corneal nerve involvement in PD is biologically plausible. This provides a mechanistic rationale for evaluating corneal nerve alterations as an adjunctive biomarker of peripheral small-fibre neurodegeneration in PD. To support this, recent studies have suggested that corneal nerve fibre alterations could serve as a novel biomarker for diagnosing and monitoring PD.

Evidence of corneal neuropathy in early PD and disease progression

Multiple studies have demonstrated that corneal nerve fibre loss is present in early and even prodromal PD, preceding overt motor symptoms. In a clinical cross-sectional study including 26 patients with newly diagnosed PD and 22 control subjects, these patients presented with significantly reduced CNFD and CNFL as compared to controls. Skin biopsies showed similar epidermal nerve densities between groups, indicating that IVCM is more sensitive for detecting early small fibre neuropathy in PD. The authors thus concluded that corneal nerve metrics provided evidence of pre-clinical peripheral neuropathy in newly diagnosed PD patients, aiding in the early diagnosis and management of PD [54].

The deterioration of corneal nerve metrics shows a consistent pattern across the disease spectrum, from healthy individuals to prodromal and manifest PD. A cross-sectional study demonstrated that CNFD and CNFL values declined progressively from healthy controls through prodromal PD to manifest PD, supporting their application as early screening markers for PD [55].

In a systematic review of IVCM and its applications in PD assessment, Che and Yang comprehensively analysed findings from multiple IVCM studies and consistently identified significantly reduced CNFD in PD patients relative to healthy controls [56]. Notably, the magnitude of corneal nerve fibre loss showed progressive worsening with advancing disease severity. The authors concluded that IVCM might serve as a supplementary tool for assessing PD progression. Through highlighting the sensitivity of IVCM in detecting early nerve fibre alterations, they proposed its use as a diagnostic tool for PD-related neuropathy.

Corneal nerve fibre loss and cognitive impairment in PD

Research has demonstrated a link between corneal nerve fibre loss and cognitive impairment in patients with PD. A study by Che et al. demonstrated that patients with PD exhibited significant corneal nerve fibre loss compared to controls, with CNFD positively correlating with cognitive performance as assessed by the Montreal Cognitive Assessment, indicating that corneal nerve fibre loss is associated with the severity of cognitive dysfunction [57].

Another study observed a marked reduction in corneal sub-basal nerve plexus density in PD patients compared to healthy controls, and this reduction was significantly correlated with cognitive impairment as measured by the Addenbrooke’s Cognitive Examination-Revised score [58]. This was the first study to directly demonstrate an association between corneal nerve loss and cognitive dysfunction in PD, supporting the involvement of the peripheral nervous system in what was previously considered a purely central nervous system disorder.

Differentiating PD with and without autonomic involvement using IVCM

Autonomic dysfunction is a common non-motor symptom in PD, often complicating the clinical picture. Identifying autonomic involvement early, especially before overt symptoms arise, is crucial for guiding tailored interventions.

Evidence shows that IVCM can distinguish PD patients with autonomic involvement from those without. Corneal nerve parameters, including CNFD, CNBD, CNFL, and the CNBD/CNFD ratio, were all significantly lower in PD patients exhibiting autonomic dysfunction compared to those without. These findings suggest that IVCM offers a non-invasive method for stratifying PD patients based on autonomic nervous system involvement [59].

In support of this, another study demonstrated that CNFD was reduced in PD compared to controls, while CNBD and CNFL were closely correlated with both autonomic symptoms and parasympathetic dysfunction. These findings reinforce the utility of IVCM in identifying autonomic involvement in PD, as the technique’s parameters were not influenced by disease duration or levodopa dose, but rather by the presence of autonomic symptoms. This supports a pathophysiological link between small fibre degeneration and impaired autonomic regulation in PD [60].

Corneal nerve fibre density in motor progression in PD

CNFD has shown promise as a prognostic biomarker for motor progression in PD. Lim et al. conducted a prospective study evaluating the association between baseline CNFD and motor decline over 12 months in PD [61]. Using the Movement Disorder Society Unified Parkinson’s Disease Rating Scale as an outcome measure, they found that patients in the lowest quartile of CNFD at baseline experienced a significantly greater worsening in motor scores compared to those in the highest quartile. These findings suggest that reduced CNFD predicts faster motor deterioration, underscoring its potential utility in identifying PD patients at risk for more rapid functional decline.

Recent investigations have demonstrated significant correlations between corneal nerve parameters and the severity of motor symptoms in PD. Notably, lower CNFD and CNFL were correlated with higher motor symptom scores, indicating that as motor impairment becomes more severe, corneal nerve loss increases [55].

Additionally, IVCM has demonstrated potential in phenotyping motor subtypes within PD. Che et al. reported that patients with the postural instability and gait disturbance subtype exhibited greater corneal nerve fibre loss than those with the tremor-dominant (TD) subtype [62]. Logistic regression analysis revealed that higher CNFD and CNFL were significantly associated with the TD phenotype (OR = 1.265 and OR = 1.706, respectively). Furthermore, receiver operating characteristic (ROC) analysis demonstrated excellent discrimination between tremor-dominant and PIGD subtypes using combined corneal nerve metrics (AUC = 0.832), indicating IVCM’s broader applicability in PD phenotyping beyond autonomic classification.

Consistent with these findings, Che and Yang also demonstrated that decreased CNFD correlated with more pronounced motor symptoms, reinforcing the role of corneal nerve status as a marker of disease progression in PD [56].

Distinguishing between multiple system atrophy and PD

Multiple system atrophy (MSA) and PD share several clinical features, making differential diagnosis challenging. Niu et al. explored the use of IVCM in distinguishing MSA from PD. Results showed that both CNFD and CNFL were significantly lower in MSA patients compared to PD patients. In the combined MSA + PD group, CNFD was negatively correlated with the Orthostatic Grading Scale. In the MSA group alone, CNFD was significantly negatively correlated with both unified multiple system atrophy rating scale (UMSARS) I and II, and regression analysis confirmed that both CNFD and CNFL were inversely associated with MSA severity, demonstrating a statistically significant relationship between corneal nerve fibre loss and disease severity. ROC analysis showed that the algorithm utilising CNFL achieved an AUC of 0.72 with 73.3% sensitivity and 73.0% specificity for distinguishing MSA from PD after adjusting for age, gender, and disease duration. The algorithm combining CNFD, CNBD, and CNFL produced a similar AUC of 0.72. Although the overall discriminatory power was modest, the greater corneal nerve fibre loss in MSA indicates that IVCM can detect a quantitatively distinct pattern of peripheral denervation. When integrated with clinical and paraclinical evaluations, IVCM may provide a useful non-invasive biomarker to aid differential diagnosis in cases where MSA and PD present with overlapping features [63].

Moreover, Lim et al. found significantly lower CNFD, CNBD and CNFL in both PD and MSA than in healthy controls [64]. These comparisons with controls do not establish a direct difference between MSA and PD.

Overall appraisal of the evidence in PD

The PD literature consists predominantly of small to moderate cross-sectional and case-control studies, with a single prospective study of motor progression over 12 months. The reduction of CNFD and CNFL in PD relative to controls has been reproduced by several independent groups and may be considered reasonably established. In contrast, the applications of greatest clinical interest - detection in the prodromal phase, motor subtype phenotyping, prediction of motor progression, and discrimination of MSA from PD - each rest on one or two single-centre studies, show only modest discriminatory performance, and have not been validated in multicentre cohorts. These findings should therefore be interpreted as promising but in the preliminary stage.

Table 2 provides a summary of studies evaluating corneal nerve fibre alterations in the context of PD.

Corneal nerve status in relation to cognitive impairment and dementia

Introduction

Dementia and mild cognitive impairment (MCI) represent a continuum of neurodegenerative disorders characterised by progressive cognitive and functional decline, with the most common subtype being Alzheimer’s disease (AD). MCI represents a stage at which cognitive decline exceeds normal ageing but does not yet meet the criteria to diagnose dementia [65]. MCI may be a precursor to dementia, and brain changes underlying AD occur at least 20 years prior to the onset of symptoms, providing an optimal window for early diagnosis and effective interventions [66]. Early diagnosis and intervention have emerged as one of the key components of the national dementia strategy in certain European countries [67]. Currently, well-established biomarkers for dementia due to AD are visualisation of beta-amyloid (Aβ) and tau deposition in cerebrospinal fluid studies and magnetic resonance imaging (MRI) or radionuclide imaging to assess the degree of atrophy, hypometabolism, or hypoperfusion [68]. IVCM is now gaining prominence in the context of neurodegenerative diseases such as dementia.

Corneal nerve fibre loss and mild cognitive impairment

The medial temporal lobe atrophy (MTA) scoring has been a well-established biomarker for dementia due to its ease of application in clinical practice [69], together with its ability to differentiate between AD and healthy controls and predict the progression of MCI to AD [70]. However, its diagnostic relevance to MCI is still indeterminate and conflicting [68].

In recent times, several studies have shown that there is an association between corneal nerve measures and cognitive impairment. In a comparison between IVCM and the widely used MTA scoring, the diagnostic accuracy of corneal nerve measures was reported to be non-inferior to MTA scoring for dementia, and higher than MTA scoring in MCI, with a higher sensitivity (81% versus 70%) and specificity (81% versus 75%) [71]. Between IVCM and brain volumetry, IVCM also had a higher diagnostic capability in identifying MCI compared to brain volumetry within the studied cohort [72].

In a study conducted by Ponirakis et al. that focused on the predictive value of IVCM parameters versus brain volumetry for progression from MCI to dementia, MCI patients who progressed to dementia had both significantly lower IVCM parameters, including CNFD, CNBD, CNFL, and lower hippocampal volume compared to those who did not [73]. IVCM parameters were significantly associated with cognitive function and functional independence in both MCI and dementia in a similar study [74].

Additionally, further evidence from Gundogan et al. showed significantly lower CNFL, CNFD, and CNBD in AD patients compared to healthy controls. Corneal sensitivity, the clinical functionality of corneal nerve metrics, was also significantly lower in AD patients compared to healthy controls, in conjunction with the aforementioned IVCM parameters [75].

Overall appraisal of the evidence in dementia and MCI

Existing studies provide preliminary evidence that IVCM parameters may represent a non-invasive biomarker of neurodegeneration and could potentially complement established cognitive and neuroimaging assessments. However, available evidence remains largely derived from single-centre cohorts with substantial overlap in investigators, limiting independent validation and generalisability of the reported diagnostic and prognostic performance.

Table 3 shows a summary of studies on corneal nerve status associated with dementia and MCI.

Corneal nerve status in relation to multiple sclerosis

Introduction

Multiple sclerosis (MS) is an autoimmune demyelinating disease of the central nervous system, particularly affecting young women [76]. It commonly involves the optic nerves, brainstem, cerebellum, and spinal cord. Axonal loss occurs early in the disease course and is a major determinant of progressive neurological disability, highlighting the need for biomarkers that facilitate early diagnosis and treatment [77].

MRI is an important imaging endpoint in MS clinical trials and is used to assess disease activity and progression [78]. Retinal nerve fibre layer thickness measured by optical coherence tomography has also been investigated as a marker of neuroaxonal damage [79], although its correlation with clinical disability remains limited. IVCM may provide an additional means of quantifying corneal axonal loss and assessing disease severity.

IVCM applications in MS

In a study of 26 patients with MS and 23 healthy controls, total corneal nerve fibre density was reduced in MS [80]. In a separate study, CNFD, CNBD and CNFL were reduced in MS, and CNFD was inversely associated with the Expanded Disability Status Scale (EDSS) and Multiple Sclerosis Severity Score (MSSS) [81]. In another study comparing clinically isolated syndrome, relapsing-remitting MS, secondary progressive MS, and healthy controls, CNFD, CNFL, and corneal nerve fractal dimension (CNFrD) were significantly reduced across MS disease courses [82].

Among IVCM parameters, CNBD showed the highest discriminative performance between healthy controls and patients with relapsing-remitting MS, while CNFD showed the highest discriminative performance between healthy controls and secondary progressive MS [83].

Monitoring of disease progression

Serial IVCM may detect changes associated with disability progression. In a 2-year prospective cohort of patients with relapsing-remitting MS, corneal nerve fibre area (CNFA) and corneal nerve fibre width (CNFW) decreased significantly as expanded disability status scale increased. Changes in CNFD and CNFL also correlated with changes in expanded disability status scale and multiple sclerosis severity score [84].

Similarly, Petropoulos et al. followed patients with relapsing-remitting MS, secondary progressive MS, and healthy controls over 2 years. CNFD, CNBD, and CNFL decreased significantly in relapsing-remitting MS, while CNFD and CNBD decreased significantly in secondary progressive MS [83].

These findings suggest that corneal nerve abnormalities are detectable across MS disease courses and may change with disability progression, supporting further evaluation of IVCM as a complementary imaging biomarker.

Overall appraisal of the evidence in MS

The MS literature includes two prospective cohort studies and several cross-sectional studies, giving it one of the stronger evidence bases for IVCM after DPN. Longitudinal studies demonstrate progressive corneal nerve loss over 2 years in association with worsening disability, suggesting that IVCM may reflect dynamic neuroaxonal change.

However, the longitudinal cohorts remain relatively small with limited follow-up. Whether IVCM provides additional information beyond MRI or retinal nerve fibre layer assessment by optical coherence tomography has not been formally established. Its diagnostic, prognostic, and disease-monitoring roles therefore remain investigational, and larger prospective multicentre studies comparing IVCM with established MS biomarkers are required before routine clinical use.

Table 4 shows a summary of studies on corneal nerve status associated with MS.

Corneal nerve status in relation to schizophrenia

IVCM applications in schizophrenia

Schizophrenia is a chronic, debilitating psychiatric condition characterised by positive, negative and cognitive symptoms most commonly attributable to genetic susceptibility and environmental influences [85]. While no objective diagnostic markers are available at present, well-accepted diagnostic criteria in the Diagnostic and Statistical Manual of Mental Disorders, fifth edition, are commonly used [86]. This reliance on clinical assessment may limit diagnostic accuracy. Changes in ocular physiology, including corneal nerve alterations, are therefore being investigated as potential biomarkers [87].

In a cross-sectional study involving 62 subjects, there were significant reductions in CNFD, CNFL, CNBD and the CNBD:CNFD ratio in patients with schizophrenia compared with healthy controls [88]. The ratio is CNBD:CNFD MRI brain volumetry also showed a significant decrease in cingulate gyrus volume in schizophrenia patients. The area under the ROC curve for distinguishing schizophrenia from healthy controls was 88% for CNFL, 84% for CNBD, 79% for CNFD and 73% for cingulate gyrus volume.

Comparison of IVCM with existing biomarkers

Neither IVCM parameters nor cingulate gyrus volume correlated with symptom severity or cognitive dysfunction [88]. Structural MRI findings may also be confounded by treatment-related factors such as antipsychotic-associated brain volume changes [89], and have limited correlations with clinical outcomes other than cognitive impairment [90], with no established diagnostic or prognostic role in psychosis [91]. Although IVCM showed greater discriminative accuracy than cingulate gyrus volumetry in a single study, further validation is required before its diagnostic utility can be established.

Overall appraisal of the evidence in schizophrenia

The evidence synthesised in this review regarding schizophrenia is confined to a single cross-sectional study of 62 participants. Although the reported discriminative performance was high, possible confounding factors including dry eye, antipsychotic exposure and metabolic syndrome were not formally assessed or adjusted for. Current limited evidence should be regarded as preliminary and hypothesis-generating rather than as evidence of diagnostic utility.

Corneal nerve status in relation to amyotrophic lateral sclerosis

Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease affecting upper and lower motor neurons, leading to muscle weakness, bulbar dysfunction and, in advanced disease, respiratory failure [92]. Diagnosis remains largely clinical and is supported by electrodiagnostic findings, due to clinical heterogeneity and the lack of reliable objective biomarkers [93]. CSF, biochemical and neurophysiological tests may support diagnosis, but none are sufficiently robust to serve as the primary outcome measure in clinical trials [94]. Although traditionally considered a motor disease, ALS is increasingly recognised to involve subclinical C-fibre sensory dysfunction [95], providing a rationale for IVCM assessment of corneal nerve degeneration.

In a cross-sectional study of ALS patients and healthy controls [96] conducted to assess corneal nerve parameters and dendritic cell densities, CNFL was significantly lower, while corneal dendritic cell densities were significantly higher in patients with ALS than in controls. In addition, inferior whorl corneal length was significantly associated with disease severity and progression in ALS. Patients with bulbar involvement also showed significantly lower inferior whorl corneal length and higher inferior whorl dendritic cell density compared to patients without bulbar involvement. In a study of a similar design, there was a significant reduction in total corneal nerve length (CNL) and an increase in nerve tortuosity in ALS patients compared to healthy controls. Importantly, there was a significant correlation between CNL and the ALS functional rating scale bulbar subscore as well as the ALS severity score bulbar subscore [97], which are commonly used clinical measures to quantify the degree of bulbar dysfunction through assessments of speech and swallowing. Use CNL for total corneal nerve length and for the bulbar-score correlations. In contrast, Turhan et al. found significantly higher corneal dendritic cell density in ALS patients than healthy controls, but no significant differences in CNFD, CNFL or CNBD [95].

Overall, the three small cross-sectional studies in this review demonstrate discordant results: two reported corneal nerve loss, while one found no difference in IVCM parameters apart from dendritic cell density. No longitudinal data are available, and no study has examined whether IVCM parameters can predict functional decline. The evidence is therefore best described as conflicting and preliminary.

Table 5 shows an overview of studies on corneal nerve status associated with ALS.

Corneal nerve status in relation to migraine

Introduction and application of IVCM

Migraine is a common, chronic neurological disorder characterised by recurrent disabling attacks of headache, and continues to be one of the world’s leading causes of disability [98] and poor quality of life. The aetiology of migraine is not well understood at present, but is believed to be due to a constellation of genetic, environmental, and neurological factors. Currently, the diagnosis of migraine is largely clinical, and is based on a comprehensive history and clinical examination [99]. The lack of objective biomarkers, together with the clinical heterogeneity of migraine, poses significant challenges to the development of effective and individualised management. The cornea receives its sensory supply from the ophthalmic division of the trigeminal nerve. The close association between the trigeminal nerve and the pathogenesis of migraine underscores the potential value of corneal nerve neuropathic changes as objective biomarkers for migraine diagnosis, disease progression, and treatment response [100].

Corneal nerve alterations have been reported in migraine, with Kinard et al. finding significantly lower CNFD in patients with chronic migraine than in healthy controls [101]. This is further corroborated by a study by Guldiken et al. who also demonstrated a significant decrease in CNFD, CNBD, and CNFL in patients with migraine compared to healthy controls, regardless of variables such as number of headache days or the presence of dry eyes. There was no significant difference in IVCM parameters for various migraine durations, number of headache days over the recent three months or pain intensity [102]. The significant corneal neuropathic changes reported in the above studies are consistent with the trigeminovascular hypothesis [103] in the pathogenesis of migraines, and suggest that corneal nerve parameters merit further evaluation as candidate biomarkers in migraine.

Overall appraisal of the evidence in migraine

Two cross-sectional studies form the basis of this section. Both report reductions in corneal nerve parameters relative to controls, but neither demonstrated a consistent relationship with headache frequency, migraine duration or pain intensity. Dry eye disease, which is more prevalent among patients with migraine and independently reduces corneal nerve metrics, was not consistently accounted for. Current evidence therefore supports an association between migraine and corneal nerve alterations rather than an established biomarker role.

Corneal nerve status in relation to tauopathies

Introduction

Tauopathies are chronic, progressive neurodegenerative disorders characterised by abnormal tau deposition in the central nervous system [104]. Tau aggregates form neuropathological inclusions within neurons, causing neuronal death [105]. Clinical manifestations include movement, cognitive and behavioural disorders with substantial heterogeneity [106]. Diagnosis is confirmed by neuropathological assessment at autopsy, except in inherited autosomal dominant tauopathies or clinically defined syndromes such as progressive supranuclear palsy (PSP) [105].

Application of IVCM in tauopathies

As substantial neuronal damage often precedes overt symptoms and approximately one-third of disease-modifying therapies in development target tau aggregation or synthesis [106], early detection of tauopathies is increasingly important [107]. Peripheral tissues may provide accessible biomarkers. Skin has emerged as a minimally invasive marker of neurodegeneration because of its bidirectional interaction with the central nervous system through the skin-brain axis [108]. Vacchi et al. demonstrated tau immunoreactivity along somatosensory and autonomic cutaneous nerve fibres in patients with degenerative tauopathies [109]. Given the similar embryological origins of skin and corneal fibres, corneal nerve degeneration has also been investigated in tauopathies.

Experimental studies have demonstrated corneal nerve abnormalities in tauopathy models. In an animal study comparing mouse models of tauopathy (P301L) and wild type [110], CNFL and CNFD were significantly decreased compared with wild-type mice, while corneal regeneration after injury was significantly impaired [110]. In another murine model, mice overexpressing human non-mutated tau showed significantly decreased stromal corneal nerve occupancy at 4 months, with no change in sub-basal plexus nerve occupancy compared with wild-type mice. Htau mice experienced corneal nerve degeneration before any memory deficit or learning impairment was apparent [111].

Overall appraisal of evidence in tauopathies

Evidence supporting IVCM as a biomarker for tauopathies remains predominantly pre-clinical. Experimental studies demonstrate corneal nerve degeneration and impaired regeneration, with some abnormalities preceding neurological deficits, suggesting potential sensitivity to early tau-associated neurodegeneration. However, human evidence remains extremely limited. Although corneal nerve abnormalities have been reported in small cohorts including patients with PSP [64], there is insufficient evidence that these changes are specific to tau pathology or correlate directly with tau burden. Longitudinal human studies incorporating tau-PET, CSF or plasma phosphorylated tau are needed to determine whether IVCM can distinguish tauopathies from other neurodegenerative disorders and detect pre-clinical disease.

Corneal nerve status in relation to chronic inflammatory demyelinating polyradiculoneuropathy (CIDP)

Chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) is a chronic, progressive immune-mediated neuropathy affecting peripheral nerves and nerve roots. The disease course is generally slow and progressive, but approximately one-third of patients with CIDP may experience a relapsing-remitting course [112]. CIDP can manifest as different clinical forms. The classical presentation is characterised by both proximal and distal symmetrical weakness with sensory loss predominantly affecting proprioception and vibration sense [113].

Diagnosis and treatment of CIDP remain challenging because reliable disease biomarkers are lacking, limiting patient stratification and individualised treatment. Although disease activity can be assessed using established severity scores, objective measures remain limited [114]. Early axonal loss has been associated with clinical outcomes and shown to predict long-term disability and muscle strength [115, 116]. Given this association, corneal nerve assessment may provide an objective measure of axonal loss in CIDP (Figure 3).

FIGURE 3

In a prospective cohort of 17 patients, IVCM was used to assess corneal nerve degeneration and immune cell infiltration over 18 months [117]. Baseline total cell infiltrates >30/mm² demonstrated 100% sensitivity and 100% specificity for predicting clinical progression over 18 months. However, total cell infiltrates did not change subsequently in either progressive or stable CIDP, raising uncertainty about their ability to monitor treatment response. Corneal parameters also did not correlate with disease duration or severity.

Among the studies included in this review, the evidence in CIDP consists of a single prospective cohort of 17 patients. A limited sample size renders the reported sensitivity and specificity of 100% for predicting clinical progression statistically fragile. The absence of correlations between corneal parameters and disease duration or severity is difficult to reconcile with a straightforward biomarker interpretation. Replication in substantially larger cohorts is required before clinical inference can be drawn.

Comparison of the evidence base across diseases

The quantity and quality of evidence supporting corneal nerve status as a biomarker vary substantially across the conditions reviewed. Amongst all conditions reviewed, DPN has the broadest and most independently replicated evidence base. Multiple independent groups have reported concordant findings in large cohorts. Prospective longitudinal follow-up extends up to 6 years and predictive thresholds for incident neuropathy have been proposed and tested. MS is also supported by two prospective cohorts with 2-year follow-up, which demonstrate that corneal nerve loss parallels disability progression. However, the longitudinal evidence base remains relatively small and includes overlapping investigators, limiting the extent of fully independent replication. In PD, the core cross-sectional observation of corneal nerve loss is well replicated, but evaluation of clinically important applications such as prognostication and phenotyping still largely rests on single-centre studies. Evidence in dementia and MCI mostly originates from one research group, with only one independent Alzheimer’s disease cohort, although findings are internally consistent. CIDP, migraine, ALS, schizophrenia and tauopathies are each supported by one to three small studies, with conflicting results in ALS and predominantly pre-clinical data in tauopathies; the evidence in these conditions should be regarded as exploratory.

Table 6 summarises the evidence according to the number and design of included publications, the largest study sample, availability and duration of longitudinal follow-up, and evidence of independent replication or multicentre validation.

Comparison between IVCM and existing biomarkers across diseases

A related and largely unanswered question is whether IVCM parameters provide incremental value over existing biomarkers. Direct comparisons are available in only a small number of contexts. In DPN, CNFD had a higher AUC than intraepidermal nerve fibre density obtained by skin biopsy within the same cohort (0.81 versus 0.73) [25], and IVCM detects abnormality earlier than nerve conduction studies [16]. In MCI, IVCM parameters showed higher diagnostic AUCs than medial temporal lobe atrophy scoring and brain volumetry within the studied cohorts, whereas performance in dementia was broadly comparable; one MCI cohort also reported higher apparent prognostic accuracy for progression to dementia than brain volumetry [71–73]. In schizophrenia, the single available study reported a higher AUC for CNFL than for cingulate gyrus volume on magnetic resonance imaging within the same cohort, although neither measure correlated with symptom severity [88]. In MS, corneal nerve measures and retinal nerve fibre layer thickness have been compared using ROC analyses in the same cohort [83], but incremental clinical value over established imaging remains unproven; and in PD no comparison has been made with dopamine transporter imaging or with skin biopsy for phosphorylated α-synuclein. In CIDP, a single 17-patient cohort reported 100% sensitivity and 100% specificity for baseline corneal immune-cell infiltration greater than 30/mm², compared with 80% sensitivity and 88% specificity for nerve ultrasound, but this has not been externally validated [117]. Demonstrating incremental value over existing modalities, rather than simply a difference from healthy controls, is the principal evidential gap facing the field.

Table 7 shows an overview of direct comparisons between IVCM and existing biomarkers across diseases.

Limitations and future directions

Limitations of the present review

This article is a narrative rather than a systematic review. Studies were identified through a single database and were selected on the basis of relevance rather than by duplicate independent screening; no formal risk-of-bias instrument was applied; and the appraisals of evidence strength presented at the end of each disease section represent the qualitative judgement of the authors. Publication bias is likely to favour positive findings, particularly in the smaller literatures such as schizophrenia, migraine and CIDP where few studies exist. In addition, several disease areas are dominated by a small number of research groups, so that apparent consistency of findings may in part reflect shared methodology rather than biological reproducibility.

A further limitation lies in the current evidence base for tauopathies, where current data are predominantly derived from animal models, with very limited human evidence. Although pre-clinical models provide valuable insight into the mechanisms of tau-related neurodegeneration, they may not fully reflect the complexity of human disease. Human data presently available are limited, and the clinical relevance of corneal nerve findings in tauopathies therefore largely remains unestablished. This limitation should be borne in mind when interpreting the section on tauopathies.

Ocular and systemic confounders of corneal nerve measurement

Ageing, ocular surface inflammation, dry eye disease, DM without neurological involvement, contact lens wear and prior ocular surgery may confound corneal nerve measurements. Ageing is associated with reduced corneal nerve fibre density and length [118], while ocular surface inflammation and dry eye disease alter corneal nerve morphology. Neuropathic corneal pain, which may be confused with dry eye disease, is associated with corneal microneuromas that could be mistaken for disease-related neurodegeneration [119].

DM may alter corneal nerve parameters before neuropathy is clinically apparent. In 37 individuals with impaired glucose tolerance compared with 20 healthy controls, CNFD, CNBD and CNFL were all significantly reduced, and 40.5% had CNFD below a neuropathy-consistent threshold despite no clinical diabetes or peripheral neuropathy [120]. Diabetic corneal neuropathy also correlates with diabetic peripheral neuropathy severity [121] and independently predicts diabetic foot risk, with corneal changes preceding clinical foot complications [45, 122]. Keratorefractive surgery [123, 124], cataract surgery [125] and orthokeratology contact lens wear may also induce corneal denervation and alter the normal architecture of the sub-basal nerve plexus [126].

Methodological heterogeneity between studies

Methodological heterogeneity across studies may affect the comparability and reproducibility of corneal nerve measurements. Different confocal microscopy platforms, such as the Nidek ConfoScan 4 and the Tandem Scanning confocal microscope, differ in resolution, field of view, and image acquisition characteristics, and have been shown to yield subbasal nerve density estimates that differ by two-to three-fold for the same tissue [127]. Laser scanning platforms, such as the Heidelberg Retina Tomograph II, provide high-resolution imaging of the corneal subbasal nerve plexus and may detect smaller-calibre nerve fibres [128]. Acquisition protocols also contribute to variability between studies. The number of images sampled per cornea affects measurement accuracy, with as many as eight non-overlapping central images required to bring nerve branch density estimates within 30% of the true mean [129], while the precise location of images sampled within the central cornea can independently produce significant differences in nerve parameter values, reflecting the heterogeneous, whorl-centred distribution of the subbasal nerve plexus [130]. In addition, manual and automated (or semi-automated) analysis software packages, such as CCMetrics and ACCMetrics, do not produce equivalent results. Automated analysis has been shown to systematically underestimate corneal nerve fibre length relative to manual tracing, despite preserving good reproducibility [131], and automated analysis has shown a non-significant tendency to underestimate corneal nerve measurements relative to manual tracing following refractive surgery-induced corneal reinnervation, attributed to thin, short regenerating fibres being more readily missed by automated software [132].

Future directions

Future research should prioritise the development of consensus-based, standardised protocols for image acquisition, sampling, parameter definition, and analysis. Moreover, studies should also report the confocal platform, image acquisition protocol, and analysis software used, adopt a fixed and consistent image-selection strategy, and, where possible, validate findings using more than one analysis method, so that results can be compared reliably across the field. Larger prospective multicentre cohorts with longitudinal follow-up should subsequently apply these standardised methods to establish robust normative reference ranges, validate diagnostic and prognostic thresholds, and assess the value of IVCM parameters over existing biomarkers. Further refinement of IVCM image acquisition may additionally improve the precision and representativeness of corneal nerve measurements. Integration of validated automated or artificial intelligence-based analysis may facilitate reproducible assessment across centres. As corneal nerve imaging becomes more widely applied as a biomarker for neurodegenerative and neurological disease, future studies should screen for and document ocular and systemic factors, such as ageing, ocular surface inflammation, dry eye disease, DM without neurological involvement, contact lens wear, and prior ocular surgery and either exclude affected individuals or statistically adjust for these variables, so that observed corneal nerve changes can be attributed to the disease under investigation.

Collectively, these developments will be critical for establishing corneal nerve status not only as a diagnostic adjunct, but also as a reliable and generalisable adjunctive biomarker in future studies of neurodegenerative and neurological diseases.

Conclusion

IVCM has emerged as a promising non-invasive approach for evaluating corneal nerve alterations across a broad spectrum of neurodegenerative and neurological diseases. Based on the available literature, reductions in corneal nerve fibre density, length, and branching have been associated with disease presence, severity, progression, and, in selected conditions, future clinical deterioration. However, the strength of this evidence is highly disease-dependent. The most robust support currently exists in diabetic peripheral neuropathy and multiple sclerosis, where findings have been reproduced across independent cohorts and supported by longitudinal data. In Parkinson’s disease and dementia, corneal nerve alterations are consistently reported and show potential diagnostic and prognostic value, although several clinically important applications remain dependent on small, single-centre cohorts and require external validation. Evidence in ALS, migraine, schizophrenia, CIDP, and tauopathies remains comparatively limited, conflicting, or predominantly pre-clinical and should therefore be considered exploratory.

Although IVCM shows promise as a non-invasive biomarker, future large-scale, multicentre longitudinal studies using standardised imaging and analysis protocols, with appropriate adjustment for confounders are needed. Advances in wide-field imaging, automation, and artificial intelligence may further improve reproducibility and scalability [133], supporting the potential integration of IVCM into the diagnosis and longitudinal monitoring of neurological diseases.

References

• 1.

  SridharMS. Anatomy of cornea and ocular surface. Indian J Ophthalmol (2018) 66(2):190–4. 10.4103/ijo.ijo_646_17

  • CrossRef
  • Google Scholar
  • View reference in article

• 2.

  Guerrero-MorenoABaudouinCMelik ParsadaniantzSRéaux-Le GoazigoA. Morphological and functional changes of corneal nerves and their contribution to peripheral and central sensory abnormalities. Front Cell Neurosci (2020) 14:610342. 10.3389/fncel.2020.610342

  • CrossRef
  • Google Scholar
  • View reference in article

• 3.

  JolivaltCGHanMMNguyenADesmondFAlves JesusCHVasconselosDCet alUsing corneal confocal microscopy to identify therapeutic agents for diabetic neuropathy. J Clin Med (2022) 11(9):2307. 10.3390/jcm11092307

  • CrossRef
  • Google Scholar
  • View reference in article

• 4.

  DabbahMAGrahamJPetropoulosITavakoliMMalikRA. Dual-model automatic detection of nerve-fibres in corneal confocal microscopy images. Med Image Comput Comput Assist Interv (2010) 13(Pt 1):300–7. 10.1007/978-3-642-15705-9_37

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 5.

  Pop-BusuiRBoultonAJMFeldmanELBrilVFreemanRMalikRAet alDiabetic neuropathy: a position statement by the American diabetes association. Diabetes Care (2017) 40(1):136–54. 10.2337/dc16-2042

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 6.

  YuY. Gold standard for diagnosis of DPN. Front Endocrinol (2021) 12:719356. 10.3389/fendo.2021.719356

  • CrossRef
  • Google Scholar
  • View reference in article

• 7.

  SoWZQi WongNSTanHCYu LinMTYu LeeIXMehtaJSet alDiabetic corneal neuropathy as a surrogate marker for diabetic peripheral neuropathy. Neural Regen Res (2022) 17(10):2172–8. 10.4103/1673-5374.327364

  • CrossRef
  • Google Scholar
  • View reference in article

• 8.

  ZhaoHHeYRenYRChenBH. Corneal alteration and pathogenesis in diabetes mellitus. Int J Ophthalmol (2019) 12(12):1939–50. 10.18240/ijo.2019.12.17

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 9.

  HafnerJKarstSSacuSScholdaCPablikESchmidt‐ErfurthU. Correlation between corneal and retinal neurodegenerative changes and their association with microvascular perfusion in type II diabetes. Acta Ophthalmol. (2019) 97(4):e545–e550. 10.1111/aos.13938

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 10.

  HanSBYangHKHyonJY. Influence of diabetes mellitus on anterior segment of the eye. Clin Interv Aging (2018) 14:53–63. 10.2147/CIA.S190713

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 11.

  LiuYCLinMTMehtaJS. Analysis of corneal nerve plexus in corneal confocal microscopy images. Neural Regen Res (2021) 16(4):690–1. 10.4103/1673-5374.289435

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 12.

  PetropoulosINPonirakisGKhanAAlmuhannadiHGadHMalikRA. Diagnosing diabetic neuropathy: something old, something new. Diabetes Metab J (2018) 42(4):255–69. 10.4093/dmj.2018.0056

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 13.

  PerkinsBALovblomLELewisEJHBrilVFerdousiMOrszagAet alCorneal confocal microscopy predicts the development of diabetic neuropathy: a longitudinal diagnostic multinational consortium study. Diabetes Care (2021) 44(9):2107–14. 10.2337/dc21-0476

  • CrossRef
  • Google Scholar
  • View reference in article

• 14.

  CarmichaelJFadaviHTavakoliM. Neurodegeneration of the cornea and retina in patients with type 1 diabetes without clinical evidence of diabetic retinopathy. Front Endocrinol (2022) 13:790255. 10.3389/fendo.2022.790255

  • CrossRef
  • Google Scholar
  • View reference in article

• 15.

  CoppeyLDavidsonEShevalyeHObrosovATorresMYorekMA. Progressive loss of corneal nerve fibers and sensitivity in rats modeling obesity and type 2 diabetes is reversible with omega-3 fatty acid intervention: supporting cornea analyses as a marker for peripheral neuropathy and treatment. Diabetes Metab Syndr Obes (2020) 13:1367–84. 10.2147/DMSO.S247571

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 16.

  PetropoulosINFerdousiMMarshallAAlamUPonirakisGAzmiSet alThe inferior whorl for detecting diabetic peripheral neuropathy using corneal confocal microscopy. Invest Ophthalmol Vis Sci (2015) 56(4):2498–504. 10.1167/iovs.14-15919

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 17.

  WangMZhangCZuoALiLChenLHouX. Diagnostic utility of corneal confocal microscopy in type 2 diabetic peripheral neuropathy. J Diabetes Investig (2021) 12(4):574–82. 10.1111/jdi.13381

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 18.

  LiQZhongYZhangTZhangRZhangQZhengHet alQuantitative analysis of corneal nerve fibres in type 2 diabetics with and without diabetic peripheral neuropathy: comparison of manual and automated assessments. Diabetes Res Clin Pract (2019) 151:33–8. 10.1016/j.diabres.2019.03.039

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 19.

  JiangMSYuanYGuZXZhuangSL. Corneal confocal microscopy for assessment of diabetic peripheral neuropathy: a meta-analysis. Br J Ophthalmol (2016) 100(1):9–14. 10.1136/bjophthalmol-2014-306038

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 20.

  HertzPBrilVOrszagAAhmedANgENwePet alReproducibility of in vivo corneal confocal microscopy as a novel screening test for early diabetic sensorimotor polyneuropathy. Diabet Med (2011) 28(10):1253–60. 10.1111/j.1464-5491.2011.03299.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 21.

  PritchardNEdwardsKDehghaniCFadaviHJeziorskaMMarshallAet alLongitudinal assessment of neuropathy in type 1 diabetes using novel ophthalmic markers (LANDMark): study design and baseline characteristics. Diabetes Res Clin Pract (2014) 104(2):248–56. 10.1016/j.diabres.2014.02.011

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 22.

  BurgessJFrankBMarshallAKhalilRSPonirakisGPetropoulosINet alEarly detection of diabetic peripheral neuropathy: a focus on small nerve fibres. Diagnostics (2021) 11(2):165. 10.3390/diagnostics11020165

  • CrossRef
  • Google Scholar
  • View reference in article

• 23.

  AhmedABrilVOrszagAPaulsonJYeungENgoMet alDetection of diabetic sensorimotor polyneuropathy by corneal confocal microscopy in type 1 diabetes: a concurrent validity study. Diabetes Care (2012) 35(4):821–8. 10.2337/dc11-1396

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 24.

  FerdousiMKaltenieceAAzmiSPetropoulosINPonirakisGAlamUet alDiagnosis of neuropathy and risk factors for corneal nerve loss in type 1 and type 2 diabetes: a corneal confocal microscopy study. Diabetes Care (2021) 44(1):150–6. 10.2337/dc20-1482

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 25.

  AlamUJeziorskaMPetropoulosINAsgharOFadaviHPonirakisGet alDiagnostic utility of corneal confocal microscopy and intra-epidermal nerve fibre density in diabetic neuropathy. PLoS One (2017) 12(7):e0180175. 10.1371/journal.pone.0180175

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 26.

  LovblomLEHalpernEMWuTKellyDAhmedABouletGet alIn vivo corneal confocal microscopy and prediction of future-incident neuropathy in type 1 diabetes: a preliminary longitudinal analysis. Can J Diabetes (2015) 39(5):390–7. 10.1016/j.jcjd.2015.02.006

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 27.

  PonirakisGAl-JanahiIElgassimEHomssiMPetropoulosINGadHet alSustained corneal nerve loss predicts the development of diabetic neuropathy in type 2 diabetes. Front Neurosci (2024) 18:1393105. 10.3389/fnins.2024.1393105

  • CrossRef
  • Google Scholar
  • View reference in article

• 28.

  TavakoliMQuattriniCAbbottCKallinikosPMarshallAFinniganJet alCorneal confocal microscopy: a novel noninvasive test to diagnose and stratify the severity of human diabetic neuropathy. Diabetes Care (2010) 33(8):1792–7. 10.2337/dc10-0253

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 29.

  PritchardNEdwardsKVagenasDRussellAWMalikRAEfronN. Corneal sensitivity is related to established measures of diabetic peripheral neuropathy. Clin Exp Optom (2012) 95(3):355–61. 10.1111/j.1444-0938.2012.00729.x

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 30.

  QuattriniCTavakoliMJeziorskaMKallinikosPTesfayeSFinniganJet alSurrogate markers of small fiber damage in human diabetic neuropathy. Diabetes (2007) 56(8):2148–54. 10.2337/db07-0285

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 31.

  PetropoulosINAlamUFadaviHMarshallAAsgharODabbahMAet alRapid automated diagnosis of diabetic peripheral neuropathy with in vivo corneal confocal microscopy. Invest Ophthalmol Vis Sci (2014) 55(4):2071–8. 10.1167/iovs.13-13787

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 32.

  TummanapalliSSIssarTKwaiNPoyntenAKrishnanAVWillcoxMet alAssociation of corneal nerve loss with markers of axonal ion channel dysfunction in type 1 diabetes. Clin Neurophysiol (2020) 131(1):145–54. 10.1016/j.clinph.2019.09.029

  • CrossRef
  • Google Scholar
  • View reference in article

• 33.

  KaltenieceAFerdousiMAzmiSKhanSUWorthingtonAMarshallAet alCorneal nerve loss is related to the severity of painful diabetic neuropathy. Eur J Neurol (2022) 29(1):286–94. 10.1111/ene.15129

  • CrossRef
  • Google Scholar
  • View reference in article

• 34.

  QuattriniCTavakoliMKallinikosPMarshallAEfronNBoultonAet alComparing skin biopsy with corneal confocal microscopy: diagnostic yield of nerve fiber density [Conference Abstract]. Diabetologia (2010) 53(Suppl. 1):S444. 10.1007/s00125-010-1872-z

  • CrossRef
  • Google Scholar
  • View reference in article

• 35.

  HerlynAPrakasamRKPeschelSAllgeierSKöhlerBWinterKet alCorneal subbasal nerve plexus changes in severe diabetic charcot foot deformity: a pilot study in search for a DNOAP biomarker. J Diabetes Res (2018) 2018:5910639. 10.1155/2018/5910639

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 36.

  KaltenieceAFerdousiMAzmiSMubitaWMMarshallALauriaGet alCorneal confocal microscopy detects small nerve fibre damage in patients with painful diabetic neuropathy. Sci Rep (2020) 10(1):3371. 10.1038/s41598-020-60422-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 37.

  WilliamsBMBorroniDLiuRZhaoYZhangJLimJet alAn artificial intelligence-based deep learning algorithm for the diagnosis of diabetic neuropathy using corneal confocal microscopy: a development and validation study. Diabetologia (2020) 63(2):419–30. 10.1007/s00125-019-05023-4

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 38.

  ScarpaFColonnaARuggeriA. Multiple-image deep learning analysis for neuropathy detection in corneal nerve images. Cornea (2020) 39(3):342–7. 10.1097/ICO.0000000000002181

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 39.

  MengYPrestonFGFerdousiMAzmiSPetropoulosINKayeSet alArtificial intelligence based analysis of corneal confocal microscopy images for diagnosing peripheral neuropathy: a binary classification model. J Clin Med (2023) 12(4):1284. 10.3390/jcm12041284

  • CrossRef
  • Google Scholar

• 40.

  PrestonFGMengYBurgessJFerdousiMAzmiSPetropoulosINet alArtificial intelligence utilising corneal confocal microscopy for the diagnosis of peripheral neuropathy in diabetes mellitus and prediabetes. Diabetologia (2022) 65(3):457–66. 10.1007/s00125-021-05617-x

  • CrossRef
  • Google Scholar

• 41.

  QiaoQCaoJXueWQianJWangCPanQet alDeep learning-based automated tool for diagnosing diabetic peripheral neuropathy. Digit Health (2024) 10:20552076241307573. 10.1177/20552076241307573

  • CrossRef
  • Google Scholar

• 42.

  ChenWLiaoDDengYHuJ. Development of a transformer-based deep learning algorithm for diabetic peripheral neuropathy classification using corneal confocal microscopy images. Front Cell Dev Biol (2024) 12:1484329. 10.3389/fcell.2024.1484329

  • CrossRef
  • Google Scholar

• 43.

  Ben RabahCPetropoulosINMalikRASeragA. Vision transformers for automated detection of diabetic peripheral neuropathy in corneal confocal microscopy images. Front Imaging (2025) 4:1542128. 10.3389/fimag.2025.1542128

  • CrossRef
  • Google Scholar

• 44.

  SalahouddinTPetropoulosINFerdousiMPonirakisGAsgharOAlamUet alArtificial intelligence-based classification of diabetic peripheral neuropathy from corneal confocal microscopy images. Diabetes Care (2021) 44(7):e151–e153. 10.2337/dc20-2012

  • CrossRef
  • Google Scholar

• 45.

  LiuCZhuLO’DonnellGPKohMHYuMLeeIXYet alPrediction of the risk of diabetic foot from corneal nerve images using deep learning algorithms. Diabetes Res Clin Pract (2025) 230:112991. 10.1016/j.diabres.2025.112991

  • CrossRef
  • Google Scholar

• 46.

  AzmiSFerdousiMAlamUPetropoulosINPonirakisGMarshallAet alSmall-fibre neuropathy in men with type 1 diabetes and erectile dysfunction: a cross-sectional study. Diabetologia (2017) 60(6):1094–101. 10.1007/s00125-017-4245-z

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 47.

  De ClerckEEBSchoutenJSAGBerendschotTTJMKoolschijnRSNuijtsRMMASchramMTet alReduced corneal nerve fibre length in prediabetes and type 2 diabetes: the maastricht study. Acta Ophthalmol (2020) 98(5):485–91. 10.1111/aos.14359

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 48.

  VidyasagarPFarrellSFColoradoLHDandoSEdwardsK. Corneal nerve morphology in painful diabetic neuropathy: a meta-analysis of in vivo confocal microscopy studies. Biomedicines (2025) 13(7):1675. 10.3390/biomedicines13071675

  • CrossRef
  • Google Scholar

• 49.

  BadianRAEkmanLPrippAHUtheimTPEnglundEDahlinLBet alComparison of novel wide-field in vivo corneal confocal microscopy with skin biopsy for assessing peripheral neuropathy in type 2 diabetes. Diabetes (2023) 72(7):908–17. 10.2337/db22-0863

  • CrossRef
  • Google Scholar

• 50.

  TibarHEl BayadKBouhoucheAAit Ben HaddouEHBenomarAYahyaouiMet alNon-motor symptoms of Parkinson’s disease and their impact on quality of life in a cohort of Moroccan patients. Front Neurol (2018) 9:170. 10.3389/fneur.2018.00170

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 51.

  SchapiraAHVChaudhuriKRJennerP. Non-motor features of parkinson disease. Nat Rev Neurosci (2017) 18(7):435–50. 10.1038/nrn.2017.62

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 52.

  Rodríguez-LeyvaICalderón-GarcidueñasALJiménez-CapdevilleMERentería-PalomoAAHernandez-RodriguezHGValdés-RodríguezRet alα-Synuclein inclusions in the skin of parkinson's disease and parkinsonism. Ann Clin Transl Neurol (2014) 1(7):471–8. 10.1002/acn3.78

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 53.

  WakabayashiK. Where and how alpha-synuclein pathology spreads in parkinson's disease. Neuropathology (2020) 40(5):415–25. 10.1111/neup.12691

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 54.

  PodgornyPJSuchowerskyORomanchukKGFeasbyTE. Evidence for small fiber neuropathy in early Parkinson’s disease. Parkinsonism Relat Disord (2016) 28:94–9. 10.1016/j.parkreldis.2016.04.033

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 55.

  YaisawangS. Corneal Nerve Parameters in Prodromal Parkinson's disease Patients (2023). 10.58837/CHULA.THE.2023.671

  • CrossRef
  • Google Scholar

• 56.

  CheN-NYangH-Q. Potential use of corneal confocal microscopy in the diagnosis of Parkinson’s disease associated neuropathy. Transl Neurodegener (2020) 9(1):28. 10.1186/s40035-020-00204-3

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 57.

  CheN-NJiangQ-HDingG-XChenS-YZhaoZ-XLiXet alCorneal nerve fiber loss relates to cognitive impairment in patients with Parkinson’s disease. Npj Parkinsons Dis (2021) 7(1):80. 10.1038/s41531-021-00225-3

  • CrossRef
  • Google Scholar

• 58.

  MisraSLKerstenHMRoxburghRHDanesh-MeyerHVMcGheeCN. Corneal nerve microstructure in Parkinson’s disease. J Clin Neurosci (2017) 39:53–8. 10.1016/j.jocn.2017.02.033

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 59.

  CheN-NChenSJiangQ-HChenS-YZhaoZ-XLiXet alCorneal confocal microscopy differentiates patients with Parkinson’s disease with and without autonomic involvement. Npj Parkinsons Dis (2022) 8(1):114. 10.1038/s41531-022-00387-8

  • CrossRef
  • Google Scholar

• 60.

  Kass-IliyyaLJavedSGosalDKobyleckiCMarshallAPetropoulosINet alSmall fiber neuropathy in Parkinson’s disease: a clinical, pathological and corneal confocal microscopy study. Parkinsonism Relat Disord (2015) 21(12):1454–60. 10.1016/j.parkreldis.2015.10.019

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 61.

  LimSHFerdousiMKaltenieceAMahfoudZRPetropoulosINMalikRAet alCorneal confocal microscopy identifies Parkinson’s disease with more rapid motor progression. Mov Disord (2021) 36(8):1927–34. 10.1002/mds.28602

  • CrossRef
  • Google Scholar

• 62.

  CheN-NJiangQ-HChenSChenS-YZhaoZ-XLiXet alThe severity of corneal nerve loss differentiates motor subtypes in patients with Parkinson’s disease. Ther Adv Neurol Disord (2023) 16:17562864231165561. 10.1177/17562864231165561

  • CrossRef
  • Google Scholar

• 63.

  NiuXYinPGuanCShaoQCuiGZanKet alCorneal confocal microscopy May help to distinguish multiple system atrophy from Parkinson’s disease. Npj Parkinsons Dis (2024) 10(1):63. 10.1038/s41531-024-00680-8

  • CrossRef
  • Google Scholar

• 64.

  LimSHFerdousiMBhattacharjeeSKaltenieceAMahfoudZRPetropoulosINet alCorneal confocal microscopy demonstrates varying degrees of neurodegeneration in atypical parkinsonian disorders. Parkinsonism Relat Disord (2023) 117:105899. 10.1016/j.parkreldis.2023.105899

  • CrossRef
  • Google Scholar

• 65.

  AnandSSchooC. Mild cognitive impairment. In: Statpearls. Treasure Island, FL: StatPearls Publishing (2025). Available online at: http://www.ncbi.nlm.nih.gov/books/NBK599514/ (Accessed March 20, 2025).

  • Google Scholar

• 66.

  PrinceMBryceRFerriC. World Alzheimer Report 2011: The Benefits of Early Diagnosis and Intervention. London, United Kngdom: Alzheimer’s Disease International (2011). Available online at: https://www.alzint.org/u/WorldAlzheimerReport2011.pdf (Accessed March 25, 2025).

  • Google Scholar

• 67.

  RobinsonLTangETaylorJ-P. Dementia: timely diagnosis and early intervention. BMJ (2015) 350:h3029. 10.1136/bmj.h3029

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 68.

  AlbertMSDeKoskySTDicksonDDuboisBFeldmanHHFoxNCet alThe diagnosis of mild cognitive impairment due to Alzheimer’s disease: recommendations from the national institute on aging-alzheimer’s association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement (2011) 7(3):270–9. 10.1016/j.jalz.2011.03.008

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 69.

  KorfESCScheltensPBarkhofFDe LeeuwF-E. Blood pressure, white matter lesions and medial temporal lobe atrophy: closing the gap between vascular pathology and Alzheimer’s disease?Dement Geriatr Cogn Disord (2005) 20(6):331–7. 10.1159/000088464

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 70.

  Van De PolLAScheltensP. Medial temporal lobe atrophy scores translated to clinical practice: editorial comment on ‘influence of age, disease onset and ApoE4 on visual medial temporal lobe atrophy cut‐offs. J Intern Med (2014) 275(3):331–3. 10.1111/joim.12176

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 71.

  Al-JanahiEPonirakisGAl HamadHVattothSElsotouhyAPetropoulosINet alCorneal nerve and brain imaging in mild cognitive impairment and dementia. J Alzheimers Dis (2020) 77(4):1533–43. 10.3233/jad-200678

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 72.

  PonirakisGHamadHAKhanAPetropoulosINGadHChandranMet alLoss of corneal nerves and brain volume in mild cognitive impairment and dementia. Alzheimers Dement (N Y) (2022) 8(1):e12269. 10.1002/trc2.12269

  • CrossRef
  • Google Scholar

• 73.

  PonirakisGAlHHOmarDAMPetropoulosINKhanAGadHet alCorneal nerve loss predicts dementia in patients with mild cognitive impairment. Ann Clin Transl Neurol (2023) 10(4):599–609. 10.1002/acn3.51747

  • CrossRef
  • Google Scholar

• 74.

  PonirakisGAl HamadHSankaranarayananAKhanAChandranMRamadanMet alAssociation of corneal nerve fiber measures with cognitive function in dementia. Ann Clin Transl Neurol (2019) 6(4):689–97. 10.1002/acn3.746

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 75.

  GundoganAOOltuluRBelviranliSTezcanAAdamMMirzaEet alCorneal innervation changes ın Alzheimer’s: implications for sensory dysfunction. Int Ophthalmol (2024) 44(1):270. 10.1007/s10792-024-03162-1

  • CrossRef
  • Google Scholar

• 76.

  CompstonAColesA. Multiple sclerosis. Lancet (2008) 372(9648):1502–17. 10.1016/s0140-6736(08)61620-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 77.

  BjartmarCWujekJRTrappBD. Axonal loss in the pathology of MS: consequences for understanding the progressive phase of the disease. J Neurol Sci (2003) 206(2):165–71. 10.1016/s0022-510x(02)00069-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 78.

  LiDKBLiMJTraboulseeAZhaoGRiddehoughAPatyD. The use of MRI as an outcome measure in clinical trials. Adv Neurol (2006) 98:203–26. Available online at: https://pubmed.ncbi.nlm.nih.gov/16400836/.

  • Pubmed Abstract
  • Google Scholar

• 79.

  SiepmanTAMWefers Bettink-RemeijerMHintzenRQ. Retinal nerve layer thickness in subgroups of multiple sclerosis, measured by optical coherence tomography and scanning laser polarimetry. J Neurol (2010) 257(10):1654–60. 10.1007/s00415-010-5589-1

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 80.

  MikolajczakJZimmermannHKheirkhahAKadasEMOberwahrenbrockTMullerRet alPatients with multiple sclerosis demonstrate reduced subbasal corneal nerve fibre density. Mult Scler (2017) 23(14):1847–53. 10.1177/1352458516677590

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 81.

  BitirgenGAkpinarZMalikRAOzkagniciA. Use of corneal confocal microscopy to detect corneal nerve loss and increased dendritic cells in patients with multiple sclerosis. JAMA Ophthalmol (2017) 135(7):777–82. 10.1001/jamaophthalmol.2017.1590

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 82.

  PetropoulosINFitzgeraldKCOakleyJPonirakisGKhanAGadHet alCorneal confocal microscopy demonstrates axonal loss in different courses of multiple sclerosis. Sci Rep (2021) 11(1):21688. 10.1038/s41598-021-01226-1

  • CrossRef
  • Google Scholar

• 83.

  PetropoulosINAl-ShibaniFBitirgenGPonirakisGKhanAGadHet alCorneal axonal loss as an imaging biomarker of neurodegeneration in multiple sclerosis: a longitudinal study. Ther Adv Neurol Disord (2023) 16:17562864221118731. 10.1177/17562864221118731

  • CrossRef
  • Google Scholar

• 84.

  BitirgenGAkpinarZUcaAUOzkagniciAPetropoulosINMalikRA. Progressive loss of corneal and retinal nerve fibers in patients with multiple sclerosis: a 2-year follow-up study. Transl Vis Sci Technol (2020) 9(13):37. 10.1167/tvst.9.13.37

  • CrossRef
  • Google Scholar

• 85.

  PatelKRCherianJGohilKAtkinsonD. Schizophrenia: overview and treatment options. P T (2014) 39(9):638–45. Available online at: https://pubmed.ncbi.nlm.nih.gov/25210417/.

  • Pubmed Abstract
  • Google Scholar

• 86.

  RahmanTLaurielloJ. Schizophrenia: an overview. Focus (Am Psychiatr Publ) (2016) 14(3):300–7. 10.1176/appi.focus.20160006

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 87.

  SilversteinSMChoiJJGreenKMBowles-JohnsonKERamchandranRS. Schizophrenia in translation: why the eye?Schizophr Bull (2022) 48(4):728–37. 10.1093/schbul/sbac050

  • CrossRef
  • Google Scholar

• 88.

  PonirakisGGhandiRAhmedAGadHPetropoulosINKhanAet alAbnormal corneal nerve morphology and brain volume in patients with schizophrenia. Sci Rep (2022) 12(1):1870. 10.1038/s41598-022-05609-w

  • CrossRef
  • Google Scholar

• 89.

  JørgensenKNNesvågRNerlandSMørch‐JohnsenLWestlyeLTLangeEHet alBrain volume change in first‐episode psychosis: an effect of antipsychotic medication independent of BMI change. Acta Psychiatr Scand (2017) 135(2):117–26. 10.1111/acps.12677

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 90.

  AndreasenNCNopoulosPMagnottaVPiersonRZiebellSHoB-C. Progressive brain change in schizophrenia: a prospective longitudinal study of first-episode schizophrenia. Biol Psychiatry (2011) 70(7):672–9. 10.1016/j.biopsych.2011.05.017

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 91.

  Fusar‐PoliPMeyer‐LindenbergA. Forty years of structural imaging in psychosis: promises and truth. Acta Psychiatr Scand (2016) 134(3):207–24. 10.1111/acps.12619

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 92.

  MasroriPVan DammeP. Amyotrophic lateral sclerosis: a clinical review. Eur J Neurol (2020) 27(10):1918–29. 10.1111/ene.14393

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 93.

  BrotmanRGMoreno-EscobarMCJosephJMunakomiSPawarG. Amyotrophic lateral sclerosis. In: Statpearls. Treasure Island, FL: StatPearls Publishing (2025). Available online at: http://www.ncbi.nlm.nih.gov/books/NBK556151/ (Accessed March 20, 2025).

  • Google Scholar

• 94.

  McMackinRBedePIngreCMalaspinaAHardimanO. Biomarkers in amyotrophic lateral sclerosis: current status and future prospects. Nat Rev Neurol (2023) 19(12):754–68. 10.1038/s41582-023-00891-2

  • CrossRef
  • Google Scholar

• 95.

  TurhanSAKarlssonPOzunYGunesHSurucuSTokerEet alIdentification of corneal and intra‐epidermal axonal swellings in amyotrophic lateral sclerosis. Muscle Nerve (2024) 69(1):78–86. 10.1002/mus.27995

  • CrossRef
  • Google Scholar

• 96.

  FuJHeJZhangYLiuZWangHLiJet alSmall fiber neuropathy for assessment of disease severity in amyotrophic lateral sclerosis: corneal confocal microscopy findings. Orphanet J Rare Dis (2022) 17(1):7. 10.1186/s13023-021-02157-w

  • CrossRef
  • Google Scholar

• 97.

  FerrariGGrisanEScarpaFFazioRComolaMQuattriniAet alCorneal confocal microscopy reveals trigeminal small sensory fiber neuropathy in amyotrophic lateral sclerosis. Front Aging Neurosci (2014) 6:278. 10.3389/fnagi.2014.00278

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 98.

  Aguilar-SheaALMembrillaMJADiaz-de-TeranJ. Migraine review for general practice. Aten Primaria (2022) 54(2):102208. 10.1016/j.aprim.2021.102208

  • CrossRef
  • Google Scholar

• 99.

  Pescador RuschelMADe JesusO. Migraine headache. In: Statpearls. Treasure Island, FL: StatPearls Publishing (2025). Available online at: http://www.ncbi.nlm.nih.gov/books/NBK560787/ (Accessed March 20, 2025).

  • Google Scholar

• 100.

  Md IsaNATummanapalliSSChiangJCBKrishnanAVZagamiASPapasEBet alNeuropathic changes in corneal nerve Endings—a potential objective biomarker for migraine frequency and response to treatment. Headache (2023) 63(9):1327–9. 10.1111/head.14630

  • CrossRef
  • Google Scholar

• 101.

  KinardKISmithAGSingletonJRLessardMKKatzBJWarnerJEAet alChronic migraine is associated with reduced corneal nerve fiber density and symptoms of dry eye. Headache (2015) 55(4):543–9. 10.1111/head.12547

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 102.

  GuldikenYCPetropoulosINMalikAMalikRAYükselRBudakFet alCorneal confocal microscopy identifies corneal nerve fiber loss in patients with migraine. Cephalalgia (2023) 43(5):3331024231170810. 10.1177/03331024231170810

  • CrossRef
  • Google Scholar

• 103.

  AshinaMHansenJMDoTPMelo-CarrilloABursteinRMoskowitzMA. Migraine and the trigeminovascular system-40 years and counting. Lancet Neurol (2019) 18(8):795–804. 10.1016/S1474-4422(19)30185-1

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 104.

  KovacsGG. Tauopathies. Handb Clin Neurol (2017) 145:355–68. 10.1016/B978-0-12-802395-2.00025-0

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 105.

  CreekmoreBCWatanabeRLeeEB. Neurodegenerative disease tauopathies. Annu Rev Pathol (2024) 19:345–70. 10.1146/annurev-pathmechdis-051222-120750

  • CrossRef
  • Google Scholar

• 106.

  CummingsJLGonzalezMIPritchardMCMayPCToledo-ShermanLMHarrisGA. The therapeutic landscape of tauopathies: challenges and prospects. Alzheimers Res Ther (2023) 15(1):168. 10.1186/s13195-023-01321-7

  • CrossRef
  • Google Scholar

• 107.

  IrwinDJ. Tauopathies as clinicopathological entities. Parkinsonism Relat Disord (2016) 22(Suppl. 1):S29–33. 10.1016/j.parkreldis.2015.09.020

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 108.

  LozziFCameraECardinaliGDi NardoA. Exploring the skin as an open window onto neurodegenerative diseases. Transl Neurodegener (2025) 14(1):61. 10.1186/s40035-025-00522-4

  • CrossRef
  • Google Scholar

• 109.

  VacchiELazzariniEPintonSChiaroGDisantoGMarchiFet alTau protein quantification in skin biopsies differentiates tauopathies from alpha-synucleinopathies. Brain (2022) 145(8):2755–68. 10.1093/brain/awac161

  • CrossRef
  • Google Scholar

• 110.

  LiSShiSLuoBXiaFHaYMerkleyKHet alTauopathy induces degeneration and impairs regeneration of sensory nerves in the cornea. Exp Eye Res (2022) 215:108900. 10.1016/j.exer.2021.108900

  • CrossRef
  • Google Scholar

• 111.

  MarquezAGuernseyLSFrizziKECundiffMConstantinoIMuttalibNet alTau associated peripheral and central neurodegeneration: identification of an early imaging marker for tauopathy. Neurobiol Dis (2021) 151:105273. 10.1016/j.nbd.2021.105273

  • CrossRef
  • Google Scholar

• 112.

  RizzutoNMorbinMCavallaroTFerrariSFallahiMGaliazzo RizzutoS. Focal lesions are a feature of chronic inflammatory demyelinating polyneuropathy (CIDP). Acta Neuropathol (1998) 96(6):603–9. 10.1007/s004010050941

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 113.

  VallatJ-MSommerCMagyL. Chronic inflammatory demyelinating polyradiculoneuropathy: diagnostic and therapeutic challenges for a treatable condition. Lancet Neurol (2010) 9(4):402–12. 10.1016/S1474-4422(10)70041-7

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 114.

  KhooAFrascaJSchultzD. Measuring disease activity and predicting response to intravenous immunoglobulin in chronic inflammatory demyelinating polyneuropathy. Biomark Res (2019) 7:3. 10.1186/s40364-019-0154-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 115.

  Al-ZuhairyAJakobsenJKrarupC. Early axonal loss predicts long-term disability in chronic inflammatory demyelinating polyneuropathy. Clin Neurophysiol (2021) 132(4):1000–7. 10.1016/j.clinph.2020.12.017

  • CrossRef
  • Google Scholar

• 116.

  Al-ZuhairyAJakobsenJMoldovanMKrarupC. Axonal loss at time of diagnosis as biomarker for long-term disability in chronic inflammatory demyelinating polyneuropathy. Muscle Nerve (2022) 66(6):715–22. 10.1002/mus.27722

  • CrossRef
  • Google Scholar

• 117.

  PitarokoiliKSturmDLabediAGreinerTEitnerLKumowskiNet alNeuroimaging markers of clinical progression in chronic inflammatory demyelinating polyradiculoneuropathy. Ther Adv Neurol Disord (2019) 12:1756286419855485. 10.1177/1756286419855485

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 118.

  ChinJYLiuCLeeIXYLinMTYChengCYWongJHFet alImpact of age on the characteristics of corneal nerves and corneal epithelial cells in healthy adults. Cornea (2024) 43(4):409–18. 10.1097/ICO.0000000000003363

  • CrossRef
  • Google Scholar

• 119.

  ChengJLiuCYuMLeeIXYWangXHsuVWTet alExploration of imaging and molecular biomarkers for differentiation of neuropathic corneal pain from dry eye syndrome. Ocul Surf (2025) 38:230–41. 10.1016/j.jtos.2025.08.002

  • CrossRef
  • Google Scholar

• 120.

  AsgharOPetropoulosINAlamUJonesWJeziorskaMMarshallAet alCorneal confocal microscopy detects neuropathy in subjects with impaired glucose tolerance. Diabetes Care (2014) 37(9):2643–6. 10.2337/dc14-0279

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 121.

  MansoorHTanHCLinMTYMehtaJSLiuYC. Diabetic corneal neuropathy. J Clin Med (2020) 9(12):3956. 10.3390/jcm9123956

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 122.

  LiuCYuMLeeIXYO’DonnellGPTanHCTakahashi-NakajimaAet alDiabetic corneal neuropathy as a window into diabetic foot risk: a large-scale association study. Diabetes Obes Metab (2026) 28(9):8219–29. 10.1111/dom.71021

  • CrossRef
  • Google Scholar

• 123.

  StewartSLiuY-CLinMT-YMehtaJS. Clinical applications of in vivo confocal microscopy in keratorefractive surgery. J Refract Surg (2021) 37(7):493–503. 10.3928/1081597X-20210419-01

  • CrossRef
  • Google Scholar

• 124.

  ChinJYLinMTYLeeIXYMehtaJSLiuYC. Tear neuromediator and corneal denervation following SMILE. J Refract Surg (2021) 37(8):516–23. 10.3928/1081597X-20210423-01

  • CrossRef
  • Google Scholar

• 125.

  KimJHChungJLKangSYKimSWSeoKY. Change in corneal sensitivity and corneal nerve after cataract surgery. Cornea (2009) 28(Suppl. 1):S20–S25. 10.1097/ICO.0b013e3181aea0e3

  • CrossRef
  • Google Scholar

• 126.

  LumEGolebiowskiBSwarbrickHA. Mapping the corneal sub-basal nerve plexus in orthokeratology lens wear using in vivo laser scanning confocal microscopy. Invest Ophthalmol Vis Sci (2012) 53(4):1803–9. 10.1167/iovs.11-8706

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 127.

  ErieEAMcLarenJWKittlesonKMPatelSVErieJCBourneWM. Corneal subbasal nerve density: a comparison of two confocal microscopes. Eye Contact Lens (2008) 34(6):322–5. 10.1097/ICL.0b013e31818b74f4

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 128.

  PatelDVMcGheeCNJ. Mapping of the normal human corneal sub-basal nerve plexus by in vivo laser scanning confocal microscopy. Invest Ophthalmol Vis Sci (2005) 46(12):4485–8. 10.1167/iovs.05-0794

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 129.

  VagenasDPritchardNEdwardsKShahidiAMSampsonGPRussellAWet alOptimal image sample size for corneal nerve morphometry. Optom Vis Sci (2012) 89(5):812–7. 10.1097/OPX.0b013e31824ee8c9

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 130.

  WinterKScheibePKöhlerBAllgeierSGuthoffRFStachsO. Local variability of parameters for characterization of the corneal subbasal nerve plexus. Curr Eye Res (2016) 41(2):186–98. 10.3109/02713683.2015.1010686

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 131.

  OstrovskiILovblomLEFarooqiMAScarrDBouletGHertzPet alReproducibility of in vivo corneal confocal microscopy using an automated analysis program for detection of diabetic sensorimotor polyneuropathy. PLoS One (2015) 10(11):e0142309. 10.1371/journal.pone.0142309

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 132.

  ChinJYYangLWYJiAJSNubileMMastropasquaLAllenJCet alValidation of the use of automated and manual quantitative analysis of corneal nerve plexus following refractive surgery. Diagnostics (Basel) (2020) 10(7):493. 10.3390/diagnostics10070493

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 133.

  Barcelo-CantonRHYuMLiuCTakahashiALeeIXYLiuYC. Applications of artificial intelligence in corneal nerve images in ophthalmology. Diagnostics (Basel) (2026) 16(4):602. 10.3390/diagnostics16040602

  • CrossRef
  • Google Scholar