Instability of Neurological Diagnosis

neuroClues TM; Paris; Sur place; EUR 200 000 +; Plein temps

Description du poste

Discover the data behind diagnostic variability and why we believe biomarkers, including oculomotor ones, could be part of the way forward.

Clinical highlights – June 2025

neuroClues has officially received itsCE-marking as a Class IIa medical device, available for clinical use in Europe.

To markBrain Awareness Month, we’re taking a closer look at a critical challenge in neurological care: thestability and accuracy of diagnosis, particularly in conditions likeParkinson’s disease and dementia.

In this month’s newsletter, we explore the data behind diagnostic variability and why we believebiomarkers, including oculomotor ones, could be part of the way forward.

1. Parkinson's Disease Remains a Clinical Challenge

Getting an accurate diagnosis of Parkinson’s Disease (PD) remains a major challenge, impacting both patient care and research outcomes.

A recent Finnish study followed1,626 people diagnosed with PD over 10 years, and the results highlight the considerableinstabilityand clinicaluncertaintyassociated with PD diagnoses [1]

Diagnosis Instability

The study revealed that13.3% of initial Parkinson’s diagnoses were later revised, mostly within the first two years. When dementia with Lewy bodies (DLB) was counted separately from PD, the rate rose to17.7%[1].

Many of the revised diagnoses turned out to bevascular parkinsonism, PSP, MSA, orclinically undetermined parkinsonism (CUPS). Other common culprits includedessential tremoranddrug-induced parkinsonism[1].

Why the revisions?

These findings reflect a broader challenge:early-stage symptoms often overlap, making it difficult to confidently diagnose PD in its early phases.

Limited role of “Gold Standard” methods

If we have “gold standards”, why do misdiagnoses persist?

Thestudy shows that“gold standard” diagnostic methods aren’t widely used in routine practice. For example,DAT imagingwas used in just36%of patients mostly in cases with atypical symptoms or diagnostic uncertainty [1].

And whilepost-mortem neuropathologyremains the definitive way to confirm PD, it was performed in only3%of deceased patients [1]. This extremely low rate limits our ability to assess the true extent of misdiagnosis, as many possible misdiagnoses may never be identified.

PD diagnosis still reliesheavily on clinical judgment,as physicians assess symptoms that often emerge gradually over time. This subjective approach leaves room for uncertainty and considerable variability between examiners. For instance, experienced neurologists may reach different diagnostic conclusions than general practitioners when evaluating the same patient.

Implications

The high rate of revised and uncertain diagnoses points to a major gap:we still lack accessible, objective and reliable tools to diagnose Parkinson’s accurately.

Even with updated clinical criteria and better imaging,diagnostic variability remains a problem. That’s why there’s growing need for robust, globally accessible diagnostic biomarker which could substantially improve diagnostic accuracy, particularly in areas with limited access to specialized neurologists or movement disorder experts.

1. Räty V, Kuusimäki T, Majuri J, Vahlberg T, Gardberg M, Noponen T, et al. Stability and Accuracy of a Diagnosis of Parkinson Disease Over 10 Years. Neurology [Internet]. 2025 May 13 [cited 2025 Jun 19];104(9):e213499. Available from: https://www.neurology.org/doi/10.1212/WNL.0000000000213499

2. Shift in Dementia Diagnosis in Routine Care

A study using German health claims data from over a decade (2004–2016) reveals thatdementia subtypes like Alzheimer’s (AD), vascular dementia (VD), and others (oD) often change over time…and more than we might expect[2].

This study considered that the ICD-10codes (codes used by physician to record diagnoses), as recorded by physicians in routine healthcare, directly represent the diagnoses.

Diagnosis Instability

The study found that a substantial number of changes occur in dementia diagnoses during the course of the disease, indicating that the initial etiological determination might not be conclusive shortly after the incident diagnosis [2].

Up to46.1% of patientshad their initial dementia diagnosis changed at least once within four years. Among those, diagnoses shifted between subtypesan average of 3 to 5 times[2].

Some Diagnoses Are Less Stable Than Others

Only65.1% of Alzheimer’s patientskept the same diagnosis from start to finish.

For VD, this dropped to53.9%, and for oD, it was73.8%.

Patients initially diagnosed with AD had thehighest number of changesand thefastest rate of change, with 30% receiving a new diagnosis within just 350 days [2].

Nature of the Diagnosis in Routine Care

While the precise clinical reasons for every diagnostic shift cannot be fully captured through health claims data alone, the study outlines several likely contributors to the observed changes in the recorded ICD-10 codes for dementia diagnoses:

• Diagnoses are often provisional, as the underlying cause of dementia may not be clear at the time of the initial diagnosis and can change as the disease progresses.
• Clinical judgment and context vary, meaning diagnoses often depend on the expertise of the individual clinician and the framework conditions under which care is delivered.
• Subtype precision has limited practical impact, since differentiating between dementia types rarely affects treatment, prognosis, or billing, reducing the incentive to refine diagnoses [2]

Implications

Together, these factors illustrate that the lack of diagnostic stability is not just a matter of clinical variability, it reflects the absence of reliable, objective tools. Without standardized, accessible biomarkers, dementia diagnosis remains subjective and inconsistent.

Dementia diagnoses can change significantly over time. Therefore,researchers relying on routine healthcare data must interpret findings with caution,especially when studying specific subtypes like Alzheimer’s or vascular dementia. Categorising patients based on a single recorded diagnosis mayoversimplify a more complex diagnostic journey.

2. Riedel O, Braitmaier M, Langner I. Stability of individual dementia diagnoses in routine care: Implications for epidemiological studies. Pharmacoepidemiol Drug Saf [Internet]. 2022 May [cited 2025 Jun 19];31(5):546–55. Available from:https://onlinelibrary.wiley.com/doi/10.1002/pds.5416

2. Need for Biomarkers

Accurately diagnosing Parkinson’s and dementia remains a challenge, especially in the early stages when symptoms overlap and current tools are limited in routine care.

Movement disorder specialists can identify subtle signs of parkinsonian syndromes, but in many regions, patients are managed by primary care physicians without access to this expertise.

Similarly, in the case of dementia, diagnoses often evolve over time and assessments may differ due to variations in clinical expertise and diagnostic frameworks.

This highlights the need fornon-invasive, objective, low-cost and accessible biomarkers that can support diagnosis in everyday clinical settings.

AtneuroClues , we see oculomotor biomarkers as a promising approach. Eye movements reflect brain function in a way that is measurable, accessible, and potentially valuable in improving diagnostic clarity.

We’re excited to see more research moving in this direction and to be part of the progress toward more accurate and accessible neurological care.

Enjoyed this article?

Join our community to receive a monthly newsletter featuring insights on the relevance of eye tracking in the clinic and exclusive company updates

Are you enthusiastic aboutNeuroClues TM and want to share it with your colleagues?

Refer them to us through this email orour LinkedIn page and we’ll be happy to have a chat.

neuroClues is powered by P3Lab SRL
VAT: BE 0750.746.841

Our development is financed with the support of the Walloon Region and the European Innovation Council