This week we will be at the ADPD 2025 International Conference talking about some of our latest biomarker research. Specifically, we are presenting an AI-based approach to differentiate different types of dementia from magnetic resonance imaging (MRI). This presentation complements our recently shown work on the combinatorial use of positron emission tomography (PET) imaging and novel blood-based biomarkers.

Such research initiatives directly translate into innovations across IXICO’s neuro imaging trial management and analysis Platform and contribute to a continually growing set of multi-modal biomarkers for Alzheimer’s Disease (AD) that can help the biopharma industry and patients drive advances in clinical trial planning.

Research differentiating types of dementia

The work we will be presenting (Presentation link) this week relates to AI-driven classification of AD and Frontotemporal Dementia (FTD) from MRI demonstrating how clinical trials can benefit from automated tools for recruitment and patient stratification.

Clinical Trials require reliable diagnosis at the earliest disease stage to improve success.​ Both conditions, AD and FTD, can present with memory impairment, behavioural changes, and executive dysfunction, leading to diagnostic ambiguity.​ While biomarkers like amyloid and tau PET scans can aid in AD diagnosis, there is no single definitive test for FTD.​ Patients with FTD may develop memory impairment later in the disease, mimicking AD.​ AD Clinical Trials may benefit from a cost-effective MRI tool for diagnosis & recruitment.

To address these challenges, IXICO have developed a framework for differentiating patients based on mixed data sources.​ In the work presented, we have applied this framework to identify cases of AD, FTD, and cognitively normal (CN) individuals for clinical trial enrolment. Our AI model for FTD/AD diagnosis based on MRI showed high performance in the test set for age-matched groups. A combination of volumetric and cortical thickness metrics with deep-learning features, showed robust differentiation between the two diseases.​​

Multi-biomarker assessment

The work presented at ADPD extends previously presented work for accurate and efficient AD phenotyping towards differential diagnosis of different dementias, providing valuable insights into emerging possibilities for efficient and reliant determination of a subject’s amyloid status, a key part of diagnostic and trial recruitment criteria.

The assessment was conducted using the Bio-Hermes study data, sponsored by the Global Alzheimer's Platform Foundation (GAP) and supported by IXICO as imaging partner.  The Bio-Hermes study https://globalalzplatform.org/biohermesstudy/ enhances our understanding of blood-based biomarkers (BBMs) like p-Tau217, as a complement to amyloid positron emission tomography (PET), a gold-standard technique for the measurement of amyloid pathology. By examining the interplay between imaging and BBMs, we highlight how BBMs offer a more accessible, less invasive, and cost-effective alternative where appropriate, acknowledging the need  for a gold-standard assessment from PET in certain cases. 

Considering the results presented at ADPD in the context of our previous findings, we have demonstrated how the active use of relevant biomarkers from PET, MRI, and blood can help to perform efficient and accurate investigation of clinical dementia symptoms towards an accurate diagnosis or reliable clinical trial inclusion criteria.

These research initiatives present further key advancements in IXICO’s portfolio of multi-modal biomarkers across AD and Parkinson’s Disease. If you would like to meet to us at AD/PD 2025 or understand more about our research progress and neuroscience expertise please contact us https://ixico.com/contact-us/

---