By Associate Biomarker Scientist Hiba Kazmi

A summary of interesting talks which we attended on Day 2 of the AAIC. In particular, there was a focus on incorporating imaging with a range of other biomarkers (for example blood plasma and retinal) and also translating AD biomarkers for clinical application. 

We saw an outline of where we are along the strategic roadmap to translate AD biomarkers to clinical application particularly for Tau-PET - Work from Alessandra Dodich (Univ Trento) and Valentina Garibotto (Univ Geneva) who showed flortaucipir to have good analytical validity and is currently undergoing further testing for clinical validity/utility. 

Interesting novel methods were reported in the neuroimaging session “Understanding Tau Progression” in this context by Joseph Giorgio (University of Cambridge) to predict future individualized rate of tau accumulation by means of the scalar projection method with Neil Oxtoby (University College London) presenting the different spatiotemporal imaging phenotypes of tau related to the AD subdivisions highlighting the idea of tailored treatment. 

Moving onto AI, there was a really interesting session on explainable AI, which was all about stripping back the black box of deep learning and understanding the why of a neural network's decision. This is crucial for building trust in machine learning as these tools are increasingly being turned towards diagnostics. One great talk presented by James Cole at UCL focussed on building a model that glimpsed at all regions in the brain, then located areas of interest that are used to classify a subject as healthy or AD. Crucially the network shows the user where is important. Such information is really useful to the clinician.

Another interesting session was on neuroinflammation and microglia activation in AD. In particular, Maura Malpetti from the University of Cambridge discussed the prognostic value of in vivo [11C]PK11195 PET of microglial activation in Alzheimer’s disease. They examined the prognostic value of TSPO PET in combination with Tau imaging and structural MRI and tested the hypothesis that inflammation accelerates the progression of AD and concluded that in vivo PET for microglia activation and tau are better predictor of cognitive decline in AD with respect to atrophy on structural MRI.  In particular in the anterior-medial temporal region microglial activation seems to be an important factor to predict cognitive decline over time.  

We also attended sessions discussing Sex and ethnoracial differences in AD. Work by Yi-Ting Wang (McGill university research centre for studies in aging) highlighted the need to consider ApoE and sex when targeting tau accumulation and clinical progression. Further talks on precision medicine included the development of a connectivity based tau spreading model by Nicolai Franzmeier and improving on more conventional approaches.

Furthermore, the main takeaway from a very interesting plenary talk discussing imaging Biomarkers and Alzheimer's Disease Prevention by Susan M. Landau was that treatment is about targeting specific groups at specific timepoints rather than a one size fits all. 

In addition, the importance of the use of longitudinal imaging to identify the optimal treatment window to target subjects at AB- before becoming AB+ (for subs on the amyloid pathway) was also highlighted.

The Day 2 talks have demonstrated the need for integrating a range of biomarkers but also taking into account the various subtypes of patients within a disease and tailoring treatment for patients.

---