icobrain aria: a breakthrough in Alzheimer’s disease management and care

Today, 32 million people are suffering from Alzheimer’s disease globally, and it is reported that more than 400 million people are on the Alzheimer’s disease continuum1. As these numbers are expected to double every two decades, increasingly impacting families, healthcare systems, and the wider society, there is a strong call to action to put brain health on the global agenda2,3. 

During the last decades, many pharmaceutical companies have been investing heavily in the development of disease-modifying therapies (DMTs) for Alzheimer’s disease4. The first amyloid beta targeting monoclonal antibody therapies have demonstrated a meaningful impact on the deposition of amyloid and the clinical course of the disease5,6. However, clinical trials have shown that these therapies can be associated with amyloid-related imaging abnormalities (ARIA). Although findings may on some occasions be mild and asymptomatic, ARIA can be symptomatic, and is on rare occasions associated with life-threatening events. Hence, ARIA safety monitoring through regular brain Magnetic Resonance Imaging (MRI) scans will play an important role in the follow-up of patients treated with amyloid beta-directed monoclonal antibodies.

The need for computer-aided diagnosis solutions to support ARIA assessment

The need for regular ARIA monitoring was emphasized by the FDA, recommending at least 3 follow-up brain MRI scans in addition to the baseline pre-treatment MRI in the label of both Aduhelm and Leqembi7,8. Given the large number of peoplƒ√e who could be eligible for the new DMTs, these ARIA monitoring brain MRI requests will impact the radiology workload significantly. Furthermore, ARIA detection and severity assessment is a challenging and time-consuming task. Indeed, ARIA (especially at the earliest stage of ‘mild’ radiological severity) can be very subtle and involves a detailed analysis and detection/quantification of cerebral edema (ARIA-E), microhemorrhages, and superficial siderosis (ARIA-H) on brain MRI FLAIR and T2* scans. 

Some clear cases of ARIA presentations are shown in the Figure below9,10.

The term ARIA was introduced in 2010 to encompass the spectrum of MRI findings observed in patients receiving investigational anti–amyloid therapies for AD. Given this specific link to these therapies, real world experience and education is very limited. In this context, it was reported that up to 84% of ARIA cases can be missed by local radiologists11.  This was also highlighted in the recent review paper on ARIA by Hampel et al. (2023)12, suggesting the need for a quantitative computer-aided diagnosis tool that can support radiological ARIA assessment. 

Introducing icobrain aria 

During the last two years, the icometrix team has been applying their state-of-the-art AI expertise in lesion detection, their unique experience in longitudinal brain MRI follow-up assessment, and their regulatory skills to develop icobrain aria (FDA pending). icobrain aria is developed based on large data sets of individuals on amyloid-targeting therapies, including ARIA cases and ARIA-free cases. The AI-based brain MRI solution allows for the automated detection and quantification of both ARIA-E (edema/sulcal effusion) and ARIA-H (hemorrhage/superficial siderosis).  Using a unique longitudinal neural network, the baseline MRI is compared with the post-dosing follow-up MRI scan, assisting radiologists with an extra pair of eyes (or AIs), allowing for robust, consistent, and standardized safety monitoring. Similar to the existing portfolio of icobrain solutions, color-coded image annotations and a quantitative summary report is seamlessly integrated into the hospital workflow and contains the following information:

• For ARIA-E, the number of sites of involvement as well as their localization and the longest axis of the largest site of involvement are measured and summarized as an ARIA-E severity according to the guidelines. 

• For ARIA-H, severity levels for microhemorrhages and superficial siderosis are calculated separately based on their count for which localization is also further specified. An example of the icobrain aria reports for ARIA-E and ARIA-H is provided below.

  
  

icobrain aria: rigorously examined for its clinical validity

This week, at the Clinical Trials in Alzheimer’s Disease (CTAD) conference in Boston, we are presenting the first results of a first-of-its-kind reader study that evaluates the clinical validity of icobrain aria. A multi-case, multi-reader (MRMC) study was performed, including 16 (ARIA-reading trained) radiologists and 398 MRI scans, which were read with and without the assistance of icobrain aria. Results were compared with a ground truth, which is defined by the consensus reading of three leading global ARIA experts. 

A significant improvement in radiological ARIA-E and ARIA-H detection and severity assessment was shown when icobrain aria was used, as summarized in the Table below13. The significant impact of the software on the radiological reading is further confirmed by the high Cohen’s d effect sizes of 2.0 for ARIA-H and 2.5 for ARIA-E. These results strongly indicate that icobrain aria significantly improves safety monitoring and management of AD patients treated with amyloid beta-directed monoclonal antibodies. 

We are incredibly excited to discuss the icobrain aria results at the CTAD2023 conference this week. The introduction of disease-modifying therapies is opening a new era for the management and care of Alzheimer’s disease, and solutions such as icobrain aria will be crucial to optimize benefits and risks in each individual. We can’t wait to integrate this solution, together with the broader icobrain portfolio, in hospital systems worldwide. Very soon, we will announce several further solutions that will impact the care of Alzheimer’s, stay tuned!

National Key Account Director, Rebecca Bartz, proudly next to our abstract poster L109

References

1.  Anders Gustavsson, et al., Global estimates on the number of persons across the Alzheimer's disease continuum. 2023, doi: 10.1002/alz.12694. 

2.    Emma Nichols, et al., Global, regional, and national burden of Alzheimer's disease and other dementias, 1990–2016: a systematic analysis for the Global Burden of Disease Study. 2016, https://doi.org/10.1016/S1474-4422(18)30403-4 

3.  Harris A. Eyre, Frédéric Destrebecq, Stephanie Kramer, Call to Action: Putting Brain Health on the Global Agenda. 2023

4.  Jeffrey Cumming, et al., Alzheimer's disease drug development pipeline: 2023 

5.   Christopher H. van Dyck et al. Lecanemab in Early Alzheimer’s Disease. 2023

6.   John R. Sims et al. Donanemab in Early Symptomatic Alzheimer Disease. 2023

7.   Leqembi US Prescribing Information 

8.  Aduhelm US Prescribing Information  

9.  Frederik Barkhof et al.  An MRI Rating Scale for Amyloid-Related Imaging Abnormalities with Edema or Effusion. 2013

10.   Petrice M. Cogswell et al. Amyloid-Related Imaging Abnormalities with Emerging Alzheimer Disease Therapeutics: Detection and Reporting Recommendations for Clinical Practice. 2022

11.   Samantha Gutt, et al., Variability of aria detection in patients receiving monoclonal antibodies against amyloid-β plaques. 2019

12.    Harald Hampel, et al., Amyloid-related imaging abnormalities (ARIA): radiological, biological and clinical characteristics. 2023 

13.   Diana Sima et al., Development and clinical validation of icobrain aria – an AI-based assistive software tool for automated detection and quantification of amyloidrelated imaging abnormalities. 2023, CAT Program