Preparing for the Future:
Why We Started With the Hippocampus

When we began building Alzevita in 2023, our mission was simple but ambitious: to replace manual hippocampal tracing with an automated, reproducible tool that radiologists, neurologists, and researchers could trust.

Anyone who has ever attempted manual tracing knows the challenges. It’s time-consuming, inconsistent, and heavily dependent on who performs the work. What should be an objective measurement often varies between experts, limiting its value in clinical practice and large-scale studies. We believed there had to be a better way.

Why the Hippocampus First?

The hippocampus represents the perfect intersection of clinical importance, technical feasibility, and strategic focus.

• Clinically: It is one of the earliest and most consistently affected regions in conditions that impact memory and cognition, making it highly relevant for neurologists and researchers.
• Technically: It is a well-defined brain structure, which makes it a strong candidate for automated segmentation using modern deep learning approaches.
• Strategically:Starting with a single, critical region allowed us to deliver meaningful value quickly, while establishing a foundation for the future.

We see the hippocampus not as the end goal, but as the first step in a broader journey — a journey that will expand to additional brain regions, whole-brain volumetry, longitudinal tracking, and eventually, multimodal integration across imaging and clinical data.

The Development Journey

Like any meaningful project, our path was not without challenges.

• Data quality: Collecting and curating high-quality MRI data with reliable expert annotations was our first major hurdle.
• Standardization: We learned that even experts sometimes disagree, which taught us the importance of building consensus and defining a “ground truth” for training.
• Reliability: Ensuring the algorithm worked consistently across different imaging conditions was another critical milestone.
• Usability: Translating the core algorithm into a cloud-based, user-friendly platform was essential to making the technology usable in real-world workflows.

Behind the Technology

At the core, our system uses a deep learning–based 3D segmentation model, designed on modern neural network architectures to identify and quantify the hippocampus with precision.

We validate our accuracy by comparing automated results with expert manual segmentations, using multiple methods to ensure both shape alignment and volume consistency. This multi-metric approach makes our results clinically meaningful, not just statistically strong.

For interoperability, Alzevita supports standard imaging formats such as DICOM and NIfTI, and delivers results as both visual overlays and structured quantitative reports through a secure, cloud-based web interface.

The User Perspective

From the beginning, we built this for the people who use it: neurologists, radiologists, researchers, and imaging labs.

Early testers told us they valued:

• Efficiency – dramatically faster than manual segmentation
• Consistency – standardized results across sites and raters
• Scalability – enabling studies and follow-ups at a scale that manual tracing simply can’t achieve.

These voices guided us to refine Alzevita not just as a powerful algorithm, but as a practical, trustworthy tool.

Looking Ahead

The hippocampus is only the beginning. Our roadmap includes expanding to additional brain regions, enabling whole-brain volumetry, supporting longitudinal tracking across timepoints, and eventually, integrating multimodal data.

At every step, one principle remains the same: trust is built on validation. That’s why we’ve invested heavily in rigorous testing and transparency from day one.

Conclusion

We started with the hippocampus because it matters — to clinicians, to researchers, and to patients. By focusing on this small but critical structure, we’ve created a replicable process for building validated, transparent, and scalable AI tools.

This is only the first step, but it’s a meaningful one. And it sets the stage for a future where quantitative brain MRI becomes routine, reliable, and transformative.

Let’s Talk

We’re building the future of quantitative brain MRI. If you’re a radiologist, researcher, or diagnostic lab interested in exploring our software, get in touch with us.