Hey readers! 👋

This week we're zooming in on a topic that's becoming impossible to ignore: the growing role of apps and artificial intelligence in celiac disease, from AI chatbots trained by dietitians to machine-learning models reading biopsy slides. Whether you're a clinician, a parent, or someone navigating the gluten-free life yourself, these tools are starting to reshape how we get diagnosed, stay safe, and manage day-to-day decisions. Let's dig in.

🤖 AI Meets Celiac: What's Real and What's Hype?

From Searching to Chatting: Is "Dr. AI" Ready for Patients with Celiac Disease? - The Canadian Celiac Association tackles the big question head-on: can general-purpose AI chatbots like ChatGPT reliably answer celiac-related health questions? - Canadian Celiac Association

This is a piece worth reading carefully. We've all seen patients (or been the patient) who type symptoms into a chatbot and get a confident-sounding but dangerously incomplete answer. General AI tools aren't trained on the nuances of cross-contact, oat safety debates, or the difference between celiac disease and non-celiac gluten sensitivity. The article is a timely reminder that "Dr. AI" still needs supervision, and that the source behind the training data matters enormously.

That said, purpose-built AI tools are a different story, and one app is making a strong case for what a specialized approach can look like.

📱 The Celiac App: AI Built by Dietitians, for Patients

The Celiac App - A subscription-based mobile tool featuring Celia, an AI assistant trained specifically by registered dietitians to answer gluten-free living questions accurately. - The Celiac Space

"Living with celiac means every bite, every label, and every menu scan matters. But when you ask a general chatbot for advice? You get half-truths, guesses, and dangerous mistakes."

What stands out here is the intentional design philosophy. The app includes a product database, dining-out guides, and even practice scripts for ordering at restaurants, but it deliberately skips a barcode-scanning feature. Why? Because the team wants users to actually learn how to read labels confidently rather than depend on a scanner that might not cover every product. The 14-day free trial makes it easy to test before committing. For clinicians, this could be a practical recommendation for newly diagnosed patients who feel overwhelmed.

🔬 AI in the Lab: Reading Biopsies with 97%+ Accuracy

Celiac Disease and Gluten Free in the News - Apr 2025 - The National Celiac Association's monthly roundup highlights a Cambridge University AI system that reads celiac biopsy images with over 97% accuracy, matching experienced pathologists. - National Celiac Association

"Over 97% accuracy, a rate comparable to experienced pathologists."

This is significant for healthcare systems where access to specialist pathologists is limited. An AI tool that can reliably interpret duodenal biopsies could help reduce diagnostic bottlenecks, particularly in rural or under-resourced settings. The same roundup also flags research on ultra-short celiac disease (USCD), a form limited to the duodenal bulb that's missed more often than we'd like. The recommendation to take at least four bulb biopsies to improve detection by 60% is a practical takeaway for gastroenterologists.

🧬 Machine Learning and Personalized Immunotherapy

TCR recognition of peptide-HLA class II complexes in autoimmunity - A review proposes that AI and machine-learning frameworks combining a patient's HLA genotype, peptide-presentation profile, and T-cell repertoire could enable personalized, HLA-stratified immunotherapies for celiac disease. - Elsevier Ltd

This is further down the research pipeline, but it's worth watching. The review draws on 38 resolved human TCR structures, most from celiac disease studies, and argues that AI could eventually help match patients to targeted therapies like peptide-HLA nanomedicines or engineered regulatory T cells. We're not there yet, but the framework is being built.

👶 Pediatric Celiac: Screening Gaps and Practical Resources

Only 1 in 10 high-risk children are being tested for celiac disease - A Beyond Celiac and Weill Cornell Medicine study found alarming screening gaps, with significant racial disparities in testing rates. - Beyond Celiac

"Testing, also called screening, was lowest in Black children, at less than 7 percent."

This is a sobering finding. When only about 10% of children who meet high-risk criteria are actually screened, we're missing diagnoses at scale. The racial disparities make it even more urgent. Could AI-driven clinical decision support tools help flag high-risk pediatric patients in electronic health records? It's exactly the kind of problem where technology could make a real difference.

For parents and practitioners looking for solid educational resources on pediatric celiac, several children's hospitals published helpful overviews this week:

📊 Research Round-Up

🍁 Advocacy Corner

Made with ❤️ by Data Drift Press

Have a question, comment, or feedback? Just hit reply - we'd love to hear from you!

Keep reading