Mode
Text Size
Log in / Sign up

Machine learning model detects cardiac amyloidosis on bone scintigraphy with high accuracyAI tool spots heart disease risk before symptoms start

AI-generated summary of the cited source, checked by automated accuracy review. How we work

Key Takeaway
Consider the Amylo-Detect model as a potential screening tool for cardiac amyloidosis, but await prospective validation.

This development and validation study evaluated the Amylo-Detect machine learning model for predicting cardiac amyloidosis (CA)-suggestive uptake (Perugini grade >=2) on bone scintigraphy. The study included 11,616 consecutive all-comer patients referred for bone scintigraphy at Vienna General Hospital (2010-2023) and University Hospital Essen. The model was compared with existing scoring systems and clinical routine.

High-grade uptake prevalence was 3.0% (388 of 11,616 patients). Model performance showed an AUC of 0.93 in the development cohort, 0.91 in internal validation, and 0.91 in external validation. Notably, Amylo-Detect identified 12 of 42 patients (29%) with CA-suggestive uptake who were missed in clinical routine.

Safety and tolerability were not reported. Limitations were not reported in the input. The model is available as a web app for further evaluation.

Clinicians should interpret these results cautiously, as the study is observational and external validation is limited. The model may help reduce missed diagnoses, but prospective clinical utility studies are needed.

Hidden heart disease in your records

Imagine walking into a doctor's office with a vague feeling of fatigue. You might feel tired or short of breath. These are common signs of many conditions. Doctors often miss the real cause until it is too late.

Cardiac amyloidosis is a rare heart condition. It happens when bad proteins build up in the heart muscle. This makes the heart stiff and weak. It can lead to heart failure if not treated.

Finding this disease early is very hard. The symptoms look like other common heart problems. Doctors usually wait for clear signs before ordering special tests. This delay can be dangerous for patients.

Why waiting for symptoms is dangerous

Waiting for symptoms to get worse is a risky strategy. By the time a patient feels sick, the heart may already be damaged. Early treatment can save lives and improve quality of life.

Current methods rely on doctors guessing who needs a test. They look at age and other risk factors. This approach misses many people who are actually at risk.

Many patients go undiagnosed for years. They suffer from symptoms that could have been managed better. The goal is to find these patients before they get sick.

The AI that spots the risk

A new tool called Amylo-Detect changes how doctors find this disease. It uses artificial intelligence to scan electronic health records. The system looks at fifty different data points from a patient file.

Think of it like a traffic jam detector. The AI watches for patterns that suggest a blockage is forming. It spots the risk before the traffic actually stops.

This tool looks at routine data like blood tests and past diagnoses. It does not need new scans or expensive equipment. It works with information doctors already have on file.

This does not mean you can use this tool at home.

The system was tested on thousands of patients in Austria and Germany. It looked at over eleven thousand people in Vienna alone. It also checked data from another hospital in Essen.

The AI predicted who would have high-grade uptake on a scan. It matched the results of a specialized bone scan very well. The model worked better than standard scoring systems used by doctors.

But the tool is not a diagnosis

The AI found patients that routine checks missed. About ten percent of patients with the disease were overlooked by standard care. The tool caught nearly thirty percent of those missed cases.

This means more people can get the right treatment sooner. It helps doctors prioritize who needs the confirmatory test first. It saves time and resources for the hospital.

The tool also predicts who might have worse outcomes. It can tell which patients are at higher risk for heart failure. This helps doctors plan care for the future.

However, the AI is not a final answer. It is a guide for doctors to make decisions. Patients still need a confirmatory scan to prove the disease is there.

What happens when the app launches

The tool is available as a web app for doctors to use. It is designed to help with referrals for testing. This makes the process faster and more accurate.

More research is needed to see how it works in other places. Doctors will need to test it in different hospitals. They must ensure it works for all types of patients.

Regulatory approval is the next step for wider use. The team plans to evaluate the tool further in real-world settings. This ensures it is safe and effective for everyone.

Timely detection is crucial to improve outcomes in patients with cardiac amyloidosis. This tool promises to address diagnostic delays and improve outcomes. It offers hope for better care in the future.

Study Details

Study typeCohort
Sample sizen = 11,616
EvidenceLevel 3
PublishedApr 2026
View Original Abstract ↓
Background Timely detection is crucial to improve outcomes in patients with cardiac amyloidosis (CA) by initiation of life-saving treatments. Although confirmatory bone scintigraphy is highly accurate for CA detection, identifying at-risk patients for referral remains challenging. Objectives This study aimed to develop and validate a machine learning model, Amylo-Detect, using structured multimodal electronic health record (EHR) data to guide referrals for confirmatory scintigraphy and monoclonal protein testing. Methods Consecutive all-comer patients (n=11,616) referred for bone scintigraphy at the Vienna General Hospital (2010-2023) were retrospectively included. Patients referred before August 2020 formed the development cohort. The remaining patients comprised the internal validation cohort. External validation was performed at the University Hospital Essen (n=1,521). Amylo-Detect was trained using 50 routinely available parameters to predict CA-suggestive uptake (Perugini grade >=2) and compared with an existing score and clinical routine. Results High-grade uptake was present in 388 patients (3.0%). Amylo-Detect demonstrated excellent performance in development (AUC 0.93), independent internal validation (AUC 0.91), and external validation cohort (AUC 0.91), outperforming existing scoring systems and clinical routine. Results were consistent across subgroups, even when crucial predictors were missing. Of the 42/388 (10.8%) patients missed in clinical routine, 12/42 (29%) were additionally detected by Amylo-Detect. The model further conveyed significant prognostic value for mortality and heart failure hospitalization. Conclusions We present Amylo-Detect, a validated EHR-based tool for CA risk prediction, available as a web app, allowing application and further evaluation. By improving timely detection and referral, Amylo-Detect promises to address diagnostic delays and improve outcomes.
Free Newsletter

Clinical research that matters. Delivered to your inbox.

Join thousands of clinicians and researchers. No spam, unsubscribe anytime.