60/100

Moderate evidence

Cite
More
EmailLinkedInX

Mayo Clinic study suggests AI-ECG score can flag vascular risk in 297 migraine patients

Study title: Vascular risk prediction in migraine: A multimodality risk score based on artificial intelligence-electrocardiogram output, echocardiogram, and detailed migraine characteristics

Evidence type
Observational study
Sample
n = 297
Sources
1 sourcePeer-reviewed
Published
2 min read

Study title: Vascular risk prediction in migraine: A multimodality risk score based on artificial intelligence-electrocardiogram output, echocardiogram, and detailed migraine characteristics

Researchers at Mayo Clinic built a risk score for vascular events in people with migraine, combining AI-enabled electrocardiogram outputs, echocardiogram measurements and detailed migraine characteristics. In a test group of 297 patients, those scoring 15 or above had a hazard ratio of 3.83 for the composite of stroke, heart attack, venous thromboembolism or death.

Read the study abstract

AimPatients with migraine, especially those with aura, have a higher risk of stroke and other vascular events. However, the underlying mechanisms are complex, multifactorial, and may require integration of multimodal data for comprehensive characterization. We have conducted prior analyses on how artificial intelligence (AI)-enabled electrocardiogram (ECG), specifically the atrial fibrillation prediction model and age estimation model output, echocardiogram parameters, and detailed migraine characteristics separ…

Vascular risk prediction in migraine: A multimodality risk score based on artificial intelligence-electrocardiogram output, echocardiogram, and detailed migraine characteristics
"A hospital emergency department entrance in the United States" by Harrison Keely, CC BY 4.0.

The gain

In a Mayo Clinic retrospective study of 297 migraine patients, those scoring 15 or higher on a risk score combining AI-ECG output, echocardiogram and migraine features had a hazard ratio of 3.83 for stroke, heart attack, clot or death versus lower scorers.

In brief

Migraine, especially with aura, is linked to higher stroke and vascular risk, but the mechanisms are complex. The study asked how much each data type adds when used together. Each modality improved reclassification: 17.7% for AI-ECG output, 16.9% for echocardiogram and 22.1% for migraine characteristics.

The authors call it a pilot evaluation. The score was tested on only 297 patients, and they say external validation and real-world clinical assessment are still needed.

Main points

  1. The score used 20 parameters: demographics, comorbidities, AI-ECG outputs, echocardiogram features and migraine characteristics.
  2. Weights came from prior Cox regression analyses on 29,928 patients (AI-ECG), 12,778 (echocardiogram) and 6,819 (migraine characteristics).
  3. The 297-patient test group was 83.2% female, 36.7% had aura, mean age 39.0 years, mean follow-up 68.9 months.
  4. Thirty-five patients (11.8%) developed the composite endpoint of stroke, heart attack, venous thromboembolism or death.
  5. In the combined model, comorbidity (HR 1.17) and migraine characteristics (HR 1.33) were significant predictors; other categories were not.
  6. Net reclassification improvement was 17.7% for AI-ECG, 16.9% for echocardiogram and 22.1% for migraine characteristics.

The full story

The study was a retrospective cohort analysis. Eligible patients had at least one 12-lead ECG, a transthoracic echocardiogram and a visit with a headache specialist at any Mayo Clinic site within three years. The AI-ECG inputs were an atrial fibrillation prediction model and ECG-estimated age minus chronological age.

Each variable received points based on hazard ratios and beta coefficients from earlier Cox regression work. The authors used Contal and O'Quigley's method to find the best cut-off, which was a score of 15, and ran a multivariate Cox regression to compare risk categories.

Patients at or above the cut-off had significantly shorter event-free survival. Detailed migraine features carried the largest reclassification gain at 22.1%, which the authors say shows the value of collecting them for vascular risk stratification.

Sources

1 on file
  1. Cephalalgiadoi:10.1177/03331024261505811
    Peer-reviewed2026-11-17

Talk it over

Three questions to start on, two more if you have time
Patients scoring 15 or higher had a hazard ratio of 3.83. How would you weigh that against the fact that the score was tested on only 297 patients?

Start with a question

Mayo Clinic study suggests AI-ECG score can flag vascular risk in 297 migraine patientshttps://truvace.com/item/vascular-risk-prediction-in-migraine-a-multimodality-risk-score-based-on-artific

  1. Patients scoring 15 or higher had a hazard ratio of 3.83. How would you weigh that against the fact that the score was tested on only 297 patients?

    Answer in the thread
  2. The 297 patients were 83.2% female with a mean age of 39.0 years, and all had seen a headache specialist at a Mayo Clinic site. Who is missing, and how could that change the score's results?

    Answer in the thread
  3. Migraine characteristics gave the largest reclassification gain at 22.1%. If you were a headache specialist, what would you do differently when collecting those details?

    Answer in the thread
If you have more time
  • Weights came from earlier Cox regression work on 29,928, 12,778 and 6,819 patients. What would you ask about how those earlier groups compare with the 297?

    Answer in the thread
  • Thirty-five patients (11.8%) had stroke, heart attack, venous thromboembolism or death. What evidence would show a high score leads to better care?

    Answer in the thread

What readers say