AI Electrocardiography Predicts Atrial Fibrillation in CLL

Patients with chronic lymphocytic leukemia (CLL) may have an increased risk of atrial fibrillation (AF); among the risk factors are history of AF, heart failure, hypertension, valvular heart disease, older age, male sex, and treatment with Bruton’s tyrosine kinase inhibitors (BTKis), like ibrutinib. A study assessed how artificial intelligence electrocardiography (AI-ECG) may help predict ibrutinib-induced AF in patients with CLL. As a reference, they also assessed patients who were not treated with ibrutinib. The results of the study were presented at the 62nd ASH Annual Meeting & Exposition.

The Mayo Clinic CLL Database was queried to collect data that allowed researchers to create two cohorts of patients: those who were evaluated within one year of CLL diagnosis who never received ibrutinib, and those who did receive ibrutinib. The AI-ECG algorithm that was utilized was developed prior with a convolutional neural network. ECGs collected within 10 years before the diagnosis of CLL (cohort 1) or 10 years before initiating ibrutinib therapy (cohort 2) were used to establish the baseline AI-ECG AF score. Patients were excluded fi they had AF at baseline, missing data, or ECGs previously used to train the AI algorithm.

A total of 2,739 patients were screened, of whom 1,149 were included in the analysis. Patients had a median of four baseline ECGs.

Cohort 1 consisted of 951 patients with a median follow-up of three years; 546 patients (57%) had a positive baseline AI-ECG. The median age was 67 years (interquartile range [IQR], 58-72 years). Most patients were male (n=681; 72%) and had low/intermediate risk CLL-International Prognostic Index (IPI; 68%); about one-third (32%) had high/very high-risk CLL-IPI.

Cohort 2 consisted of 198 patients with a median follow-up of 1.6 years; 91 patients (46%) had a positive baseline AI-ECG. The median age was 69 years (IQR, 62-75 years). Most patients were male (n=139; 70%) and had high/very high-risk CLL-IPI (87%). Only 13% had low/intermediate risk CLL-IPI.

During follow-up, 164 patients in cohort 1 (17%) and 46 patients in cohort 2 (23%) sustained AF. Positive baseline AI-ECG was a significant risk factor for AF in both groups. In cohort 1, the hazard ratio (HR) for positive versus negative AI-ECG was 33.9 (95% confidence interval [CI], 15.0-76.6); in cohort 2, the HR was 14.8 (95% CI, 5.3-41.3).

“The addition of AI to a standard 12-lead ECG obtained during normal sinus rhythm—an inexpensive and ubiquitous test—predicts the occurrence of future AF in patients with CLL. This holds true irrespective of BTKi-based therapy and has important implications for the management of CLL patients,” the study authors concluded.