Atrial Fibrillation Among Patients with Hematologic Malignancies on Tyrosine Kinase Inhibitor Therapy - PROJECT SUMMARY Bruton’s tyrosine kinase inhibitors (BTKIs) have revolutionized the treatment of hematologic malignancies, significantly improving patient survival. However, as their clinical use expands, an increasing number of patients are being diagnosed with BTKI-associated arrhythmias, leading to treatment interruptions, discontinuation, bleeding, systemic embolism, and even sudden death. Atrial fibrillation (AF) is the most common arrhythmia in BTKI-treated patients, contributing to oncologic treatment interruption in up to 78% of cases, as well as increased risks of thromboembolic and bleeding complications. Clinical trials evaluating BTKI efficacy report an AF incidence of 5–16%, based solely on clinically symptomatic episodes documented during the studies. However, this likely underestimates the true burden, as AF can be asymptomatic, paroxysmal, or clinically silent in 10–27% of patients, and comprehensive monitoring has been lacking. Furthermore, the arrhythmia burden associated with BTKIs—defined as the proportion of time spent in AF—has never been prospectively evaluated. Additionally, no AF risk scores have been specifically developed for BTKI-treated patients. Existing risk models, designed for the general population or patients with hematologic malignancies, are rarely used in this setting due to their poor predictive accuracy. With the widespread and chronic use of BTKIs in a growing population of cancer patients, it is of critical concern to improve ways to identify those at risk for arrhythmias. To address this challenge, we have assembled a multidisciplinary team spanning five leading healthcare systems: Stanford University, UC San Diego, City of Hope, Wake Forest University and the Mayo Clinic, bringing together experts in cardiology, oncology, and machine learning. Aim 1 will prospectively characterize the incidence and burden of AF in patients on BTKI therapy compared to controls through continuous monitoring with smartwatches and patch monitors. Aim 2 will develop a multimodal AF prediction algorithm specifically trained on patients with hematologic malignancies on BTKI therapy, leveraging machine learning algorithms to incorporate 12-lead ECG data. Aim 3 will externally validate our model at the Mayo Clinic and test two separate AF prediction models developed at Mayo Clinic using our cohort from Aim 1. The ability to identify subsets of patients at low, intermediate, and high risk for AF will allow preferential selection of a non-BTKI-based treatment for those at high risk, more widespread use among patients with low risk, and optimal monitoring for rhythm problems and prompt intervention to decrease AF-related comorbidities among those with intermediate risk.
