Authored by Ann Liu, PhD
Medically Reviewed by Brian Koffman, MDCM (retired), MSEd
The Bottom Line:
Researchers used artificial intelligence to develop a risk score for predicting atrial fibrillation in treatment-naïve patients with CLL. Patients who are categorized as at high risk of atrial fibrillation should be closely monitored.
Who Performed the Research and Where Was it Presented:
Dr. Tamar Tadmor from Bnai-Zion Medical Center and colleagues presented the results at the American Society for Hematology (ASH) Annual Meeting 2024.
Background:
Bruton tyrosine kinase inhibitors (BTKi) are very effective in treating chronic lymphocytic leukemia (CLL) and small lymphocytic lymphoma (SLL). However, there have been concerns about cardiovascular side effects with their use, especially with first-generation BTKi ibrutinib, which has been linked to an increased risk of abnormal heart rhythms (atrial fibrillation) and high blood pressure. While this risk is reduced with second-generation BTKi, such as acalabrutinib and zanubrutinib, it is not completely eliminated. Additionally, CLL mainly affects older adults, who are already at higher risk for cardiac disease. For this study, researchers used artificial intelligence to identify factors associated with a high risk for atrial fibrillation.
Methods and Participants:
Data from electronic medical records were used to develop a machine learning-based risk score for predicting atrial fibrillation in treatment-naïve patients with CLL. Over 100 variables were evaluated to develop the risk score, which could contain a maximum of 10 variables.
Results:
- Nearly 4,000 patients diagnosed with CLL were in the electronic database, and 208 patients had started a BTKi during the study period.
- 125 patients started ibrutinib therapy, and 83 patients began acalabrutinib therapy.
- Well-established risk factors for atrial fibrillation were identified by the model, including advanced age (80+ years), past history of diabetes or hypertension, and being male.
- Ibrutinib use was identified as a risk factor, which was not surprising since it is well-known that ibrutinib increases the risk of atrial fibrillation.
- More novel risk factors were also identified, including monocytes (a type of white blood cell), C-reactive protein (a marker of inflammation), CK (creatine kinase, an enzyme used to measure damage to muscle, including heart muscle), and β2-microglobulin (a protein that is correlated with tumor mass).
- The total score was used to classify patients into three atrial fibrillation risk groups: low risk (0-6), intermediate risk (7-11), and high risk (≥12).
Conclusions:
Researchers used artificial intelligence to develop a risk score for predicting atrial fibrillation in treatment-naïve patients with CLL. This risk score consisted of 10 key variables out of over 100 that were initially considered. Patients who are categorized as at high risk of atrial fibrillation should be closely monitored. However, the development of atrial fibrillation does not affect overall survival or progression-free survival.
Links and Resources:
Watch the interview on the abstract here:
You can read the actual ASH abstract here: Development of a Machine Learning-Based Risk Score for Predicting Atrial Fibrillation in Treatment-Naïve CLL Patients Initiating BTK Inhibitor Therapy
