Digital Health Technologies for Detecting and Tracking Cancer Treatment-Related Cardiotoxicities

About This Grant

PROJECT SUMMARY/ABSTRACT Several cancer treatments, including chemotherapy, radiation, and modern targeted therapies, can damage the heart and cardiovascular (CV) system (cardiotoxicity). Up to 20% of cancer patients may develop cardiotoxicities like arrhythmias, cardiomyopathy, and heart failure (HF) months or years post-treatment. Overlapping risk factors for CV disease and cancer, along with increased longevity, results in more patients with both conditions. Echocardiography is commonly used to assess cardiotoxicities clinically. However, cardiorespiratory fitness (CRF), a key marker for CV function and functional capacity, can decline by up to 26% post-cancer treatment without any change in echo parameters. Early diagnosis and proactive management can improve treatment success and patients’ quality of life and survival. However, frequent and recurrent echocardiography and CRF assessments are costly and inaccessible for the more than 18M cancer survivors in the US, particularly in underserved communities. Thus, there is a need for a feasible solution to remotely and serially monitor CRF, CV health, and cardiotoxicity, enabling early detection and proactive care for cancer patients and survivors. The short-term goals of this research are to develop and validate artificial intelligence algorithms to detect and track cardiovascular diseases in cancer patients using clinical data with and without wearable data to demonstrate digital health technologies’ potential in proactive tracking and managing cancer treatment-related cardiotoxicity. The long-term goal is to enhance accessibility to longitudinal cardiovascular care and enable remote screening and management of cardiotoxicity for the growing number of cancer survivors. Toward this goal, we recently demonstrated the potential of wearable seismocardiogram (SCG, chest wall vibrations originating from heart movements) signals in differentiating breast cancer survivors with diagnosed cardiotoxicity from controls. We also demonstrated the potential of machine learning using SCG and electrocardiogram (ECG) data in assessing CV health (CRF and hemodynamics) and detecting arrhythmias in HF patients. Based on these prior works, we hypothesize that wearable SCG and ECG features will significantly differ between cancer survivors with cardiotoxicity and controls. We will validate and extend our preliminary findings on cardiotoxicity monitoring using wearable SCG and ECG to distinguish cancer survivors with diagnosed cardiotoxicity from controls. Furthermore, we will develop and validate ML models to early detect CV risk and cardiotoxicity in cancer patients using clinical data (echocardiography, blood biomarker, prior medical history, etc.). Early detection of subclinical cardiotoxicity can enable proactive adjustment of care to improve the quality and long-term survival of cancer survivors. This research addresses inaccessible and infrequent routine CV health follow-ups required throughout the continuum of cancer treatments through novel wearable sensors, which can enable personalized remote management of cardiotoxicity. It has the potential to significantly impact the cardio-oncology field and improve access to healthcare in the US and beyond, especially for underserved and rural communities.