Abstract: 

The efficacy and toxicity of diverse drugs including more than 50 anticancer drugs largely depend on dosing time. Despite the potential benefits to patients from time-of-day treatment, current clinical practice guidelines have largely ignored it. This is mainly due to the lack of reliable and efficient methods to identify the patient’s internal time in the real world. Wearables (e.g., Apple Watch) provide an opportunity for non-invasive continuous monitoring of physiological signals, such as activity and heart rate. In this talk, I will present a Kalman filter approach that assimilates wearable data into the model of the human circadian (~24 hr) clock and estimates the internal circadian time. I will also introduce a Kalman filter-assisted neural network approach for early detection of aberrant changes in circadian physiology related to disease progression from wearable measurements. The mathematics of the wearables can pave the way toward precision medicine in the real world.