One of the topics identified as an important minimally invasive technique in cancer patients is the use of wearables. This project focuses on the use of wearable technology to improve toxicity management for people with cancer undergoing treatments with a high toxicity profile. Chemotherapy, chemoradiation therapy or immunotherapy are cancer treatments that come with a high risk for acute, sometimes life-threatening treatment toxicity. Patient-report toxicity grading has been shown to aid early detection and adequate treatment of toxicity but puts an additional burden on patients and is only capable of detecting clinically overt toxicity. Personalized monitoring of health status during cancer treatment by wearable technology can likely capture many of the (pre)clinical signs of (imminent) toxicity with limited burden to patients. Machine-learning derived algorithms applied to this data could contribute to earlier detection and management of treatment toxicity, potentially improving treatment outcomes and ultimately quality of life. Before further development and implementation of this approach can be undertaken, it is important to understand the technical feasibility and patient acceptance of data collection via wearables, as well as the optimal type of data and data-collection timeframe(s), in the context of cancer care (e.g. timing relative to treatment cycles). We will carry out an explorative developmental study focusing on these outcomes. Among head and neck and melanoma cancer patients undergoing treatment, we will
- Collect data using both high-grade experimental (i.e. biosensors capable of continuously collecting many physiological parameters) and consumer-grade (i.e. Fitbits) wearables.
- Wearable data will be correlated with clinical toxicity data collected via medical files and via patients’ self-report by using an app (i.e. Kaiku), as is already current clinical practice in different clinics.
- Preliminary algorithms for early detection of treatment toxicity will be developed based on pilot data
The results of this pilot study will inform a large-scale study to further develop AI algorithms for this purpose and to evaluate the clinical effectiveness of this minimally-invasive approach to personalized toxicity management.