Exploring the potential of using Digital Phenotyping for monitoring late effects after breast cancer

Background: With an increasing number of cancer survivors requiring follow-up and the burdens of intensive face-to-face monitoring, finding alternative solutions is imperative. This study explores the potential of digital phenotyping, defined as ``moment-by-moment quantification of the individual-level human phenotype in situ using data from personal digital devices, in particular smartphones” (Torous et al., 2016) for unobtrusively monitoring late effects after breast cancer.

Methods: Following Mohr et al.'s (2017) hierarchical model, the study investigates smartphone sensors, low-level features, and high-level behavioral markers for passive monitoring of various late effects, including cardiotoxicity, neuropathy, lymphedema, fatigue, insomnia, anxiety, post-traumatic stress disorder, depression, and fear of recurrence. A literature search identified features and markers connecting sensor data with late effects. Additionally, three eHealth and monitoring experts were consulted. The number of features, markers and sensor systems identified per late effect was used to create a ranking.

Findings: The literature review uncovered diverse smartphone sensors and electronic activities generating passive data for monitoring late effects, like physical activity, location, and sleep rhythms. As an example, depression has several symptoms, including sleep disturbance, and fatigue. The associated markers include reduced participation in activities and difficulty falling asleep. Associated features include, less socializing that is tracked through GPS locations, irregular sleep cycles through screen-use times etc. Neuropathy, cardiotoxicity, and depression emerge as promising candidates based on the number of associated high-level markers, low-level features and sensors, highlighting the possibility of smartphone-based monitoring for late effects in cancer survivors.

Conclusions: This research explores the potential of digital phenotyping to track late effects after breast cancer and provides an overview of associated high-level behavioral markers, low-level features, and sensors . The ranking identifies neuropathy as most promising for monitoring through digital phenotyping. While offering a framework for subsequent studies, further data-driven validation for smartphone-based monitoring of late effects is crucial.