Augmenting Prognostication: Utilizing Activity Trackers to Enhance Survival Prediction in Metastatic Non-Small Cell Lung Cancer

Introduction/background: Prognostication by performance status (PS) assessment is a fundamental element of treatment decisions and clinical trial design in oncology, but it is limited by subjectivity and potential miscommunication between patient, physician, and family. Activity tracker offers the potential to collect a broad range of patient-generated data to supplement the assessment of PS.

Patients and methods: Patients with metastatic NSCLC (mNSCLC) participated in a single institute, prospective, observational feasibility study conducted at Huntsman Cancer Institute. Patients were given a Fitbit® activity tracker, which collects their steps taken, distance moved, heart rate, and activity intensity. At baseline, PS was assessed by physicians and patients, and demographics and clinical data were collected. We defined novel indices of health: Heart rate Activity zone Mismatch (HAM) and excessive Sedentary Heart Rate (eSHR). We used multivariable Cox proportional hazards models adjusted for age, sex, and treatment line to estimate and test the prognostic ability of clinical and fitness metrics on overall survival (OS). Each prognostic model was evaluated using Harrell's concordance index (C-index).

Results: Fifty-five patients with mNSCLC were enrolled. The median OS was 10.4 months (95% CI: 7.2, 15.2). PS-physician (HR = 2.0; P < .001) and Fitbit metrics were associated with OS, including daily total steps (1,000-steps) (HR = 0.8; P = .004), HAM (HR = 2; P = .02), eSHR (HR = 0.3; P = .001). The prognostic model that includes PS-physician was associated with the best concordance (C-index = 0.75), followed by daily total distance (C-index = 0.74) and steps (C-index = 0.73) CONCLUSIONS: Tracker-based measures were prognostic of survival in mNSCLC and may be useful as a supplement or alternative to PS in practice and clinical trials.