Untapping the Prognostic Value of Patient-Generated Health Data in Locally Advanced Non-small Cell Lung Cancer

Abstract

Patient-generated health data (PGHD), which includes patient-reported outcomes (PROs) and wearable device data, may have prognostic value for cancer patients. We tested that hypothesis using data from several prospective trials where patients with locally advanced non-small cell lung cancer (LA-NSCLC) were treated with definitive chemoradiotherapy. Cox proportional hazards models were utilized to identify the baseline patient-reported symptom that best predicted progression-free survival (PFS) duration in a trial that involved PRO-CTCAE collection (Cohort 1). Using data from trials that included EORTC QLQ-C30 questionnaires and wearable devices (Cohort 2), the same symptom was tested as a predictor of PFS. Baseline physical inactivity was also tested as a predictor of PFS. A simple risk stratification tool utilizing PROs and physical activity was proposed. In Cohort 1 (n = 50), anorexia was the only pretreatment PRO that was significantly associated with PFS after Bonferroni correction (HR = 3.94, = .002). In Cohort 2 (n = 58), baseline anorexia was also significantly associated with PFS (HR = 2.48, = .018), as was physical inactivity (HR = 3.11, < .001). Median PFS duration for patients in Cohort 2 with anorexia or physical inactivity was 6 months, compared to 18 months for other patients (HR = 3.08, < .001). Median overall survival duration for patients with anorexia or physical inactivity was 19 months, compared to 65 months for other patients (HR = 2.44, = .021). PGHD, including PROs and wearable device data, can provide valuable prognostic information for LA-NSCLC patients treated with definitive chemoradiotherapy. These findings should be validated using larger datasets.