A feasibility study of ApricityCarefor early detection and management of immune-related adverse events in cancer patients receiving immune checkpoint therapies.

Abstract

e13711 Background: Immune-related adverse events (irAEs) is a leading cause of immune checkpoint therapy (ICT) interruption or discontinuation. Earlier detection of irAEs can improve outcomes and extend ICT continuation. ApricityCare is a HIPAA-compliant AI-powered remote symptom monitoring platform designed to facilitate early irAE detection and standardize management. Hypothesis: ApricityCare in a routine care setting can enable remote patient symptom monitoring and irAE detection. Methods: This prospective, single-center pilot feasibility study was conducted Feb 1, 2019 to Feb 1, 2020. Enrolled patients receiving ICT were instructed to report symptoms and biometric data via ApricityCare over a 12-week period. Pre-specified algorithms based off inputted data generated alerts categorized by suspected irAE type and severity: green (no concerning symptoms), yellow (mild-moderate symptoms), and red (moderate-severe symptoms). The co-primary endpoints were frequency of symptom reporting and educational video access. Feasibility was defined as 80% of patients reporting ≥3 times per week. Secondary endpoints included distribution of generated alerts. Post-hoc exploratory analysis was performed to assess irAE incidence based off documentation and lab results, detection accuracy, and interventions. Results: 44.4% of participants (n = 18) checked in via ApricityCare ≥ 3 times per week, and 38% chose to continue ApricityCare after the trial. Of the 705 total alerts recorded, 72% were green, 24% yellow, and 5% red. Ten confirmed or suspected irAEs were documented, of which 6 were detected via ApricityCare: dermatitis Grade (G) 1, dermatitis G2, thyroiditis G2, myositis G2, arthralgia G2, and colitis G1. All received or initiated treatment within 48 hours of alert with no delays in ICT. The 4 irAEs not detected included hepatitis G2, 2 cases of thyroiditis G2, and pneumonitis G3 which required ICT discontinuation. Rate of irAE incidence was 50% (4/8) for patients who frequently checked in and 30% (3/10) for those who did not. Out of 8 patients who prematurely stopped check-ins prior trial completion, one experienced a subsequent irAE. Conclusions: While frequency of remote symptom reporting via ApricityCare did not meet its co-primary endpoint, a subset of patients showed high engagement and majority of irAEs were captured. Lessons learned included: (1) patient adherence and accurate data input are critical; (2) rigid check-in schedules limit patient uptake, especially when asymptomatic; (3) proactive monitoring creates primarily non-urgent alerts that increase clinic workload. Building on these, ApricityCare was revised with flexible check-ins and first-response triage by remote nurses. A prospective clinical trial is underway to evaluate the updated ApricityCare program in routine care and clinical trial populations (NCT06693687).