EJC Skin Cancer (Dec 2024)

Electronic patient-reported outcomes, fever management, and symptom prediction among patients with BRAF V600 mutant stage III–IV melanoma: The Kaiku Health platform

  • Peter Mohr,
  • Paolo Ascierto,
  • Alfredo Addeo,
  • Maria Grazia Vitale,
  • Paola Queirolo,
  • Christian Blank,
  • Jussi Ekström,
  • Joonas Vainio,
  • Vesa Kataja,
  • Sibel Gunes,
  • Mia Engström-Risku,
  • Henriette Thole,
  • Ailis Fagan,
  • Frederico Calado,
  • Ruben Marques,
  • Judith Lijnsvelt

Journal volume & issue
Vol. 2
p. 100254

Abstract

Read online

Background: The Kaiku Health platform is an electronic patient-reported outcome (ePRO) system aimed at monitoring symptoms and quality of life (QoL) among cancer patients. ePRO systems allow timely interventions to symptoms, which has been shown to improve treatment adherence, QoL, and survival. We examined symptoms and QoL in stage III–IV melanoma patients using the Kaiku ePRO module. Additionally, we evaluated prediction of symptom onset or continuation using a machine learning (ML) model. Methods: Patients ≥18 years old with high-risk stage III (adjuvant) or unresectable/metastatic stage IV BRAF V600E/K–mutant melanoma treated with BRAF/MEK inhibitor combination were included in a descriptive cohort study. Data on symptoms and QoL were collected through questionnaires in 2021–2022 using the digital Kaiku ePRO platform. Results: Altogether, 44 patients were included. The weighted average for compliance with the symptom questionnaire was 73 %. The compliance with the QoL questionnaire was 77 % at the 1–3 months interval and 50 % at the 7–9 months interval. The most common symptoms were fatigue (85.4 %), headache (56.1 %), muscle/limb pain (56.1 %), skin changes (56.1 %), and cough (51.2 %). The median global health status was 66.7 at baseline and 75.0 at 3 months follow-up. ML-based prediction performance for nine symptoms (blurred vision, fatigue, cough, shortness of breath, skin changes, fever, itching, diarrhoea, and lymphoedema) was good to excellent. Conclusion: The Kaiku ePRO platform was well-accepted and had high compliance, with both the symptom and QoL questionnaires. The ML algorithm accurately predicted the onset and continuity of nine symptoms.

Keywords