An AI-Assisted Conceptual Framework for Personalized Nutrition Recommendation Based on Oral Mucositis Severity in Head and Neck Cancer Patients

Abstract

Oral mucositis often occurs in head and neck cancer (HNC) patients during treatment and can make eating difficult because of pain and swallowing problems. A recent meta-analysis reported oral mucositis in approximately 94% of HNC patients receiving radiotherapy, with 37% experiencing severe mucositis. In Sri Lanka, cancers of the lip, tongue and mouth accounted for 18.8% of reported cancers among males and 4.1% among females in 2022. Nutritional needs may change according to mucositis severity, treatment related symptoms, dietary restrictions and individual preferences. This paper proposes an AI assisted conceptual framework for personalized nutritional recommendation for HNC patients with oral mucositis. The framework integrates clinical information, nutritional requirements, patient reported symptoms, dietary preferences and mucositis severity to support individualized meal recommendations. Its main components include patient data collection, mucositis assessment, nutritional requirement estimation, meal recommendation, patient feedback and healthcare professional supervision. The proposed framework was developed using findings from the literature and initial interviews with 16 patients, 4 doctors, and 5 nurses. This study mainly focuses on identifying system requirements and developing the conceptual framework. System implementation, AI model training and clinical validation are considered future work. Future evaluation will focus on accuracy, safety, usability and clinical relevance.