The study investigates taste disorders that commonly occur during or after cancer treatment, often leading to issues such as malnutrition and treatment discontinuation. Although many non-pharmacological recommendations exist, it is unclear which methods are suitable for which individuals. This pilot study aims to use an AI-based, or rule-driven system to generate personalized recommendations based on patients' specific impairments and taste disorder. For the pilot study, the AI will be trained on recipes to create individualized meal plans, helping to identify foods that are likely to be better. The primary endpoint is the assessment of the nutritional intervention as helpful at the second visit within 12 weeks of inclusion.

The primary goal of this pilot study is to evaluate the feasibility and perceived helpfulness of a web application using AI-supported, personalized dietary recommendations for patients experiencing taste disorders during or after cancer therapy. The primary endpoint is the assessment of the nutritional intervention as helpful at the second visit within 12 weeks of inclusion.

The study is a prospective, single-arm, non-blinded intervention study with a maximum duration of 12 weeks for individual participants. Participants will be identified on the ward or in outpatient centers during their clinical treatment routine. Study measures and progress assessments will also only take place during the chemotherapy administrations planned as part of the clinical treatment routine or during the follow-up visits.

Data is entered into the Gustabor portal together with study staff. Initially, anthropometric data, eating habits, taste experiences, and taste alterations are recorded via questionnaires. Taste thresholds for sweet, sour, salty, bitter, and umami are then assessed using taste strips. Based on these inputs, an AI system generates personalized recipe recommendations as well as recommendations for actions (e.g., cold meals recommended in case of metallic taste), which are printed and reviewed with the participant. Additionally, participants receive a flyer containing general dietary advice for oncology patients. At the start of the study and again after a period of at least three and up to twelve weeks (timed to fit clinical routines), the Patient-Generated Subjective Global Assessment (PG-SGA) score will be recorded, evaluating weight, symptoms, dietary intake, and physical activity. At the second visit it will be assessed which recommendations were implemented and how helpful they were perceived overall (primary endpoint) and in detail.

Inclusion Criteria:

• Age ≥ 18
• Suffering from one of the following tumor entities: Multiple myeloma or malignant tumor of the gastrointestinal tract
• Mainly oral nutrition
• Subjectively perceived tumor therapy-related taste disorder
• At least two clinical presentations planned within 12 weeks with a minimum interval of three weeks
• Ability to participate in nutritional intervention, including use of the online portal and implementation of recipe suggestions (e.g. resources and access to a kitchen), either independently or with third-party support (e.g. by relatives, outpatient care services).

Exclusion Criteria:

• Pregnancy
• Taste disorder explained by other causes (e.g. existing before therapy or COVID disease)
• Placement in an inpatient care facility