Treatment Recommendations for Gastrointestinal Cancers Via Large Language Models

Application of Large Language Models in the Recommendation of Treatment Plans for Gastrointestinal Cancers

ZhuHai Hospital, Fifth Affiliated Hospital, Sun Yat-Sen University, Jiangmen Central Hospital, Peking University Cancer Hospital (Inner Mongolia Campus), San Raffaele University Hospital, Italy, University Hospital Magdeburg, Germany, City of Hope Medical Center

This study will evaluate the utility of ChatGPT in recommending treatment plans for patients with gastrointestinal cancers, using both retrospective and prospective data.

The medical records of over 1,200 patients with gastrointestinal cancers will be collected retrospectively from participating hospitals. This data will be split into an exploratory dataset (n=200) and a validation dataset (n>=1,000). Within the exploratory dataset, various prompt methods will be used to determine the treatment plans suggested by ChatGPT. Additionally, several clinicians of varied seniority levels will provide their treatment recommendations. For the validation dataset, ChatGPT's suggestions for treatment plans will undergo both qualitative and quantitative assessments by a multidisciplinary consultation (MDT) team. The recommendations from ChatGPT will then be compared with those from the clinicians. Furthermore, this study will incorporate a prospective dataset comprising 400 participants with gastrointestinal cancers. The participants will be randomly allocated to either a control group (n=200) or a ChatGPT-Assisted group (n=200). In the control group, treatment plan recommendations will solely be provided by the clinicians and will guide subsequent treatments. In the ChatGPT-Assisted group, initial treatment plan recommendations will be independently proposed by both ChatGPT and the clinicians. Based on ChatGPT's suggestions, clinicians might selectively adjust their initial plans. Participants will then receive treatments as per these refined plans. Within the ChatGPT-Assisted group, the treatment plans of the initial 100 participants will be evaluated to determine the percentage of patients whose treatment plans are influenced by ChatGPT. Subsequently, the proportion of participants in the entire ChatGPT-Assisted group with treatment plans modified by ChatGPT will be calculated. The study will further monitor the 3-year progression-free survival (PFS) and the 5-year overall survival (OS) rates, contrasting the outcomes between the control and ChatGPT-assisted groups.

In this arm, participants receive treatment plans directly from clinicians without the assistance of ChatGPT.

In this approach, clinicians do not employ any technological assistance and rely solely on their professional expertise and experience to formulate treatment plans for participants.

In this approach, clinicians utilize the ChatGPT technological tool, formulating treatment plans for participants based on its suggestions and their own professional expertise.

The percentage of the initial 100 participants and the overall participants in the ChatGPT-Assisted group whose treatment plans are adjusted by clinicians after consulting with ChatGPT.

Percentage of participants without disease progression over a period of 3 years from the start of treatment.

Inclusion Criteria: Age ≥18 years, both male and female. Pathologically confirmed diagnosis of gastrointestinal cancer (gastric cancer or colorectal Cancer). Detailed medical records available prior to treatment (including chief complaint, history of present illness, radiological examinations, pathological examinations, laboratory tests, etc.). Participants will receive complete treatment in the participating hospitals. Exclusion Criteria: Participants with cancers other than gastrointestinal cancers. Participants who receive treatment in multiple hospitals.

Individual participant data (IPD) may be made available to other researchers upon request. Interested researchers should present a reasonable research proposal and a data usage application. All participating units of this study will review and assess the proposal and application to determine whether to share the data.

Data will become available 6 months after study completion and will remain available for a period of 5 years.

Interested researchers should submit a detailed research proposal and a data usage application for review. All participating units of this study will assess the application to determine eligibility for data access.