Skip to content

Can ChatGPT Serve as a Supportive Tool for Multidisciplinary Tumor Boards? A Single-Center, Retrospective Analysis

Can ChatGPT Serve as a Supportive Tool for Multidisciplinary Tumor Boards? A Single-Center, Retrospective Analysis

Status
Recruiting
Phases
Unknown
Study type
Observational
Source
DRKS
Registry ID
DRKS00034227
Enrollment
3000
Registered
2024-09-27
Start date
2024-10-01
Completion date
Unknown
Last updated
2025-04-07

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

C49 C49 C25 C15 C16 C18 C19 C22 C50 C43 C25

Interventions

Group 1: Historical Tumor Board Decisions: This arm contains the actual historical decisions made by multidisciplinary tumor boards for oncological patients, based on the existing anonymized clinical

Sponsors

Universitätsmedizin mannheim
Lead Sponsor

Eligibility

Sex/Gender
All
Age
18 Years to No maximum

Inclusion criteria

Inclusion criteria: - Patients with confirmed diagnoses of cancers. - Cases with comprehensive documentation of tumor board decisions and clinical data.

Exclusion criteria

Exclusion criteria: - Cases lacking comprehensive documentation of tumor board decisions or missing significant clinical data. - Patients whose cases were discussed in emergency settings outside regular tumor board meetings.

Design outcomes

Primary

MeasureTime frame
"Degree of concordance between the historical tumor board decisions and recommendations generated by various LLM approaches (Base LLM, RAG-enhanced LLM, and Fine-tuned LLM), measured using a 5-point Likert scale (1 = no concordance at all, 5 = complete concordance).

Secondary

MeasureTime frame
- Comparison of the concordance rate between the three LLM approaches (Base LLM vs. RAG-enhanced LLM vs. Fine-tuned LLM). - Evaluation of the clinical applicability of LLM-generated recommendations by expert specialists. - Analysis of concordance rates by tumor entity (colorectal carcinoma, primary liver tumors, soft tissue sarcomas). - Assessment of the accuracy of LLM recommendations in relation to current medical guidelines. - Identification of specific clinical scenarios or patient groups in which LLMs generate particularly precise or imprecise recommendations.

Countries

Germany

Contacts

Public ContactCui Yang

Universitätsmedizin Mannheim

cui.yang@umm.de+496213835152

Outcome results

None listed

Source: DRKS (via WHO ICTRP) · Data processed: Feb 4, 2026