A Learning Program for Treatment Recommendations by Molecular Tumor Boards and Artificial Intelligence.

Area of Science:

• Oncology
• Biomedical Informatics
• Quality Improvement

Background:

• Molecular tumor boards (MTBs) exhibit significant variability in treatment recommendations, particularly for biomarkers with limited evidence.
• Standardization of MTB recommendations is crucial for consistent patient care.

Purpose of the Study:

• To evaluate if a learning program enhances the standardization of MTB treatment recommendations.
• To assess the efficacy of an artificial intelligence (AI)-based annotation system in improving recommendation concordance.

Main Methods:

• A prospective quality improvement study involving 42 participants and 50 simulated cases.
• Cases were divided into pre- and post-learning evaluation groups to assess concordance before and after the intervention.
• Participants included MTBs, treating physicians, and an AI system.

Main Results:

• MTB accreditation rate was 55.6%.
• Concordance of treatment recommendations among MTBs increased significantly post-learning (58.7% to 67.9%).
• The AI system achieved significantly higher concordance (88.0%) compared to MTBs.

Conclusions:

• The learning program effectively improved the concordance of treatment recommendations within MTBs.
• AI systems show promise for enhancing accuracy and consistency in clinical decision-making for cancer treatment recommendations.