Deep learning ensemble models for CT-based differentiation of malignant and benign sacral bone tumors: development and evaluation

Key Points

• The aim is to develop an artificial intelligence model for classifying sacral bone tumors using NCCT scans.
• Developed an AI-radiologist ensemble model for tumor classification.
• Evaluated performance based on CT images.
• Compared improvements across radiologists of varying experience.
• All radiologists showed improved classification performance.
• The most significant improvements were observed in junior radiologists.
• The model reduces reliance on contrast and MRI for diagnosis.