Deep Learning Model for Differentiating Between Neoplastic Pathologic Fracture and Nonpathologic Fracture Using Hip...

Area of Science:

• Orthopaedic Surgery
• Medical Imaging
• Artificial Intelligence

Background:

• Radiographs are the primary imaging method for fractures.
• Distinguishing neoplastic pathologic fractures from nonpathologic fractures on radiographs can be difficult.
• Diagnostic accuracy is crucial for effective patient management.

Purpose of the Study:

• To develop and evaluate a deep learning model for differentiating neoplastic pathologic fractures from nonpathologic fractures on hip radiographs.
• To enhance diagnostic accuracy in fracture classification.
• To provide a tool to assist clinicians in interpreting radiographic findings.

Main Methods:

• A retrospective, multicenter study analyzed anteroposterior hip radiographs.
• A deep learning model was trained and tested on data from 338 patients.
• External validation was performed on 488 patients across three additional institutions.

Main Results:

• The model achieved an accuracy of 0.880 internally and 0.848 externally.
• Internal sensitivity was 0.882 and specificity was 0.879.
• External validation showed a sensitivity of 0.910 and specificity of 0.786, comparable to orthopaedic surgeons.

Conclusions:

• The deep learning model reliably distinguishes neoplastic pathologic fractures from nonpathologic fractures.
• The model can support orthopaedic surgeons, particularly in resource-limited environments.
• The tool has the potential to optimize radiographic interpretation and improve patient care.