Evaluating the diagnostic performance of 3D radiographs in emergency department extremity imaging: an early comparative study.

OBJECTIVES

To compare diagnostic performance of novel three-dimensional advanced reconstruction (3D AR) with conventional radiography (CR) in detecting extremity fractures, and to assess its multiplanar reconstruction (MPR) image quality.

METHODS

This prospective, single-center study included 88 extremity exams from 73 adult emergency department (ED) patients who underwent both CR and 3D AR. Two blinded radiologists independently reviewed all exams, with cross-sectional/follow-up imaging as the reference standard. Diagnostic performance metrics and receiver operating characteristics (ROC) were calculated. MPR image quality was graded on a 5-point Likert scale. Inter-reader reliability was assessed with the intraclass correlation coefficient (ICC).

RESULTS

Fractures were confirmed in 30 of 88 exams (34%). CR detected 14 (sensitivity: 46.7%), while 3D AR identified all 30 (sensitivity: 100%). Specificity was high for both modalities: 96.6% for CR, 98.3% for 3D AR. ROC analysis demonstrated significantly superior performance for 3D AR (area under curve, 0.99 vs. 0.71, p<.001). MPR quality was adequate or higher than needed in 85% of cases, with an excellent ICC of 0.88.

CONCLUSION

3D AR significantly improves diagnostic accuracy for extremity fractures compared to CR while maintaining high image quality. Its integration into ED workflows may reduce the need for additional imaging and expedite clinical decision-making. These findings support 3D AR as a valuable tool in the evaluation of acute extremity trauma.