Preoperative virtual reduction method for pelvic fractures based on statistical shape models and partial surface data

Abstract

Virtual reduction is crucial for successful and accurate reduction of pelvic fractures. Various methods have been proposed in this regard. However, not all of them are applicable to every pelvic fracture. Among these methods, the efficiency and accuracy of the method based on statistical shape models in clinical applications require further improvement. This study proposes a virtual reduction method for pelvic fractures that uses statistical shape models and partial surface data of a broken pelvis. Simulated fracture and clinical case experiments were conducted to validate the accuracy and effectiveness of the proposed method. The simulated fracture experiments yielded an average error of 1.57 ± 0.39 mm and a maximum error of 12.82 ± 3.54 mm. The virtual reduction procedure takes approximately 40 s. Based on three clinical case experiments, the proposed method achieves an acceptable level of accuracy compared with manual reduction by a surgeon. The proposed method offers the advantages of shorter virtual reduction times and satisfactory reduction accuracy. In the future, it will be integrated into the preoperative planning system for pelvic fracture reduction, thereby improving patient outcomes.