A Simple Method to Improve Detection of Femoral Nail Abutment in the Distal Femur: A Computer Modeling Study.

Abstract

Proximal femur fractures are more frequently treated with long femoral nails. Lateral radiographs are used to assess the nail position in the distal femur. However, because of the asymmetric shape of the distal femur, standard lateral radiographs alone are suboptimal for assessing anteriorly positioned nails in the distal femur. Consequently, instances of nail abutment or even perforation might be missed intraoperatively. Using a three-dimensional (3D) modelling approach, we asked: When the nail is in the anterior fifth of the canal, will rotating the femur to align the simulated x-ray beam with the anterior femoral condyles instead of the posterior femoral condyles increase the diagnostic accuracy of detecting nail perforation of the anterior cortex? 3D models of 42 unilateral femora from a population sample representative of patients with hip fractures (mean age of 76 ± 7 years, 10 males and 32 females, and 16 left and 26 right femora) were used. The patients had a mean height of 158 ± 9 cm; 27 femora were from Asians and 15 were from Caucasians. Clinically appropriately sized nails had already been virtually implanted previously as part of a quantitative nail fit assessment study. In a preliminary step, the mean angles of inclination of the distal anterior supracondylar region were quantified using four axial sections of the distal femur. For the femora with the nail tip in the anterior fifth of the canal, projections representing a lateral radiograph were generated along with rotated projections at mean angle (5°, 8°) rotations, with the anterior femoral condyles aligned, and anterior femoral condyle alignment followed by internal and external rotation to detect maximum nail perforation. The distance from the nail to the distal anterior cortex was measured for each rotational projection and used to detect anterior nail perforation. The accuracy of detection was assessed using the area under the curve (AUC) of the receiver operating characteristic (ROC) analysis. Rotating the 3D models by aligning the x-ray beam with the anterior femoral condyles improves the diagnostic ability of detecting anterior nail perforation compared with standard lateral radiographs. The AUC increased with rotation from 0.50 (95% confidence interval 0.50 to 0.50) on the lateral projection to 0.73 (95% CI 0.57 to 0.88, difference = -0.23; p = 0.004) at 5° of rotation, 0.77 (95% CI 0.62 to 0.93, difference = -0.27; p = 0.001) at both 8° of rotation and with the anterior femoral condyles aligned, and to 0.82 (95% CI 0.67 to 0.97, difference = -0.32; p < 0.001) with internal/external rotations past the anterior femoral condyles. There were no differences in accuracy between the four methods of rotation. This study shows that rotating the femur to align the anterior femoral condyles on a lateral radiograph and then internally/externally rotating it, improves the accuracy assessed via the AUC of detecting anterior perforation when long nails are positioned in the anterior fifth of the distal femur. This approach can easily be used in a clinical setting. Intraoperatively, the image intensifier can be rotated around the leg to produce an image with the anterior femoral condyle aligned, providing surgeons with an opportunity to identify and improve the nail's position or exchange the nail while the patient is still under anesthesia.