Measurements of bone tunnel size in anterior cruciate ligament reconstruction: 2D versus 3D computed tomography model

Abstract

Background: Revision anterior cruciate ligament (ACL) reconstruction requires a precise evaluation of previous tunnel locations and diameters. Enlargement of the tunnels, despite not usually affecting primary reconstruction outcomes, plays an important role in revision ACL management. Three dimensional (3D) computed tomography (CT) models are reported to be the most accurate method for identifying the tunnel position and possible conflicts with a revision tunnel placement. However, the ability of 3D CT to measure the tunnel size is still not proven. The goal of this study was to evaluate the ability of measuring the size of the bone tunnels in ACL reconstructed knees with 3D CT compared to the traditional two dimensional (2D) CT method.

Methods: Twenty-four patients had CT scans performed immediately following ACL reconstruction surgery. Their femoral tunnels size were measured by a standard 2D CT measurement and then compared with three novel 3D CT measuring methods: the best transverse section method, the best fit cylinder method and the wall thickness method. The drill size used during surgery was used as a control measure for the tunnel width. Intra-class correlation coefficients were obtained.

Results: The intra-class correlation coefficient and respective 95% confidence interval range (ICC [95%CI]) for the three methods compared with the drill sizes were 0.899 [0.811-0.947] for the best transverse section method, 0.745 [0.553-0.862] for the best fit cylinder method, -0.004 [-0.081 to -0.12] for the wall thickness method and 0.922 [0.713-0.97] for the 2D CT method. The mean differences compared to the drill size were 0.02 mm for the best fit transverse section method, 0.01 mm for the best fit cylinder diameter method, 3.34 mm for the wall thickness method and 0.29 mm for the 2D CT method. The intra-rater agreement (ICC [95%CI]) was excellent for the best transverse section method 0.999 [0.998-0.999] and the 2D CT method 0.969 [0.941-0.984].

Conclusions: The 3D best transverse section method presented a high correlation to the drill sizes and high intra-rater agreement, and was the best method for ACL tunnel evaluation in a 3D CT based model.

Keywords: 3D CT model; Anterior cruciate ligament; Revision ACL; Tunnel enlargement.

Figure 1

Segmentation process of 2D CT images. Segmentation process performed on the original 2D CT images viewed in axial (A), coronal (B) and sagittal (C) CT images.

Figure 2

3D model acquisition. CT based 3D model showing all bone structures segmented (A). Same image with the femur hidden (B), and a closer view of the segmented femoral tunnels (C).

Figure 3

Best fit cylinder method. The original tunnel 3D model was exported to the 3D-matic® software (A). An analytical cylinder was generated (B) and, a surface cylinder was created guided by the analytical cylinder. The cylinder diameter was measured (C).

Figure 4

Wall thickness method. The original tunnel 3D model exported to the 3D-matic® software (A) and the Wall Thickness function was applied. A color scale showed the distance between the triangles and the opposite wall (B).

Figure 5

Best transverse section method. The segmented tunnel (A) was used as a guide for the automated centerline drawing made by the Mimics® software. Centerline half-way distances were measured and the best fit diameter on this points was evaluated (B). Image of the tunnel model with the measurements on its surface (C).

Figure 6

2D CT method. The measurements were made based on the original CT images, and the tunnels were assessed in all three CT planes: coronal (A), sagittal (B), and axial (C).

Figure 7

Difference from Drill Size Mean. Shows the difference between the average of the drill size and the mean of the measurements for each method with respective standard deviations.