Are Two- and Three-Dimensional Computed Tomographic Scan Measurements of Tibial Tubercle-Trochlear Groove Distance Equivalent? A Comparative Study

Abstract

Background: Accurate planning for patellar instability correction is important in obtaining good post-operative outcome. The main challenge in the current two-dimensional (2-D) computed tomographic (CT) scans method is the difficulty in choosing reliable bony landmarks. This study aimed to compare the reliabilities between the 2-D and three-dimensional (3-D) methods of measuring tibial tubercle-trochlear groove (TT-TG) distance. We hypothesize that the proposed 3-D method will result in measurements with narrower error margin, providing higher reliability and accuracy.

Materials and methods: We traced CT scans of 106 knees with no patellofemoral pathology from 59 subjects from the database system and converted all 2-D images into 3-D models to determine the values for each parameter. We compared the intra- and interobserver reliability of each method using intraclass correlation (ICC) and Bland-Altman method.

Results: The values of TT-TG measured by 2-D and 3-D methods were 16.1 ± 4.6 mm and 16.2 ± 4.2 mm, respectively. The ICC values of both methods were comparable (95% limits of agreement between the same observer: - 3.3 to 3.8 mm versus - 2.4 to 2.7 mm and different observers: - 4.3 to 4.9 mm versus - 3.9 to 2.7 mm), with 3-D method results in narrower limits of agreement.

Conclusion: TT-TG measurement is reliable using the 2-D method without using advanced radiographic software. The 3-D method of measuring TT-TG provides measurement with narrower variation when compared with the 2-D method. However, both TT-TG distances' measurement methods in the current study were comparable as the variations are not significant.

Keywords: CT scan; Inter- and intraobserver reliability; Intraclass correlation; Patellar instability; Three-dimensional model.

Fig. 1

Identification of TG (Point A) from the 2-D image and the resulting point in the 3-D image. The identified Point A from 2-D image resulted in an inaccurate location of TG when transformed into the 3-D image

Fig. 2

Determination of TT–TG distance through the 2-D method. A The a–b and c–d lines were drawn to determine the TG and were later copied on the axial view B of the TT to determine the TT–TG distance (c–e)

Fig. 3

Images used in the determination of TT–TG distance on the 3-D image. TG was identified from the axial view (A) and 3-D model (B) of the distal femur. TT was identified from the axial view (C), sagittal view (D), and 3-D image (E) of the proximal tibia. The points for TG (A) and TT (C) on the 3-D images were identified and the distance between the points (points c and e) was measured

Fig. 4

The Bland–Altman plots following the analysis of 95% limits of agreement for intraobserver analyses for 2-D (A) and 3-D methods (B) and interobserver analyses for 2-D (C) and 3-D methods (D)