The CPAK classification in three-dimensional measurements is consistent with those in two-dimensional measurements

This study aimed to evaluate whether the coronal plane alignment of the knee (CPAK) classification between three-dimensional (3D) measurement using computed tomography (CT) and two-dimensional (2D) measurement using long leg radiographs (LLR) match.

A retrospective radiographic study compared pre-operative CT and LLR measurements in 69 patients undergoing primary total knee arthroplasty. The arithmetic hip-knee-ankle angle (aHKA), joint line obliquity (JLO), and CPAK types were calculated. A match in the CPAK classification was defined as "complete agreement" and agreement was considered partial if the CPAK was located in an adjacent column. If the discrepancy was more than one column, then it was recorded as no agreement.

The mean aHKA was - 0.6° ± 4.4° in 3D measurement and - 1.1° ± 4.3° in 2D measurement (p = 0.42). The mean JLO was 174.2° ± 3.5° in 3D measurement and 175.2° ± 3.6° in 2D measurement (p = 0.07). There was no statistically significant difference in either aHKA or JLO obtained by the two measurement methods, each within 1°. However, the complete agreement was confirmed in only 60.8% of cases. After incorporating even partial agreements including neighboring columns, the agreement rate was 91.2%. The disagreement was noticed in 8.8% of the cases.

The 3D measurement method enables aHKA and JLO calculations similar to those of 2D measurements. However, complete agreement with the CPAK classification in all cases was not obtained. On the other hand, 2D measurement has the advantage that JLO can be easily appreciated and remais more intuitive for surgeons.

III.