Abstract
Object
CT-based navigation system in total hip arthroplasty(THA) is widely used to achieve accurate implant placement. However, its internal structure was a trade secret. Therefore, it was hard to analyze optimal reference points. Now, we developed our own CT-based navigation system originally, and since then we have been conducting various analyses in order to use the system more effectively. The purpose of this study was to analyze the optimal area and the number of registration points, which enables to move initial errors into the acceptable range.
Methods
We set the anterior pelvic plane(APP) as the reference plane, and defined the coordinates as follows: X-axis for external direction, Y-axis for anterior direction, and Z-axis for proximal direction. We made pelvic bone models after THA, a normal shape and acetabular dysplasia model, and performed registration using an originally developed CT based navigation system. At first, we registered point paired matching at 4 points, and surface matching was performed at 53 points, which were placed around the acetabulum. 20 points were on anterosuperior, 10 points were on posterosuperior, 20 points were on posterior around the acetabulum, and 3 points were on the pubis. We selected surface matching points based on the actual operation approach, calculated the accuracy of the error correction, and searched the optimal area and the number of surface matching points.
Results
Initial error could correct under 2mm gradually more than 20 surface matching points on anterosuperior area of the acetabular, assumed anterolateral approach. Initial error correction was improved by adding the reference points on the pubis. In comparison with the initial error on the X-axis, corrections of the Y-axis and Z-axis were more difficult. Initial error could correct under 2mm with around 20 surface matching points on posterior area of the acetabulum, assumed posterior approach, initial error on the Y-axis direction had difficulty in correction similarly. The difference of the shape of the pelvic effected initial error correction, it was more difficult to correct initial error on Y, Z-axis on dysplasia hip.
Discussion
To correct initial error needs more than 20 surface matching points at least on the both anterolateral and posterior approaches. However, it is hard to correct initial error depending on surface matching points on the Y-axis, so it is desirable to return to point paired matching. On the registration of anterosuperior area of the acetabulum, correction of the initial error improved effectively by adding reference points on the pubis in this study, so to get reference points on the pubis is recommended if possible. Moreover, it was hard to correct initial error on the Z-axis because reference points on the anterosuperior of the acetabulum were flat, therefore it is necessary to reduce the initial error on Z-axis, we should consider choosing other reference points such as on the iliac crests which can define Z-axis direction. When using the navigation system on pelvic, some tendencies and characteristics exist, and understanding them can make the operation smooth and precisely.
