Aims

This study was designed to develop a model for predicting bone mineral density (BMD) loss of the femur after total hip arthroplasty (THA) using artificial intelligence (AI), and to identify factors that influence the prediction. Additionally, we virtually examined the efficacy of administration of bisphosphonate for cases with severe BMD loss based on the predictive model.

Rotational acetabular osteotomy (RAO), one of periacetabular osteotomies, is an effective joint-preserving surgical treatment for developmental dysplasia of the hip. Since 2013, we have been using a CT-based navigation for RAO to perform safe and accurate osteotomy. CT-based navigation allows precise osteotomy during surgery but cannot track the bony fragment after osteotomy. Thus, it is an issue to achieve successful reorientation in accordance with preoperative planning. In this presentation, we introduce a new method to achieve reorientation and evaluate its accuracy.

Thirty joints in which CT-based navigated RAO was performed were included in this study. For the first 20 joints, reorientation was confirmed by tracing the lateral aspect of rotated fragment with navigation and checked if it matched with the preoperative planning. For the latter 10 joints, a new method was adopted. Four fiducial points were made on lateral side of the acetabulum in the preoperative 3-dimensional model and intraoperatively, rotation of the osteotomized bone was performed so that the 4 fiducial points match the preoperative plan.

To assess the accuracy of position of rotated fragment in each group, preoperative planning and postoperative CT were compared. A total of 24 radial reformat images of postoperative CT were obtained at a half-hour interval following the clockface system around the acetabulum. In every radial image, femoral head coverage of actual postop- and planned were measured to evaluate the accuracy of acetabular fragment repositioning.

The 4-fiducial method significantly reduced the reorientation error. Especially in the 12:00 to 1:00 position of the acetabulum, there were significantly fewer errors (p<0.01) and fewer cases with under-correction of the lateral acetabular coverage.

With the new method with 4 reference fiducials, reorientation of the acetabulum could be obtained as planned with lesser errors.

SL-PLUS MIA stem (Smith & Nephew Orthopaedics AG) is a modified implant of Zweymuller type SL-PLUS standard stem (Smith & Nephew Orthopaedics AG). We constructed finite element (FE) models and analysed equivalent stresses in the femur. In addition, we measured bone mineral density (BMD) in the femur by dual-energy X-ray absorptiometry (DEXA) after THA. The purpose of this study was to investigate the equivalent stress and to compare the results of the FE analyses with changes in BMD after THA. Twenty-one patients (18 women and 3 men) who underwent primary cementless THA with SL-PLUS MIA or SL-PLUS formed the basis of this study. Eleven patients received SL-PLUS MIA and ten patients received SL-PLUS. Zones were defined according to Gruen's system (zones 1∼7). Computed-tomography (CT) images of the femur were taken before and at 1 week after THA. FE models of the femur and prosthesis were obtained from CT data by Mechanical Finder (Research Center of Computational Mechanics Inc., Tokyo, Japan), software that creates FE models showing individual bone shape and density distribution. Equivalent stresses were analysed in zones 1 to 7 and compared to the DEXA data. FE studies revealed that there was no significant difference in equivalent stress between SL-PLUS MIA and SL-PLUS. BMD was maintained after THA in zones 3, 4, and 5, whereas BMD decreased in zones 2, 6, and 7. In zone 1, BMD decreased in SL-PLUS MIA stem group by 14%, while BMD was maintained in SL-PLUS standard stem.

In this study, we evaluated the labrum tear using radial sequence 3D Multiple Echo Recombined Gradient Echo (MERGE) MRI without arthrography based on modified Czerny's classification, comparing with actual arthroscopic findings. A total of 61 hips including 27 hips of femoroacetabular impingement (FAI), 19 hips of borderline development dysplasia of the hip (BDDH) and 15 hips of early stage osteoarthritis (OA) were enrolled this retrospective study. MRI findings evaluated in each three regions of interest; anterior region, anterolateral region, and lateral region. The cases with severe degeneration that is not concordant with any original Czerny's classification is defined as stage4. We compared MRI findings with arthroscopic findings and calculated the sensitivity, specificity, and likelihood ratio in terms of the existence of labrum tear. MRI findings revealed labrum tear more frequently in anterolateral than lateral (p<0.001). Especially in FAI group, labrum tear was more frequently observed by MRI in anterolateral than lateral (p=0.006). In comparison with MRI findings and arthroscopic findings, the sensitivity was 97%, specificity was 79% and likelihood ratio was 4.59 as average of all regions in terms of the existence of labrum tear. In each region, sensitivity and specificity was 97% and 50% in anterior, 97% and 100% specificity in anterolateral, 94% and 81% in lateral, respectively. Thus, MERGE MRI revealed excellent sensitivity and specificity for diagnosis of labrum tear, especially in anterolateral region. The cases with severely degenerated labrum were classified as newly defined stage 4, which was recognized frequently in OA cases.

While stable long-term clinical results have been achieved in total joint arthroplasty, periprosthetic joint infection (PJI) has been actualized as difficult issue in this decade. For accurate diagnosis, it is important to establish standard criteria such as MSIS criteria, and it is prevailing now. As an issue involving PJI, however, the existence of viable, but non-culturable (VNC) bacteria must be noticed. It is difficult to identify the VNC state infection, because microbiologic culture result shows negative and other markers tend to be negative. Here, molecular diagnosis based on polymerase chain reaction (PCR) has certain role as potential diagnostic tools for such VNC infection. We have applied a real-time PCR system for the diagnosis of PJI, which is able to detect methicillin-resistant Staphylococcus (MRS) and distinguish gram-positive from gram-negative bacteria. The prominent advantage is that PCR is the singular way to identify MRS in such culture negative cases. Recent development of full-automatic PCR system may improve the time efficiency for routine application. In this presentation, we will show the overall sensitivity and specificity of our PCR system for diagnosing PJI and discuss the current problem and future prospect.