ABSTRACT. Background: Cruciate-retaining total knee arthroplasty (CR-TKA) aims to preserve native knee kinematics and proprioception by maintaining an intact posterior cruciate ligament (PCL). However, intra-operative PCL release is frequently required to achieve balanced flexion and extension gaps. Robotic-assisted TKA (RA-TKA) has been introduced to improve surgical accuracy and soft-tissue balancing, potentially reducing the need for PCL release. Objective: This study compared the incidence of intra-operative PCL release between robotic-assisted and conventional CR-TKA. Methods: A retrospective–prospective cohort study was conducted in patients aged 50 years or older undergoing primary CR-TKA for end-stage knee osteoarthritis between 2023 and 2026. Patients were allocated to conventional TKA (cTKA) or robotic-assisted TKA (rTKA) groups. The primary outcome was the incidence of intra-operative PCL release. Secondary outcomes included visual analog scale (VAS) pain scores, Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), and knee range of motion (ROM). Relative risk (RR) was calculated for PCL release, and linear mixed-effects models were used to analyze longitudinal ROM outcomes. Results: A total of 138 knees were included (cTKA = 86, rTKA = 52). The incidence of intra-operative PCL release was significantly lower in the rTKA group compared with the cTKA group (25.0% vs. 43.0%; RR = 0.58, 95% CI 0.35–0.97; p = 0.036). Baseline demographic characteristics were comparable between groups. No significant differences were observed in postoperative VAS or WOMAC scores at any follow-up time point. Linear mixed-effects analysis demonstrated a significant effect of time on knee ROM (p < 0.001) without a significant effect of surgical technique or group-by-time interaction. Conclusion: Robotic-assisted CR-TKA significantly reduced the incidence of intra-operative PCL release compared with the conventional technique, without compromising postoperative pain, functional outcomes, or knee range of motion. These findings suggest that robotic assistance may facilitate more precise soft-tissue balancing and improved preservation of the PCL in cruciate-retaining TKA.

ABSTRACT. Secure fixation of pelvic reference pins is essential for accuracy in computer-assisted total hip arthroplasty (THA). While pin-related complications are uncommon, serious adverse events have been reported. This study aimed to evaluate pelvic reference pin placement and associated complications using postoperative CT imaging for robotic-arm-assisted (Mako) and CT-navigated THA. We retrospectively analyzed 937 hips (823 patients) that underwent computer-assisted THA between April 2023 and December 2025, including 374 robotic-arm-assisted THA and 563 CT-navigated THA. Postoperative CT images were utilized to assess pin trajectory, intraosseous pin length, cortical penetration, and pin-related complications. The distance from the anterior superior iliac spine (ASIS) to the most anterior pin and intraosseous pin length were measured using Vincent software. Robotic-arm-assisted cases employed three parallel pins inserted with a guide sleeve, while navigation cases utilized two manually inserted pins arranged in an inverted V-shaped configuration. In the Mako group, pin tips were located external to the pelvis in 35%, within the pelvis in 42%, and internal to the pelvis in 23%. In comparison, the corresponding proportions in navigation cases were 41%, 51%, and 8%, respectively. The most posterior pin in the Mako group exhibited a significantly higher rate of penetration beyond the outer cortex of the ilium compared to other pins (P < 0.05). Despite differences in pin trajectory and spatial configuration between the techniques, no pin-related complications, such as fracture, hematoma, nerve injury, or infection, were identified on postoperative CT in any case. These findings suggest that appropriate pin insertion technique is critical for achieving secure fixation and mitigating the risk of pin-related complications in computer-assisted THA.

ABSTRACT. Unicompartmental knee arthroplasty (UKA) is an established surgical option for isolated medial compartment knee disease and provides preservation of native joint structures and faster postoperative recovery compared with total knee arthroplasty. However, higher revision rates have been reported, and inaccurate implant positioning is recognized as a major cause of early failure. Robotic-assisted systems have been developed to improve the accuracy of implant placement, but the accuracy of implant placement in UKA using a handheld robotic system has not been sufficiently evaluated three-dimensionally. The purpose of this study was to perform a three-dimensional (3D) evaluation of implant placement accuracy in UKA using CORI. Sixteen consecutive knees underwent UKA using CORI between January 2024 and February 2025. Postoperative femoral and tibial component alignment in the coronal and sagittal planes was evaluated using CT-based 3D templating software. Accuracy was defined as the absolute deviation between postoperative component alignment and the intraoperative target alignment angles. Mean femoral component alignment was 0.7°±1.5° of varus in the coronal plane and 1.4°±1.7° of flexion in the sagittal plane. Mean tibial component alignment was 1.7°±0.7° of varus in the coronal plane and 8.3°±1.8° of posterior slope in the sagittal plane. Mean absolute deviations from target alignment were 1.3°±0.9° and 0.8 ±0.9° for the femoral component and 0.6°±0.7° and 1.4°±1.6° for the tibial component in the coronal and sagittal planes, respectively. No outliers exceeding 3° were observed. Accuracy of implant placement in UKA using CORI was high, even when evaluated with 3D measurement.

ABSTRACT. Accurate long-bone fracture reduction is challenging due to its reliance on repeated C-arm fluoroscopy and the difficulty of correcting rotational malalignment. While robotic systems offer improved precision, existing prototypes often utilize wired configurations that complicate sterile management and disrupt operating-room (OR) workflow. This study presents a modular, wireless fracture-reduction robot designed for practical OR integration, featuring sterilization readiness and medical-grade electromagnetic compatibility (EMC). The system comprises a 6-DOF Stewart-platform robot with an autoclavable mechanism pack and a draped, wireless driver pack. Mechanical performance was verified through bench testing of traction force, repeatability, and IEC 60601-1-2 EMC compliance. A cadaveric study using six paired tibiae compared three approaches: conventional manual reduction, robotassisted manual reduction, and automated autopilot reduction. Endpoints included alignment accuracy, reduction time, and C-arm exposure counts. Bench testing confirmed a maximum traction force of 378.6 N and repeatability within 0.3 mm and 0.2°. The system successfully passed all EMC standards. In cadaveric trials, the autopilot mode significantly improved workflow efficiency, reducing mean reduction time from 18.83 to 8.50 minutes (p=0.009) and C-arm exposure counts from 20.00 to 10.67 (p=0.006). Autopilot also achieved the lowest average rotational misalignment, although alignment differences were not statistically significant in this sample. The modular, wireless robot demonstrates high mechanical reliability and clinical practicality. By significantly reducing operative time and radiation exposure, the autopilot system addresses key barriers to the adoption of robotic fracture reduction in surgical practice.

ABSTRACT. While robotic assistance in spinal surgery is well-established for pedicle screw placement, its application in decompression procedures like laminectomy remains limited. This study aims to verify the feasibility of automated laminectomy using a collaborative robot in a laboratory setting. Seven deer vertebral specimens were potted, and 3D models were generated from CT scans. A surgical plan was created targeting a residual cortical bone thickness of 1.0 mm. A 6-axis collaborative robot equipped with a high-speed spindle and a 6-mm steel burr performed the laminectomy based on manual registration. Evaluation metrics included drilling time, residual lamina thickness (measured at 28 points), and presence of dural injury. The average drilling time was 60.3 ± 9.1 seconds. The mean residual lamina thickness was 0.85 ± 0.97 mm, compared to the planned 1.0 mm. Residual bone was confirmed in 78.6% of measurement points, while 21.4% showed over-resection (0 mm thickness). One significant deviation (4.52 mm) was observed. Importantly, no dural injury occurred in any specimen. Conclusion: This study demonstrated the feasibility of robotic laminectomy using a collaborative robot. The average error of 0.15 mm suggests that while the concept is valid, improvements in registration accuracy and the integration of force feedback are essential for future "Physical AI" applications to ensure safety.

ABSTRACT. Robotic-assisted total knee arthroplasty (TKA) is increasingly used to improve the reproducibility of component positioning, but most comparisons between platforms rely on radiographs rather than computed tomography (CT). We compared implant-positioning accuracy and early clinical outcomes between two clinically used robotic TKA workflows, Mako and ROSA, using postoperative three-dimensional (3D) CT. This retrospective single-surgeon study included 60 knees that underwent primary robotic-assisted TKA during a period when both systems were used concurrently at a single institution (30 Mako, 30 ROSA). Mako used CT-based planning with intraoperative adjustment based on soft-tissue balance, whereas ROSA used an imageless modified gap-balancing workflow. Intraoperative target angles were compared with postoperative 3D-CT measurements for femoral coronal, sagittal, and rotational alignment and tibial coronal alignment and posterior slope. Accuracy was primarily assessed using absolute differences, and variability of absolute differences was analyzed secondarily. Mako showed smaller absolute differences than ROSA in femoral sagittal alignment (1.4 ± 1.0° vs 2.2 ± 1.3°, p < 0.05) and femoral rotational alignment (1.3 ± 1.0° vs 5.9 ± 4.4°, p < 0.0001), whereas the other alignment parameters were similar. Variability was significantly greater in the ROSA group only for femoral rotational alignment (p < 0.001). At 1 year, ROM and 2011 KSS were comparable between groups. Signed-difference analysis additionally demonstrated between-workflow differences in femoral coronal, femoral sagittal, femoral rotational, and tibial coronal alignment. Postoperative 3D-CT demonstrated workflow-dependent differences in implant-positioning accuracy, most notably for femoral rotation, without between-group differences in early clinical outcomes.

ABSTRACT. Computer-assisted surgery (CAS) is widely used in total hip arthroplasty (THA) to improve acetabular component placement and biomechanical reconstruction, but its relevance to postoperative sport-inclusive activity remains unclear. This study evaluated 1-year activity after CAS-THA and examined associations with cup placement accuracy and center of rotation (COR) reproducibility. We retrospectively reviewed 196 patients (202 hips) who underwent primary THA via the anterolateral supine approach from 2021 onward. Cases were categorized by CAS modality: robot-assisted THA (Mako; R group, 73 hips), CT-based navigation-assisted THA (C group, 83 hips), and portable/AR navigation-assisted THA (P group, 46 hips). Sport-inclusive activity was assessed using the UCLA activity score preoperatively and at 1 year; improvement was calculated as postoperative minus preoperative score. Cup alignment was measured on 3-month postoperative CT (ZedHip®) in the functional pelvic plane. Accuracy was defined as absolute error from target radiographic inclination and anteversion (ΔRI, ΔRA). COR reproducibility was quantified as anteroposterior COR deviation (ΔAP, mm). Postoperative UCLA was higher in the R group (5.1±1.8) than in the C (4.5±1.2) and P groups (4.2±1.0) (p=0.04), and improvement was greater in R than P (2.8±1.7 vs 1.6±1.1; p<0.03). ΔRI and ΔAP were smallest in R (both p<0.01), whereas ΔRA did not differ. UCLA improvement correlated with ΔAP (r=−0.62, p=0.01) but not with ΔRI/ΔRA. COR reproducibility, rather than cup angle error, may be a key determinant of postoperative sport-inclusive activity after CAS-THA

ABSTRACT. Background: Computer-assisted navigation (CAN) improves glenoid implant positioning in total shoulder arthroplasty (TSA), but its accuracy in severe glenoid deformities requiring augments remains understudied. This study evaluated CAN precision in replicating preoperative plans for version, inclination, and entry point in TSA with augmented implants.

Methods: Retrospective analysis of 19,054 navigated TSA cases (2,266 aTSA, 16,788 rTSA) using the Advita GPS system. Glenoid deformities were stratified by glenoid retroversion (≥15° or <15°) and glenoid inclination (≥10° or <10°). Primary outcomes were deviations from preoperative plans in version, inclination, and entry point, analyzed by de formity severity and augment type (posterior augment, superior augment, posterosuperior augment).

Results: In rTSA, glenoids with glenoid retroversion ≥15° showed slightly greater version (0.35° vs. 0.33°, p=0.0057) and entry point (1.24 mm vs. 1.18 mm, p=0.0001) deviations. Posterior augment had greater version deviation than superior augment (0.35° vs. 0.32°, p<0.0001). In aTSA, no significant differences were observed across glenoid retroversion subgroups. All deviations remained submillimetric or subdegree, indicating high precision.

Conclusions: CAN enables precise intraoperative replication of preoperative plans regardless of glenoid deformity severity in TSA thanks to augmented components. Statistical differences in deviations were clinically negligible, supporting CAN reliability even in complex reconstructions.

ABSTRACT. Background: Navbit Sprint is a portable navigation system for total hip arthroplasty (THA) that constructs a three-dimensional coordinate system by laterally tilting the operating table without anterior superior iliac spine registration. The purpose of this study was to compare the accuracy of cup placement in THA between Navbit Sprint and an augmented reality (AR) CT-based navigation system (HoloNavi One).

Methods: Thirty-four primary cementless THAs were retrospectively analyzed. All procedures were performed via an anterolateral approach in the supine position. Using HoloNavi One, pelvic coronal tilt was corrected to 0°, and sagittal and axial pelvic tilts were recorded. Thereafter, table tilting was performed to complete Navbit Sprint registration. Acetabular cups were placed using HoloNavi One and measured using Navbit Sprint. Postoperative CT was used to measure radiographic inclination (RI) and anteversion (RA). Navigation errors were defined as absolute differences between intraoperative navigation values and postoperative CT measurements. Correlations between navigation errors of Navbit Sprint and pelvic tilt at registration were evaluated.

Results: The mean postoperative RI and RA were 39.3° and 15.4°, respectively, and all cups were within the Lewinnek safe zone. Navigation errors for HoloNavi One were 1.9° for both inclination and anteversion, whereas navigation errors of Navbit Sprint were 2.6° and 4.3°, respectively. The anteversion error of Navbit Sprint was significantly larger than that of HoloNavi One and showed negative correlation with axial pelvic tilt.

Conclusions: Navbit Sprint demonstrated inferior anteversion accuracy compared with AR CT-based navigation. Axial pelvic tilt during registration may affect anteversion accuracy with Navbit Sprint.

ABSTRACT. This study investigated whether intraoperative navigation in reverse total shoulder arthroplasty enables screw placement as planned preoperatively. A total of 82 shoulders treated with the same implant system were included. Patients were divided into two groups: 44 shoulders without navigation (non-navigation group) and 38 shoulders with navigation (navigation group). Using CT-based preoperative planning, the planned lengths of the superior and inferior screws were compared with the lengths of the screws actually used in each group. In the preoperative plan, the superior screw length was 27.2 ± 4.4 mm in the non-navigation group and 25.9 ± 4.4 mm in the navigation group, and the inferior screw length was 28.3 ± 4.7 mm and 27.3 ± 4.8 mm, respectively, with no significant differences between the two groups. The actual screw lengths used were 21.7 ± 5.8 mm (non-navigation group) and 24.8 ± 6.3 mm (navigation group) for the superior screw, and 28.0 ± 6.6 mm (non-navigation group) and 28.0 ± 4.7 mm (navigation group) for the inferior screw. The superior screw was significantly shorter in the non-navigation group than in the navigation group, whereas no significant difference was observed in the inferior screw length.

ABSTRACT. Introduction: Femoroacetabular impingement syndrome (FAIS) requires precise cam resection to restore hip range of motion while avoiding under- or over-resection. Conventional freehand arthroscopic techniques depend on fluoroscopy and surgeon experience, which may result in discrepancies between preoperative planning and intraoperative execution. CT-based computer navigation enables quantitative planning and real-time guidance; however, its impact on the accuracy of surgical execution relative to the preoperative plan remains unclear. Methods: A retrospective comparative study was performed in patients with cam- or combined-type FAIS who underwent arthroscopic osteochondroplasty with or without CT-based computer navigation. Preoperative impingement points and virtual osteochondroplasty were defined using CT-based three-dimensional range-of-motion simulation. In the navigation-assisted group, planned resections were transferred intraoperatively for real-time guidance. Achievement of the preoperative plan was evaluated using postoperative CT-based simulations at 90°, 70°, and 45° of hip flexion. Patient-reported outcome measures were assessed as secondary outcomes. Results: Fifty-five hips were analyzed. Navigation-assisted osteochondroplasty demonstrated significantly higher achievement of the preoperative plan at 90° (92.0% vs 46.7%, P<0.001) and 70° (80.0% vs 50.0%, P=0.027) of hip flexion compared with the freehand technique, with no significant difference at 45°. At 1 year, the navigation group achieved higher Non-Arthritic Hip Scores and greater rates of clinically meaningful improvement, while differences in other clinical outcomes were limited. Conclusion: Computer navigation—assisted osteochondroplasty may improve the accuracy of cam resection and contribute to better short-term outcomes such as the NAHS at 1 year.

ABSTRACT. Purpose: Curved varus osteotomy (CVO) is a joint-preserving procedure for osteonecrosis of the femoral head (ONFH), requiring precise 3D transposition of the necrotic lesion. This study aimed to determine whether a CT-based navigation improves osteotomy accuracy compared to conventional fluoroscopic technique. Methods: We retrospectively reviewed 26 hips with Type C1 ONFH treated with CVO. The Patients consisted of two groups: the conventional group (16 hips) and the navigation group (10 hips). We evaluated the postoperative intact ratio, which is defined as the proportion of the transposed intact articular surface of the femoral head to the weight-bearing acetabular surface, identifying cases falling below the 34% threshold necessary to prevent collapse as outliers, the accuracy of the osteotomy on the axial plane relative to the femoral neck axis, and the accuracy of the varus angle using postoperative 3D CT images. Results: While the postoperative intact ratio wasn’t significantly differences between the groups (Conventional: 36.8% vs. Navigation: 44.2%, P = 0.24), the variance was significantly reduced in the navigation group (P = 0.009). Notably, 38% of the conventional group were outliers (< 33.6% intact ratio), whereas no outliers were found in the navigation group. The navigation group achieved significantly higher accuracy in the axial osteotomy angle relative to the femoral neck axis (Conventional group: 16.7° vs. Navigation group: -1.0°, P < 0.001). Conclusion: CT-based navigation system significantly reduces variability and eliminates outliers in CVO. By integrating 3D planning with precise osteotomy, navigation transforms CVO into a highly predictable and reproducible procedure.

ABSTRACT. Virtual surgical planning and 3D-printed cutting guides (PCGs) have improved operative times and bony union in mandibular reconstruction, which is critical for restoring jaw function and facial aesthetics following oncological resection. However, current systems have limited flexibility to accommodate intraoperative changes arising from tumour growth or cut errors. We are designing a “day-of-surgery” image-guided system to adjust the surgical plan in response to such changes. The system was previously tested in a cadaver trial and showed comparable accuracy to guide-based reconstructions, but took significantly longer.

To improve operative efficiency, we made targeted refinements to the system, including automating the surgical planning and registering the cut planes with a planar tool. Here, we report on the performance of the refined system in a surgical simulation using anatomical models. We performed five two-segment central or right mandibular reconstructions on Sawbone models spanning the stages from preoperative preparation through cutting the fibula segments. Workflow tasks were timed, and the resulting segments were scanned. Procedure time decreased by 86.5 minutes compared to a cadaver study and 8.5 minutes compared to a previous Sawbone model study. However, it still took 14 minutes to position the cutting guides and the bone cut angles deviated significantly from the targeted values (4.1±5.2 degrees), which indicates that finer incremental control over the cutting guide position is necessary prior to clinical evaluation.

ABSTRACT. High plantar pressures (PP) are associated with pain, impaired mobility, and tissue breakdown, playing a central role in foot ulcer development. This could be prevented provided the plantar pressure (PP) can be predicted quantitatively. Traditionally, high-fidelity finite element (FE) models achieve accurate biomechanical predictions. However, they require extensive manual preparation and long computation times, limiting routine clinical use. The aim of this study was to develop and evaluate a clinically feasible workflow that automatically generates a targeted FE foot model from non-weight-bearing (non-WB) MRI and transforms it into a weight-bearing (WB) configuration for PP simulation. The FE Model was constructed by representing all bones as a lumpsum rigid structure and the plantar soft tissue as a homogeneous bulk material with mesh resolution optimized for computational efficiency. The model was evaluated on a healthy subject by comparing simulated PP distribution with actual pressure plate measurements, focusing on both spatial and amplitude agreement of peak pressures across forefoot, midfoot, and hindfoot regions. Results showed that FE Model consistently localized high-pressure regions within correct anatomical subregions, demonstrating good spatial agreement. However, simulated peak amplitudes deviated from actual pressure plate measurements. Total simulation time was under five minutes on standard laptop hardware. These findings indicate that targeted FE models derived from routinely acquired MRI may provide rapid, clinically relevant insights into PP distribution. The workflow supports potential integration in clinical workflow, while emphasizing the need for further refinement and larger-scale validation.

ABSTRACT. Posttraumatic Segmental bone defects of the femur and tibia pose major challenges for joint-sparing reconstruction. Patient-specific implants (PSIs), enabled by 3D imaging and additive manufacturing, offer tailored solutions, but clinical evidence remains fragmented. This review systematically evaluated PSI applications for femoral and tibial shaft reconstruction and introduced a classification framework to standardize reporting.

A systematic search of PubMed and Embase identified 53 studies (317 patients) reporting joint-sparing reconstructions using PSIs in the lower extremities. Data were extracted on implant design, fixation method, material, biological augmentation, and outcomes. Implants were categorized by defect location (unicortical, diaphyseal, meta-extended) and fixation type (screws, plates, nails, stems, hybrid). Kaplan–Meier analysis assessed five-year mechanical implant survival across fixation subgroups.

Titanium was the predominant material, often with porous lattice structures to support bone ingrowth. Hybrid fixation demonstrated the highest five-year survival (97%), followed by nails/stems (95%) and plate-only fixation (89%). Structural failure was the most common complication, while most implants remained failure-free at last follow-up. The classification system enabled structured comparison across heterogeneous reports.

ABSTRACT. Lower trapezius tendon (LTT) transfer can restore shoulder function, particularly active external rotation (ER) following massive rotator cuff tears (MRCT). However, the optimal graft type and its transfer location on the greater tuberosity remains unclear. This biomechanical study aimed to investigate the optimal graft type and location in LTT transfer for MRCT. Eight fresh frozen cadaver shoulders were tested on a shoulder simulator. LTT transfer was performed with an Achilles tendon fixed over the superior-middle facets (LTT-Achilles), a semitendinosus (ST) tendon to the superior facet (LTT-ST-S), or to the middle facet (LTT-ST-M). A 24N load was applied to each transferred graft. Under each condition (intact rotator cuff, MRCT, LTT-Achilles, ST-S, and ST-M), humeral head translation and functional abduction force (FAF) were evaluated at 0°, 30°, and 60° of glenohumeral elevation. ER torque was assessed across five rotational angles at 0°, 30°, and 60° of glenohumeral elevation. None of the LTT conditions significantly depressed humeral head superior migration as compared to MRCT. Compared to MRCT, FAF improved significantly with LTT-Achilles at 0° elevation and LTT-ST-S at 0 and 30° elevation. LTT-ST-M improved ER torque compared to MRCT at 0° elevation and at 30° elevation. LTT-ST-M showed higher ER torque than both LTT-Achilles and LTT-ST-S at 0° elevation. In LTT transfer for MRCT, LTT-ST-M most effectively restored ER torque, whereas LTT-Achilles and LTT-ST-S improved FAF. None of the conditions of LTT transfer suppressed humeral translation. Graft selection and transfer location in LTT transfer can be tailored to patient goals and graft availability.

ABSTRACT. In reverse shoulder arthroplasty (RSA), preoperative planning for implant placement often focuses primarily on the bony morphology, with insufficient consideration regarding the dynamic and complex interactions between the implant position and the surrounding soft tissues. This study aimed to determine the optimal implant position in RSA, taking into account the residual rotator cuff condition. Eight cadaveric shoulders were mounted on a biomechanical simulator with a custom RSA implant. The glenoid baseplate allowed adjustment of the anterior–posterior glenosphere offset. Functional abduction force (FAF), axial rotation torque, and corresponding joint reaction force (JRF) were evaluated. Joint stability on the glenoid side was assessed using the shear-to-compressive (S/C) ratio. FAF was measured with four rotator cuff conditions: anterior/posterior cuff intact, anterior cuff intact, posterior cuff intact, and complete cuff tear. Internal rotation (IR) torque was measured with pectoralis major and latissimus dorsi/teres major load, with either the anterior cuff intact/torn. External rotation (ER) torque was measured with a posterior cuff load. FAF and S/C ratio during abduction and rotation showed only minimal effects of anterior–posterior glenosphere offset, regardless of the residual rotator cuff condition. In contrast, rotational torque showed a trend, with posterior offset favoring IR and anterior offset favoring ER. These findings indicate that adjusting anterior–posterior glenosphere offset according to patient rotator cuff condition and functional demands may enable more patient specific optimization of treatment, and that implant positioning should consider rotator cuff status in addition to bony factors such as impingement-free range of motion.

ABSTRACT. Open wedge high tibial osteotomy (owHTO) aims to redistribute tibiofemoral loading in patients with medial compartment osteoarthritis; however, substantial inter-individual variability in postoperative outcomes persists despite alignment-based planning. To explore the contribution of bone morphology, a normative morphological reference space was constructed using principal component analysis (PCA) of ten normalized tibial and femoral shape parameters derived from 92 healthy cadaveric knees. Eleven owHTO knees were projected into this space after z-standardization, and multivariate deviations were quantified using squared Mahalanobis distances. PCA revealed a consistent shift of owHTO knees towards positive values of the third principal component (PC3), reflecting sagittal and coronal tibial morphology. Pronounced multivariate deviations occurred when this shift was combined with additional alterations along PC2 (tibial slope asymmetry) or PC1 (global shape proportions). Three of the eleven owHTO knees exceeded the 95 % normative threshold. These findings indicate that, apart from alignment parameters, combinations of tibial and femoral shape characteristics may influence the patient specific biomechanical response to owHTO.

ABSTRACT. Medial congruent (MC) bearings in Total Knee Arthroplasty (TKA) aims to mimic the native knee, by introducing further posterior congruency in the design of the medial compartment. Our primary aim was to compare dwell point of the medial and lateral femoral condyle during step-up and lunge motions, in patients operated with an MC and a cruciate retaining (CR) bearing. Additionally, clinical outcomes of the groups were compared at time of the movement analysis. In a double-blinded RCT, 52 (mean age 70 ± SD 8 years, 63% female) of originally 60 patients were evaluated during step-up and lunge motions, using a single-plane, 15 Hz fluoroscopy, 1 year post-surgery. Step-up went from >90° flexion to full extension. Lunge went from full extension to deepest possible flexion. Dwell point of the medial and lateral femoral compartment, expressed in mm from center of the tibial bearing, was calculated using a 2D to 3D approach, utilizing a CAD model of the actual implants. Dwell point positions were compared using one-dimensional statistical parametric mapping. We measured clinical outcomes of the TKA surgery with Oxford Knee Score (OKS) and Forgotten Joint Score (FJS) 1 year post-surgery. Primary outcome was available for 52 patients (MC:28, CR:24). Lateral dwell point was significantly different from 58-100% of the motion, medial dwell point from 0-100% (p< 0.001 in both cases). During lunge, we saw same posterior shift in dwell point, however, bigger variation led to significance being more inconsistent. The differences in lateral dwell point occurred in the beginning of the lunge (0-26%). Medial dwell point differed significantly through the entire motion, except for small fluctuations in significance from 52-60% of the motion. We saw overlapping 95% confidence intervals in OKS and FJS. We found a posterior placement of the lateral and medial femoral compartment during step-up and lunge, when comparing patients with MC to patients with a CR bearing. An MC bearing seems to mimic the movement of the native knee better than the CR bearing.

ABSTRACT. Mandibular reconstruction, following tumor-related resection, is essential for restoring form and function of the jaw by replacing the removed segment with a donor bone. Achieving bony union is critical; however, mechanisms affecting union rates remain poorly understood. Recent computational work has demonstrated that bone cut angle significantly affects predicted bone union propensity (BUP) when using full-length reconstruction plates. Mini-plate-based fixation offers a promising alternative to full-length plates, however, the optimal bone cut angles for these reconstructions remains unknown. The purpose of this study is to determine whether variations in bone cut angle has a significant effect on predicted BUP when using mini-plates. Using a physics-based simulation and finite-element modelling, the cut angle was varied, covering a total of 9 cases including the baseline conventional cut angles as reference. The results indicated that varying cut angle can increase the BUP metric from a baseline value of ~23% to as much as ~42%. Across all configurations, apposition followed a consistent two-peak pattern, occurring at maximum mandibular opening and during bolus engagement. Results demonstrated that bone cut angle has a significant and direction-dependent effect on predicted BUP. Specifically, negative left pitch, positive right pitch, negative right yaw, and positive left yaw increased overall apposition at both interfaces. While limited to a single defect model and one-segment reconstructions, this study extends prior findings on angled osteotomies to mini-plate fixation. Overall, optimizing donor bone cut angle represents a promising strategy for improving predicted bone union in mini-plate-based mandibular reconstruction.

ABSTRACT. Background: AR-Hip is a smartphone-based AR navigation system for total hip arthroplasty (THA), launched in Japan in 2021. Over 45,000 procedures have used this system, with international rollout beginning in 2025. Despite previous studies and reports, we performed a multicenter randomized evaluation of AR-Hip for lateral-decubitus THA. This report presents the interim results of that study. Methods: At each participating institution, patients were allocated to either AR-Hip–assisted THA (10 cases) or THA performed using a conventional mechanical alignment guide (10 cases). Across five institutions, a total of 100 hips were enrolled; of these, 64 hips have been analyzed to date. Accuracy was evaluated using postoperative CT images obtained 3 months after surgery. Absolute errors from the target cup orientation were measured using three-dimensional templating software (ZedHip). Results: In the manual alignment group (n = 31), the mean absolute error was 3.9 ± 2.2° for Radiographic inclination (RI) and 6.5 ± 4.7° for Radiographic anteversion (RA). In the AR-Hip group (n = 33), the mean absolute error was 2.3 ± 1.7°for RI, and he mean absolute error was 2.7 ± 2.8°for RA. Significant differences were observed between the two groups for both inclination (p = 0.003) and anteversion (p < 0.001). No complications such as dislocation, infection, or fracture were observed in either group. Conclusions: Consistent with previous reports, AR-Hip demonstrated a placement accuracy of approximately 3 degrees. Although data demonstrating direct clinical superiority remain limited, AR-Hip appears to be a practical and reliable tool functioning as a “safety belt” in THA, supporting accurate component placement in routine clinical practice.

ABSTRACT. The purpose of this study was to compare the accuracy of acetabular cup placement in total hip arthroplasty (THA) between a CT-based AR navigation system and a pinless portable AR navigation system used simultaneously during surgery. Both navigation systems can be used with acetabular components from multiple manufacturers. This study included 43 hips in 43 patients who underwent primary cementless THA in the supine position. Acetabular cup placement was guided by the CT-based AR navigation system, after which the pinless portable navigation system was additionally applied. Target angles for cup placement were set at 40° for radiographic inclination (RI) and 15° for radiographic anteversion (RA). Navigation error was defined as the absolute difference between intraoperative navigation values and postoperative CT measurements and compared between groups. Intraoperative RI and RA values were 40.8° ± 2.1° and 14.6° ± 1.8°, in the CT-based group, and 39.0° ± 3.7° and 11.5° ± 4.1° in the pinless portable group. Postoperative RI and RA values were 39.8° ± 3.6° and 14.3° ± 3.5°. RI navigation errors were similar (2.5° ± 2.1° in the CT-based group vs. 2.9° ± 2.5° in the pinless portable group; not significant). In contrast, RA navigation error was significantly smaller in the CT-based group (2.3° ± 1.9°) than in the pinless portable group (3.4° ± 2.5°) (p < 0.05). In conclusion, both navigation systems demonstrated good accuracy in RI during supine THA. However, the pinless portable AR navigation system showed inferior accuracy in RA compared with the CT-based AR navigation system.

ABSTRACT. Planning orthopedic tumor surgery requires significant mental effort to translate 3D plans and decisions from 2D preoperative images into the patient’s actual anatomy. Mixed Reality (MR), an immersive technology that merges real and virtual environments, allows users to interact with digital objects. This study builds on our earlier proof-of-concept by including more cases involving bone tumors and other orthopedic tumors. From July 2021 to November 2025, we retrospectively reviewed 72 patients with bone or soft tissue tumors or those with unsuccessful tumor reconstruction surgeries. An MR platform was developed to generate holographic applications tailored for each case. The surgeon performed surgical planning using both traditional 2D methods and MR 3D holograms on the HoloLens 2. For each approach, the surgeon completed a qualitative survey: 1) a Likert-scale (LS) questionnaire evaluating spatial awareness of bone structures and the effectiveness of surgical planning, and 2) a NASA-TLX score assessing cognitive workload. The LS questionnaire revealed that the MR 3D hologram group outperformed the 2D group in all aspects of spatial awareness related to pathoanatomy and was regarded as more effective for preoperative planning. NASA-TLX scores indicated that the overall cognitive workload during preoperative assessment was lower in the MR 3D hologram group. Expanding on a previous MR proof-of-concept study in orthopedic oncology, this largest series suggests that the technology may also benefit patients with soft-tissue sarcomas, benign bone tumors, and revision surgeries. With reduced cognitive effort and improved ergonomics, surgeons can focus more on patient care and procedures.

ABSTRACT. Purpose: Total hip arthroplasty (THA) for severe acetabular defects presents considerable technical challenges and high complication risks. While portable augmented reality (AR) navigation is effective in cementless THA, its utility in complex reconstructions using reinforcement devices and cemented dual mobility (DM) cups remains unexplored. This study evaluated the precision of a portable AR system in such cases. Methods: This study included 19 patients underwent THA for severe acetabular defects using AR navigation and cemented DM cups, compared to 19 manual cemented cup THA patients as control group. Accuracy was assessed by comparing intraoperative AR measurements with postoperative 3D-CT data. Outcomes included accuracy of cup placement and postoperative complications including dislocation rates. Results: The AR group showed high precision, with mean absolute discrepancies of 2.2° (SD 1.6) for inclination and 3.7° (SD 2.2) for anteversion. The AR group achieved a higher safe zone rate (89% vs. 73%) and a significantly lower dislocation rate (0% vs. 31%, p < 0.05) than the manual group. Other clinical outcomes, including operative time and modified Harris Hip Score, were comparable between the two groups. Conclusion: Our findings suggest that portable AR navigation potentially offers high precision and safety for complex acetabular reconstructions, which could enhance component placement accuracy and minimize postoperative dislocations in challenging surgical scenarios.

ABSTRACT. Background: Accurate glenoid component placement is essential to the success of total shoulder arthroplasty, yet it remains technically demanding. Augmented-reality (AR) navigation using a head-mounted display (HMD) provides real-time intraoperative feedback and may improve precision. This study evaluated the accuracy and technical impact of HMD-guided navigation by comparing intraoperative estimations with postoperative CT measurements. Methods: Thirteen glenoid pin-placement procedures were performed on cadaveric shoulders using an HMD-based AR navigation system. Intraoperative values for entry point, version, and inclination were recorded, and the final pin position was assessed on postoperative CT using the same parameters. Agreement between intraoperative and postoperative measurements was analyzed using Bland–Altman plots and complementary predictive probability analyses. Results: Postoperative CT scans showed a mean global entry-point deviation of 1.98 mm, with mean angular deviations of 1.60° for version and 1.67° for inclination. Intraoperative HMD values were consistently smaller (0.69 mm for entry point, 1.08° for version, 0.85° for inclination). Bland–Altman analysis demonstrated small systematic biases but noticeable dispersion at the individual-case level, indicating that intraoperative values did not reproduce postoperative deviations exactly. Predictive analyses showed that when the HMD deviation was ≤ 1 mm, the CT deviation remained ≤ 2 mm in most linear measurements, and when the HMD deviation was ≤ 2°, the CT deviation was ≤ 4° in the majority of angular measurements. These findings support a probabilistic intraoperative “safe zone,” in which small deviations displayed by the HMD correspond to a high likelihood of acceptable postoperative alignment. Conclusion: The HMD provided accurate guidance with limited residual error, but small intraoperative deviations could not be used to predict exact postoperative values. A tolerance-based safe-zone approach (≤ 1 mm or ≤ 2°) may prevent unnecessary corrections that risk weakening fixation while preserving surgical precision and efficiency.

ABSTRACT. Introduction Optimizing glenoid component positioning is a key determinant of stability, function and implant survival in total shoulder arthroplasty. Postoperative computed tomography (CT) is considered the reference method for assessing guide-pin or implant alignment, but it remains a retrospective evaluation and is often difficult to obtain in cadaveric research settings. Optical navigation systems such as the NDI Polaris provide immediate intraoperative geometric measurements and may represent a pragmatic alternative. This cadaveric validation study aimed to evaluate the agreement between NDI Polaris optical-navigation measurements and postoperative CT analysis for glenoid guide-pin positioning. Methods Eight cadaveric shoulders underwent augmented-reality–guided placement of a glenoid guide-pin, followed by postoperative CT acquisition. Intraoperative optical-navigation measurements were recorded using an NDI Polaris tracking station. Two accuracy domains were analysed relative to CT: (1) linear entry-point deviation (mm) and (2) angular deviation in version and inclination (degrees). Agreement between both measurement modalities was assessed using Bland–Altman analysis, complemented by descriptive accuracy statistics. Results Sixteen paired measurements were available for comparison. The mean absolute NDI–CT difference was 0.98 mm (SD 0.55; range 0.15–2.17 mm) for entry-point localisation and 1.58° (SD 0.97; range 0.03–3.28°) for trajectory orientation. Bland–Altman analysis demonstrated a small systematic bias for both linear (−0.15 mm) and angular (−0.75°) parameters, with most discrepancies contained within approximately ±2 mm and ±4°, respectively. These results suggest that, under standardised acquisition conditions, optical navigation provides geometric measurements that closely approximate CT-based reference values and may serve as a practical surrogate in cadaveric research when CT access is limited. Conclusion NDI Polaris optical tracking demonstrated good agreement with postoperative CT for evaluating glenoid pin positioning in this cadaveric model. Although the small sample size warrants cautious interpretation, these findings support the potential use of optical navigation as a complementary or alternative accuracy-assessment tool, pending confirmation in larger validation studies.

ABSTRACT. Background : Functional alignment (FA) in total knee arthroplasty (TKA) aims to restore individual knee anatomy and preserve periarticular soft tissues, whereas mechanical alignment (MA) pursues neutral hip-knee-ankle (HKA) axes. Femoral component rotation (FCR) is a critical determinant of knee kinematics, patellar tracking, and functional outcomes. Objective : This study compares FCR and clinical outcomes between FA-TKA and MA-TKA in bilateral TKA patients. Methods : A prospective randomized controlled trial enrolled 37 patients with primary bilateral knee osteoarthritis. Patients underwent bilateral TKA with one knee randomized to FA and the contralateral knee to MA, all performed using robotic-arm-assisted surgery with patient and outcome assessor blinding. The primary outcome was FCR measured on postoperative CT scans. Secondary outcomes included pain (VAS), knee function (KOOS), and patient-perceived joint awareness (Forgotten Joint Score, FJS). Result : FCR in the FA group was significantly lower than in the MA group (2.50±1.74° vs. 3.65±1.38°, p=0.002), indicating FA achieves more conservative external rotation to optimize flexion-gap balance. At 6-month follow-up, FJS scores were significantly higher in the FA group (66.92±9.08 vs. 62.25±6.90, p=0.015), suggesting improved patient-perceived natural knee function. Both groups demonstrated significant pain reduction and functional improvement on KOOS from immediate postoperative to 6-month assessment, with no statistically significant differences. Conclusion : Robot-assisted FA permits precise personalized FCR adjustment within a safe anatomical range and achieves superior patient-perceived functional outcomes compared to MA at 6-month follow-up, despite equivalent pain relief and objective knee function.

ABSTRACT. Restricted kinematic alignment (rKA) has been developed as an alternative to mechanical alignment (MA) in total knee arthroplasty (TKA) that aims to restore native knee alignment. However, its effect on patellofemoral alignment has not been studied before. This study compared patellofemoral radiographic parameters and early clinical outcomes between rKA and MA in bilateral robotic-assisted TKA. A prospective, double-blinded randomized controlled trial was conducted in patients undergoing simultaneous bilateral robotic TKA. Each patient received rKA in one knee and MA in the contralateral knee. Radiographic patellar tilt angle and lateral patellar shift were measured. Postoperative clinical outcomes were assessed at 3 and 6 months including Visual Analogue Scale (VAS), Knee Injury and Osteoarthritis Outcome Score (KOOS), and Forgotten Joint Score (FJS). Twenty patients (40 knees) were included. The rKA group demonstrated significantly greater patellar tilt compared with the MA group (median 5.0°(2.0-8.0) vs 3.0°(2.0-6.0), p = 0.007). Lateral patellar shift was also higher in the rKA group (median 3.9 mm (2.0-8.1) vs 2.8 mm (1.3-5.1), p = 0.017). There were no significant differences in VAS pain scores or KOOS between groups at all follow-up periods. At 6 months postoperatively, the rKA group showed a significantly higher FJS than the MA group (p = 0.018). In robotic-assisted TKA, rKA resulted in increased patellar tilt and lateral patellar shift compared with MA; however, these radiographic differences did not adversely affect early clinical outcomes. rKA demonstrated superior functional outcomes compared to mechanical alignment at short-term follow-up.

ABSTRACT. Aims To classify intraoperative dynamic hip–knee–ankle alignment (dHKA) using unsupervised machine learning (ML) and to determine how often a patient’s post cut alignment phenotype matches their pre cut phenotype in a tibia first total knee arthroplasty (TKA) workflow.

Methods We analyzed 1,890 tibia first TKAs performed by 11 surgeons, excluding flexion contractures. An intra articular, force-controlled distraction device integrated with computer assisted orthopaedic surgery (CAOS) recorded varus–valgus (VV) angles at 12 flexion positions (0°–120°) before the femoral cut (pre cut) and after the femoral cut with trials (post cut). Each case was represented as a vector of VV angles. K means clustering was trained on pre cut data; internal validity metrics selected the optimal feature subset and number of clusters. The trained model labelled post cut data to quantify phenotype preservation overall and by surgeon.

Results An eight-angle feature set (10°, 20°, 30°, 45°, 60°, 75°, 90°, 105°) and four distinct phenotypes were optimal. Centroid trajectories characterized phenotypes as valgus/neutral, neutral, low to moderate varus, and moderate to high varus. Pre cut vs post cut distributions were 15.3%/14.2%, 35.9%/30.1%, 34.2%/35.6%, and 14.5%/20.2%, respectively. Phenotype preservation was 69.4% overall, with surgeon specific rates 61%–88%. Transitions occurred predominantly between adjacent phenotypes; extreme shifts were rare.

Conclusion Intraoperative, force controlled dHKA measurement combined with unsupervised ML yields reproducible, patient specific dynamic alignment phenotypes and quantifies the surgical impact on alignment within a tibia first TKA workflow. This approach provides real time feedback and a foundation for automated intraoperative classification to support personalized alignment strategies and potentially improve functional outcomes.

ABSTRACT. Kinematic alignment (KA) total knee arthroplasty (TKA) has demonstrated favorable early clinical outcomes compared with mechanical alignment; however, concerns persist regarding coronal plane alignment outliers, particularly excessive tibial varus. Restricted KA has been proposed to address these concerns, but its clinical impact within robotic-assisted workflows remains unclear. This retrospective study included 75 patients (80 knees) who underwent primary robotic-assisted TKA using the ROSA Knee System between June 2024 and September 2025. All procedures were performed using a femoral-first, calipered technique with tibial resection based on a soft-tissue respecting approach. During the first half of the study period, knees with intraoperatively assessed tibial varus angles exceeding 3° were adjusted to ≤3° and classified as the restricted KA (rKA) group. Knees with tibial varus angles ≤3°, as well as those treated without restriction in the latter half of the study period, were classified as the unrestricted KA (uKA) group. Early postoperative outcomes, including time to cane-assisted ambulation, range of motion, operative time, blood loss, and postoperative complications, were compared. Thirty-six knees were assigned to the uKA group and 44 knees to the rKA group. No significant differences were observed in days to cane ambulation (uKA 9.7±3.3 vs rKA 9.8±3.3 days), range of motion at postoperative week 1 or week 2, operative time, or blood loss. No postoperative complications occurred in either group. Restricting tibial varus to ≤3° in robotic-assisted KA TKA did not adversely affect early functional recovery or short-term clinical outcomes, supporting the feasibility of incorporating a tibial varus “safe zone” within robotic KA workflows.

ABSTRACT. Introduction An international Delphi consensus recently defined fundamental principles for a 3D leg alignment analysis framework, but its clinical implications and comparability with established 2D alignment parameters remain unclear. This study aimed to quantify agreement and systematic differences between knee-related leg alignment parameters measured using the conventional 2D framework and a consensus-based 3D framework. Methodology Patients were retrospectively included if they underwent both 2D radiographic and 3D CT-based alignment analysis between 2022 and 2025. Coronal, sagittal, and axial alignment parameters relevant for knee osteotomy planning were measured on the affected limb using a Paley-based 2D workflow and two distinct 3D workflows adhering to the Delphi consensus framework. Systematic differences between methods were assessed. Results 61 patients were included, of whom 41 had complete sagittal plane radiological data. Significant differences between 2D and 3D measurements were found for 10 of 14 alignment parameters (71%). Six parameters demonstrated systematic differences of at least 1° (MPTA, mPPTA, mMPPTA, aPPTA, aMPPTA, and TTA). One parameter (FVA) differed by more than 0.5° between the two 3D workflows. Conclusion Although many knee-related alignment parameters differed statistically between 2D and 3D frameworks, clinically relevant differences were mainly observed for tibial alignment parameters. Existing 2D-based normal values may therefore not be directly applicable to 3D clinical decision-making, emphasizing the need for 3D-specific reference standards.

ABSTRACT. Background: Precise registration of the femoral mechanical axis (FMA) serves as the foundation for successful alignment and optimal gap balancing in Total Knee Arthroplasty. Imageless robotic systems depend on the registration of intraoperative bony landmarks; however, the ideal distal exit point for the FMA registration remains a subject of debate.

Objective: To evaluate the accuracy of FMA registration by comparing distal femoral resection thickness differences (Medial - Lateral of the distal femoral resection thickness) between the traditional center of the intercondylar notch and a 5-mm medial offset approach, using preoperative X-ray plans as the control.

Methods: The prospective multi-center comparative study involved 53 Imageless Robotic-assisted TKA cases across two institutions: Veterans General Hospital and Prince of Songkla Hospital. Preoperative long-leg standing radiographs established the control "Resection Difference" (Medial - Lateral of the distal femoral resection thickness). Intraoperatively, the FMA was registered at two distinct points: the geometric center of the notch and a 5-mm medial offset. The primary metric was the absolute error of the intraoperative resection difference compared to the preoperative plan. Statistical analysis was performed using the Wilcoxon signed-rank test.

Results: The 5-mm medial offset registration was closer to the preoperative plan in 83.02% of cases. The preoperative control group showed a mean resection difference of +0.79 mm. The Center of Notch group demonstrated a significant systematic shift toward a negative resection difference (-0.88 mm), indicating a pronounced valgus bias. Conversely, the 5-mm medial offset yielded a mean difference of -0.29 mm, which was significantly closer to the anatomical baseline (p=0.000132).

Conclusion: For imageless robotic-assisted TKA, registration using a 5-mm medial offset is significantly more accurate than using the notch center itself. This technique reduces systematic valgus bias and ensures that the distal resection difference more closely reflects the preoperative anatomical plan.

ABSTRACT. Comprehensive evaluation of gait-related determinants remains limited. We aimed to elucidate gait characteristics and their associated determinants in patients with knee osteoarthritis (KOA). Twenty-seven KOA patients were enrolled. Preoperative gait parameters: gait speed, stride length, step width, cadence, and double support ratio, were assessed using a smartphone-based markerless motion capture system, “OpenCap.” As part of the patient background assessment, passive knee range of motion (ROM), the Numeric Rating Scale, and the Knee injury and Osteoarthritis Outcome Score were measured. The results indicated that hip and knee joint maximum ROM during gait, as well as the degree of knee flexion contracture, were significantly correlated with a broad range of gait parameters. Conversely, patient-reported outcomes showed no significant correlations with gait parameters. The findings support the potential effectiveness of kinematics-focused interventions for improving gait in patients with KOA, compared with approaches that primarily target pain or satisfaction.

ABSTRACT. Progress in three-dimensional (3D) printing technology has expanded various optional surgical methods for orthopaedic surgeons, such as surgical planning and creating patient specific instruments. However, a strategy for intramedullary pinning of femoral fractures using 3D printers has seldom been reported. Case presentation: A 5-year-old female presented with right femoral diaphyseal fracture caused by indoor fall. The patient had an anterior bowing deformity of the right femur associated with osteogenesis imperfecta. A 3D-printed femoral model was created prior to the surgery, to plan the osteotomy site and the entry point for intramedullary pin fixation. Planned deformity correction and pin fixation were successfully performed intraoperatively. At the 2-year postoperative follow-up, the patient showed favorable outcome without requiring reoperation. Our case demonstrated the usefulness of 3D printing in surgery, showing that the surgeon could recognise reliable 3D orientation for the smooth flow of the intramedullary pin fixation surgical procedure against fractures with osteogenesis imperfecta.

ABSTRACT. Introduction: Medial open-wedge high tibial osteotomy (OWHTO) is widely performed for medial unicompartmental knee osteoarthritis. Although it effectively shifts the mechanical axis laterally, it may unintentionally increase the posterior tibial slope (PTS). The sagittal inclination of the osteotomy plane has been proposed as a factor influencing this change, as it alters the geometry of the opening gap. This study evaluated how different anterior–posterior sagittal inclinations affect anterior and posterior opening distances and their ratio using three-dimensional (3D) simulation analysis.

Methods: Ten patients scheduled for OWHTO or unicompartmental arthroplasty were analyzed using 3D planning software. Five sagittal osteotomy inclinations relative to the tibial shaft were simulated: 20° anterior tilt (AT-20), 10° anterior tilt (AT-10), neutral (AP-0), 10° posterior tilt (PT-10), and 20° posterior tilt (PT-20). For each condition, the anterior opening distance (AD), posterior opening distance (PD), and the AD/PD ratio were measured. A correction angle of 10° was applied uniformly in all simulations.

Results: The mean AD values were 7.6 ± 0.9 mm (AT-20), 7.9 ± 0.9 mm (AT-10), 8.1 ± 0.8 mm (AP-0), 8.3 ± 0.7 mm (PT-10), and 9.1 ± 0.6 mm (PT-20). The mean PD values were 13.1 ± 1.0 mm (AT-20), 12.2 ± 0.8 mm (AT-10), 11.6 ± 0.9 mm (AP-0), 10.7 ± 0.6 mm (PT-10), and 10.3 ± 0.6 mm (PT-20). Accordingly, calculated AD/PD ratios were 0.58 (AT-20), 0.65 (AT-10), 0.70 (AP-0), 0.77 (PT-10), and 0.89 (PT-20). Differences in the AD/PD ratios across the tested inclinations were statistically significant (p < 0.01, in all).

Conclusion: Greater anterior tilt of the osteotomy plane results in a disproportionately large posterior opening, which may increase PTS. Conversely, posterior tilt yields a more balanced anterior–posterior gap and may help preserve the native PTS. Careful adjustment of the sagittal osteotomy inclination during OWHTO planning may therefore minimize unintended PTS changes and improve postoperative biomechanical outcomes.

ABSTRACT. The aim of this study was to clarify the relationship between pain catastrophizing, patient satisfaction, and objectively measured daily physical activity in patients with end-stage knee osteoarthritis. This study evaluated 48 patients (48 knees) with end-stage knee osteoarthritis prior to total knee arthroplasty. Pain intensity, pain catastrophizing, and patient satisfaction were assessed preoperatively using the visual analog scale, Pain Catastrophizing Scale, and the 2011 Knee Society Score, respectively. Daily physical activity was objectively measured for one month before surgery using a wearable physical activity monitor, including step count and time spent at different metabolic equivalent levels. Pain Catastrophizing Scale was significantly associated with lower patient satisfaction, whereas no significant association was observed between patient satisfaction and daily physical activity. These findings suggest that psychological factors, rather than objectively measured physical activity, play a key role in determining preoperative patient satisfaction in end-stage knee osteoarthritis.

ABSTRACT. Background Ankylosed hips are often associated with minimal pain; however, severe restriction of range of motion and increased stress on adjacent joints can lead to impaired gait and functional decline. Total hip arthroplasty (THA) is an effective treatment, but accurate surgical separation of the fused pelvis and femur is essential. Preoperative planning based solely on plain radiographs or computed tomography (CT) is often insufficient to fully understand complex fusion morphology. We report a case of THA for an ankylosed hip in which a three-dimensional (3D) printed bone model was combined with CT-based navigation for osteotomy planning. Methods A 63-year-old woman presented with left knee pain. She had a history of septic arthritis of the right hip at the age of three and subsequently developed right hip ankylosis with leg-length discrepancy. At presentation, the right hip was fixed at 50° of flexion and 15° of adduction, with a leg-length discrepancy of 4.5 cm. Radiographs showed a high-riding greater trochanter and obscured acetabular–femoral boundaries. THA with subtrochanteric shortening osteotomy via a posterolateral approach was planned to restore hip mobility. Results The 3D printed model revealed extensive bony fusion involving both posterior and anterior acetabular regions. Intraoperatively, osteotomy was guided by tactile landmarks identified on the model and confirmed using CT-based navigation. A curved osteotomy line enabled safe joint separation and accurate reproduction of the planned femoral neck osteotomy. Stable acetabular reaming, cup placement, and femoral reconstruction were successfully achieved. Conclusions The combination of a 3D printed bone model and CT-based navigation allows precise understanding of fusion morphology and facilitates safe, accurate osteotomy in THA for ankylosed hips, potentially improving surgical precision and outcomes.

ABSTRACT. Introduction: Full-endoscopic cervical foraminotomy (FECF) is a minimally invasive and effective surgical option for cervical radiculopathy. However, several challenges remain, including intraoperative spatial disorientation, radiation exposure associated with fluoroscopy, and difficulty in accurately confirming the adequacy of decompression. This study aimed to evaluate the clinical outcomes and safety of FECF assisted by O-arm navigation.

Methods: Twelve patients who underwent FECF assisted by O-arm navigation for cervical radiculopathy were retrospectively reviewed. The affected cervical levels were recorded. Clinical outcomes were assessed using the Japanese Orthopaedic Association Cervical Myelopathy Evaluation Questionnaire (JOACMEQ) and the Visual Analog Scale (VAS) for neck and upper limb pain, preoperatively and at 1 year postoperatively.

Results: All patients demonstrated postoperative improvement in JOACMEQ scores. Mean VAS scores significantly decreased from 59.5 to 31.6 for neck pain and from 58.3 to 24.5 for upper limb pain. No patients required revision surgery during the follow-up period, and no severe perioperative complications were observed.

Conclusions: The use of O-arm navigation enabled safe and accurate decompression during FECF, even in the early phase of the learning curve, resulting in favorable clinical outcomes. O-arm–assisted navigation appears to be a valuable adjunct for enhancing the safety and precision of FECF.

ABSTRACT. Revision total knee arthroplasty is a challenging procedure, resulting in a gradual increase in its necessity. Some main obstacles that we have encountered are bone loss and quality of soft tissue. Usually, a metaphyseal sleeve is used in metaphyseal bone loss. However, different constrains utilizes the same metaphyseal sleeves. Factors which determine constrain type includes collateral ligament insufficiency, severity of varus or valgus deformity, gross flexion, extension gap imbalance, and hyper-laxity. These problems can be fixed with constraint prosthesis such as varus-valgus constrain and rotational hinge knee. Our study was designed to demonstrate the level of constrain that affects functional outcome of femoral sleeve in various types of revision total knee arthroplasty. We measure KOOS score, canal filled ratio and survival rate in different level of constrains to find correlation among these types of prosthesis. From September 2012 to December 2020, we retrospectively analyzed all revision total knee arthroplasty in our institute. This retrospective study involved moderate to severe bone loss as classified by the AORI; 100 cases were included into the study. Mean postoperative KOOS score was lowest in rotational hinge knee prosthetic especially those caused by hyper-laxity. Most failure cases in all groups were peri prosthetic infection resulting from 4 out of 5 cases. A satisfactory canal filled ratio of 0.77 was achieved in the metaphyseal areas of all the groups, without any significant differences between them. Semi constrain; we do posterior stabilization design(PS) , varus-valgus constrain(VVC) and a rotational hinge knee(RHK) with metal sleeve. VVC and PS have better functional outcome than RHK at minimal one year follow up.VVC, PS and RHK have no difference in canal filled ratio.

ABSTRACT. [Introduction] For the surgery of malignant bone tumors, securing surgical margins while preserving function is of importance. Especially for complex osteotomies, such as pelvic resections or resections near joints, a three-dimensional understanding of bone shape is more useful than traditional two-dimensional planning with X-ray. Since 2025, our hospital has been utilizing a CAD system for preoperative planning of bone tumor resections and for creating patient-specific 3D printing cutting guides (PSGs). This paper reports a representative case. [Method] The preoperative CT image was transformed to DICOM data. With CAD software (Materialise), the annotations of bone and tumor were performed to construct a three-dimensional model. The optimal osteotomy plane was determined within the software, and a PSG was designed to align with this osteotomy plane. The designed PSG was printed using a 3D printer, sterilized, and then utilized in the actual surgery. [Case] Nine-year-old boy, Ewing's sarcoma in the left pelvis. He had visited a local hospital complaining of fever and left buttock pain and diagnosed with Ewing's sarcoma via open biopsy. The tumor originated in the P1-2 region of left pelvis (80 × 50 × 45 mm) with a massive extraosseous mass. Following neoadjuvant chemotherapy, a PSG was created to make osteotomy lines at the sacroiliac joint and the lateral half of the acetabulum, allowing for a partial hip-joint sparing osteotomy. After the tumor was removed, reconstruction was performed using liquid nitrogen-treated bone auto-grafts and plates. The operative time was 8.5 hours, with a blood loss of 400 mL. Negative margins were achieved, and precise osteotomy was carried out as planned. [Discussion] PSGs possess potential advantages, including ensuring safe surgical margins, facilitating accurate reconstruction, enhancing understanding of complex anatomical structures, and reducing surgical invasiveness. However, their implementation faces challenges, such as lead time, cost, the learning curve, material properties, establishing appropriate sterilization methods, and a lack of insurance coverage in Japan. To promote the widespread adoption of this method in Japan, it is essential to encourage multi-institutional collaborative research in the future.

ABSTRACT. Background: Preoperative computed tomography (CT)-based navigation is used for cervical pedicle screw (CPS) insertion to mitigate the risk of spinal cord and vertebral artery injury. In vertebrae with osteosclerosis due to degeneration or other factors, however, probing may not proceed easily, with difficulty creating the CPS insertion hole. This study investigated the impact of osteosclerosis on the accuracy of CPS insertion. Methods: A total of 138 patients with CPS inserted at the C3 to C7 level using preoperative CT navigation were retrospectively analyzed. Pre- and postoperative CT was employed to investigate screw position and Hounsfield Unit (HU) values at the lateral mass to evaluate the degree of osteosclerosis in the CPS insertion pathway. Results: Among 561 CPS insertions, the Grade 3 perforation rate was 1.8%, and the Grade 2 or higher perforation rate was 8.0%. When comparing insertions with and without CPS perforation, HU values were significantly higher in the perforation group (578 ± 191 vs. 318 ± 191, p<0.01). The frequency of CPS insertion into the mid-cervical spine was also significantly greater in the perforation group (68.9% vs. 62.5%, p<0.01). Logistic regression analysis revealed that a high HU value at the lateral mass (odds ratio 1.09, 95% confidence interval: 1.07-1.11, p<0.01) was a significant independent factor associated with CPS deviation. Conclusions: The screw perforation rate of Grade 2 or higher in CPS insertion using preoperative CT-based navigation was 8.0%. Since osteosclerosis was an independent factor related to CPS deviation, additional care may be required during insertion into affected vertebrae.

ABSTRACT. Background The direct superior approach (DSA), a muscle-sparing technique derived from the posterior approach (PA), has received little attention despite its potential advantages. This study compared the impact of DSA and PA on patient-reported pain and postoperative opioid consumption with medical and surgical complications. Methods We reviewed 451 primary total hip arthroplasties for osteonecrosis of femoral head and osteoarthritis between January 2016 and December 2022, categorized as DSA or PA. Demographic data, including age, sex, preoperative opioid usage, smoking status, chronic alcoholism, and underlying diseases were collected. Propensity score matching balanced DSA and PA groups. Maximum and minimum pain score on the postoperative day (POD), total opioid consumption were compared between the two groups. Inflammation-related serum markers, medical, surgical complications, and length of hospital stay were also analyzed. Results After matching, 139 patients were included in each group. Patients with DSA reported lower average maximum pain on POD #1 and #4, and lower average minimum pain on POD #1, #2, and #4. DSA showed a significant reduction in opioid consumption. In addition, DSA showed a significant reduction in C-reactive protein (CRP) on POD #5, 14, and 28 after index surgery compared to PA. Postoperative delirium (3.6 vs. 9.4%, P=0.05) and length of stay (5.4 vs 8.4 days, P<0.001) showed significant difference between two groups. No significant differences were observed in chronic opioid use, medical complications, other surgical complications. Conclusion DSA was associated with lower patient-reported pain and a marked reduction in opioid consumption, delirium and length of hospital stay.

ABSTRACT. A Smartphone-Based Augmented Reality–Aided Navigation System for Total Knee Arthroplasty

Background: We developed AR-KNEE, a smartphone-based augmented reality (AR)–aided navigation system for total knee arthroplasty (TKA) and unicompartmental knee arthroplasty (UKA). The system projects bone resection angles, the femoral head center, the functional axes of the femur and tibia, and the tibial anteroposterior axis directly onto the surgical field via a smartphone.

Methods: Femoral-side accuracy was assessed by measuring coronal alignment on standing long-leg radiographs in patients undergoing TKA and comparing absolute resection angle errors between AR-aided TKA (n = 31) and TKA performed using a conventional intramedullary guide (n = 41). Tibial-side accuracy was evaluated using postoperative CT images by comparing axial, coronal, and sagittal alignments of the tibial baseplate between TKAs performed with AR-aided navigation (n = 58) and those in which tibial alignment was determined manually (n = 64).

Results: On the femoral side, AR-aided TKA demonstrated significantly smaller absolute errors from the target resection angle compared with the intramedullary guide technique (1.1 ± 1.0° vs. 2.2 ± 1.6°, p < 0.001). On the tibial side, mean absolute errors were 1.1 ± 0.9° for coronal alignment, 1.9 ± 1.3° for posterior slope, and 3.7 ± 3.4° for rotational alignment. The rotational alignment error was significantly smaller with AR assistance than with manual determination (3.7 ± 3.4° vs. 5.1 ± 4.4°, p = 0.047).

Conclusions: The smartphone-based AR-aided navigation system provides acceptable accuracy for knee arthroplasty and may represent a practical alternative to conventional navigation methods.

ABSTRACT. Three-dimensional evaluation of the proximal femur is essential for implant planning and anatomical assessment. However, most conventional coordinate systems rely on posterior reference planes, which are invalid in intertrochanteric femoral fractures due to displacement of the greater trochanter. Considering the simplicity and robustness of coordinate system construction, a plane defined by the proximal femoral axis and the femoral head center—the theoretical plane to place the implant (TPI)—provides a practical and robust alternative. This study aimed to evaluate the reproducibility and validity of a TPI-based coordinate system. Postoperative computed tomography images of 25 patients with intertrochanteric fractures with no history of contralateral femoral fracture were retrospectively analyzed. Parameters included templated stem size during TPI construction, femoral neck anteversion and mediolateral offset measured in the TPI coordinate system, and contralateral femoral neck anteversion measured in a posterior plane–based coordinate system. Reproducibility was assessed using intraobserver and interobserver reliability, while validity was evaluated by comparing contralateral femoral neck anteversion between coordinate systems. Templated stem size showed moderate intraobserver and interobserver reliability (ICC = 0.79 and 0.54, respectively). In contrast, femoral neck anteversion and mediolateral offset demonstrated excellent reproducibility (all ICC > 0.80). Mean contralateral femoral neck anteversion was 28° in the TPI coordinate system and 27° in the posterior plane–based coordinate system, with TPI measurements slightly larger by 1.3 ± 1.0°. These results indicate that the TPI-based coordinate system provides reproducible and valid measurements in fractures where posterior reference plane–based systems fail. Its simplicity and constructability make it suitable for fracture-specific three-dimensional evaluation and potential integration into computer-assisted surgical workflows.

ABSTRACT. Background: Polymethylmethacrylate (PMMA) cement augmentation improves pedicle screw fixation in osteoporotic spines, but its effect on interbody fusion remains controversial. This study compared fusion outcomes between cement-augmented and non-augmented pedicle screws in transforaminal lumbar interbody fusion (TLIF). Methods: A retrospective matched cohort study was conducted in patients with osteopenia or osteoporosis (quantitative CT–derived bone mineral density <120 mg/cm³) who underwent primary TLIF. Patients were matched 1:1 for age, sex, body mass index, and bone mineral density. Interbody fusion was assessed at 3 and 12 months postoperatively using computed tomography and graded according to the Bridwell classification. Fusion rates and perioperative parameters were compared between groups. Results: Eighty-four matched pairs were included. Early fusion at 3 months did not differ significantly between groups (P = 0.266). At 12 months, the cement-augmented group demonstrated a significantly lower rate of complete fusion (Bridwell Grade I–II) compared with the non-cement group (63.5% vs. 76.9%, P = 0.022), with a higher proportion of delayed or incomplete fusion (Grade III–IV). Rates of definite nonunion (Grade IV) were low and comparable between groups. Conclusions: In patients with low bone mineral density undergoing TLIF, PMMA cement augmentation of pedicle screws was associated with delayed or less complete interbody fusion at 12 months despite providing immediate mechanical stability. Cement augmentation should therefore be selectively applied, balancing mechanical benefits against potential adverse effects on fusion maturation.

ABSTRACT. Conventional evaluation of acetabular morphology has relied primarily on two-dimensional radiographic parameters, such as the Sharp angle and center–edge angle. However, these measurements depend on the femoral head center and pelvic reference lines, making it difficult to assess intrinsic acetabular morphology independently of femoral head position or deformation. The aim of this study was to quantitatively evaluate bilateral symmetry of periacetabular morphology along the acetabular height using axial CT images, independent of the femoral head center. Nine patients with minimally deformed osteonecrosis of the femoral head who underwent total hip arthroplasty were included. Preoperative bilateral hip CT data were analyzed using a pelvic coordinate system based on the functional pelvic plane, and image orientation was adjusted to achieve bilateral height alignment. Axial CT slices were extracted at 5-mm intervals from the inferior to the superior acetabular margin. On each slice, bony contours of the bilateral acetabula were manually delineated. For bilateral comparison, the left acetabulum was mirrored and translated to the right without rotational adjustment, and shape similarity was quantified using the Dice coefficient at each normalized acetabular height. Dice coefficients remained consistently high throughout the entire acetabular height. The pooled Dice coefficient across all slices was 0.896 ± 0.037, and mean Dice coefficients at each normalized height ranged from 0.881 to 0.942, with no specific level demonstrating marked bilateral asymmetry. These findings demonstrate that periacetabular morphology exhibits high bilateral symmetry throughout the acetabular height, supporting the validity of a height-dependent, CT-based approach for three-dimensional morphological assessment independent of femoral head deformation or displacement.

ABSTRACT. The obturator externus (OE) is an important dynamic stabilizer against posterior dislocation after total hip arthroplasty (THA). While posterior approaches typically detach the OE, anterior-based approaches allow its preservation. However, surgeons frequently report a tendency for femoral stems inserted through anterior approaches to be driven from anterior to posterior relative to the proximal femoral axis. This raises concern that attempts to achieve neutral sagittal stem alignment may place the femoral entry point in close proximity to the OE footprint at the trochanteric fossa.  This study aimed to construct a three-dimensional coordinate system to quantify the relationship between the OE footprint and the proximal femoral axis, and to evaluate whether a stem placed in a neutral sagittal position would intersect the OE footprint.  We analyzed preoperative CT scans of 100 hips undergoing THA for osteoarthritis, excluding severe deformity or trauma. Using ZedHip®, the largest fit-and-fill stem that matched the femoral canal was templated, and its long axis was defined as the proximal femoral axis. A theoretical plane to place implants (TPI), defined as the plane including the proximal femoral axis and the femoral head center, was constructed, and a TPI-based coordinate system was established. OE location was defined on coronal DRR images using the 6.4-mm footprint described by Vles et al., and on sagittal DRRs using the percentage distance reported by Ito et al. A CT-based landmark was additionally used to define an alternative sagittal OE’ point. Distances between the OE (and OE’) location and the proximal axis were measured.  OE was located a mean of 1.7 ± 2.6 mm lateral to the proximal femoral axis in the coronal plane and 2.4 ± 2.3 mm posterior to the axis in the sagittal plane, whereas the sagittal OE’ was located 5.0 ± 1.9 mm posterior to the axis. The proximal axis passed within a 6.4-mm OE footprint in 20–43% of cases, depending on the definition used.  These findings indicate that, when the anteroposterior width of the femoral stem is taken into account, neutral sagittal stem insertion frequently intersects the OE footprint. To preserve the OE, stems may need to be inserted slightly from anterior to posterior rather than in strict neutral alignment.

ABSTRACT. Introduction: Intramedullary nailing (IMN) of distal tibia fracture was reported low rate of postoperative wound complications, but malalignment was frequently seen in this procedure. We applied three-dimensional computed tomography (3DCT) based simulation for fracture reduction and implanting of IMN.The aim of this study was to assess the improvement of the postoperative alignment for distal tibia fracture treated with IMN using 3DCT preoperative simulation. Methods: Retrospective cohort. Six cases of distal tibial fractures were included. 3DCT multidata fusion workstation (Ziostation Revoras®, Ziosoft, Tokyo) was used for preoperative simulation. Alignment in post-operative coronal and sagittal plane radiographs was measured and >5°deformity was defined as malalignment. Simulation time, operation time, blood loss, fracture distance from ankle plafond, number of screws for distal fragment, need for fibula fixation, use of additional screws, clinical result (walking ability) and complications were investigated. Results: There was no case of malalignment >5°. Mean 2.2° (0 to 4°) of deformity were measured in the coronal plane and its tendency was valgus, on the other hands mean 1.7° (0 to 4°) of deformity in the sagittal. It took mean 40minute (18 to 75minute) for preoperative simulation. Operation time was mean 142minute, blood loss was mean 67.5ml. Distance from joint line was mean 22mm. Number of screws inserted into distal fragment was mean 2.7 (2.5 to 3 screws). Clinically all cases achieved same walking ability as before injury. All cases achieved bone union. Conclusion: 3DCT based simulation for reduction and implanting will improve accuracy of operation for distal tibial fracture.

ABSTRACT. Background Augmented reality (AR) superimposes computer-generated information onto the real world and has been applied in total knee arthroplasty (TKA) to improve surgical accuracy. Smartphone-based AR navigation is a low-cost, less invasive alternative that visualizes the femoral head center and functional axis in real time. This study evaluated the accuracy of distal femoral osteotomy in TKA using a smartphone-based AR system.

Methods Thirty-three knees (9 males, 24 females) underwent primary TKA with the OrthoRaptor® AR system (Shell-Ha Medical Co., Ltd.) between January and August 2025. The system uses a smartphone camera to guide distal femoral resection without guide pins. Postoperative evaluation included standard anteroposterior and lateral knee radiographs and full-length lower limb images. Component placement angles were measured: the lateral angle (α) in the coronal plane and posterior inclination angle (β) in the sagittal plane. Lower limb alignment was assessed using the mechanical axis (MA) and femoro-tibial angle (FTA). The target lateral distal femoral angle (LDFA) was 90°, and errors within ±3° were defined as a safe zone.

Results Mean patient age was 73.9 years, and mean operative time was 98.2 minutes. All procedures were completed without intraoperative complications. Postoperative measurements showed a mean α of 86.0±3.9°, β of 2.5±2.2°, MA of −1.2±3.7°, and FTA of 174.2±4.5°. The postoperative LDFA was 91.3±2.3°, and 78.8% of knees were within the safe zone, indicating high osteotomy accuracy.

Discussion Smartphone-based AR navigation achieved accurate femoral component placement and satisfactory lower limb alignment, comparable to conventional navigation, while eliminating invasive guide pins. Its low cost and portability may facilitate adoption in settings without conventional navigation or robotic systems. Limitations include the small sample size, single-center design, and reliance on radiographic evaluation.

Conclusion The smartphone-based AR system enabled precise distal femoral resection and implant placement in TKA. This minimally invasive, low-cost technology represents a promising alternative to conventional navigation systems in clinical practice.

ABSTRACT. Background: Total hip arthroplasty (THA) using robotic assistance aims to accurate implant positioning and overall outcomes. Robotic assisted THA is written to be particularly effective for difficult cases with severe deformity.

Case Description: A case is 34-year-old woman. She underwent wide excision of pubic chondrosarcoma and the tumor went into remission. 4years later her left hip became severe osteoarthritis and admitted to our hospital. Her ischium and pubis had been widely resected and the fibula had been transplanted as a substitute for the pubis. Therefore achieving sufficient contact between the cup and either the anterior or posterior column of the pelvis was difficult. we underwent robotic-assisted THA (RA-THA) via a anterolateral approach following the MAKO hip protocol [Mako robotic-arm system (Stryker)]. Postoperative radiographs showed no immediate complications, the patient could walk with cane 4 days after surgery, could run 6weeks after surgery.

Conclusions: Robotic-assisted technology can help in the accurate cup placement of acetabular defect in complex hip surgery.

ABSTRACT. Restoring physiological "medial pivot" kinematics is a fundamental objective in total knee arthroplasty (TKA). However, osteoarthritic knees with varus deformity often exhibit "anterior paradoxical motion" rather than normal rollback. This study aimed to determine which implant designs—Cruciate-Retaining (CR), Posterior-Stabilized (PS), or condylar stabilizing (CS)—can effectively eliminate this pathological motion and restore a stable medial pivot. We evaluated intraoperative kinematics in 41 knees across seven TKA models using a navigation system under load stress. Bone resections were performed using the measured resection technique. Based on the translation of the femoral center of rotation, kinematic patterns were classified into Medial Pivot (MP), Parallel Motion, or Lateral Pivot (LP). Significant differences were observed among designs (p=0.04). The Attune PS and Sphere CS models demonstrated high rates of MP (100% and 75%, respectively). In contrast, CR designs were predominantly characterized by Parallel Motion. The persistence of Parallel Motion in CR knees suggests a failure to correct the anterior paradoxical motion inherent to varus osteoarthritis, likely due to residual rotational laxity and sensitivity to soft-tissue balance. Conversely, PS and CS designs successfully enforced a medial pivot, supported by their intrinsic stability mechanisms. Given that anterior paradoxical motion is known to negatively impact patient satisfaction, selecting implants with intrinsic stability is crucial for optimizing joint dynamics and improving clinical outcomes.

ABSTRACT. While various reports indicate that total hip arthroplasty (THA) improves the range of motion (ROM) of the hip joint—typically quantified via goniometry—there remains a paucity of literature utilizing intraoperative navigation for these measurements. This study evaluated perioperative ROM changes immediately before and after component implantation using CT-based navigation. We analyzed 29 cases (21 females, 8 males; mean age: 64.8 years) of THA performed via an anterolateral approach. Following registration, which was executed without femoral head dislocation, baseline ROM was recorded using the navigation system; post-implantation ROM was measured immediately following component placement. Our findings indicate that changes in flexion and extension angles were negligible. However, adduction at maximum extension decreased by 5°in 80% of cases. Conversely, internal rotation at 90° flexion increased by 5° in 74% of cases, with 64.5% exhibiting an increase of 10°. These results suggest that ROM improvements following THA are direction-specific, and clinicians should remain cognizant of posterior dislocation risks, even when utilizing an anterolateral approach.

ABSTRACT. Introduction: Wrist arthroplasty aims to restore wrist function by replicating native joint geometry and biomechanics. Accurate characterization of wrist articular surface geometry is essential for improving implant fit, load transfer, and postoperative mobility.

Objective: The objective of this study was to extract and quantitatively analyze wrist articular surface geometry from three-dimensional (3D) bone meshes using two computational strategies and to compare their performance.

Method: 3D wrist bone surface meshes from an Open-Source Carpal Database (120 wrists from 90 healthy subjects) were analyzed. Articular surfaces were extracted using (1) a template-based approach employing Optimal-Step Non-Rigid Iterative Closest Point (OS-NRICP) registration and (2) a geometric interaction–based strategy using surface proximity and vertex normal intersections. Geometric parameters including articular surface area, dimensions, and curvature were extracted. Statistical analyses were performed to compare the two methods.

Results: While both strategies produced comparable mean geometric measurements, the surface areas in particular showed high deviations. The capitate–hamate articulation exhibited the largest surface area (190,48 mm2), whereas the capitate-lunate articulation was the smallest (55,28 mm2). Curvature analysis revealed slight concave and convex patterns across articulations, with a maximum curvature of 0,15 mm-1. Overall, the template-based OS-NRICP approach demonstrated greater robustness.

Conclusion: This study provides a geometric extraction and characterization of 3D wrist articular surfaces and demonstrates the advantages of template-based non-rigid registration for automated articular surface extraction.

ABSTRACT. Soft tissue sarcomas (STS) are rare and heterogeneous malignancies in which therapy-induced necrosis has prognostic value but is typically estimated from sampled histopathology. We assessed a novel automated MRI-based sub-segmentation tool for quantitative tumor phenotyping in thigh STS, focusing on total tumor volume and percent necrosis. In a retrospective cohort of 20 patients treated with neoadjuvant therapy followed by resection, contrast-enhanced T1-weighted MRI was processed with automated voxel-intensity histogram modeling using a bi-Gaussian fit to separate necrotic from viable tissue. Automated measurements were compared with radiology qualitative reports and pathology. Automated volume estimates showed strong agreement with both pathology (R2 = 0.92) and radiology (R2 = 0.93), exceeding radiology-pathology agreement (R2 = 0.88). Necrosis category agreement was fair between pathology and automated sub-segmentation (weighted kappa = 0.333, p = 0.035), while radiology demonstrated no agreement with pathology or automated estimates. These interim findings suggest that full-volume analysis combined with sub-segmentation, may provide a more reliable characterization of prognostic necrosis percentage. An expanded cohort analysis and ROC-based outcome correlation analysis are ongoing.

ABSTRACT. Hip osteoarthritis is associated with muscle atrophy and fatty degeneration around the hip joint, which negatively affect postoperative outcomes. However, the relationship between symptom duration and muscle atrophy or degeneration in individual hip muscles remains unclear. This study investigated the association between hip pain duration and muscle atrophy or fatty degeneration in patients with unilateral hip osteoarthritis using AI-based CT analysis. Sixty-one patients undergoing primary total hip arthroplasty were included and stratified into short-term (<3 years) and long-term (≥3 years) pain duration groups. AI-based software quantified muscle volume as a percentage of the contralateral side and fatty degeneration using Hounsfield unit (HU) reduction. The mean pain duration was 3.2 years, with average muscle volume reduced to 87.6% and muscle density decreased by 6.9 HU on the affected side. Compared with the short-term group, the long-term group showed significantly greater muscle atrophy (84.6% vs. 90.2%) and fatty degeneration (−7.9 vs. −5.9 HU). Significant atrophy in the long-term group was observed in the iliacus, gluteus maximus, sartorius, adductors, semimembranosus, semitendinosus, and vastus medialis, while significant fatty degeneration was detected in the iliacus, gluteus medius and minimus, obturator externus, pectineus, sartorius, and adductors. Patients with longer symptom duration exhibited greater reductions in muscle volume and muscle density.The iliacus, sartorius, and adductor muscles were particularly susceptible, indicating that prolonged symptoms adversely affect muscle mass and quality in a muscle-specific manner. These findings suggest that symptom duration may help guide targeted preoperative rehabilitation strategies.

ABSTRACT. Purpose: Curved varus osteotomy (CVO), a joint-preserving procedure commonly performed for osteonecrosis of the femoral head (ONFH), aims to move the necrotic lesion of the femoral head from a weight-bearing position to a less weight-bearing portion by rotating it into the varus position. However, it is unclear how effectively CVO increases the non-necrotic weight-bearing area of the femoral head during daily activities. The present study aimed to clarify how CVO increases the non-necrotic weight-bearing area during stair climbing and to identify its determinants using CT-based simulation. Method: The subjects were 20 cases diagnosed with Type C1 ONFH according to the Japanese Investigation Committee classification, which is considered a good indication for CVO. We simulated a 30-degree varus osteotomy and maximum hip flexion of 67 degrees during stair climbing using CT. Results: The percentage of non-necrotic weight-bearing area of the femoral head during stair climbing after CVO was 17.1% (-20.6% to 80.3%), whereas it was 48.6% (23.8% to 84.4%) in the neutral position. The preoperative anterior sagittal necrosis angle exhibited a moderate negative correlation with the percentage of the non-necrotic weight-bearing area of the femoral head during stair climbing following CVO (ρ = -0.66, p = 0.0016). Additionally, a stronger negative correlation was observed with the postoperative anterior sagittal necrosis angle (ρ = -0.88, p < 0.001). Conclusion: The effect of CVO on increasing the non-necrotic weight-bearing area during stair climbing varied among patients, and the postoperative anterior sagittal necrotic angles was a good predictor.

ABSTRACT. Background: Adequate alignment and gap balance are the most important in successful primary total knee arthroplasty. However they are c losely related with component size, joint line level and obliquity, MPTA, LDFA, degree of ER, and degree of ligament laxity. Due to advantages of robotic assistance in terms of knee balance and alignment.robotic primary TKA became easier than manual TKA in term of achieving proper knee balance and alignment. The purpose of this study was to demonstrate the detailed technique of robotic control of all in solution for knaa balance and alignmnet of primary TKA and to report and its clinical results Methods: 503 TKAs for gonarthritis with varus deformed knees were involved in this study. Pts ROM was 12-115 degree, 9 varus, HSS 75 in average. Peroperatively CT was taken and alignment, MPTA and LDFA were measured. After arthrotomy position tracker was fixed and osteophyte was removed. Coronal alignment was optimized according to the patient anatomy within 5 degree of varus with 20 lB distraction of med & lat jt space while the knee joint was maintained in full extrension. MCL was not released usually but released minimally when necessary. And same distraction force was applied in 90 degree flexion. According to the extension-flexion space and fem trochlear alignment which MAKO Robot showed, control and optimize extension gap(0mm) and flexion gap(2-3mm) and trochlear alignment with functional positioning of the fem & tibial implant whining 5 degree of MPTA, LDFA and TEA. PCL recession was not considered because of risk of instability. When CR implantation was not good enough it was converted to PS design. It were cases when 5° or greater ER of TEA , 14mm or greater bony cut from posterior medial femoral condyle was necessary and flexion gap was 3mm or smaller than ext. gap. Postop 2ys clinical evaluation was performed. Results: CR implantation were successful in 477/503 cases(95 %) ROM was improved to 5-120 degree, HSS was also improves to 93 point at 2ys f-up No Infection was found in all 477 cases

Conclusion:Robotic control of all in one sloution for knee balance and alignment is easy and simple way of successful primary TKA, and its clinical results was excellent without periprosthetic infection.

ABSTRACT. In patient-specific implant design, three-dimensional (3D) anatomical bone models are required in parametric formats, most represented by Non-Uniform Rational B-Splines (NURBS). However, 3D bone models are commonly generated as polygonal surface meshes, limiting their direct use in engineering design. Therefore, this work aims to develop and evaluate a fully automated pipeline for converting polygonal bone meshes into NURBS surfaces. To achieve this, a conversion pipeline was implemented in Rhino3D using its Python interface, combining mesh preprocessing, quad-remeshing, surface creation, and NURBS conversion. Evaluation was performed on femoral bone meshes from n=37 subjects. The accuracy was assessed using Chamfer and Hausdorff distances. The results showed that for quad-mesh edge lengths below 3.5 mm, the pipeline achieved an average surface Chamfer distance of approximately 1.5 mm and an average Hausdorff distance of about 2.9 mm. The automated pipeline enables the generation of CAD-ready anatomical models, supporting downstream applications such as patient-specific implant design, and bridging the gap between medical imaging and engineering design tasks.

ABSTRACT. Psoas syndrome is a common complication following THA. In most patients, an overhanging cup can be diagnosed, but there are numerous other factors that can lead to irritation of the psoas tendon. Based on a literature review, an analysis of possible risk factors was carried out with reference to relevant parameters of preoperative planning. A total of 20 parameters associated with psoas syndrome was identified in the literature. By structuring the data, it was possible to develop a recommendation for action, in which the patient's risk is first assessed on the basis of anthropometric data. In a subsequent analysis of morphological, implant and implantation parameters, a more detailed examination of the risk parameters can then be carried out and a reduction in risk can be achieved through adaptations in planning. The heterogeneity of the condition often makes it difficult to identify the cause directly, and even small changes can cause irritation. For this reason, threshold-based risk classification is usually not directly applicable in practice. Our approach can therefore be interpreted as a guideline that can help to assess and reduce the risk of psoas syndrome.

ABSTRACT. Purpose: Functionally aligned total knee arthroplasty (FA-TKA) aims to preserve the native coronal plane anatomy and has demonstrated favorable clinical outcomes. However, concerns remain about the impact of rotational changes in FA-TKA on patellofemoral joint. Therefore, this study aimed to compare radiographic patellofemoral parameters and clinical outcomes between patients undergoing mechanically aligned TKA (MA-TKA) and FA-TKA. Methods: After 1:1 propensity score matching, this retrospective study included 115 patients from both MA-TKA and FA-TKA groups. Pre- and post-operative 3D computed tomography (CT) scans were analyzed to evaluate radiographic parameters of the patellofemoral morphology, including mediolateral and anteroposterior shift of trochlear sulcus. Clinical outcomes were assessed at one year post-operatively. Results: FA-TKA showed significantly fewer cases of trochlear understuffing (2.6% vs. 14.8%), fewer medio-lateral outliers (25.2% vs. 43.5%), and better sulcus restoration (58.3% vs. 39.1%). Although FCR was more internally rotated in FA-TKA, it was closer to the surgical trans-epicondylar axis (1.56° ± 1.39° vs. -0.16° ± 1.37°, p < 0.001). Clinical outcomes were comparable between groups. Conclusion: FA-TKA demonstrated more favorable radiographic restoration of the trochlear anatomy, alleviating concerns about potential patellofemoral compromise compared to MA-TKA. The slight differences in FCR in FA-TKA likely reflect a more anatomical orientation of the trochlear groove, without negatively affecting clinical outcomes. These findings support recent evidence that FA-TKA optimizes trochlear morphology without compromising patellofemoral function. Further long-term studies are needed to confirm these results.

ABSTRACT. Conventional computer-assisted orthopedic surgery (CAOS) systems rely on external displays and instrument-mounted tracking markers, requiring surgeons to mentally map two-dimensional navigation data onto a three-dimensional operative field. Augmented reality (AR) systems have been proposed to address this limitation; however, many existing solutions suffer from fixed focal planes, vergence–accommodation conflict (VAC), unstable overlays, and workflow disruption. We present a binocular AR surgical navigation platform based on laser beam scanning (LBS) retinal projection combined with depth-controlled stereoscopic imaging. Virtual anatomical information is projected directly onto the surgeon’s retina, enabling natural multi-focal depth perception without VAC. Spatial registration between virtual content and patient anatomy is achieved using one to three anatomical or computational landmarks and integrated optical tracking, without requiring instrument-mounted markers. The system dynamically compensates for changes in viewing geometry, magnification, and surgeon motion to maintain stable, depth-accurate overlays. System accuracy was evaluated using an optical tracking reference (mean tracking error 0.12 mm). Mean virtual-to-real registration error was 0.72 mm. In model surgery validation conducted at a craniofacial center, mean navigation error was 0.62 ± 0.22 mm for the mandible and 1.08 ± 0.61 mm for the maxilla, outperforming previously reported navigation accuracy (1–2 mm) while simplifying visualization and workflow. By centering surgical navigation on depth-accurate retinal AR visualization rather than instrument-based tracking complexity, this platform delivers intuitive “what-you-see-is-what-you-get” navigation. The system preserves existing surgical workflows, reduces cognitive load, and demonstrates clinically acceptable accuracy for high-precision orthopedic applications.

ABSTRACT. [Purpose] Transtrochanteric rotational osteotomy (TRO) is an effective treatment for osteonecrosis of the femoral head (ONFH); however, the procedure is technically demanding. In recent years, the usefulness of navigation (Navi) systems has been reported in terms of achieving accurate osteotomy and reducing complications during joint preservation osteotomies. The purpose of this study was to evaluate the clinical outcomes of TRO combined with Navi for ONFH. [Subjects and Methods] This retrospective study included 32 patients (32 hips) who underwent TRO for ONFH between 2014 and 2023 and were followed for at least 3 years. Patients who underwent TRO with Navi after 2019 were classified as Group N (18 patients, 18 hips), whereas those who underwent conventional manual TRO before 2019 were classified as Group C (14 patients, 14 hips). The male-to-female ratio was 10:8 in Group N and 7:7 in Group C. The mean age at surgery was 27.2 years in Group N and 26.2 years in Group C. The mean follow-up period was 4.5 years in Group N and 7.8 years in Group C. All hips were classified as JIC type C2 ONFH. Evaluation parameters included the Harris Hip Score (HHS) preoperatively and at final follow-up, the postoperative healthy weight-bearing area, complications, conversion to total hip arthroplasty (THA), and survival rate, with radiographic failure (recollapse ≥2 mm or joint space narrowing ≤2 mm) defined as the endpoint. [Results] The mean preoperative HHS was 69.6 in Group N and 70.8 in Group C, which improved significantly to 88.3 and 86.2 at final follow-up, respectively. The healthy weight-bearing area was significantly greater in Group N (38.2%) than in Group C (33.2%). Complications included two cases of greater trochanteric pseudarthrosis in Group N, and one case each of infection and greater trochanteric pseudarthrosis in Group C. Conversion to THA occurred in three hips (16%) in Group N and three hips (22%) in Group C. The 4-year survival rate, with radiographic failure as the endpoint, was 80.6% in Group N and 72.3% in Group C, with no significant difference between the groups. [Conclusion] Although Navi-assisted TRO for ONFH was useful in achieving a larger healthy weight-bearing area through accurate osteotomy, the overall clinical outcomes were comparable to those of conventional manual TRO.

ABSTRACT. Background: Instability after total knee arthroplasty (TKA) remains a leading cause of revision, often due to suboptimal ligament balancing or malalignment. Modern techniques, including robotic and navigation-assisted surgery, aim to optimize stability and outcomes. This study investigates the relationship between demographic data, implanted material, and intraoperative planning parameters on 2-year functional outcomes in TKA using a navigation system and a tibia-first surgical workflow.

Methods: De-identified records from 49 patients (26 women, 23 men; mean age 70.7±11.1 years) who underwent primary TKA were retrospectively reviewed. All surgeries were performed by a single surgeon using a standardized workflow: a mid-vastus approach, navigation-assisted tibial preparation, and femoral planning based on soft-tissue laxity assessed with an intra-articular distractor. Demographic data, implant sizes, planning parameters (femur varus, flexion, distal cut height; tibia slope, proximal cut height), and Knee Injury and Osteoarthritis Outcome Scores (KOOS) were collected preoperatively and at 1- and 2-year follow-ups. Pearson correlation analyses were conducted to evaluate associations between variables and KOOS scores.

Results: KOOS scores improved significantly from preoperative (55.79±10.01) to 1 year (81.41±13.94) and 2 years (84.49±9.89), with a mean 2-year improvement of 28.70±12.00, exceeding the minimal clinically important difference. BMI showed a negative correlation with KOOS at both 1 year (r=-0.39, p=0.0053) and 2 years (r=-0.43, p=0.0020). Planned femoral flexion and distal femoral cut height were positively correlated with KOOS at 2 years (r=0.29, p=0.0450 and r=0.31, p=0.0287, respectively).

Conclusion: Higher BMI is associated with poorer functional outcomes, while individualized femoral planning may modestly improve long-term results. These findings underscore the importance of patient-specific planning and preoperative counseling, particularly regarding BMI, to optimize TKA outcomes.

ABSTRACT. Medial joint opening represents a challenging postoperative complication following medial open-wedge high tibial osteotomy (owHTO). Although empirical definitions and preoperative planning strategies have been proposed to avoid reaching the coronal hypomochlion, also referred to as the “tipping point”, the underlying biomechanical mechanisms remain insufficiently understood. The aim of this study was to investigate the distance of the tibiofemoral resultant force to a biomechanical tipping point after correction to the Fujisawa point and an additional 10 % overcorrection. Patient-specific multibody simulations were performed for 11 owHTO patients. The mediolateral (ML) distance between the center of pressure (CoP) of the tibiofemoral resultant force and the tipping point were evaluated at near-extension (5 °) and flexed (90 °) knee position. Under the investigated conditions, the tipping point was not reached in either near extension or flexion, and no medial joint opening occurred. Notably, substantial inter-individual variability in postoperative distances to the tipping point were observed despite identical target alignment.

ABSTRACT. Using computed tomography (CT)-based navigation for total hip arthroplasty (THA) enables accurate reproduction of preoperative three-dimensional (3D) planning and may reduce dislocation risk by improving implant alignment. Although automated preoperative planning has attracted increasing attention to reduce surgeon’s time and effort, the variability of 3D preoperative planning across surgeons remains unclear. Therefore, we aimed to evaluate inter-surgeon variation in preoperative THA planning. Preoperative CT data from 25 patients who underwent unilateral THA using the same implant (Accolade II stem; Stryker) were analyzed. Eleven hip surgeons independently performed 3D preoperative planning using Stryker planning software. All surgeons selected and placed the implant, determining the combination of cup anteversion in relation to stem anteversion with cup inclination fixed at 40°, while targeting equal leg length to the contralateral side, and implant size was fixed to the size used during the actual surgery. Variation in implant positioning was assessed along the mediolateral, anteroposterior, and craniocaudal axes. Angular variation was evaluated for cup anteversion, and stem flexion, anteversion, and varus. Each parameter was compared using the Mann-Whitney U test within the angular and positional parameter groups, and agreement in the combination of cup and stem anteversion was assessed using Fleiss’ kappa coefficient. Mean variation in cup positioning ranged from 2.9 to 4.7 mm, and stem positioning ranged from 2.2 to 5.9 mm. Cup anteversion variation was 5.1 ± 2.5°, whereas stem angular variation ranged from 2.2° to 11.2°. Stem anteversion showed the greatest angular variation, and stem craniocaudal position showed the greatest positional variation (P <0.001). The agreement for combination of cup and stem anteversion was moderate (Fleiss’ kappa coefficient of 0.40). Substantial inter-surgeon variation exists in preoperative THA planning even under a shared planning philosophy. In particular, stem anteversion and stem position along the craniocaudal axis showed large variation. These findings suggest that each surgeon should perform preoperative 3D planning on their own.

ABSTRACT. The integration of large language models (LLMs) into healthcare has demonstrated significant potential for enhancing clinical outcomes and patient engagement, particularly in the complex domain of reverse shoulder arthroplasty, where intricate patient data and precise surgical decisions are paramount. This research examines how LLM tools can improve physician-patient communication, streamline clinical documentation, and bolster preoperative education. Employing a mixed-methods approach, the study incorporates a systematic literature review alongside an analysis of existing practices to pinpoint gaps in LLM applications within orthopedic surgery, while combining quantitative evaluations of surgical results with qualitative insights into patient experiences to gauge the tools' efficacy. Initial results reveal that LLMs markedly boost patients' comprehension of procedures, enabling better-informed choices, with experimental tests on skeletal models confirming the system's accuracy, efficiency, and viability for real-world surgical support. Overall, these implementations not only refine communication and workflows but also heighten patient involvement and satisfaction, underscoring LLMs' revolutionary role in reverse shoulder surgery and providing key insights for elevating patient care and clinical standards. The findings emphasize the need to further develop AI applications in orthopedics and allied fields to build on these promising outcomes. Ultimately, the study calls for embedding sophisticated LLM tools into healthcare frameworks to maximize surgical success and patient contentment, with notable implications for health policy in areas like Hong Kong, where such innovations could drive advancements in care and efficiency, and urges expanded future investigations into LLMs' wider medical uses to unlock their full advantages.

ABSTRACT. Purpose: In conversion Total Hip Arthroplasty (cTHA) following intramedullary nailing for intertrochanteric fractures, stress concentration around distal screw holes is a risk factor for periprosthetic fracture. However, no clear consensus exists regarding the appropriate stem length. This study aimed to determine the optimal stem length in cTHA using finite element analysis (FEA).

Methods: Based on preoperative CT data from 30 patients who underwent THA, 3D femoral models post-nail removal were created. We constructed 3D-cTHA models by inserting stems with bypass lengths of 10, 20, 30, and 40 mm relative to the distal screw holes. Simulations were performed under loading conditions of normal walking and stair climbing to calculate the maximum equivalent stress around the distal screw holes. Statistical analysis was performed using one-way repeated measures ANOVA with Bonferroni correction (p < 0.05).

Results: The mean femoral diameter at the distal screw holes was 25.7 ± 1.7 mm. Maximum equivalent stress around the distal screw holes was significantly lower with stem bypass lengths of 30 and 40 mm compared to 10 mm. There was no significant difference between the 30 mm and 40 mm bypass lengths.

Discussion: Stress around the distal screw holes decreased as stem length increased, plateauing at a bypass length of 30 mm (approximately 1.2 times the femoral diameter). This length was considered necessary and sufficient. The mechanism likely involves physical reinforcement of the weakened area, suppression of the local hinge effect, and stress dispersion due to the high rigidity of the stem.

Conclusion: The optimal stem length for cTHA involves a bypass length of 30 mm past the distal screw holes. This length effectively mitigates stress concentration and may reduce the risk of periprosthetic fracture.

ABSTRACT. Recurrent anterior dislocation after revision total hip arthroplasty (THA) remains a challenging complication, with implant impingement playing a major role. Although implant alignment is critical, the functional implant-specific range of motion (ROM) may be more predictive of postoperative stability. This study assessed whether achieving implant ROM exceeding the required range for activities of daily living (ADLs) could help prevent recurrent dislocation. A retrospective review identified five patients who underwent revision THA for anterior instability. Pre- and postoperative implant ROM—including extension (Ext), external rotation (ER), flexion, and internal rotation at 90° flexion—was measured using CT-based 3D analysis. Outcomes were compared to functional ROM thresholds (≥40° for Ext and ER) and categorized into stable (n=4) or unstable (n=1) groups based on postoperative dislocation recurrence. In the stable group, postoperative Ext improved from 27° to 58° and ER from 32° to 62° (both p < 0.05). The unstable patient failed to achieve sufficient improvement (Ext: 13°→16°, ER: 17°→22°), remaining below the functional thresholds. Only the stable group exceeded ROM requirements postoperatively. Restoring extension and external rotation ROM beyond 40° appears essential for dislocation prevention after revision THA. These findings suggest the implant selection and alignment exceeding the required ROM for ADLs appear essential for successfully controlling dislocation following revision THA.

ABSTRACT. The present study aims to analyze the incidence of intraoperative mid-flexion laxity using continuous flexion-arc gap assessment, risk factors for mid-flexion laxity, and clinical results in navigation-assisted total knee arthroplasty (TKA). Ninety posterior-stabilized TKAs were performed under navigation guidance for patients with degenerative arthritis and varus deformity. Intraoperatively, the gap between the trial femoral component and insert was evaluated in the navigation system with continuous flexion-arc gap assessment. Each medial and lateral gap at flexion (90°) and extension (0°) were made to be less than 3 mm. Mid-flexion laxity was determined when the gap in the flexion range between 15° and 60° was 3mm or more. The proportion of knees with mid-flexion laxity was investigated. The factors affecting mid-flexion laxity were identified in terms of demographics, preoperative convergence angle, and change in joint line height and posterior femoral offset. The Knee Society Score and Western Ontario and McMaster Universities Osteoarthritis Index were evaluated. There were 31 cases (34.4%) of lateral mid-flexion laxity (average peak mid-flexion gap=3.7 mm). The other 59 cases did not show mid-flexion laxity. The preoperative convergence angle was the only significant factor affecting lateral mid-flexion laxity (odds ratio=1.466, p=0.002). There were no significant differences in the clinical results between the groups with and without mid-flexion laxity. The continuous flexion-arc gap assessment was useful in evaluating mid-flexion laxity using navigation-assisted TKA. The preoperative convergence angle, reflecting soft tissue laxity, can be a practical and simple radiographic finding for predicting lateral mid-flexion laxity.

ABSTRACT. Surgical treatment for capitellar osteochondritis dissecans aims to restore radiocapitellar biomechanics and prevent long-term joint degeneration; however, the mechanisms of postoperative joint adaptation during skeletal growth remain incompletely understood. In this study, we quantitatively analyzed longitudinal changes in subchondral bone density distribution and whole-joint morphology using three-dimensional computed tomography in 51 adolescent patients treated surgically for capitellar osteochondritis dissecans. Subchondral bone density within the lesion was significantly reduced preoperatively but increased after surgery and normalized toward the distribution observed in the contralateral elbow. The degree of postoperative density normalization showed a significant correlation with improvements in clinical outcomes, suggesting that restoration of physiological joint loading promotes recovery of subchondral bone adaptation. In contrast, morphological remodeling characterized by enlargement of the radial head and secondary adaptation of the ulnohumeral joint was evaluated by postoperative comparisons between the affected and contralateral elbows. These structural changes were associated with restriction of elbow flexion and reflected coordinated remodeling of the entire elbow joint rather than isolated local deformation. Taken together, our findings indicate that surgical restoration of radiocapitellar biomechanics initiates a progressive whole-joint stabilization process, in which early normalization of subchondral bone adaptation is followed by secondary morphological remodeling of the surrounding joint structures. We conclude that this coordinated sequence of biomechanical correction and joint adaptation may represent a fundamental mechanism underlying favorable long-term clinical outcomes after surgical treatment for capitellar osteochondritis dissecans in growing patients.

ABSTRACT. Background: Articular cartilage is poorly visualized on single-energy CT (SECT), and MRI may be unavailable in acute musculoskeletal settings. Dual-energy CT (DECT) can generate material-pair images that may enhance cartilage conspicuity. Purpose: To evaluate DECT for cartilage visualization in pediatric elbow imaging. Methods: Five pediatric patients (10 elbows) underwent DECT. We evaluated 272 material-pair conditions from 17 candidate basis materials. Contrast-to-noise ratio (CNR) was calculated from material-density measurements within regions of interest placed in articular cartilage and adjacent soft tissue. Conditions were ranked by CNR, and top-ranked conditions were compared with SECT using the Wilcoxon signed-rank test (p<0.05). In one elbow with MRI, cartilage was segmented on the top-ranked DECT images to create a three-dimensional CT-derived cartilage model; an MRI-based model served as the reference. Agreement was quantified using the Dice coefficient, recall (sensitivity), and precision (positive predictive value). Results: CNR on SECT was 5.7±0.69 (mean±SD). The three highest-CNR conditions were Brushite (HAP) 13.4±3.01, HAP (Calc_Bone) 13.3±2.32, and Calc_Bone (Calcium) 12.4±1.89; all exceeded SECT (all p<0.01) and showed clear qualitative cartilage visualization. For MRI-referenced cartilage extent, Calc_Bone (Calcium) showed the best agreement (Dice 0.62; recall 50.5%; precision 78.6%). Conclusion: DECT material-pair imaging significantly improved cartilage–soft tissue CNR versus SECT, supporting feasibility for enhancing cartilage conspicuity; however, boundary accuracy versus MRI was limited, suggesting use as a conspicuity aid rather than for precise delineation. This technique may be useful for pediatric intra-articular fractures with substantial cartilaginous components that can be difficult to diagnose.

ABSTRACT. This study aimed to evaluate the effects of a pressure-dispersion pad on pain and activities of daily living (ADL) in patients with difficulty kneeling after total knee arthroplasty (TKA). Among 125 patients who underwent primary TKA, 25 patients who reported kneeling difficulty at the 1-year postoperative follow-up were enrolled. A pressure-dispersion pad was prescribed, and patient-reported outcomes were assessed before pad use and reassessed after 3 months of routine use. Kneeling ability was evaluated using a 10-point kneeling score (0 = completely unable, 10 = no difficulty). Kneeling-related ADL was assessed using a 7-item questionnaire (sitting on the floor, cleaning the floor, folding laundry, standing up from a futon on the floor, picking up objects from the floor, gardening, and caregiving). The mean kneeling score declined significantly from 4.9 preoperatively to 2.4 at 1 year postoperatively prior to pad use (p = 0.02). After pad use, the score improved significantly from 2.4 to 7.5 (p < 0.001). ADL improvements were observed in sitting on the floor (63%), folding laundry (63%), cleaning the floor (80%), picking up objects (83%), and gardening (60%). In contrast, improvement rates were low for standing up from a futon (16%) and caregiving (0%). These findings suggest that a pressure-dispersion pad improves kneeling ability and several kneeling-related ADL tasks after TKA, although some functional limitations may persist.

ABSTRACT. This study establishes a comprehensive seven-phase automated analysis pipeline that integrates bone shape and quality parameters to quantify age-related structural changes in the proximal femur and lumbar vertebrae. Utilizing Quantitative Computed Tomography (QCT) data from 110 subjects aged 51 to 91 years, we constructed Statistical Shape Models (SSM) by combining Bayesian Coherent Point Drift (BCPD)-based registration with StradView Cortical Bone Mapping. Spatial distribution analysis revealed a significant decline in femoral cortical thickness at a rate of -0.009 mm per year (p = 0.008), representing a reduction of approximately -3.0% per decade. To enhance detection sensitivity, Statistical Shape and Appearance Models (SSAM) were developed, integrating cortical thickness and density into unified feature vectors. This multi-modal approach successfully identified age-correlated patterns in lumbar vertebrae (L1-L3) that remained undetected in shape-only models (FDR-corrected p < 0.05), demonstrating the superior power of appearance integration in osteoporotic assessment. Furthermore, site-specific correlation analysis indicated independent aging mechanisms between the femur and vertebrae (r < 0.2). While the femur exhibited uniform thinning, the vertebrae showed heterogeneous changes influenced by osteophyte formation and cortical-trabecular boundary ambiguity. These findings underscore the fundamental structural differences in bone remodeling across anatomical sites. The proposed pipeline provides a reproducible and high-precision framework for identifying site-specific fracture risks, emphasizing the necessity of localized evaluation rather than unified models for a comprehensive understanding of skeletal aging and osteoporosis.

ABSTRACT. Single-view wrist radiography remains the clinical standard for the assessment of carpal deformities due to its low cost, low radiation exposure, and wide availability. However, radiographs provide only a two-dimensional projection of a highly complex three-dimensional (3D) anatomy, limiting accurate interpretation of carpal bone geometry and spatial relationships. This limitation is particularly critical in the carpal region, where multiple small bones overlap in projection and exhibit substantial inter-subject anatomical variability, making reliable 2D–3D reconstruction from a single view highly challenging. In this work, we present an anatomy-driven framework for depthmap-based 2D–3D reconstruction of the carpal bones from a single-view wrist radiograph. Rather than directly addressing end-to-end reconstruction, the proposed approach focuses on establishing robust methodological foundations for quantitative depth estimation in this anatomically complex region. The framework integrates automatic carpal bone segmentation and anatomical landmark generation with a constrained CT-to-X-ray registration pipeline. Deep learning models trained on CT data are employed to extract anatomically meaningful constraints, which are used to substantially reduce the search space and ambiguity of the mono-view registration problem. CT-to-X-ray registration is formulated as a gradient-based optimization task using a gradient correlation similarity metric to emphasize structural consistency between digitally reconstructed radiographs and input X-ray images. Pose estimation is performed using the Covariance Matrix Adaptation Evolution Strategy (CMA-ES), enabling robust convergence in the presence of strong projection overlap and non-convex optimization landscapes. Following registration, per-pixel depth information is extracted from the aligned CT volumes to generate quantitative depth maps of the carpal region. The framework is validated on a large dataset comprising 510 wrist CT volumes and 476 single-view radiographs, including expert-annotated carpal bone masks. Experimental results demonstrate stable and reproducible carpal alignment across subjects and acquisition conditions, and enable the generation of millimeter-scale depth maps. These results establish essential foundations for future single-view 2D–3D carpal reconstruction and geometry-aware diagnostic and treatment planning applications.

ABSTRACT. Osteoporosis is a globally prevalent disease characterized by decreased bone mineral density (BMD), significantly increasing fracture risk. While dual-energy X-ray absorptiometry (DXA) and quantitative computed tomography (QCT) are standard diagnostic methods, they face challenges regarding cost and radiation exposure. Recently, BMD estimation using accessible, low-dose plain X-rays has gained attention. However, most existing methods focus on areal BMD. In contrast, volumetric BMD (vBMD) enables three-dimensional evaluation, offering a superior assessment of bone strength. Therefore, estimating vBMD from plain X-rays is of significant clinical value. This study proposes a method for high-precision lumbar vBMD estimation. Specifically, we automatically extract the Region of Interest (ROI) covering the first to fourth lumbar vertebrae (L1–L4) from frontal plain X-ray images. This extracted ROI is then fed into a GAN-based model to generate a BMD distribution map, which enables the direct inference of volumetric BMD values. Experiments involved pre-training with a large-scale dataset of 9,006 simulated X-ray images and BMD maps. Subsequently, we trained and evaluated the model using real X-ray images from 212 cases, achieving a Pearson correlation coefficient (PCC) of 0.835. Furthermore, to validate clinical effectiveness, we evaluated accuracy using real X-ray images and actual measured BMD values from 110 cases. The proposed method demonstrated high accuracy with a PCC of 0.852.

ABSTRACT. Opportunities to learn advanced surgical techniques are limited in rural areas, requiring medical professionals to travel to large urban centers for training. This paper presents the development of an online surgical training system based on omnidirectional video streaming, enabling physicians, nurses, and medical device engineers in remote regions to participate in immersive surgical training from distant locations. The proposed system simultaneously streams a 360-degree view of the operating room and multiple detailed video feeds of the surgical field and medical instruments. Low-latency transmission is achieved using a 5G-compatible router, and participants view the content through head-mounted displays. In addition, a management application for instructors allows real-time interaction, including responding to questions, switching video feeds, and broadcasting announcements. The system supports up to ten auxiliary cameras, enabling training from multiple professional perspectives. Experimental deployment demonstrates that the system provides a highly immersive and interactive learning environment, suggesting that omnidirectional video streaming is a practical alternative to conventional on-site surgical training.

ABSTRACT. Objective Developmental dysplasia of the hip (DDH) is one of the major causes of degenerative hip disease. Rotational acetabular osteotomy (RAO) is widely performed as a joint-preserving procedure for DDH. However, RAO is technically demanding, requiring precise osteotomy and optimal acetabular fragment rotation. This study aimed to evaluate the clinical and radiographic outcomes of RAO with CT-based navigation in patients with DDH. Methods This retrospective study included 104 hips in patients with DDH who underwent RAO with CT-based navigation between 2000 and 2022, with a minimum follow-up of 3 years. The mean age at surgery was 33 years, and the mean follow-up period was 10 years. Radiographic evaluation included measurement of the center-edge (CE) angle and acetabular roof obliquity (ARO). Clinical outcomes were assessed using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) score preoperatively and at final follow-up. Intraoperative complications and cumulative survival rate, with conversion to total hip arthroplasty as the endpoint, were also evaluated. Results The mean CE angle improved from 4° preoperatively to 35° postoperatively, and the mean ARO improved from 27° to −1°. The mean WOMAC score significantly improved at final follow-up compared with the preoperative assessment. No serious intraoperative complications, including major vascular or nerve injury, were observed. The cumulative survival rate at 10 years was 96%.

ABSTRACT. Background: Highly porous acetabular components are increasingly used in total hip arthroplasty (THA); however, their early radiological behavior remains controversial. This study compared short-term clinical and radiological outcomes between a three-dimensional (3D)-printed highly porous cup and a hydroxyapatite (HA)-coated non-highly porous cup implanted using computed tomography (CT)-based navigation. Methods: After propensity score matching for age, sex, diagnosis, and surgical approach, 66 hips were analyzed in each group. Clinical outcomes were evaluated using Japanese Orthopaedic Association (JOA) and Japanese Orthopaedic Association Hip Disease Evaluation Questionnaire (JHEQ) scores preoperatively and at 6, 12, and 24 months postoperatively. Radiological evaluation of cup orientation, cup positioning, initial gaps, and radiolucent lines (RLLs) was performed. Results: Clinical outcomes did not differ significantly between groups at the 2-year follow-up. The mean absolute error of cup orientation did not differ significantly between the Trident II and Trident HA groups (inclination: 2.9 ± 2.3° vs. 2.6 ± 2.2°, respectively; anteversion: 2.7 ± 2.0° vs. 2.7 ± 2.0°, respectively). The incidence of initial gaps was significantly higher in the Trident II group than in the Trident HA group (18/66 cases, 27.3% vs. 8/66 cases, 12.1%; p = 0.027). Radiolucent lines showed a progressive decline during the 2-year follow-up in both groups, with no significant differences at 2 years. Conclusion: Short-term clinical outcomes and radiological evaluations of the 3D-printed highly porous and non-highly porous cups were comparable. Initial gaps in the 3D-printed highly porous cups diminished over time, and the highly porous cups demonstrated excellent radiological outcomes.

ABSTRACT. Early intervention is critical for managing osteonecrosis of the femoral head (ONFH), yet manual severity assessment remains time-consuming and prone to inter-observer variability. This study proposes a fully automated diagnostic pipeline that integrates deep learning with explicit geometric analysis for multi-standard classification and collapse risk prediction. Using T1-weighted MRI of 92 hips, we developed an integrated framework employing 3D nnU-Net and SwinUNETR for segmentation and landmark detection, followed by the Geometric ONFH Classification Algorithm (GOCA) to mathematically replicate JIC, ARCO, and Steinberg definitions. Unlike conventional deep learning studies relying on qualitative Grad-CAM heatmaps, GOCA provides explicit visual and numerical grounds for classification. The automated pipeline demonstrated high consistency with expert diagnosis, achieving accuracies of 69.6% for JIC and 82.6% for ARCO, with Cohen's Kappa values indicating moderate to substantial agreement. Furthermore, Kaplan-Meier analysis confirmed that the automated framework provided robust risk stratification (p < 0.001), demonstrating significant risk separation comparable to manual diagnosis (Log-rank p < 0.001) and effectively stratifying patient risks. These findings demonstrate that objective geometric quantification effectively mitigates diagnostic subjectivity and serves as a reliable tool for clinical screening and prognosis prediction.

ABSTRACT. The use of navigation systems has become increasingly common in bone tumor surgery. We performed bone tumor resection surgery using two different types navigation systems: CT-based hip navigation system or O-arm navigation system, depending on the tumor characteristics, and investigated their applicability and usefulness of each.

ABSTRACT. Osteoarthritis (OA) of the hip and knee is a major cause of pain, functional impairment, and disability in the aging population. Beyond joint degeneration, increasing evidence indicates that periarticular and lower-limb muscle pathology plays an important role in disease progression and functional outcomes. However, direct comparisons of muscle volume and myosteatosis between hip osteoarthritis (HOA) and knee osteoarthritis (KOA) across the entire lower limb remain limited. This study aimed to clarify joint-specific differences in lower-limb muscle volume and myosteatosis between patients with HOA and KOA using artificial intelligence (AI)–based computed tomography (CT) muscle segmentation. This retrospective study included 58 female patients with HOA and 53 female patients with KOA undergoing primary joint arthroplasty. Preoperative whole-length lower-limb CT images were analyzed using a Bayesian U-Net–based deep learning system for fully automated muscle segmentation. Muscle volume and mean CT attenuation values (Hounsfield units) were calculated as indices of muscle quantity and muscle quality, respectively. In the pelvic and proximal thigh region, CT attenuation values of the adductor and iliopsoas muscle groups were significantly lower in the HOA group, indicating increased myosteatosis, and adductor muscle volume was also reduced. In contrast, the KOA group demonstrated significantly smaller quadriceps femoris volume in the thigh and reduced muscle volume of the superficial posterior compartment in the lower leg. Furthermore, CT attenuation values of both the deep and superficial posterior compartment muscles were significantly lower in KOA. HOA and KOA exhibit distinct, joint-specific patterns of lower-limb muscle pathology. AI-based CT muscle segmentation enables comprehensive and objective assessment of muscle pathology and may support joint-specific perioperative evaluation and rehabilitation strategies.

ABSTRACT. This retrospective study assessed whether lower-limb CT–derived muscle cross-sectional area (CSA) can identify DXA-defined low muscle mass more accurately than the conventional L3 psoas measure in women with end-stage hip osteoarthritis (HOA). Seventy-three female patients scheduled for total hip arthroplasty (mean age 61.4±12.6 years) who underwent pelvic-to-lower-limb CT and DXA within an acceptable interval were included. Muscle CSA (cm²) was quantified at standardized anatomical levels spanning the pelvis to the proximal lower leg—anterior superior iliac spine (ASIS), femoral head center, mid-thigh, distal thigh, and proximal lower leg—and compared with L3 psoas major CSA when available. Muscle segmentation and CSA extraction were performed using an AI-based automated method built on a Bayesian U-Net architecture. Low muscle mass was defined as a DXA-derived skeletal muscle mass index (SMI) <5.4 kg/m². Receiver operating characteristic analyses demonstrated that limb-level CSA measures consistently outperformed L3 psoas for discriminating low muscle mass. Among unadjusted CSAs, distal thigh total CSA showed the highest accuracy (AUC 0.941), followed by proximal lower leg (AUC 0.937) and mid-thigh total CSA (AUC 0.926), each significantly superior to L3 psoas (AUC 0.732). After height-squared normalization, discrimination remained strongest at limb levels, with distal thigh CSA/height² achieving the highest AUC (0.970) and similarly high performance at mid-thigh and proximal lower leg. In contrast, pelvic-level gluteal measures showed only moderate accuracy (AUC ~0.86–0.88) and were not clearly superior to L3 psoas. These findings suggest that, in women with end-stage HOA, a single lower-extremity CT slice—particularly at the distal thigh—provides a practical and physiologically relevant opportunistic marker for identifying DXA-defined low muscle mass in preoperative settings, while height indexing offers only modest incremental benefit.

ABSTRACT. Aims: To compare early postoperative functional recovery between conventional jig-based total knee arthroplasty (TKA) and robotic-arm assisted TKA. Methods: This retrospective cohort study included 170 consecutive primary TKAs with a minimum 1-year follow-up: 112 conventional TKAs (Group 1) followed by 58 robotic-arm assisted TKAs (Group 2). Inpatient functional outcomes and follow-up clinical and radiographic outcomes were collected. Clinical outcomes were assessed using the Knee Society Score (KSS), WOMAC, and Feller scores. Radiographic alignment parameters were measured using an AI-based program, including medial proximal tibial angle (MPTA), joint line convergence angle (JLCA), mechanical lateral distal femoral angle (mLDFA), hip–knee–ankle (HKA) angle, weight-bearing line (WBL, %), and joint line obliquity angle (JLOA). Baseline demographics and preoperative radiographic parameters were compared between groups. Results: There were no significant baseline differences between groups in age, sex, body mass index, ASA score, preoperative haemoglobin level, or preoperative radiographic parameters (all p > 0.05). Robotic-arm assisted TKA was associated with significantly lower pain scores on postoperative days (POD) 1 and 2 (p < 0.001), reduced analgesic requirements on POD 1 and 2 (p < 0.001), and greater maximum knee flexion at discharge (POD 7) (p < 0.001) compared with conventional TKA. At final follow-up, clinical outcomes (KSS, WOMAC, Feller) did not differ significantly between groups. Postoperative MPTA was statistically different (Group 1 vs Group 2: 91.8 ± 1.8° vs 90.5 ± 2.1°, p = 0.001), whereas other radiographic parameters were not. Alignment outliers (>3°) were similar between groups (13.4% vs 8.6%, p = 0.361). Conclusion: Robotic-arm assisted TKA improved early postoperative pain and short-term functional recovery versus conventional TKA, but did not demonstrate superior 1-year clinical outcomes or meaningful differences in most alignment measures. Longer-term studies are warranted.

ABSTRACT. Rigid spinal orthoses are prescribed for vertebral fractures and postoperative stabilization, but conventional fabrication requires in-person casting and multiple manual steps. We developed a scan-to-print workflow using patient geometry, computer-aided design (CAD), and 3D printing in polypropylene and assessed feasibility in a pilot clinical series. Between September 2024 and May 2025, twelve patients with thoracolumbar and lumbar vertebral compression fractures (mean age 78.3 years) received 3D-printed orthoses; conventional devices were prepared in parallel. Scan inputs were standing surface scans (n=9) or CT-derived models (n=3). Hands-on fabrication time decreased from approximately 4 hours to 1 hour; printing took about 3 hours and could be performed unattended. Mechanical testing showed orientation-dependent strength, highlighting the importance of build orientation. During three months of wear, no orthosis broke and no progressive collapse was observed; one patient switched because of discomfort. These findings support feasibility of 3D-printed rigid spinal orthoses with reduced waste, while prospective studies are required to optimize comfort, evaluate outcomes, and establish secure data-transfer systems.

ABSTRACT. Objective: To evaluate the efficacy of Bilateral Contralateral Laminoplasty (FE-BCL) using a full-endoscopic spine surgery (FESS) system for lumbar spinal stenosis (LSS) regardless of anatomical variations. Methods: A retrospective analysis was conducted on 27 consecutive patients (54 joints) who underwent FESS. The surgical procedure—either conventional unilateral laminotomy for bilateral decompression (ULBD) or FE-BCL—was selected based on preoperative anatomical evaluation, particularly facet joint orientation. In FE-BCL, staged resection of the ipsilateral ligamentum flavum was performed via bilateral entry to ensure a safe working space and optimize the decompression angle. Postoperative CT was used to evaluate facet preservation rates and the incidence of iatrogenic fractures. Results: FE-BCL was primarily selected for cases with sagittalized facets (facet angle <20°). Postoperative CT confirmed symmetrical "trumpet-shaped" decompression in both groups. Both techniques achieved high facet preservation rates exceeding 85%, and a significant correlation was found between facet morphology and the choice of surgical technique ($R = 0.004$). While one iatrogenic fracture and one reoperation due to insufficient decompression occurred in the ULBD group, no such complications were observed in the FE-BCL group. Conclusion: FE-BCL maximizes the advantages of the uniportal FESS system to achieve ideal spinal canal enlargement while preserving a large portion of the facet joint. By utilizing a bilateral approach and staged ligamentum flavum resection, this technique offers a safe and reliable minimally invasive treatment option for anatomically challenging cases where conventional unilateral approaches might carry a risk of articular damage.

ABSTRACT. Background: Anterior knee pain (AKP) after total knee arthroplasty (TKA) results in dissatisfaction with the surgical outcomes. This study aimed to investigate risk factors for AKP after TKA using radiographic assessments. Methods: This retrospective matched-pair case-control study included 284 patients who underwent primary fixed-bearing posterior-stabilized TKA (Persona, Zimmer) from 2016–2020, with a minimum follow-up of 2 years. The inclusion criterion was varus osteoarthritis in the knees without patellar resurfacing. The exclusion criteria were the use of bone graft or metal block, valgus knees, and patellar resurfacing. The patellar tilt angle, lateral patellofemoral angle, length of the tibial tuberosity to the trochlear groove (TT-TG), hip-knee-ankle (HKA) angle, patellar height, and trochlear dysplasia index were measured. Patients were divided into two groups according to the presence of AKP based on the Feller score (≤20 points) at 2-year follow-up, then a 1:2 propensity matching was performed. Risk factors for AKP were identified through a logistic regression analysis. Results: A total of 135 patients were enrolled in this study after matching. The preoperative patellar tilt angle, lateral patellofemoral angle, TT-TG, and trochlear dysplasia index significantly differed between the groups (P<0.05), but not the postoperative parameters (P>0.05). The WOMAC total score and Feller score were significantly different postoperatively. (P<0.05) Logistic regression analysis revealed that the preoperative patellar tilt angle, preoperative trochlear dysplasia index, and preoperative lateral patellofemoral angle were significant risk factors for AKP. Conclusion: Care should be exercised when performing TKA in patients with a large patellar tilt angle, a small lateral patellofemoral angle, and a small trochlear dysplasia index preoperatively, as there can be an increased risk of postoperative anterior knee pain in TKA with unresurfaced patellae. Surgeons may consider counseling patients about their increased risk of AKP if these factors are present preoperatively

ABSTRACT. Background: Perioperative changes in femorotibial rotational alignment following total knee arthroplasty (TKA) are not well characterized. This study aimed to evaluate femorotibial rotational alignment change after TKA and its effects on intraoperative knee kinematics and postoperative patient-reported outcome measures (PROMs). Methods: We analyzed 250 knees undergoing TKA for osteoarthritis. Axial femorotibial rotational alignment was determined using the anatomical femorotibial rotational angle (FTRA) on preoperative and postoperative computed tomography scans. A >10° perioperative change in FTRA (preoperative to postoperative) was defined as a femorotibial rotational change (RC). Intraoperative tibial internal rotation relative to the femur during knee flexion was recorded using a navigation system. PROMs were assessed using the pain, symptoms, and activities of daily living (ADL) subscales of the Knee injury and Osteoarthritis Outcome Score (KOOS) preoperatively and at 1 year postoperatively. Results: RC occurred in 10% of knees (n = 25) and was associated with significantly smaller improvements in KOOS pain, symptoms, and ADL subscales compared with knees without RC (p < 0.01, p = 0.02, and p = 0.01, respectively). Knees with RC showed lower intraoperative tibial internal rotation during knee flexion than knees without RC. Conclusions: Perioperative changes in femorotibial rotational alignment worsened intraoperative knee kinematics, as reflected by reduced tibial internal rotation during knee flexion, and were associated with poorer clinical outcomes 1 year after TKA. To minimize postoperative changes in femorotibial rotational alignment, components should be positioned with accurate rotational alignment.

ABSTRACT. Loss of skeletal muscle mass in older adults is associated with functional decline, disability, and mortality. Oral hypofunction, a multidimensional impairment of oral function, has been increasingly recognized as a factor related to nutritional status and systemic health; however, evidence remains limited among patients with end-stage knee osteoarthritis (KOA), who are at high risk of sarcopenia due to pain-related mobility restriction. This study investigated whether oral hypofunction is independently associated with skeletal muscle mass index (SMI) in elderly women with end-stage KOA scheduled for total knee arthroplasty (TKA), with particular emphasis on adjustment for objectively measured physical activity. In this retrospective observational study, 41 elderly women scheduled for primary TKA were analyzed. Oral hypofunction was diagnosed using standardized criteria based on impairment in three or more of seven oral functional domains. Preoperative physical activity was objectively measured for 30 consecutive days using a wearable activity monitor. Body composition, including SMI, was assessed by bioelectrical impedance analysis. Multivariable linear regression analysis was performed with SMI as the dependent variable, adjusting for age, body weight, oral hypofunction, and physical activity. Oral hypofunction was present in 22 patients (53.7%). Age and objectively measured physical activity did not differ significantly between patients with and without oral hypofunction. In multivariable analysis, oral hypofunction was independently associated with lower SMI (β = 0.257, p = 0.003) after adjustment for age, body weight, and physical activity. The regression model demonstrated strong explanatory power (adjusted R² = 0.749). Oral hypofunction was independently associated with reduced skeletal muscle mass in elderly women with end-stage KOA scheduled for TKA, irrespective of objectively measured physical activity. These findings suggest that oral functional impairment may represent a clinically relevant and potentially modifiable factor related to sarcopenia risk in this surgical population.