Authors: Jaden S. Queen, John AR. Ferrara, Jeffrey A. Geller
Categories: Original Research, Total hip arthroplasty, Hip dislocation, AI
Source: Arthroplasty Today
Authors: Jaden S. Queen, John AR. Ferrara, Jeffrey A. Geller
Dislocation after total hip arthroplasty (THA) is a devastating complication. The hip-spine relationship is a significant contributor to hip instability and dislocation after THA but is predominantly evaluated with static radiographs, limiting its utility. This study evaluated a novel artificial intelligence (AI)-based application for real-time analysis of hip-spine motion prior to THA to dynamically evaluate patients’ hip-spine stiffness in real-time prior to THA.
Preoperative hip and spine flexibility were assessed using an AI application that recorded patients performing sit-to-stand, forward flexion, and standing posture maneuvers. Minimum and maximum neck, spine, trunk, and knee angles were measured preoperatively. Preoperative radiographs were also evaluated for spinal stiffness indicators. Acetabular component abduction and anteversion angles were measured to confirm adequate positioning.
Nineteen patients underwent THA via an anterior-based muscle-sparing approach with a minimum 12-month follow-up. The mean preoperative forward flexion trunk angle was 95.7° ± 14.4° (25th ≤87.2°). During sit-to-stand, mean maximum and minimum spine angles were 38.3° ± 13.3° (25th ≤27.6°) and 5.1° ± 5.9° (75th ≥6.2°), respectively. Fifteen patients (78.9%) received 36-mm femoral heads. Mean abduction and anteversion was 43.9° and 26.4°, respectively. No postoperative hip dislocations occurred.
This AI-based hip joint assessment tool may serve as a clinic-based tool to evaluate the hip-spine relationship as a dynamic predictor of dislocation risk. It may offer greater accuracy than static radiographs, which cannot comprehensively capture real-time functional movements. This tool may improve surgical planning, particularly in higher-risk patients. Larger studies are needed to validate its predictivity and clinical utility.
Total hip arthroplasty (THA) is a widely performed procedure to improve hip function in patients with degenerative joint conditions such as osteoarthritis. Although THA is generally successful, it is associated with several postoperative complications, including but not limited to dislocation, infection, fracture, and aseptic loosening [1]. Among these, dislocation remains one of the most prevalent complications, with reported rates ranging from 0.12% to 16.13% [2,3]. This variation in dislocation rates may be associated with factors such as operative approach, surgical technique, as well as prosthetic type [4,5].
There has been increased attention on potential preoperative planning tools and the use of technology to evaluate preoperative risk factors for dislocation after THA [5]. The hip-spine relationship, as well as sites of bony impingement, has become an important consideration in assessing postoperative dislocation risk. Patients with lumbar spine deformities, spinal stiffness, or a history of lumbar spine fusion are at an increased risk of dislocation following THA [6,7]. Currently, the evaluation of the hip-spine relationship for assessing dislocation risk is primarily conducted through either just physical examination, radiographic analysis, or both, which has notable limitations in its utility. Though flexion and extension radiographic views attempt to recreate a more dynamic sense of a patient's range of motion via more static pictures, these images may not take into account muscle tone, body habitus, or a patient’s inherent flexibility.
In this study, we aim to utilize a novel digital application powered by artificial intelligence (AI) to assess a patient’s preoperative hip-spine flexibility and characterize their risk of dislocation after THA. An AI-driven application could offer a powerful and efficient tool for improving our understanding of a patient’s dislocation risk in clinical practice.
This study was conducted with institutional review board approval (IRB-AAAU9353). Patients being evaluated for primary THA due to end-stage osteoarthritis were screened using the in-clinic digital application. After clinical evaluation and radiographic review, patients indicated for THA were offered enrollment. Informed consent was obtained prior to participation.
Preoperative hip-spine flexibility was assessed using the ExCalibur AI application (Exer Health, Inc. Denver, CO), which captures video of prescribed patient movements and extracts hip joint and spine angle data. Patients were instructed to perform 3 standardized functional maneuvers, including the sit-to-stand (STS) and stand-to-sit transitions, a forward flexion or toe-touch maneuver, and upright standing posture. Patients were positioned in front of an Apple iPad in the sagittal plane. They are asked to complete the aforementioned series of movements which are captured by the optical camera and digitized in the application into a completely blinded stick figure depiction highlighting key joints for identifying and capturing full body motion (Fig. 1a-d). The application’s AI algorithm then automatically calculated minimum and maximum angles of the spine, trunk, neck, and knees for each maneuver, generating a comprehensive biomechanical profile reflective of functional hip-spine mobility (Figure 2, Figure 3, Figure 4).Figure 1(a-d) Sequential images of an example of the data capturing process.Figure 2Example of sit-to-stand maneuver and measurements.Figure 3Example of forward flexion maneuver and measurements.Figure 4Example of standing posture maneuver and measurements.
All THA procedures were performed by a single surgeon using an anterolateral, abductor-sparing/anterior-based muscle-sparing approach, entering between the gluteus medius and tensor fascia lata on a standard operating room table. Preoperative templating guided surgical planning. Acetabular component positioning was confirmed with intraoperative fluoroscopy to ensure appropriate cup abduction and anteversion positioning. Cementless stems were inserted at approximately 15 degrees of anteversion relative to the native femoral epicondylar axis. The largest feasible femoral heads, typically 36 mm, were used to optimize THA stability. Intraoperative verification of leg length, offset, and hip stability was performed fluoroscopically and clinically.
Standard anteroposterior and lateral radiographs were obtained postoperatively. Radiographic analysis focused on measuring acetabular cup abduction and anteversion angles to confirm proper prosthetic positioning with a general bias toward the Lewinnek safe zone, but intraoperative matching of patients’ native acetabular position for the acetabular component took precedence. Preoperative lumbar spine radiographs were also evaluated for indicators of spinal stiffness, including degenerative changes and/or prior spinal fusion.
Descriptive statistics were calculated for demographic data, motion metrics, and radiographic measurements. Percentile thresholds were explored to identify potentially meaningful cutoffs in preoperative flexibility. All data analysis was performed using Python, and due to the exploratory nature of the study, no formal hypothesis testing was conducted.
Twenty patients completed the preoperative ExCalibur assessments. Nineteen of those patients subsequently underwent THA (one patient ultimately canceled their THA). The cohort had a nearly equal distribution of men (47.3%) and women (52.6%), with a mean age of 66.4 years (range, 43 to 84) and a mean body mass index of 30.1 (range, 20.4 to 41.6) (Table 1). All participants were successfully assessed using the AI application. There were no patients with a history of surgical spinal fusion, although one patient demonstrated radiographic signs of multilevel degenerative spinal deformities.Table 1Patient demographics.DemographicN = 19Age, mean (y)66.4Sex, n (%) Men9 (47.3) Women10 (52.6)BMI, mean (range)30.1 (20.4–41.6)
During the forward flexion maneuver, the mean maximum trunk angle was 95.7 degrees with a standard deviation of 14.4, ranging from 73.0 to 117.9 degrees. The 25th percentile cutoff for trunk flexion was 87.2 degrees or less. In the STS task, the mean maximum spine angle was 38.3 degrees with a standard deviation of 13.4 (range, 23.7 to 70.6 degrees), and the 25th percentile threshold was 27.6 degrees. The mean minimum spine angle during STS was 5.1 degrees (range, 0.2 to 24.4 degrees), with a 75th percentile threshold of 6.2 degrees. The mean maximum knee angle during STS was 93.5 degrees with a standard deviation of 9.4 (range, 79.4 to 111.8 degrees). The mean minimum knee angle during STS was 13.0 degrees with a standard deviation of 5.4 (range, 2.4 to 26.7 degrees) (Table 2). These motion metrics were automatically collected and consistently measured across all patients, indicating the feasibility and reproducibility of the AI-assisted assessment process.Table 2Maneuver Measurements.ManeuverMeasureMean ± SDRangePercentile thresholdFFTrunk angle (°)95.7 ± 14.473.0–117.925th ≤ 87.2°STSMax spine angle (°)38.3 ± 13.323.7–70.625th ≤ 27.6°STSMin spine angle (°)5.1 ± 5.90.2–24.475th ≥ 6.2°STSMax knee angle (°)93.5 ± 9.479.4–111.825th ≤ 86.7STSMin knee angle (°)13.0 ± 5.42.4–26.725th ≤ 10.2BMI, body mass index; FF, forward flexion; SD, standard deviation.
All patients underwent successful surgery with no intraoperative complications. Fifteen patients (78.9%) received 36-mm femoral heads, with 3 32-mm (15.8%) and 1 40-mm (5.3%) femoral heads used in the remaining 4 patients due to the smaller or larger sized acetabular components, respectively. Intraoperative and postoperative radiographs confirmed accurate component placement. The mean cup abduction angle was 43.9 degrees, and the mean anteversion was 26.4 degrees. No patient experienced a postoperative dislocation during the 1-year follow-up period.
This study demonstrates the feasibility of using AI technology to evaluate dynamic hip-spine motion in patients undergoing THA. In this study, the studied application (ExCalibur) provided an efficient, real-time method for quantifying patient-specific flexibility patterns during common functional maneuvers. Unlike traditional radiographs, which offer only static alignment, this tool enables dynamic analysis of movement sequences associated with dislocation risk.
This pilot study indicates that this technology could potentially be applied to classify thresholds of dynamic hip-spine flexibility. Notably, trunk flexion angles below 87.2 degrees and STS spine angles below 38.7 degrees characterized the lower quartile of patients, suggesting these individuals may have more restricted motion (stiffness) and potentially altered lumbopelvic kinematics. While no dislocations were observed in this 1-year follow-up cohort, larger, prospective studies could validate the accuracy and predictive potential of these mobility thresholds. This is notable, as recent research has revealed the limitations of prosthesis planning guidelines such as the Lewinnek safe zone used in this study. Tezuka et al. reported that 14.2% of hip replacements within the traditional Lewinnek safe zone were outside of the sagittal plane functional safe zone, and that femoral and spinopelvic mobility, as well as pelvic incidence, are predictive of hip dislocation after THA [8]. The present study demonstrates this application’s ability to accurately measure dynamic, sagittal plane hip alignment variables, though its predictive capacity remains unknown.
The absence of postoperative dislocations in this series may reflect not only the value of functional assessment, but is also likely related to the anterior approach surgical technique, which historically has a lower dislocation rate compared to posterior-based hip approaches. During this study period, the AI-derived mobility data may have also contributed to surgical planning by identifying limited flexibility in patients who may not have had more defined identifiers such as degenerative lumbar stiffness or simply a lack of spinal radiographic imaging. However, the degree to which the application improved surgical planning is uncertain; all procedures used an anterior-based approach with intraoperative fluoroscopic guidance, cementless components, and large-diameter femoral heads, all factors known to improve surgical outcomes [[9], [10], [11], [12], [13], [14], [15]]. For these reasons, this pilot study is unable to definitively quantify the utility of this technology. Larger studies are needed to compare this technology to standard radiographic or functional spinopelvic assessments, particularly in high-risk patients. However, this study does highlight the practicality and feasibility of this AI-enhanced functional assessment, as its relatively simple, space-and-time-efficient administration is well-suited to the modern clinical setting. If further investigation yields predictive capability comparable to current standards, this technology could be a useful boon in an environment where time is a valuable resource that must be allocated efficiently across many patients.
This study has limitations, including a small sample size, relatively short-term follow-up, and lack of comparative data. It is also subject to confounding, as this was only a single tertiary care surgeon’s case series, with a single surgical anterior-based technique to THA. It is not yet known whether the angle thresholds identified here can reliably predict dislocation or other adverse events. Nonetheless, these findings provide the groundwork for the clinical utility of incorporating functional biomechanical analysis into the preoperative evaluation of THA candidates. Larger, multicenter studies with longer-term outcomes will be essential to validate these metrics and define their role in risk stratification and surgical decision-making.
This initial evaluation supports the exploration of further integration of AI-based motion analysis into the preoperative workflow for THA as a feasible and informative adjunct to conventional imaging. By capturing real-time hip-spine dynamics through this AI-enhanced functional assessment, this application (ExCalibur) may offer a promising approach to identifying patients with limited mobility who may be at increased risk for instability after THA. Future research with broader cohorts and extended follow-up will be critical to establishing clinically meaningful thresholds and determining whether this technology can enhance surgical planning and reduce postoperative complications.
J.A. Geller receives royalties from Smith & Nephew; is on the speakers bureau/paid presentations for Smith & Nephew; is a paid consultant for Smith & Nephew; receives research support from Orthopaedic Scientific Research Foundation, OrthoSensor, and Smith & Nephew; is a member of the medical/orthopaedic publications editorial/governing board at Clinical Orthopaedics and Related Research, Journal of Arthroplasty, and Journal of Bone and Joint Surgery - British; all other authors declare no potential conflicts of interest.
For full disclosure statements refer to https://doi.org/10.1016/j.artd.2026.102070.
Jaden S. Queen: Writing – review & editing, Writing – original draft, Formal analysis, Data curation. John AR. Ferrara: Writing – review & editing, Writing – original draft, Formal analysis, Data curation. Jeffrey A. Geller: Writing – review & editing, Supervision, Methodology, Formal analysis, Conceptualization.