Summary
This study revealed that the pre-operative surgical planning software overestimates predicted ROMs compared to measured ROMs.
Abstract
Introduction
The advent of pre-operative three-dimensional (3D) planning software offers significant potential for enhancing the precision of implant alignment, sizing, and positioning, thereby optimizing glenohumeral range of motion (ROM). These predictions often involve glenohumeral motion around a fixed scapula, and fail to account for soft tissue constraints, variations in muscle strength, and tissue quality. Consequently, the accuracy of predicted ROMs may be compromised. Therefore, this study aimed to compare predicted ROMs with the measured glenohumeral ROMs of RSA patients using biplane video radiography (BVR). Given the limitations of current ROM predictions, it was hypothesized that the measured ROMs would be smaller than the predicted ROMs.
Methods
Twelve subjects who underwent RSA (7 males and 5 females; mean age 69 ± 8.8 years; body mass index (BMI) 29.3 ± 7.9 kg/m²; 71.4 ± 17.5 weeks post-surgery) provided informed consent to participate in this IRB-approved study. All subjects adhered to a standardized post-operative protocol. Pre-operative planning was performed using a commercial 3D CT software to optimize implant positioning and generate predicted impingement-free ROMs along anatomical planes. Full active ROM for axial rotation, lateral elevation (abduction), and frontal elevation (flexion) of the shoulder were measured using BVR at a sampling frequency of 40 Hz. Subjects were instructed to perform these movements at a rate of 70 bpm controlled by a metronome.
Three-dimensional, patient-specific models of the humerus and glenoid implants were provided by the manufacturer, each featuring pre-defined anatomical coordinate systems for the humerus and scapula, respectively. These models were registered to bilateral X-ray images obtained from BVR using the DSX Suite (HAS-motion) software. Following registration, the models were imported into Visual3D (HAS-motion) to compute Euler angles, representing the orientation of the humeral implant relative to the glenoid implant. The range of motion (ROM) for each trial was calculated and averaged for each movement type. Post-operative ROMs were compared to the pre-operative predicted impingement-free ROMs for each movement using a paired t-test, with significance level (α) set at 0.05. Bonferroni corrections were applied to account for multiple comparisons.
Results
The predicted impingement-free ROMs was significantly larger compared to what was measured post-operatively. Specifically, the differences were identified during axial rotation (measured ROM mean ± SD = 62.6 ± 22.4 vs. predicted ROM = 140.7 ± 15.2; p < 0.01, η2 = 0.96, O. power = 1.00), abduction (measured ROM = 44.9 ± 24.3 vs. predicted ROM = 101.1 ± 12.5; p < 0.01, η2 = 0.84, O. power = 1.00), and flexion (measured ROM = 133.7 ± 58.1 vs. predicted ROM = 201.0 ± 42.45; p < 0.01, η2 = 0.57, O. power = 0.93).
Discussion
This study revealed that the pre-operative surgical planning software overestimates predicted ROMs compared to measured ROMs. This finding contrasts with previous studies that reported underestimations when using clinical measurements, which cannot differentiate between glenohumeral and scapulothoracic contributions to shoulder motion. The precise isolation of glenohumeral kinematics in this study provides a more accurate assessment. Despite the small sample size, the large effect sizes suggest practical significance, though further research with larger cohorts and long-term follow-up is needed to validate these results and assess their impact on patient outcomes. Addressing the factors contributing to these discrepancies and developing advanced preoperative planning tools that incorporate dynamic factors and patient-specific data could significantly enhance the accuracy of ROM predictions and improve surgical outcomes.