Measurement of the Balance Stability Angle to Predict the Stability Ratio in Patients With Recurrent Anterior Shoulder Dislocation: A Novel Computed Tomography–Based Protocol

Abstract

Background:

The stability ratio (SR) is an important biomechanical parameter for evaluating glenoid stability in patients with recurrent anterior shoulder dislocation (RASD), and it cannot be practically and conveniently measured in clinical scenarios.

Purpose:

To investigate a novel computed tomography (CT)–based protocol to estimate the SR efficiently.

Study Design:

Descriptive laboratory study.

Methods:

A total of 102 patients with RASD were included. Demographic information, CT scans, and bone defect area (BDA) were collected. The new protocol, based on balance stability angle (BSA) measurements on CT, was conducted to estimate the SR (SR_CT) by 2 surgeons independently. Biomechanical testing was then performed on patient-specific 3-dimensional (3D)–printed glenoid models to calculate the SR (SR_3Dprint), which was used to (1) analyze the reliability of SR_CT and (2) examine if the BDA could predict SR_3Dprint. To validate whether the 3D-printed glenoid could reflect the actual biomechanical properties of the shoulder, the SR from 5 cadaveric glenoid specimens (SR_cadaver) was also calculated and compared with that from the 3D-printed glenoid (SR_3Dprint) under 6 osteotomy conditions. Linear regression and intraclass correlation coefficients (ICCs) were used for statistical analysis.

Results:

The interrater reliability of SR_CT measurements was high (ICC = 0.95). SR_CT was highly correlated with SR_3Dprint (R ² = 0.86; ICC = 0.92). The mean BDA was 11.44% ± 6.72% by the linear ratio method, with a weak correlation with SR_3Dprint (R ² = 0.31; ICC = -0.46). The cadaveric validation experiment indicated that SR_cadaver was highly correlated with SR_3Dprint (R ² = 0.86; ICC = 0.77).

Conclusion:

Results indicated that (1) the proposed CT-based protocol of obtaining BSA measurements is promising for the SR estimation in patients with RASD, (2) the BDA was not an effective parameter to predict the biomechanical SR, and (3) the 3D-printed glenoid could reflect the biomechanical properties of cadaveric shoulders regarding the SR estimation.

Clinical Relevance:

Traditional BDA measurements cannot accurately reflect the biomechanical stability of the glenoid. The newly proposed CT-based protocol is practical for surgeons to estimate the SR.

The stability ratio (SR) is an important biomechanical parameter in evaluating the concavity-compression effect in shoulder midrange motion. ⁸ The SR is defined as the ratio of 2 forces: one is the force necessary to displace the humeral head and the other is the joint compressive load when dislocation occurs. ¹² Because the SR is highly correlated with the 3-dimensional (3D) shape of the articular surface, it is able to account for more 3D biomechanical implications in patients with recurrent anterior shoulder dislocation (RASD) than traditional 2-dimensional measurements of the bone defect area (BDA). ¹⁵

Although the SR has existed as an important biomechanical parameter for decades, it is still difficult to apply in clinical scenarios because of limitations in measuring methodologies. To be specific, traditional cadaveric tests require customized biomechanical testing platforms ^{7,10,11,16,23,24} and cannot be performed on patients. Finite element analysis (FEA) can be conducted to calculate the SR for specific patients; nevertheless, it also requires technology and is time-consuming for surgeons. ^{9,15,16,18,21,26} In clinical practice, the load and shift test is usually performed to examine the SR of patients with RASD, but the result is qualitative and dependent on the experience of the examiner.

A method to efficiently estimate the SR of patients with RASD is necessary. In this study, we investigated a practical and convenient method for calculating the SR and to promote its clinical application among patients with RASD. Our hypothesis was that the SR estimated by the newly proposed computed tomography (CT)–based protocol would be highly correlated with the SR directly calculated by biomechanical testing. This novel protocol could potentially be an efficient method for evaluating the SR.

Methods

The study protocol received ethics committee approval. Patients who underwent surgery for RASD at our institution between January 1, 2016, and December 31, 2019, were retrospectively screened from a clinical database. Informed consent was signed by final patients included. The inclusion criteria were as follows: RASD with or without visible bony glenoid defects and available preoperative narrow-sliced CT scans of the shoulder. The exclusion criteria were scapular fractures, joint arthropathy, and incomplete radiological data.

The following clinical data were collected: demographic information (sex, age, affected side), BDA (determined according to linear and area ratio methods proposed by Barchilon et al ² ), and preoperative narrow-sliced shoulder CT scans (thickness, 0.6 mm; resolution, 512 × 512 pixels; DICOM format). As this study only focused on the relationship between radiological parameters and biomechanical testing results, irrelevant clinical information (eg, number of instability events before surgery, physical examination findings, surgical details, clinical function scores, etc) was not obtained.

There were 2 types of protocols adopted to calculate the SR: (1) a novel CT-based protocol for SR_CT based on balance stability angle (BSA) measurements after multiple-plane reconstruction of CT data and (2) a biomechanical protocol using a customized inclination platform that could directly calculate the SR of the glenoid. For the included patients, patient-specific 3D-printed glenoid models were used for biomechanical testing to calculate SR_3Dprint. Based on the protocols mentioned above, 2 questions were investigated: (1) Could SR_CT obtained by the CT-based protocol predict SR_3Dprint obtained from biomechanical testing? (2) Could the BDA (measured using the methods of Barchilon et al) predict SR_3Dprint obtained from biomechanical testing? As there are morphological differences between a 3D-printed glenoid model and the real human glenoid, a cadaveric validation experiment was additionally conducted to explore if a patient-specific 3D-printed glenoid is a reliable model regarding the SR calculation.

A total of 114 patients were assessed for eligibility, and 12 patients were excluded (7 patients did not have complete CT data involving the whole scapula, and 5 patients had preoperative CT scans that could not be read) (Figure 1).

Figure 1.

Overview of study procedures. 3D, 3-dimensional; BDA, bone defect area; BSA, balance stability angle; CT, computed tomography; ICC, intraclass correlation coefficient; SR_CT, stability ratio based on BSA measurement on CT; SR_3Dprint, stability ratio calculated by biomechanical testing using 3D-printed glenoid model; SR_cadaver, stability ratio calculated by biomechanical testing using cadaveric shoulder specimen.

CT-Based Protocol of BSA Measurements for Calculating SR_CT

The BSA is the maximal angle formed by the net force on the humeral head and the glenoid centerline before dislocation occurs. ¹¹ According to the geometric properties, the tangent value of the BSA equals the SR. ¹¹ This is the theoretical foundation of the CT-based protocol (Figure 2A). The BSA measurement was independently performed by 2 qualified sports medicine surgeons (Q.H. and G.Z.), and the average BSA value was used for SR_CT calculation (SR_CT = tan[BSA]).

Figure 2.

(A) Geometric diagram of the balance stability angle (BSA) and (B) laboratory setup of measuring the stability ratio (SR) using an inclination platform. CT, computed tomography; F_C , compressive force; F_D , displacing force; mg, gravity of the humeral head; α, the angle formed by the glenoid centerline and the tangent line (a 3-mm line in this study) at the steepest point of the glenoid rim; γ, inclination angle.

The radiological measurement was performed using Radiant software. The multiple-plane reconstruction function was adopted to measure the BSA (for a detailed protocol, see Supplementary Video). Briefly, there were 3 steps to measure the BSA. The first step was to identify the scapular coronal plane, which was performed by locating 3 specific anatomic points in turn: the inferior scapular tip, the medial root of the scapular spine, and the center of the glenoid surface (the lower glenoid circle). The second step was to identify the glenoid centerline by moving or rotating the axis to ensure that the axis, passing through the glenoid center point, was perpendicular to the glenoid surface in the scapular coronal plane. The third step was to measure the BSA. To eliminate errors of measurement, the “Avg” mode in the “Thickness” drop-down menu was chosen. In this study, the SR in the 3-o’clock direction in the right shoulder (9-o’clock direction in the left shoulder) was studied, as it was deemed that the 3-o’clock direction was an important direction when dislocation occurred. ¹⁵ The anterior glenoid rim area was magnified, and the steepest area was located. Then, a short line (3-mm length for this study) was drawn, with one end at the aforementioned point and the other at the glenoid joint surface, to form angle α with the glenoid centerline. According to the geometric properties, angle α minus 90° equals the BSA, and SR_CT equals the tangent value of the BSA (Figure 2A).

Biomechanical Protocol for Calculating SR_3Dprint

SR_3Dprint was calculated via biomechanical testing of the 3D-printed glenoid model of corresponding patients on the customized inclination platform.

3D-Printed Glenoid Model

All design and production of the 3D-printed glenoid models were performed by a senior medical engineer (S.Z.) based on CT data. The whole scapula was reconstructed in Mimics 21.0 (Materialise) based on CT data and exported in a stereolithography (STL) format into 3-matic 13.0 (Materialise).

The 3D-printed glenoid models were designed based on the major principles of the CT-based protocol. The first step was to locate the scapular coronal plane, and 3 anatomic reference points were selected: the center of the circle fitted to the rim of the lower glenoid (point 1 in Figure 3A), the medial root of the scapular spine (point 2 in Figure 3A), and the inferior scapular tip (point 3 in Figure 3A). ¹³ Then, the scapular model was cut in the coronal plane. The second step was to locate the glenoid centerline. At this time, the perspective was switched to surface 1 as shown in Figure 3D, and a circle was drawn at point 1 (circle center), ensuring that the circle passed through the steepest point (point b in Figure 3D; the location of point b did not require high accuracy as long as it was in the region of the lower glenoid rim) in the inferior margin of the glenoid and intersected the glenoid outline at the 12-o’clock direction at point a (Figure 3D). The vertical bisector of line ab was deemed the glenoid centerline (line 1 in Figure 3D).

Figure 3.

A 3-dimensional (3D)–printed glenoid model for biomechanical testing. (A, B) The first step was to locate the scapular coronal plane. Then, 3 anatomic reference points were selected to construct the coronal plane: the center of the circle fitted to the rim of the inferior glenoid (point 1), the medial root of the scapular spine (point 2), and the inferior scapular tip (point 3). (C) The scapular model was cut in the coronal plane (surface 1). (D) The glenoid centerline in surface 1 was located (line 1). A circle was drawn with point 1 at the center and passing through the steepest point in the inferior margin of the glenoid (point b) and intersecting the glenoid outline in the 12-o’clock direction (point a). The vertical bisector of line ab was deemed the glenoid centerline. Along the glenoid centerline, point c was positioned at a distance of 15 mm downward from point 1. The osteotomy surface (surface 2) was perpendicular to line 1 and crossed point c. (E, F) After osteotomy of the glenoid, the lateralpart was reserved as the glenoid model and was added to a base. It was assumed that surface 1 divided the glenoid in the 12- to 6-o’clock direction. Next, twelve 2-mm holes were drilled from 1 o’clock to 12 o’clock on the base for orientation. At this point, the model was ready for 3D printing. (G) The 3D glenoid model after printing.

After determining the glenoid centerline, the glenoid was osteotomized perpendicular to line 1. Along line 1, point c was positioned 15 mm downward from point 1 (Figure 3D). The osteotomy surface (surface 2 in Figure 3, D and E) was perpendicular to line 1 and crossed point c (Figure 3, D and E). After osteotomy of the glenoid, the lateral part was reserved as the glenoid model, which was added to a base (Figure 3F). It was assumed that surface 1 divided the glenoid in the 12- to 6-o’clock direction. Next, twelve 2-mm holes were drilled from 1 o’clock to 12 o’clock on the base for orientation and fixation to the biomechanical connecting board (Figure 3F). Then, the STL format of the model was exported for 3D printing (Figure 3G).

Biomechanical Testing

A customized biomechanical testing platform was constructed in this study based on the design of Weldon et al. ²² The inclination angle of the platform could be manually altered. When testing, the 3D-printed glenoid model was mounted on the connecting board via 12 holes in the 3-o’clock direction (right shoulder). A previous study showed that the SR mainly derives from the glenoid arc and has no relation to the size of the humeral head. ⁶ Therefore, a 40 mm–diameter humeral steel ball was chosen and placed on the glenoid to represent the humeral head. ^14,19 If the steel ball remained stable on the glenoid model placed horizontally, the glenoid model was recorded as “stable.” Then, the inclination angle was enlarged gently until the steel ball suddenly rolled off the glenoid (Figure 4, A-C). At the last moment of the quasi-equilibrium state, according to the force analysis, the gravity of the steel ball was decomposed into the compressive force perpendicular to the board (or the glenoid) and the displacing force along the board (Figure 2B). Therefore, the tangent value of the inclination angle is the ratio between the displacing force and the compressive force (Figure 2B), which is SR_3Dprint.

Figure 4.

Biomechanical testing. For a stable glenoid, (A) the stability ratio (SR) in the 3-o’clock direction was measured on an inclination board, and (B, C) the inclination angle was increased until the steel ball rolled off. The tangent value of the inclination angle is the SR. For an unstable glenoid, (D) the SR was measured at 3 o’clock in the opposite direction, (E) the inclination angle was set to a certain angle to keep the steel ball in a balanced state, and (F) the inclination angle was decreased and noted when the steel ball rolled off in the 3-o’clock direction. The negative tangent value of the inclination angle is the SR.

If the steel ball could not keep stable and rolled off the glenoid when set on the 3D-printed model placed horizontally, it was recorded as “unstable.” Then, the 3D-printed glenoid model was remounted on the connecting board at 3 o’clock (right shoulder) in the opposite direction (Figure 4, D-F). The inclination angle was increased until the steel ball on the glenoid was “stable.” At this point, the inclination angle was adjusted toward zero and was noted when the steel ball rolled off (Figure 4F). The negative tangent value of the inclination angle is SR_3Dprint.

Cadaveric Validation of 3D-Printed Glenoid Models

As there are some differences between a cadaveric glenoid and 3D-printed glenoid models (eg, cartilage, friction coefficient, etc), a validation experiment was performed. For both the cadaveric glenoid and the corresponding 3D-printed glenoid model, the SR in the 3-o’clock direction was calculated in the intact condition and after 2, 4, 6, 8, and 10 mm of bone loss. ^1,3,4

Overall, 5 shoulder specimens were collected, and a narrow-sliced CT scan was obtained for 3D printing of the scapula and guide plate design. There were 2 guide plates adopted during the experiment (Figure 5A). The first plate was the base plate, which was used to fix the scapular neck via anatomic matching, thus guiding accurate osteotomy along the plane identical to surface 2 in Figure 3D and 3E and Figure 5B. In addition, the base plate could mount the glenoid toward the connecting board. The second plate was an osteotomy plate (Figure 5, A, C, D). It could be fixed to the base plate guiding glenoid osteotomy to simulate various conditions of bone defects (intact and with 2, 4, 6, 8, and 10 mm of bone loss).

Figure 5.

Cadaveric experiment to validate the biomechanical results of the 3-dimensional (3D)–printed glenoid model. (A) Design of base plates and osteotomy plates based on computed tomography of a right shoulder. (B) Osteotomy guided by base plates. (C, D) Osteotomy to create bone defects guided by the osteotomy plate. (E) A left shoulder specimen and its customized 3D-printed glenoid model under the various testing conditions.

Before testing, the specimens were thawed overnight at room temperature. The scapular region was dissected. The muscles, capsule, and labrum were removed, and only scapular bone and cartilage were preserved for further testing. Apart from the specimens, bony scapular models for each specimen were also produced by 3D printing.

The customized base plates were used to guide preliminary glenoid osteotomy in both the specimens and the 3D bony scapular models along surface 2 in Figure 5B. Next, the glenoids (intact condition) were mounted to the connecting board. The SRs (SR_3Dprint and SR_cadaver) in the 3-o’clock direction were calculated and compared for each 3D-printed glenoid model and specimen as described in Biomechanical Testing. Then, glenoid bone defects (2, 4, 6, 8, and 10 mm of bone loss) were created by a 0.5 mm–thick saw aided by an osteotomy guide. In each condition, the SRs (SR_3Dprint and SR_cadaver) in the 3-o’clock direction were calculated.

Statistical Analysis

Statistical analysis was performed using SPSS Statistics 25.0.0.0 (IBM) and GraphPad Prism 9 (GraphPad Software). Continuous variables are displayed as mean ± SD. Linear regression was conducted to analyze the correlations between parameters. To further examine the consistency between parameters, the intraclass correlation coefficient (ICC) was calculated, and a Bland-Altman plot was produced. ICC values were also calculated for interrater reliability of the BSA measurements. ICC values were interpreted as poor (<0.40), fair/moderate (0.40-0.75), or high (>0.75). A negative ICC may occur when the mean square between the groups is smaller than the mean square within the groups (eg, the 2 parameters were negatively correlated). P < .05 was considered indicative of statistical significance.

Results

Of the 102 included patients (82 male and 20 female), 63 had a dislocated right shoulder, and 39 had a dislocated left shoulder; the mean age was 30.08 ± 12.57 years. The BDAs of the included patients are summarized in Table 1.

Table 1

Patient Demographics (N = 102) ^a

	Value
Sex
Male	82 (80.39)
Female	20 (19.61)
Age, y, mean ± SD	30.08 ± 12.57
Affected side
Left	39 (38.24)
Right	63 (61.76)
Bone defect area, linear ratio method/area ratio method
0%-≤5%	17 (16.67)/38 (37.25)
>5%-≤10%	22 (21.57)/40 (39.22)
>10%-≤15%	37 (36.27)/16 (15.69)
>15%-≤20%	18 (17.65)/5 (4.90)
>20%	8 (7.84)/3 (2.94)

^a Data are reported as n (%) unless otherwise indicated.

Relationship Between SR_CT and SR_3Dprint

The mean SR_CT of 102 patients was 10.52% ± 12.24% (range, –14.78% to 40.65%). The mean SR_3Dprint of 102 patients was 12.11% ± 12.80% (range, –17.63% to 38.39%). According to linear regression, SR_CT was highly correlated with SR_3Dprint (r = 0.97, R ² = 0.86) (Figure 6A). The equation relating SR_CT to SR_3Dprint is as follows: SR_3Dprint = 0.97 × SR_CT + 0.02.

Figure 6.

Linear regression plots, Bland-Altman plots, and intraclass correlation coefficient (ICC) values for (A) SR_CT (stability ratio calculated by computed tomography) versus SR_3Dprint (stability ratio calculated by biomechanical testing using 3-dimensional–printed glenoid model), (B) SR_{CT(observer 1)} (stability ratio calculated by computed tomography for observer 1) versus SR_{CT(observer 2)} (stability ratio calculated by computed tomography for observer 2), (C) SR_3Dprint versus SR_cadaver (stability ratio calculated by biomechanical testing using cadaveric shoulder specimen), (D) BDA_{area ratio} (bone defect area measured by area ratio method of Barchilon et al ² ) versus SR_3Dprint, and (E) BDA_{linear ratio} (bone defect area measured by linear ratio method of Barchilon et al) versus SR_3Dprint.

The ICC between SR_CT and SR_3Dprint was 0.92 (95% CI, 0.89 to 0.95; P < .001). The Bland-Altman plot is shown in Figure 6A with a bias of –0.01 (95% CI, –0.11 to 0.08). Regarding the reproducibility of the protocol (Figure 6B), the ICC between SR_CT obtained from 2 separate observers (Q.H. and G.Z.) was 0.95 (95% CI, 0.93 to 0.97; P < .001), indicating high reproducibility of the newly proposed protocol.

Relationship Between BDA and SR_3Dprint

The mean BDA of the glenoid was 11.44% ± 6.72% (range, 0.00%-27.45%) by the linear ratio method and 7.10% ± 5.23% by the area ratio method (range, 0.00%-22.29%). Between SR_3Dprint and the BDA according to the area ratio method, the R ² was 0.33, and the ICC was –0.36 (Figure 6D). Between SR_3Dprint and the BDA according to the linear ratio method, the R ² was 0.31, and the ICC was –0.46 (Figure 6E). The results indicated that the BDA could not be used to predict the biomechanical SR.

Cadaveric Validation of 3D-Printed Glenoid Models

According to linear regression, SR_cadaver was highly correlated with SR_3Dprint (R ² = 0.86) (Figure 6C). The equation relating SR_cadaver to SR_3Dprint is as follows: SR_cadaver = 0.58 × SR_3Dprint + 0.06. The ICC between SR_cadaver and SR_3Dprint was 0.77 (95% CI, 0.37 to 0.91; P < .001). The Bland-Altman plot is shown in Figure 6C with a bias of –0.07 (95% CI, –0.24 to 0.11).

Discussion

The most important finding of this study is that the proposed CT-based protocol was reliable with high reproducibility. It was demonstrated that the relationship between SR_CT and SR_3Dprint (r = 0.97; R ² = 0.86) was strong, with an ICC of 0.92, indicating excellent reliability. In addition, the high ICC value (0.95) between SR_CT from independent observers showed good reproducibility.

Regarding previous radiological studies on the SR estimation, Sigrist et al ¹⁸ explored the relationship between CT-based glenoid parameters (glenoid depth, width, and retroversion) and the SR calculated by FEA. The reported correlation coefficients were 0.7 to 0.8 between the glenoid depth and the SR from 2 o’clock to 5 o’clock. These authors also demonstrated that it is possible to predict the SR by geometric parameters. Moroder et al ¹⁶ proposed a CT-based protocol for the SR prediction by measuring the glenoid width and depth. However, this protocol was not validated by direct biomechanical experiments. Therefore, the reliability of the Moroder et al ¹⁶ protocol is unknown. To the best of our knowledge, this study is the first to report a high correlation between the SR obtained in a radiological manner and that obtained in a biomechanical manner among patients with RASD.

In addition, the results of this study indicated that there was a weak relationship between the BDA and the biomechanical SR (linear ratio method: R ² = 0.31, ICC = –0.46; area ratio method: R ² = 0.33, ICC = –0.36). This finding demonstrated that the BDA was not sufficient to account for the biomechanical impact of the 3D shape of the glenoid surface in patients with RASD, which is also in line with the results of the studies by Moroder et al ¹⁵ and Wermers et al. ²³

Another important finding of this study is that biomechanical testing using a patient-specific 3D-printed glenoid model was found to be reliable. The cadaveric validation experiment demonstrated a strong correlation between SR_cadaver and SR_3Dprint (R ² = 0.86; ICC = 0.77) in the 3-o’clock direction, which indicated that a 3D-printed glenoid could be used to measure the SR on biomechanical testing, especially when a cadaveric glenoid is unavailable. Regarding the biomechanical testing platform that we used (see Figure 4), the adjustable inclined board, automatically decomposing gravity into a compressive load and a displacing load, could even examine unstable glenoid models (severe glenoid bone defects), which cannot be managed by traditional biomechanical platforms. ²² Due to the biomechanical platform, the domain of the SR has been successfully broadened: a negative SR indicates the minimal ratio of a displacing load and a compressive load to maintain the balance of the steel ball, which is extremely helpful to evaluate glenoids with severe bone defects in future studies.

Clinical Relevance

The newly proposed CT-based protocol is an important alternative for the SR measurement. The protocol, based on downloadable DICOM software and patient CT data, does not need any biomechanical platform or FEA. From the perspective of clinical application, the newly proposed protocol could also be used among patients with glenoid bone defects, which cannot be realized by previous methods. This is mainly because of the determination of the glenoid centerline (direction of the compressive load), which was overlooked by Moroder et al. ¹⁶ In the newly proposed protocol, only 1 parameter, the BSA, was measured for the SR calculation. The average measuring time is approximately 7 minutes with high reproducibility. It is practical and convenient for surgeons to preoperatively evaluate patients with RASD in clinical scenarios.

In addition, the 3D-printed glenoid model and inclination board platform make it possible to assess the SR for a specific patient because of the high reliability confirmed by the cadaveric validation experiment. Furthermore, the processing of CT data in Mimics and 3-matic software was simplified by the “script” function, which takes only 5 minutes to design a patient’s customized glenoid model. It is convenient for medical centers equipped with 3D printing devices to examine the SR of the glenoid according to the biomechanical protocol reported in this study.

Limitations

There are some limitations in this study. First, the radiological measurements and the design of 3D-printed glenoids were based on CT data, which did not consider the labrum, cartilage, and other soft tissue. However, the bony structure provides the main portion of the SR of the glenoid, and many previous studies have simplified the influence of the labrum or cartilage. ^{5,16,17,20,25} In addition, biomechanical testing indicated that a strong correlation was found between the cadaveric specimen and the related 3D-printed glenoid model when measuring the SR. Second, the study used a 40-mm steel ball to mimic the humeral head, which meant that the impact of the humeral head diameter on the SR might be neglected. However, a previous study demonstrated that the size of the humeral head is not associated with the SR. ⁶ Indeed, in an incongruent system, it is recommended to calculate the SR based on the glenoid concavity radius. ⁶

A third limitation is that our steel ball–based model could not reflect Hill-Sachs lesions or off-track mechanisms. An off-track mechanism indicates a mechanism of instability when the arm is in the end-range position, but our study mainly focused on midrange motion in which the concavity-compression effect rather than the Hill-Sachs lesion and off-track mechanism plays the major role. ⁸ Fourth, it was impossible to create ideal conditions without friction for SR_3Dprint. Yet, an inclination board was used to minimize the influence of friction as much as possible. Fifth, the study was a laboratory exploration of methodology using only BDA and CT data. As the primary aim of this study was to propose a practical and convenient protocol for surgeons to preoperatively evaluate the SR in patients with RASD, clinical information including surgical details, dislocation times before surgery, or functional scores was not included. In the future, the clinical significance of the SR needs to be further addressed, for example, whether there is a cutoff value for the SR when determining the surgical procedure (bone augmentation or single Bankart repair) that should be performed.

Conclusion

The study findings indicated that (1) a newly proposed CT-based protocol, based on BSA measurements, is promising for the SR estimation in patients with RASD; (2) the BDA was not an effective parameter to predict the biomechanical SR; and (3) a 3D-printed glenoid could reflect the true biomechanical properties of the shoulder regarding the SR estimation.

A Video Supplement for this article is available at https://journals.sagepub.com/doi/full/10.117723259671221140908#supplementary-materials