Evaluating Bone Loss in Anterior Shoulder Instability

Eric C. Makhni, MD, MBA; Joseph S. Tramer, MD; Matthew J.J. Anderson, MD; William N. Levine, MD


J Am Acad Orthop Surg. 2022;30(12):563-572. 

In This Article

Assessment of Glenoid Bone Loss


After an episode of traumatic anterior shoulder instability, standard radiographs of the glenohumeral joint including AP, Grashey, axillary, and scapular y views should be obtained. In addition to assessing for congruity of the glenohumeral joint, radiographs can be helpful in identifying glenoid and humeral bone loss. On the AP view, for instance, an intact anterior glenoid rim typically appears as a continuous sclerotic line. Jankauskas et al[6] demonstrated that loss of the sclerotic glenoid line on the AP radiograph was highly specific but only moderately sensitive for identifying anterior glenoid rim deficiencies compared with CT. Consistent with this finding, Auffarth et al asked six observers to review conventional radiographs (AP and axillary) of patients with a first-time shoulder dislocation and found that of the 10 patients who presented with a glenoid rim fracture (confirmed on CT), each investigator overlooked at least one fracture (range, 1 to 4) based on radiographs alone. Accordingly, the authors recommended CT evaluation in all patients after primary dislocation.[7] However, additional radiographic projections have also been devised to facilitate the identification and evaluation of glenoid bone.

The West Point view (Figure 1) is helpful for identifying osseous Bankart lesions of the anteroinferior glenoid rim. The West Point view is obtained with the patient lying prone, the shoulder slightly elevated, and the arm abducted to 90° hanging over the edge of the table. The radiograph tube is oriented inferosuperior, 25° medial, and 25° anterior so that it is tangential to the anteroinferior rim of the glenoid. Itoi et al conducted a cadaveric study that involved obtaining radiographs of progressively larger glenoid defects and found that changes in the glenoid width were more appreciable on the West Point view compared with the axillary view. However, the West Point view was still less accurate than CT for evaluating anteroinferior bony Bankart lesions, leading the authors to conclude that although radiographs represent an acceptable screening tool, CT is notably more accurate for measuring bone loss.[8]

Figure 1.

West point radiograph view demonstrating patient positioning. The patient is in the prone position with the forearm hanging off the table. The radiograph beam is centered on the axilla and aimed at 25° downward from the horizon (A) and 25° medial to the plate (B). Radiograph view (C) demonstrating a view of the anterior glenoid rim with an osseous lesion at the anterior-inferior glenoid (arrow). Figure modified from Rockwood and Green Fractures in Adults, 9th edition.

The Bernageau glenoid profile view (Figure 2) is another radiographic projection that can be used to assess for bony defects that are more anterior along the glenoid face.[9,10] For the Bernageau view, the patient is standing with the affected arm forward flexed to 160° and the thorax in contact with the cassette at an angle of 70°. The radiograph tube is centered over the scapular spine with a caudal inclination of 30°. Murachovsky et al compared the Bernageau view with three-dimensional (3D) CT for quantifying glenoid bone loss and found no difference between the two imaging modalities. Although the authors did not recommend radiographs in lieu of 3D CT for preoperative planning, the Bernageau view was identified as an accurate and reproducible technique for identifying and measuring glenoid bone loss.[11] Among patients undergoing surgical treatment of chronic anterior shoulder instability, Edwards et al[9] found osseous abnormalities of the glenoid in 78.8% of shoulders using the Bernageau view but noted that very inferior fractures can be difficult to visualize on this projection.

Figure 2.

A: Patient positioning for taking the Bernageau view of the shoulder. B: Bernageau view demonstrating an intact anterior glenoid rim (blue arrows). Reference: Rockwood and Green Fractures in Adults, 9th edition.


Given the overall difficulty of reliably identifying and measuring glenoid bony defects using radiographs alone, CT should be obtained in patients with evidence of bone loss on radiograph, patients who have experienced recurrent instability, and patients who have failed prior instability surgery.[12] A multitude of methods have been devised for quantifying glenoid bone loss based on both two-dimensional (2D) and 3D CT. A benefit of 3D CT is that the humerus can be subtracted from the digital reconstruction of the shoulder joint, allowing for a perfect en face sagittal view of the glenoid surface. Most methods of measuring glenoid bone loss use a "best-fit circle" technique, which is based on a cadaveric study by Huysmans et al[13] that found the inferior glenoid roughly constitutes a true circle. Viewing the glenoid articular surface en face, a circle is drawn centered about the glenoid bare area, using the intact posteroinferior glenoid as a reference (Figure 3, A). The area of the circle that does not overlap with the glenoid anteriorly is presumed to be bone loss and can be compared with the area of the entire circle to calculate percent bone loss (Figure 3, B).[14] The Pico method, developed by Baudi et al,[15] uses a 3D CT of the patient's contralateral uninjured glenoid to generate a best-fit circle, which is superimposed on the injured glenoid to determine percent bone loss. Barchilon et al[16] devised a relatively simple mathematical function to estimate glenoid bone loss based on the ratio of the depth of the defect (a perpendicular line from the center of the best-fit circle to the anterior edge of the glenoid) and the radius of the best-fit circle (Figure 4). Similarly, Dumont et al[17] described a method for determining percent bone loss by measuring the arc angle that subtends the area of glenoid bone loss as defined by a best-fit circle (Figure 5).

Figure 3.

Sagittal oblique projection of the glenoid fossa from a three-dimensional CT reconstruction. The intact posteroinferior glenoid has been used as a template to overlay a "best-fit circle" centered over the glenoid bare area (A). The "best-fit circle" method (B) can be used to estimate glenoid bone loss (blue). Percent bone loss may be determined by dividing the area in blue by the entire area within the "best-fit circle" (blue + red). Image courtesy of Dr. Eric Makhni.

Figure 4.

Linear method of estimating glenoid bone loss as described by Barchilon et al.14 A mathematical equation is used to measure the area of defect as a function of the ratio between the depth (d) to the glenoid defect margin (a perpendicular line from the erosion edge (red line) to the center of the "best-fit circle") and the radius (R) of the intact inferior glenoid rim, as defined by the "best-fit circle." This ratio is then used to determine the percent bone loss. Image courtesy of Dr. Eric Makhni.

Figure 5.

Arc angle method of calculating glenoid bone loss as described by Dumont et al. A circle is superimposed on a sagittal view of the glenoid, using the inferior border of the glenoid as a reference. The glenoid arc angle (α) that subtends the area of bone loss (shaded in yellow) is measured and used to calculate percent bone loss with the equation provided.17

CT may also be beneficial in differentiating acute bony Bankart lesions amenable to repair from attritional glenoid bone loss with resorption of the bone fragments, which often requires grafting to reconstitute the bony architecture of the glenoid (Figure 6). When possible, incorporation of the bony Bankart fragment into the capsulolabral repair leads to improved patient-reported outcomes and lower rates of recurrence, particularly when the glenoid defect is large (>20%).[18] Nakagawa et al[19] examined serial CT scans of patients who underwent arthroscopic bony Bankart repair and found that larger fragments were more likely to unite and that union of the bony fragment decreased the glenoid defect from 18.6% to 4.7% on average. As expected, nonunion of the bony Bankart fragment was a positive predictor of recurrent instability. However, longitudinal assessment of bony Bankart fragments has demonstrated a correlation between bone loss and time from the initial trauma, with severe resorption observed at 1 year after primary dislocation.[20] Accordingly, timely treatment of acute bony Bankart lesions with incorporation of the bone fragment into the repair may decrease the risk of recurrent instability without the need for bone grafting.

Figure 6.

En face views of three-dimensional CT reconstructions of the glenoid. A, The glenoid of a patient suffering from recurrent instability demonstrating attritional bone loss with several small bone fragments. B, The glenoid of a patient who experienced a traumatic first-time dislocation event resulting in a large bony Bankart lesion. Image courtesy of Dr. Eric Makhni.


Although MRI is used to evaluate for soft-tissue injuries associated with shoulder instability, there is increasing evidence that MRI can also be used to reliably and accurately assess bone loss.[21] Lee et al[22] compared MR arthrography with CT in the evaluation of glenoid bone loss and found excellent correlation, with strong interobserver and intraobserver correlations of MR arthrography–derived measurements of bone loss. In addition to allowing for concomitant evaluation of soft tissues and bone, MRI does not involve ionizing radiation, and the introduction of 3 tesla(T) magnets has dramatically improved the acquisition speed, signal, and spatial resolution of MRI.[23] Moreover, it is now possible to generate 3D reconstructions using MRI, further increasing its utility in assessing glenohumeral bone loss. For all these reasons, MRI is now included in the basic workup of anterior shoulder instability for many orthopaedic surgeons.[23]

Gyftopoulos et al[21] described using 3T and 1.5T MRI with dedicated 16-channel shoulder array coils to produce an axial 3D dual echo time T1-weighted sequence with Dixon-based fat-water separation, which was then used to generate a 3D reconstruction of the glenohumeral joint. Glenoid bone loss was calculated from the 3D reconstruction using the best-fit circle method and found to be consistent with findings on arthroscopy. Using a similar imaging protocol with Dixon-based fat-water separation MRI, Lansdown et al[24] found a strong correlation between estimates of bone loss based on 3D MRI reconstructions and 3D CT reconstructions (Figure 7), further supporting the notion that CT may not be necessary if a 3D MRI will be obtained. However, 3D MRI is not yet widely available, and CT remains the benchmark for detecting significant bone loss in patients with anterior shoulder instability, with a sensitivity approaching 100% compared with 35.3% for standard 2D MRI.[25]

Figure 7.

Views of three-dimensional reconstructions of the glenoid from the same patient using (A) CT versus (B) MRI. Reference: Lansdown DA, Cvetanovich GL, Verma NN, et al: Automated 3-Dimensional Magnetic Resonance Imaging Allows for Accurate Evaluation of Glenoid Bone Loss Compared With 3-Dimensional Computed Tomography. Arthroscopy 2019;35:734–740.