Use of Preoperative Three-Dimensional Computed Tomography to Quantify Glenoid Bone Loss in Shoulder Instability Tai-Yuan Chuang, M.D., Christopher R. Adams, M.D., and Stephen S. Burkhart, M.D. Purpose: The purpose of this study was to determine if three-dimensional computed tomography (3-D CT) scans of the glenoid can be used to accurately quantify, by means of a glenoid index, bone loss in patients with anterior glenohumeral instability, and to compare the results with arthroscopic measurements to determine if the 3-D CT scan can preoperatively predict which patients with anterior glenohumeral instability will benefit from a bone graftingMethods: procedure. From 2003 to 2006, 188 patients with anterior glenohumeral instability underwent arthroscopic evaluation and treatment by the senior author (S.S.B.). Of 188 total patients, there were 25 patients ranging in age from 15 to 43 years (median, 19 years) who underwent 3-D CT evaluations of both shoulders followed by arthroscopy of the unstable shoulder. For an arthroscopically measured bone loss of less than 25% of the inferior glenoid diameter, an arthroscopic Bankart repair was performed; for a glenoid bone loss of greater than or equal to 25%, an open Latarjet reconstruction was performed. We defined the glenoid index as the ratio of the maximum inferior diameter of the injured glenoid compared to the maximum inferior diameter of the uninjured contralateral glenoid as calculated from the 3-D CT scans. If the glenoid index was greater than 0.75, the patient was predicted to benefit from an arthroscopic Bankart repair (the need for surgery and the type of surgery having been determined on the basis of arthroscopic measurements). However, if the glenoid index was less than or equal to 0.75, the patient was predicted to benefit from an open Latarjet procedure. The results of each patient’s glenoid index were compared with the arthroscopic decision to perform either an arthroscopic Bankart repair or an open Latarjet Results:procedure. Of the 25 patients included in this study, 13 patients underwent an open Latarjet procedure and 12 patients underwent an arthroscopic Bankart repair. The 3-D CT scans accurately predicted the arthroscopic decisions to perform an arthroscopic Bankart repair or open Latarjet in 24 (96%) of 25 casesP Ͻ(Fisher .001). Conclusions:exact test; The glenoid index as calculated from the 3-D CT scan accurately predicted the requirement of a bone grafting procedure for 24 (96%) of 25 patients when the benchmark value of 0.75 was used. The 3-D CT scan can therefore be used by surgeons as an additional diagnostic tool for preoperative planning and patient counseling.Level of Evidence:Level III, development of diagnostic criteria with universally applied reference (nonconsecutive patients).Key Words: Bankart repair—Bone graft— Computed tomography—Glenoid index—Instability—Shoulder arthroscopy—Shoulder instability. e believe that the treatment of anterior recurrencegleno- of instability or dislocation and ultimately Whumeral instability with significant bone toloss improve the clinical outcome.1-3 The difficulty lies requires a bone grafting procedure to preventin determiningthe the amount of bone loss that represents a clinically significant deficit. The arthroscopic appre- ciation of an inverted pear–shaped glenoid and quan- From The San Antonio Orthopaedic Group, San Antonio, Texas, U.S.A. tification of the amount of glenoid bone loss using the S.S.B. is a consultant for, and receives royalties from, Arthrex, glenoid bare spot as a reference point have been Inc. The other authors report no conflict of interest. previously described as accurate and well-accepted Address correspondence and reprint requests to Stephen S. Burkhart, M.D., 400 Concord Plaza Dr, Ste 300, San Antonio, TX guides to determine significant bone loss.3-5 78216, U.S.A. E-mail: [email protected] Burkhart and DeBeer3 showed that in patients with © 2008 by the Arthroscopy Association of North America 0749-8063/08/2404-7372$34.00/0 anterior glenohumeral instability and an inverted pear– doi:10.1016/j.arthro.2007.10.008 shaped glenoid, which represents a loss of greater than 376 Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 24, No 4 (April), 2008: pp 376-382 PREOPERATIVE 3-D CT TO QUANTIFY GLENOID BONE LOSS 377 25% of the maximum inferior glenoid width, the recur- Of the 188 patients, there were 25 nonconsecutive rence rate of shoulder instability or dislocation after an patients ranging in age from 15 to 43 years (median, arthroscopic Bankart repair was 67%. However, pa- 19 years) who underwent 3-D CT evaluations of both tients with less than 25% loss of the maximum inferior of their shoulders. The scans were done by a General glenoid width had a recurrence rate of only 4%. Electric Lightspeed Ultra (Milwaukee, WI) with Three-dimensional computed tomography (3-D CT) 8-slice scans converted to 3-D CT reconstructions by has been used for measuring bone loss in patients with combined standard and bone algorithms. Our criteria anterior glenohumeral instability.6 However, to our for obtaining a 3-D CT scan were: (1) at least 3 known knowledge, there have been no CT studies in the shoulder dislocations; (2) total duration of all disloca- literature that quantify anterior glenoid bone loss and tions more than 4 hours; or (3) suspicion of significant compare it to the measured bone loss with an arthro- bone loss on plain films or magnetic resonance imag- scopically validated method. ing. In this study, we used 3-D CT imaging of both the A linear method was used to calculate the amount of injured and intact shoulders to measure the amount of bone loss of the injured glenoid compared to the glenoid bone loss in the injured shoulder. We then uninjured contralateral glenoid with the 3-D CT scans. compared the CT measurements and CT-derived gle- The first step in the process was obtaining an en face noid index to the direct arthroscopic measurements view of the uninjured glenoid with the 3-D CT scan. and arthroscopically derived glenoid index of patients The most superior aspect of the glenoid is just poste- who underwent treatment for anterior glenohumeral rior to the base of the coracoid and is labeled as A1 instability. If the arthroscopically derived glenoid in- (Fig 1A). The most inferior aspect of the glenoid is the dex was greater then 0.75, then an arthroscopic Ban- farthest point from A1 and usually corresponds with kart repair was done. If the arthroscopically derived the lateral ridge of the scapular body. This point is glenoid index was less than 0.75, then an open Latarjet labeled B1. A line is then drawn connecting A1 to B1 procedure with coracoid bone grafting of the glenoid (labeled A1B1 in Fig 1A). The line A1B1 corresponds was performed. All surgical treatment decisions were to the normal glenoid height (H1) of the uninjured based on the arthroscopic findings and measurements. shoulder (A1B1 ϭ H1). We compared the CT-derived glenoid index of each The second line drawn (C1D1) is perpendicular to patient with the arthroscopically derived glenoid index. A1B1 and is adjusted up or down until it is at the We hypothesized that the 3-D CT–derived glenoid widest portion of the inferior glenoid (Fig 1B). The index would correlate well with the arthroscopically line C1D1 corresponds to the normal inferior glenoid derived glenoid index and that the CT measurements width (W1) of the uninjured shoulder (C1D1 ϭ W1). would therefore be useful in accurately predicting the The intersection of lines A1B1 and C1D1 is labeled preferred treatment (arthroscopic Bankart repair v O1 (Fig 1C), which represents the geometric center of open bone grafting). the inferior glenoid circle (Fig 1D). The line O1D1 corresponds to the radius of the normal inferior glenoid ϭ METHODS circle (R1) of the uninjured shoulder (O1D1 R1). An en face view of the injured glenoid is then The study was done with institutional approval as obtained. The length of the glenoid is measured in a per the research requirements of The Orthopaedic similar fashion as the uninjured glenoid and line A2B2 Institute in San Antonio, Texas. From 2003 to 2006, (H2) is drawn (Fig 2A). We made the assumption that 188 patients with anterior glenohumeral instability the injured glenoid with the Bankart lesion had mor- underwent arthroscopic evaluation and treatment by phology identical to the uninjured glenoid before the the senior author (S.S.B.). All of these patients had injury.6 The radius of the inferior glenoid circle was diagnostic arthroscopic evaluations of their shoulders then calculated through basic proportions. Because to determine the morphology of their glenoids. R1/H1 ϭ R2/H2, R2 could be calculated by the fol- Through a standard anterosuperolateral viewing portal lowing formula: R2 ϭ (R1/H1)*H2. Therefore, the the glenoids were viewed en face and classified as geometric center of the inferior glenoid circle (O2) either pear-shaped or inverted pear–shaped. Further- could be found a distance equal to R2 from the point more, the amount of glenoid bone loss was quantified B2 (Fig 2B). arthroscopically using a validated methodology based A new line is drawn that crosses point O2 and is on the glenoid bare spot, which is the center of the perpendicular to line A2B2. This line is labeled C2D2 inferior glenoid circle.4 (Fig 2C) and represents the inferior width of the 378 T-Y. CHUANG ET AL. FIGURE 1. Normal glenoid, left shoulder. (A) The long axis of the glenoid is defined by line A1B1. (B) The widest portion of the inferior glenoid is line C1D1, which is perpendicular to line A1B1. (C) The intersec- tion of lines A1B1 and C1D1 is labeled O1, and it represents the geometric center of the in- ferior glenoid circle (D).