Olecranon plates used for the internal fixation of complex olecranon fractures are applied directly over the triceps tendon on the posterior aspect of the olecranon. The aim of the study is to describe the relationship of the plates and screws to the triceps tendon at the level of the olecranon.

Eight cadaveric elbows were used. Dimensions of the triceps tendon at the insertion and 1cm proximal were measured. A long or a short olecranon plate was then applied over the olecranon and the most proximal screw applied. The length of the plate impinging on the tendon and the level of the screw tract on the tendon and bone were measured.

The mean olecranon height was 24.3cm (22.4-26.9cm) with a tip-to-tendon distance of 14.5cm (11.9-16.2cm). The triceps tendon footprint averaged 13.3cm (11.7-14.9cm) and 8.8cm (7.6-10.2cm) in width and length, respectively. The mean width of the central tendon 1 cm proximal to the footprint was 6.8 cm. The long olecranon plate overlay over more movable tendon length than did the short plate and consequently the superior screw pierced the triceps tendon more proximally with the long plate. Using the Mann-Whitney U test, the differences were significant.

The long olecranon plates encroach on more triceps tendon than short plates. This may be an important consideration for olecranon fractures with regards implant loosening or triceps tendon injury.

Local infiltration analgesia is a relatively novel technique developed for effective pain control following total knee replacement, reducing requirements of epidural or parenteral post-operative analgesia. The study aimed to investigate the anatomical spread of Local Infiltration Analgesia (LIA) used intra-operatively in total knee arthroplasty (TKA) and identify the nerve structures reached by the injected fluid.

Six fresh-frozen cadaveric lower limbs were injected with 180ml of a solution of latex and India ink to enable visualisation. Injections were done according to our standardised LIA technique. Wounds were closed and limbs were placed flat in a freezer at −20°C for two weeks. Limbs were then either sliced or dissected to identify solution locations.

Injected solution was found from the proximal thigh to the middle of the lower leg. The main areas of concentration were the popliteal fossa, the anterior aspect of the femur and the subcutaneous tissue of the anterior aspect of the knee. There was less solution in the lower popliteal fossa. The solution was found to reach the majority of the terminal branches of the tibial, fibular and obturator nerves.

Overall, there was good infiltration of nerves supplying the knee. The lack of infiltration into the lower popliteal fossa suggests more fluid or a different injection point could be used. The solution that travelled distally to the extensor muscles of the lower leg probably has no beneficial analgesic effect for a TKA patient. This LIA technique reached most nerves that innervate the knee joint which supports the positive clinical results from this LIA technique. However, there may be scope to optimise the injection sites.

The purpose of the study was to describe the normal anatomy of glenoid labrum. Twenty dry bone scapulas and 19 cadaveric shoulders were examined. Light microscopy was performed on 12 radial slices through the glenoid.

An external capsular circumferential ridge, 7–8mm medial to the glenoid rim marks the attachment of the capsule. A separate internal labral circumferential ridge 4mm central to the glenoid rim marks the interface for the labrum and articular cartilage. A superior-posterior articular facet contains the superior labrum. Two thirds of the long head of biceps arise from the supraglenoid tubercle, the remainder from the labrum. The superior labrum is concave and is loosely attached to the articular cartilage and glenoid rim. In contrast the anterior-inferior labrum is convex, attaches 4mm central to the glenoid rim and has a strong attachment to articular cartilage and bone.

The anatomy of the superior and anteroinferior labrum are fundamentally different. Suture anchor repair of the superior labrum should be 7mm medial to the glenoid rim whereas the anterior-inferior labrum should be repaired to the face of the glenoid. By defining the normal anatomy of the superior labrum, pathological tears can be identified.

Purpose: Despite a number of recently published reports on the success of meniscal repair devices, there are no anatomic studies documenting their safety. The purpose of this cadaveric and radiographic study was to anatomically determine the proximity of a common commercially available meniscal repair device to the popliteal neurovascular structures

Methods: Five human cadaveric knees were obtained and procured from the medical school anatomy lab. Two Biostingers (Linvatec) measuring 16mm in length were placed in the posterior one third of the medial meniscus. Each specimen was then placed prone with the knee extended to expose the posterior aspect of the knee. The distance to the neurovascular bundle for each device was then measured with a ruler calibrated to the nearest 0.1cm. To validate our anatomic dissection results, fifty calibrated human knee MRI scans were reviewed by two independent radiologists. The distances measured were from the popliteal artery to the closest point at the lateral meniscus periphery/capsule and the medial meniscus periphery/capsule. The average distance as measured by the two radiologists was calculated as was the average for the entire population of fifty subjects

Results: The mean distance in the cadaveric study was 15.6mm (14.0–18.0mm) between the tip of the repair device needle and the neurovascular bundle. The mean distance on MRI from the popliteal neurovascular bundle to the closest point in the posterior medial meniscus was 20.0 mm (13.0 mm–28.7 mm). The mean distance from the popliteal structures to the posterior lateral meniscus was 9.4 mm (3.2 mm–16.5 mm).

Conclusions: Considering the potential for significant morbidity, we recommend medial meniscal repair should be performed carefully with repair devices. Specifically, one should limit posterior capsule penetrations to less than 15 mm based on these findings.

Background:

In anterior approaches for total hip arthroplasty (THA), the femoral part of the procedure requires the release of the capsule from the greater trochanter. However, it is unknown whether any other tendons of the short external rotator muscles are also damaged during capsular release procedures. The aim of this cadaveric study was to identify the bony landmarks on the greater trochanter, which indicate the individual short external rotator muscle insertions.

The purpose of the study was to describe the normal anatomy of glenoid labrum.

20 dry bone scapulas and 19 cadaveric shoulders were examined. Light microscopy was performed on 12 radial slices through the glenoid.

An external capsular circumferential ridge, 7–8mm medial to the glenoid rim marks the attachment of the capsule. A separate internal labral circumferential ridge 4mm central to the glenoid rim marks the interface for the labrum and articular cartilage. A superior-posterior articular facet contains the superior labrum. Two thirds of the long head of biceps arise from the supraglenoid tubercle, the remainder from the labrum.

The superior labrum is concave and is loosely attached to the articular cartilage and glenoid rim. In contrast the anterior-inferior labrum is convex, attaches 4mm central to the glenoid rim and has a strong attachment to articular cartilage and bone.

The anatomy of the superior and anterior-inferior labrum are fundamentally different.

Shoulder girdles of 20 cadavers (68–94yrs) were harvested. The anterior (ACHA) and posterior circumflex humeral arteries (PCHA) were injected with ink and the extra and intraosseous courses of the dyed vasculature dissected through the soft tissues and bone to the osteotendinous junctions of the rotator cuff. The ink injection and bone dissection method was newly developed for the study.

Rates of cross-over at the osteotendinous juntion were 75% in the supraspinatus, 67% in subscapularis, 33% in infraspinatus and 20% in teres minor. The supraspinatus and subscapularis insertions were vascularised by the arcuate artery, a branch of the ACHA. The insertions of the infraspinatus and teres minor were supplied by an unnamed terminal branch of the PCHA.

The insertions of the rotator cuff receive an arterial supply across their OTJ's in 50% of cases. This may explain observed rates of AVN in comminuted proximal humeral fractures. The terminal branch of the PCHA supplying the infraspinatus and teres minor insertions was named the “Posterolateral Artery”. Finally, the new method employed for this study which allowed for direct visualisation of intraosseous vasculature, will enhance our understanding of skeletal vascular anatomy and have clinical applications in orthopaedic and reconstructive surgery.

Purpose of the study: The posterior paraspinal approach to the lumbar spine was initially described and promoted by Wiltse for posterolateral arthrodesis of the lumbosacral junction in patients with spondylolisthesis. Despite technical improvements proposed by Wiltse, the muscular cleavage is still poorly localized in the sacrospinalis muscle. The purpose of this work was to provide a more accurate anatomic description of this spinal approach and to describe anatomic landmarks to facilitate execution of the procedure.

Material and methods: Fifty anatomic specimens were dissected (27 male and 23 female cadavers); 33 had been embalmed. The anatomy study used a bilateral approach to the spine. The exact anatomic localization of the muscle cleavage was noted. Measures were taken in relation to the mid line of the L4 spinatus process.

Results: In all specimens, the muscle cleavage lay between the multifidus and longissimus heads of the sacrospinalis muscle. A fibrous partition was noted in 88 of the 100 specimens. The mean distance from the mid line to the cleavage line was 4.04 cm (range 2.4–7.0 cm). The surface of the sacrospinalis muscle presented fine perforating arteries and veins in all specimens, directly in line with the cleavage plane. In 12 cases, a major posterior sensorial branch of the L3 nerve running to the skin was identified in the cranial portion of the approach.

Discussion: The muscle cleavage plane appears to be easy to localize for the paraspinal approach to the lumbosacral junction. Opening the aponeurosis of the latissimus dorsi near the mid line enables visualization of the perforating vessels in line with the anatomic cleavage plane of the sacrospinalis muscle. In our experience, this plane is situated on average 4 cm from the mid line. Hemostasis of these vessels is acceptable since the sacrospinalis muscle has a rich supply of anastomosed vessels. Care must be taken to avoid injury to the posterior sensorial branch of the L3 nerve which runs along the plane of the muscle cleavage.

Conclusion: In our opinion, this minimally hemorrhagic approach is perfectly adapted to non-instrumented fusion of the lumbosacral junction, particularly for spondylolisthesis in children and adults. Precise knowledge of the anatomy of this approach is a necessary prerequisite for successful execution.

Purpose of the study: The anterior rim of the arthroplasty cup can overhang the bone in certain situations: oversized cup, insufficient anteversion, insufficient ace-tabular reaming, cylindrospherical reaming overriding the acetabular opening. The straight or concave shape of the anterior wall of the acetabulum also affects prosthetic overhang. The purpose of this anatomic study was to use a navigation system to quantify, in vitro, the height of the iliopubic psoas notch.

Material and methods: Sixty-eight acetabuli from 34 cadaver pelvi free of osteoarthritis (13 male and 21 female) were analyzed using the Stryker™ hip navigation system. Morphological data were collected for mathematical processing which defined the diameter and the center of the articular surface.

Results: Considerable intra- and inter-individual differences in the shape of the acetabular rim were noted and quantified, particularly concerning the psoas notch. When the lateral view of the acetabular rim is projected onto a plane it produces a succession of three summits and three valleys explaining the difficulty encountered in obtaining a precise mean plane for the acetabular opening.

Discussion: More or less pronounced protrusion of the cup could explain potentially painful anterior impingement of the psoas, especially for certain types of acetabular morphology.

Purpose

Current coronoid fracture classification systems are based on fragment size and configuration using plain radiographs and/or CT. During surgery, coronoid fracture fragments appear much larger than anticipated because cartilage is radiolucent and therefore not taken into account with preoperative imaging. The purpose of this study was to quantify the articular cartilage thickness of the coronoid process, with reference to coronoid fracture classifications.

Purpose of the study: The communicating branch of the lateral plantar nerve is an anastomotic branch between the medial and lateral plantar nerves. Morton’s pseudoneuroma is usually described as resulting from the combination of one of the divisions of the medial plantar nerve and the communicating branch of the lateral plantar nerve. Surgical treatment of Morton’s pseudoneuroma can fail, leading to recurrent neuroma, or digital hypoesthesia. We performed an anatomic dissection study to search for the anatomic basis for an improved surgical technique.

Material and methods: The study included 35 feet of embalmed cadavers. A standard protocol was used for dissection. We studied the communicating branch, its frequency and size, and measured its attachment on the 3rd and 4th plantar common digital nerves. All other nerve ramifications were noted.

Results: None of the 35 dissected feet presented a macroscopic Morton pseudoneuroma. The communicating branch was present in 77% of the specimens with frequent anatomic variations concerning the size, ramifications, orientation and distance from the intermetatarsal ligament. A fine plantar cutaneous branch was often found under the intermetatarsal ligament. The bifurcation of the 2^nd and 3^rd nerves was occasions not far from the junction of the communicating branch on the 3^rd nerve, raising the risk of injury to the 2^nd nerve during surgical excision of a Morton pseudoneuroma.

Discussion: The communicating branch appears to play a role in recurrence after excision of the Morton pseudoneuroma by preventing the retraction of the 3^rd digital nerve in the muscle zone if it is not sectioned. However, wide resection of the proximal part of the 3^rd nerve and the communicating branch could lead to digital hypoesthesis if the 2^nd digital nerve is injured. The proximity of the bifurcations of the 2^nd and 3^rd nerves and of the 3^rd nerve with the communicating branch could explain this type of complication.

Conclusion: We advocate resection of the pseudoneuroma under visual control in order to carefully resect the communicating branch without injuring the adjacent nerve branches.

Purpose: Isolated paralysis of the great toe long extensor is a rare complication of leg fractures. In certain patients, an erroneous diagnosis of compartment syndrome or muscle incarceration may be made.

Material and methods: We dissected ten fresh cadavers.

Results: The great toe long extensor was innervated by a branch of the deep fibular nerve which arose 15 cm from the talocrural joint space and directly in contact with the periosteum of the tibial shaft. In this localisation, the branch can be directly sectioned during trauma, reduction, or reaming.

Discussion: We had one patient aged 30 years with a fracture of the mid third of the leg who was treated by centromedullar nailing. Postoperatively, this patient developed isolated paralysis of the great toe long extensor. The isolated neurogenic origin of this paralysis was confirmed by electromyography.

Conclusion: This case is illustrative of direct injury of the great toe long extensor innervation, a cause which to our knowledge has not been previously described.

Purpose: The purpose of this work was to study arterial blood supply to the coxal bone in order to minimise the risk of postsurgical necrosis during acetabular osteosynthesis. Necrosis of the coxal bone is a rare but well-known complication of acetabular fracture surgery.

Material and methods: Ten fresh cadavers were dissected after intra-arterial injection of coloured resin. All collaterals feeding the bone were described and counted. An arterial map was drawn.

Results: The acetabulum is supplied by four main arterial sources: 1) the ischion artery, a collateral of the pudendal artery, which supplies the posterior and lateral part of the acetabulum; 2) the artery of the roof of the acetabulum, a collateral of the superior gluteal artery, supplies the upper and lateral part of the acetabulum: 3) branches of the anterior and posterior division of the obturator artery which supply the upper part and the rim of the obturated foramen and the anteroinferior and posteroinferior parts of the acetabulum; 4) branches issuing from the obturator artery supply the quadrilateral surface.

Discussion: The Kocher approach can easily injure the ischion artery. The wide lateral approach described by Letournel and the triradiate approach described by Mears can injure the ischion artery and the artery of the roof of the acetabulum. Theoretically, the risk of bony necrosis would be greater if an endopelvic approach is associated due to the risk of injury to the endopelvic arteries issuing from the obturator artery. The anterior approach to the acetabulum appears to carry the less risk of ischemia, theoretically, than the other approaches to the acetabulum.

Purpose: Due to its proximity to the glenohumeral joint, the suprascapular nerve may be at risk of iatrogenic nerve injury during arthroscopic labral repair. Our primary objective is to evaluate the risk of suprascapular nerve injury during standard drilling techniques utilized in arthroscopic superior labral repairs. Secondarily, we evaluated the correlation between this risk and scapular size.

Method: Forty-two cadaveric shoulders were dissected to isolate their scapulae. A surgical drill and guide was used to create suture anchor holes in 3 locations in the superior rim of the glenoids as typically done in arthroscopic superior labral repairs. The orientation of these drill holes correspond to common shoulder arthroscopic portals. The suprascapular nerve was then dissected from the suprascapular notch to the spinoglenoid notch. The presence of drill perforations through the medial cortex of the glenoid vault was recorded along with the corresponding hole depth and distance to the suprascapular nerve.

Results: Medial glenoid vault perforations occurred in 8/21(38%) cadavers with a total of 18/126(14%) perforations. The suprascapular nerve was in line of the drill path in 5/18(28%) perforations. Female specimens and smaller scapulae had a statistically higher risk of having a perforation (p< 0.05).

Conclusion: The results of this anatomic study suggest that there is a substantial risk of medial glenoid vault perforation. When a perforation does occur, the suprascapular nerve appears to be at high risk for injury especially with more posterior drill holes. The risk is significantly higher in females and in smaller scapulae.

Purpose of the study: The piriform syndrome is treated medically: functional rehabilitation and injections. If the medical treatment fails, tenotomy of the piriform muscle can be proposed. Published studies report good outcome in 66 to 87% Of patients. The purpose of this study was to examine the extrapelvic innervations of this muscle in order to assess the feasibility of neurotomy of he piriform muscle.

Material and method: Twenty gluteal regions were dissected. We studied first the relations between the piriform muscle and the ischiatic nerve. Then the innervations branches of the piriform muscle were localized in three landmarks.

Results: We found the of the six types of relation between the ischiatic nerve and the piriform muscle described by Beaton, with frequencies comparable to reports in the literature. Innervation of the piriform muscle does not follow a standard pattern, even though the innervations generally comes from the ischiatic nerve; the nerve branches come from the superior and inferior gluteal pedicles. In addition, these nerve branches penetrate the deep aspect of the muscle in random fashion. In addition, accessibility to the deep aspect of the piriform muscle cannot be achieved easily but requires prior section of its insertion on the greater trochanter.

Discussion: In our opinion, these results suggest that isolated neurotomy of the piriform muscle is not clinically feasible; it might be possible to improve function results of isolated tenotomy by performing a neurotomy of the nerve branches visible during the tenotomy procedure. A greater benefit might be expected in forms where the ischiatic nerve crosses the piriform muscle.

Purpose: Cage fusion of the L5–S1 segment is a controversial issue due to the weak stabilisation of the spine during extension and axial rotation. Complementary fixation appears to be needed to improve stability, but the presence of the bifurcation of the great vessels is an anatomic limitation. We studied the anatomy of this area to examine the feasibility of anterior plate fixation.

Material and methods: According to the recognised anatomic references (Rouvière, Bouchet and Cuilleret, Louis), a 33-mm safety zone was described at the aortoiliac and iliocaval bifurcation. This space free from contact with the greater vessels lies in front of the L5–S1 disc. Pre-operative angio-MRI was used to assess the size of this safety zone. A triangular anterior plate was designed for arthrodesis (Pyramid Sofamor Danek, USA). An L5–S1 arthrodesis was performed in 15 consecutive patients using this plate. Follow-up was one year. A video-assisted anterior retroperitoneal approach was used in all cases. The Prolo and Oswestry scores were used for the preoperative and last follow-up evaluations.

Results: According to the anatomic study and the MRI views, 89% of the patients had a sufficient safety zone for plate fixation. The method was contraindicated in two patients who had a low bifurcation. Angio-MRI was found to be simple and reproducible. The MRI analysis was confirmed at surgery (no false negatives). The plate was successfully implanted in 15 patients with no contact with the great vessels. There were no serious complications (vascular, neurological, urological, digestive). The economic Prolo score was improved from 2.7 to 4.2 and the functional score from 2.6 to 4.3. The Oswestry score improved 33%. The rate of clinical success was 93% (14/15 patients).

Discussion and conclusion: Implantation of an anterior plate for L5-S1 fusion is feasible. The key points are: 1) rigorous preoperative evaluation of the greater vessel bifurcation; 2) anatomic plate design; 3) appropriate surgical technique. The risk of retrograde ejaculation is related to retraction of the hypogastric plexus and should be investigated with a prospective study. This osteosynthesis technique can avoid secondary operations for fixation with pedicular or transarticular screwing.

Background: The positioning of the acetabular component is of critical importance in total hip arthroplasty. Due to the orientation of the acetabulum and limitations of observation imposed at the operative site mal-positioning is common. We believe that by utilising the transverse acetabular ligament (TAL) and acetabular labrum, we are able to anatomically position our cup. In this study, we evaluate the correlation between placement of the acetabular component by reference to the TAL and the acetabular labrum with the taught safe zones for cup placement.

Method: 7 embalmed hips were studied. Following disarticulation the labrum and TAL were digitised and their plane was calculated. Orientation of cup placement in this plane was calculated from a pre-dissection pelvic CT.

Results: The plane of the labrum/TAL varied between 5–26° of anteversion and 32–59° of inclination. Interob-server differences in acetabular cup placement based on the TAL/labral plane indicate reasonable coherence. Almost all components were inside the documented “safe zone” 0–40° of anteversion and 30–55° of inclination of placement.

Conclusion: The acetabular labrum and TAL form a plane that reflects the documented “safe zones” for acetabular component placement. We feel that this plane allows a surgeon to determine optimal patient specific acetabular component placement, irrespective of patient position.

Introduction: Reports on nerve injury after arthroscopic ACL reconstruction using hamstring tendon autograft had mainly focused on injury to the infrapatellar branch of the saphenous nerve (IPBSN), with few reports on injury of the sartorial branch of the saphenous nerve (SBSN).

Aim of the work: was to define the level of anatomical termination of the saphenous nerve in relation to knee joint level and the relation of its sartorial branch to the surrounding tendons so that it could be avoided during hamstring tendon harvesting.

Materials and Methods: This anatomical study included cadaveric dissection of the medial aspect of the knee joint of 25 preserved knees. The saphenous nerve was dissected proximal to the knee joint and followed distally till it was divided into its two terminal branches.

Results: In 68 %, the saphenous nerve gave its two terminal branches at a mean distance of 8 cm above the knee joint line. In 32 %, the level of termination of the saphenous nerve was below the knee joint line by a mean distance of 3 cm.

In 92 % the saphenous nerve or the SBSN was passing posterior to the sartorius tendon by a mean distance of 19.8 mm. In 68 % the saphenous nerve or the SBSN continued distally anterior to the gracilis tendon, while in 16 % the SBSN continued distally posterior to the gracilis tendon. In 20 % the distance between the saphenous nerve or the SBSN and gracilis tendon was 5 mm or less. In 12 % the saphenous nerve or the SBSN was lying directly anterior to the gracilis and in 4 %, the SBSN was lying directly behind the gracilis tendon at the knee joint line.

In all the knees the saphenous nerve or the SBSN was passing distally anterior to the semitendinosus tendon at a mean distance of 23.1 mm.

Conclusion: The saphenous nerve or its terminal branch the SBSN, is at a close anatomical relation with the gracilis tendon. This might predispose the nerve to be damaged during passage of the tendon stripper over the tendon.

Clinical correlation: The saphenous nerve or its terminal branch the SBSN, are at a risk of injury during arthroscopic ACL reconstruction using hamstring tendon autograft. The nerve Injury of the saphenous nerve or its terminal branch (SBSN) might be an intrinsic problem associated with the technique itself.

Ventral screw osteosynthesis is a common surgical method for treating fractures of the odontoid peg, but there is still no consensus about the number and diameter of the screws to be used. The purpose of this study was to develop a more accurate measurement technique for the morphometry of the odontoid peg (dens axis) and to provide a recommendation for ventral screw osteosynthesis.

Images of the cervical spine of 44 Caucasian patients, taken with a 64-line CT scanner, were evaluated using the measuring software MIMICS. All measurements were performed by two independent observers. Intraclass correlation coefficients were used to measure inter-rater variability.

The mean length of the odontoid peg was 39.76 mm (sd 2.68). The mean screw entry angle α was 59.45° (sd 3.45). The mean angle between the screw and the ventral border of C2 was 13.18° (sd 2.70), the maximum possible mean converging angle of two screws was 20.35° (sd 3.24). The measurements were obtained at the level of 66% of the total odontoid peg length and showed mean values of 8.36 mm (sd 0.84) for the inner diameter in the sagittal plane and 7.35 mm (sd 0.97) in the coronal plane. The mean outer diameter of the odontoid peg was 12.88 mm (sd 0.91) in the sagittal plane and 11.77 mm (sd 1.09) in the coronal plane. The results measured at the level of 90% of the total odontoid peg length were a mean of 6.12 mm (sd 1.14) for the sagittal inner diameter and 5.50 mm (sd 1.05) for the coronal inner diameter. The mean outer diameter of the odontoid peg was 11.10 mm (sd 1.0) in the sagittal plane and 10.00 mm (sd 1.07) in the coronal plane. In order to calculate the necessary screw length using 3.5 mm cannulated screws, 1.5 mm should be added to the measured odontoid peg length when anatomical reduction seems possible.

The cross-section of the odontoid peg is not circular but slightly elliptical, with a 10% greater diameter in the sagittal plane. In the majority of cases (70.5%) the odontoid peg offers enough room for two 3.5 mm cannulated cortical screws.

Cite this article: Bone Joint J 2013;95-B:536–42.

Anatomic studies have demonstrated that bipolar glenoid and humeral bone loss have a cumulative impact on shoulder instability, and that these defects may engage in functional positions depending on their size, location, and orientation, potentially resulting in failure of stabilisation procedures. Determining which lesions pose a risk for engagement remains a challenge, with Itoi's 3DCT based glenoid track method and arthroscopic assessment being the accepted approaches at this time. The purpose of this study was to investigate the interaction of humeral and glenoid bone defects on shoulder engagement in a cadaveric model. Two alternative approaches to predicting engagement were evaluated; 1) CT scanning the shoulder in abduction and external rotation 2) measurement of Bankart lesion width and a novel parameter, the intact anterior articular angle (IAAA), on conventional 2D multi-plane reformats. Hill-Sachs and Bony Bankart defects of varying size were created in 12 cadaveric upper limbs, producing 45 bipolar defect combinations. The shoulders were assessed for engagement using cone beam CT in various positions of function, from 30 to 90 degrees of both abduction and external rotation. The humeral and glenoid defects were characterised by measurement of their size, location, and orientation. The abduction external rotation scan and 2D IAAA approaches were compared to the glenoid track method for predicting engagement. Engagement was predicted by Itoi's glenoid track method in 24 of 45 specimens (53%). The abduction external rotation CT scan performed at 60 degrees of glenohumeral abduction (corresponding to 90 degrees of abduction relative to the trunk) and 90 degrees of external rotation predicted engagement accurately in 43 of 45 specimens (96%), with sensitivity and specificity of 92% and 100% respectively. A logistic model based on Bankart width and IAAA provided a prediction accuracy of 89% with sensitivity and specificity of 91% and 87%. Inter-rater agreement was excellent (Kappa = 1) for classification of engagement on the abduction external rotation CT, and good (intraclass correlation = 0.73) for measurement of IAAA. Bipolar lesions at risk for engagement can be identified using an abduction external rotation CT scan at 60 degrees of glenohumeral abduction and 90 degrees of external rotation, or by performing 2D measurements of Bankart width and IAAA on conventional CT multi-plane reformats. This information will be useful for peri-operative decision making around surgical techniques for shoulder stabilisation in the setting of bipolar bone defects