Purpose: The identification of anatomical landmarks is an important aspect of joint surgery, to ensure proper placement and alignment for implants and other reconstructive procedures. At the elbow, the center of the capitellum (derived via a digitization of the surface and subsequent sphere fitting) has been well established as a key landmark to identify the axis of rotation of the joint. For some cases, and in particular minimally invasive surgery, only small regions of the capitellum may be exposed which may lead to errors in determining the centre. The purpose of this study was to identify the optimal location of digitizations of the capitellum.

Method: Twenty-five fresh frozen cadaveric distal humeri (19 left, 6 right) were studied. Using an x-ray computed tomography scanner, volumetric images of each specimen were acquired and used to reconstruct a 3-dimensional digital model of the specimen using the Visualization Toolkit (VTK). A sphere-fit algorithm was used to determine the centre of the spherical capitellum based on manually chosen (digitized) points across the 3D capitellar surface. The true geometric centre was located by digitizing points across the entire capitellar surface. Three sub-regions of the capitellum, commensurate with typical surgical approaches with minimal dissection, were then digitized. These were superior anterior lateral (SAL), inferior anterior lateral (IAL) and a combination of these two regions. These regions were compared to the true center using a 1-way Repeated Measures ANOVA with significance set to p = 0.05.

Results: Digitizations of only SAL and IAL sub-regions resulted in the largest differences relative to the true centre: SAL = 3.9±3.4 mm, IAL = 4.2±3.4 mm, (p < 0.0005). There was no difference between SAL and IAL (p = 1.0). Digitization of the combined SAL + IAL regions, while significantly different from the entire capitellum, resulted in the smallest mean difference of 0.87±0.84 mm.

Conclusion: These data show that the region of digitization affects the accuracy of predicting the capitellum centre. In a previous study by our group, we showed that an accurate determination of the centre of a sphere can be achieved with a small surface area of digitization. In the current study, the large errors that occurred when a small surface was digitized (i.e. SAL and IAL alone), are in all likelihood, due the non-spherical nature of the capitellum. In summary, while the most precise method in locating the true centre is to digitize the entire capitellar surface where possible, an alternative approach is to digitize both the superior and inferior anterior lateral regions.

Purpose: Current techniques for the investigation of elbow stability following injury or surgical interventions rely on kinematic descriptors. Typically, the motion pathways of the bones are employed to describe the effect of various clinical variables on alignment joint stability. This study describes a new approach to better visualize joint motion pathways that relates the anatomical geometry of the joint, obtained using medical imaging, with the recorded motion of the joint. The clinical aim of our study was to use this approach to investigate the effect of radial head resection and subsequent radial head arthroplasty on joint kinematics and elbow stability.

Method: Five fresh-frozen cadaveric specimens were employed. Computed tomography (CT) scans of each upper extremity were obtained to create a three-dimensional model of the joint. Simulated active elbow flexion with the arm in the valgus gravity loaded position was achieved using an upper arm simulator previously developed in our laboratory. Receivers from an electromagnetic tracking device were attached to the humerus and ulna in order to record their relative motion. Sutures were secured to the tendons of relevant muscles, which were connected to servomotors and pneumatic actuators, used to simulate motion. Kinematic data was collected with the radial head intact, radial head resected and following placement of metallic radial head implant. A repeated-measures analysis of variance was used to detect statistical differences. After testing, each specimen was denuded of all soft tissue and disarticulated. Fiducial markers were attached to the humerus and the ulna. The joint was then re-imaged in the CT scanner to obtain a volumetric image of each fiducial. Using the kinematic data recorded during simulated motion, and the knowledge of the position of each fiducial, a direct visualization of the recorded motion, using the 3D models was obtained. The bony position was then compared to the traditional graphical kinematic analysis examining changes in valgus angulations throughout the arc of motion.

Results: We observed a close agreement between the kinematic output and the registered bony 3D models showing the joint position. Following resection of the radial head, in the valgus dependent position, there was an increase in the valgus angulation of the ulna with respect to the humerus (p< 0.05).

Conclusion: Using this visualization approach, these changes in bony alignment were readily observed and understood visually in the 3D model of the ulna. Unlike the traditional graphical approach used to investigate elbow stability, this technique allows for the representation of coupled motion (rotation) of the bones. This technique also permits direct visualization the relative position of the bones within the joint, hence improving the overall understanding of joint motion.

Purpose: The purpose of this study was to determine the effect of serial olecranon resections on elbow stability.

Method: Eight fresh, previously frozen cadaveric arms underwent CT scanning. The specimens were mounted in an in-vitro motion simulator, and kinematic data was obtained using an electromagnetic tracking system. Simulated active and passive flexion was produced with servo-motors and pneumatic pistons attached to specific muscles. Flexion was studied in the dependent, horizontal, varus, and valgus positions. Custom computer navigation software was utilized to guide serial resection of the olecranon in 12.5% increments. A triceps advancement repair was performed following each resection.

Results: Serial olecranon resections resulted in a significant increase in valgus-varus (V-V) laxity for both passive (p< 0.001) and active (p=0.04) flexion. For passive motion this increase reached statistical significance following the 12.5% resection. This corresponded to an increase in V-V laxity of 1.4 ± 0.1o and a total laxity of 7.5 ± 1.0o. For active flexion this increase reached significance following the 62.5% resection. This corresponded to an increase in V-V laxity of 5.6 ± 1.1o and a total laxity of 11.2 ± 1.5. There was no significant effect of sequential olecranon excision on elbow kinematics or stability with the elbow in the vertical or horizontal positions. The elbows became grossly unstable after resection of greater than 75% of the olecranon.

Conclusion: A progressive increase in the varus-valgus laxity of the elbow was seen with sequential excision of the olecranon. Laxity of the elbow was increased with excision of 75% of the olecranon, likely due to the loss of the bony congruity and attachment site of the posterior band of the medial collateral ligament. Gross instability resulted when 87.5% or greater was removed, likely due to damage to the anterior band of the medial collateral ligament as it inserts on the sublime tubercle of the ulna. Rehabilitation of the elbow with the arm in the dependant position should be considered following excision of the olecranon; varus and valgus orientations should be avoided. The contribution of the olecranon to elbow stability may be even more important in patients with associated ligament injuries or fractures of the elbow.

Purpose: The role of the posterior bundle of the medial collateral ligament (PMCL) in stability of the elbow remains poorly defined. The purpose of this study was to determine the effect of sectioning the PMCL on the stability of the elbow.

Method: Varus and valgus gravity-loaded passive elbow motion and simulated active vertical elbow motion were performed on 11 cadaveric arms. An in-vitro elbow motion simulator, utilizing computer-controlled pneumatic actuators and servo-motors sutured to tendons, was used to simulate active elbow flexion. Varus/valgus angle and internal/external rotation of the ulna with respect to the humerus were recorded using an electromagnetic tracking system. Testing was performed on the intact elbow and following sectioning of the PMCL.

Results: With active flexion in the vertical position the varus/valgus kinematics were unchanged after PMCL sectioning (p=0.08). However, with the forearm in pronation, there was a significant increase in internal rotation after PMCL sectioning compared to the intact elbow (p< 0.05) which was most evident at 0° and 120° degrees of flexion (p< 0.05). This rotational difference was not statistically significant with the forearm in supination (p=0.07). During supinated passive flexion in the varus position, PMCL sectioning resulted in increased varus angulation at all flexion angles (p< 0.05). In pronation varus angulation was only increased at 120° of flexion (p< 0.05). However, internal rotation was increased at flexion angles of 30° to 120° (p< 0.05). In supination, sectioning the PMCL had no significant effect on maximum varus-valgus laxity or maximum internal rotation (p=0.1). However, in pronation, the maximum varus-valgus laxity increased by 3.5° (30%) and maximum internal rotation increased by 1.0° (29%) (p< 0.05).

Conclusion: These results indicate that isolated sectioning of the PMCL causes a small increase in varus angulation and internal rotation during both passive varus and active vertical flexion. This study suggests that isolated sectioning of the PMCL may not be completely benign and may contribute to varus and rotation instability of the elbow. In patients with insufficiency of the PMCL appropriate rehabilitation protocols (avoiding forearm pronation and shoulder abduction) should be followed when other injuries permit.

Purpose: This study evaluated the accuracy of humeral component alignment in total elbow arthroplasty. An image-based navigated approach was compared against a conventional non-navigated technique. We hypothesized that an image-based navigation system would improve humeral component positioning, with navigational errors less than or approaching 2.0mm and 2.0°.

Method: Eleven cadaveric distal humeri were imaged using a CT scanner, from which 3D surface models were reconstructed. Non-navigated humeral component implantation was based on a visual estimation of the flexion-extension (FE) axis on the medial and lateral aspects of the distal humerus, followed by standard instrumentation and positioning of a commercial prosthesis by an experienced surgeon. Positioning was based on the estimated FE axis and surgeon judgment. The stem length was reduced by 75% to evaluate the navigation system independent of implant design constraints. For navigated alignment, the implant was aligned with the FE axis of the CT surface model, which was registered to landmarks of the physical humerus using the iterative closest point algorithm. Navigated implant positioning was based on aligning a 3D computer model calibrated to the implant with a 3D model registered to the distal humerus. Each alignment technique was repeated for a bone loss scenario where distal landmarks were not available for FE axis identification.

Results: Implant alignment error was significantly lower using navigation (P< 0.001). Navigated implant alignment error was 1.2±0.3 mm in translation and 1.3±0.3° in rotation for the intact scenario, and 1.1±0.5 mm and 2.0±1.3° for the bone loss scenario. Non-navigated alignment error was 3.1±1.3 mm and 5.0±3.8° for the intact scenario, and 3.0±1.6 mm and 12.2±3.3° for the bone loss scenario. Without navigation, 5 implants were aligned outside 5° for intact bone while 9 were aligned outside 10° for the bone loss scenario.

Conclusion: Image-based navigation improved the accuracy of humeral component placement to less than 2.0 mm and 2.0°. Further, outliers in implant positioning were reduced using image-based navigation, particularly in the presence of bone loss. Implant malalignment may well increase the likelihood of early implant wear, instability and loosening. It is likely that improved implant positioning will lead to fewer implant related complications and greater prosthesis longevity.

Purpose: While computer-assisted techniques can improve the alignment of the implant articulation with the native structure, stem abutment in the intramedullary canal may impede achievement of this alignment. In the current study, the effect of a fixed valgus (6 degree) stemmed humeral component on the alignment of navigated total elbow arthroplasty was investigated. Our hypothesis was that implantation of a humeral component with a reduced stem length would be more accurate than implantation of the humeral component with a standard length stem.

Method: Thirteen cadaveric distal humeri were imaged using a CT scanner, and a 3D surface model was reconstructed from each scan. Implantation was performed using two implant configurations. The first set was unmodified (Regular) while the second set was modified by reducing the length of the humeral stem to 25% of the original stem (Reduced). A surface model of the humeral component was aligned with the flexion-extension (FE) axis of the CT-based surface model, which was registered to the landmarks of the physical humerus using the iterative closest point algorithm. Navigated implant positioning was based on aligning a 3D computer model calibrated to the implant with a 3D model registered to the distal humerus.

Results: Implant alignment error was significantly lower for the Reduced implant, averaging 1.3±0.5 mm in translation and 1.2±0.4° in rotation, compared with 1.9±1.1 mm and 3.6±2.1° for the Regular implant. Abutment of the implant stem with the medullary canal of the humerus prevented optimal alignment of the Regular humeral component as only four of the 13 implantations were aligned to within 2.0° using navigation.

Conclusion: These results demonstrate that a humeral component with a fixed valgus angulation cannot be accurately positioned in a consistent fashion within the medullary canal of the distal humerus without sacrificing alignment of the FE axis due to stem abutment. Improved accuracy of implant placement can be achieved by introducing a family of humeral components, with three valgus angulations of 0°, 4° and 8°. Based on humeral morphology for these specimens, 12 of the 13 implants may be positioned to within 2° of the native FE axis using one of these 3 valgus angulations.

Purpose: Coronal shear fractures of the humerus include the Kocher-Lorenz fracture, an osteochondral fracture of the capitellar articular surface, the Hahn-Steinthal fracture, a substantial shear fragment, extension into the trochlea, and complete involvement of the capitellum and trochlea. If the fracture proves irreparable, it is not known what the impact of fragment excision would have on the biomechanics of the elbow. The purpose of this study was to examine the effect of the sequential loss of the capitellum and trochlea on the kinematics and stability of the elbow.

Method: Eight fresh-frozen cadaveric arms were mounted in an upper extremity joint testing system, with cables attaching the tendons of the major muscles to motors and pneumatic actuators. Electromagnetic receivers attached to the radius and ulna enabled quantification of the kinematics of both bones with respect to the humerus. The distal humeral articular surface was sequentially excised to replicate clinically relevant coronal shear fractures while leaving the collateral ligaments intact. Active flexion in both the vertical and valgus-loaded positions, and passive rotation in the vertical position was conducted for each excision.

Results: Excision of the capitellum had no effect on ulnohumeral stability or kinematics in both the vertical or valgus positions (p=1.0). Excision of the entire capitellum and trochlea led to significant valgus instability with the arm in the valgus position (p=0.01), while excision of the lateral trochlea led to increased valgus instability with pronated flexion in the valgus position (p=0.049). Progressive loss of the articular surface led to posterior, inferior, and medial displacement of the radial head with respect to the capitellum and increased external rotation of the ulna with respect to the humerus in the vertical position (p< 0.05).

Conclusion: Excision of the capitellum did not result in valgus or rotational instability, while excision of the trochlea resulted in multiplanar instability. The radial head displaced medially because it is constrained to the ulna by the annular ligament, and the ulna pivoted into valgus and external rotation on the residual trochlea and medial collateral ligament. In patients with coronal shear fractures, the trochlea must be reconstructed to prevent instability and the potential for secondary degenerative change.

Purpose: The successful placement of elbow prostheses, external fixators and ligament reconstructions is dependent on the accurate identification of the elbow’s flexion-extension (FE) axis. In the case of periarticular bone loss, the FE axis must be visually estimated, as the necessary anatomical landmarks may not be available. Hence, referencing the uninjured elbow anatomy may prove beneficial in accurately defining this axis. However, this is contingent on the morphological features being similar between the two sides. Our objective was to compare distal humeral morphology between paired specimens. Our hypothesis was that anthropometric measurements from the distal humerus would be similar to the contralateral side.

Method: CT Images of 25 paired distal humeri were obtained. A right-to-left surface registration was then performed on each pair using the iterative closest point (icp) least-squares algorithm, thus placing each specimen in the same coordinate system.. Anthropometric characteristics measured (and compared between the left and right sides) included the angles of the FE and epicondylar axes in both the coronal and transverse planes, the anterior offset of the FE axis with respect to the humeral shaft axis, the length of the FE axis and the radius of curvature of the capitellum and trochlea.

Results: There was no statistically significant difference between the left and right humeri for the eight anthropometric characteristics studied (p > 0.05). The mean difference in magnitude for the FE axis angle was approximately 1.0° in both the coronal and transverse planes and the difference in magnitude for 80% of the paired specimens was less than 1.5°.

Conclusion: The anthropometric features of the distal humerus that are typically employed during elbow surgery are similar from side to side. Preoperative imaging of the contralateral normal elbow should be considered in patients with periarticular bone loss where referencing anatomical landmarks of the injured side is not possible. This information can be used as part of a preoperative plan to determine the ideal position of the implant, ligament reconstruction or external fixator during surgery. Contralateral imaging should be particularly useful when combined with computer-assisted elbow surgery.