Clinical decision support tools are software that match the input characteristics of an individual patient to an established knowledge base to create patient-specific assessments that support and better inform individualized healthcare decisions. Clinical decision support tools can facilitate better evidence-based care and offer the potential for improved treatment quality and selection, shared decision making, while also standardizing patient expectations. Predict+ is a novel, clinical decision support tool that leverages clinical data from the Exactech Equinoxe shoulder clinical outcomes database, which is composed of >11,000 shoulder arthroplasty patients using one specific implant type from more than 30 different clinical sites using standardized forms. Predict+ utilizes multiple coordinated and locked supervised machine learning algorithms to make patient-specific predictions of 7 outcome measures at multiple postoperative timepoints (from 3 months to 7 years after surgery) using as few as 19 preoperative inputs. Predict+ algorithms predictive accuracy for the 7 clinical outcome measures for each of aTSA and rTSA were quantified using the mean absolute error and the area under the receiver operating curve (AUROC).Introduction
Methods
Variability in placement of total shoulder arthroplasty (TSA) glenoid implants has led to the increased use of 3D CT preoperative planning software. Computer assisted surgery (CAS) offers the potential of improved accuracy in TSA while following a preoperative plan, as well as the flexibility for intraoperative adjustment during the procedure. This study compares the accuracy of implantation of reverse total shoulder arthroplasty (rTSA) glenoid implants using a CAS TSA system verses traditional non-navigated techniques in 30 cadaveric shoulders relative to a preoperative plan from 3D CT software. High resolution 1mm slice thickness CT scans were obtained on 30 cadaveric shoulders from 15 matched pair specimens. Each scan was segmented and the digital models were incorporated into a preoperative planning software. Five fellowship trained orthopedic shoulder specialists used this software to virtually place a rTSA glenoid implant as they deemed best fit in six cadavers each. The specimens were randomized with respect to side and split into a cohort utilizing the CAS system and a cohort utilizing conventional instrumentation, for a total of three shoulders per cohort per surgeon. A BaSO4 PEEK surrogate implant identical in geometry to the metal implant used in the preoperative plan was used in every specimen, to maintain high CT resolution while minimizing CT artifact. The surgeons were instructed to implant the rTSA implants as close to their preoperative plans as possible for both cohorts. In the CAS cohort, each surgeon used the system to register the native cadaveric bones to each respective CT, perform the TSA procedure, and implant the surrogate rTSA implant. The surgeons then performed the TSA procedure on the opposing side of the matched pair using conventional instrumentation. Postoperatively, CT scans were repeated on each specimen and segmented to extract the digital models. The pre- and postoperative scapulae models were aligned using a best fit match algorithm, and variance between the virtual planned position of the implant and the executed surgical position of the implant was calculated [Fig 1].INTRODUCTION
METHODS
Preoperative planning software for anatomic total shoulder arthroplasty (ATSA) allows surgeons to virtually perform a reconstruction based off 3D models generated from CT scans of the glenohumeral joint. The purpose of this study was to examine the distribution of chosen glenoid implant as a function of glenoid wear severity, and to evaluate the inter-surgeon variability of optimal glenoid component placement in ATSA. CT scans from 45 patients with glenohumeral arthritis were planned by 8 fellowship trained shoulder arthroplasty specialists using a 3D preoperative planning software, planning each case for optimal implant selection and placement. The software provided three implant types: a standard non-augmented glenoid component, and an 8° and 16° posterior augment wedge glenoid component. The software interface allowed the surgeons to control version, inclination, rotation, depth, anterior- posterior and superior-inferior position of the glenoid components in 1mm and 1° increments, which were recorded and compared for final implant position in each case.INTRODUCTION
METHODS
Preoperative planning software for reverse total shoulder arthroplasty (RTSA) allows surgeons to virtually perform a reconstruction based off 3D models generated from CT scans of the glenohumeral joint. While anatomical studies have defined the range of normal values for glenoid version and inclination, there is no clear consensus on glenoid component selection and position for RTSA. The purpose of this study was to examine the distribution of chosen glenoid implant as a function of glenoid wear severity, and to evaluate the inter-surgeon variability of optimal glenoid component placement in RTSA. CT scans from 45 patients with glenohumeral arthritis were planned by 8 fellowship trained shoulder arthroplasty specialists using a 3D preoperative planning software, planning each case for optimal implant selection and placement. The software provided four glenoid baseplate implant types: a standard non-augmented component, an 8° posterior augment wedged component, a 10° superior augment wedged component, and a combined 8° posterior and 10° superior wedged augment component. The software interface allowed the surgeons to control version, inclination, rotation, depth, anterior-posterior and superior-inferior position of the glenoid components in 1mm and 1° increments, which were recorded and compared for final implant position in each case.INTRODUCTION
METHODS
As computer navigated surgery continues to progress to the forefront of orthopedic care, the application of a navigated total shoulder arthroplasty has yet to appear. However, the accuracy of these systems is debated, as well as the dilemma of placing an accurate tool in an inaccurate hand. Often times a system's accuracy is claimed or validated based on postoperative imaging, but the true positioning is difficult to verify. In this study, a navigation system was used to preoperatively plan, guide, and implant surrogate shoulder glenoid implants and fiducials in nine cadaveric shoulders. A novel method to validate the position of these implants and accuracy of the system was performed using pre and post operative high resolution CT scans, in conjunction with barium sulfate impregnated PEEK surrogate implants. Nine cadaveric shoulders were CT scanned with .5mm slice thickness, and the digital models were incorporated into a preoperative planning software. Five orthopedic shoulder specialists used this software to virtually place aTSA and rTSA glenoid components in two cadavers each (one cadaver was omitted due to incomplete implantation), positioning the components as they best deemed fit. Using a navigation system, each surgeon registered the native cadaveric bone to each respective CT. Each surgeon then used the navigation system to guide him or her through the total shoulder replacement, and implant the barium sulfate impregnated PEEK surrogate implants. Four cylindrical PEEK fiducials were also implanted in each scapula to help triangulate the position of the surrogate implants. Previous efforts were attempted with stainless steel alloy fiducials, but position and image accuracy were limited by CT artifact. BaSO4 PEEK provided the highest resolution on a postoperative CT with as little artifact as possible. All PEEK fiducials and surrogate implants were registered by probing points and planes with the navigation system to capture the digital position. A high resolution post operative CT scan of each specimen was obtained, and variance between the executed surgical plan and PEEK fiducials was calculated.INTRODUCTION
METHODS
Total knee arthroplasty (TKA) is a common procedure with good success rates. The literature shows resection accuracy plays a crucial role in device longevity1. Computer guidance is used by some surgeons to enhance accuracy. This study reports on a continuous series of Optetrak knee prostheses (Exactech Inc., FL, USA) implanted by three senior surgeons between October 2010 and December 2013. 324 TKA were implanted at the Joseph Ducuing Hospital, Toulouse, France (Site 1), the Cleveland Clinic, Cleveland, OH, USA (Site 2) and the Riverview Hospital, Noblesville, IN, USA (Site 3) using Exactech GPS (Blue-Ortho, Grenoble, FR), a new computer-assisted guidance system. Each centre in this study used different surgical profiles defined specifically for their surgeical preferences. Planned tibial and femoral cuts were compared to actual cuts digitised using GPS. Operating time was analyzed and post-operative leg alignment was compared to pre-operative. The mean error between planned and digitised proximal tibial cuts was 0.06°±0.89 of valgus and 0.53°±0.90 of anterior slope for Site 1, 0.18°±0.85 of varus and 0.25°±1.18 of posterior slope for Site 2, and 0.02°±0.51 of valgus and 0.60°±1.15 of anterior slope for Site 3. The mean error between planned and digitised femoral distal cuts was 0.14°±0.85 of valgus and 0.49°±0.93 of flexion for Site 1, 0.15°±0.96 of varus and 0.04°±1.54 of extension for Site 2, and 0.09°±0.54 of varus and 0.48°±1.21 of extension for Site 3. Average operating time was 29 minutes for Site 1, 39 minutes for Site 2, and 33 minutes for Site 3. Post-operative Hip-Knee-Ankle angle (HKA) varied between 172° and 184° with an average of 179° for Site 1, 177° to 183° with an average of 179° for Site 2, and 177° to 185° with an average of 180° for Site 3. Pre-operative HKA ranged from 162 to 189°. Site 1 was already reporting in the series presented at ISTA 20132. Sites 2 and 3 were added later and could therefore benefit from the early feedback the analysis of site 1 cases provided. The use of the computer guidance at the new sites was associated with promising results and it did not take long to the surgeons to reach a reproducibility equivalent to the one of site 1. Average surgical time was similar in all three sites. GPS guidance added an average of 10 minutes to standard surgical times. All surgeons agreed the increased accuracy justified the additional time. Average post-operative HKA was 179°. HKA scores were within 3° of perfect alignment in 96% of the cases of Site 1, 99% of Site 2 and 97% of Site 3. According to the literature1, HKA between 177° and 183° is linked with high implant survival. Participating surgeons still associated Exactech GPS with satisfactory immediate post-operative results.
Clinical outcomes for total knee arthroplasty (TKA) are sensitive to lower extremity alignment, implant positioning, and implant size. Accurate determination of femoral implant size is the focus of this paper. As existing methods (conventional instrumentation, preoperative images, navigation) can be limited by issues including inaccuracy, time required, exposure, and cost, this study assesses a novel method for determining femoral component size using navigation. We used a commercially available navigation system (Exactech GPS, Blue Ortho, Grenoble, FR, with Total Knee V1.13 software). The system uses surface patches to collect small point clouds, and then computes points that match a given criteria (e.g. the most distal point). For femoral component sizing, the proposed method automatically defines a target area to be digitised on the anterior cortex. To do this, the surgeon acquires anatomical landmarks (i.e., knee centre, distal condyles, etc.) for all femoral implant parameters but the size. The surgeon then moves the tip of the acquisition instrument near the anterior cortex, and the system computes the distance between the virtual posterior cut and the tip in real time. The theoretical implant size increases in real time as the instrument tip moves anteriorly and decreases as it moves posteriorly. The target area is displayed on the anterior cortex such that it covers all the bone in the medio-lateral direction, is centred on the most proximal part of the theoretical implant in the proximal-distal direction, and covers the current size plus or minus one size. As a result, the target area virtually moves in the proximal-distal direction as the surgeon moves the instrument tip closer to the anterior cortex surface. When the tip is in contact with the anterior surface, acquisition of the point cloud is performed. From a user point of view, the system does not move the target area relative to the bone on the display, but instead adjusts the relative position of the instrument tip, creating the impression that no matter the bone size, the target area does not move and the instrument tip is always guided to the right spot. The method has been successfully implemented and used on more than 1,400 patients. A preliminary analysis on 189 surgical reports shows in 188 cases (99,5%) the proximal point of the selected implant is inside the target area (which means that the selected size is the one by default, plus or minus one). We conclude the proposed method as implemented in the Exactech GPS has proven to be clinically effective. It can easily be extended to determination of other points when global criteria can be used to define an optimal area of digitisation determined from previously acquired data.
Total knee arthroplasty (TKA) is a common procedure with good success rates. The literature shows resection accuracy plays a crucial role in device longevity1. Computer guidance is used by some surgeons to enhance accuracy. This study reports on a continuous series of Optetrak knee prostheses (Exactech Inc., FL, USA) implanted by three senior surgeons between July 2010 and April 2013. 259 TKA were implanted at the Haguenau Hospital, Haguenau, France (Site 1), Joseph Ducuing Hospital, Toulouse, France (Site 2) and Saint Michaels Medical Center, Newark, NJ, USA (Site 3) using Exactech GPS (Blue-Ortho, Grenoble, FR), a new computer-assisted guidance system. Surgeons can use the unique Exactech GPS profiler to define steps to be computer-assisted during surgery. Each centre in this study used different surgical profiles. Planned tibial and femoral cuts were compared to actual cuts digitised using GPS. Operating time and external femoral rotation were analyzed and post-operative leg alignment was compared to pre-operative. The mean error between planned and digitised proximal tibial cuts was 0.26° ± 1.11 of valgus and 0.06° ± 0.99 of posterior slope for Site 1, 0.07° ± 0.89 of varus and 0.53° ± 0.90 of anterior slope for Site 2, and 0.19° ± 0.73 of varus and 0.10° ± 1.17 of posterior slope for Site 3 (see Fig. 1). The mean error between planned and digitised femoral distal cuts was 0.03° ± 0.99 of varus and 0.67° ± 1.36 of extension for Site 1, 0.14° ± 0.85 of varus and 0.49° ± 0.94 of extension for Site 2, and 0.26° ± 0.86 of varus and 0.09° ± 1.22 of flexion for Site 3. Average operating time was 38 minutes for Site 1, 29 minutes for Site 2, and 34 minutes for Site 3. External femoral component rotation ranged from 0° to 18° with an average of 3.7° degrees for Site 1 and from −3° to 8° with an average of 3.0° for Site 2. External rotation was fixed at 3° for Site 3. Post-operative Hip-Knee-Ankle angle (HKA) varied between 177° and 182° with an average of 179° for Site 1, 172° to 184° with an average of 179° for Site 2, and 178° to 185° with an average of 180° for Site 3. Pre-operative HKA ranged from 162 to 191°. Despite different techniques and teams, all surgeons experienced similar results. Cuts were aligned in the frontal plane, while guidance was harder to follow in the sagittal plane, possibly due to saw blade bending during resection. Average surgical time was similar. GPS guidance added an average of 10 minutes to standard surgical times. All surgeons agreed the increased accuracy justified the additional time. Regardless the site, all average femoral rotations were close to the accepted 3° standard. Average post-operative HKA was 179°. HKA scores were within 3° of perfect alignment in all Site 1 cases and 96% of Site 2 and Site 3 cases. According to the literature1, HKA between 177° and 183° is linked with high implant survival. Participating surgeons associated Exactech GPS with satisfactory immediate post-operative results.
