Does commercially available shoulder arthroplasty preoperative planning software agree with surgeon measurements of version, inclination, and subluxation?

Machine learning can predict anterior elevation after reverse total shoulder arthroplasty: A new tool for daily outpatient clinic?

The aim of the present study was to individuate and compare specific machine learning algorithms that could predict postoperative anterior elevation score after reverse shoulder arthroplasty surgery at different time points. Data from 105 patients who underwent reverse shoulder arthroplasty at the same institute have been collected with the purpose of generating algorithms which could predict the target. Twenty-eight features were extracted and applied to two different machine learning techniques: Linear regression and support vector regression (SVR). These two techniques were also compared in order to define to most faithfully predictive. Using the extracted features, the SVR algorithm resulted in a mean absolute error (MAE) of 11.6° and a classification accuracy (PCC) of 0.88 on the test-set. Linear regression, instead, resulted in a MAE of 13.0° and a PCC of 0.85 on the test-set. Our machine learning study demonstrates that machine learning could provide high predictive algorithms for anterior elevation after reverse shoulder arthroplasty. The differential analysis between the utilized techniques showed higher accuracy in prediction for the support vector regression. Level of Evidence III: Retrospective cohort comparison; Computer Modeling.

Current clinical applications of artificial intelligence in shoulder surgery: what the busy shoulder surgeon needs to know and what's coming next

Background

Artificial intelligence (AI) is a continuously expanding field with the potential to transform a variety of industries-including health care-by providing automation, efficiency, precision, accuracy, and decision-making support for simple and complex tasks. Basic knowledge of the key features as well as limitations of AI is paramount to understand current developments in this field and to successfully apply them to shoulder surgery. The purpose of the present review is to provide an overview of AI within orthopedics and shoulder surgery exploring current and forthcoming AI applications.

Methods

PubMed and Scopus databases were searched to provide a narrative review of the most relevant literature on AI applications in shoulder surgery.

Results

Despite the enormous clinical and research potential of AI, orthopedic surgery has been a relatively late adopter of AI technologies. Image evaluation, surgical planning, aiding decision-making, and facilitating patient evaluations over time are some of the current areas of development with enormous opportunities to improve surgical practice, research, and education. Furthermore, the advancement of AI-driven strategies has the potential to create a more efficient medical system that may reduce the overall cost of delivering and implementing quality health care for patients with shoulder pathology.

Conclusion

AI is an expanding field with the potential for broad clinical and research applications in orthopedic surgery. Many challenges still need to be addressed to fully leverage the potential of AI to clinical practice and research such as privacy issues, data ownership, and external validation of the proposed models.

Navigation in reverse shoulder arthroplasty: how the lateralization of glenosphere can affect the clinical outcome

INTRODUCTION

One of the main causes of RSA failure is attributable to the malpositioning of the glenoid component. Initial experiences with computer-assisted surgery have shown promising results in increasing the accuracy and repeatability of placement of the glenoid component and screws. The aim of this study was to evaluate the functional clinical results, in terms of joint mobility and pain, by correlating them with intraoperative data regarding the positioning of the glenoid component. The hypothesis was that the lateralization more than 25 mm of the glenosphere can led to better stability of the prosthesis but should pay in term of a reduced range of movement and increased pain.

MATERIALS AND METHODS

50 patients were enrolled between October 2018 and May 2022; they underwent RSA implantation assisted by GPS navigation system. Active ROM, ASES score and VAS pain scale were recorded before surgery. Preoperative data about glenoid inclination and version were collected by pre-op X-Rays an CT. Intraoperative data-inclination, version, medialization and lateralization of the glenoid component-were recorded using computer-assisted surgery. 46 patients had been further clinically and radiographically re-evaluated at 3-months, 6-months, 1-year, and 2-years follow-up.

RESULTS

We found a statistically significant correlation between anteposition and glenosphere lateralization value (DM - 6.057 mm; p = 0.043). Furthermore a statistically significant correlation has been shown between abduction movement and the lateralization value (DM - 7.723 mm; p = 0.015). No other statistically significant associations were found when comparing the values of glenoid inclination and version with the range of motion achieved by the patients after reverse shoulder arthroplasty.

CONCLUSION

We observed that the patients with the best anteposition and abduction results had a glenosphere lateralization between 18 and 22 mm. When increasing the lateralization above 22 mm or reducing it below 18 mm, on the other hand, both movements considered decreased their range.

LEVEL OF EVIDENCE

Level IV; Case Series; Treatment Study.

Emerging Technologies in Shoulder Arthroplasty: Navigation, Mixed Reality, and Preoperative Planning

Shoulder arthroplasty is a rapidly improving and utilized management for end-stage arthritis that is associated with improved functional outcomes, pain relief, and long-term implant survival. Accurate placement of the glenoid and humeral components is critical for improved outcomes. Traditionally, preoperative planning was limited to radiographs and 2-dimensional computed tomography (CT); however, 3-dimensional CT is becoming more commonly utilized and necessary to understand complex glenoid and humeral deformities. To further increase accurate component placement, intraoperative assistive devices-patient-specific instrumentation, navigation, and mixed reality-minimize malpositioning, increase surgeon accuracy, and maximize fixation. These intraoperative technologies likely represent the future of shoulder arthroplasty.

Can We Completely Trust in Automated Software for Preoperative Planning of Shoulder Arthroplasty? Software Update May Modify Glenoid Version, Glenoid Inclination and Humeral Head Subluxation Values

Background: The purpose of this study was to evaluate the impact of software updating on measurements of the glenoid inclination and version, along with humeral head subluxation performed by an automated 3D planning program. The hypothesis was that the software update could significantly modify the values of the glenoid inclination and version, as well as of the humeral head subluxation. Methods: A comprehensive pool of 76 shoulder computed tomography (CT) scans of patients who underwent total shoulder arthroplasty (TSA) or reverse total shoulder arthroplasty (RTSA) were analyzed with the automated program Blueprint in 2018 and again in 2020 after a software update. Results: A statistically significant difference of 8.1 ± 8.2 and 5.4 ± 7.8 (mean difference of −2.8 ± 5.0, p < 0.001) was indeed reached when comparing the mean glenoid inclination achieved with Blueprint 2018 and Blueprint 2020, respectively. The glenoid version, as well as the humeral head subluxation evaluations, were not significantly different between the two software versions, with mean values being −9.4 ± 8.9 and −9.0 ± 7.4 and 60.1 ± 12.6 and 61.8 ± 12.0, respectively (p = 0.708 and p = 0.115, respectively). In 22% of CT scans, the software update determined a variation of the glenoid inclination of more than 5° or 10°. Conclusion: The present study shows the software update of an automated preoperative planning program may significantly modify the values of glenoid inclination. Even though without a significant difference, variations were also found for the glenoid version and humeral head subluxation. Accordingly, these results should further advise surgeons to carefully and critically evaluate data acquired with automated software.

Accuracy of Blueprint software in predicting range of motion 1 year after reverse total shoulder arthroplasty

HYPOTHESIS AND BACKGROUND

Blueprint 3-dimensional computed tomography software has a functionality that predicts impingement-free range of motion (ROM) with determination of the limits of ROM at which bone and/or prosthetic impingement occurs. To our knowledge, only 1 previously published study has assessed the ability of Blueprint software to predict actual postoperative ROM after reverse total shoulder arthroplasty (RTSA). The hypotheses of this study were that (1) mean Blueprint-predicted impingement-free ROM would be statistically similar to the mean actual ROM 1 year after RTSA and (2) there would be a correlation between Blueprint-predicted impingement-free ROM and the actual ROM 1 year after RTSA.

MATERIALS AND METHODS

A retrospective review of patients who underwent Blueprint planning prior to undergoing RTSA from March 2017 through May 2021 was performed. At 1-year follow-up, flexion, external rotation at the side, abduction, external rotation in the abducted position, internal rotation in the abducted position, and internal rotation behind the back were measured. The preoperatively predicted flexion, extension, abduction, external rotation, and internal rotation were recorded using Blueprint software. The group 1 analysis examined the predicted vs. actual ROM of all 127 patients regardless of whether intraoperative component modifications were made. The group 2 analysis examined the predicted vs. actual ROM of only the patients who did not undergo intraoperative changes that would affect the preoperative ROM prediction (n = 97). The group 3 analysis examined the predicted vs. actual ROM of group 2 combined with the 30 patients who underwent post hoc Blueprint planning modifications to account for the changes made intraoperatively (combined sample size of 127).

RESULTS

Of the 141 patients, 127 (90%) were available for 1-year follow-up. When the mean values of all 3 groups were examined, the actual ROM and predicted ROM were statistically significantly different (P < .0001) for flexion, external rotation, abduction, abduction-external rotation, and abduction-internal rotation. In group 1, a very weak or poor correlation was found between predicted internal rotation and actual abducted internal rotation (r = 0.19, P = .04). For all other ROM metrics in groups 1, 2, and 3, there were no correlations between predicted and actual ROM (P ≥ .07).

CONCLUSIONS

In its current state, preoperative Blueprint 3-dimensional computed tomography planning software is unable to accurately predict ROM 1 year after RTSA.

Difference analysis of the glenoid centerline between 3D preoperative planning and 3D printed prosthesis manipulation in total shoulder arthroplasty

INTRODUCTION

Excessive version and inclination of the glenoid component during total shoulder arthroplasty can lead to glenohumeral instability, early loosening, and even failure. The orientation and position of the central pin determine the version and inclination of the glenoid component. The purpose of this study was to compare the differences in centerline position and orientation obtained using "3D preoperative planning based on the best-fit method for glenoid elements" and the surgeon's manipulation.

MATERIALS AND METHODS

Twenty-nine CT images of glenohumeral osteoarthritis of the shoulder were reconstructed into a 3D model, and a 3D printer was used to create an in vitro model for the surgeon to drill the center pin. The 3D shoulder model was also used for 3D preoperative planning (3DPP) using the best-fit method for glenoid elements. The in vitro model was scanned and the version, inclination and center position were measured to compare with the 3DPP results.

RESULTS

The respective mean inclinations (versions) of the surgeon and 3DPP were -2.63° ± 6.60 (2.87° ± 5.97) and -1.96° ± 4.24 (-3.21° ± 4.00), respectively. There was no significant difference in the inclination and version of the surgeon and 3DPP. For surgeons, the probability of the inclination and version being greater than 10° was 13.8% (4/29) and 10.3% (3/29), respectively. Compared to the 3DPP results, the surgeon's center position was shifted down an average of 1.63 mm. There was a significant difference in the center position of the surgeon and 3DPP (p < 0.05).

CONCLUSION

The central pin drilled by surgeons using general instruments was significantly lower than those defined using 3D preoperative planning and standard central definitions. 3D preoperative planning prevents the version and inclination of the centerline from exceeding safe values (± 10°).

Preoperative Planning for Anatomic Total Shoulder Arthroplasty

The success of total shoulder arthroplasty is dependent on both proper patient selection and restoration of the native anatomy. After proper patient selection, preoperative planning is essential to select implants that will allow the surgeon to properly restore soft-tissue tension and correct for deformity. Although it is possible to template implants with plain radiographs, these do not allow accurate measurements of the complex three-dimensional anatomy of the glenohumeral joint. CT can be used to further examine version of the glenoid and humerus, as well as humeral head subluxation. Three-dimensional reconstructions also allow for virtual implantation, resulting in a more reliable prediction of implant appearance. Commercial software is available that calculates parameters such as version; however, these have been shown to have variability when compared with measurements obtained by surgeons. Patient-specific instrumentation can also be obtained based on preoperative measurements; however, although it allowed for improved measurements when compared with two-dimensional imaging, there has been no difference in version error, inclination error, or positional offset of the glenoid implant when comparing patient-specific instrumentation with standard instrumentation. Intraoperative navigation can also be used to give real-time feedback on implant positioning; however, additional studies are needed to fully evaluate its benefit.

Concordance of Preoperative 3D Templating in Stemless Anatomic Total Shoulder Arthroplasty

INTRODUCTION

Recent advances in preoperative 3D templating software allow surgeons to plan implant size and position for stemless total shoulder arthroplasty (TSA). Whether these preoperative plans accurately reflect intraoperative decisions is yet unknown, and the purpose of this study was to evaluate concordance between planned and actual implant sizes in a series of patients undergoing stemless TSA.

METHODS

A retrospective cohort of consecutive, anatomic, stemless TSA cases performed by two surgeons between September 2019 and February 2021 was examined. Preoperative templated plans were collected using 3D planning software, and the sizes of planned glenoid, humeral head, and nucleus "stem" implants and other procedural data were recorded, along with sociodemographic information. These predicted parameters were compared with the implant sizes, and the concordance of these templated plans was quantified by direct comparison and bootstrapped simulations.

RESULTS

Fifty cases met inclusion criteria, among which perfect concordance across all three implants was observed in 11 cases (22%). The glenoid implant had the highest concordance (80%) relative to the humeral head and nucleus implants (38% and 60%, respectively), which was statistically significant ( P < 0.001). Planned humeral head implants were more often oversized relative to their actual implanted size. However, 84% of the planned humeral heads were within 1 diameter size; in addition, 98% of the planned glenoid implants were within one size and all were within 10 mm of the implanted glenoid backside radius. All nucleus implants were within one size.

DISCUSSION

Final implant sizes demonstrated variable concordance relative to preoperative plans, with glenoid implants having the highest accuracy and humeral heads having the highest variability. Multiple factors contributed to the varying concordances for the different implants, suggesting possible areas of improvement in this technology. These results may have implications for logistics, intraoperative efficiency, and overall cost and underscore the potential value of this technology.

LEVEL OF EVIDENCE

Level III.

Failed Reverse Total Shoulder Arthroplasty: What Are Our Bailouts?

PURPOSE OF REVIEW

As the population continues to age and indications continue to expand, the number of reverse total shoulder arthroplasty (RSTA) procedures has increased significantly. While RTSA is an effective solution to many shoulder problems, it is not without complications. Furthermore, as the number of RTSA procedures increases, so will the number of complications following this procedure. While some complications can be managed with revision RTSA, there are some complications that, unfortunately, cannot. The purpose of this review is to discuss the revision options for failed RTSA.

RECENT FINDINGS

While there has been a significant amount of recent literature surrounding RTSA, much of this literature has been aimed at improving outcomes for primary RTSA by improving glenoid placement, maximizing range of motion, etc., or improving outcomes following conversion of another surgery to RTSA [1••, 2, 3]. There has been little evidence surrounding options for failed RTSA that cannot be salvaged to a revision RTSA. These options are limited and often involve resection arthroplasty and hemiarthroplasty, although neither option provides patients with significant function of the shoulder [4, 5•]. Complications following RTSA are becoming more common as the number of RTSA continues to increase. Furthermore, as the indications for RTSA expand, the complications will continue to increase as this implant is used to tackle more difficult problems about the shoulder. When possible, the etiology of the problem with the RTSA should be addressed and may involve component revision, bone grafting, etc. When the problem cannot be solved with revision RTSA, then the patient can be converted to a hemiarthroplasty, or have a resection arthroplasty, with the understanding that their shoulder function will be limited.

Computer-Assisted Surgery in Reverse Shoulder Arthroplasty: Early Experience

BACKGROUND

In the last decade, new technologies have been applied to shoulder arthroplasty. The aim of this work was to show that navigated RSA allows the surgeon to reach the planned version/inclination in all cases. In this article are shown preliminary data, advantages, disadvantages and limits of the technique.

METHODS

Eighteen computer-assisted reverse shoulder arthroplasty were performed. Preoperative glenoid version and inclination were evaluated with preoperative CT scan using Orthoblue^® (Exactech, Gainesville, FL,USA) software, as well as baseplate type, planned glenoid component seating, planned postoperative version, planned postoperative inclination, intraoperative glenoid version/inclination, screw length and surgical time. A senior shoulder surgeon has analyzed the advantages, disadvantages and limitation of this kind of surgery.

RESULTS

Mean surgical time of the primary implants was 92 ± 12 min (min 75-max 110). Mean preoperative inclination was + 2.6° ± 6.4, mean preoperative version was - 7.6° ± 8.4. Mean planned postoperative inclination was - 2.7° ± 2.3, mean planned postoperative version was - 1.6° ± 2.9 and mean planned glenoid seating was 89% ± 8%. Planned settings were reached in all cases during surgery. Baseplate implanted were in nine cases 8° posterior augmented, in six cases standard and in three cases 10° superior augmented. Mean screw length was 33.5 mm ± 4.2 mm. No GPS system failure has been recorded. One coracoid fracture occurred during the first case.

DISCUSSION

Intraoperative navigation system is a reliable and user-friendly technology that allows the surgeon to reach planned glenoid positioning during surgery. Furthermore, this technology will allow the surgeon to compare clinical outcomes to component positioning data. The lack of humeral implant navigation is the main limit of this technique.