Impact of Deltoid Computer Tomography Image Data on the Accuracy of Machine Learning Predictions of Clinical Outcomes after Anatomic and Reverse Total Shoulder Arthroplasty.

Background: Despite the importance of the deltoid to shoulder biomechanics, very few studies have quantified the three-dimensional shape, size, or quality of the deltoid muscle, and no studies have correlated these measurements to clinical outcomes after anatomic (aTSA) and/or reverse (rTSA) total shoulder arthroplasty in any statistically/scientifically relevant manner. Methods: Preoperative computer tomography (CT) images from 1057 patients (585 female, 469 male; 799 primary rTSA and 258 primary aTSA) of a single platform shoulder arthroplasty prosthesis (Equinoxe; Exactech, Inc., Gainesville, FL) were analyzed in this study. A machine learning (ML) framework was used to segment the deltoid muscle for 1057 patients and quantify 15 different muscle characteristics, including volumetric (size, shape, etc.) and intensity-based Hounsfield (HU) measurements. These deltoid measurements were correlated to postoperative clinical outcomes and utilized as inputs to train/test ML algorithms used to predict postoperative outcomes at multiple postoperative timepoints (1 year, 2-3 years, and 3-5 years) for aTSA and rTSA. Results: Numerous deltoid muscle measurements were demonstrated to significantly vary with age, gender, prosthesis type, and CT image kernel; notably, normalized deltoid volume and deltoid fatty infiltration were demonstrated to be relevant to preoperative and postoperative clinical outcomes after aTSA and rTSA. Incorporating deltoid image data into the ML models improved clinical outcome prediction accuracy relative to ML algorithms without image data, particularly for the prediction of abduction and forward elevation after aTSA and rTSA. Analyzing ML feature importance facilitated rank-ordering of the deltoid image measurements relevant to aTSA and rTSA clinical outcomes. Specifically, we identified that deltoid shape flatness, normalized deltoid volume, deltoid voxel skewness, and deltoid shape sphericity were the most predictive image-based features used to predict clinical outcomes after aTSA and rTSA. Many of these deltoid measurements were found to be more predictive of aTSA and rTSA postoperative outcomes than patient demographic data, comorbidity data, and diagnosis data. Conclusions: While future work is required to further refine the ML models, which include additional shoulder muscles, like the rotator cuff, our results show promise that the developed ML framework can be used to evolve traditional CT-based preoperative planning software into an evidence-based ML clinical decision support tool.

Planned Glenoid Implant Utilization: A Comparison of Preoperative Planning with and Without Computer-Assisted Navigation.

BACKGROUND: Preoperative planning for anatomic total shoulder arthroplasty (aTSA) and reverse total shoulder arthroplasty (rTSA) is becoming increasingly common. While preoperative planning allows surgeons to determine individualized implant types, utilization of intraoperative navigation improves the accuracy of implant placement and may increase confidence in the preoperative plan. The purpose of this study was to evaluate and compare the rate at which surgeons use a glenoid implant different than their preoperative plan with and without the use of computer navigation. METHODS: A retrospective review of a multicenter prospectively collected shoulder arthroplasty database was conducted between 2016 and 2022. Inclusion criteria were primary aTSA or rTSA with an available preoperative plan and record of the actual implant used. Change in glenoid implant was defined as a deviation in the final implant from the preoperative plan in regard to backside shape (nonaugmented vs augment or differing augment shape). RESULTS: We included 1,915 shoulder arthroplasties (525 aTSA, 1,390 rTSA) performed with preoperative planning and intraoperative navigation and 110 shoulder athroplasties (37 aTSA, 73 rTSA) performed with preoperative planning alone. Overall, the final glenoid implant deviated from the preoperative plan less frequently when intraoperative navigation was used compared with preoperative planning alone (1.9% [n = 36] versus 7.3% [n = 8], P = 0.002). When stratified by procedure, deviation from the preoperative plan occurred significantly less for rTSA when preoperative planning was used with intraoperative navigation versus planning alone (2% [n = 29] versus 11% [n = 8], P < 0.001; OR = 0.17 [95% CI = 0.07 to 0.46]), but not aTSA (1% [n = 7] versus 0% [n = 0], P = 1). Use of intraoperative navigation was independently associated with lower odds of deviation from the preoperative plan on multivariable logistic regression (OR = 0.25 [95% CI = 0.11 to 0.56], P = 0.001). CONCLUSION: Use of intraoperative navigation is associated with increased adherence to the preoperative plan for primary rTSA. Use of navigation may increase surgeon confidence despite known limitations of glenoid visualization during this procedure. This may offer advantages in outpatient surgery centers and smaller hospitals where inventory space may be limited. LEVEL OF EVIDENCE: Ⅲ, retrospective cohort study.

Initial clinical experience with a predictive clinical decision support tool for anatomic and reverse total shoulder arthroplasty.

PURPOSE: Clinical decision support tools (CDSTs) are software that generate patient-specific assessments that can be used to better inform healthcare provider decision making. Machine learning (ML)-based CDSTs have recently been developed for anatomic (aTSA) and reverse (rTSA) total shoulder arthroplasty to facilitate more data-driven, evidence-based decision making. Using this shoulder CDST as an example, this external validation study provides an overview of how ML-based algorithms are developed and discusses the limitations of these tools. METHODS: An external validation for a novel CDST was conducted on 243 patients (120F/123M) who received a personalized prediction prior to surgery and had short-term clinical follow-up from 3 months to 2 years after primary aTSA (n = 43) or rTSA (n = 200). The outcome score and active range of motion predictions were compared to each patient's actual result at each timepoint, with the accuracy quantified by the mean absolute error (MAE). RESULTS: The results of this external validation demonstrate the CDST accuracy to be similar (within 10%) or better than the MAEs from the published internal validation. A few predictive models were observed to have substantially lower MAEs than the internal validation, specifically, Constant (31.6% better), active abduction (22.5% better), global shoulder function (20.0% better), active external rotation (19.0% better), and active forward elevation (16.2% better), which is encouraging; however, the sample size was small. CONCLUSION: A greater understanding of the limitations of ML-based CDSTs will facilitate more responsible use and build trust and confidence, potentially leading to greater adoption. As CDSTs evolve, we anticipate greater shared decision making between the patient and surgeon with the aim of achieving even better outcomes and greater levels of patient satisfaction.

Two-year clinical outcomes and complication rates in anatomic and reverse shoulder arthroplasty implanted with Exactech GPS intraoperative navigation.

INTRODUCTION: We compared the 2-year clinical outcomes of both anatomic and reverse total shoulder arthroplasty (ATSA and RTSA) using intraoperative navigation compared to traditional positioning techniques. We also examined the effect of glenoid implant retroversion on clinical outcomes. HYPOTHESIS: In both ATSA and RTSA, computer navigation would be associated with equal or better outcomes with fewer complications. Final glenoid version and degree of correction would not show outcome differences. MATERIAL AND METHODS: A total of 216 ATSAs and 533 RTSAs were performed using preoperative planning and intraoperative navigation with a minimum of 2-year follow-up. Matched cohorts (2:1) for age, gender, and follow-up for cases without intraoperative navigation were compared using all standard shoulder arthroplasty clinical outcome metrics. Two subanalyses were performed on navigated cases comparing glenoids positioned greater or less than 10° of retroversion and glenoids corrected more or less than 15°. RESULTS: For ASTA, no statistical differences were found between the navigated and non-navigated cohorts for postoperative complications, glenoid implant loosening, or revision rate. No significant differences were seen in any of the ATSA outcome metrics besides higher internal and external rotation in the navigated cohort. For RTSA, the navigated cohort showed an ARR of 1.7% (95% CI 0%, 3.4%) for postoperative complications and 0.7% (95% CI 0.1%, 1.2%) for dislocations. No difference was found in the revision rate, glenoid implant loosening, acromial stress fracture rates, or scapular notching. Navigated RTSA patients demonstrated significant improvements over non-navigated patients in internal rotation, external rotation, maximum lifting weight, the Simple Shoulder Test (SST), Constant, and Shoulder Arthroplasty Smart (SAS) scores. For the navigated subcohorts, ATSA cases with a higher degree of final retroversion showed significant improvement in pain, Constant, American Shoulder and Elbow Surgeons Standardized Shoulder Assessment Form (ASES), SST, University of California-Los Angeles shoulder score (UCLA), and Shoulder Pain and Disability Index (SPADI) scores. No significant differences were found in the RTSA subcohort. Higher degrees of version correction showed improvement in external rotation, SST, and Constant scores for ATSA and forward elevation, internal rotation, pain, SST, Constant, ASES, UCLA, SPADI, and SAS scores for RTSA. CONCLUSION: The use of intraoperative navigation shoulder arthroplasty is safe, produces at least equally good outcomes at 2 years as standard instrumentation does without any increased risk of complications. The effect of final implant position above or below 10° of glenoid retroversion and correction more or less than 15° does not negatively impact outcomes.

High intraoperative accuracy and low complication rate of computer-assisted navigation of the glenoid in total shoulder arthroplasty.

BACKGROUND: Preoperative planning software with intraoperative guidance technology is increasingly being used to manage complex glenoid deformity in anatomic total shoulder arthroplasty (TSA) and reverse TSA. The aim of this study was to review the intraoperative efficacy and complications of computer-assisted navigation (CAN) surgery for the treatment of glenoid deformity in TSA. METHODS: We performed a retrospective review of all TSAs implanted using a single computer navigation shoulder system. All patients underwent preoperative planning with computed tomography-based preoperative planning software. The starting point on the glenoid and the final version and inclination of the central post (cage) of the glenoid component were reviewed on the intraoperative navigation guidance report and compared with these parameters on the preoperative plan for each patient. The intraoperative accuracy of CAN for glenoid positioning was determined by the deviation of the starting point and final position of the central cage drill in the glenoid compared with the preoperative plan. Data regarding intraoperative complications and the number of times the navigation system was abandoned intraoperatively were collected. RESULTS: A total of 16,723 anatomic TSAs and reverse TSAs performed worldwide with the aforementioned navigation system were included in this review. In 16,368 cases (98%), every step of the navigation procedure was completed without abandoning use of the system intraoperatively. There was minimal deviation in the intraoperative execution of the preoperative plan with respect to version (0.6° ± 1.96°), inclination (0.2° ± 2.04°), and the starting point on the glenoid face (1.90 ± 1.2 mm). In this cohort, 9 coracoid fractures (0.05%) were reported. CONCLUSION AND DISCUSSION: This study demonstrates the safety and efficacy of CAN for glenoid implantation in TSA. Future studies should focus on assessing the impact of CAN on the longevity and survival of glenoid components and improving the cost-effectiveness of this technology.

Intersurgeon and intrasurgeon variability in preoperative planning of anatomic total shoulder arthroplasty: a quantitative comparison of 49 cases planned by 9 surgeons.

BACKGROUND: Preoperative planning software is widely available for most anatomic total shoulder arthroplasty (ATSA) systems. It can be most useful in determining implant selection and placement with advanced glenoid wear. The purpose of this study was to quantify inter- and intrasurgeon variability in preoperative planning of a series of ATSA cases. METHODS: Forty-nine computed tomography scans were planned for ATSA by 9 fellowship-trained shoulder surgeons using the ExactechGPS platform (Exactech Inc., Gainesville, FL, USA). Each case was planned a second time between 4 and 12 weeks later. Variability within and between surgeons was measured for implant type, size, version and inclination correction, and implant face position. Interclass correlation coefficients, Pearson, and Light's kappa coefficients were used for statistical analysis. RESULTS: There was considerable variation in the frequency of augment use between surgeons and between rounds for the same surgeon. Thresholds for augment use also varied between surgeons. Interclass correlation coefficients for intersurgeon variability were 0.37 for version, 0.80 for inclination, 0.36 for implant type, and 0.36 for implant size. Pearson coefficients for intrasurgeon variability were 0.17 for version and 0.53 for inclination. Light's kappa coefficient for implant type was 0.64. CONCLUSIONS: This study demonstrates substantial inter- and intrasurgeon variability in preoperative planning of ATSA. Although the magnitude of differences in correction was small, surgeons differed significantly in the use of augments to achieve the resultant plan. Surgeons differed from each other on thresholds for augment use and maximum allowable residual retroversion. This suggests that there may a range of acceptable corrections for each shoulder rather than a single optimal plan.

Assessment of surgeon variability in preoperative planning of reverse total shoulder arthroplasty: a quantitative comparison of 49 cases planned by 9 surgeons.

BACKGROUND: Preoperative planning software is gaining utility in reverse total shoulder arthroplasty (RTSA), particularly when addressing pathologic glenoid wear. The purpose of this study was to quantify inter- and intrasurgeon variability in preoperative planning a series of RTSA cases to identify differences in how surgeons consider optimal implant placement. This may help identify opportunities to establish consensus when correlating plan differences with clinical data. METHODS: A total of 49 computed tomography scans from actual RTSA cases were planned for RTSA by 9 fellowship-trained shoulder surgeons using the same platform (Exactech GPS, Exactech Inc., Gainesville, FL, USA). Each case was planned a second time 6-12 weeks later. Variability within and between surgeons was measured for implant selection, version correction, inclination correction, and implant face position. Interclass correlation coefficients, and Pearson and Light's kappa coefficient were used for statistical analysis. RESULTS: There was considerable variation in the frequency of augmented baseplate selection between surgeons and between rounds for the same surgeon. Thresholds for augment use also varied between surgeons. Interclass correlation coefficients for intersurgeon variability ranged from 0.43 for version, 0.42 for inclination, and 0.25 for baseplate type. Pearson coefficients for intrasurgeon variability were 0.34 for version and 0.30 for inclination. Light's kappa coefficient for baseplate type was 0.61. CONCLUSIONS: This study demonstrates substantial variability both between surgeons and between rounds for individual surgeons when planning RTSA. Although average differences between plans were relatively small, there were large differences in specific cases suggesting little consensus on optimal planning parameters and opportunities to establish guidelines based on glenoid pathoanatomy. The correlation of preoperative planning with clinical outcomes will help to establish such guidelines.

Distribution of Glenoid Implant Options for Correcting Deformities Using a Preoperative Planning Tool.

Preoperative planning tools in shoulder arthroplasty are a recently developing technology with the advantage of being able to clearly assess patient anatomy and deformities before entering the OR. Addressing retroverted glenoids remains one of the most difficult aspects of primary shoulder arthroplasty. In this study, five surgeons were provided with a preoperative planning tool with posterior augmented glenoid implant options (0°, 8°, and 16°) to treat 10 cadaveric cases with a range of versions from 7.8° anteversion to 25.1° retroversion. Surgeons were able to remove less bone using 8° augmented implants over standard non-augmented implants (2.8° reamed vs. 6.4° reamed) and were able to correct each case on average within ± 1.8° of neutral version. Slight glenoid vault perforation was observed in 18% of the plans. Eight degrees posterior augmented implants were used in scans averaging 9.0° retroversion, and 16° posterior augmented implants were used in scans averaging 20.6° retroversion. Results were then compared to 14 preoperative CT scans provided by one of the surgeons in which both 8° and 16° posterior augmented glenoid implants were used in actual patients, showing 8° posterior augmented implants were used in cases averaging 12.3° retroversion, and 16° posterior augmented implants were used in cases averaging 20.7° retroversion. The study shows that surgeons can effectively and predictably use a preoperative planning tool to correct glenoid abnormalities using augmented implant solutions while minimizing both scapular bone removal and vault perforation and maximizing version correction.

Evaluation of Preoperative Implant Placement in Total Shoulder Arthroplasty.

INTRODUCTION: New technology to assist with glenoid placement in shoulder arthroplasty has evolved to include preoperative planning tools and intraoperative guides. These tools provide surgeons with a more complete understanding of glenoid anatomy prior to surgery. However, there have been no studies identifying the information that most influences surgical decision making. Further, there have been few studies that quantify intraoperative identification of scapular landmarks required to execute a preoperative plan. The purpose of this study is to examine the variables that are considered when making a preoperative plan in shoulder arthroplasty. METHODS: The first part of this study was a cadaveric lab in which three surgeons identified the neutral axis in surgical simulation. The second part of the study utilized a preliminary software tool in which surgeons were able to place glenoid implants in a set of CT reconstructions utilizing standard pegged glenoid components. In the third part of the study, surgeons utilized a novel planning software that included the ability to view the 3D reconstructed glenoid in all planes simultaneously and place either standard or augmented glenoid implants. The results of these three studies were compared. RESULTS: The center of the glenoid identified in the cadaver lab was 1.69 mm ± 1.58 mm anterior and 1.99 mm ± 2.49 mm superior to center. The identified neutral axis was tilted 14.2° ± 9.2° superior to the Friedman axis with 11.8° ± 7.9° of retroversion relative to that axis. Using the novel preoperative planning tool, the surgeons placed implants less than 0.5 mm from the center of the glenoid (AP = -0.07 mm ± 0.42 mm, SI = 0.44 mm ± 0.82 mm) with an average retroversion of less than 1° (-0.96° ± 3.04°). CONCLUSION: There was a discernible difference between the neutral axis identified in the cadaveric simulation (aver age of 14.2° superior and 11.8° retroverted) and the implant orientation planned using preoperative software (average of 3.26° superior and 0.96° retroverted). Based on the variability of position and orientation seen cadaverically, it is concluded that additional intraoperative guidance is needed alongside a preoperative plan in order to execute ideal placement of the glenoid component.