A practical guide to the implementation of AI in orthopaedic research, Part 6: How to evaluate the performance of AI research?

Artificial intelligence's (AI) accelerating progress demands rigorous evaluation standards to ensure safe, effective integration into healthcare's high-stakes decisions. As AI increasingly enables prediction, analysis and judgement capabilities relevant to medicine, proper evaluation and interpretation are indispensable. Erroneous AI could endanger patients; thus, developing, validating and deploying medical AI demands adhering to strict, transparent standards centred on safety, ethics and responsible oversight. Core considerations include assessing performance on diverse real-world data, collaborating with domain experts, confirming model reliability and limitations, and advancing interpretability. Thoughtful selection of evaluation metrics suited to the clinical context along with testing on diverse data sets representing different populations improves generalisability. Partnering software engineers, data scientists and medical practitioners ground assessment in real needs. Journals must uphold reporting standards matching AI's societal impacts. With rigorous, holistic evaluation frameworks, AI can progress towards expanding healthcare access and quality. Level of Evidence: Level V.

Fragment size of lateral Hoffa fractures determines screw fixation trajectory: a human cadaveric cohort study.

BACKGROUND AND PURPOSE: Recommendations regarding fragment-size-dependent screw fixation trajectory for coronal plane fractures of the posterior femoral condyles (Hoffa fractures) are lacking. The aim of this study was to compare the biomechanical properties of anteroposterior (AP) and crossed posteroanterior (PA) screw fixations across differently sized Hoffa fractures on human cadaveric femora. PATIENTS AND METHODS: 4 different sizes of lateral Hoffa fractures (n = 12 x 4) were created in 48 distal human femora according to the Letenneur classification: (i) type I, (ii) type IIa, (ii) type IIb, and (iv) type IIc. Based on bone mineral density (BMD), specimens were assigned to the 4 fracture clusters and each cluster was further assigned to fixation with either AP (n = 6) or crossed PA screws (n = 6) to ensure homogeneity of BMD values and comparability between the different test conditions. All specimens were biomechanically tested under progressively increasing cyclic loading until failure, capturing the interfragmentary movements via motion tracking. RESULTS: For Letenneur type I fractures, kilocycles to failure (mean difference [∆] 2.1, 95% confidence interval [CI] -1.3 to 5.5), failure load (∆ 105 N, CI -83 to 293), axial displacement (∆ 0.3 mm, CI -0.8 to 1.3), and fragment rotation (∆ 0.5°, CI -3.2 to 2.1) over 5.0 kilocycles did not differ significantly between the 2 screw trajectories. For each separate subtype of Letenneur type II fractures, fixation with crossed PA screws resulted in significantly higher kilocycles to failure (∆ 6.7, CI 3.3-10.1 to ∆ 8.9, CI 5.5-12.3) and failure load (∆ 275 N, CI 87-463 to ∆ 438, CI 250-626), as well as, less axial displacement from 3.0 kilocycles onwards (∆ 0.4°, CI 0.03-0.7 to ∆ 0.5°, CI 0.01-0.9) compared with AP screw fixation. CONCLUSION: Irrespective of the size of Letenneur type II fractures, crossed PA screw fixation provided greater biomechanical stability than AP-configured screws, whereas both screw fixation techniques demonstrated comparable biomechanical competence for Letenneur type I fractures. Fragment-size-dependent treatment strategies might be helpful to determine not only the screw configuration but also the surgical approach.

A practical guide to the implementation of artificial intelligence in orthopaedic research-Part 2: A technical introduction.

Recent advances in artificial intelligence (AI) present a broad range of possibilities in medical research. However, orthopaedic researchers aiming to participate in research projects implementing AI-based techniques require a sound understanding of the technical fundamentals of this rapidly developing field. Initial sections of this technical primer provide an overview of the general and the more detailed taxonomy of AI methods. Researchers are presented with the technical basics of the most frequently performed machine learning (ML) tasks, such as classification, regression, clustering and dimensionality reduction. Additionally, the spectrum of supervision in ML including the domains of supervised, unsupervised, semisupervised and self-supervised learning will be explored. Recent advances in neural networks (NNs) and deep learning (DL) architectures have rendered them essential tools for the analysis of complex medical data, which warrants a rudimentary technical introduction to orthopaedic researchers. Furthermore, the capability of natural language processing (NLP) to interpret patterns in human language is discussed and may offer several potential applications in medical text classification, patient sentiment analysis and clinical decision support. The technical discussion concludes with the transformative potential of generative AI and large language models (LLMs) on AI research. Consequently, this second article of the series aims to equip orthopaedic researchers with the fundamental technical knowledge required to engage in interdisciplinary collaboration in AI-driven orthopaedic research. Level of Evidence: Level IV.

Analysis of postoperative complications 5 years after osteosynthesis of patella fractures-a retrospective, multicenter cohort study.

PURPOSE: The study aims to investigate the influence of patient- and fracture-specific factors on the occurrence of complications after osteosynthesis of patella fractures and to compare knee joint function, activity, and subjective pain levels after a regular postoperative course and after complications in the medium term. METHODS: This retrospective, multicenter cohort study examined patients who received surgery for patella fracture at level 1 trauma centers between 2013 and 2018. Patient demographics and fracture-specific variables were evaluated. Final follow-up assessments included patient-reported pain scores (NRS), subjective activity and knee function scores (Tegner Activity Scale, Lysholm score, IKDC score), complications, and revisions. RESULTS: A total of 243 patients with a mean follow-up of 63.4 ± 21.3 months were included. Among them, 66.9% of patients underwent tension band wiring (TBW), 19.0% received locking plate osteosynthesis (LPO), and 14.1% underwent screw osteosynthesis (SO). A total of 38 patients (15.6%) experienced complications (TBW: 16.7%; LPO: 15.2%; SO: 11.8%). Implant-related complications of atraumatic fragment dislocation and material insufficiency/dislocation, accounted for 50% of all complications, were significantly more common after TBW than LPO (p = 0.015). No patient-specific factor was identified as a general cause for increased complications. Overall, particularly following complications such as limited range of motion or traumatic refracture, functional knee scores were significantly lower and pain levels were significantly higher at the final follow-up when a complication occurred. Implant-related complications, however, achieved functional scores comparable to a regular postoperative course without complications after revision surgery. CONCLUSION: The present study demonstrated that implant-related complications occurred significantly more often after TBW compared to LPO. The complication rates were similar in all groups.

Biomechanical Stability of Third-Generation Adjustable Suture Loop Devices Versus Continuous Loop Button Device for Cortical Fixation of ACL Tendon Grafts.

Background: Concerns regarding the primary stability of early adjustable loop button (ALB) devices for cortical fixation of tendon grafts in anterior cruciate ligament reconstruction (ACLR) have led to the development of new implant designs. Purpose: To evaluate biomechanical stability of recent ALB implants in comparison with a continuous loop button (CLB) device. Study Design: Controlled laboratory study. Methods: ACLR was performed in a porcine model (n = 40) using 2-strand porcine flexor tendons with a diameter of 8 mm. Three ALB devices (Infinity Button [ALB1 group]; Tightrope II RT [ALB2 group]; A-TACK [ALB3 group]) and 1 CLB device (FlippTack with polyethylene suture) were used for cortical tendon graft fixation. Cyclic loading (1000 cycles up to 250 N) with complete unloading were applied to the free end of the tendon graft using a uniaxial testing machine, followed by load to failure. Elongation, stiffness, yield load, and ultimate failure load were recorded and compared between the groups using a Kruskal-Wallis test with post hoc Dunn correction. Results: Elongation after 1000 cycles at 250 N was similar between groups (ALB1, 4.5 ± 0.7 mm; ALB2, 4.8 ± 0.8 mm; ALB3, 4.5 ± 0.6 mm; CLB, 4.5 ± 0.8 mm), as was load to failure (ALB1, 838 ± 109 N; ALB2, 930 ± 89 N; ALB3, 809 ± 103 N; CLB, 842 ± 80 N). Stiffness was significantly higher in the ALB1 group compared with the CLB group (262.3 ± 21.6 vs 229.3 ± 15.1 N/mm; P < .05). No significant difference was found between the 4 groups regarding yield load. Constructs failed either by rupture of the loop, breakage of the button, or rupture of the tendon. Conclusion: The tested third-generation ALB devices for cortical fixation in ACLR withstood cyclic loading with complete unloading without significant differences to a CLB device. Clinical Relevance: The third-generation ALB devices tested in the present study provided biomechanical stability comparable with that of a CLB device. Furthermore, ultimate failure loads of all tested implants exceeded the loads expected to occur in the postoperative period after ACLR.

Effects of transcranial direct current stimulation and sensorimotor training in anterior cruciate ligament patients: a sham-controlled pilot study.

BACKGROUND: Studies showed changes in the central nervous system in patients who sustained an anterior cruciate ligament tear. There is a lack of evidence regarding the effectiveness of transcranial direct-current stimulation in such patients. METHODS: A sham-controlled randomised study. One group of patients (n = 6) underwent 6 weeks of sensorimotor training after an anterior cruciate ligament tear during transcranial direct-current stimulation. The stimulation consisted of 20 minutes (3 sessions/week; 2 weeks) of 2 mA anodal transcranial direct-current stimulation over the primary motor and premotor cortex. The second group (n = 6) received sham stimulation with 6 weeks of sensorimotor training. Centre of pressure deviations in the medio-lateral and anterior-posterior direction and centre of pressure velocity were measured. RESULTS: The results demonstrated a significant effect of sensorimotor training on the centre of pressure in medio-lateral and anterior-posterior direction (p=0.025) (p=0.03) in the leg in which an anterior cruciate ligament tear occurred. The type of training did not affect the results. Post-hoc tests showed no significant effect of training in the subgroups (p≥0.115). CONCLUSION: Sensorimotor training led to a decrease in sway of the centre of pressure in patients who sustained an anterior cruciate ligament tear, but the addition of anodal transcranial direct-current stimulation placed over the primary motor cortex did not potentiate the adaptive responses of the sensorimotor training.

Comparison of Time-Zero Primary Stability Between a Biodegradable Magnesium Bone Staple and Metal Bone Staples for Knee Ligament Fixation: A Biomechanical Study in a Porcine Model.

Background: Bone staples have been shown previously to be a viable modality for cortical tendon graft fixation in ligament knee surgery. However, soft tissue reactions have been reported, making implant removal necessary. Magnesium alloys are a promising material for biodegradable orthopaedic implants, with mechanical properties closely resembling those of human bone. Purpose: To compare the primary stability of a biodegradable bone staple prototype made from magnesium to bone staples made from metal in the cortical fixation of tendon grafts during knee surgery. Study Design: Controlled laboratory study. Methods: Primary stability of peripheral tendon graft fixation was assessed in a porcine model of medial collateral ligament reconstruction. Two commercially available metal bone staples (Richards fixation staple with spikes [Me1] and spiked ligament staple [Me2]) were compared with a magnesium bone staple prototype for soft tissue fixation. Primary stability was assessed using a uniaxial materials testing machine. Cyclic loading at 50 and 100 N was applied for 500 cycles each, followed by load-to-failure testing. Results: After 500 cycles at 50 N, elongation was 1.5 ± 0.5 mm in the Me1 group, 1.9 ± 0.5 mm in the Me2 group, and 1.8 ± 0.4 mm in the magnesium group. After 1000 cycles of loading (500 cycles at 50 N and 500 at 100 N), elongation was 3.6 ± 0.9 mm in the Me1 group, 3.5 ± 0.6 mm in the Me2 group, and 4.1 ± 1.0 mm in the magnesium group. No significant differences regarding elongation were found between the groups. Load to failure was 352 ± 115 N in the Me1 group, 373 ± 77 N in the Me2 group, and 449 ± 92 N in the magnesium group, with no significant difference between the groups. Conclusion: In this study, the magnesium bone staples provided appropriate time-zero biomechanical primary stability in comparison with metal bone staples and may therefore be a feasible alternative for cortical fixation of tendon grafts in knee surgery. Clinical Relevance: The biodegradability of magnesium bone staples would eliminate the need for later implant removal.

Isometry of anteromedial reconstructions mimicking the deep medial collateral ligament depends on the femoral insertion.

PURPOSE: This study aimed to investigate the length change patterns of the native deep medial collateral ligament (dMCL) and potential anteromedial reconstructions (AMs) that might be added to a reconstruction of the superficial MCL (sMCL) to better understand the control of anteromedial rotatory instability (AMRI). METHODS: Insertion points of the dMCL and potential AM reconstructions were marked with pins (tibial) and eyelets (femoral) in 11 cadaveric knee specimens. Length changes between the pins and eyelets were then tested using threads in a validated kinematics rig with muscle loading of the quadriceps and iliotibial tract. Between 0° and 100° knee flexion, length change pattern of the anterior, middle and posterior part of the dMCL and simulated AM reconstructions were analysed using a rotary encoder. Isometry was tested using the total strain range (TSR). RESULTS: The tibiofemoral distance of the anterior dMCL part lengthened with flexion (+12.7% at 100°), whereas the posterior part slackened with flexion (-12.9% at 100°). The middle part behaved almost isometrically (maximum length: +2.8% at 100°). Depending on the femoral position within the sMCL footprint, AM reconstructions resulted in an increase in length as the knee flexed when a more centred position was used, irrespective of the tibial attachment position. Femoral positioning in the posterior aspect of the sMCL footprint exhibited <4% length change and was slightly less tight in flexion (min TSR = 3.6 ± 1.5%), irrespective of the tibial attachment position. CONCLUSION: The length change behaviour of potential AM reconstructions in a functionally intact knee is mainly influenced by the position of the femoral attachment, with different tibial attachments having a minimal effect on length change. Surgeons performing AM reconstructions to control AMRI would be advised to choose a femoral graft position in the posterior part of the native sMCL attachment to optimise graft length change behaviour. Given the high frequency of MCL injuries, sufficient restoration of AMRI is essential in isolated and combined ligamentous knee injuries. LEVEL OF EVIDENCE: There is no level of evidence as this study was an experimental laboratory study.

Valgus malalignment causes increased forces on a medial collateral ligament reconstruction under dynamic valgus loading: A biomechanical study.

PURPOSE: To investigate the forces on a medial collateral ligament (MCL) reconstruction (MCLR) relative to the valgus alignment of the knee. METHODS: Eight fresh-frozen human cadaveric knees were subjected to dynamic valgus loading at 400 N using a custom-made kinematics rig. After resection of the superficial medial collateral ligament, a single-bundle MCLR with a hamstring tendon autograft was performed. A medial opening wedge distal femoral osteotomy was performed and fixed with an external fixator to gradually adjust the alignment in 5° increments from 0° to 10° valgus. For each degree of valgus deformity, the resulting forces acting on the MCLR were measured through a force sensor and captured in 15° increments from 0° to 60° of knee flexion. RESULTS: Irrespective of the degree of knee flexion, increasing valgus malalignment resulted in significantly increased forces acting on the MCLR compared to neutral alignment (p < 0.05). Dynamic loading at 5° valgus resulted in increased forces on the MCLR at all flexion angles ranging between 16.2 N and 18.5 N (p < 0.05 from 0° to 30°; p < 0.01 from 45° to 60°). A 10° valgus malalignment further increased the forces on the MCLR at all flexion angles ranging between 29.4 N and 40.0 N (p < 0.01 from 0° to 45°, p < 0.05 at 60°). CONCLUSION: Valgus malalignment of the knee caused increased forces acting on the reconstructed MCL. In cases of chronic medial instabilities accompanied by a valgus deformity ≥ 5°, a realigning osteotomy should be considered concomitantly to the MCLR to protect the graft and potentially reduce graft failures. LEVEL OF EVIDENCE: Level III.

The anteromedial retinaculum in ACL-injured knees: An overlooked injury?

PURPOSE: The purpose of this study was to retrospectively analyse the pattern of injury to the medial knee structures in anterior cruciate ligament (ACL) injured patients. It was hypothesised that anteromedial injuries would be more common than posteromedial lesions. METHODS: One hundred and twenty subjects aged 18-25 years with a primary ACL injury were included. Patients were excluded if the time between injury and magnetic resonance imaging (MRI) was more than 28 days or if a knee dislocation or fracture was present. The MRIs were analysed with particular emphasis on injuries to the medial knee structures, menisci and bone bruise patterns. Injuries to the ligaments and anteromedial retinaculum (AMR) were graded according to severity, ranging from periligamentous oedema (grade I), partial fibre disruption of less or more than 50% (grade IIa or IIb) to complete tears (grade III). RESULTS: AMR injury was seen in 87 subjects (72.5%) on the coronal plane and in 88 (73.3%) on the axial plane, with grade III lesions observed in 27 (22.5%) and 29 knees (24.2%). Injuries to the superficial medial collateral ligament (sMCL), deep MCL (dMCL) and posterior oblique ligament (POL) were detected in 60 patients (50%), 93 patients (77.5%) and 38 patients (31.6%). However, grade III injuries to the POL were observed in only seven knees (5.8%). Medial meniscus injuries were associated with lesions of the sMCL and AMR (p < 0.05), while lateral meniscus injuries were significantly more common in patients with dMCL rupture (p < 0.05). CONCLUSION: Data from this study suggest that injuries to the AMR are much more common than posteromedial lesions in subjects with ACL injuries. LEVEL OF EVIDENCE: Level IV.

The Effect of Varying Sizes of Ramp Lesions in the ACL-Deficient and Reconstructed Knee: A Biomechanical Robotic Investigation.

BACKGROUND: Conflicting evidence has been reported regarding the biomechanical relevance of ramp lesions (RLs) on knee kinematics. Furthermore, the influence of the defect size of the RLs on anterior tibial translation (ATT) and external rotation (ER) is currently unknown. PURPOSE: To evaluate the influence of RL defect size on knee kinematics in anterior cruciate ligament (ACL) deficiency and after simulated ACL reconstruction (sACLR). STUDY DESIGN: Controlled laboratory study. METHODS: Eight cadaveric knee specimens were tested in a 6 degrees of freedom robotic test setup. Force-controlled clinical laxity tests were performed with 200 N of axial compression in 0°, 30°, 60°, and 90° of flexion: 5 N·m internal rotation (IR)/ER torque, 134 N ATT force, and an anteromedial drawer test consisting of 134 N ATT force under 5 N·m ER torque. After determining the native knee kinematics, the ACL was cut at the tibial insertion, followed by a transosseous refixation to simulate a surgical repair or reconstruction (simulated ACL reconstruction; sACLR). An RL was sequentially created with a length of 1, 2, and 3 cm. Each state of the RL was evaluated in the ACL-deficient state and after sACLR. RESULTS: In the ACL-deficient state, only an RL of 3 cm length resulted in a significant increase of ATT in 30° of flexion (mean difference 0.73 mm; 95% CI, 0.36-1.1 mm). After sACLR, an RL had no significant effect. When looking at ER, an RL significantly increased ER in full extension in the ACL-deficient state in 2 cm (mean difference 0.9°; 95% CI, 0.08°-1.74°) and 3 cm length (mean difference 1.9°; 95% CI, 0.57-3.25). Furthermore, a 3-cm RL significantly increased IR in 0° of flexion in the ACL-deficient state (mean difference 1.9°; 95% CI, 0.2°-3.6°). No effect of ramp lesions on rotation was found after sACLR. CONCLUSION: RLs result in a small increase in ATT, ER, and IR in ACL-deficient knees at early flexion angles, but not after sACLR. CLINICAL RELEVANCE: Small RLs did not change time-zero knee kinematics and may, therefore, be left untreated, especially when the ACL is reconstructed.

Exposure of Hoffa Fractures Is Improved by Posterolateral and Posteromedial Extensile Approaches: A Qualitative and Quantitative Anatomical Study.

BACKGROUND: The current literature lacks recommendations regarding surgical approaches to best visualize and reduce Hoffa fractures. The aims of this study were to (1) define surgical corridors to the posterior portions of the lateral and medial femoral condyles and (2) compare the articular surface areas visible with different approaches. METHODS: Eight fresh-frozen human cadaveric knees (6 male and 2 female donors; mean age, 68.2 ± 10.2 years) underwent dissection simulating 6 surgical approaches to the distal femur. The visible articular surface areas for each approach were marked using an electrocautery device and subsequently analyzed using image-processing software. The labeled areas of each femoral condyle were statistically compared. RESULTS: At 30° of flexion, visualization of the posterior portions of the lateral and medial femoral condyles was not possible by lateral and medial parapatellar approaches, as only the anterior 29.4% ± 2.1% of the lateral femoral condyle and 25.6% ± 2.8% of the medial condyle were exposed. Visualization of the lateral femoral condyle was limited by the posterolateral ligamentous structures, hence a posterolateral approach only exposed its central (13.1% ± 1.3%) and posterior (12.4% ± 1.1%) portions. Posterolateral extension by an osteotomy of the lateral femoral epicondyle significantly improved the exposure to 53.4% ± 2.7% and, when combined with a Gerdy's tubercle osteotomy, to 70.9% ± 4.1% (p < 0.001). For the posteromedial approach, an arthrotomy between the anteromedial retinaculum and the superficial medial collateral ligament, and one between the posterior oblique ligament and the medial gastrocnemius tendon, allowed visualization of the central (13.5% ± 2.2%) and the posterior (14.6% ± 2.3%) portions of the medial femoral condyle, while a medial femoral epicondyle osteotomy significantly improved visualization to 66.1% ± 5.5% (p < 0.001). CONCLUSIONS: Visualization of the posterior portions of the femoral condyles is limited by the specific anatomy of each surgical corridor. Extension by osteotomy of the femoral epicondyles and Gerdy's tubercle significantly improved articular surface exposure of the femoral condyles. CLINICAL RELEVANCE: Knowledge of the surgical approach-specific visualization of the articular surface of the femoral condyles might be helpful to properly reduce small Hoffa fragments.

Treatment of cartilage defects in the patellofemoral joint with matrix-associated autologous chondrocyte implantation effectively improves pain, function, and radiological outcomes after 5-7 years.

INTRODUCTION: The aim of the present study was to evaluate midterm outcomes 5-7 years after matrix-associated autologous chondrocyte implantation (MACI) in the patellofemoral joint. MATERIALS AND METHODS: Twenty-six patients who had undergone MACI using the Novocart® 3D scaffold were prospectively evaluated. Clinical outcomes were determined by measuring the 36-Item Short-Form Health Survey (SF-36) and International Knee Documentation Committee (IKDC) scores and the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) values preoperatively and 3, 6, and 12 months, and a mean of 6 years postoperatively. At the final follow-up, the Magnetic Resonance Observation of Cartilage Repair Tissue (MOCART) score was evaluated. RESULTS: Twenty-two patients with 23 focal cartilage defects (19 patella and four trochlea) were available for the final follow-up. The mean defect size was 4.0 ± 1.9 cm2 (range 2.4-9.4 cm2). All clinical outcome scores improved significantly until 5-7 years after MACI (SF-36 score, 61.2 ± 19.6 to 83.2 ± 11.6; P = 0.001; IKDC score, 47.5 ± 20.6 to 74.7 ± 15.5; P < 0.001; and WOMAC, 29.8 ± 15.7 to 8.2 ± 10.3; P < 0.001). The mean MOCART score was 76.0 ± 11.0 at the final follow-up. Nineteen of the 22 patients (86.4%) were satisfied with the outcomes after 5-7 years and responded that they would undergo the procedure again. CONCLUSION: MACI in the patellofemoral joint demonstrated good midterm clinical results with a significant reduction in pain, improvement in function, and high patient satisfaction. These clinical findings are supported by radiological evidence from MOCART scores. LEVEL OF EVIDENCE: IV-case series.

[Injuries of the medial side of the knee : When and how should they be treated?] / Verletzungen des medialen Bandapparats des Kniegelenks : Wann und wie therapieren?

Different medial structures are responsible for restraining valgus rotation, external rotation, and anteromedial rotation. When injured this can result in various degrees of isolated and combined instabilities. In contrast to earlier speculation, the posterior oblique ligament (POL) is no longer considered to be the main stabilizer of anteromedial rotatory instability (AMRI). Acute proximal medial ruptures are typically managed conservatively with very good clinical results. Conversely, acute distal ruptures usually require a surgical intervention. Chronic instabilities mostly occur in combination with instabilities of the anterior cruciate ligament (ACL). The clinical examination is a particularly important component in these cases to determine the indications for surgery for an additional medial reconstruction. In cases of severe medial and anteromedial instabilities, surgical treatment should be considered. Biomechanically, a combined medial and anteromedial reconstruction appears to be superior to other reconstruction methods; however, there is currently a lack of clinical studies to confirm this biomechanical advantage.

A practical guide to the implementation of AI in orthopaedic research - part 1: opportunities in clinical application and overcoming existing challenges.

Artificial intelligence (AI) has the potential to transform medical research by improving disease diagnosis, clinical decision-making, and outcome prediction. Despite the rapid adoption of AI and machine learning (ML) in other domains and industry, deployment in medical research and clinical practice poses several challenges due to the inherent characteristics and barriers of the healthcare sector. Therefore, researchers aiming to perform AI-intensive studies require a fundamental understanding of the key concepts, biases, and clinical safety concerns associated with the use of AI. Through the analysis of large, multimodal datasets, AI has the potential to revolutionize orthopaedic research, with new insights regarding the optimal diagnosis and management of patients affected musculoskeletal injury and disease. The article is the first in a series introducing fundamental concepts and best practices to guide healthcare professionals and researcher interested in performing AI-intensive orthopaedic research studies. The vast potential of AI in orthopaedics is illustrated through examples involving disease- or injury-specific outcome prediction, medical image analysis, clinical decision support systems and digital twin technology. Furthermore, it is essential to address the role of human involvement in training unbiased, generalizable AI models, their explainability in high-risk clinical settings and the implementation of expert oversight and clinical safety measures for failure. In conclusion, the opportunities and challenges of AI in medicine are presented to ensure the safe and ethical deployment of AI models for orthopaedic research and clinical application. Level of evidence IV.

Additional Plate Fixation of Hinge Fractures After Varisation Distal Femoral Osteotomies Provides Favorable Torsional Stability: A Biomechanical Study.

BACKGROUND: Hinge fractures are considered risk factors for delayed or nonunion of the osteotomy gap in distal femoral osteotomies (DFOs). Limited evidence exists regarding the treatment of hinge fractures after DFO, which could improve stability and thus bone healing. PURPOSE: To (1) examine the effect of hinge fractures on the biomechanical properties of the bone-implant construct, (2) evaluate the biomechanical advantages of an additional fixation of a hinge fracture, and (3) test the biomechanical properties of different types of varisation DFOs. STUDY DESIGN: Controlled laboratory study. METHODS: A total of 32 fresh-frozen human distal femora equally underwent medial closing wedge DFO or lateral opening wedge DFO using a unilateral locking compression plate. The following conditions were serially tested: (1) preserved hinge; (2) hinge fracture along the osteotomy plane; (3) screw fixation of the hinge fracture; and (4) locking T-plate fixation of the hinge fracture. Using a servo-hydraulic materials testing machine, we subjected each construct to 15 cycles of axial compression (400 N; 20 N/s) and internal and external rotational loads (10 N·m; 0.5 N·m/s) to evaluate the stiffness. The axial and torsional hinge displacement was recorded using a 3-dimensional optical measuring system. Repeated-measures 1-way analysis of variance and post hoc Bonferroni correction were used for multiple comparisons. Statistical significance was set at P < .05. RESULTS: Independent from the type of osteotomy, a fractured hinge significantly (P < .001) increased rotational displacement and reduced stiffness of the bone-implant construct, resulting in ≥1.92 mm increased displacement and ≥70% reduced stiffness in each rotational direction, while the axial stiffness remained unchanged. For both procedures, neither a screw nor a plate could restore intact rotational stiffness (P < .01), while only the plate was able to restore intact rotational displacement. However, the plate always performed better compared with the screw, with significantly higher and lower values for stiffness (+38% to +53%; P < .05) and displacement (-55% to -72%; P < .01), respectively, in ≥1 rotational direction. At the same time, the type of osteotomy did not significantly affect axial and torsional stability. CONCLUSION: Hinge fractures after medial closing wedge DFO and lateral opening wedge DFO caused decreased bone-implant construct rotational stiffness and increased fracture-site displacement. In contrast, the axial stiffness remained unchanged in the cadaveric model. CLINICAL RELEVANCE: When considering an osteosynthesis of a hinge fracture in a DFO, an additional plate fixation was the construct with the highest stiffness and least displacement, which could restore intact hinge rotational displacement.

Interference screws manufactured from magnesium display similar primary stability for soft tissue anterior cruciate ligament graft fixation compared to a biocomposite material - a biomechanical study.

PURPOSE: Biodegradable interference screws (IFS) can be manufactured from different biomaterials. Magnesium was previously shown to possess osteoinductive properties, making it a promising material to promote graft-bone healing in anterior cruciate ligament reconstruction (ACLR). The purpose of this study was to compare IFS made from magnesium to a contemporary biocomposite IFS. METHODS: In a porcine model of ACL reconstruction, deep porcine flexor tendons were trimmed to a diameter of 8 mm, sutured in Krackow technique, and fixed with either 8 × 30 mm biocomposite IFS (Bc-IFS) or 8 × 30 mm magnesium IFS (Mg-IFS) in an 8 mm diameter bone tunnel in porcine tibiae. Cyclic loading for 1000 cycles from 0 to 250 N was applied, followed by load to failure testing. Elongation, load to failure and stiffness of the tested constructs was determined. RESULTS: After 1000 cycles at 250 N, elongation was 4.8 mm ± 1.5 in the Bc-IFS group, and 4.9 mm ± 1.5 in the Mg-IFS group. Load to failure was 649.5 N ± 174.3 in the Bc-IFS group, and 683.8 N ± 116.5 in the Mg-IFS group. Stiffness was 125.3 N/mm ± 21.9 in the Bc-IFS group, and 122.5 N/mm ± 20.3 in the Mg-IFS group. No significant differences regarding elongation, load to failure and stiffness between Bc-IFS and Mg-IFS were observed. CONCLUSION: Magnesium IFS show comparable biomechanical primary stability in comparison to biocomposite IFS and may therefore be an alternative to contemporary biodegradable IFS.

S2k Guideline for Tibial Plateau Fractures - Classification, Diagnosis, and Treatment. / S2k-Leitlinie Tibiakopffraktur ­ Klassifikation, Diagnostik und Therapie.

Tibial plateau fractures are mostly complex and surgically demanding joint fractures, which require a comprehensive understanding of the fracture morphology, ligamentous and neurovascular injuries, as well as the diagnostic and therapeutic options for an optimal clinical outcome. Therefore, a standardised and structured approach is required. The success of the treatment of tibial plateau fractures relies on the interdisciplinary cooperation between surgical and conservative physicians in an outpatient and inpatient setting, physical therapists, patients and service providers (health insurance companies, statutory accident insurance, pension providers). On behalf of the German Society for Orthopaedics and Trauma Surgery (DGOU), the German Trauma Society (DGU) and the Society for Arthroscopy and Joint Surgery (AGA), under the leadership of the Fracture Committee of the German Knee Society (DKG), a guideline for tibial plateau fractures was created, which was developed in several voting rounds as part of a Delphi process. Based on the current literature, this guideline is intended to make clear recommendations and outline the most important treatment steps in diagnostics, therapy and follow-up treatment. Additionally, 25 statements were revised by the authors in several survey rounds using the Likert scale in order to reach a final consensus.