Background: Large language models (LLMs) can understand natural language and generate corresponding text, images, and even videos based on prompts, which holds great potential in medical scenarios. Orthopedics is a significant branch of medicine, and orthopedic diseases contribute to a significant socioeconomic burden, which could be alleviated by the application of LLMs. Several pioneers in orthopedics have conducted research on LLMs across various subspecialties to explore their performance in addressing different issues. However, there are currently few reviews and summaries of these studies, and a systematic summary of existing research is absent. Objective: The objective of this review was to comprehensively summarize research findings on the application of LLMs in the field of orthopedics and explore the potential opportunities and challenges. Methods: PubMed, Embase, and Cochrane Library databases were searched from January 1, 2014, to February 22, 2024, with the language limited to English. The terms, which included variants of “large language model,” “generative artificial intelligence,” “ChatGPT,” and “orthopaedics,” were divided into 2 categories: large language model and orthopedics. After completing the search, the study selection process was conducted according to the inclusion and exclusion criteria. The quality of the included studies was assessed using the revised Cochrane risk-of-bias tool for randomized trials and CONSORT-AI (Consolidated Standards of Reporting Trials–Artificial Intelligence) guidance. Data extraction and synthesis were conducted after the quality assessment. Results: A total of 68 studies were selected. The application of LLMs in orthopedics involved the fields of clinical practice, education, research, and management. Of these 68 studies, 47 (69%) focused on clinical practice, 12 (18%) addressed orthopedic education, 8 (12%) were related to scientific research, and 1 (1%) pertained to the field of management. Of the 68 studies, only 8 (12%) recruited patients, and only 1 (1%) was a high-quality randomized controlled trial. ChatGPT was the most commonly mentioned LLM tool. There was considerable heterogeneity in the definition, measurement, and evaluation of the LLMs’ performance across the different studies. For diagnostic tasks alone, the accuracy ranged from 55% to 93%. When performing disease classification tasks, ChatGPT with GPT-4’s accuracy ranged from 2% to 100%. With regard to answering questions in orthopedic examinations, the scores ranged from 45% to 73.6% due to differences in models and test selections. Conclusions: LLMs cannot replace orthopedic professionals in the short term. However, using LLMs as copilots could be a potential approach to effectively enhance work efficiency at present. More high-quality clinical trials are needed in the future, aiming to identify optimal applications of LLMs and advance orthopedics toward higher efficiency and precision.

Background: The role of paraspinal muscle degeneration in the cascade of sagittal imbalance is still unclear. This study aimed to compare paraspinal muscle degeneration in the 4 stages of sagittal imbalance: sagittal balance (SB), compensated sagittal balance (CSB), decompensated sagittal imbalance (DSI), and sagittal imbalance with failure of pelvic compensation (SI-FPC). In addition, it aimed to compare the effects paraspinal muscle endurance and morphology on sagittal spinopelvic alignment in patients with lumbar spinal stenosis. Methods: A cross-sectional study of 219 patients hospitalized with lumbar spinal stenosis was performed. The isometric paraspinal extensor endurance test and evaluation of atrophy and fat infiltration of the paraspinal extensor muscles and psoas major on magnetic resonance imaging were performed at baseline. Spinopelvic parameters including lumbar lordosis, pelvic tilt, sacral slope, pelvic incidence, and the sagittal vertical axis were measured. Results: The patients with lumbar spinal stenosis were divided into 67 with SB, 85 with CSB, 49 with DSI, and 17 with SI-FPC. There were significant differences in paraspinal muscle endurance and morphology among the 4 groups. Furthermore, the SI-FPC group had poorer paraspinal muscle endurance than either the SB or the CSB group. In multiple linear regression analysis, paraspinal muscle endurance and the relative functional cross-sectional area of the paraspinal extensor muscles were the independent predictors of the sagittal vertical axis, and the relative functional cross-sectional area of the psoas major was the independent predictor of relative pelvic version. Conclusions: This study indicated that paraspinal muscle degeneration is not only an initiating factor in pelvic retroversion but also a risk factor for progression from a compensated to a decompensated stage. Specifically, the impairment of muscle endurance in the CSB stage may be the reason why patients experience failure of pelvic compensation. In addition, paraspinal muscle endurance and muscle morphology (relative functional cross-sectional area of the paraspinal extensor muscles and psoas major) had different clinical consequences. Level of Evidence: Prognostic Level II. See Instructions for Authors for a complete description of levels of evidence.

Objective: This study aimed to describe the changes in spinopelvic sagittal alignment in the sitting position after posterior lumbar fusion, and to identify the factors influencing unfused adjacent segment lordosis. Methods: Consecutive patients with lumbar degenerative disease who underwent posterior lumbar interbody fusion between December 2010 and April 2012 were recruited. Lateral full spine radiographs were obtained in the standing, erect sitting, and natural sitting positions. Spinopelvic parameters were measured preoperatively and at the final follow-up. Results: The data of 63 patients were analyzed in this study. The average age was 61.6 ± 11.0 years. When changing from standing to sitting at the final follow-up, all spinopelvic sagittal parameters with the exceptions of pelvic incidence and thoracic kyphosis were significantly altered. The most noticeable changes occurred in the natural sitting position, with the spine slumped toward achieving a C-shaped sagittal profile. Multiple linear regression analysis revealed that when changing to a natural sitting position, age and fusion levels reflected the changes in lumbar lordosis (ΔLL), age and lumbosacral fusion reflected the changes in upper residual lordosis (ΔURL). Conclusion: The most noticeable changes in spinopelvic sagittal alignment occurred in the natural sitting position after lumbar fusion. Age, fusion levels, and lumbosacral fusion significantly influenced the differences in LL and URL between the standing and natural sitting position. These characteristics should be fully considered when planning spinal realignment surgery and investigating the etiological factors of junctional complications.

Study Design.: A retrospective cohort study of consecutive patients. Objective.: To investigate the clinical value of thoracic tilt (TT) in characterizing thoracic compensation and predicting proximal junctional kyphosis (PJK) in degenerative lumbar scoliosis (DLS). Summary of Background Data.: Thoracic compensation has been shown to be associated with the development of PJK, while thoracic shape and morphology in patients with DLS remain understudied. Methods.: Patients with DLS who underwent long-segment fusion were divided into a PJK group and a non-PJK group. Asymptomatic elderly volunteers were recruited as healthy controls. Thoracic parameters were measured in both cohorts, including the TT, T1-L1 pelvic angle (TLPA), T12 slope, thoracic kyphosis (TK, T4-T12), global thoracic kyphosis (GTK, T1-T12), and thoracolumbar kyphosis (TLK, T10-L2). Multivariate logistic regression was used to assess the association between TT and the development of PJK, adjusting for confounders. Multivariate linear regression was used to establish the predictive formula for TT. Results.: A total of 126 patients with DLS were enrolled, of which 37 (29.4%) developed PJK. Compared with 110 healthy controls, DLS patients had significantly greater TT, TLPA, T12 slope, and TLK as well as smaller TK and GTK (all P<0.001). Preoperatively, the PJK group showed significantly greater TT (P=0.013), TLPA (P<0.001), and TLK (P=0.034) than the non-PJK group. No significant differences were found in TK and GTK before surgery. Postoperatively, the PJK group showed significantly greater TT (P<0.001), TLPA (P<0.001), TLK (P<0.001), and proximal junctional angle (P<0.001). Multivariate logistic regression analysis showed that greater postoperative TT was associated with the development of PJK. Multivariate linear regression analysis suggested that the regression formula was postoperative TT=0.675×T12slope+0.412×TK+0.158×TLK−4.808 (R ²=0.643, P<0.001). Conclusions.: The novel sagittal parameter TT can be used for the evaluation of thoracic compensation. Greater preoperative TT might represent a decompensated state of TK. Rebalancing the TT in a sagittal neutral position might help to prevent PJK in patients with DLS.

BACKGROUND Large language models (LLMs) can understand natural language and generate corresponding text, images, and even videos based on prompts, which holds great potential in medical scenarios. Orthopedics is a significant branch of medicine, and orthopedic diseases contribute to a significant socioeconomic burden, which could be alleviated by the application of LLMs. Several pioneers in orthopedics have conducted research on LLMs across various subspecialties to explore their performance in addressing different issues. However, there are currently few reviews and summaries of these studies, and a systematic summary of existing research is absent. OBJECTIVE The objective of this review was to comprehensively summarize research findings on the application of LLMs in the field of orthopedics and explore the potential opportunities and challenges. METHODS PubMed, Embase, and Cochrane Library databases were searched from January 1, 2014, to February 22, 2024, with the language limited to English. The terms, which included variants of “large language model,” “generative artificial intelligence,” “ChatGPT,” and “orthopaedics,” were divided into 2 categories: large language model and orthopedics. After completing the search, the study selection process was conducted according to the inclusion and exclusion criteria. The quality of the included studies was assessed using the revised Cochrane risk-of-bias tool for randomized trials and CONSORT-AI (Consolidated Standards of Reporting Trials–Artificial Intelligence) guidance. Data extraction and synthesis were conducted after the quality assessment. RESULTS A total of 68 studies were selected. The application of LLMs in orthopedics involved the fields of clinical practice, education, research, and management. Of these 68 studies, 47 (69%) focused on clinical practice, 12 (18%) addressed orthopedic education, 8 (12%) were related to scientific research, and 1 (1%) pertained to the field of management. Of the 68 studies, only 8 (12%) recruited patients, and only 1 (1%) was a high-quality randomized controlled trial. ChatGPT was the most commonly mentioned LLM tool. There was considerable heterogeneity in the definition, measurement, and evaluation of the LLMs’ performance across the different studies. For diagnostic tasks alone, the accuracy ranged from 55% to 93%. When performing disease classification tasks, ChatGPT with GPT-4’s accuracy ranged from 2% to 100%. With regard to answering questions in orthopedic examinations, the scores ranged from 45% to 73.6% due to differences in models and test selections. CONCLUSIONS LLMs cannot replace orthopedic professionals in the short term. However, using LLMs as copilots could be a potential approach to effectively enhance work efficiency at present. More high-quality clinical trials are needed in the future, aiming to identify optimal applications of LLMs and advance orthopedics toward higher efficiency and precision.