Using Machine Learning to Predict Surgical Site Infection After Lumbar Spine Surgery.

Objective: The objective of this study was to utilize machine learning techniques to analyze perioperative factors and identify blood glucose levels that can predict the occurrence of surgical site infection following posterior lumbar spinal surgery. Methods: A total of 4019 patients receiving lumbar internal fixation surgery from an institute were enrolled between June 2012 and February 2021. First, the filtered data were randomized into the test and verification groups. Second, in the test group, specific variables were screened using logistic regression analysis, Lasso regression analysis, support vector machine, and random forest. Specific variables obtained using the four methods were intersected, and a dynamic model was constructed. ROC and calibration curves were constructed to assess model performance. Finally, internal model performance was verified in the verification group using ROC and calibration curves. Results: The data from 4019 patients were collected. In total, 1327 eligible cases were selected. By combining logistic regression analysis with three machine learning algorithms, this study identified four predictors associated with SSI, namely Modic changes, sebum thickness, hemoglobin, and glucose. Using this information, a prediction model was developed and visually represented. Then, we constructed ROC and calibration curves using the test group; the area under the ROC curve was 0.988. Further, calibration curve analysis revealed favorable consistency of nomogram-predicted values compared with real measurements. The C-index of our model was 0.986 (95% CI 0.981-0.994). Finally, we used the validation group to validate the model internally; the AUC was 0.987. Calibration curve analysis revealed favorable consistency of nomogram-predicted values compared with real measurements. The C-index was 0.982 (95% CI 0.974-0.999). Conclusion: Logistic regression analysis and machine learning were employed to select four risk factors: Modic changes, sebum thickness, hemoglobin, and glucose. Then, a dynamic prediction model was constructed to help clinicians simplify the monitoring and prevention of SSI.

Identification of osteosarcoma m6A-related prognostic biomarkers using artificial intelligence: RBM15.

Osteosarcoma has the worst prognosis among malignant bone tumors, and effective biomarkers are lacking. Our study aims to explore m6A-related and immune-related biomarkers. Gene expression profiles of osteosarcoma and healthy controls were downloaded from multiple public databases, and their m6A-based gene expression was utilized for tumor typing using bioinformatics. Subsequently, a prognostic model for osteosarcoma was constructed using the least absolute shrinkage and selection operator and multivariate Cox regression analysis, and its immune cell composition was calculated using the CIBERSORTx algorithm. We also performed drug sensitivity analysis for these two genes. Finally, analysis was validated using immunohistochemistry. We also examined the RBM15 gene by qRT-PCR in an in vitro experiment. We collected routine blood data from 1738 patients diagnosed with osteosarcoma and 24,344 non-osteosarcoma patients and used two independent sample t tests to verify the accuracy of the CIBERSORTx analysis for immune cell differences. The analysis based on m6A gene expression tumor typing was most reliable using the two typing methods. The prognostic model based on the two genes constituting RNA-binding motif protein 15 (RBM15) and YTDC1 had a much lower survival rate for patients in the high-risk group than those in the low-risk group (P < 0.05). CIBERSORTx immune cell component analysis demonstrated that RBM15 showed a negative and positive correlation with T cells gamma delta and activated natural killer cells, respectively. Drug sensitivity analysis showed that these two genes showed varying degrees of correlation with multiple drugs. The results of immunohistochemistry revealed that the expression of these two genes was significantly higher in osteosarcoma than in paraneoplastic tissues. The results of qRT-PCR experiments showed that the expression of RBM15 was significantly higher in both osteosarcomas than in the control cell lines. Absolute lymphocyte value, lymphocyte percentage, hematocrit and erythrocyte count were lower in osteosarcoma than in the control group (P < 0.001). RBM15 and YTHDC1 can serve as potential prognostic biomarkers associated with m6A in osteosarcoma.

Application of machine learning in prediction of bone cement leakage during single-level thoracolumbar percutaneous vertebroplasty.

BACKGROUND: In the elderly, osteoporotic vertebral compression fractures (OVCFs) of the thoracolumbar vertebra are common, and percutaneous vertebroplasty (PVP) is a common surgical method after fracture. Machine learning (ML) was used in this study to assist clinicians in preventing bone cement leakage during PVP surgery. METHODS: The clinical data of 374 patients with thoracolumbar OVCFs who underwent single-level PVP at The First People's Hospital of Chenzhou were chosen. It included 150 patients with bone cement leakage and 224 patients without it. We screened the feature variables using four ML methods and used the intersection to generate the prediction model. In addition, predictive models were used in the validation cohort. RESULTS: The ML method was used to select five factors to create a Nomogram diagnostic model. The nomogram model's AUC was 0.646667, and its C value was 0.647. The calibration curves revealed a consistent relationship between nomogram predictions and actual probabilities. In 91 randomized samples, the AUC of this nomogram model was 0.7555116. CONCLUSION: In this study, we invented a prediction model for bone cement leakage in single-segment PVP surgery, which can help doctors in performing better surgery with reduced risk.

Risk Factors in the Prediction of Leg Numbness after Spinal Endoscopic Surgery: Evaluation and Development of a Nomogram.

Purpose: This study aims at constructing a clinical predictive model that predicted the risk factors for leg numbness after spinal endoscopic surgery. Methods: We collected the clinical data of patients, including general information, imaging parameters, and clinical score, from our hospital's electronic database. Based on the postoperative leg numbness visual analog scale (LN-VAS), the clinical data were divided into the leg numbness group (≥25) and the improvement group (<25). All parameters were included in the least absolute shrinkage and selection operator (LASSO) regression analysis, while the parameters with the area under the curve (AUC) greater than 0.7 were selected to construct nomograms. Furthermore, the accuracy and validity of the model were evaluated using the C-index, decision curve analysis (DCA), calibration curve, and receiver operating characteristic curve (ROC). Results: A total of 73 patients' clinical data were included in the training set, where 51 patients were assigned to the improvement group and 22 to the leg numbness group. The nomogram was constructed using four selected parameters, including symptom duration, lumbar spinal stenosis (LSS), pelvic incidence (PI), and preoperative low back pain visual analog scale (LBP-VAS). The nomogram predictions were found to range between 0.01 and 0.99. The values of the C-index, AUC, and internally validated C-index were 0.96, 0.96, and 0.94, respectively. Our result showed that the clinical net benefit of the nomogram ranged between 0.01 and 0.99. Conclusion: Our clinical prediction model demonstrated high predictive ability and clinical validity. Moreover, we found that symptom duration, LSS, PI, and preoperative LBP-VAS were the predictive risk factors for leg numbness after spinal endoscopic surgery.

Machine learning-based clustering in cervical spondylotic myelopathy patients to identify heterogeneous clinical characteristics.

Background: Anterior cervical decompression and fusion can effectively treat cervical spondylotic myelopathy (CSM). Accurately classifying patients with CSM who have undergone anterior cervical decompression and fusion is the premise of precision medicine. In this study, we used machine learning algorithms to classify patients and compare the postoperative efficacy of each classification. Methods: A total of 616 patients with cervical spondylotic myelopathy who underwent anterior cervical decompression and fusion were enrolled. Unsupervised machine learning algorithms (UMLAs) were used to cluster subjects according to similar clinical characteristics. Then, the results of clustering were visualized. The surgical outcomes were used to verify the accuracy of machine learning clustering. Results: We identified two clusters in these patients who had significantly different baseline clinical characteristics, preoperative complications, the severity of neurological symptoms, and the range of decompression required for surgery. UMLA divided the CSM patients into two clusters according to the severity of their illness. The repose to surgical treatment between the clusters was significantly different. Conclusions: Our results showed that UMLA could be used to rationally classify a heterogeneous cohort of CSM patients effectively, and thus, it might be used as the basis for a data-driven platform for identifying the cluster of patients who can respond to a particular treatment method.

Predicting Surgical Site Infection Risk after Spinal Tuberculosis Surgery: Development and Validation of a Nomogram.

Background: The purpose of this study was to predict the surgical site infection risk after spinal tuberculosis surgery based on a nomogram. Patients and Methods: We collected the clinical data of patients who underwent spinal tuberculosis surgery in our hospital and included all the data in the least absolute shrinkage and selection operator (LASSO) regression analysis. Next, the selected parameters were analyzed using logistic regression. The logistic regression analysis and receiver operating characteristic (ROC) curve analysis were further used to obtain statistically significant parameters. These parameters were then used to construct a nomogram. The C-index, ROC curve, and decision curve analysis (DCA) were used to assess the predictive ability and accuracy of the nomogram, whereas internal verification was used to calculate the C-index by bootstrapping with 1,000 resamples. Results: A total of 394 patients with spinal tuberculosis surgery were included in the study, of whom 76 patients had surgical site infections whereas 318 patients did not. The predicted risk of surgical site infection in the nomogram ranged between 0.01 and 0.98. Both the value of the C-index of the nomogram (95% confidence interval [CI], 0.62-0.76) and the area under the curve (AUC) were found to be 0.69. The net benefit of the model ranged between 0.01 and 0.99. In contrast, the C-index calculated by the internal verification method of the nomogram was found to be 0.68. Conclusions: The risk factors predicting surgical site infection after spinal tuberculosis surgery included albumin, lesion segment, operation time, and incision length.

Multiple Spinal Tuberculosis with Severe Kyphosis: A Case Report.

Background: The purpose of this study was to analyze the clinical efficacy of a patient with multiple tuberculosis of the spine combined with severe kyphosis. Case Summary: A 56-year-old male patient presented with low back pain with numbness and fatigue in both lower extremities for 5 months. Chest and back showed intermittent acid pain. The patient had not a history of constitutional symptoms. Preoperative X-ray and CT examination revealed multiple vertebral segmental bone destruction, multiple abscess calcification, and severe kyphosis. Preoperative MRI examination showed that the tuberculous abscess broke through the spinal canal and compressed the spinal cord and nerve roots. The patient underwent posterior lumbar abscess debridement, expanded decompression of the spinal canal, and nerve lysis in our hospital. The operation time was 70 min, and the intraoperative blood loss was 200 ml. The postoperative drainage volume was 250 ml. The patient was hospitalized for a total of 13 days, and the patient's vital signs were stable before and after surgery. The patient was satisfied with the treatment. Conclusion: For the patient with multiple spinal tuberculosis complicated with severe kyphosis and multiple calcified abscesses in this study, we considered performing abscess debridement to relieve the symptoms of back pain and achieved good clinical efficacy.

Novel prognostic biomarkers, METTL14 and YTHDF2, associated with RNA methylation in Ewing's sarcoma.

Ewing's sarcoma has a poor prognosis and high metastasis rate; thus, it is critical to explore prognostic biomarkers of m6A-related genes. Two datasets were downloaded from the Gene Expression Omnibus database, m6A-related genes were extracted, and prognostic models were constructed using the least absolute shrinkage and selection operator and multivariate COX regression analyses. Immune cell composition and drug sensitivity analyses were performed, and our analysis was validated using laboratory methods of immunohistochemical specific staining and qRT-PCR. Ewing's sarcoma prognostic model demonstrated that the survival rate of cases in the high-risk group was much lower than that of the low-risk group. Naïve B cells, macrophages M0, macrophages M1, and resting mast cells are closely associated with Ewing's sarcoma. METTL14 and YTHDF2 are strongly associated with multiple drug sensitivity. Immunohistochemical specific staining revealed higher expression of both METTL14 and YTHDF2 in Ewing's sarcoma than in the paraneoplastic tissues. The results of qRT-PCR showed that METTL14 expression was significantly higher in both ES cell lines than in the control cell line. The prognostic model constructed using m6A-related genes METTL14 and TYHDF2, can be a potential prognostic biomarker for Ewing's sarcoma, with the survival rate of cases in the high-risk group being much lower than that of the low-risk group.

Comparison of Clinical Data Between Patients With Complications and Without Complications After Spinal Tuberculosis Surgery: A Propensity Score Matching Analysis.

Purpose: This study used a propensity score matching (PSM) analysis to explore the risk factors of post-operative complications and compared the differences in clinical data between them following spinal tuberculosis surgery. Methods: The clinical data of patients with spinal tuberculosis were collected in our hospital from June 2012 to June 2021, including general information, laboratory results, surgical information, and hospitalization costs. The data were divided into two groups: complication and without complication groups. The baseline data of the two groups were obtained using the PSM analysis. Univariate and multivariate logistic analyses were used to analyze the differences between the two groups. Results: A total of 292 patients were included in the PSM analysis: 146 patients with complications and 146 patients without complications. The operation time, incision length, hospital stay, and albumin quantity in the complications group were 162 ± 74.1, 11.2 ± 4.76, 14.7 ± 9.34, and 1.71 ± 2.82, respectively, and those in the without complication group were 138 ± 60.5, 10.2 ± 3.56, 11.7 ± 7.44, and 0.740 ± 2.44, respectively. The laboratory costs, examination costs, guardianship costs, oxygen costs, and total costs in the complications group were higher than those in the without complication group. A significant difference was observed in the albumin quantity by logistic regression analysis (P < 0.05). Conclusion: Several costs in the complication group were higher than in the without complication group. The albumin quantity may be an independent factor to predict post-operative complications of spinal tuberculosis by logistic regression analysis.

Blood transfusion risk prediction in spinal tuberculosis surgery: development and assessment of a novel predictive nomogram.

OBJECTIVE: The present study attempted to predict blood transfusion risk in spinal tuberculosis surgery by using a novel predictive nomogram. METHODS: The study was conducted on the clinical data of 495 patients (167 patients in the transfusion group and 328 patients in the non-transfusion group) who underwent spinal tuberculosis surgery in our hospital from June 2012 to June 2021. The least absolute shrinkage and selection operator (LASSO) and multivariable logistic regression analyses were used to screen out statistically significant parameters, which were included to establish a novel predictive nomogram model. The receiver operating characteristic (ROC) curve, calibration curves, C-index, and decision curve analysis (DCA) were used to evaluate the model. Finally, the nomogram was further assessed through internal validation. RESULTS: The C-index of the nomogram was 0.787 (95% confidence interval: 74.6%-.82.8%). The C-value calculated by internal validation was 0.763. The area under the curve (AUC) of the predictive nomogram was 0.785, and the DCA was 0.01-0.79. CONCLUSION: A nomogram with high accuracy, clinical validity, and reliability was established to predict blood transfusion risk in spinal tuberculosis surgery. Surgeons must prepare preoperative surgical strategies and ensure adequate availability of blood before surgery.

Pulmonary embolism following the third thoracic tuberculosis surgery: A case report and literature review.

BACKGROUND: The study aimed to analyze the clinical effects of pulmonary embolism succeeding a third surgery conducted for multiple recurrences in thoracic tuberculosis (TB). CASE REPORT: A 74-year-old female patient developed thoracic tuberculosis and was subsequently treated in our hospital in March 2019, October 2020, and February 2021. The third surgical intervention included anterolateral thoracic lesion resection, internal fixation, posterior spinal tuberculous sinus resection, and debridement with suture. The operative time was 172 min resulting in a substantial intraoperative blood loss (2321 ml). Postoperative re-examination of chest CTPA indicated a strip filling defect and pulmonary embolism in the external branch of the right middle lobe of the lung. After completing the active treatment, the D-dimer quantification, WBC, CRP, and ESR values were 1261 ng/ml, 7.71 × 109 /L, 74.66 mg/L, and 63 mm, respectively. Chest CTPA re-examination after the treatment showed no signs of pulmonary embolism. CONCLUSION: Patients with a long-term history of multiple operations, high BMI, cerebral infarction, diabetes, and older age group were more likely to develop pulmonary embolism after spinal tuberculosis surgery. Thus, the possibility of postoperative pulmonary embolism should be thoroughly analyzed before any subsequent surgical treatment in patients with recurrent spinal tuberculosis.

Feasibility of 3.5 mm C2 pedicle screws in children: A computerized tomography analysis.

There have been no studies with large sample sizes on growth of the pedicle of C2 in children. In the present study we measured the pedicle of C2 through computed tomography (CT) imaging in children aged less than 14 years and evaluated the suitability of the 3.5-mm screw for the pedicle in such children. The study was conducted on CT morphometric images of 420 children in our hospital between June 2018 and June 2020. The width (D1), length (D2), height (D3), inclination angle (α), and tail angle (ß) of the C2 pedicle were measured. One-way analysis of variance and Student's t test were used for statistical analyses. The least-square method was used to analyze the curve fitting the trend of anatomical change in the pedicle. The largest degree of goodness of fit determined the best-fitting curve. The size of the pedicle of C2 increased with age. The median ranges of D1, D2, D3, α, and ß were 3.312-5.431 mm, 11.732-23.645 mm, 3.597-8.038 mm, 32.583°-36.640°, and 24.867°-31.567°, respectively. The curves fitting the trends of D1 and D3 were power functions, whereas D2 was fitted by a logarithmic curve. However, no curve fitted α or ß. A 3.5-mm screw can be placed in the pedicle of C2 in children aged more than 1 year. The growth and development trend of this pedicle can provide an anatomical reference for deciding on posterior cervical surgery and for selecting and designing pedicle screws for children.

Development and Validation of a Nomogram for Predicting Albumin Transfusion After Spinal Tuberculosis Surgery: Based on Propensity Score Matching Analysis.

BACKGROUND: There have been few literature reports on the use of perioperative parameters to predict the risk of albumin transfusion after spinal tuberculosis surgery based on the application of nomogram and propensity score matching (PSM) analysis. OBJECTIVE: The purpose was to predict the risk of albumin transfusion after spinal tuberculosis surgery based on a combination of PSM and nomogram. METHODS: The clinical data of the patients were collected in our hospital, including preoperative clinical data, preoperative laboratory tests, and postoperative clinical data. All data were divided into 2 groups, including the albumin transfusion group and the non-albumin transfusion group. The PSM analysis was used to adjust the baseline data of the 2 groups. The nomogram was further constructed. The practicability and predictive ability of the model were evaluated. RESULTS: A total of 494 cases were collected in this article; 102 pairs by PSM analysis were used to construct the nomogram. There were statistical differences in surgical approach, aspartate aminotransferase/alanine aminotransferase levels, drainage, and kyphosis by logistic analysis, and these parameters were included in the construction of the nomogram. The C-index of the prediction model was 0.734. The area under the curve was 0.73 and the net benefit was between 0.13 and 0.99. The calculated C-index was 0.71 by the internal verification method. CONCLUSIONS: The PSM analysis had a good matching effect and the nomogram had a good predictive ability. Surgical approach, aspartate aminotransferase/alanine aminotransferase levels, drainage, and kyphosis might be predictors of albumin transfusion after spinal tuberculosis surgery.

Can the Risk of Postoperative Cerebrospinal Fluid Leakage Be Predicted for Patients Undergoing Cervical Spine Surgery? Development and Evaluation of a New Predictive Nomogram.

OBJECTIVE: Previous studies have retrospectively analyzed the likely causes of cerebrospinal fluid leakage (CSFL) during cervical spine surgery and the management of CSFL after its occurrence. In the present study, we aimed to develop and validate a nomogram for the risk of CSFL in Chinese patients who had undergone cervical decompression and internal fixation (CDIF) surgery. METHODS: We performed a retrospective analysis of patients who had undergone CDIF surgery. Of the 1286 included patients, 54 were in the CSFL group and 1232 were in the normal group. The patients were randomly divided into training and validation tests. The risk assessment for CSFL included 21 characteristics. The feature selection for the CSFL model was optimized using the least absolute shrinkage and selection operator regression model in the training test. Multivariate logistic regression analysis was performed to construct the model according to the selected characteristics. The clinical usefulness of the predictive model was assessed using the C-index, calibration curve, and decision curve analysis with identification and calibration. RESULTS: The risk prediction nomogram included the diagnosis, revision surgery, ossification of the posterior longitudinal ligament, cervical instability, and a history of malignancy in the training test. The model demonstrated high predictive power, with a C-index of 0.914 (95% confidence interval, 0.876-0.951) and an area under the curve of 0.914. The results of the decision curve analysis demonstrated the clinical usefulness of the CSFL risk nomogram when the probability threshold for CSFL was 1%-62%. CONCLUSIONS: Our proposed nomogram for CSFL risk includes the diagnosis, revision surgery, ossification of the posterior longitudinal ligament, cervical instability, and a history of malignancy. The nomogram can be used to evaluate the risk of CSFL for patients undergoing CDIF surgery.

Difference between the blood samples of patients with bone and joint tuberculosis and patients with tuberculosis studied using machine learning.

Background: Tuberculosis (TB) is a chronic infectious disease. Bone and joint TB is a common type of extrapulmonary TB and often occurs secondary to TB infection. In this study, we aimed to find the difference in the blood examination results of patients with bone and joint TB and patients with TB by using machine learning (ML) and establish a diagnostic model to help clinicians better diagnose the disease and allow patients to receive timely treatment. Methods: A total of 1,667 patients were finally enrolled in the study. Patients were randomly assigned to the training and validation cohorts. The training cohort included 1,268 patients: 158 patients with bone and joint TB and 1,110 patients with TB. The validation cohort included 399 patients: 48 patients with bone and joint TB and 351 patients with TB. We used three ML methods, namely logistic regression, LASSO regression, and random forest, to screen the differential variables, obtained the most representative variables by intersection to construct the prediction model, and verified the performance of the proposed prediction model in the validation group. Results: The results revealed a great difference in the blood examination results of patients with bone and joint TB and those with TB. Infectious markers such as hs-CRP, ESR, WBC, and NEUT were increased in patients with bone and joint TB. Patients with bone and joint TB were found to have higher liver function burden and poorer nutritional status. The factors screened using ML were PDW, LYM, AST/ALT, BUN, and Na, and the nomogram diagnostic model was constructed using these five factors. In the training cohort, the area under the curve (AUC) value of the model was 0.71182, and the C value was 0.712. In the validation cohort, the AUC value of the model was 0.6435779, and the C value was 0.644. Conclusion: We used ML methods to screen out the blood-specific factors-PDW, LYM, AST/ALT, BUN, and Na+-of bone and joint TB and constructed a diagnostic model to help clinicians better diagnose the disease in the future.

A seven-gene signature and the C-C motif chemokine receptor family genes are the sarcoma-related immune genes.

Cells of the tumor microenvironment exert a vital influence on sarcoma prognosis. This study aimed to analyze and identify differentially expressed genes (DEGs) related to immunity and their significance as immune biomarkers for the accurate prediction of overall survival of patients with sarcoma. The Cancer Genome Atlas was adopted for obtaining sarcoma gene microarray and corresponding clinical information. ESTIMATE algorithm was used to calculate tumor immune microenvironment indices. Immune-associated DEGs were identified using the limma packages and were further analyzed using the ClusterProfiler package and STRING website. Based on the results of these analyses, we constructed a prognostic model. Furthermore, we assessed the prognosis prediction model through functional evaluation and analysis of GSE17674. The functional analysis revealed that the upregulated immune DEGs were related to immune-related aspects. Chemokine ligands/receptors and immune-related genes were found to be vital for sarcoma formation and progression. We established a prognostic signature of seven genes, which indicated that high-risk cases exhibit poor prognostic outcome. The prognostic signature constructed in this study can accurately predict the overall prognosis in patients with sarcoma. Moreover, the novel immune gene expression analysis may provide clinical guidance for predicting prognosis in patients with sarcoma.

Glycolysis- and immune-related novel prognostic biomarkers of Ewing's sarcoma: glucuronic acid epimerase and triosephosphate isomerase 1.

INTRODUCTION: Owing to the poor prognosis of Ewing's sarcoma, reliable prognostic biomarkers are highly warranted for clinical diagnosis of the disease. MATERIALS AND METHODS: A combination of the weighted correlation network analysis and differentially expression analysis was used for initial screening; glycolysis-related genes were extracted and subjected to univariate Cox, LASSO regression, and multivariate Cox analyses to construct prognostic models. The immune cell composition of each sample was analysed using CIBERSORT software. Immunohistochemical analysis was performed for assessing the differential expression of modelled genes in Ewing's sarcoma and paraneoplastic tissues. RESULTS: A logistic regression model constructed for the prognosis of Ewing's sarcoma exhibited that the patient survival rate in the high-risk group is much lower than in the low-risk group. CIBERSORT analysis exhibited a strong correlation of Ewing's sarcoma with naïve B cells, CD8+ T cells, activated NK cells, and M0 macrophages (P < 0.05). Immunohistochemical analysis confirmed the study findings. CONCLUSIONS: GLCE and TPI1 can be used as prognostic biomarkers to predict the prognosis of Ewing's sarcoma, and a close association of Ewing's sarcoma with naïve B cells, CD8+ T cells, activated NK cells, and M0 macrophages provides a novel approach to the disease immunotherapy.

TRIM68, PIKFYVE, and DYNLL2: The Possible Novel Autophagy- and Immunity-Associated Gene Biomarkers for Osteosarcoma Prognosis.

INTRODUCTION: Osteosarcoma is among the most common orthopedic neoplasms, and currently, there are no adequate biomarkers to predict its prognosis. Therefore, the present study was aimed to identify the prognostic biomarkers for autophagy-and immune-related osteosarcoma using bioinformatics tools for guiding the clinical diagnosis and treatment of this disease. MATERIALS AND METHODS: The gene expression and clinical information data were downloaded from the Public database. The genes associated with autophagy were extracted, followed by the development of a logistic regression model for predicting the prognosis of osteosarcoma using univariate and multivariate COX regression analysis and LASSO regression analysis. The accuracy of the constructed model was verified through the ROC curves, calibration plots, and Nomogram plots. Next, immune cell typing was performed using CIBERSORT to analyze the expression of the immune cells in each sample. For the results obtained from the analysis, we used qRT-PCR validation in two strains of human osteosarcoma cells. RESULTS: The screening process identified a total of three genes that fulfilled all the screening criteria. The survival curves of the constructed prognostic model revealed that patients with the high risk presented significantly lower survival than the patients with low risk. Finally, the immune cell component analysis revealed that all three genes were significantly associated with the immune cells. The expressions of TRIM68, PIKFYVE, and DYNLL2 were higher in the osteosarcoma cells compared to the control cells. Finally, we used human pathological tissue sections to validate the expression of the genes modeled in osteosarcoma and paracancerous tissue. CONCLUSION: The TRIM68, PIKFYVE, and DYNLL2 genes can be used as biomarkers for predicting the prognosis of osteosarcoma.