Development and Internal Validation of Supervised Machine Learning Algorithm for Predicting the Risk of Recollapse Following Minimally Invasive Kyphoplasty in Osteoporotic Vertebral Compression Fractures

Risk factors for secondary vertebral compression fracture after percutaneous vertebral augmentation: a single-centre retrospective study

OBJECTIVE

To determine the incidence of secondary vertebral compression fracture (SVCF) after percutaneous vertebral augmentation (PVA) and its correlative risk factors, and to provide theoretical evidence for clinical practice.

METHODS

A retrospective analysis of 288 cases of PVA completed in our hospital from June 2020 to June 2023 was performed, and the patients were divided into the non-secondary vertebral compression fracture group (N-SVCF group) and the secondary vertebral compression fracture group (SVCF group) according to whether SVCF occurred during the postoperative follow-up review. Gender, age, body mass index (BMI), T value of bone mineral density (BMD-T), underlying diseases (hypertension, diabetes mellitus, coronary heart disease, chronic obstructive pulmonary disease), intravertebral vacuum cleft (IVC), amount of bone cement injected, classification of cement diffusion, anterior vertebral recovery ratio, local Cobb angle correction rate, leakage of bone cement into the intervertebral space, and fat infiltration rate (FIR) of paraspinal muscles were collected from the patients. The incidence and risk factors of SVCF after PVA were evaluated using univariate and multivariate logistic regression analysis, and the predictive value of the independent risk factors was evaluated using receiver operating characteristic curve (ROC) to determine the cut-off points at which they were meaningful for the development of SVCF.

RESULTS

In our study, the incidence of SVCF was 14.60% (42/288) in 288 patients who underwent PVA. Univariate analysis showed that age, BMI, fat infiltration rate of paraspinal muscles, cement leakage into the intervertebral space, unilateral/bilateral pedicle puncture approach and presence of IVC were statistically different between N-SVCF and SVCF (P < 0.05). Multifactorial regression analysis and ROC regression analysis revealed that the fat infiltration rate of the psoas major and erector spinae muscles, cement leakage into the intervertebral space, and IVC (P < 0.05) were risk factors for the incident of SVCF after PVA (P < 0.05). Psoas major (FIR) more than 5.490% and erector spinae (FIR) more than 52.413% had a high possibility of the occurrence of SVCF after PVA.

CONCLUSION

In this study, logistic regression combined with ROC curve analysis indicated that FIR of psoas major and erector spinae, cement leakage in the intervertebral space, and IVC were risk factors for the occurrence of SVCF after PVA. Psoas major (FIR) more than 5.490% and erector spinae (FIR) more than 52.413% had a high possibility of the occurrence of SVCF after PVA.

Risk Factors for Bone Cement Displacement After Percutaneous Kyphoplasty in Osteoporotic Vertebral Fractures: A Retrospective Analysis

BACKGROUND Bone cement displacement (BCD), which has received increasing attention from scholars, is a serious complication following vertebroplasty in patients with osteoporotic vertebral fractures (OVFs), and percutaneous kyphoplasty (PKP) might promote its occurrence. However, few studies have systematically explored the risk factors of BCD after PKP. This research aimed to study the risk factors for BCD following PKP. MATERIAL AND METHODS The clinical data of patients with OVFs treated with PKP from June 2016 to August 2022 in our department were retrospectively reviewed. Patients were categorized into the bone cement displacement group and the bone cement non-displacement group. Data on the subjects and their radiologic images were gathered for univariate analysis and binary logistic regression analysis. The receiver operating characteristic (ROC) curve and confusion matrix were utilized to assess the discrimination ability. RESULTS We included 181 patients, of which 12 had BCD after PKP. Binary logistic regression analysis revealed that independent risk factors associated with BCD after PKP were: high BMI, high restoration rate of the Cobb angle, high distance between the bone cement and the vertebral endplates, and presence of bone cement leakage. The ROC curve and confusion matrix indicates that logistic regression exhibited a strong predictive value for BCD. CONCLUSIONS Patients with a high BMI, a high restoration rate of the Cobb angle, a high distance between the bone cement and the vertebral endplates, and bone cement leakage have an increased risk of BCD after PKP.

Radiomics Based on Multimodal magnetic resonance imaging for the Differential Diagnosis of Benign and Malignant Vertebral Compression Fractures

OBJECTIVES

Recent studies have indicated that radiomics may have excellent performance and clinical application prospects in the differential diagnosis of benign and malignant vertebral compression fractures (VCFs). However, multimodal magnetic resonance imaging (MRI)-based radiomics model is rarely used in the differential diagnosis of benign and malignant VCFs, and is limited to lumbar. Herein, this study intends to develop and validate MRI radiomics models for differential diagnoses of benign and malignant VCFs in patients.

METHODS

This cross-sectional study involved 151 adult patients diagnosed with VCF in The First Affiliated Hospital of Soochow University in 2016-2021. The study was conducted in three steps: (i) the original MRI images were segmented, and the region of interest (ROI) was marked out; (ii) among the extracted features, those features with Pearson's correlation coefficient lower than 0.9 and the top 15 with the highest variance and Lasso regression coefficient less than and more than 0 were selected; (iii) MRI images and combined data were studied by logistic regression, decision tree, random forest and extreme gradient boosting (XGBoost) models in training set and the test set (ratio of 8:2), respectively; and the models were further verified and evaluated for the differential diagnosis performance. The evaluated indexes included area under receiver (AUC) of operating characteristic curve, accuracy, sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), and 95% confidence intervals (CIs). The AUCs were used to assess the predictive performance of different machine learning modes for benign and malignant VCFs.

RESULTS

A total of 1144 radiomics features, and 14 clinical features were extracted. Finally, 12 radiomics features were included in the radiomics model, and 12 radiomics features with 14 clinical features were included in the combined model. In the radiomics model, the differential diagnosis performance in the logistic regression model with the AUC of 0.905 ± 0.026, accuracy of 0.817 ± 0.057, sensitivity of 0.831 ± 0.065, and negative predictive value of 0.813 ± 0.042, was superior to the other three. In the combined model, XGBoost model had the superior differential diagnosis performance with specificity (0.979 ± 0.026) and positive predictive value (0.971 ± 0.035).

CONCLUSION

The multimodal MRI-based radiomics model performed well in the differential diagnosis of benign and malignant VCFs, which may provide a tool for clinicians to differentially diagnose VCFs.

Development and validation of a prediction model for vertebral recompression and adjacent vertebral fracture after kyphoplasty in geriatric patients

PURPOSE

Short-term efficacy of percutaneous kyphoplasty (PKP) for treating osteoporotic vertebral compression fracture (OVCF) in elderly patients is good, but long-term complications such as vertebral recompression (VRC) and adjacent vertebral fracture (AVF) may arise. Identifying risk factors in patients with poor prognoses, we developed a nomogram model to mitigate these potential complications.

METHODS

Patients with OVCFs who underwent PKP had their medical data retrospectively evaluated. Analysis of the sample included their pre- and postoperative conditions. Stepwise logistic regression analyses were conducted to identify independent risk factors for postoperative complications. For forecasting the likelihood of postoperative comorbidities, we offered a nomogram. The prognostic performance was assessed using the receiver operating characteristic curve (ROC), calibration curve, and decision curve analyses (DCA). Internal model validation using the Bootstrap method.

RESULTS

A total of 235 patients were included in this study. Among them, 147 patients were utilized to develop nomograms and for internal validation, while the remaining 88 patients from a different time period were designated as the external validation cohort. The results of stepwise logistic regression analysis showed that thoracolumbar (TL) fracture, posterior wall of vertebral fracture, vertebral compression > 30%, and lack of continuous anti-osteoporosis therapy after surgery as independent risks associated with poor prognosis. The nomogram exhibited outstanding predictive accuracy and clinical utility.

CONCLUSIONS

This study identified four independent predictors of poor prognosis following PKP and devised a straightforward yet efficient predictive model. This model offers valuable insights for guiding clinical decision-making in the management of elderly patients with OVCFs.

Application of machine learning algorithms to predict 30-day hospital readmission following cement augmentation for osteoporotic vertebral compression fractures

Objective

Osteoporosis is a common skeletal disease that greatly increases the risk of pathologic fractures and accounts for approximately 700,000 vertebral compression fractures (VCFs) annually in the United States. Cement augmentation procedures such as balloon kyphoplasty (KP) and percutaneous vertebroplasty (VP) have demonstrated efficacy in the treatment of VCFs, however, some studies report rates of readmission as high as 10.8% following such procedures. The purpose of this study was to employ Machine Learning (ML) algorithms to predict 30-day hospital readmission following cement augmentation procedures for the treatment of VCFs using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database.

Methods

ACS-NSQIP was queried to identify patients undergoing either KP or VP from 2011 to 2014. Three ML algorithms were constructed and tasked with predicting post-operative readmissions within this cohort of patients. Results: Postoperative pneumonia, ASA Class 2 designation, age, partially-dependent functional status, and a history of smoking were independently identified as highly predictive of readmission by all ML algorithms. Among these variables postoperative pneumonia (p < 0.01), ASA Class 2 designation (p < 0.01), age (p = 0.002), and partially-dependent functional status (p < 0.01) were found to be statistically significant. Predictions were generated with an average AUC value of 0.757 and an average accuracy of 80.5%.

Conclusions

Postoperative pneumonia, ASA Class 2 designation, partially-dependent functional status, and age are perioperative variables associated with 30-day readmission following cement augmentation procedures. The use of ML allows for quantification of the relative contributions of these variables toward producing readmission.

Limitations in Evaluating Machine Learning Models for Imbalanced Binary Outcome Classification in Spine Surgery: A Systematic Review

Clinical prediction models for spine surgery applications are on the rise, with an increasing reliance on machine learning (ML) and deep learning (DL). Many of the predicted outcomes are uncommon; therefore, to ensure the models' effectiveness in clinical practice it is crucial to properly evaluate them. This systematic review aims to identify and evaluate current research-based ML and DL models applied for spine surgery, specifically those predicting binary outcomes with a focus on their evaluation metrics. Overall, 60 papers were included, and the findings were reported according to the PRISMA guidelines. A total of 13 papers focused on lengths of stay (LOS), 12 on readmissions, 12 on non-home discharge, 6 on mortality, and 5 on reoperations. The target outcomes exhibited data imbalances ranging from 0.44% to 42.4%. A total of 59 papers reported the model's area under the receiver operating characteristic (AUROC), 28 mentioned accuracies, 33 provided sensitivity, 29 discussed specificity, 28 addressed positive predictive value (PPV), 24 included the negative predictive value (NPV), 25 indicated the Brier score with 10 providing a null model Brier, and 8 detailed the F1 score. Additionally, data visualization varied among the included papers. This review discusses the use of appropriate evaluation schemes in ML and identifies several common errors and potential bias sources in the literature. Embracing these recommendations as the field advances may facilitate the integration of reliable and effective ML models in clinical settings.

Lower ratio of adjacent to injured vertebral bone quality scores can predict augmented vertebrae recompression following percutaneous kyphoplasty for osteoporotic vertebral fractures with intravertebral clefts

BACKGROUND

Despite the favorable clinical outcome of percutaneous kyphoplasty (PKP) in symptomatic osteoporotic vertebral fractures (OVFs) patients with intravertebral clefts (IVCs), previous studies have demonstrated a high incidence of augmented vertebrae recompression (AVR). We aim to evaluate the usefulness of the adjacent and injured vertebral bone quality scores (VBQS) based on T1-weighted MRI images in AVR after PKP for OVFs with IVCs.

METHODS

Patients who underwent PKP for single OVFs with IVCs between January 2014 and September 2020 were reviewed and met the inclusion criteria. The follow-up period was at least 2 years. Relevant data affecting AVR were collected. Pearson and Spearman correlation coefficients were used to calculate the correlation between the injured and adjacent VBQS and BMD T-score. We determined independent risk factors and critical values using binary logistic regression analysis and receiver operating characteristic curves (ROC).

RESULTS

A total of 165 patients were included. Recompression group was found in 42 (25.5%) patients. The independent risk factors for AVR were lumbar BMD T-score (OR = 2.53, p = 0.003), the adjacent VBQS (OR = 0.79, p = 0.016), the injured VBQS (OR = 1.27, p = 0.048), the ratio of adjacent to injured VBQS (OR = 0.32, p < 0.001), and cement distribution pattern. Among these independent significant risk factors, the prediction accuracy of the ratio of adjacent to injured VBQS was the highest (Cutoff = 1.41, AUC = 0.753). Additionally, adjacent and injured VBQS were negatively correlated with lumbar BMD T-scores.

CONCLUSION

For the patients after PKP treatment for OVFs with IVCs, the ratio of adjacent to injured VBQS had the best prediction accuracy in predicting recompression and when the ratio of adjacent to injured VBQS was <1.41, the augmented vertebrae were more likely to have recompression in the future.

Prediction of subsequent vertebral compression fractures after thoracolumbar kyphoplasty: a multicenter retrospective analysis

OBJECTIVE

Second fractures at the cemented vertebrae (SFCV) are often seen after percutaneous kyphoplasty, especially at the thoracolumbar junction. Our study aimed to develop and validate a preoperative clinical prediction model for predicting SFCV.

METHODS

A cohort of 224 patients with single-level thoracolumbar osteoporotic vertebral fractures (T11-L2) from 3 medical centers was analyzed between January 2017 and June 2020 to derive a preoperative clinical prediction model for SFCV. Backward-stepwise selection was used to select preoperative predictors. We assigned a score to each selected variable and developed the SFCV scoring system. Internal validation and calibration were conducted for the SFCV score.

RESULTS

Among the 224 patients included, 58 had postoperative SFCV (25.9%). The following preoperative measures on multivariable analysis were summarized in the 5-point SFCV score: bone mineral density (≤-3.05), serum 25-hydroxy vitamin D3 (≤17.55 ng/mL), standardized signal intensity of fractured vertebra on T1-weighted images (≤59.52%), C7-S1 sagittal vertical axis (≥3.25 cm), and intravertebral cleft. Internal validation showed a corrected area under the curve of 0.794. A cutoff of ≤1 point was chosen to classify a low risk of SFCV, for which only 6 of 100 patients (6%) had SFCV. A cutoff of ≥4 points was chosen to classify a high risk of SFCV, for which 28 of 41 (68.3%) had SFCV.

CONCLUSION

The SFCV score was found to be a simple preoperative method for identification of patients at low and high risk of postoperative SFCV. This model could be applied to individual patients and aid in the decision-making before percutaneous kyphoplasty.