Hybrid deep learning model for risk prediction of fracture in patients with diabetes and osteoporosis

Fully Automatic Deep Learning Model for Spine Refracture in Patients with OVCF: A Multi-Center Study

BACKGROUND

The reaserch of artificial intelligence (AI) model for predicting spinal refracture is limited to bone mineral density, X-ray and some conventional laboratory indicators, which has its own limitations. Besides, it lacks specific indicators related to osteoporosis and imaging factors that can better reflect bone quality, such as computed tomography (CT).

OBJECTIVE

To construct a novel predicting model based on bone turn-over markers and CT to identify patients who were more inclined to suffer spine refracture.

METHODS

CT images and clinical information of 383 patients (training set = 240 cases of osteoporotic vertebral compression fractures (OVCF), validation set = 63, test set = 80) were retrospectively collected from January 2015 to October 2022 at three medical centers. The U-net model was adopted to automatically segment ROI. Three-dimensional (3D) cropping of all spine regions was used to achieve the final ROI regions including 3D_Full and 3D_RoiOnly. We used the Densenet 121-3D model to model the cropped region and simultaneously build a T-NIPT prediction model. Diagnostics of deep learning models were assessed by constructing ROC curves. We generated calibration curves to assess the calibration performance. Additionally, decision curve analysis (DCA) was used to assess the clinical utility of the predictive models.

RESULTS

The performance of the test model is comparable to its performance on the training set (dice coefficients of 0.798, an mIOU of 0.755, an SA of 0.767, and an OS of 0.017). Univariable and multivariable analysis indicate that T_P1NT was an independent risk factor for refracture. The performance of predicting refractures in different ROI regions showed that 3D_Full model exhibits the highest calibration performance, with a Hosmer-Lemeshow goodness-of-fit (HL) test statistic exceeding 0.05. The analysis of the training and test sets showed that the 3D_Full model, which integrates clinical and deep learning results, demonstrated superior performance with significant improvement (p-value < 0.05) compared to using clinical features independently or using only 3D_RoiOnly.

CONCLUSION

T_P1NT was an independent risk factor of refracture. Our 3D-FULL model showed better performance in predicting high-risk population of spine refracture than other models and junior doctors do. This model can be applicable to real-world translation due to its automatic segmentation and detection.

Multi-modal molecular determinants of clinically relevant osteoporosis subtypes

Due to a rapidly aging global population, osteoporosis and the associated risk of bone fractures have become a wide-spread public health problem. However, osteoporosis is very heterogeneous, and the existing standard diagnostic measure is not sufficient to accurately identify all patients at risk of osteoporotic fractures and to guide therapy. Here, we constructed the first prospective multi-omics atlas of the largest osteoporosis cohort to date (longitudinal data from 366 participants at three time points), and also implemented an explainable data-intensive analysis framework (DLSF: Deep Latent Space Fusion) for an omnigenic model based on a multi-modal approach that can capture the multi-modal molecular signatures (M3S) as explicit functional representations of hidden genotypes. Accordingly, through DLSF, we identified two subtypes of the osteoporosis population in Chinese individuals with corresponding molecular phenotypes, i.e., clinical intervention relevant subtypes (CISs), in which bone mineral density benefits response to calcium supplements in 2-year follow-up samples. Many snpGenes associated with these molecular phenotypes reveal diverse candidate biological mechanisms underlying osteoporosis, with xQTL preferences of osteoporosis and its subtypes indicating an omnigenic effect on different biological domains. Finally, these two subtypes were found to have different relevance to prior fracture and different fracture risk according to 4-year follow-up data. Thus, in clinical application, M3S could help us further develop improved diagnostic and treatment strategies for osteoporosis and identify a new composite index for fracture prediction, which were remarkably validated in an independent cohort (166 participants).

Predicting fracture risk for elderly osteoporosis patients by hybrid machine learning model

Background and Objective

Osteoporotic fractures significantly impact individuals's quality of life and exert substantial pressure on the social pension system. This study aims to develop prediction models for osteoporotic fracture and uncover potential risk factors based on Electronic Health Records (EHR).

Methods

Data of patients with osteoporosis were extracted from the EHR of Xinhua Hospital (July 2012-October 2017). Demographic and clinical features were used to develop prediction models based on 12 independent machine learning (ML) algorithms and 3 hybrid ML models. To facilitate a nuanced interpretation of the results, a comprehensive importance score was conceived, incorporating various perspectives to effectively discern and mine critical features from the data.

Results

A total of 8530 patients with osteoporosis were included for analysis, of which 1090 cases (12.8%) were fracture patients. The hybrid model that synergistically combines the Support Vector Machine (SVM) and XGBoost algorithms demonstrated the best predictive performance in terms of accuracy and precision (above 90%) among all benchmark models. Blood Calcium, Alkaline phosphatase (ALP), C-reactive Protein (CRP), Apolipoprotein A/B ratio and High-density lipoprotein cholesterol (HDL-C) were statistically found to be associated with osteoporotic fracture.

Conclusions

The hybrid machine learning model can be a reliable tool for predicting the risk of fracture in patients with osteoporosis. It is expected to assist clinicians in identifying high-risk fracture patients and implementing early interventions.

AI algorithms for accurate prediction of osteoporotic fractures in patients with diabetes: an up-to-date review

Osteoporotic fractures impose a substantial burden on patients with diabetes due to their unique characteristics in bone metabolism, limiting the efficacy of conventional fracture prediction tools. Artificial intelligence (AI) algorithms have shown great promise in predicting osteoporotic fractures. This review aims to evaluate the application of traditional fracture prediction tools (FRAX, QFracture, and Garvan FRC) in patients with diabetes and osteoporosis, review AI-based fracture prediction achievements, and assess the potential efficiency of AI algorithms in this population. This comprehensive literature search was conducted in Pubmed and Web of Science. We found that conventional prediction tools exhibit limited accuracy in predicting fractures in patients with diabetes and osteoporosis due to their distinct bone metabolism characteristics. Conversely, AI algorithms show remarkable potential in enhancing predictive precision and improving patient outcomes. However, the utilization of AI algorithms for predicting osteoporotic fractures in diabetic patients is still in its nascent phase, further research is required to validate their efficacy and assess the potential advantages of their application in clinical practice.

Machine learning algorithms for predicting the risk of fracture in patients with diabetes in China

Background

Patients with diabetes are more likely to be predisposed to fractures compared to those without diabetes. In clinical practice, predicting fracture risk in diabetics is still difficult because of the limited availability and accessibility of existing fracture prediction tools in the diabetic population. The purpose of this study was to develop and validate models using machine learning (ML) algorithms to achieve high predictive power for fracture in patients with diabetes in China.

Methods

In this study, the clinical data of 775 hospitalized patients with diabetes was analyzed by using Decision Tree (DT), Gradient Boosting Decision Tree (GBDT), Logistic Regression (LR), Random Forest (RF), Support Vector Machine (SVM), eXtreme Gradient Boosting (XGBoost) and Probabilistic Classification Vector Machines (PCVM) algorithms to construct risk prediction models for fractures. Moreover, the risk factors for diabetes-related fracture were identified by the feature selection algorithms.

Results

The ML algorithms extracted 17 most relevant factors from raw clinical data to maximize the accuracy of the prediction results, including bone mineral density, age, sex, weight, high-density lipoprotein cholesterol, height, duration of diabetes, total cholesterol, osteocalcin, N-terminal propeptide of type I, diastolic blood pressure, and body mass index. The 7 ML models including LR, SVM, RF, DT, GBDT, XGBoost, and PCVM had f1 scores of 0.75, 0.83, 0.84, 0.85, 0.87, 0.88, and 0.97, respectively.

Conclusions

This study identified 17 most relevant risk factors for diabetes-related fracture using ML algorithms. And the PCVM model proved to perform best in predicting the fracture risk in the diabetic population. This work proposes a cheap, safe, and extensible ML algorithm for the precise assessment of risk factors for diabetes-related fracture.

Vitamin D ameliorates insulin resistance-induced osteopenia by inactivating the nucleotide-binding oligomerization domain-like receptor protein 3 inflammasome

Objective

Osteoporosis (OP) can be considered a chronic complication of type 2 diabetes mellitus (T2DM). Aberrant activation of the nucleotide-binding oligomerization domain-like receptor protein 3 (NLRP3) inflammasome is associated with the pathogenesis of various inflammation-related diseases, e.g., T2DM and OP. Vitamin D affects the inflammatory pathway and inhibits an excessive inflammatory response. The current study investigated the inter-relationship between vitamin D and inflammasome activation in T2DM.

Method

Hepatocellular carcinoma (HepG2) cells and bone marrow stromal cells (BMSCs) were treated with Conditioned Medium of bone marrow mesenchymal stem cells after VitD treatment (CM-VitD), as well as phosphoinositide 3-kinase (PI3K) specific agonist, 740Y-P, or the PI3K specific inhibitor, LY294002, respectively, or both. 40 Eight-week-old female Sprague Dawley rats were selected and established as a DM model. The rats were injected with CM-VitD, as well as the 740Y-P specific agonist, or the LY294002 inhibitor, respectively, or both. A quantitative reverse transcription polymerase chain reaction and western blotting were conducted to evaluate the expression of messenger ribonucleic acid and protein in the RUX2 gene, alkaline phosphatase (ALP), OsteoPontiN (OPN), peroxisome proliferator-activated receptor gamma (PPARγ), fatty acid-binding protein 4 (FABP4), protein kinase B (AKT), PI3K, NLRP3, apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC), caspase-1, interleukin (IL)-1 beta (β), IL-18, and tumor necrosis factor alpha (TNF-α) in the BMSCs and liver tissue of rats. Enzyme-linked immunosorbent assay was used to detect the concentration of inflammatory factors in the cell supernatant and serum of rats.

Results

An isolated co-culture of HepG2/insulin-resistance cells and BMSCs promoted the adipogenic transformation of the latter and inhibited the transformation of BMSCs into osteogenesis. The PI3K specific agonist, 740Y-P, significantly increased the expression of PI3K, AKT, NLRP3, ASC and Caspase-1 while the PI3K specific inhibitor, LY294002, does the opposite. Additionally, CM-VitD reduced the expression of NLRP3, ASC, caspase-1, IL-1β, and IL-18 in BMSCs and rat liver via the PI3K/AKT pathway.

Conclusion

Vitamin D can inhibit the inflammatory response induced by T2DM and promote the osteogenesis of BMSCs, which may play a key role in the treatment of type 2 diabetes patients with OP.

Interpretable Deep-Learning Approaches for Osteoporosis Risk Screening and Individualized Feature Analysis Using Large Population-Based Data: Model Development and Performance Evaluation

BACKGROUND

Osteoporosis is one of the diseases that requires early screening and detection for its management. Common clinical tools and machine-learning (ML) models for screening osteoporosis have been developed, but they show limitations such as low accuracy. Moreover, these methods are confined to limited risk factors and lack individualized explanation.

OBJECTIVE

The aim of this study was to develop an interpretable deep-learning (DL) model for osteoporosis risk screening with clinical features. Clinical interpretation with individual explanations of feature contributions is provided using an explainable artificial intelligence (XAI) technique.

METHODS

We used two separate data sets: the National Health and Nutrition Examination Survey data sets from the United States (NHANES) and South Korea (KNHANES) with 8274 and 8680 respondents, respectively. The study population was classified according to the T-score of bone mineral density at the femoral neck or total femur. A DL model for osteoporosis diagnosis was trained on the data sets and significant risk factors were investigated with local interpretable model-agnostic explanations (LIME). The performance of the DL model was compared with that of ML models and conventional clinical tools. Additionally, contribution ranking of risk factors and individualized explanation of feature contribution were examined.

RESULTS

Our DL model showed area under the curve (AUC) values of 0.851 (95% CI 0.844-0.858) and 0.922 (95% CI 0.916-0.928) for the femoral neck and total femur bone mineral density, respectively, using the NHANES data set. The corresponding AUC values for the KNHANES data set were 0.827 (95% CI 0.821-0.833) and 0.912 (95% CI 0.898-0.927), respectively. Through the LIME method, significant features were induced, and each feature's integrated contribution and interpretation for individual risk were determined.

CONCLUSIONS

The developed DL model significantly outperforms conventional ML models and clinical tools. Our XAI model produces high-ranked features along with the integrated contributions of each feature, which facilitates the interpretation of individual risk. In summary, our interpretable model for osteoporosis risk screening outperformed state-of-the-art methods.