Application of deep learning upon spinal radiographs to predict progression in adolescent idiopathic scoliosis at first clinic visit

Deep Learning-Based Prediction Model for the Cobb Angle in Adolescent Idiopathic Scoliosis Patients

Scoliosis, characterized by spine deformity, is most common in adolescent idiopathic scoliosis (AIS). Manual Cobb angle measurement limitations underscore the need for automated tools. This study employed a vertebral landmark extraction method and Feedforward Neural Network (FNN) to predict scoliosis progression in 79 AIS patients. The novel intervertebral angles matrix format showcased results. The mean absolute error for the intervertebral angle progression was 1.5 degrees, while the Pearson correlation of the predicted Cobb angles was 0.86. The accuracy in classifying Cobb angles (<15°, 15-25°, 25-35°, 35-45°, >45°) was 0.85, with 0.65 sensitivity and 0.91 specificity. The FNN demonstrated superior accuracy, sensitivity, and specificity, aiding in tailored treatments for potential scoliosis progression. Addressing FNNs' over-fitting issue through strategies like "dropout" or regularization could further enhance their performance. This study presents a promising step towards automated scoliosis diagnosis and prognosis.

Artificial Intelligence in Spinal Imaging and Patient Care: A Review of Recent Advances

Artificial intelligence (AI) is transforming spinal imaging and patient care through automated analysis and enhanced decision-making. This review presents a clinical task-based evaluation, highlighting the specific impact of AI techniques on different aspects of spinal imaging and patient care. We first discuss how AI can potentially improve image quality through techniques like denoising or artifact reduction. We then explore how AI enables efficient quantification of anatomical measurements, spinal curvature parameters, vertebral segmentation, and disc grading. This facilitates objective, accurate interpretation and diagnosis. AI models now reliably detect key spinal pathologies, achieving expert-level performance in tasks like identifying fractures, stenosis, infections, and tumors. Beyond diagnosis, AI also assists surgical planning via synthetic computed tomography generation, augmented reality systems, and robotic guidance. Furthermore, AI image analysis combined with clinical data enables personalized predictions to guide treatment decisions, such as forecasting spine surgery outcomes. However, challenges still need to be addressed in implementing AI clinically, including model interpretability, generalizability, and data limitations. Multicenter collaboration using large, diverse datasets is critical to advance the field further. While adoption barriers persist, AI presents a transformative opportunity to revolutionize spinal imaging workflows, empowering clinicians to translate data into actionable insights for improved patient care.

3D prediction of curve progression in adolescent idiopathic scoliosis based on biplanar radiological reconstruction

Aims

This systematic review aims to identify 3D predictors derived from biplanar reconstruction, and to describe current methods for improving curve prediction in patients with mild adolescent idiopathic scoliosis.

Methods

A comprehensive search was conducted by three independent investigators on MEDLINE, PubMed, Web of Science, and Cochrane Library. Search terms included "adolescent idiopathic scoliosis","3D", and "progression". The inclusion and exclusion criteria were carefully defined to include clinical studies. Risk of bias was assessed with the Quality in Prognostic Studies tool (QUIPS) and Appraisal tool for Cross-Sectional Studies (AXIS), and level of evidence for each predictor was rated with the Grading of Recommendations, Assessment, Development, and Evaluations (GRADE) approach. In all, 915 publications were identified, with 377 articles subjected to full-text screening; overall, 31 articles were included.

Results

Torsion index (TI) and apical vertebral rotation (AVR) were identified as accurate predictors of curve progression in early visits. Initial TI > 3.7° and AVR > 5.8° were predictive of curve progression. Thoracic hypokyphosis was inconsistently observed in progressive curves with weak evidence. While sagittal wedging was observed in mild curves, there is insufficient evidence for its correlation with curve progression. In curves with initial Cobb angle < 25°, Cobb angle was a poor predictor for future curve progression. Prediction accuracy was improved by incorporating serial reconstructions in stepwise layers. However, a lack of post-hoc analysis was identified in studies involving geometrical models.

Conclusion

For patients with mild curves, TI and AVR were identified as predictors of curve progression, with TI > 3.7° and AVR > 5.8° found to be important thresholds. Cobb angle acts as a poor predictor in mild curves, and more investigations are required to assess thoracic kyphosis and wedging as predictors. Cumulative reconstruction of radiographs improves prediction accuracy. Comprehensive analysis between progressive and non-progressive curves is recommended to extract meaningful thresholds for clinical prognostication.

Development of a Machine Learning Algorithm to Correlate Lumbar Disc Height on X-rays with Disc Bulging or Herniation

Lumbar disc bulging or herniation (LDBH) is one of the major causes of spinal stenosis and related nerve compression, and its severity is the major determinant for spine surgery. MRI of the spine is the most important diagnostic tool for evaluating the need for surgical intervention in patients with LDBH. However, MRI utilization is limited by its low accessibility. Spinal X-rays can rapidly provide information on the bony structure of the patient. Our study aimed to identify the factors associated with LDBH, including disc height, and establish a clinical diagnostic tool to support its diagnosis based on lumbar X-ray findings. In this study, a total of 458 patients were used for analysis and 13 clinical and imaging variables were collected. Five machine-learning (ML) methods, including LASSO regression, MARS, decision tree, random forest, and extreme gradient boosting, were applied and integrated to identify important variables for predicting LDBH from lumbar spine X-rays. The results showed L4-5 posterior disc height, age, and L1-2 anterior disc height to be the top predictors, and a decision tree algorithm was constructed to support clinical decision-making. Our study highlights the potential of ML-based decision tools for surgeons and emphasizes the importance of L1-2 disc height in relation to LDBH. Future research will expand on these findings to develop a more comprehensive decision-supporting model.

Predicting Progression in Adolescent Idiopathic Scoliosis at the First Visit by Integrating 2D Imaging and 1D Clinical Information

STUDY DESIGN

Retrospective observational study.

OBJECTIVES

The prediction of curve progression in patients with adolescent idiopathic scoliosis (AIS) remains an unresolved area in orthopedic surgery. To make a rapid meaningful prediction, easily accessible multi-dimensional data at the patient's first consultation should be used. Current studies use clinical growth parameters and numerical values extracted from radiographs to compile a predictive model, leaving out the radiographs themselves. Such practice inevitably wastes a lot of information. Thus, this study aims to create a neural network that can predict AIS progression among patients with curves indicated for bracing by integrating both one-dimensional (1D) clinical and two-dimensional (2D) radiological data collected at the patient's first visit in a fully automated manner.

METHODS

513 idiopathic scoliosis patients indicated for and managed with bracing orthosis were recruited. After exclusion, 463 patients were included in deep learning analysis. Processed first-visit growth parameters and posteroanterior radiographs are used as training inputs and the curve progression outcomes obtained in follow ups are used as binary training outputs. The CapsuleNet architecture was modified and trained accordingly to make a prediction.

RESULTS

The final model achieved 90% sensitivity with an overall accuracy of 73.9% in the prediction of AIS in-brace curve progression by using first-visit multi-dimensional data, outperforming conventional convolutional neural networks.

CONCLUSIONS

This first-ever multidimensional-input model shows promise in serving as a screening tool for AIS in-brace curve progression. The incorporation of such a model into routine AIS diagnostic pipeline can assist orthopedics clinicians in personalizing the most appropriate management for each patient.

An intelligent composite model incorporating global / regional X-rays and clinical parameters to predict progressive adolescent idiopathic scoliosis curvatures and facilitate population screening

BACKGROUND

Adolescent idiopathic scoliosis (AIS) affects up to 5% of the population. The efficacy of school-aged screening remains controversial since it is uncertain which curvatures will progress following diagnosis and require treatment. Patient demographics, vertebral morphology, skeletal maturity, and bone quality represent individual risk factors for progression but have yet to be integrated towards accurate prognostication. The objective of this work was to develop composite machine learning-based prediction model to accurately predict AIS curves at-risk of progression.

METHODS

1870 AIS patients with remaining growth potential were identified. Curve progression was defined by a Cobb angle increase in the major curve of ≥6° between first visit and skeletal maturity in curves that exceeded 25°. Separate prediction modules were developed for i) clinical data, ii) global/regional spine X-rays, and iii) hand X-rays. The hand X-ray module performed automated image classification and segmentation tasks towards estimation of skeletal maturity and bone mineral density. A late fusion strategy integrated these domains towards the prediction of progressive curves at first clinic visit.

FINDINGS

Composite model performance was assessed on a validation cohort and achieved an accuracy of 83.2% (79.3-83.6%, 95% confidence interval), sensitivity of 80.9% (78.2-81.9%), specificity of 83.6% (78.8-84.1%) and an AUC of 0.84 (0.81-0.85), outperforming single modality prediction models (AUC 0.65-0.78).

INTERPRETATION

The composite prediction model achieved a high degree of accuracy. Upon incorporation into school-aged screening programs, patients at-risk of progression may be prioritized to receive urgent specialist attention, more frequent follow-up, and pre-emptive treatment.

FUNDING

Funding from The Society for the Relief of Disabled Children was awarded to GKHS.

Deep Learning Model to Classify and Monitor Idiopathic Scoliosis in Adolescents Using a Single Smartphone Photograph

Importance

Adolescent idiopathic scoliosis (AIS) is the most common pediatric spinal disorder. Routine physical examinations by trained personnel are critical to diagnose severity and monitor curve progression in AIS. In the presence of concerning malformation, radiographs are necessary for diagnosis or follow-up, guiding further management, such as bracing correction for moderate malformation and spine surgery for severe malformation. If left unattended, progressive deterioration occurs in two-thirds of patients, leading to significant health concerns for growing children.

Objective

To assess the ability of an open platform application (app) using a validated deep learning model to classify AIS severity and curve type, as well as identify progression.

Design, Setting, and Participants

This diagnostic study was performed with data from radiographs and smartphone photographs of the backs of adolescent patients at spine clinics. The ScolioNets deep learning model was developed and validated in a prospective training cohort, then incorporated and tested in the AlignProCARE open platform app in 2022. Ground truths (GTs) included severity, curve type, and progression as manually annotated by 2 experienced spine specialists based on the radiographic examinations of the participants' spines. The GTs and app results were blindly compared with another 2 spine surgeons' assessments of unclothed back appearance. Data were analyzed from October 2022 to February 2023.

Exposure

Acquisitions of unclothed back photographs using a mobile app.

Main Outcomes and Measures

Outcomes of interest were classification of AIS severity and progression. Quantitative statistical analyses were performed to assess the performance of the deep learning model in classifying the deformity as well as in distinguishing progression during 6-month follow-up.

Results

The training data set consisted of 1780 patients (1295 [72.8%] female; mean [SD] age, 14.3 [3.3] years), and the prospective testing data sets consisted of 378 patients (279 [73.8%] female; mean [SD] age, 14.3 [3.8] years) and 376 follow-ups (294 [78.2%] female; mean [SD] age, 15.6 [2.9] years). The model recommended follow-up with an area under receiver operating characteristic curve (AUC) of 0.839 (95% CI, 0.789-0.882) and considering surgery with an AUC of 0.902 (95% CI, 0.859-0.936), while showing good ability to distinguish among thoracic (AUC, 0.777 [95% CI, 0.745-0.808]), thoracolumbar or lumbar (AUC, 0.760 [95% CI, 0.727-0.791]), or mixed (AUC, 0.860 [95% CI, 0.834-0.887]) curve types. For follow-ups, the model distinguished participants with or without curve progression with an AUC of 0.757 (95% CI, 0.630-0.858). Compared with both surgeons, the model could recognize severities and curve types with a higher sensitivity (eg, sensitivity for recommending follow-up: model, 84.88% [95% CI, 75.54%-91.70%]; senior surgeon, 44.19%; junior surgeon, 62.79%) and negative predictive values (NPVs; eg, NPV for recommending follow-up: model, 89.22% [95% CI, 84.25%-93.70%]; senior surgeon, 71.76%; junior surgeon, 79.35%). For distinguishing curve progression, the sensitivity and NPV were comparable with the senior surgeons (sensitivity, 63.33% [95% CI, 43.86%-80.87%] vs 77.42%; NPV, 68.57% [95% CI, 56.78%-78.37%] vs 72.00%). The junior surgeon reported an inability to identify curve types and progression by observing the unclothed back alone.

Conclusions

This diagnostic study of adolescent patients screened for AIS found that the deep learning app had the potential for out-of-hospital accessible and radiation-free management of children with scoliosis, with comparable performance as spine surgeons experienced in AIS management.

Regional disparity in epidemiological characteristics of adolescent scoliosis in China: Data from a screening program

Objective

We investigated regional disparities in rates of scoliosis among adolescents in western and eastern China and the dominant factors underlying these disparities.

Methods

This cross-sectional study used data from a school scoliosis screening program conducted in two typical areas: Yangpu District of Shanghai (eastern China) and Tianzhu Tibetan Autonomous County of Gansu Province (western China), during October 2020 to February 2021. Participants included adolescents aged 12-16 years (4,240 in Shanghai and 2,510 in Gansu Province). School scoliosis screening data were obtained on age, sex, height, weight and BMI, and region as well. We screened angles of trunk rotation in level of proximal thoracic (T1-T4), main thoracic (T5-T12), and lumbar (T12-L4) by the forward bend test with scoliometer. An angle of trunk rotation ≥5° was used as the criterion to identify suspected scoliosis.

Results

The proportion of suspected scoliosis was lower in Shanghai (6.9%) than in Gansu (8.6%). Angle of trunk rotation tended to increase with age in Shanghai, peaking at 15 years, but decreased with age in Gansu, and bottomed at 15 years. The angle of trunk rotation in the proximal thoracic, main thoracic, and lumbar part of the spine appeared to be larger in Gansu adolescents and in Shanghai female adolescents. Age was a relevant factor in angle trunk rotation in regression models and interacted with region as well.

Conclusion

We found regional and age- and sex-related disparities in rates of suspected scoliosis.

Current models to understand the onset and progression of scoliotic deformities in adolescent idiopathic scoliosis: a systematic review

PURPOSE

To create an updated and comprehensive overview of the modeling studies that have been done to understand the mechanics underlying deformities of adolescent idiopathic scoliosis (AIS), to predict the risk of curve progression and thereby substantiate etiopathogenetic theories.

METHODS

In this systematic review, an online search in Scopus and PubMed together with an analysis in secondary references was done, which yielded 86 studies. The modeling types were extracted and the studies were categorized accordingly.

RESULTS

Animal modeling, together with machine learning modeling, forms the category of black box models. This category is perceived as the most clinically relevant. While animal models provide a tangible idea of the biomechanical effects in scoliotic deformities, machine learning modeling was found to be the best curve-progression predictor. The second category, that of artificial models, has, just as animal modeling, a tangible model as a result, but focusses more on the biomechanical process of the scoliotic deformity. The third category is formed by computational models, which are very popular in etiopathogenetic parameter-based studies. They are also the best in calculating stresses and strains on vertebrae, intervertebral discs, and other surrounding tissues.

CONCLUSION

This study presents a comprehensive overview of the current modeling techniques to understand the mechanics of the scoliotic deformities, predict the risk of curve progression in AIS and thereby substantiate etiopathogenetic theories. Although AIS remains to be seen as a complex and multifactorial problem, the progression of its deformity can be predicted with good accuracy. Modeling of AIS develops rapidly and may lead to the identification of risk factors and mitigation strategies in the near future. The overview presented provides a basis to follow this development.

A Population-Based 3D Atlas of the Pathological Lumbar Spine Segment

The spine is the load-bearing structure of human beings and may present several disorders, with low back pain the most frequent problem during human life. Signs of a spine disorder or disease vary depending on the location and type of the spine condition. Therefore, we aim to develop a probabilistic atlas of the lumbar spine segment using statistical shape modeling (SSM) and then explore the variability of spine geometry using principal component analysis (PCA). Using computed tomography (CT), the human spine was reconstructed for 24 patients with spine disorders and then the mean shape was deformed upon specific boundaries (e.g., by ±3 or ±1.5 standard deviation). Results demonstrated that principal shape modes are associated with specific morphological features of the spine segment such as Cobb’s angle, lordosis degree, spine width and height. The lumbar spine atlas here developed has evinced the potential of SSM to investigate the association between shape and morphological parameters, with the goal of developing new treatments for the management of patients with spine disorders.

A deep convolutional neural network to predict the curve progression of adolescent idiopathic scoliosis: a pilot study

Background

Adolescent idiopathic scoliosis (AIS) is a three-dimensional spinal deformity that predominantly occurs in girls. While skeletal growth and maturation influence the development of AIS, accurate prediction of curve progression remains difficult because the prognosis for deformity differs among individuals. The purpose of this study is to develop a new diagnostic platform using a deep convolutional neural network (DCNN) that can predict the risk of scoliosis progression in patients with AIS.

Methods

Fifty-eight patients with AIS (49 females and 9 males; mean age: 12.5 ± 1.4 years) and a Cobb angle between 10 and 25 degrees (mean angle: 18.7 ± 4.5) were divided into two groups: those whose Cobb angle increased by more than 10 degrees within two years (progression group, 28 patients) and those whose Cobb angle changed by less than 5 degrees (non-progression group, 30 patients). The X-ray images of three regions of interest (ROIs) (lung [ROI1], abdomen [ROI2], and total spine [ROI3]), were used as the source data for learning and prediction. Five spine surgeons also predicted the progression of scoliosis by reading the X-rays in a blinded manner.

Results

The prediction performance of the DCNN for AIS curve progression showed an accuracy of 69% and an area under the receiver-operating characteristic curve of 0.70 using ROI3 images, whereas the diagnostic performance of the spine surgeons showed inferior at 47%. Transfer learning with a pretrained DCNN contributed to improved prediction accuracy.

Conclusion

Our developed method to predict the risk of scoliosis progression in AIS by using a DCNN could be a valuable tool in decision-making for therapeutic interventions for AIS.

Using deep transfer learning to detect scoliosis and spondylolisthesis from x-ray images

Recent years have witnessed wider prevalence of vertebral column pathologies due to lifestyle changes, sedentary behaviors, or injuries. Spondylolisthesis and scoliosis are two of the most common ailments with an incidence of 5% and 3% in the United States population, respectively. Both of these abnormalities can affect children at a young age and, if left untreated, can progress into severe pain. Moreover, severe scoliosis can even lead to lung and heart problems. Thus, early diagnosis can make it easier to apply remedies/interventions and prevent further disease progression. Current diagnosis methods are based on visual inspection by physicians of radiographs and/or calculation of certain angles (e.g., Cobb angle). Traditional artificial intelligence-based diagnosis systems utilized these parameters to perform automated classification, which enabled fast and easy diagnosis supporting tools. However, they still require the specialists to perform error-prone tedious measurements. To this end, automated measurement tools were proposed based on processing techniques of X-ray images. In this paper, we utilize advances in deep transfer learning to diagnose spondylolisthesis and scoliosis from X-ray images without the need for any measurements. We collected raw data from real X-ray images of 338 subjects (i.e., 188 scoliosis, 79 spondylolisthesis, and 71 healthy). Deep transfer learning models were developed to perform three-class classification as well as pair-wise binary classifications among the three classes. The highest mean accuracy and maximum accuracy for three-class classification was 96.73% and 98.02%, respectively. Regarding pair-wise binary classification, high accuracy values were achieved for most of the models (i.e., > 98%). These results and other performance metrics reflect a robust ability to diagnose the subjects’ vertebral column disorders from standard X-ray images. The current study provides a supporting tool that can reasonably help the physicians make the correct early diagnosis with less effort and errors, and reduce the need for surgical interventions.