Opportunistic bone density screening for the abdominal radiologist using colored CT images: a pilot retrospective study

Development and validation of a radiomics-based model for predicting osteoporosis in patients with lumbar compression fractures

BACKGROUND

Osteoporosis, a metabolic bone disorder, markedly elevates fracture risks, with vertebral compression fractures being predominant. Antiosteoporotic treatments for patients with osteoporotic vertebral compression fractures (OVCF) lessen both the occurrence of subsequent fractures and associated pain. Thus, diagnosing osteoporosis in OVCF patients is vital.

PURPOSE

The aim of this study was to develop a predictive radiographic model using T1 sequence MRI images to accurately determine whether patients with lumbar spine compression fractures also have osteoporosis.

STUDY DESIGN

Retrospective cohort study.

PATIENT SAMPLE

Patients over 45 years of age diagnosed with a fresh lumbar compression fracture.

OUTCOME MEASURES

Diagnostic accuracy of the model (area under the ROC curve).

METHODS

The study retrospectively collected clinical and imaging data (MRI and DEXA) from hospitalized lumbar compression fracture patients (L1-L4) aged 45 years or older between January 2021 and June 2023. Using the pyradiomics package in Python, features from the lumbar compression fracture vertebral region of interest (ROI) were extracted. Downscaling of the extracted features was performed using the Mann-Whitney U test and the least absolute shrinkage selection operator (LASSO) algorithm. Subsequently, six machine learning models (Naive Bayes, Support Vector Machine [SVM], Decision Tree, Random Forest, Extreme Gradient Boosting [XGBoost], and Light Gradient Boosting Machine [LightGBM]) were employed to train and validate these features in predicting osteoporosis comorbidity in OVCF patients.

RESULTS

A total of 128 participants, 79 in the osteoporotic group and 49 in the nonosteoporotic group, met the study's inclusion and exclusion criteria. From the T1 sequence MRI images, 1906 imaging features were extracted in both groups. Utilizing the Mann-Whitney U test, 365 radiologic features were selected out of the initial 1,906. Ultimately, the lasso algorithm identified 14 significant radiological features. These features, incorporated into six conventional machine learning algorithms, demonstrated successful prediction of osteoporosis in the validation set. The NaiveBayes model yielded an area under the receiver operating characteristic curve (AUC) of 0.84, sensitivity of 0.87, specificity of 0.70, and accuracy of 0.81.

CONCLUSIONS

A NaiveBayes machine learning algorithm can predict osteoporosis in OVCF patients using t1-sequence MRI images of lumbar compression fractures. This approach aims to obviate the necessity for further osteoporosis assessments, diminish patient exposure to radiation, and bolster the clinical care of patients with OVCF.

Using advanced imaging to measure bone density, compression fracture risk, and risk for construct failure after spine surgery

Low bone mineral density (BMD) can predispose to vertebral body compression fractures and postoperative instrumentation failure. DEXA is considered the gold standard for measurement of BMD, however it is not obtained for all spine surgery patients preoperatively. There is a growing body of evidence suggesting that more routinely acquired spine imaging studies such as computed tomography (CT) and magnetic resonance imaging (MRI) can be opportunistically used to measure BMD. Here we review available studies that assess the validity of opportunistic screening with CT-derived Hounsfield Units (HU) and MRI-derived vertebral vone quality (VBQ) to measure BMD of the spine as well the utility of these measures in predicting postoperative outcomes. Additionally, we provide screening thresholds based on HU and VBQ for prediction of osteopenia/ osteoporosis and postoperative outcomes such as cage subsidence, screw loosening, proximal junctional kyphosis, and implant failure.

Evaluation of bone mineral density in adolescent idiopathic scoliosis using a three-dimensional finite element model: a retrospective study

BACKGROUND

Adolescent idiopathic scoliosis (AIS) is often accompanied by osteopenia and osteoporosis, which can cause serious complications. The aim of this study was to determine the specific bone mineral density (BMD) of each vertebral body in patients with AIS using biomechanical finite element modeling based on three-dimensional (3D) reconstruction.

METHODS

This retrospective study involved 56 patients with AIS. Computed tomography (CT) and radiography were performed. Spinal vertebrae were segmented from the spinal CT images of patients with AIS to reconstruct 3D vertebral models. The vertebral models were meshed into tetrahedral finite elements to assess the BMD.

RESULTS

The mean main curve Cobb angle was 88.6 ± 36.7°, and the mean kyphosis angle was 36.8 ± 31.5°. The mean BMD of the global spine was 0.83 ± 0.15 g/cm2. The highest BMD was measured on the concave side of the apex (0.98 ± 0.16 g/cm2). Apical vertebral BMD was negatively correlated with age and height (r = - 0.490, p = 0.009 and r =  - 0.478, p = 0.043, respectively). There were no significant differences in BMD values between the concave and convex sides (p > 0.05).

CONCLUSIONS

The 3D finite element modeling of BMD in patients with AIS is a reliable and accurate BMD measurement method. Using this method, the overall BMD of patients with AIS was shown to gradually decrease from the top to the bottom of the spine. Our findings provide valuable insights for surgical planning, choice of screw trajectories, and additional biomechanical analyzes using finite element models in the context of scoliosis.

Correlation of bone density to screw loosening in dynamic stabilization: an analysis of 176 patients

Although osteoporosis has negative impacts on lumbar fusion, its effects on screw loosening in dynamic stabilization remain elusive. We aimed to correlate bone mineral density (BMD) with screw loosening in Dynesys dynamic stabilization (DDS). Consecutive patients who underwent 2- or 3-level DDS for spondylosis, recurrent disc herniations, or low-grade spondylolisthesis at L3-5 were retrospectively reviewed. BMD was assessed by the Hounsfield Unit (HU) in vertebral bodies (VB) and pedicles with and without cortical bone (CB) on pre-operative computed tomography (CT). Screw loosening was assessed by radiographs and confirmed by CT. HU values were compared between the loosened and intact screws. 176 patients and 918 screws were analyzed with 78 loosened screws found in 36 patients (mean follow-up: 43.4 months). The HU values of VB were similar in loosened and intact screws (p = 0.14). The HU values of pedicles were insignificantly less in loosened than intact screws (including CB: 286.70 ± 118.97 vs. 297.31 ± 110.99, p = 0.45; excluding CB: 238.48 ± 114.90 vs. 240.51 ± 108.91, p = 0.88). All patients had clinical improvements. In conclusion, the HU values, as a surrogate for BMD, were unrelated to screw loosening in DDS. Therefore, patients with compromised BMD might be potential candidates for dynamic stabilization rather than fusion.

Prospective validation of a rapid CT-based bone mineral density screening method using colored spinal images

PURPOSE

To prospectively validate a method to accurately and rapidly differentiate normal from abnormal spinal bone mineral density (BMD) using colored abdominal CT images.

METHODS

For this prospective observational study, 196 asymptomatic women ≥ 50 years of age presenting for screening mammograms underwent routine nonenhanced CT imaging of the abdomen. The CT images were processed with software designed to generate sagittal colored images with green vertebral trabecular bone indicating normal BMD and red indicating abnormal BMD (low BMD or osteoporosis). Four radiologists evaluated L1/L2 BMD on sagittal images using visual assessment of grayscale images, quantitative measurements of mean vertebral attenuation, and visual assessment of colored images. Mean BMD values at L1/L2 using quantitative CT with a phantom served as the reference standard. The average accuracy and time of interpretation were calculated. Inter-observer agreement was assessed using intraclass correlation coefficient (ICC).

RESULTS

Mean attenuation at L1/L2 was highly correlated with mean BMD (r = 0.96/0.91, p < 0.001 for both). The average accuracy and mean time to assess BMD among four readers for differentiating normal from abnormal BMD was 66% and 6.0 s using visual assessment of grayscale images, 88% and 15.2 s using quantitative measurements of mean vertebral attenuation, and 92% and 2.1 s using visual assessment of colored images (p < 0.001 and p < 0.001, respectively). Inter-observer agreement was poor using visual assessment of grayscale images (ICC:0.31), good using quantitative measurements of mean vertebral attenuation (ICC:0.73), and excellent using visual assessment of colored images (ICC:0.90).

CONCLUSION

Detection of abnormal BMD using colored abdominal CT images was highly accurate, rapid, and had excellent inter-observer agreement.

Vertebral bone quality score predicts fragility fractures independently of bone mineral density

BACKGROUND

Current evidence suggests that dual-energy x-ray absorptiometry (DXA) scans, the conventional method defining osteoporosis, is underutilized and, when used, may underestimate patient risk for skeletal fragility. It has recently been suggested that other imaging modalities may better estimate bone quality, such as the magnetic resonance imaging (MRI)-based vertebral bone quality (VBQ) score which also may assess vertebral compression fracture risk in patients with spine metastases.

PURPOSE

To evaluate whether VBQ score is predictive of fragility fractures in a population with pre-existing low bone density and at high-risk for fracture.

STUDY DESIGN/SETTING

Retrospective single-center cohort.

PATIENT SAMPLE

Patients followed at a metabolic bone clinic for osteopenia and/or osteoporosis.

OUTCOME MEASURES

Radiographically-documented new-onset fragility fracture.

METHODS

Patients with a DXA and MRI scans at the time of consultation and ≥2-year follow-up were included. Details were gathered about patient demographics, health history, current medication use, and serological studies of kidney function and bone turnover. For each patient, VBQ score was calculated using T1-weighted lumbar MRI images. Univariable and multivariable analyses were used to identify the independent predictors of a new fragility fracture. To support the construct validity of VBQ, patient VBQ scores were compared to those in a cohort of 45 healthy adults.

RESULTS

Seventy-two (39.1%) study participants suffered fragility fractures, the occurrence of which was associated with higher VBQ score (3.50 vs. 3.01; p<.001), chronic glucocorticoid use (30.6% vs. 15.2%; p=.014), and a history of prior fragility fracture (36.1% vs. 21.4%; p=.030). Mean VBQ score across all patients in the study cohort was significantly higher than the mean VBQ score in the healthy controls (p<.001). In multivariable analysis, new-onset fracture was independently associated with history of prior fracture (OR=6.94; 95% confidence interval [2.48-19.40]; p<.001), higher VBQ score (OR=2.40 per point; [1.30-4.44]; p=.003), higher body mass index (OR=1.09 per kg/m²; [1.01-1.17]; p=.03), and chronic glucocorticoid use (OR=2.89; [1.03-8.17]; p=0.043). Notably, DXA bone mineral density (BMD) was not found to be significantly predictive of new-onset fractures in the multivariable analysis (p=.081).

CONCLUSIONS

Here we demonstrate the novel, MRI-derived VBQ score is both an independent predictor of fragility fracture in at-risk patients and a superior predictor of fracture risk than DXA-measured BMD. Given the frequency with which MRIs are obtained by patients undergoing spine surgery consultation, we believe the VBQ score could be a valuable tool for estimating bone quality in order to optimize the management of these patients.

Diagnostic accuracy of quantitative dual-energy CT-based bone mineral density assessment in comparison to Hounsfield unit measurements using dual x-ray absorptiometry as standard of reference

PURPOSE

To assess the diagnostic accuracy of phantomless dual-energy computed tomography (DECT)-based volumetric material decomposition to assess bone mineral density (BMD) of the lumbar spine for the detection of osteoporosis compared to Hounsfield unit (HU) measurements with dual x-ray absorptiometry (DXA) as reference standard.

METHOD

A total of two hundred lumbar vertebrae in 53 patients (28 men, 25 women; mean age, 52 years, range, 23-87 years) who had undergone clinically-indicated third-generation dual-source DECT and DXA within 30 days were retrospectively analyzed. For volumetric BMD assessment, dedicated DECT postprocessing software using material decomposition was applied, which enables color-coded three-dimensional mapping of the trabecular BMD distribution. Manual HU measurements were performed by defining five trabecular regions of interest (ROI) per vertebra as suggested by literature. The DXA T-score served as standard of reference (osteoporosis: T < -2.5). Sensitivity, specificity and the area under the curve (AUC) were primary metrics of diagnostic accuracy.

RESULTS

An optimal patient-based DECT-derived BMD cut-off of 84 mg/cm³ yielded 96 % sensitivity (22/23) and 93 % specificity (28/30) for detecting osteoporosis, while an optimal CT attenuation cut-off of 139 HU showed 65 % sensitivity (15/23) and 93 % specificity (28/30) for the detection of osteoporosis. Overall patient-based AUC were 0.930 (volumetric DECT) and 0.790 (HU analysis) (p < .001). Pearson's product-moment correlation showed higher correlation between DECT BMD and DXA values (r=0.780) compared to HU and DXA values (r=0.528) (p < .001).

CONCLUSIONS

Phantomless volumetric DECT yielded significantly more accurate BMD assessment of the lumbar spine and superior diagnostic accuracy of osteoporosis compared to HU measurements.

Novel MRI-based score for assessment of bone density in operative spine patients

BACKGROUND

Good bone quality is key in avoiding a multitude of afflictions, including osteoporotic fragility fractures and poor outcomes after spine surgery. In patients undergoing instrumented spine fusion, bone quality often dictates screw pullout strength, insertional torque, and vertebral body loading properties. While dual-energy X-ray absorptiometry (DEXA) screening is the current method of assessing bone mineral density, the majority of patients do not have DEXA measurements available before undergoing surgical instrumentation.

PURPOSE

To create a simple magnetic resonance imaging (MRI)-based score to evaluate bone quality and evaluate the degree to which it correlates with conventional DEXA scores.

STUDY DESIGN/SETTING

Retrospective cohort.

PATIENT SAMPLE

Patients ≥18 years of age undergoing spine surgery for degenerative conditions between 2013 and 2018.

OUTCOME MEASURES

Correlation of the vertebral bone quality (VBQ) score with DEXA T-scores, and association between VBQ score and presence of osteopenia/osteoporosis.

METHODS

Using noncontrast T1-weighted MRIs of the lumbar spine, the novel VBQ score was calculated for each patient. DEXA T-scores of the femoral neck and total hip were obtained and were compared with patient VBQ scores using linear regression and Pearson's correlation.

RESULTS

Among 68 patients included in this study, 37 were found to have osteopenia/osteoporosis (T-score < -1.0) based on DEXA. A greater VBQ score was significantly associated with the presence of osteopenia/osteoporosis with a predictive accuracy of 81%. VBQ scores correlated moderately with femoral neck T-scores, the lowest overall T-scores of each patient, and correlated fairly with total hip T-scores.

CONCLUSIONS

This is the first study to correlate the novel VBQ score obtained from MRIs with DEXA T-score. We found this score to be a significant predictor of healthy versus osteopenic/osteoporotic bone with an accuracy of 81%, and found that VBQ score was moderately correlated with femoral neck and overall lowest T-score.