Simultaneous automated ascertainment of prevalent vertebral fracture and abdominal aortic calcification in clinical practice: role in fracture risk assessment.

Whether simultaneous automated ascertainments of prevalent vertebral fracture (auto-PVFx) and abdominal aortic calcification (auto-AAC) on vertebral fracture assessment (VFA) lateral spine bone density (BMD) images jointly predict incident fractures in routine clinical practice is unclear. We estimated the independent associations of auto-PVFx and auto-AAC primarily with incident major osteoporotic and secondarily with incident hip and any clinical fractures in 11 013 individuals (mean [SD] age 75.8 [6.8] years, 93.3% female) who had a BMD test combined with VFA between March 2010 and December 2017. Auto-PVFx and auto-AAC were ascertained using convolutional neural networks (CNNs). Proportional hazards models were used to estimate the associations of auto-PVFx and auto-AAC with incident fractures over a mean (SD) follow-up of 3.7 (2.2) years, adjusted for each other and other risk factors. At baseline, 17% (n = 1881) had auto-PVFx and 27% (n = 2974) had a high level of auto-AAC (≥ 6 on scale of 0 to 24). Multivariable-adjusted hazard ratios (HR) for incident major osteoporotic fracture (95% C.I.) were 1.85 (1.59, 2.15) for those with compared to those without auto-PVFx, and 1.36 (1.14, 1.62) for those with high compared to low auto-AAC. The multivariable-adjusted HRs for incident hip fracture were 1.62 (95% C.I. 1.26 to 2.07) for those with compared to those without auto-PVFx, and 1.55 (95% C.I. 1.15 to 2.09) for those high auto-AAC compared to low auto-AAC. The 5-year cumulative incidence of major osteoporotic fracture was 7.1% in those with no auto-PVFx and low auto-AAC, 10.1% in those with no auto-PVFx and high auto-AAC, 13.4% in those with auto-PVFx and low auto-AAC, and 18.0% in those with auto-PVFx and high auto-AAC. While physician manual review of images in clinical practice will still be needed to confirm image quality and provide clinical context for interpretation, simultaneous automated ascertainment of auto-PVFx and auto-AAC can aid fracture risk assessment.

Individuals with calcification of their abdominal aorta (AAC) and vertebral fractures seen on lateral spine bone density images (easily obtained as part of a bone density test) are much more likely to have subsequent fractures. Prior studies have not shown if both AAC and prior vertebral fracture both contribute to fracture prediction in routine clinical practice. Additionally, a barrier to using these images to aid fracture risk assessment at the time of bone density testing has been the need for expert readers to be able to accurately detect both AAC and vertebral fractures. We have developed automated computer methods (using artificial intelligence) to accurately detect vertebral fracture (auto-PVFx) and auto-AAC on lateral spine bone density images for 11 013 older individuals having a bone density test in routine clinical practice. Over a 5-year follow-up period, 7.1% of those with no auto-PVFx and low auto-AAC, 10.1% of those with no auto-PVFx and high auto-AAC, 13.4% of those with auto-PVFx and low auto-AAC, and 18.0% of those with auto-PVFx and high auto-AAC had a major osteoporotic fracture. Auto-PVFx and auto-AAC, ascertained simultaneously on lateral spine bone density images, both contribute to the risk of subsequent major osteoporotic fractures in routine clinical practice settings.

Machine learning for abdominal aortic calcification assessment from bone density machine-derived lateral spine images.

BACKGROUND: Lateral spine images for vertebral fracture assessment can be easily obtained on modern bone density machines. Abdominal aortic calcification (AAC) can be scored on these images by trained imaging specialists to assess cardiovascular disease risk. However, this process is laborious and requires careful training. METHODS: Training and testing of model performance of the convolutional neural network (CNN) algorithm for automated AAC-24 scoring utilised 5012 lateral spine images (2 manufacturers, 4 models of bone density machines), with trained imaging specialist AAC scores. Validation occurred in a registry-based cohort study of 8565 older men and women with images captured as part of routine clinical practice for fracture risk assessment. Cox proportional hazards models were used to estimate the association between machine-learning AAC (ML-AAC-24) scores with future incident Major Adverse Cardiovascular Events (MACE) that including death, hospitalised acute myocardial infarction or ischemic cerebrovascular disease ascertained from linked healthcare data. FINDINGS: The average intraclass correlation coefficient between imaging specialist and ML-AAC-24 scores for 5012 images was 0.84 (95% CI 0.83, 0.84) with classification accuracy of 80% for established AAC groups. During a mean follow-up 4 years in the registry-based cohort, MACE outcomes were reported in 1177 people (13.7%). With increasing ML-AAC-24 scores there was an increasing proportion of people with MACE (low 7.9%, moderate 14.5%, high 21.2%), as well as individual MACE components (all p-trend <0.001). After multivariable adjustment, moderate and high ML-AAC-24 groups remained significantly associated with MACE (HR 1.54, 95% CI 1.31-1.80 & HR 2.06, 95% CI 1.75-2.42, respectively), compared to those with low ML-AAC-24. INTERPRETATION: The ML-AAC-24 scores had substantial levels of agreement with trained imaging specialists, and was associated with a substantial gradient of risk for cardiovascular events in a real-world setting. This approach could be readily implemented into these clinical settings to improve identification of people at high CVD risk. FUNDING: The study was supported by a National Health and Medical Research Council of Australia Ideas grant and the Rady Innovation Fund, Rady Faculty of Health Sciences, University of Manitoba.

Transitions between versions of the International Classification of Diseases and chronic disease prevalence estimates from administrative health data: a population-based study.

BACKGROUND: Diagnosis codes in administrative health data are routinely used to monitor trends in disease prevalence and incidence. The International Classification of Diseases (ICD), which is used to record these diagnoses, have been updated multiple times to reflect advances in health and medical research. Our objective was to examine the impact of transitions between ICD versions on the prevalence of chronic health conditions estimated from administrative health data. METHODS: Study data (i.e., physician billing claims, hospital records) were from the province of Manitoba, Canada, which has a universal healthcare system. ICDA-8 (with adaptations), ICD-9-CM (clinical modification), and ICD-10-CA (Canadian adaptation; hospital records only) codes are captured in the data. Annual study cohorts included all individuals 18 + years of age for 45 years from 1974 to 2018. Negative binomial regression was used to estimate annual age- and sex-adjusted prevalence and model parameters (i.e., slopes and intercepts) for 16 chronic health conditions. Statistical control charts were used to assess the impact of changes in ICD version on model parameter estimates. Hotelling's T2 statistic was used to combine the parameter estimates and provide an out-of-control signal when its value was above a pre-specified control limit. RESULTS: The annual cohort sizes ranged from 360,341 to 824,816. Hypertension and skin cancer were among the most and least diagnosed health conditions, respectively; their prevalence per 1,000 population increased from 40.5 to 223.6 and from 0.3 to 2.1, respectively, within the study period. The average annual rate of change in prevalence ranged from -1.6% (95% confidence interval [CI]: -1.8, -1.4) for acute myocardial infarction to 14.6% (95% CI: 13.9, 15.2) for hypertension. The control chart indicated out-of-control observations when transitioning from ICDA-8 to ICD-9-CM for 75% of the investigated chronic health conditions but no out-of-control observations when transitioning from ICD-9-CM to ICD-10-CA. CONCLUSIONS: The prevalence of most of the investigated chronic health conditions changed significantly in the transition from ICDA-8 to ICD-9-CM. These results point to the importance of considering changes in ICD coding as a factor that may influence the interpretation of trend estimates for chronic health conditions derived from administrative health data.

Mapping three versions of the international classification of diseases to categories of chronic conditions.

INTRODUCTION: Administrative health data capture diagnoses using the International Classification of Diseases (ICD), which has multiple versions over time. To facilitate longitudinal investigations using these data, we aimed to map diagnoses identified in three ICD versions - ICD-8 with adaptations (ICDA-8), ICD-9 with clinical modifications (ICD-9-CM), and ICD-10 with Canadian adaptations (ICD-10-CA) - to mutually exclusive chronic health condition categories adapted from the open source Clinical Classifications Software (CCS). METHODS: We adapted the CCS crosswalk to 3-digit ICD-9-CM codes for chronic conditions and resolved the one-to-many mappings in ICD-9-CM codes. Using this adapted CCS crosswalk as the reference and referring to existing crosswalks between ICD versions, we extended the mapping to ICDA-8 and ICD-10-CA. Each mapping step was conducted independently by two reviewers and discrepancies were resolved by consensus through deliberation and reference to prior research. We report the frequencies, agreement percentages and 95% confidence intervals (CI) from each step. RESULTS: We identified 354 3-digit ICD-9-CM codes for chronic conditions. Of those, 77 (22%) codes had one-to-many mappings; 36 (10%) codes were mapped to a single CCS category and 41 (12%) codes were mapped to combined CCS categories. In total, the codes were mapped to 130 adapted CCS categories with an agreement percentage of 92% (95% CI: 86%-98%). Then, 321 3-digit ICDA-8 codes were mapped to CCS categories with an agreement percentage of 92% (95% CI: 89%-95%). Finally, 3583 ICD-10-CA codes were mapped to CCS categories; 111 (3%) had a fair or poor mapping quality; these were reviewed to keep or move to another category (agreement percentage = 77% [95% CI: 69%-85%]). CONCLUSIONS: We developed crosswalks for three ICD versions (ICDA-8, ICD-9-CM, and ICD-10-CA) to 130 clinically meaningful categories of chronic health conditions by adapting the CCS classification. These crosswalks will benefit chronic disease studies spanning multiple decades of administrative health data.

A comparison of regularized logistic regression and random forest machine learning models for daytime diagnosis of obstructive sleep apnea.

A major challenge in big and high-dimensional data analysis is related to the classification and prediction of the variables of interest by characterizing the relationships between the characteristic factors and predictors. This study aims to assess the utility of two important machine-learning techniques to classify subjects with obstructive sleep apnea (OSA) using their daytime tracheal breathing sounds. We evaluate and compare the performance of the random forest (RF) and regularized logistic regression (LR) as feature selection tools and classification approaches for wakefulness OSA screening. Results show that the RF, which is a low-variance committee-based approach, outperforms the regularized LR in terms of blind-testing accuracy, specificity, and sensitivity with 3.5%, 2.4%, and 3.7% improvement, respectively. However, the regularized LR was found to be faster than the RF and resulted in a more parsimonious model. Consequently, both the RF and regularized LR feature reduction and classification approaches are qualified to be applied for the daytime OSA screening studies, depending on the nature of data and applications' purposes. Graphical Abstract.

Regularized logistic regression for obstructive sleep apnea screening during wakefulness using daytime tracheal breathing sounds and anthropometric information.

Obstructive sleep apnea (OSA) is a prevalent health problem. Developing a technology for quick OSA screening is momentous. In this study, we used regularized logistic regression to predict the OSA severity level of 199 individuals (116 males) with apnea/hypopnea index (AHI) ≥ 15 (moderate/severe OSA) and AHI < 5 (non-OSA) using their tracheal breathing sounds (TBS) recorded during daytime, while they were awake. The participants were guided to breathe through their nose, and then through their mouth at their deep breathing rate. The least absolute shrinkage and selection operator (LASSO) feature selection approach was used to select the discriminative features from the power spectra of the TBS and the anthropometric information. Using a five-fold cross-validation procedure, five different training sets and their corresponding blind-testing sets were formed. The average blind-testing classification accuracy over the five different folds was found to be 79.3% ± 6.1 with the sensitivity (specificity) of 82.2% ± 7.2% (75.8% ± 9.9%). The accuracy for the entire dataset was found to be 81.1% with sensitivity (specificity) of 84.4% (77.0%). The feature selection and classification procedures were intelligible and fast. The selected features were physiologically meaningful. Overall, the results show that TBS analysis can be used as a quick and reliable prediction of the presence and severity of OSA during wakefulness without a sleep study. Graphical abstract Wakefulness screening of obstructive sleep apnea using tracheal breathing sounds and anthropometric information by means of regularized logistic regression with the least absolute shrinkage and selection operator approach for feature selection and classification.

Identification of Vertebral Fractures by Convolutional Neural Networks to Predict Nonvertebral and Hip Fractures: A Registry-based Cohort Study of Dual X-ray Absorptiometry.

Background Detection of vertebral fractures (VFs) aids in management of osteoporosis and targeting of fracture prevention therapies. Purpose To determine whether convolutional neural networks (CNNs) can be trained to identify VFs at VF assessment (VFA) performed with dual-energy x-ray absorptiometry and if VFs identified by CNNs confer a similar prognosis compared with the expert reader reference standard. Materials and Methods In this retrospective study, 12 742 routine clinical VFA images obtained from February 2010 to December 2017 and reported as VF present or absent were used for CNN training and testing. All reporting physicians were diagnostic imaging specialists with at least 10 years of experience. Randomly selected training and validation sets were used to produce a CNN ensemble that calculates VF probability. A test set (30%; 3822 images) was used to assess CNN agreement with the human expert reader reference standard and CNN prediction of incident non-VFs. Statistical analyses included area under the receiver operating characteristic curve, two-tailed Student t tests, prevalence- and bias-adjusted κ value, Kaplan-Meier curves, and Cox proportional hazard models. Results This study included 12 742 patients (mean age, 76 years ± 7; 12 013 women). The CNN ensemble demonstrated an area under the receiver operating characteristic curve of 0.94 (95% confidence interval [CI]: 0.93, 0.95) for VF detection that corresponded to sensitivity of 87.4% (534 of 611), specificity of 88.4% (2838 of 3211), and prevalence- and bias-adjusted κ value of 0.77. On the basis of incident fracture data available for 2813 patients (mean follow up, 3.7 years), hazard ratios adjusted for baseline fracture probability were 1.7 (95% CI: 1.3, 2.2) for CNN versus 1.8 (95% CI: 1.3, 2.3) for expert reader-detected VFs for incident non-VF and 2.3 (95% CI: 1.5, 3.5) versus 2.4 (95% CI: 1.5, 3.7) for incident hip fracture. Conclusion Convolutional neural networks can identify vertebral fractures on vertebral fracture assessment images with high accuracy, and these convolutional neural network-identified vertebral fractures predict clinical fracture outcomes. © RSNA, 2019 Online supplemental material is available for this article.