Advanced Sampling Technique in Radiology Free-Text Data for Efficiently Building Text Mining Models by Deep Learning in Vertebral Fracture

This study aims to establish advanced sampling methods in free-text data for efficiently building semantic text mining models using deep learning, such as identifying vertebral compression fracture (VCF) in radiology reports. We enrolled a total of 27,401 radiology free-text reports of X-ray examinations of the spine. The predictive effects were compared between text mining models built using supervised long short-term memory networks, independently derived by four sampling methods: vector sum minimization, vector sum maximization, stratified, and simple random sampling, using four fixed percentages. The drawn samples were applied to the training set, and the remaining samples were used to validate each group using different sampling methods and ratios. The predictive accuracy was measured using the area under the receiver operating characteristics (AUROC) to identify VCF. At the sampling ratios of 1/10, 1/20, 1/30, and 1/40, the highest AUROC was revealed in the sampling methods of vector sum minimization as confidence intervals of 0.981 (95%CIs: 0.980-0.983)/0.963 (95%CIs: 0.961-0.965)/0.907 (95%CIs: 0.904-0.911)/0.895 (95%CIs: 0.891-0.899), respectively. The lowest AUROC was demonstrated in the vector sum maximization. This study proposes an advanced sampling method, vector sum minimization, in free-text data that can be efficiently applied to build the text mining models by smartly drawing a small amount of critical representative samples.

Establish and validate the reliability of predictive models in bone mineral density by deep learning as examination tool for women

While FRAX with BMD could be more precise in estimating the fracture risk, DL-based models were validated to slightly reduce the number of under- and over-treated patients when no BMD measurements were available. The validated models could be used to screen for patients at a high risk of fracture and osteoporosis.

PURPOSE

Fracture risk assessment tool (FRAX) is useful in classifying the fracture risk level, and precise prediction can be achieved by estimating both clinical risk factors and bone mineral density (BMD) using dual X-ray absorptiometry (DXA). However, DXA is not frequently feasible because of its cost and accessibility. This study aimed to establish the reliability of deep learning (DL)-based alternative tools for screening patients at a high risk of fracture and osteoporosis.

METHODS

Participants were enrolled from the National Bone Health Screening Project of Taiwan in this cross-sectional study. First, DL-based models were built to predict the lowest T-score value in either the lumbar spine, total hip, or femoral neck and their respective BMD values. The Bland-Altman analysis was used to compare the agreement between the models and DXA. Second, the predictive model to classify patients with a high fracture risk was built according to the estimated BMD from the first step and the FRAX score without BMD. The performance of the model was compared with the classification based on FRAX with BMD.

RESULTS

Approximately 10,827 women (mean age, 65.4 ± 9.4 years) were enrolled. In the prediction of the lumbar spine BMD, total hip BMD, femoral neck BMD, and lowest T-score, the root-mean-square error (RMSE) was 0.099, 0.089, 0.076, and 0.68, respectively. The Bland-Altman analysis revealed a nonsignificant difference between the predictive models and DXA. The FRAX score with femoral neck BMD for major osteoporotic fracture risk was 9.7% ± 6.7%, whereas the risk for hip fracture was 3.3% ± 4.6%. Comparison between the classification of FRAX with and without BMD revealed the accuracy rate, positive predictive value (PPV), and negative predictive value (NPV) of 78.8%, 64.6%, and 89.9%, respectively. The area under the receiver operating characteristic curve (AUROC), accuracy rate, PPV, and NPV of the classification model were 0.913 (95% confidence interval: 0.904-0.922), 83.5%, 71.2%, and 92.2%, respectively.

CONCLUSION

While FRAX with BMD could be more precise in estimating the fracture risk, DL-based models were validated to slightly reduce the number of under- and over-treated patients when no BMD measurements were available. The validated models could be used to screen for patients at a high risk of fracture and osteoporosis.

Evaluation of the 10 Years' Association Between DXA Screening and Mortality in Patients Who Received Vertebroplasty in the Taiwan Population

Background

Osteoporotic fractures are a significant cause of morbidity and mortality affecting population worldwide. All guidelines recommended vertebral fracture assessment by dual-energy X-ray absorptiometry (DXA). This study aimed at evaluation of any associated benefits of screening with DXA in patients who had received vertebroplasty in Taiwan.

Methods

Data were obtained from the National Health Insurance Research Database (NHIRD) in Taiwan. We retrospectively compared the data of patients, who were admitted for vertebroplasty, whether they received DXA screening or not. The outcomes of interest were recurrence of spinal fracture and mortality during a follow-up period of 10 years.

Results

From this Taiwan national database, the screening rate of osteoporosis in patient who received vertebroplasty was 11.7%. The mean age in the non-DXA screened cohort (n=32,986) was 74.03±12.21 years (71.98% female). In the DXA screened cohort (n=4361), the mean age was 76.43±9.19 years (79.91% female). During the 10-year follow-up period, after matching, non-DXA patients had significantly higher mortality rates than their DXA counterparts, which were 42.37% and 37.73% (p-value < 0.0001), respectively. The re-fracture rates between non-DXA and DXA patients were not significantly different at 17.26% and 16.89% (p-value = 0.1766), respectively.

Conclusion

The rate of DXA screening before patients receiving vertebroplasty was extremely low, at 11.7%. Our results showed that DXA screening before vertebroplasty in spinal fractures patients had lower mortality. From this national retrospective cohort study, routine screening of osteoporosis in spinal fracture patients can lead to reduction in mortality.

Decreased Long-Term Respiratory Infection Risk After Bariatric Surgery: a Comprehensive National Cohort Study

PURPOSE

This study aims to determine whether bariatric surgery (BS) decreases the risk of respiratory tract infections (RTIs) in obese patients.

MATERIALS AND METHODS

This retrospective, population-based, matched cohort study utilized data from the Taiwan National Health Insurance Research Database. All patients 18 to 55 years of age with obesity diagnosed between 2005 and 2010 were enrolled. Patients were separated into two groups based on whether they underwent BS. Two groups were selected using 1:1 propensity score matching according to age, sex, and comorbidities. The general population was also enrolled for comparison. The primary endpoint was the incidence of RTIs, including pneumonia, influenza, and bronchitis. All patients were followed up until the end of 2013, the primary endpoint, or death.

RESULTS

Compared to the non-surgery group, the BS group was at significantly lower risk for RTIs (aHR 0.432, 95% CI 0.340-0.549, p < 0.001) with shorter length of hospital stay (LOH) and lower cost. Regardless of the RTI-related mortality, pneumonia, influenza, and bronchitis rates, BS did have significant protective effects on the non-surgery group. Compared to the general population, the BS group was at higher risk for RTIs (aHR 3.601, 95% CI 2.742-4.728, p < 0.001) with similar LOH and lower cost.

CONCLUSION

Patients with obesity who underwent BS were at significantly lower risk for RTIs than obese patients who did not undergo BS but were at higher risk for RTIs than the general population. BS may result in a long-term reduction of the RTI risk.

Diagnostic accuracy of exercise thallium-201 scintigraphy in men with asymptomatic essential hypertension

Coronary artery disease is responsible for much of the morbidity and mortality in patients with essential hypertension, and these complications have proven to be relatively resistant to antihypertensive therapy. However, the diagnosis of coronary disease in the hypertensive population has been considered problematic. In the present study, 30 asymptomatic patients with mild to moderate hypertension with positive exercise electrocardiograms (ECG) or stress thallium-201 scintigrams underwent coronary angiography to determine the accuracy of these tests for coronary artery disease. The exercise ECG was positive in 25 subjects, of whom 15 had significant coronary lesions and 10 did not. Thallium-201 scintigraphy proved more accurate: 17 of 18 patients with reversible abnormalities had significant obstructive coronary disease anatomically corresponding to the defect, one patient with a fixed defect had normal coronary arteries and was found to have an idiopathic cardiomyopathy, and 9 of 11 without defects had no significant lesions. The results were similar in populations with and without echocardiographic criteria for left ventricular hypertrophy. These findings indicate that despite previous suggestions to the contrary, thallium-201 scintigraphy can accurately diagnose coronary artery disease in most patients with asymptomatic essential hypertension, and that most asymptomatic hypertensive patients with physiologic evidence of myocardial ischemia have associated coronary artery disease.

Effect of antibiotics on L-form induction of Neisseria meningitidis

Conditions have been developed that permit highly efficient conversion in vitro of all serogroups of Neisseria meningitidis to the L-form. During the course of the studies, optimal parameters for cultivation and propogation were determined and compared with those found in studies done by previous workers on meningococci. The plasma expander polyvinylpyrrolidone (PVP) (molecular weight, 40,000) has been incorporated into the medium as an osmotic stabilizer. A method is outlined for removing toxicity (due to impurities in PVP) reported in some earlier studies done in this laboratory. Light and scanning electron microscopy of the L-form inductants and of a stable L-form have shown characteristics distinctive among the strains observed. Finally, the effects of both cell wall and other antibiotics were studied. It was shown that cell wall antibiotics other than the commonly used potassium penicillin G could induce N. meningitidis to the L-form, using the PVP-stabilized medium base. Furthermore, the antibiotics vancomycin, colistin, nystatin, and trimethoprim, used at the concentrations found in commercially prepared Transgrow formulations, did not adversely affect meningococcal L-form induction. The system described herein will hopefully serve as an effective tool, specifically in the study of "aseptic" meningitis and possible meningococcal L-form etiology and, more generally, in evaluating the ability of bacterial L-forms to cause disease.