Optimizing prediction accuracy for early recurrent lumbar disc herniation with a directional mutation-guided SVM model.

Percutaneous endoscopic lumbar discectomy (PELD) is one of the main means of minimally invasive spinal surgery, and is an effective means of treating lumbar disc herniation, but its early recurrence is still difficult to predict. With the development of machine learning technology, the auxiliary model based on the prediction of early recurrent lumbar disc herniation (rLDH) and the identification of causative risk factors have become urgent problems in current research. However, the screening ability of current models for key factors affecting the prediction of rLDH, as well as their predictive ability, needs to be improved. Therefore, this paper presents a classification model that utilizes wrapper feature selection, developed through the integration of an enhanced bat algorithm (BDGBA) and support vector machine (SVM). Among them, BDGBA increases the population diversity and improves the population quality by introducing directional mutation strategy and guidance-based strategy, which in turn allows the model to secure better subsets of features. Furthermore, SVM serves as the classifier for the wrapper feature selection method, with its classification prediction results acting as a fitness function for the feature subset. In the proposed prediction method, BDGBA is used as an optimizer for feature subset filtering and as an objective function for feature subset evaluation based on the classification results of the support vector machine, which improves the interpretability and prediction accuracy of the model. In order to verify the performance of the proposed method, this paper proves the performance of the model through global optimization experiments and prediction experiments on real data sets. The accuracy of the proposed rLDH prediction model is 93.49% and sensitivity is 88.33%. The experimental results show that Level of herniated disk, Modic change, Disk height, Disk length, and Disk width are the key factors for predicting rLDH, and the proposed method is an effective auxiliary diagnosis method.

Comparative study of two models predicting the risk of deep vein thrombosis progression in spinal trauma patients after operation.

OBJECTIVE: Patients with preoperative deep vein thrombosis (DVT) exhibit a notable incidence of postoperative deep vein thrombosis progression (DVTp), which bears a potential for silent, severe consequences. Consequently, the development of a predictive model for the risk of postoperative DVTp among spinal trauma patients is important. METHODS: Data of 161 spinal traumatic patients with preoperative DVT, who underwent spine surgery after admission, were collected from our hospital between January 2016 and December 2022. The least absolute shrinkage and selection operator (LASSO) combined with multivariable logistic regression analysis was applied to select variables for the development of the predictive logistic regression models. One logistic regression model was formulated simply with the Caprini risk score (Model A), while the other model incorporated not only the previously screened variables but also the age variable (Model B). The model's capability was evaluated using sensitivity, specificity, positive predictive value, negative predictive value, accuracy, F1 score, and receiver operating characteristic (ROC) curve. Nomograms simplified and visually presented Model B for the clinicians and patients to understand the predictive model. The decision curve was used to analyze the clinical value of Model B. RESULTS: A total of 161 DVT patients were enrolled in this study. Postoperative DVTp occurred in 48 spinal trauma patients, accounting for 29.81% of the total patient enrolled. Model A inadequately predicted postoperative DVTp in spinal trauma patients, with ROC AUC values of 0.595 for the training dataset and 0.593 for the test dataset. Through the application of LASSO regression and multivariable logistic regression, a screening process was conducted for seven risk factors: D-dimer, blood platelet, hyperlipidemia, blood group, preoperative anticoagulant, spinal cord injury, lower extremity varicosities. Model B demonstrated superior and consistent predictive performance, with ROC AUC values of 0.809 for the training dataset and 0.773 for the test dataset. According to the calibration curves and decision curve analysis, Model B could accurately predict the probability of postoperative DVTp after spine surgery. The nomograms enhanced the interpretability of Model B in charts and graphs. CONCLUSIONS: In summary, we established a logistic regression model for the accurate predicting of postoperative deep vein thrombosis progression in spinal trauma patients, utilizing D-dimer, blood platelet, hyperlipidemia, blood group, preoperative anticoagulant, spinal cord injury, lower extremity varicosities, and age as predictive factors. The proposed model outperformed a logistic regression model based simply on CRS. The proposed model has the potential to aid frontline clinicians and patients in identifying and intervening in postoperative DVTp among traumatic patients undergoing spinal surgery.

Perceived brightness of incandescent and LED aviation signal lights.

INTRODUCTION: Aviation signal lights using light emitting diodes (LEDs) are commonly perceived as brighter than those using incandescent sources, even at the same measured intensity. In general, saturated colors, like those produced by LEDs, appear brighter than less saturated lights, like those produced by incandescent sources. METHODS: We describe a series of experiments quantifying the brightness of simulated blue, white, and green LED signal lights relative to incandescent signal lights of the same hue. Simulated signal lights and arrays were compared against dark and against dimly lighted backgrounds, and through simulated fog. RESULTS: The results confirm that LED signal lights are brighter than incandescent signals at matched luminous intensities. Brightness relationships were unaffected by background light level, and by the number of signals viewed, but the simulated fog reduced the brightness difference between the incandescent and LED signal lights. CONCLUSIONS: The present results could not be accurately predicted by several previously published models of brightness appearance, probably because of differences in experimental conditions. We present a new model that can be used to predict signal light brightness for blue, white, and green signal colors. Except for very short-wavelength blue signal lights, the model was able to accurately predict the present brightness data as well as those from previously published independent experiments. This validation lends confidence to the generality of the model for predicting blue, white, and green signal light brightness, but different colors (e.g., yellow or red) remain to be tested and modeled using this approach.