Comparison of different machine learning classification models for predicting deep vein thrombosis in lower extremity fractures

Deep vein thrombosis (DVT) is a common complication in patients with lower extremity fractures. Once it occurs, it will seriously affect the quality of life and postoperative recovery of patients. Therefore, early prediction and prevention of DVT can effectively improve the prognosis of patients. This study constructed different machine learning models to explore their effectiveness in predicting DVT. Five prediction models were applied to the study, including Extreme Gradient Boosting (XGBoost) model, Logistic Regression (LR) model, RandomForest (RF) model, Multilayer Perceptron (MLP) model, and Support Vector Machine (SVM) model. Afterwards, the performance of the obtained prediction models was evaluated by area under the curve (AUC), accuracy, sensitivity, specificity, F1 score, and Kappa. The prediction performances of the models based on machine learning are as follows: XGBoost model (AUC = 0.979, accuracy = 0.931), LR model (AUC = 0.821, accuracy = 0.758), RF model (AUC = 0.970, accuracy = 0.921), MLP model (AUC = 0.830, accuracy = 0.756), SVM model (AUC = 0.713, accuracy = 0.661). On our data set, the XGBoost model has the best performance. However, the model still needs external verification research before clinical application.

Central Venous Catheter as a Novel Approach to Postoperative Thrombolysis in Patients with Acute Iliofemoral Deep Venous Thrombosis

BACKGROUND AND AIMS

Percutaneous mechanical thrombectomy (PMT) along with postoperative thrombolysis (POT) has been the standard treatment for acute iliofemoral deep venous thrombosis (IFDVT). However, commonly used catheter directed thrombolysis (CDT) approaches for POT carry certain disadvantages, including the need for a sheath, inferior comfortability, and catheter-related complications. Therefore, we propose a new simplified method of POT using a central venous catheter (CVC).

METHODS

The retrospective study analyzed patients with IFDVT who underwent POT using CVC from January 2020 to August 2021. The treatment modalities included filter placement, thrombus removal, iliac vein obstruction release, postoperative CVC thrombolysis, filter retrieval, and adequate full course anticoagulation.

RESULTS

A total of 39 patients were included in this retrospective study. All patients underwent PMT surgery with a procedure success rate of 100%. In the post-PMT CVC thrombolysis, the puncture sites were located in the below-knee vein, including 58.97% in the peroneal vein. The mean duration of CVC-directed thrombolysis was 3.69 ± 1.08 days, and the total urokinase dose was 2.27 ± 0.71 MIU. A total of 37 patients (94.87%) had successful thrombolysis with a length of hospital stay of 5.82 ± 2.21 days. During CVC-directed thrombolysis, only four minor bleeding complications occurred, two of which were indwelling catheter-related. During the 12-month follow-up period, the patency rate and post-thrombotic syndrome incidences were 97.44% and 2.56%, respectively.

CONCLUSION

Thrombolysis through a CVC is a feasible, safe, and effective POT method, and could be an alternative to the conventional CDT approach for patients with IFDVT.

Analysis of Related Influencing Factors of Deep Vein Thrombosis after Lumbar Internal Fixation and Treatment Strategy

Lumbar internal fixation is a traditional surgical method for the treatment of degenerative diseases of the lumbar spine. However, due to its large surgical trauma, it easily causes complications such as deep venous thrombosis (DVT) after the operation. DVT refers to the abnormal coagulation of blood in deep veins, blocking the lumen, causing venous blood return disorder, causing venous blood return disorder to cause swelling and pain, which affects the recovery of the patient's lumbar spine function. In severe cases, even complicated pulmonary embolism endangers the life and health of the patient. Therefore, it is extremely important to explore the related influencing factors and effective treatment of DVT. The purpose of this study was to investigate the influencing factors and effective treatment of DVT after lumbar internal fixation. Univariate analysis and multivariate a logistic regression model were used to analyze the related factors affecting DVT after lumbar internal fixation. Conventional treatments such as anticoagulation, promotion of venous blood return, and improvement of limb circulation were given to patients with DVT, and functional exercise was guided to compare the hypercoagulability and hyperviscosity of blood in patients with DVT before and after treatment. The results showed that the incidence of DVT after lumbar internal fixation was related to age, BMI, and bed time. Getting out of bed for functional exercise in time after surgery can effectively prevent the formation of DVT. Preoperative grading examination, intraoperative intervention, postoperative physical exercise, and other preventive guidance can be carried out according to different individuals during the perioperative period to prevent the formation of postoperative DVT.

Assessment of the Probability of Post-thrombotic Syndrome in Patients with Lower Extremity Deep Venous Thrombosis

This study was performed to assess the probability of post-thrombotic syndrome (PTS) after treatment of lower extremity deep venous thrombosis (LEDVT). Patients with LEDVT undergoing their first treatments in Nanjing First Hospital from January 2013 to December 2014 were enrolled in this study (156 patients were enrolled in the training cohort, and 135 patients were enrolled in the validation cohort). 51 and 45 patients developed PTS in the two cohorts, respectively. Independent risk factors for PTS were investigated in the training cohort, and these independent risk factors were employed to develop the APTSD scoring system with which to predict the probability of PTS. Four independent risk factors for PTS were identified: iliac vein compression syndrome, residual iliac-femoral vein thrombosis, residual femoral-popliteal vein thrombosis and insufficient anticoagulation. Patients in the training cohort were divided into 2 groups according to the APTSD score of ≤7.0 and >7.0 points regarding the probability of PTS (median PTS-free time, 21.82 vs. 18.84 months; P < 0.001). The accuracy of this score system was 81.7% for the training cohort and 82.5% for the validation cohort. Patients with an APTSD score of >7.0 points may have an increased probability of developing PTS.

Detecting stent geometry changes after venous recanalization using duplex ultrasound

Background

Patients with post-thrombotic syndrome due to chronic venous obstruction and resistant to conservative management can benefit from endovenous revascularization. The current study investigated the usefulness of duplex ultrasound in monitoring the stent changes over the time.

Method

All duplex ultrasound images of treated patients were reviewed retrospectively. The stent diameter and area during the follow-up visits have been analyzed.

Result

A total of 210 stents were placed in 137 limbs. Duplex ultrasound findings showed a decrease in area of stent in all patients (mean: 0.69 cm2). Reduction of stent area over the time was a predictor of stent patency (odds ratio: 0.910; confidence interval: 0.832–0.997).

Conclusion

Duplex ultrasound has sufficient accuracy in detection of stent changes and its patency. There is a discrepancy between diameter of the stent lumen in vitro and after deployment in all patients. Stent occlusion is related to reduction of stent lumen over the time rather than the percent of the stenosis.