Prediction and Diagnosis of Venous Thromboembolism Using Artificial Intelligence Approaches: A Systematic Review and Meta-Analysis

Venous thromboembolism performance measurement in the United States: An evolving landscape with many stakeholders

Venous thromboembolism (VTE), including deep vein thrombosis and pulmonary embolism, is a life-threatening, costly, and common preventable complication associated with hospitalization. Although VTE prevention strategies such as risk assessment and prophylaxis are available, they are not applied uniformly or systematically across US hospitals and healthcare systems. Hospital-level performance measurement has been used nationally to promote standardized approaches for VTE prevention and incentivize the adoption of guideline-based care management. Though most measures reflect care processes rather than outcomes, certain domains including diagnosis, treatment, and continuity of care remain unmeasured. In this article, we describe the development of VTE prevention measures from various stakeholders, measure strengths and limitations, publicly reported rates, the impact of technology and health policy on measure use, and perspectives on future options for surveillance and performance monitoring.

The diagnostic value of artificial intelligence-assisted imaging for developmental dysplasia of the hip: a systematic review and meta-analysis

OBJECTIVE

To clarify the efficacy of artificial intelligence (AI)-assisted imaging in the diagnosis of developmental dysplasia of the hip (DDH) through a meta-analysis.

METHODS

Relevant literature on AI for early DDH diagnosis was searched in PubMed, Web of Science, Embase, and The Cochrane Library databases until April 4, 2024. The Quality Assessment of Diagnostic Accuracy Studies tool was used to assess the quality of included studies. Revman5.4 and StataSE-64 software were used to calculate the combined sensitivity, specificity, AUC value, and DOC value of AI-assisted imaging for DDH diagnosis.

RESULTS

The meta-analysis included 13 studies (6 prospective and 7 retrospective) with 28 AI models and a total of 10,673 samples. The summary sensitivity, specificity, AUC value, and DOC value were 99.0% (95% CI: 97.0-100.0%), 94.0% (95% CI: 89.0-96.0%), 99.0% (95% CI: 98.0-100.0%), and 1342 (95% CI: 469-3842), respectively.

CONCLUSION

AI-assisted imaging demonstrates high diagnostic efficacy for DDH detection, improving the accuracy of early DDH imaging examination. More prospective studies are needed to further confirm the value of AI-assisted imaging for early DDH diagnosis.

Cohort study of prediction of venous thromboembolism in emergency department patients with extremity symptoms

Despite diagnostic algorithms, identification of venous thromboembolism (VTE) in emergency departments (ED) remains a challenge. We evaluated symptoms, background, and laboratory data in 27,647 ED patients presenting with pain, swelling, or other symptoms from the extremities, and identified predictors of VTE diagnosis within one year. Predictors of a clinical decision to perform phlebography, ultrasound, or computer tomography (CT) angiography of pelvic, lower, or upper extremity veins, CT of pulmonary arteries, or pulmonary scintigraphy at the ED or within 30 days, and the results of such investigations were also evaluated. A total of 3195 patients (11.6%) were diagnosed with VTE within one year. In adjusted analysis of patients in whom all laboratory data were available, a d-dimer value ≥ 0.5 mg/l (odds ratio [OR]: 2.602; 95% confidence interval [CI] 1.894-3.575; p < 0.001) at the ED and a previous diagnosis of VTE (OR: 6.037; CI 4.465-8.162; p < 0.001) independently predicted VTE within one year. Of diagnosed patients, 2355 (73.7%) had undergone imaging within 30 days after the ED visit and 1730 (54.1%) were diagnosed at this examination. Lower age (OR: 0.984; CI 0.972-0.997; p = 0.014), higher blood hemoglobin (OR: 1.023; CI 1.010-1.037; p < 0.001), C-reactive protein (OR: 2.229; CI 1.433-3.468; p < 0.001), d-dimer (OR: 8.729; CI 5.614-13.574; p < 0.001), and previous VTE (OR: 7.796; CI 5.193-11.705; p < 0.001) predicted VTE on imaging within 30 days, whereas female sex (OR 0.602 [95% CI 0.392-0.924]; p = 0.020) and a previous diagnosis of ischemic heart disease (OR 0.254 [95% CI 0.113-0.571]; p = 0.001) were negative predictors of VTE. In conclusion, analysis of 27,647 ED patients with extremity symptoms confirmed the importance of well-established risk factors for VTE. Many patients developing VTE within one year had initial negative imaging, highlighting the importance of continued symptom vigilance.

Do Decision Support Tools Decrease the Prevalence of Hospital-Acquired Venous Thromboembolisms When Compared to Clinical Judgement? A Single-Center Pre-Post Study

Introduction: Hospital-acquired venous thromboembolisms (HA-VTEs) carry a significant health burden on patients and a financial burden on hospitals due to reimbursement penalties. VTE prophylaxis at our institute was performed through utilizing an order set based on healthcare professionals' perceived level of risk. However, the use of standardized risk assessment models is recommended by multiple professional societies. Furthermore, integrating decision support tools (DST) based on the standardized risk assessment models has been shown to increase the administration of appropriate deep vein thrombosis (DVT) prophylaxis. Nonetheless, such scoring systems are not inherently flawless and their integration into EMR as a mandatory step can come at the risk of healthcare professional fatigue and burnout. We conducted a study to evaluate the incidence of HA-VTE and length of stay pre- and post implementation of a DST. Methods: We conducted a retrospective, pre-post-implementation observational study at a tertiary medical center after implementing a mandatory DST. The DST used Padua scores for medical patients and Caprini scores for surgical patients. Patients were identified through ICD-10 codes and outcomes were collected from electronic charts. Healthcare professionals were surveyed through an anonymous survey and stored securely. Statistical analysis was conducted by using R (version 3.4.3). Results: A total of 343 patients developed HA-VTE during the study period. Of these, 170 patients developed HA-VTE in the 9 months following the implementation of the DST, while 173 patients were identified in the 9 months preceding the implementation. There was no statistically significant difference in mean HA-VTE/1000 discharge/month pre- and post implementation (4.4 (SD 1.6) compared to 4.6 (SD 1.2), confidence interval [CI] -1.6 to 1.2, p = 0.8). The DST was used in 73% of all HA-VTE cases over the first 6 months of implementation. The hospital length of stay (LOS) was 14.2 (SD 1.9) days prior to implementation and 14.1 (SD 1.6) days afterwards. No statistically significant change in readmission rates was noted (8.8% (SD 2.6) prior to implementation and 15.53% (SD 9.6) afterwards, CI -14.27 to 0.74, p = 0.07). Of the 56 healthcare professionals who answered the survey, 84% (n = 47) reported to be dissatisfied or extremely dissatisfied with the DST, while 91% (n = 51) reported that it slowed them down. Conclusions: There were no apparent changes in the prevalence of HA-VTE, length of stay, or readmission rates when VTE prophylaxis was mandated through DST compared to a prior model which used order sets based on perceived risk. Further studies are needed to further evaluate the current risk assessment models and improve healthcare professionals' satisfaction with DST.

Incidence of venous thromboembolism in patients with ovarian cancer receiving neoadjuvant chemotherapy: systematic review and meta-analysis

OBJECTIVE

Venous thromboembolism is associated with significant patient morbidity, mortality, and can lead to delays in treatment for patients with cancer. The objectives of this study were to identify the incidence of venous thromboembolism in patients with advanced ovarian cancer receiving neoadjuvant chemotherapy, and identify risk factors for venous thromboembolism.

METHODS

A systematic literature search of biomedical databases, including Ovid Medline, Web of Science, Scopus, CINAHL, and Embase was performed on December 6, 2022 and updated on December 21, 2023 for peer reviewed articles. Studies were included if they were cohort studies or randomized controlled trials that evaluated the incidence of venous thromboembolism for patients with ovarian cancer receiving neoadjuvant chemotherapy. Risk of bias assessment was performed using the Newcastle Ottawa Scale for cohort studies and the Cochrane risk of bias tool for randomized controlled trials. Random effects meta-analysis was used to pool results across studies.

RESULTS

A total of 2636 studies were screened, and 11 were included in the review. Ten were retrospective cohort studies, and one was a randomized controlled trial. The incidence of venous thromboembolism in the included studies ranged from 0% to 18.9%. The pooled incidence rate of venous thromboembolism was 10% (95% confidence interval (CI) 7% to 13%). This remained significant when restricted to only studies with a low risk of bias (pooled incidence of 11%, 95% CI 9% to 14%). Body mass index of ≥30 kg/m2 was a significant risk factor for venous thromboembolism with a pooled odds ratio of 1.76 (95% CI 1.13 to 2.76) CONCLUSIONS: The results from this study demonstrated a 10% incidence of venous thromboembolism for patients with advanced ovarian cancer receiving neoadjuvant chemotherapy. This suggests that there may be a role for universal thromboprophylaxis in this population.

TRIAL REGISTRATION

PROSPERO CRD42022339602.

The potential use of artificial intelligence for venous thromboembolism prophylaxis and management: clinician and healthcare informatician perspectives

Venous thromboembolism (VTE) is the leading cause of preventable death in hospitalized patients. Artificial intelligence (AI) and machine learning (ML) can support guidelines recommending an individualized approach to risk assessment and prophylaxis. We conducted electronic surveys asking clinician and healthcare informaticians about their perspectives on AI/ML for VTE prevention and management. Of 101 respondents to the informatician survey, most were 40 years or older, male, clinicians and data scientists, and had performed research on AI/ML. Of the 607 US-based respondents to the clinician survey, most were 40 years or younger, female, physicians, and had never used AI to inform clinical practice. Most informaticians agreed that AI/ML can be used to manage VTE (56.0%). Over one-third were concerned that clinicians would not use the technology (38.9%), but the majority of clinicians believed that AI/ML probably or definitely can help with VTE prevention (70.1%). The most common concern in both groups was a perceived lack of transparency (informaticians 54.4%; clinicians 25.4%). These two surveys revealed that key stakeholders are interested in AI/ML for VTE prevention and management, and identified potential barriers to address prior to implementation.

Artificial Intelligence in Perioperative Planning and Management of Liver Resection

Artificial intelligence (AI) is a speciality within computer science that deals with creating systems that can replicate the intelligence of a human mind and has problem-solving abilities. AI includes a diverse array of techniques and approaches such as machine learning, neural networks, natural language processing, robotics, and expert systems. An electronic literature search was conducted using the databases of "PubMed" and "Google Scholar". The period for the search was from 2000 to June 2023. The search terms included "artificial intelligence", "machine learning", "liver cancers", "liver tumors", "hepatectomy", "perioperative" and their synonyms in various combinations. The search also included all MeSH terms. The extracted articles were further reviewed in a step-wise manner for identification of relevant studies. A total of 148 articles were identified after the initial literature search. Initial review included screening of article titles for relevance and identifying duplicates. Finally, 65 articles were reviewed for this review article. The future of AI in liver cancer planning and management holds immense promise. AI-driven advancements will increasingly enable precise tumour detection, location, and characterisation through enhanced image analysis. ML algorithms will predict patient-specific treatment responses and complications, allowing for tailored therapies. Surgical robots and AI-guided procedures will enhance the precision of liver resections, reducing risks and improving outcomes. AI will also streamline patient monitoring, better hemodynamic management, enabling early detection of recurrence or complications. Moreover, AI will facilitate data-driven research, accelerating the development of novel treatments and therapies. Ultimately, AI's integration will revolutionise liver cancer care, offering personalised, efficient and effective solutions, improving patients' quality of life and survival rates.

Artificial intelligence in the prediction of venous thromboembolism: A systematic review and pooled analysis

BACKGROUND

Accurate diagnostic and prognostic predictions of venous thromboembolism (VTE) are crucial for VTE management. Artificial intelligence (AI) enables autonomous identification of the most predictive patterns from large complex data. Although evidence regarding its performance in VTE prediction is emerging, a comprehensive analysis of performance is lacking.

AIMS

To systematically review the performance of AI in the diagnosis and prediction of VTE and compare it to clinical risk assessment models (RAMs) or logistic regression models.

METHODS

A systematic literature search was performed using PubMed, MEDLINE, EMBASE, and Web of Science from inception to April 20, 2021. Search terms included "artificial intelligence" and "venous thromboembolism." Eligible criteria were original studies evaluating AI in the prediction of VTE in adults and reporting one of the following outcomes: sensitivity, specificity, positive predictive value, negative predictive value, or area under receiver operating curve (AUC). Risks of bias were assessed using the PROBAST tool. Unpaired t-test was performed to compare the mean AUC from AI versus conventional methods (RAMs or logistic regression models).

RESULTS

A total of 20 studies were included. Number of participants ranged from 31 to 111 888. The AI-based models included artificial neural network (six studies), support vector machines (four studies), Bayesian methods (one study), super learner ensemble (one study), genetic programming (one study), unspecified machine learning models (two studies), and multiple machine learning models (five studies). Twelve studies (60%) had both training and testing cohorts. Among 14 studies (70%) where AUCs were reported, the mean AUC for AI versus conventional methods were 0.79 (95% CI: 0.74-0.85) versus 0.61 (95% CI: 0.54-0.68), respectively (p < .001). However, the good to excellent discriminative performance of AI methods is unlikely to be replicated when used in clinical practice, because most studies had high risk of bias due to missing data handling and outcome determination.

CONCLUSION

The use of AI appears to improve the accuracy of diagnostic and prognostic prediction of VTE over conventional risk models; however, there was a high risk of bias observed across studies. Future studies should focus on transparent reporting, external validation, and clinical application of these models.

Prediction of Venous Thromboembolism in Patients With Cancer Using Machine Learning Approaches: A Systematic Review and Meta-Analysis

PURPOSE

Recent studies have suggested that machine learning (ML) could be used to predict venous thromboembolism (VTE) in cancer patients with high accuracy.

METHODS

We aimed to evaluate the performance of ML in predicting VTE events in patients with cancer. PubMed, Web of Science, and EMBASE to identify studies were searched.

RESULTS

Seven studies involving 12,249 patients with cancer were included. The combined results of the different ML models demonstrated good accuracy in the prediction of VTE. In the training set, the global pooled sensitivity was 0.87, the global pooled specificity was 0.87, and the AUC was 0.91, and in the test set 0.65, 0.84, and 0.80, respectively.

CONCLUSION

The prediction ML models showed good performance to predict VTE. External validation to determine the result's reproducibility is necessary.

Venous thromboembolism prophylaxis for hospitalized adult patients: a survey of US health care providers on attitudes and practices

Background

Venous thromboembolism (VTE) is a leading cause of preventable mortality among hospitalized patients, but appropriate risk assessment and thromboprophylaxis remain underutilized or misapplied.

Objectives

We conducted an electronic survey of US health care providers to explore attitudes, practices, and barriers related to thromboprophylaxis in adult hospitalized patients and at discharge.

Results

A total of 607 US respondents completed the survey: 63.1% reported working in an academic hospital, 70.7% identified as physicians, and hospital medicine was the most frequent specialty (52.1%). The majority of respondents agreed that VTE prophylaxis is important (98.8%; 95% CI: 97.6%-99.5%) and that current measures are safe (92.6%; 95% CI: 90.2%-94.5%) and effective (93.8%; 95% CI: 91.6%-95.6%), but only half (52.0%; 95% CI: 47.9%-56.0%) believed that hospitalized patients at their institution are on appropriate VTE prophylaxis almost all the time. One-third (35.4%) reported using a risk assessment model (RAM) to determine VTE prophylaxis need; 44.9% reported unfamiliarity with RAMs. The most common recommendation for improving rates of appropriate thromboprophylaxis was to leverage technology. A majority of respondents (84.5%) do not reassess a patient's need for VTE prophylaxis at discharge, and a minority educates patients about the risk (16.2%) or symptoms (18.9%) of VTE at discharge.

Conclusion

Despite guideline recommendations to use RAMs, the majority of providers in our survey do not use them. A majority of respondents believed that technology could help improve VTE prophylaxis rates. A majority of respondents do not reassess the risk of VTE at discharge or educate patients about this risk of VTE at discharge.

A prospective marker for the prediction of postoperative deep venous thrombosis: Neutrophil extracellular traps

Deep venous thrombosis (DVT) is a common medical complication in patients with lumbar fractures. The current study aimed to investigate the predictive value of neutrophil extracellular traps (NETs) in postoperative DVT formation in patients with lumbar fractures and to develop a nomogram relating clinical admission information for prediction. Patients who underwent open reduction and pedicle screw internal fixation in the treatment of single-segment lumbar fracture in the Department of Spine Surgery, the First Affiliated Hospital of Nanjing Medical University, from December 2020 to June 2022 were enrolled in this study. Baseline data and laboratory results were collected from enrollees, and the primary study endpoint event was the occurrence of DVT in patients after surgery. Multivariable logistic regression analysis was used to identify risk factors associated with higher odds of DVT after surgery. A nomogram was constructed using the results of the multivariable model. The calibration plot and receiver operating characteristics (ROC) curve were used to show the satisfactory predictive capacity of the model. Of these 393 patients who did not have DVT preoperatively, 79 patients developed it postoperatively, and 314 did not, respectively. Multivariate analysis showed that higher body mass index (BMI) (BMI between 24 and 28: RR = 1.661, 95% CI = 0.891-3.094; BMI ≤28: RR = 5.625, 95% CI = 2.590-12.217; reference: BMI <24), neutrophils (RR = 1.157, 95% CI 1.042-1.285), D-dimer (RR = 1.098, 95% CI 1.000-1.206), and citrullinated histone H3 (CitH3) (RR = 1.043, 95% CI 1.026-1.060) were independent risk factors for postoperative DVT. Using the multivariable analysis, we then constructed a nomogram to predict DVT, which was found to have an area under the curve of 0.757 (95% CI = 0.693-0.820). Calibration plots also showed the satisfied discrimination and calibration of the nomogram. In conclusion, patients with lumbar fractures with postoperative DVT had higher levels of NETs in the circulation preoperatively compared to those without postoperative DVT. Furthermore, based on BMI, D-dimer, neutrophils, and CitH3, we developed a predictive model to predict postoperative DVT incidence in these patients.