Endovascular therapy for acute severe pulmonary embolism

nnU-Net-based deep-learning for pulmonary embolism: detection, clot volume quantification, and severity correlation in the RSPECT dataset

OBJECTIVES

CT pulmonary angiography is the gold standard for diagnosing pulmonary embolism, and DL algorithms are being developed to manage the increase in demand. The nnU-Net is a new auto-adaptive DL framework that minimizes manual tuning, making it easier to develop effective algorithms for medical imaging even without specific expertise. This study assesses the performance of a locally developed nnU-Net algorithm on the RSPECT dataset for PE detection, clot volume measurement, and correlation with right ventricle overload.

MATERIALS & METHODS

User input was limited to segmentation using 3DSlicer. We worked with the RSPECT dataset and trained an algorithm from 205 PE and 340 negatives. The test dataset comprised 6573 exams. Performance was tested against PE characteristics, such as central, non-central, and RV overload. Blood clot volume (BCV) was extracted from each exam. We employed ROC curves and logistic regression for statistical validation.

RESULTS

Negative studies had a median BCV of 1 μL, which increased to 345 μL in PE-positive cases and 7,378 μL in central PEs. Statistical analysis confirmed a significant BCV correlation with PE presence, central PE, and increased RV/LV ratio (p < 0.0001). The model's AUC for PE detection was 0.865, with an 83 % accuracy at a 55 μL threshold. Central PE detection AUC was 0.937 with 91 % accuracy at 850 μL. The RV overload AUC stood at 0.848 with 79 % accuracy.

CONCLUSION

The nnU-Net algorithm demonstrated accurate PE detection, particularly for central PE. BCV is an accurate metric for automated severity stratification and case prioritization.

CLINICAL RELEVANCE STATEMENT

The nnU-Net framework can be utilized to create a dependable DL for detecting PE. It offers a user-friendly approach to those lacking expertise in AI and rapidly extracts the Blood Clot Volume, a metric that can evaluate the PE's severity.

Venous thromboembolism and intracranial hemorrhage in patients undergoing glioblastoma surgery

In the perioperative management of patients with glioblastoma (GBM), physicians face the question of whether and when to administer prophylactic or therapeutic anticoagulation (AC). In this study, we investigate the effects of the timing of postoperative heparinization on thromboembolic events (TE) and postoperative hemorrhage (bleeding, PH) as well as the interactions between the two in the context of an underlying intracerebral malignancy. For this retrospective data analysis, 222 patients who underwent surgery for grade IV glioblastoma, IDH-wildtype (2016 CNS WHO) between 01/01/2014 and 31/12/2019 were included. We followed up for 12 months. We assessed various biographical and clinical data for risk factors and focused on the connection between timepoint of AC and adverse events. Subgroup analyses were performed for pulmonary artery embolism (PE), deep vein thrombosis, and postoperative intracranial hemorrhage (PH) that either required surgical intervention or was controlled radiologically only. Statistical analysis was performed using Mann-Whitney U-Test, Chi-square test, Fisher's exact test and univariate binomial logistic regression. p values below 0.05 were considered statistically significant. There was no significant association between prophylactic AC within 24 h and more frequent major bleeding (p = 0.350). AC in patients who developed major bleeding was regularly postponed by the physician/surgeon upon detection of the re-bleeding; therefore, patients with PH were anticoagulated significantly later (p = 0.034). The timing of anticoagulant administration did not differ significantly between patients who experienced a thromboembolic event and those who did not (p = 0.634). There was considerable overlap between the groups. Three of the six patients (50%) with PE had to be lysed or therapeutically anticoagulated and thereafter developed major bleeding (p < 0.001). Patients who experienced TE were more likely to die during hospitalization than those with major bleeding (p = 0.022 vs. p = 1.00). Prophylactic AC within 24 h after surgery does not result in more frequent bleeding. Our data suggests that postoperative intracranial hemorrhage is not caused by prophylactic AC but rather is a surgical complication or the result of antithrombotic therapy. However, thromboembolic events worsen patient outcomes far more than postoperative bleeding. The fact that bleeding may occur as a complication of life-saving lysis therapy in the setting of a thromboembolic event should be included in this cost-benefit consideration.

New trend in artificial intelligence-based assistive technology for thoracic imaging

Although there is no solid agreement for artificial intelligence (AI), it refers to a computer system with intelligence similar to that of humans. Deep learning appeared in 2006, and more than 10 years have passed since the third AI boom was triggered by improvements in computing power, algorithm development, and the use of big data. In recent years, the application and development of AI technology in the medical field have intensified internationally. There is no doubt that AI will be used in clinical practice to assist in diagnostic imaging in the future. In qualitative diagnosis, it is desirable to develop an explainable AI that at least represents the basis of the diagnostic process. However, it must be kept in mind that AI is a physician-assistant system, and the final decision should be made by the physician while understanding the limitations of AI. The aim of this article is to review the application of AI technology in diagnostic imaging from PubMed database while particularly focusing on diagnostic imaging in thorax such as lesion detection and qualitative diagnosis in order to help radiologists and clinicians to become more familiar with AI in thorax.

Multimodal fusion models for pulmonary embolism mortality prediction

Pulmonary embolism (PE) is a common, life threatening cardiovascular emergency. Risk stratification is one of the core principles of acute PE management and determines the choice of diagnostic and therapeutic strategies. In routine clinical practice, clinicians rely on the patient's electronic health record (EHR) to provide a context for their medical imaging interpretation. Most deep learning models for radiology applications only consider pixel-value information without the clinical context. Only a few integrate both clinical and imaging data. In this work, we develop and compare multimodal fusion models that can utilize multimodal data by combining both volumetric pixel data and clinical patient data for automatic risk stratification of PE. Our best performing model is an intermediate fusion model that incorporates both bilinear attention and TabNet, and can be trained in an end-to-end manner. The results show that multimodality boosts performance by up to 14% with an area under the curve (AUC) of 0.96 for assessing PE severity, with a sensitivity of 90% and specificity of 94%, thus pointing to the value of using multimodal data to automatically assess PE severity.

Pulmonary embolism and 529 human blood metabolites: genetic correlation and two-sample Mendelian randomization study

BACKGROUND

The incidence of pulmonary embolism complications in the literature ranges from 10 to 50%, with a 0.5-10% risk of fatal pulmonary embolism. However, the biological cause of pulmonary embolism is unknown.

METHODS

This study used data from the Genome-Wide Association Study (GWAS) of Pulmonary Embolism and Human Blood Metabolites from the UK Biobank, and the data from subjects of European ancestry were analyzed. We explored the relationship between pulmonary embolism and blood metabolites in three ways. We first analyzed the genetic correlation between pulmonary embolism and human blood metabolites using the linkage disequilibrium score regression (LDSC) and then analyzed the causal relationship between pulmonary embolism and meaningful blood metabolites obtained from the LDSC, a procedure for which we used Mendelian randomization analysis. Finally, we obtained transcriptome sequencing data for patients with a pulmonary embolism from the GEO database, analyzed differentially expressed genes (DEGs) in patients with pulmonary embolism versus healthy populations, and compared the DEGs with the resulting blood metabolite genes to further validate the relationship between pulmonary embolism and blood metabolites.

RESULT

We found six human blood metabolites genetically associated with pulmonary embolism, stearic acid glycerol phosphate ethanolamine (correlation coefficient = 0.2582, P = 0.0493), hydroxytryptophan (correlation coefficient = 0.2894, P = 0.0435), and N1-methyladenosine (correlation coefficient = 0.0439, P = 0.3728), and a significant causal relationship was discovered between hydroxytryptophan and pulmonary embolism. After screening microarray data from the GEO database, we performed differential gene analysis on the GSE19151 dataset and screened a total of 22,216 genes with P values less than 0.05, including 17,361 upregulated genes and 4854 downregulated genes. By comparing the resulting differentially expressed genes with six genes encoding blood metabolites, LIPC and NAT2 were found to be differentially expressed in association with pulmonary embolism.

Short- and Long-Term Outcomes of Catheter-Directed Thrombolysis versus Pulmonary Artery Embolectomy in Pulmonary Embolism: A National Population-Based Study

OBJECTIVES

This study aimed to compare the short-term and long-term follow-up outcomes of catheter-directed thrombolysis (CDT) with those of pulmonary artery embolectomy (PAE) for patients with acute pulmonary embolism (PE) included in a nationwide cohort.

BACKGROUND

Data allowing direct comparisons between CDT and PAE are lacking in the literature, and the optimal management of high-risk and intermediate-risk PE is still debated.

METHODS

A retrospective cohort study was conducted with data for 2001 through 2013 collected from the Taiwan National Health Insurance Research Database (NHIRD). Patients who were first admitted for PE and treated with either CDT or PAE were included and compared. In-hospital outcomes included in-hospital death and safety (bleeding and cardiac arrhythmias) outcomes. Follow-up outcomes included all-cause mortality and recurrent PE during the 1- and 2-year follow-up periods and through the last follow-up. Inverse probability of treatment weighting (IPTW) based on the propensity score was used to minimize possible selection bias, including indices for multimorbidity such as the Charlson's Comorbidity Index (CCI) and HAS-BLED scores.

RESULTS

A total of 389 patients treated between January 1, 2001, and December 31, 2013, were identified; 169 underwent CDT and 220 underwent PAE. After IPTW, there were no significant differences in in-hospital mortality (18.2% vs 21.3%; odds ratio 1.07, 95% confidence interval [CI]: 0.70-1.62) or the incidence of safety outcomes between the CDT and PAE groups. The risks of all-cause mortality (30% vs 29.5%; hazard ratio 1.16, 95% CI: 0.89-1.53), recurrent PE (7.2% vs 8.7%; subdistribution hazard ratio [SHR] 0.68, 95% CI: 0.39-1.21) and new-onset pulmonary hypertension (SHR 0.25, 95% CI: 0.05-1.32) were also not significantly different between the CDT and PAE groups at 2 years of follow-up. Subgroup analysis indicated that PAE may be associated with a more favorable 2-year mortality in patients <65 years old, patients with CCI scores of <3, patients with HAS-BLED scores of 1 to 2, and patients without cardiogenic shock (all P for interaction <.05).

CONCLUSIONS

In patients with PE who required reperfusion therapy, CDT and PAE resulted in similar in-hospital and long-term all-cause mortality rates and long-term rates of recurrent PE. Bleeding risks were also comparable in the 2 groups.

Deep learning for pulmonary embolism detection on computed tomography pulmonary angiogram: a systematic review and meta-analysis

Computed tomographic pulmonary angiography (CTPA) is the gold standard for pulmonary embolism (PE) diagnosis. However, this diagnosis is susceptible to misdiagnosis. In this study, we aimed to perform a systematic review of current literature applying deep learning for the diagnosis of PE on CTPA. MEDLINE/PUBMED were searched for studies that reported on the accuracy of deep learning algorithms for PE on CTPA. The risk of bias was evaluated using the QUADAS-2 tool. Pooled sensitivity and specificity were calculated. Summary receiver operating characteristic curves were plotted. Seven studies met our inclusion criteria. A total of 36,847 CTPA studies were analyzed. All studies were retrospective. Five studies provided enough data to calculate summary estimates. The pooled sensitivity and specificity for PE detection were 0.88 (95% CI 0.803-0.927) and 0.86 (95% CI 0.756-0.924), respectively. Most studies had a high risk of bias. Our study suggests that deep learning models can detect PE on CTPA with satisfactory sensitivity and an acceptable number of false positive cases. Yet, these are only preliminary retrospective works, indicating the need for future research to determine the clinical impact of automated PE detection on patient care. Deep learning models are gradually being implemented in hospital systems, and it is important to understand the strengths and limitations of these algorithms.

Impact of renal function on the prognosis of acute pulmonary embolism patients: a systematic review and meta-analysis

OBJECTIVES

We conduct a study to explore the influence of impaired renal function on prognosis in Acute pulmonary embolism (APE) patients.

METHODS

A meta-analysis was performed using the EMBASE and PubMed databases for relevant publications reporting the impact of impaired renal function on the clinical outcomes in patients with APE.

RESULTS

Eventually, 17 articles were included in our analysis. The results suggested that renal insufficiency (RI) is a predictor of poor prognosis in APE patients(short-term mortality: pooled OR = 2.83, 95%CI: 2.20-3.63; long-term mortality: pooled OR = 2.30, 95%CI: 1.72-3.08; adverse outcomes: pooled OR = 3.02, 95%CI: 2.60-3.51). The short-term and long-term mortality rates of APE patients with RI were both higher than those in patients without RI. In addition, acute kidney injury(AKI) could serve as a predictive factor of poor prognosis (pooled OR = 2.75, 95%CI: 2.45-3.08), and it doubles the overall mortality rate in APE patients. However, chronic kidney disease (CKD) did not predict poor prognosis in APE patients (pooled OR = 1.94, 95%CI: 0.99-3.81), although it could slightly increase the overall mortality rate in APE patients.

CONCLUSIONS

RI and AKI could be included in the prognosis evaluation for APE, but the impact of CKD in APE patients has yet to be determined.

Acute pulmonary embolism multimodality imaging prior to endovascular therapy

The manuscript discusses the application of CT pulmonary angiography, ventilation–perfusion scan, and magnetic resonance angiography to detect acute pulmonary embolism and to plan endovascular therapy. CT pulmonary angiography offers high accuracy, speed of acquisition, and widespread availability when applied to acute pulmonary embolism detection. This imaging modality also aids the planning of endovascular therapy by visualizing the number and distribution of emboli, determining ideal intra-procedural catheter position for treatment, and signs of right heart strain. Ventilation–perfusion scan and magnetic resonance angiography with and without contrast enhancement can also aid in the detection and pre-procedural planning of endovascular therapy in patients who are not candidates for CT pulmonary angiography.

Interventional Radiology Therapy: Inferior Vena Cava Filter and Catheter-based Therapies

Endovascular management of pulmonary embolism can be divided into therapeutic and prophylactic treatments. Prophylactic treatment includes inferior vena cava filter placement, whereas endovascular therapeutic interventions include an array of catheter-directed therapies. The indications for both modalities have evolved over the last decade as new evidence has become available.