CT Angiography Radiomics Combining Traditional Risk Factors to Predict Brain Arteriovenous Malformation Rupture: a Machine Learning, Multicenter Study

The current landscape of machine learning-based radiomics in arteriovenous malformations: a systematic review and radiomics quality score assessment

Background

Arteriovenous malformations (AVMs) are rare vascular anomalies involving a disorganization of arteries and veins with no intervening capillaries. In the past 10 years, radiomics and machine learning (ML) models became increasingly popular for analyzing diagnostic medical images. The goal of this review was to provide a comprehensive summary of current radiomic models being employed for the diagnostic, therapeutic, prognostic, and predictive outcomes in AVM management.

Methods

A systematic literature review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines, in which the PubMed and Embase databases were searched using the following terms: (cerebral OR brain OR intracranial OR central nervous system OR spine OR spinal) AND (AVM OR arteriovenous malformation OR arteriovenous malformations) AND (radiomics OR radiogenomics OR machine learning OR artificial intelligence OR deep learning OR computer-aided detection OR computer-aided prediction OR computer-aided treatment decision). A radiomics quality score (RQS) was calculated for all included studies.

Results

Thirteen studies were included, which were all retrospective in nature. Three studies (23%) dealt with AVM diagnosis and grading, 1 study (8%) gauged treatment response, 8 (62%) predicted outcomes, and the last one (8%) addressed prognosis. No radiomics model had undergone external validation. The mean RQS was 15.92 (range: 10-18).

Conclusion

We demonstrated that radiomics is currently being studied in different facets of AVM management. While not ready for clinical use, radiomics is a rapidly emerging field expected to play a significant future role in medical imaging. More prospective studies are warranted to determine the role of radiomics in the diagnosis, prediction of comorbidities, and treatment selection in AVM management.

Stacking learning based on micro-CT radiomics for outcome prediction in the early-stage of silica-induced pulmonary fibrosis model

Silicosis is a progressive pulmonary fibrosis disease caused by long-term inhalation of silica. The early diagnosis and timely implementation of intervention measures are crucial in preventing silicosis deterioration further. However, the lack of screening and diagnostic measures for early-stage silicosis remains a significant challenge. In this study, silicosis models of varying severity were established through a single exposure to silica with different doses (2.5mg/mice or 5mg/mice) and durations (4 weeks or 12 weeks). The diagnostic performance of computed tomography (CT) quantitative analysis was assessed using lung density biomarkers and the lung density distribution histogram, with a particular focus on non-aerated lung volume. Subsequently, we developed and evaluated a stacking learning model for early diagnosis of silicosis after extracting and selecting features from CT images. The CT quantitative analysis reveals that while the lung densitometric biomarkers and lung density distribution histogram, as traditional indicators, effectively differentiate severe fibrosis models, they are unable to distinguish early-stage silicosis. Furthermore, these findings remained consistent even when employing non-aerated areas, which is a more sensitive indicator. By establishing a radiomics stacking learning model based on non-aerated areas, we can achieve remarkable diagnostic performance to distinguish early-stage silicosis, which can provide a valuable tool for clinical assistant diagnosis. This study reveals the potential of using non-aerated lung areas as a region of interest in stacking learning for early diagnosis of silicosis, providing new insights into early detection of this disease.

Current and Future Applications of Arterial Spin Labeling MRI in Cerebral Arteriovenous Malformations

Arterial spin labeling (ASL) has emerged as a promising noninvasive tool for the evaluation of both pediatric and adult arteriovenous malformations (AVMs). This paper reviews the advantages and challenges associated with the use of ASL in AVM assessment. An assessment of the diagnostic workup of AVMs and their variants in both adult and pediatric populations is proposed. Evaluation after treatments, whether endovascular or microsurgical, was similarly examined. ASL, with its endogenous tracer and favorable safety profile, offers functional assessment and arterial feeder identification. ASL has demonstrated strong performance in identifying feeder arteries and detecting arteriovenous shunting, although some studies report inferior performance compared with digital subtraction angiography (DSA) in delineating venous drainage. Challenges include uncertainties in sensitivity for specific AVM features. Detecting AVMs in challenging locations, such as the apical cranial convexity, is further complicated, demanding careful consideration due to the risk of underestimating total blood flow. Navigating these challenges, ASL provides a noninvasive avenue with undeniable merits, but a balanced approach considering its limitations is crucial. Larger-scale prospective studies are needed to comprehensively evaluate the diagnostic performance of ASL in AVM assessment.

Overloaded transnidal pressure gradient as the hemodynamic mechanism leading to arteriovenous malformation rupture: a quantitative analysis using intravascular pressure monitoring and color-coded digital subtraction angiography

BACKGROUND

The hemodynamics of brain arteriovenous malformations (AVMs) may have implications for hemorrhage. This study aimed to explore the hemodynamics of ruptured AVMs by direct microcatheter intravascular pressure monitoring (MIPM) and indirect quantitative digital subtraction angiography (QDSA).

METHODS

We recruited patients with AVMs at a tertiary neurosurgery center from October 2020 to March 2023. In terms of MIPM, we preoperatively super-selected a predominant feeding artery and main draining vein through angiography to measure intravascular pressure before embolization. In processing of QDSA, we adopted previously standardized procedure for quantitative hemodynamics analysis of pre-embolization digital subtraction angiography (DSA), encompassing main feeding artery, nidus, and the main draining vein. Subsequently, we investigated the correlation between AVM rupture and intravascular pressure from MIPM, as well as hemodynamic parameters derived from QDSA. Additionally, we explored the interrelationships between hemodynamic indicators in both dimensions.

RESULTS

After strict screening of patients, our study included 10 AVMs (six ruptured and four unruptured). We found that higher transnidal pressure gradient (TPG) (53.00±6.36 vs 39.25±8.96 mmHg, p=0.042), higher feeding artery pressure (FAP) (72.83±5.46 vs 65.00±6.48 mmHg, p=0.031) and higher stasis index of nidus (3.54±0.73 vs 2.43±0.70, p=0.043) were significantly correlated with AVM rupture. In analysis of interrelationships between hemodynamic indicators in both dimensions, a strongly positive correlation (r=0.681, p=0.030) existed between TPG and stasis index of nidus.

CONCLUSIONS

TPG and FAP from MIPM platform and nidus stasis index from QDSA platform were correlated with AVM rupture, and both were positively correlated, suggesting that higher pressure load within nidus may be the central mechanism leading to AVM rupture.

Increased Collagen I/Collagen III Ratio Is Associated with Hemorrhage in Brain Arteriovenous Malformations in Human and Mouse

Background: The increase in the collagen I (COL I)/COL III ratio enhances vessel wall stiffness and renders vessels less resistant to blood flow and pressure changes. Activated microglia enhance inflammation-induced fibrosis. Hypotheses: The COL I/COL III ratio in human and mouse brain arteriovenous malformations (bAVMs) is associated with bAVM hemorrhage, and the depletion of microglia decreases the COL I/COL III ratio and hemorrhage. Method: COL I, COL III, and hemorrhages were analyzed in 12 human bAVMs and 6 control brains, and mouse bAVMs induced in three mouse lines with activin receptor-like kinase 1 (n = 7) or endoglin (n = 7) deleted in the endothelial cells or brain focally (n = 5). The controls for the mouse study were no-gene-deleted litter mates. Mouse bAVMs were used to test the relationships between the Col I/Col III ratio and hemorrhage and whether the transient depletion of microglia reduces the Col I/Col III ratio and hemorrhage. Results: The COL I/COL III ratio was higher in the human and mouse bAVMs than in controls. The microhemorrhage in mouse bAVMs was positively correlated with the Col I/Col III ratio. Transient depletion of microglia reduced the Col I/Col III ratio and microhemorrhage. Conclusions: The COL I/COL III ratio in the bAVMs was associated with bAVM hemorrhage. The depletion of microglia reduced the bAVM Col I/Col III ratio and hemorrhage.