Umeå University's logo

Methods: In the multicentre, multidisciplinary PAROKRANK study, 805 patients with a first myocardial infarction (MI) and 805 matched controls without MI were recruited at 17 hospitals in Sweden. At baseline, the participants were examined with panoramic radiography, and in 737 patients and 739 controls, the carotid artery region was assessable. CCAAs were identified in 251 (34%) patients and 205 (28%) controls at baseline. The primary endpoint was defined as the first occurrence of all-cause mortality, non-fatal MI, non-fatal stroke, or hospitalization following heart failure after a mean follow-up of 10 years. The risks of cardiovascular events and death were evaluated using event survival analysis and regression models.

Results: Participants with bilateral CCAAs, regardless of whether they were patients or controls, had significantly higher mortality and morbidity than those without CCAA (P < 0.001). The risk of cardiovascular events was increased in the presence of bilateral CCAAs among both controls (hazard ratio, 1.96 [95% confidence interval : 1.21–3.16], P = 0.006) and patients (1.67 [1.15–2.43], P = 0.007).

Conclusion: Bilateral CCAAs on PRs were an indicator of an increased risk of future cardiovascular events and early death among both controls and patients in the PAROKRANK study. Therefore, dentists can detect CCAA on PR and contribute to identifying individuals in need of medical attention and treatment of cardiovascular disease to prevent morbidity and early death.

Deep learning has been used extensively for medical image analysis applications, assuming the training and test data adhere to the same probability distributions. However, a common challenge arises when dealing with medical images generated by different systems or even the same system with varying parameter settings. Such images often contain diverse textures and noise patterns, violating the assumption. Consequently, models trained on data from one machine or setting usually struggle to perform effectively on data from another. To address this issue in ultrasound images, we proposed a Generative Adversarial Network (GAN) based model in this paper. We formulated image harmonization and denoising tasks as an image-to-image translation task, wherein we adapt the texture pattern and reduced noise in Carotid ultrasound images while keeping the image content (the anatomy) unchanged. The performance was evaluated using feature distribution and pixel-space similarity metrics. In addition, blood-to-tissue contrast and influence on computed risk markers (Grey scale median, GSM) were evaluated. The results showed that domain adaptation was achieved in both tasks (histogram correlation 0.920 (0.043) and 0.844 (0.062)), as compared to no adaptation (0.890 (0.077) and 0.707 (0.098)), and that the anatomy of the images was retained (structure similarity index measure e.g. the arterial wall 0.71 (0.09) and 0.80 (0.08)). In addition, the image noise level (contrast) did not change in the image harmonization task (-34.1 (3.8) vs -35.2 (4.1) dB) but was improved in the noise reduction task (-23.5 (3.2) vs -46.7 (18.1) dB). To validate the performance of the proposed model, we compare its results with CycleGAN, the current state-of-the-art model. Our model outperformed CycleGAN in both tasks. Finally, the risk marker GSM was significantly changed in the noise reduction but not in the image harmonization task. We conclude that domain translation models are powerful tools for improving ultrasound image while retaining the underlying anatomy, but downstream calculations of risk markers may be affected.

Background: The ABO blood group system has shown an association with cardiovascular disease. The susceptibility to CVD is proposed to be partly mediated by dyslipidaemia in non-O individuals. Previous studies are scarce for the RhD blood group, but we recently showed that RhD − young individuals are associated with subclinical atherosclerosis. Hence, we sought to examine whether the ABO blood groups and RhD factor are associated with dyslipidaemia.

Methods: All participants were part of the VIPVIZA study, including 3532 individuals with available plasma lipid levels. Lipids were assessed as total, LDL, HDL, remnant, non-HDL cholesterol and triglycerides. Information about ABO and RhD was retrieved by linking VIPVIZA with the SCANDAT-3 database, where 85% of VIPVIZA participants were registered.

Results: For the ABO blood groups, no significant differences in lipid levels between non-O and O individuals were seen. In 40-year-old males, RhD − individuals compared to RhD + had higher levels of non-HDL cholesterol, LDL cholesterol, and remnant cholesterol, with ratios of geometric means of 1.21 (CI95% 1.03; 1.43), 1.20 (1.02; 1.41) and 1.38 (1.00; 1.92), respectively. No differences in lipid levels depending on the RhD blood group were seen in women or the older age groups.

Conclusion: Our study indicates that younger RhD − men have increased non-HDL, LDL, and remnant cholesterol levels. Thus, the RhD blood group, but not ABO, seems to be associated with dyslipidaemia and may act as a future possible risk marker of cardiovascular disease.

Detecting and segmenting arterial plaques in ultrasound images is essential for the early diagnosis and prevention of cardiovascular diseases. This paper presents Hybrid Stable AnoGAN (HS-AnoGAN), an enhanced anomaly detection framework based on AnoGAN (Anomaly Generative Adversarial Network), which utilizes generative adversarial learning to model normal anatomical structures and identify abnormal regions indicative of pathology. The proposed approach introduces key improvements, including direct latent space encoding, hybrid reconstruction loss, feature matching in the discriminator, and adaptive thresholding, leading to more precise anomaly localization. Additionally, spectral normalization and Wasserstein loss with gradient penalty are incorporated to improve training stability and prevent mode collapse. To the best of our knowledge, this is the first attempt to apply anomaly detection techniques for arterial plaque detection and segmentation in ultrasound images. Comparative experiments demonstrate that HS-AnoGAN outperforms state-of-the-art methods, achieving a 9.8% increase in detection accuracy, and a 7.5% enhancement in Dice score for segmentation quality. These results highlight the effectiveness of HS-AnoGAN in improving both plaque detection and segmentation in ultrasound imaging, making it a promising tool for clinical applications.

Objective: To test the hypothesis of low emotional support being associated with lifestyle and biomedical cardiovascular disease (CVD) risk factors, estimated risk of CVD morbidity and mortality, and subclinical atherosclerosis in middle-aged healthy adults.

Methods and measures: Cross-sectional data were obtained from participants aged 40–60 years who had one or more conventional CVD risk factor. They underwent assessment based on questionnaires, clinical examination, blood sampling, and carotid ultrasound of plaque formation and carotid intima-media wall thickness (cIMT). Based on the Interview Schedule for Social Interaction, the participants were categorised as either low in emotional support (n = 884) or as a referent (n = 2570). Logistic regression analyses were conducted to study the associations.

Results: Logistic regression analyses showed that low emotional support was significantly associated with smoking, alcohol consumption and physical inactivity (OR = 1.53 − 1.94), estimated risk of CVD morbidity and mortality (OR = 1.56 − 1.68), and plaque formation (OR = 1.39). No significant associations were found regarding biomedical CVD risk factors or cIMT.

Conclusion: The findings suggest that low social support is associated with lifestyle CVD risk factors, estimated risk of CVD morbidity and mortality, and subclinical atherosclerosis in middle-aged healthy adults, encouraging causal evaluation with longitudinal data investigating an impact of emotional support on mechanisms underlying CVD.

Atherosclerosis is a leading cardiovascular disease typified by fatty streaks accumulating within arterial walls, culminating in potential plaque ruptures and subsequent strokes. Existing clinical risk scores, such as systematic coronary risk estimation and Framingham risk score, profile cardiovascular risks based on factors like age, cholesterol, and smoking, among others. However, these scores display limited sensitivity in early disease detection. Parallelly, ultrasound-based risk markers, such as the carotid intima media thickness, while informative, only offer limited predictive power. Notably, current models largely focus on either ultrasound image-derived risk markers or clinical risk factor data without combining both for a comprehensive, multimodal assessment. This study introduces a multimodal ensemble learning framework to assess atherosclerosis severity, especially in its early sub-clinical stage. We utilize a multi-objective optimization targeting both performance and diversity, aiming to integrate features from each modality effectively. Our objective is to measure the efficacy of models using multimodal data in assessing vascular aging, i.e., plaque presence and vascular age, over a six-year period. We also delineate a procedure for optimal model selection from a vast pool, focusing on best-suited models for classification tasks. Additionally, through eXplainable Artificial Intelligence techniques, this work delves into understanding key model contributors and discerning unique subject subgroups.

Background/Introduction: Cardiovascular diseases (CVDs) are the leading cause of death worldwide, with atherosclerosis being a major contributor. Early detection of carotid artery plaques is crucial for CVD prevention and management. While ultrasound imaging is widely used, its manual analysis is time-consuming and prone to variability. Automating plaque detection with deep learning can improve diagnostic consistency and efficiency. Recent advancements in unsupervised anomaly detection models offer a promising approach to detect plaque-related anomalies in carotid ultrasound images without labeled data. This study is, to the best of our knowledge, the first to apply these models for plaque detection in ultrasound images.

Purpose: This work aims to evaluate the effectiveness of three unsupervised anomaly detection methods—Autoencoder (AE), Variational Autoencoder (VAE), and Anomaly Detection Generative Adversarial Networks (AnoGAN)—for detecting plaques in longitudinal carotid artery ultrasound images. We used carotid ultrasound images from a large Randomized Controlled Trial (RCT) on subclinical atherosclerosis, comprising images from a diverse group of participants collected at both baseline and follow-up periods.

Methods: The generative models were trained to recognize normal anatomical structures using 3,000 plaque-free (normal) images. These models were then tested on 400 images, which included both normal and abnormal images, to distinguish them and detect anomalies in abnormal images. As shown in Figure 1, the anomaly scores were computed by comparing the original images to the generated ones, with a distribution-based cutoff applied to classify the images as normal or abnormal. The threshold for this cutoff was determined based on the distribution of the anomaly scores.

Results: AnoGAN achieved the highest accuracy, as shown in Figure 2, which compares the input image and generated output of AnoGAN, VAE, and AE for an abnormal image. Also, AnoGAN achieved the highest Area Under the Curve of the Receiver Operating Characteristic (AUC-ROC) at 91%, followed by VAE at 79% and AE at 66%. Additionally, AnoGAN achieved the best sensitivity-specificity balance, resulting in the highest AUC-ROC, while VAE demonstrated moderate performance, and AE, despite lower sensitivity, maintained competitive specificity.

Conclusion(s): This study shows the effectiveness of anomaly detection models, especially AnoGAN, in automating the detection of plaques in carotid ultrasound imaging. These models have the potential to reduce diagnostic variability, improve early detection of CVD risk factors, and enhance clinical efficiency. Future research should aim to expand the use of these models to larger datasets and explore their clinical integration to assess real-world applicability. Additionally, optimizing the computational performance and interpretability of these models will be crucial for their widespread adoption in routine clinical settings.

Background and Aim: The study ‘Periodontitis and Its Relation to Coronary Artery Disease’ (PAROKRANK) reported an association between periodontitis (PD) and the first myocardial infarction (MI). This follow-up study aims to test the hypothesis that those with PD—compared to periodontally healthy individuals—are at increased risk for cardiovascular (CV) events and death.

Methods: A total of 1587 participants (age <75 years; females 19%) had a dental examination including panoramic radiographs between 2010 and 2014. PD was categorized as healthy (≥80% alveolar bone height), mild/moderate (79%–66%) or severe (<66%). A composite CV event (first of all-cause death, non-fatal MI or stroke and hospitalization following to heart failure) was investigated during a mean follow-up period of 9.9 years (range 0.2–12.5 years). Participants were divided into two groups: those with and without PD. The primary event rate, stratified by periodontal status at baseline, was calculated using the Kaplan–Meier method and Cox regression.

Results: The number of events was 187 in the 985 periodontally healthy participants (19%) and 174 in the 602 participants with PD (29%; p < 0.0001). Those with PD had a higher likelihood for a future event (hazard ratio [HR] = 1.26; 95% CI: 1.01–1.57; p = 0.038), following adjustment for age, smoking and diabetes.

Conclusion: The PAROKRANK follow-up revealed that CV events were more common among participants with PD, which supports the assumption that there might be a direct relation between PD and CV disease.

Objectives: We examined the magnitude of transcription errors in lipid variables in the VIPVIZA study and assessed whether education among the research personnel reduced the error frequency at follow-up. We also examined how the errors affected the SCORE2 risk prediction algorithm for cardiovascular disease, which includes lipid parameters, as this could lead to an incorrect treatment decision.

Methods: The VIPVIZA study includes assessment of lipid parameters, where results for total cholesterol, triglycerides, HDL cholesterol, and calculated LDL cholesterol are transcribed into the research database by research nurses. Transcription errors were identified by recalculating LDL cholesterol, and a difference>0.15 indicated a transcription error in any of the four lipid parameters. To assess the presence of risk category misclassification, we compared the individual's SCORE2 risk category based on incorrect lipid levels to the SCORE2 categories based on the correct lipid levels.

Results: The transcription error frequency was 0.55 % in the 2019 VIPVIZA research database and halved after the educational intervention to 0.25 % in 2023. Of the 39 individuals who had a transcription error in total or HDL cholesterol (with the possibility of affecting the SCORE2 risk category based on non-HDL cholesterol), six individuals (15 %) received an incorrect risk category due to the error.

Conclusions: Transcription errors persist despite digitalisation improvements. It is essential to minimise transcriptions in fields outside the laboratory environment, as we observed that critical decisions also rely on accurate information such as the SCORE2-risk algorithm, which is dependent on lab results but not necessarily reported by the laboratory.

The silent progression of cardiovascular disease (CVD) is a major problem particularly in CVD prevention. New techniques enabled by the rise of electronic health records may facilitate CVD prevention. Both public health research and big data applications, such as digital twins, are dependent on access to longitudinal and sensitive data; a challenge which may be facilitated by access to longitudinal synthetic data. In this study, we establish a fidelity benchmark for longitudinal synthetic data by extending a well-known method for cross-sectional synthetic data to a longitudinal application within CVD. We find that the univariate distributional difference between the real and the synthetic data is kept low and that pairwise relations are preserved in the synthetic data. Further, we see that the variable-wise temporal trends are preserved, yet may be more extensively studied and have some room for improvement. The results of this study is important to enable future studies within public health prevention and cardiovascular digital twins.