Search

In real-world scenarios, medical image segmentation models encounter input images that may deviate from the training images in various ways. These differences can arise from changes in image scanners and acquisition protocols, or even the images can come from a different modality or domain. When the model encounters these out-of-distribution (OOD) images, it can behave unpredictably. Therefore, it is important to develop a system that handles such out-of-distribution images to ensure the safe usage of the models in clinical practice. In this paper, we propose a post-hoc out-of-distribution (OOD) detection method that can be used with any pre-trained segmentation model. Our method utilizes multi-scale representations extracted from the encoder blocks of the segmentation model and employs Mahalanobis distance as a metric to measure the similarity between the input image and the in-distribution images. The segmentation model is pre-trained on a publicly available cardiac short-axis cine MRI dataset. The detection performance of the proposed method is evaluated on 13 different OOD datasets, which can be categorized as near, mild, and far OOD datasets based on their similarity to the in-distribution dataset. The results show that our method outperforms state-of-the-art feature space-based and uncertainty-based OOD detection methods across the various OOD datasets. Our method successfully detects near, mild, and far OOD images with high detection accuracy, showcasing the advantage of using the multi-scale and semantically rich representations of the encoder. In addition to the feature-based approach, we also propose a Dice coefficient-based OOD detection method, which demonstrates superior performance for adversarial OOD detection and shows a high correlation with segmentation quality. For the uncertainty-based method, despite having a strong correlation with the quality of the segmentation results in the near OOD datasets, they failed to detect mild and far OOD images, indicating the weakness of these methods when the images are more dissimilar. Future work will explore combining Mahalanobis distance and uncertainty scores for improved detection of challenging OOD images that are difficult to segment. • Introduces a post-hoc OOD detection method for any pre-trained segmentation model. • Uses multi-scale encoder features and Mahalanobis distance to detect OOD images. • Outperforms state-of-the-art feature space and uncertainty-based OOD detection methods. • Proposes a Dice-based OOD method, excelling in adversarial OOD detection and quality correlation. [ABSTRACT FROM AUTHOR]

People living with HIV are at higher risk of heart failure and associated left atrial remodeling compared to people without HIV. Mechanisms are unclear but have been linked to inflammation and premature aging. Here we obtain plasma proteomics concurrently with cardiac magnetic resonance imaging in two independent study populations to identify parallels between HIV-related and aging-related immune dysfunction that could contribute to atrial remodeling and clinical heart failure. We discover a plasma proteomic signature that may in part reflect or contribute to HIV-associated atrial remodeling, many features of which are associated with older age and time to incident heart failure among an independent community-based cohort without HIV. This proteomic profile was statistically enriched for immune checkpoint proteins, tumor necrosis factor signaling, ephrin signaling, and extracellular matrix organization, identifying possible shared pathways in HIV and aging that may contribute to risk of heart failure. Mechanisms underlying excess risk of heart failure among people with HIV are unclear. Here, the authors identify a proteomic signature enriched in immune activation and extracellular matrix remodeling that may reflect shared pathways in HIV and aging that contribute to risk of heart failure. [ABSTRACT FROM AUTHOR]

Myocyte disarray and fibrosis are underlying pathologies of hypertrophic cardiomyopathy (HCM) caused by genetic mutations. However, the extent of their contributions has not been extensively evaluated. In this study, we investigated the effects of genetic mutations on myofiber function and fibrosis patterns in HCM. A total of 133 patients with HCM underwent chamber geometry, late gadolinium enhancement (LGE), and T1-mapping evaluation using 1.5T cardiac magnetic resonance (CMR) imaging, echo-derived diastolic function analyses, and genetic testing. Left ventricular (LV) segmental and global longitudinal strain (LS), circumferential strain (CS), and rotation were measured using feature tracking analysis. Patients with sarcomere-associated mutation (SM, n = 41) exhibited lower LV-CS (all three slices) and higher basal rotationendo, along with a higher prevalence of midepicardial LGE. The relationship between SM and LV-CSmyo was independent of LGE amount (ß = 0.239, p = 0.008). However, global LS and E/e' were not correlated with SM but were associated with LV mass index and LGE extent. SM was significantly correlated with the presence of midepicardial LGE (odds ratio 5.81, 95% confidence interval 2.15–15.72, p = 0.001), independent of LV mass index, hypertrophy pattern and E/e'. Augmented LV basal segmental rotation was significantly associated with dynamic obstruction. Circumferential fiber dysfunction and midepicardial fibrosis were related to SM, independent of the extent of LV hypertrophy. However, longitudinal fiber function was correlated to the extent of hypertrophy and fibrosis, regardless of SM. Subendocardial fibrosis did not show a significant association with SM. [ABSTRACT FROM AUTHOR]

The editorial discusses sex differences in right ventricular function, particularly in the context of pulmonary arterial hypertension (PAH). Females generally have smaller right ventricular volumes and higher ejection fraction, leading to better survival rates in PAH. The study by Harbaum and colleagues explores genetic factors, identifying BMPR1A and DMRT2 as candidate gene loci associated with RV adaptation. While the study sheds light on potential therapeutic targets for RV dysfunction, limitations include the focus on load-dependent measures and the overrepresentation of females in the UK Biobank cohort, which may limit generalizability. The findings suggest exciting new avenues for research and potential sex-specific treatments for PAH patients. [Extracted from the article]