24 July 2025 in Computer Methods and Programs in Biomedicine
The extraction of head-neck blood vessels from 3D medical images plays a vital role in diagnosing vascular diseases. While many existing methods rely on convolutional neural networks (CNNs), they can encounter challenges in maintaining the continuity of extracted vessels, particularly when segmenting these slender tubular structures within 3D images. This study aims to address these challenges by introducing a novel inference method that preserves vessel continuity taking into consideration the global geometry of vascular structures, while also reducing the number of high resolution patches being processed. The proposed approach employs a two-stage CNN process complimented by a centerline-aware thresholding step. First, a CNN performs preliminary localization of arteries within a highly downsampled volume, identifying seed points. These seed points serve as input for a second CNN, specialized in capturing local artery appearances, to perform precise segmentation. A centerline-based artery tracking algorithm is then applied to guide patch segmentation along the artery until the entire vascular structure is segmented. Physical connectivity is ensured by our novel centerline-aware thresholding strategy which is used to construct the final segmentation mask. The proposed method effectively reduces the number of high-resolution patches processed by the neural network, thereby not only addressing the issue of class imbalance by primarily concentrating on patches containing arteries but also reducing computational complexity. The performance is comparable to state of the art methods across various metrics, while the algorithm additionally recovers missing segments of falsely interrupted arteries, thereby facilitating automatic extraction of medically relevant quantities like for example artery length. The method requires significantly less memory and performs approximately one-tenth of the computations needed to segment a patient. The proposed approach offers an effective solution for extracting blood vessels from 3D medical images, overcoming the limitations of traditional CNN-based methods by preserving vessel continuity and addressing the challenge of class imbalance. This approach enables the selective segmentation of specific arteries in a resource-efficient manner, thereby enhancing the diagnosis and treatment of vascular diseases by providing more accurate vascular segmentations.
19 September 2026 in European Journal of Neurology
Background: The relationship between intrathecal kappa free light chain (κ‐FLC) synthesis and retinal layer atrophy in multiple sclerosis (MS) is unknown. Objective: To investigate whether the κ‐FLC index is associated with peripapillary retinal nerve fiber layer (pRNFL) and ganglion cell plus inner plexiform layer (GCIPL) thickness. Methods: Patients with newly diagnosed clinically isolated syndrome or relapsing MS and available cerebrospinal fluid (CSF) analysis and optical coherence tomography (OCT) were included. Clinical and magnetic resonance imaging (MRI) data were also assessed. κ‐FLC concentrations were measured by nephelometry, and the κ‐FLC index was calculated as (CSF κ‐FLC/serum κ‐FLC)/albumin quotient. Results: A total of 100 patients at a median age of 33 (25th–75th percentile: 25–39) years and a female predominance (58%) were included. In multivariable linear regression analysis adjusted for age, sex, disease duration, number of T2‐hyperintense and number of contrast‐enhancing MRI lesions, the κ‐FLC index was associated with pRNFL (log‐transformed κ‐FLC index: β : −2.51; 95% confidence interval [CI]: −4.01, −1.01; p = 0.001) and GCIPL thinning (log‐transformed κ‐FLC index: β : −1.24, 95% CI: −2.32, −0.17; p = 0.024). Conclusions: Intrathecal plasma cell activity, as measured by the κ‐FLC index, is associated with neuroaxonal damage, as reflected by pRNFL and GCIPL atrophy.
1 January 2026 in International Journal of Biomedical Imaging
Brain volume changes and infratentorial involvement are key predictors of disability in multiple sclerosis (MS) and can be assessed using magnetic resonance imaging (MRI) planimetry. Although MRI planimetry is less susceptible to methodological and patient‐related confounders than volumetry, it currently depends on manual measurements by unblinded experts, an approach that is time‐consuming and vulnerable to bias. In this study, we present a fully automated deep learning framework for deriving brainstem planimetric measurements from MRI. The pipeline integrates an automated midsagittal plane (MSP) detection algorithm with a convolutional neural network trained to perform the segmentations required for planimetry. The automated method shows strong agreement with manual measurements and remains robust across scanners and acquisition protocols. These findings suggest that the proposed framework enables reliable, reproducible, and scalable MRI planimetry, supporting objective assessment of disease progression and treatment response in patients with MS.