Improved detection of small pulmonary embolism on unenhanced computed tomography using an artificial intelligence-based algorithm - a single centre retrospective study.

To preliminarily verify the feasibility of a deep-learning (DL) artificial intelligence (AI) model to localize pulmonary embolism (PE) on unenhanced chest-CT by comparison with pulmonary artery (PA) CT angiography (CTA). In a monocentric study, we retrospectively reviewed 99 oncological patients (median age in years: 64 (range: 28-92 years); percentage of female: 39.4%) who received unenhanced and contrast-enhanced chest CT examinations in one session between January 2020 and October 2022 and who were diagnosed incidentally with PE. Findings in the unenhanced images were correlated with the contrast-enhanced images, which were considered the gold standard for central, segmental and subsegmental PE. The new algorithm was trained and tested based on the 99 unenhanced chest-CT image data sets. Based on them, candidate boxes, which were output by the model, were post-processed by evaluating whether the predicted box intersects with the patient's lung segmentation at any position. The AI-based algorithm proved to have an overall sensitivity of 54.5% for central, of 81.9% for segmental and 80.0% for subsegmental PE if taking n = 20 candidate boxes into account. Depending on the localization of the pulmonary embolism, the detection rate for only one box was: 18.1% central, 34.7% segmental and 0.0% subsegmental. The median volume of the clots differed significantly between the three subgroups and was 846.5 mm3 (IQR:591.1-964.8) in central, 201.3 mm3 (IQR:98.3-390.9) in segmental and 110.6 mm3 (IQR:94.3-128.0) in subsegmental PA (p < 0.05). The new algorithm proved to have high sensitivity in detecting PE in particular in segmental/subsegmental localization and may guide to decide whether a second contrast enhanced CT is necessary.

Uncertainty-aware deep learning for trustworthy prediction of long-term outcome after endovascular thrombectomy.

Acute ischemic stroke (AIS) is a leading global cause of mortality and morbidity. Improving long-term outcome predictions after thrombectomy can enhance treatment quality by supporting clinical decision-making. With the advent of interpretable deep learning methods in recent years, it is now possible to develop trustworthy, high-performing prediction models. This study introduces an uncertainty-aware, graph deep learning model that predicts endovascular thrombectomy outcomes using clinical features and imaging biomarkers. The model targets long-term functional outcomes, defined by the three-month modified Rankin Score (mRS), and mortality rates. A sample of 220 AIS patients in the anterior circulation who underwent endovascular thrombectomy (EVT) was included, with 81 (37%) demonstrating good outcomes (mRS ≤ 2). The performance of the different algorithms evaluated was comparable, with the maximum validation under the curve (AUC) reaching 0.87 using graph convolutional networks (GCN) for mRS prediction and 0.86 using fully connected networks (FCN) for mortality prediction. Moderate performance was obtained at admission (AUC of 0.76 using GCN), which improved to 0.84 post-thrombectomy and to 0.89 a day after stroke. Reliable uncertainty prediction of the model could be demonstrated.

Unsupervised Deep Learning for Advanced Forming Limit Analysis in Sheet Metal: A Tensile Test-Based Approach.

An accurate description of the formability and failure behavior of sheet metal materials is essential for an optimal forming process design. In this respect, the forming limit curve (FLC) based on the Nakajima test, which is determined in accordance with DIN EN ISO 12004-2, is a wide-spread procedure for evaluating the formability of sheet metal materials. Thereby the FLC is affected by influences originating from intrinsic factors of the Nakajima test-setup, such as friction, which leads to deviations from the linear strain path, biaxial prestress and bending superposition. These disadvantages can be circumvented by an alternative test combination of uniaxial tensile test and hydraulic bulge test. In addition, the forming limit capacity of many lightweight materials is underestimated using the cross-section method according to DIN EN ISO 12004-2, due to the material-dependent occurrence of multiple strain maxima during forming or sudden cracking without prior necking. In this regard, machine learning approaches have a high potential for a more accurate determination of the forming limit curve due to the inclusion of other parameters influencing formability. This work presents a machine learning approach focused on uniaxial tensile tests to define the forming limit of lightweight materials and high-strength steels. The transferability of an existing weakly supervised convolutional neural network (CNN) approach was examined, originally designed for Nakajima tests, to uniaxial tensile tests. Additionally, a stereo camera-based method for this purpose was developed. In our evaluation, we train and test materials, including AA6016, DX54D, and DP800, through iterative data composition, using cross-validation. In the context of our stereo camera-based approach, strains for different materials and thicknesses were predicted. In this cases, our method successfully predicted the major strains with close agreement to ISO standards. For DX54D, with a thickness of 0.8 mm, the prediction was 0.659 (compared to ISO's 0.664). Similarly, for DX54D, 2.0 mm thickness, the predicted major strain was 0.780 (compared to ISO 0.705), and for AA6016, at 1.0 mm thickness, a major strain of 0.314 (in line with ISO 0.309) was estimated. However, for DP800 with a thickness of 1.0 mm, the prediction yielded a major strain of 0.478 (as compared to ISO 0.289), indicating a divergence from the ISO standard in this particular case. These results in general, generated with the CNN stereo camera-based approach, underline the quantitative alignment of the approach with the cross-section method.

Bifurcation matching for consistent cerebral vessel labeling in CTA of stroke patients.

PURPOSE: Vessel labeling is a prerequisite for comparing cerebral vasculature across patients, e.g., for straightened vessel examination or for localization. Extracting vessels from computed tomography angiography scans may come with a trade-off in segmentation accuracy. Vessels might be neglected or artificially created, increasing the difficulty of labeling. Related work mainly focuses on magnetic resonance angiography without stroke and uses trainable approaches requiring costly labels. METHODS: We present a robust method to identify major arteries and bifurcations in cerebrovascular models generated from existing segmentations. To localize bifurcations of the Circle of Willis, candidate paths for the adjacent vessels of interest are identified using registered landmarks. From those paths, the optimal ones are extracted by recursively maximizing an objective function for all adjacent vessels starting from a bifurcation to avoid erroneous paths and compensate for stroke. RESULTS: In 100 CTA stroke data sets for evaluation, 6 bifurcation locations are placed correctly in 85% of cases; 92.5% when allowing a margin of 5 mm. On average, 14 vessels of interest are found in 90% of the cases and traced correctly end-to-end in 73.5%. The baseline achieves similar detection rates but only 35.5% of the arteries are traced in full. CONCLUSION: Formulating the vessel labeling process as a maximization task for bifurcation matching can vastly improve accurate vessel tracing. The proposed algorithm only uses simple features and does not require expensive training data.

An algorithm for the labeling and interactive visualization of the cerebrovascular system of ischemic strokes.

During the diagnosis of ischemic strokes, the Circle of Willis and its surrounding vessels are the arteries of interest. Their visualization in case of an acute stroke is often enabled by Computed Tomography Angiography (CTA). Still, the identification and analysis of the cerebral arteries remain time consuming in such scans due to a large number of peripheral vessels which may disturb the visual impression. We propose VirtualDSA++, an algorithm designed to segment and label the cerebrovascular tree on CTA scans. Especially with stroke patients, labeling is a delicate procedure, as in the worst case whole hemispheres may not be present due to impeded perfusion. Hence, we extended the labeling mechanism for the cerebral arteries to identify occluded vessels. In the work at hand, we place the algorithm in a clinical context by evaluating the labeling and occlusion detection on stroke patients, where we have achieved labeling sensitivities comparable to other works between 92% and 95%. To the best of our knowledge, ours is the first work to address labeling and occlusion detection at once, whereby a sensitivity of 67% and a specificity of 81% were obtained for the latter. VirtualDSA++ also automatically segments and models the intracranial system leading to further processing possibilities. We present the generic concept of iterative systematic search for pathways on all nodes of said model, which enables new interactive features. Exemplary, we derive in detail, firstly, the interactive planning of vascular interventions like the mechanical thrombectomy and secondly, the interactive suppression of vessel structures that are not of interest in diagnosing strokes (like veins). We discuss both features as well as further possibilities emerging from the proposed concept.

Magnetic Resonance Fingerprinting Reconstruction Using Recurrent Neural Networks.

Magnetic Resonance Fingerprinting (MRF) is an imaging technique acquiring unique time signals for different tissues. Although the acquisition is highly accelerated, the reconstruction time remains a problem, as the state-of-the-art template matching compares every signal with a set of possible signals. To overcome this limitation, deep learning based approaches, e.g. Convolutional Neural Networks (CNNs) have been proposed. In this work, we investigate the applicability of Recurrent Neural Networks (RNNs) for this reconstruction problem, as the signals are correlated in time. Compared to previous methods based on CNNs, RNN models yield significantly improved results using in-vivo data.