Intracranial aneurysms are a major cause of morbidity and mortality worldwide, and detecting them manually is a complex, time-consuming task. Albeit automated solutions are desirable, the limited availability of training data makes it difficult to develop such solutions using typical supervised learning frameworks. In this work, we propose a novel pre-training strategy using more widely available unannotated head CT scan data to pre-train a 3D Vision Transformer model prior to fine-tuning for the aneurysm detection task. Specifically, we modify masked auto-encoder (MAE) pre-training in the following ways: we use a factorized self-attention mechanism to make 3D attention computationally viable, we restrict the masked patches to areas near arteries to focus on areas where aneurysms are likely to occur, and we reconstruct not only CT scan intensity values but also artery distance maps, which describe the distance between each voxel and the closest artery, thereby enhancing the backbone's learned representations. Compared with SOTA aneurysm detection models, our approach gains +2-7% absolute Sensitivity at a false positive rate of 0.5. Code and weights are available at: https://github.com/alceballosa/aneurysm-mae.

a confident diagnosis can be made. We use an LLM to retrieve an initial pool of evidence, but then refine this set of evidence according to correlations learned by the model. We conduct an in-depth evaluation of the usefulness of our approach by simulating how it might be used by a clinician to decide between a pre-defined list of differential diagnoses.

Pretraining multimodal models on Electronic Health Records (EHRs) provides a means of learning representations that can transfer to downstream tasks with minimal supervision. Recent multimodal models induce soft local alignments between image regions and sentences. This is of particular interest in the medical domain, where alignments might highlight regions in an image relevant to specific phenomena described in free-text. While past work has suggested that attention "heatmaps" can be interpreted in this manner, there has been little evaluation of such alignments. We compare alignments from a state-of-the-art multimodal (image and text) model for EHR with human annotations that link image regions to sentences. Our main finding is that the text has an often weak or unintuitive influence on attention; alignments do not consistently reflect basic anatomical information. Moreover, synthetic modifications - such as substituting "left" for "right" - do not substantially influence highlights. Simple techniques such as allowing the model to opt out of attending to the image and few-shot finetuning show promise in terms of their ability to improve alignments with very little or no supervision. We make our code and checkpoints open-source.

Blood circulation in the human brain is supplied through a network of cerebral arteries. If a clinician suspects a patient has a stroke or other cerebrovascular condition, they order imaging tests. Neuroradiologists visually search the resulting scans for abnormalities. Their visual search tasks correspond to the abstract network analysis tasks of browsing and path following. To assist neuroradiologists in identifying cerebral artery abnormalities, we designed CerebroVis, a novel abstract-yet spatially contextualized-cerebral artery network visualization. In this design study, we contribute a novel framing and definition of the cerebral artery system in terms of network theory and characterize neuroradiologist domain goals as abstract visualization and network analysis tasks. Through an iterative, user-centered design process we developed an abstract network layout technique which incorporates cerebral artery spatial context. The abstract visualization enables increased domain task performance over 3D geometry representations, while including spatial context helps preserve the user's mental map of the underlying geometry. We provide open source implementations of our network layout technique and prototype cerebral artery visualization tool. We demonstrate the robustness of our technique by successfully laying out 61 open source brain scans. We evaluate the effectiveness of our layout through a mixed methods study with three neuroradiologists. In a formative controlled experiment our study participants used CerebroVis and a conventional 3D visualization to examine real cerebral artery imaging data to identify a simulated intracranial artery stenosis. Participants were more accurate at identifying stenoses using CerebroVis (absolute risk difference 13%). A free copy of this paper, the evaluation stimuli and data, and source code are available at osf.io/e5sxt.

In many applications, image deblurring is a pre-requisite to improve the sharpness of an image before it can be further processed. Iterative methods are widely used for deblurring images but care must be taken to ensure that the iterative process is robust, meaning that the process does not diverge and reaches the solution reasonably fast, two goals that sometimes compete against each other. In practice, it remains challenging to choose parameters for the iterative process to be robust. We propose a new approach consisting of relaxed initialization and pixel-wise updates of the step size for iterative methods to achieve robustness. The first novel design of the approach is to modify the initialization of existing iterative methods to stop a noise term from being propagated throughout the iterative process. The second novel design is the introduction of a vectorized step size that is adaptively determined through the iteration to achieve higher stability and accuracy in the whole iterative process. The vectorized step size aims to update each pixel of an image individually, instead of updating all the pixels by the same factor. In this work, we implemented the above designs based on the Landweber method to test and demonstrate the new approach. Test results showed that the new approach can deblur images from noisy observations and achieve a low mean squared error with a more robust performance.