Scott Bruce Raymond

Purpose Automated C-arm positioning ensures timely treatment in patients requiring emergent interventions. When a conventional deep learning (DL) approach for C-arm control fails, clinicians must revert to manual operation, resulting in additional delays. Consequently, an agentic C-arm control framework based on multimodal large language models (MLLMs) is highly desirable, as it can incorporate clinician feedback and use reasoning to make adjustments toward more accurate positioning. Skeletal landmark localization is essential for C-arm control, and we investigate adapting MLLMs for autonomous landmark localization. Methods We used an annotated synthetic X-ray dataset and a real X-ray dataset. Each X-ray in both datasets is paired with several skeletal landmarks. We fine-tuned two MLLMs and tasked them with retrieving the closest landmarks from each X-ray. Quantitative evaluations of landmark localization were performed and compared against a leading DL approach. We further conducted qualitative experiments demonstrating: (1) how an MLLM can correct an initially incorrect prediction through reasoning, and (2) how the MLLM can sequentially navigate the C-arm toward a target location. Results On both datasets, fine-tuned MLLMs demonstrate competitive performance across all localization tasks when compared with the DL approach. In the qualitative experiments, the MLLMs provide evidence of reasoning and spatial awareness. Conclusion This study shows that fine-tuned MLLMs achieve accurate skeletal landmark localization and hold promise for agentic autonomous C-arm control. Our code is available at https://github.com/marszzibros/C-arm-localization-LLMs.git.

Accurate and reliable C-arm positioning is essential for fluoroscopy-guided interventions. However, clinical workflows rely on manual alignment that increases radiation exposure and procedural delays. In this work, we present a pipeline that autonomously navigates the C-arm to predefined anatomical landmarks utilizing X-ray images. Given an input X-ray image from an arbitrary starting location on the operating table, the model predicts a 3D displacement vector toward each target landmark along the body. To ensure reliable deployment, we capture both aleatoric and epistemic uncertainties in the model's predictions and further calibrate them using conformal prediction. The derived prediction regions are interpreted as 3D confidence regions around the predicted landmark locations. The training framework combines a probabilistic loss with skeletal pose regularization to encourage anatomically plausible outputs. We validate our approach on a synthetic X-ray dataset generated from DeepDRR. Results show not only strong localization accuracy across multiple architectures but also well-calibrated prediction bounds. These findings highlight the pipeline's potential as a component in safe and reliable autonomous C-arm systems. Code is available at https://github.com/AhmadArrabi/C_arm_guidance_APAH

Accurate MRI-to-CT translation promises the integration of complementary imaging information without the need for additional imaging sessions. Given the practical challenges associated with acquiring paired MRI and CT scans, the development of robust methods capable of leveraging unpaired datasets is essential for advancing the MRI-to-CT translation. Current unpaired MRI-to-CT translation methods, which predominantly rely on cycle consistency and contrastive learning frameworks, frequently encounter challenges in accurately translating anatomical features that are highly discernible on CT but less distinguishable on MRI, such as bone structures. This limitation renders these approaches less suitable for applications in radiation therapy, where precise bone representation is essential for accurate treatment planning. To address this challenge, we propose a path- and bone-contour regularized approach for unpaired MRI-to-CT translation. In our method, MRI and CT images are projected to a shared latent space, where the MRI-to-CT mapping is modeled as a continuous flow governed by neural ordinary differential equations. The optimal mapping is obtained by minimizing the transition path length of the flow. To enhance the accuracy of translated bone structures, we introduce a trainable neural network to generate bone contours from MRI and implement mechanisms to directly and indirectly encourage the model to focus on bone contours and their adjacent regions. Evaluations conducted on three datasets demonstrate that our method outperforms existing unpaired MRI-to-CT translation approaches, achieving lower overall error rates. Moreover, in a downstream bone segmentation task, our approach exhibits superior performance in preserving the fidelity of bone structures. Our code is available at: https://github.com/kennysyp/PaBoT.

A patient presented with acute onset headache and subsequent unconsciousness. The neurologic exam showed left-sided myoclonic jerking and right flaccid hemiparalysis. Noncontrast computed tomography revealed diffuse subarachnoid hemorrhage (SAH) with acute hydrocephalus. Initial digital subtraction angiography (DSA) showed no culprit source for SAH. Repeat DSA on day 7 after initial presentation raised suspicion for left internal carotid artery ophthalmic segment and left lateral lenticulostriate artery (LSA) aneurysms. A magnetic resonance vessel wall imaging (VWI) exam was performed given the presence of multiple potential culprit aneurysms. Vessel wall enhancement around the dome of the left LSA aneurysm suggested rupture, which then facilitated treatment with surgical clipping. LSA aneurysms are exceedingly rare and challenging to treat. Given the associated high degree of morbidity, expedient diagnosis is critical to direct management. VWI could be a valuable tool for detecting ruptured aneurysms in the setting of angiogram-negative SAH.

A stroke volume of arterial blood that arrives to the brain housed in the rigid cranium must be matched over the cardiac cycle by an equivalent volume of ejected venous blood. We hypothesize that the brain maintains this equilibrium by organizing coherent arterial and venous pulse waves. To test this hypothesis, we applied wavelet computational methods to diagnostic cerebral angiograms in four human patients, permitting the capture and analysis of cardiac frequency phenomena from fluoroscopic images acquired at faster than cardiac rate. We found that the cardiac frequency reciprocal phase of a small region of interest (ROI) in a named artery predicts venous anatomy pixel-wise and that the predicted pixels reconstitute venous bolus passage timing. Likewise, a small ROI in a named vein predicts arterial anatomy and arterial bolus passage timing. The predicted arterial and venous pixel groups maintain phase complementarity across the bolus travel. We thus establish a novel computational method to analyze vascular pulse waves from minimally invasive cerebral angiograms and provide the first direct evidence of arteriovenous coupling in the intact human brain. This phenomenon of arteriovenous coupling may be a physiologic mechanism for how the brain precisely maintains mechanical equilibrium against volume displacement and kinetic energy transfer resulting from cyclical deformations with each heartbeat. The study also paves the way to study deranged arteriovenous coupling as an underappreciated pathophysiologic disturbance in a myriad of neurological pathologies linked by mechanical disequilibrium.

Since 2012, the Joint Commission and American Heart Association created the Comprehensive Stroke Center (CSC) certification to identify select institutions capable of treating the most complex strokes. Certification requires sophisticated clinical capabilities including advanced imaging, neurointensive care nursing, and 24/7 coverage by stroke neurologists, neuroradiologists,

Introduction: We describe our initial experience with the React 68 catheter (Medtronic, Dublin, Ireland), an FDA-approved catheter designed for aspiration in cases of emergent large vessel occlusion, as compared with the ACE 68 catheter (Penumbra, Alameda, CA). Methods: This observational study followed consecutive patients treated with the React catheter over a seven-month period at a comprehensive stroke center. Use of the device was per discretion of the operator. Patient demographics, thrombectomy technique, reperfusion scoring, and disposition were assessed. Performance was compared with patients treated with the ACE 68 catheter over a comparable period. Results: We treated 47 patients using the React 68 catheter using either aspiration alone or a combination of aspiration and stent retriever technique. The catheter was used in a variety of circumstances including proximal and distal occlusions involving the anterior and posterior circulation. Modified TICI 2b-3 was achieved in 45 of the 47 patients. The React 68 was comparable to the ACE 68 by all performance measures. Conclusion: The React 68 catheter is a large-bore reperfusion catheter with trackability suitable for use in direct aspiration for recanalization of emergent large vessel occlusion.