Purpose: The purpose of this study was to determine if an ensemble of multiple LLM agents could be used collectively to provide a more reliable assessment of a pixel-based AI triage tool than a single LLM. Methods: 29,766 non-contrast CT head exams from fourteen hospitals were processed by a commercial intracranial hemorrhage (ICH) AI detection tool. Radiology reports were analyzed by an ensemble of eight open-source LLM models and a HIPAA compliant internal version of GPT-4o using a single multi-shot prompt that assessed for presence of ICH. 1,726 examples were manually reviewed. Performance characteristics of the eight open-source models and consensus were compared to GPT-4o. Three ideal consensus LLM ensembles were tested for rating the performance of the triage tool. Results: The cohort consisted of 29,766 head CTs exam-report pairs. The highest AUC performance was achieved with llama3.3:70b and GPT-4o (AUC= 0.78). The average precision was highest for Llama3.3:70b and GPT-4o (AP=0.75 & 0.76). Llama3.3:70b had the highest F1 score (0.81) and recall (0.85), greater precision (0.78), specificity (0.72), and MCC (0.57). Using MCC (95% CI) the ideal combination of LLMs were: Full-9 Ensemble 0.571 (0.552-0.591), Top-3 Ensemble 0.558 (0.537-0.579), Consensus 0.556 (0.539-0.574), and GPT4o 0.522 (0.500-0.543). No statistically significant differences were observed between Top-3, Full-9, and Consensus (p > 0.05). Conclusion: An ensemble of medium to large sized open-source LLMs provides a more consistent and reliable method to derive a ground truth retrospective evaluation of a clinical AI triage tool over a single LLM alone.

Traumatic brain injuries present significant diagnostic challenges in emergency medicine, where the timely interpretation of medical images is crucial for patient outcomes. In this paper, we propose a novel AI-based approach for automatic radiology report generation tailored to cranial trauma cases. Our model integrates an AC-BiFPN with a Transformer architecture to capture and process complex medical imaging data such as CT and MRI scans. The AC-BiFPN extracts multi-scale features, enabling the detection of intricate anomalies like intracranial hemorrhages, while the Transformer generates coherent, contextually relevant diagnostic reports by modeling long-range dependencies. We evaluate the performance of our model on the RSNA Intracranial Hemorrhage Detection dataset, where it outperforms traditional CNN-based models in both diagnostic accuracy and report generation. This solution not only supports radiologists in high-pressure environments but also provides a powerful educational tool for trainee physicians, offering real-time feedback and enhancing their learning experience. Our findings demonstrate the potential of combining advanced feature extraction with transformer-based text generation to improve clinical decision-making in the diagnosis of traumatic brain injuries.

Background: There are many challenges and opportunities in the clinical deployment of AI tools in radiology. The current study describes a radiology software platform called NeoMedSys that can enable efficient deployment and refinements of AI models. We evaluated the feasibility and effectiveness of running NeoMedSys for three months in real-world clinical settings and focused on improvement performance of an in-house developed AI model (VIOLA-AI) designed for intracranial hemorrhage (ICH) detection. Methods: NeoMedSys integrates tools for deploying, testing, and optimizing AI models with a web-based medical image viewer, annotation system, and hospital-wide radiology information systems. A prospective pragmatic investigation was deployed using clinical cases of patients presenting to the largest Emergency Department in Norway (site-1) with suspected traumatic brain injury (TBI) or patients with suspected stroke (site-2). We assessed ICH classification performance as VIOLA-AI encountered new data and underwent pre-planned model retraining. Performance metrics included sensitivity, specificity, accuracy, and the area under the receiver operating characteristic curve (AUC). Results: NeoMedSys facilitated iterative improvements in the AI model, significantly enhancing its diagnostic accuracy. Automated bleed detection and segmentation were reviewed in near real-time to facilitate re-training VIOLA-AI. The iterative refinement process yielded a marked improvement in classification sensitivity, rising to 90.3% (from 79.2%), and specificity that reached 89.3% (from 80.7%). The bleed detection ROC analysis for the entire sample demonstrated a high area-under-the-curve (AUC) of 0.949 (from 0.873). Model refinement stages were associated with notable gains, highlighting the value of real-time radiologist feedback.

Patients with Intracranial Hemorrhage (ICH) face a potentially life-threatening condition, and patient-centered individualized treatment remains challenging due to possible clinical complications. Deep-Learning-based methods can efficiently analyze the routinely acquired head CTs to support the clinical decision-making. The majority of early work focuses on the detection and segmentation of ICH, but do not model the complex relations between ICH and adjacent brain structures. In this work, we design a tailored object detection method for ICH, which we unite with segmentation-grounded Scene Graph Generation (SGG) methods to learn a holistic representation of the clinical cerebral scene. To the best of our knowledge, this is the first application of SGG for 3D voxel images. We evaluate our method on two head-CT datasets and demonstrate that our model can recall up to 74% of clinically relevant relations. This work lays the foundation towards SGG for 3D voxel data. The generated Scene Graphs can already provide insights for the clinician, but are also valuable for all downstream tasks as a compact and interpretable representation.

Purpose: To develop and evaluate a semi-supervised learning model for intracranial hemorrhage detection and segmentation on an out-of-distribution head CT evaluation set. Materials and Methods: This retrospective study used semi-supervised learning to bootstrap performance. An initial "teacher" deep learning model was trained on 457 pixel-labeled head CT scans collected from one US institution from 2010-2017 and used to generate pseudo-labels on a separate unlabeled corpus of 25000 examinations from the RSNA and ASNR. A second "student" model was trained on this combined pixel- and pseudo-labeled dataset. Hyperparameter tuning was performed on a validation set of 93 scans. Testing for both classification (n=481 examinations) and segmentation (n=23 examinations, or 529 images) was performed on CQ500, a dataset of 481 scans performed in India, to evaluate out-of-distribution generalizability. The semi-supervised model was compared with a baseline model trained on only labeled data using area under the receiver operating characteristic curve (AUC), Dice similarity coefficient (DSC), and average precision (AP) metrics. Results: The semi-supervised model achieved statistically significantly higher examination AUC on CQ500 compared with the baseline (0.939 [0.938, 0.940] vs. 0.907 [0.906, 0.908]) (p=0.009). It also achieved a higher DSC (0.829 [0.825, 0.833] vs. 0.809 [0.803, 0.812]) (p=0.012) and Pixel AP (0.848 [0.843, 0.853]) vs. 0.828 [0.817, 0.828]) compared to the baseline. Conclusion: The addition of unlabeled data in a semi-supervised learning framework demonstrates stronger generalizability potential for intracranial hemorrhage detection and segmentation compared with a supervised baseline.

Background: Accurate volumetric assessment of spontaneous subarachnoid hemorrhage (SAH) is a labor-intensive task performed with current manual and semiautomatic methods that might be relevant for its clinical and prognostic implications. In the present research, we sought to develop and validate an artificial intelligence-driven, fully automated blood segmentation tool for SAH patients via noncontrast computed tomography (NCCT) scans employing a transformer-based Swin UNETR architecture. Methods: We retrospectively analyzed NCCT scans from patients with confirmed aneurysmal subarachnoid hemorrhage (aSAH) utilizing the Swin UNETR for segmentation. The performance of the proposed method was evaluated against manually segmented ground truth data using metrics such as Dice score, intersection over union (IoU), the volumetric similarity index (VSI), the symmetric average surface distance (SASD), and sensitivity and specificity. A validation cohort from an external institution was included to test the generalizability of the model. Results: The model demonstrated high accuracy with robust performance metrics across the internal and external validation cohorts. Notably, it achieved high Dice coefficient (0.873), IoU (0.810), VSI (0.840), sensitivity (0.821) and specificity (0.996) values and a low SASD (1.866), suggesting proficiency in segmenting blood in SAH patients. The model's efficiency was reflected in its processing speed, indicating potential for real-time applications. Conclusions: Our Swin UNETR-based model offers significant advances in the automated segmentation of blood after aSAH on NCCT images. Despite the computational intensity, the model operates effectively on standard hardware with a user-friendly interface, facilitating broader clinical adoption. Further validation across diverse datasets is warranted to confirm its clinical reliability.

Multiple Instance Learning (MIL) has been widely applied to medical imaging diagnosis, where bag labels are known and instance labels inside bags are unknown. Traditional MIL assumes that instances in each bag are independent samples from a given distribution. However, instances are often spatially or sequentially ordered, and one would expect similar diagnostic importance for neighboring instances. To address this, in this study, we propose a smooth attention deep MIL (SA-DMIL) model. Smoothness is achieved by the introduction of first and second order constraints on the latent function encoding the attention paid to each instance in a bag. The method is applied to the detection of intracranial hemorrhage (ICH) on head CT scans. The results show that this novel SA-DMIL: (a) achieves better performance than the non-smooth attention MIL at both scan (bag) and slice (instance) levels; (b) learns spatial dependencies between slices; and (c) outperforms current state-of-the-art MIL methods on the same ICH test set.

This week Aidoc announced that they have raised $110 million in their Series D expansion round. This round of funding was co-led by TCV and Alpha Intelligence Capital with participation from CDIB Capital. Funding raised in this round will go toward expansion of Aidoc's first of its kind AI Care Platform. The platform offers health systems a singular platform solution designed to help doctors manage the entire patient lifecycle--from diagnostic aid, to consultation, to suggested treatment paths, to patient follow-up tools. In clinical studies, this platform has proven to reduce turnaround time, shorten patient length of stay and improve patient outcomes.

Radiology artificial intelligence company Aidoc scored FDA 510(k) clearance for its tool for flagging and triaging cases of pneumothorax, or a collapsed lung, on X-rays. Aidoc said the software could run on all X-rays, including portable machines, and automatically notes positive cases of pneumothorax so physicians can focus on these images more quickly. Some of Aidoc's other FDA-cleared tools include software for triaging and notification of incidental pulmonary embolism, triaging cervical spine fractures and flagging acute intracranial hemorrhage. "We're very excited about this important milestone," CEO Elad Walach said in a statement. "This FDA clearance further validates the breadth of our AI platform, going beyond specific AI algorithms to act as a healthcare AI hub for the enterprise's cross-specialty needs. This includes ER, ICU, outpatient centers, inpatient admissions, and the coordination of care and communication among providers. By bringing radiologists and proceduralists to the same AI platform, we enable enhanced collaboration across departments and systems to deliver patients with the right treatment at the right time."

Over the last several years, artificial intelligence (AI) has become one of the highest profile topics in radiology, recognized in part by the creation of this journal (1). This focus and interest has been driven largely by the potential AI shows to broadly change the way we practice radiology across every subspecialty. That potential has been demonstrated by a flood of manuscripts describing technical advances, algorithms, and proofs of concept aimed at a wide variety of radiologic tasks. However, no amount of demonstrated potential has a direct impact on patient care or clinical practice; achieving such an impact requires moving beyond the creation of AI to the deployment of AI into clinical environments for routine use. It is probably not surprising to those who practice radiology or work in radiology information technology that achieving this translational goal is challenging and has occurred at a much slower pace than suggested by some who feverishly predicted that AI would bring an end to radiology as a profession in a few short years.