Considerations in Specialty Selections: What Medical Students Think, Know, and Think They Know about Interventional Radiology.

PURPOSE: To compare the motivation, deterrents, knowledge, exposure, and other specialty considerations of first- to fourth-year medical students interested in interventional radiology (IR) with those who are not. MATERIALS AND METHODS: Matriculants of 5 medical schools varying by region, public/private, class size, and National Institutes of Health research ranking received a 19-question survey with questions about demographics, specialty interests, motivations/deterrents, knowledge, and exposure to IR. RESULTS: A total of 25.8% (611/2370) of students completed the survey, of which 20.5% (125/611) expressed interest in IR, and 25% (47/186), 26% (40/153), 24% (34/143), and 3% (3/117) of first-year, second-year, third-year, and fourth-year medical students, respectively, were seriously considering IR. Those interested in IR were less motivated by direct patient care (mean, 2.8/5; 95% confidence interval [CI], 2.6-3.0) and longitudinal patient care (mean, 1.6/5; 95% CI, 1.4-1.7) (both, P < .01) and more motivated by salary (2.6/5; 95% CI, 2.3-2.9), job market (2.8/5; 95% CI, 2.6-2.9), and procedures (3.1/5; 95% CI, 2.8-3.4) compared with their peers (all P < .05). Those interested in IR were more certain about their IR knowledge (mean range, 1.6-2.0/3.0; 95% CI, 1.3-2.3) than their peers (mean range, 1.9-2.4/3.0; 95% CI, 1.6-2.1, in which 0 = certain, P ≤ .01); however, both groups scored low in actual knowledge (those considering IR: 35.0-73.2% correct; 95% CI, 23.5-81.4; those who were not: 26.6-66.7% correct; 95% CI, 24.3-75.9, P > .05). CONCLUSIONS: Although medical students showed interest in IR, they had a limited understanding of IR. IR educators may increase IR interest and understanding among medical students by clarifying the procedural aspects and longitudinal care present in a comprehensive IR practice.

Machine Learning Offers Exciting Potential for Predicting Postprocedural Outcomes: A Framework for Developing Random Forest Models in IR.

PURPOSE: To demonstrate that random forest models trained on a large national sample can accurately predict relevant outcomes and may ultimately contribute to future clinical decision support tools in IR. MATERIALS AND METHODS: Patient data from years 2012-2014 of the National Inpatient Sample were used to develop random forest machine learning models to predict iatrogenic pneumothorax after computed tomography-guided transthoracic biopsy (TTB), in-hospital mortality after transjugular intrahepatic portosystemic shunt (TIPS), and length of stay > 3 days after uterine artery embolization (UAE). Model performance was evaluated with area under the receiver operating characteristic curve (AUROC) and maximum F1 score. The threshold for AUROC significance was set at 0.75. RESULTS: AUROC was 0.913 for the TTB model, 0.788 for the TIPS model, and 0.879 for the UAE model. Maximum F1 score was 0.532 for the TTB model, 0.357 for the TIPS model, and 0.700 for the UAE model. The TTB model had the highest AUROC, while the UAE model had the highest F1 score. All models met the criteria for AUROC significance. CONCLUSIONS: This study demonstrates that machine learning models may suitably predict a variety of different clinically relevant outcomes, including procedure-specific complications, mortality, and length of stay. Performance of these models will improve as more high-quality IR data become available.

Acute Lower Gastrointestinal Bleeding: Temporal Factors Associated With Positive Findings on Catheter Angiography After (99m)Tc-Labeled RBC Scanning.

OBJECTIVE: The objective of the study was to determine if time to positive (TTP), defined as the time from the start of (99m)Tc-labeled RBC scanning to the appearance of a radionuclide blush (considered to be a positive finding for acute lower gastrointestinal bleeding [LGIB]), and lag time (LT), defined as the time from the appearance of a radionuclide blush to the start of catheter angiography (CA), affected the yield of CA for the detection of acute LGIB. MATERIALS AND METHODS: TTP and LT were retrospectively evaluated in 120 patients who had positive findings for acute LGIB on (99m)Tc-labeled RBC scanning and subsequently underwent CA for the diagnosis and localization of gastrointestinal bleeding. Two nuclear medicine fellowship-trained radiologists independently reviewed the (99m)Tc-labeled RBC scans. Two fellowship-trained interventional radiologists independently reviewed the angiograms. All data were analyzed using SAS software. RESULTS: When a TTP threshold of ≤ 9 minutes was used, the sensitivity, specificity, positive predictive value, and negative predictive value for a positive CA study were 92%, 35%, 27%, and 94%, respectively. In addition, the odds of detecting bleeding on CA increased 6.1-fold with a TTP of ≤ 9 minutes relative to a TTP of > 9 minutes (p = 0.020). A significant inverse relationship was found between LT and a positive CA study (p = 0.041). CONCLUSION: TTP and LT impact the rate of positive CA studies. A TTP threshold of ≤ 9 minutes allows the detection of almost all patients who would benefit from CA for treatment and allows a reduction in unnecessary negative CA studies. The likelihood of positive findings on CA decreases with a delay in the performance of CA.

The Application of Large Language Models for Radiologic Decision Making.

BACKGROUND AND PURPOSE: Large language models (LLMs) have seen explosive growth, but their potential role in medical applications remains underexplored. Our study investigates the capability of LLMs to predict the most appropriate imaging study for specific clinical presentations in various subspecialty areas in radiology. METHODS AND MATERIALS: Chat Generative Pretrained Transformer (ChatGPT), by OpenAI and Glass AI by Glass Health were tested on 1,075 clinical scenarios from 11 ACR expert panels to determine the most appropriate imaging study, benchmarked against the ACR Appropriateness Criteria. Two responses per clinical presentation were generated and averaged for the final clinical presentation score. Clinical presentation scores for each topic area were averaged as its final score. The average of the topic scores within a panel determined the final score of each panel. LLM responses were on a scale of 0 to 3. Partial scores were given for nonspecific answers. Pearson correlation coefficient (R-value) was calculated for each panel to determine a context-specific performance. RESULTS: Glass AI scored significantly higher than ChatGPT (2.32 ± 0.67 versus 2.08 ± 0.74, P = .002). Both LLMs performed the best in the Polytrauma, Breast, and Vascular panels, and performed the worst in the Neurologic, Musculoskeletal, and Cardiac panels. Glass AI outperformed ChatGPT in 10 of 11 panels, except Obstetrics and Gynecology. Maximum agreement was in the Pediatrics, Neurologic, and Thoracic panels, and the most disagreement occurred in the Vascular, Breast, and Urologic panels. CONCLUSION: LLMs can be used to predict imaging studies, with Glass AI's superior performance indicating the benefits of extra medical-text training. This supports the potential of LLMs in radiologic decision making.

De-Biased Disentanglement Learning for Pulmonary Embolism Survival Prediction on Multimodal Data.

Health disparities among marginalized populations with lower socioeconomic status significantly impact the fairness and effectiveness of healthcare delivery. The increasing integration of artificial intelligence (AI) into healthcare presents an opportunity to address these inequalities, provided that AI models are free from bias. This paper aims to address the bias challenges by population disparities within healthcare systems, existing in the presentation of and development of algorithms, leading to inequitable medical implementation for conditions such as pulmonary embolism (PE) prognosis. In this study, we explore the diverse bias in healthcare systems, which highlights the demand for a holistic framework to reducing bias by complementary aggregation. By leveraging de-biasing deep survival prediction models, we propose a framework that disentangles identifiable information from images, text reports, and clinical variables to mitigate potential biases within multimodal datasets. Our study offers several advantages over traditional clinical-based survival prediction methods, including richer survival-related characteristics and bias-complementary predicted results. By improving the robustness of survival analysis through this framework, we aim to benefit patients, clinicians, and researchers by enhancing fairness and accuracy in healthcare AI systems.

Using ChatGPT to Improve Readability of Interventional Radiology Procedure Descriptions.

PURPOSE: This project examines ChatGPT's potential to enhance the readability of patient educational materials about interventional radiology (IR) procedures. METHODS AND MATERIALS: The descriptions of IR procedures from the Cardiovascular and Interventional Radiological Society of Europe (CIRSE) were used as the original text. Readability scores were calculated using three metrics: Flesch Reading Ease (FRE), Gunning Fog (GF), and the Automated Readability Index (ARI) using an online calculator ( https://readabilityformulas.com ). FRE is scored on a scale of 0-100, where 100 indicates easy-to-read texts, and GF and ARI represent the grade level required to comprehend the text. The DISCERN instrument measured credibility and reliability. ChatGPT was prompted to simplify the texts to a fifth-grade reading level, with subsequent recalculation of readability and DISCERN scores for comparison. Statistical significance was determined using a Wilcoxon Signed-Rank Test. Articles were subsequently organized by subgroups and analyzed. RESULTS: 73 interventional radiology procedures from CIRSE were analyzed. The original FRE score was 47.2 (Difficult), improved to 78.4 (Fairly Easy) by ChatGPT. GF and ARI scores dropped from 14.4 and 11.2 to 7.8 and 5.8, respectively, after simplification, showing significant improvement (p < 0.001). However, the average DISCERN score decreased from 3.73 to 2.99 (p < 0.001) post-ChatGPT simplification. CONCLUSION: This study shows ChatGPT's ability to make interventional radiology descriptions more readable but highlights its struggle to maintain the original's reliability, suggesting the need for human review and prompt engineering to enhance outcomes. LEVEL OF EVIDENCE: Level 6.

Preoperative prediction of the stage, size, grade, and necrosis score in clear cell renal cell carcinoma using MRI-based radiomics.

PURPOSE: Clear cell renal cell carcinoma (ccRCC) is the most common subtype of renal cell carcinoma. Currently, there is a lack of noninvasive methods to stratify ccRCC prognosis prior to any invasive therapies. The purpose of this study was to preoperatively predict the tumor stage, size, grade, and necrosis (SSIGN) score of ccRCC using MRI-based radiomics. METHODS: A multicenter cohort of 364 histopathologically confirmed ccRCC patients (272 low [< 4] and 92 high [≥ 4] SSIGN score) with preoperative T2-weighted and T1-contrast-enhanced MRI were retrospectively identified and divided into training (254 patients) and testing sets (110 patients). The performance of a manually optimized radiomics model was assessed by measuring accuracy, sensitivity, specificity, area under receiver operating characteristic curve (AUROC), and area under precision-recall curve (AUPRC) on an independent test set, which was not included in model training. Lastly, its performance was compared to that of a machine learning pipeline, Tree-Based Pipeline Optimization Tool (TPOT). RESULTS: The manually optimized radiomics model using Random Forest classification and Analysis of Variance feature selection methods achieved an AUROC of 0.89, AUPRC of 0.81, accuracy of 0.89 (95% CI 0.816-0.937), specificity of 0.95 (95% CI 0.875-0.984), and sensitivity of 0.72 (95% CI 0.537-0.852) on the test set. The TPOT using Extra Trees Classifier achieved an AUROC of 0.94, AUPRC of 0.83, accuracy of 0.89 (95% CI 0.816-0.937), specificity of 0.95 (95% CI 0.875-0.984), and sensitivity of 0.72 (95% CI 0.537-0.852) on the test set. CONCLUSION: Preoperative MR radiomics can accurately predict SSIGN score of ccRCC, suggesting its promise as a prognostic tool that can be used in conjunction with diagnostic markers.

Vertebroplasty.

Percutaneous vertebroplasty has become widely accepted as a safe and effective minimally invasive procedure for the treatment of painful vertebral body compression fractures refractory to medical therapy. In this article, the authors review the indications and contraindications for vertebroplasty, principles of appropriate patient selection, useful techniques to achieve optimal outcomes, and the potential risks and complications of the procedure.

Basic vascular neuroanatomy of the brain and spine: what the general interventional radiologist needs to know.

This article is intended to provide a review of clinically relevant neurovascular anatomy. A solid understanding of the vascular anatomy of the brain and spine are essential for the safe and effective performance of neurointerventional radiology. Key concepts to master include collateral pathways and anastomoses between the external and internal carotid circulation, the Circle of Willis as a route to otherwise inaccessible intracranial vascular distributions, and the origin of spinal arterial blood supply. These concepts will be highlighted using clinical angiographic examples with discussion of relevant embryology and pathology as needed.

Interventional management of head and neck emergencies: carotid blowout.

Involvement of the carotid artery by malignant processes of the head and neck with compromise of vessel integrity and rupture-"carotid blowout syndrome" (CBS)-is one of the most devastating complications of malignancy. Most often, it is associated with squamous cell cancer and almost always in patients who have undergone prior radiation therapy. CBS is classified as threatened, impending, or acute. Bleeding into the oral cavity or from areas of skin breakdown is a frightening experience for patients and their families and often a terminal event. Prognosis is poor with up to 50% mortality and morbidity, and surgical options are limited and risky. Endovascular management with vessel sacrifice or stent placement has become the principle treatment option in this patient population, though still associated with procedural complications, often neurologic, that can occur acutely or in a delayed fashion. This article reviews techniques and outcomes associated with endovascular treatment of CBS.

Interventional management of vertebral body metastases.

Metastases to the vertebral column are often due to hypervascular primary tumors, the most common of which is renal cell carcinoma. Clinical symptoms attributed to vertebral body metastases include localized pain, mechanical instability of the vertebral column, and neurologic deficits resulting from mass effect. Treatment options include targeted radiotherapy, percutaneous vertebral augmentation with or without thermal ablation, and surgical resection with subsequent fusion. Overall, surgical resection of the tumor and stabilization of the vertebral column provide the best prognosis for the patient in terms of symptomatic improvement and long-term survival; however, resection of hypervascular vertebral body metastases can result in significant intraoperative blood loss that can add to the morbidity of the procedure. Preoperative embolization of hypervascular metastases of the vertebral column has been shown to significantly reduce intraoperative blood loss at the time of surgery. The goal of this manuscript is to describe the role of embolization therapy in the management of patients with vertebral body metastases.

Basic neuroangiography: review of technique and perioperative patient care.

Neuroangiography (NA) is an important part of diagnosis and treatment of patients with neurological disease. Although NA may be performed for diagnostic purposes, in many instances NA is performed with the intent to treat. Indications for NA range from extracranial diseases (vertebrobasilar insufficiency from subclavian steal, extracranial carotid stenosis, cavernous-carotid fistula, neck trauma, epistaxis, tumor invasion of the carotid artery, and tumor embolization) to intracranial diseases (nontraumatic subarachnoid hemorrhage, cerebral aneurysms, cerebral arteriovenous malformations, cerebral vasospasm, acute stroke, tumor embolization, and WADA test). Similar to peripheral angiography, appropriate preprocedural assessment and postprocedural care, along with understanding of anatomy, catheter technique, and disease processes, are vital to successful outcomes. This article will review the basic technique, equipment, and patient management in NA. With appropriate skill and knowledge, interventional radiologists can perform NA with safe and successful results.

Transcatheter embolization in the management of epistaxis.

A majority of the population will experience epistaxis at some time in their life. Most cases will be from an anterior source and can be treated with pressure, anterior nasal packing, or cautery. Intractable epistaxis is generally posterior in origin and may require endoscopic cautery, posterior packing, surgical ligation, or embolization. Embolization has been used to treat epistaxis for more than 30 years and success can be achieved in approximately 90% of patients, with major complications occurring in approximately 2%. These excellent results require thorough knowledge of the regional anatomy, familiarity with the equipment and various agents used to achieve this type of embolization, as well as attention to detail and meticulous technique. There remains debate on several aspects of embolization, including the agent of choice, preferred size of the embolic, and the number of vessels to embolize. Advances in endoscopic surgery have evolved to the point that similar success rates for embolization and modern surgical techniques in treating epistaxis may be expected. This detailed review of pertinent vascular anatomy, embolization technique, and surgical alternatives should allow practitioners to formulate treatment algorithms that result in optimal outcomes at their institutions.

Contemporary endovascular embolotherapy for meningioma.

Preoperative endovascular tumor embolization has been used for 40 years. Meningiomas are the most common benign intracranial tumor in which preoperative embolization has been most extensively described in the literature. Advocates of embolization report that it reduces operative blood-loss, and softens the tumor, thus making surgery safer and easier. Opponents suggest that it adds additional risk and cost for patients without controlled studies showing conclusive benefit. The literature suggests a 3 to 6% neurological complication rate related to embolization. The combined external and internal carotid artery blood supply and complex anastomoses of the meninges can make embolization challenging. Positive outcomes require thorough knowledge of the pertinent vascular anatomy, familiarity with the neurovascular equipment and embolics, and meticulous technique. There remains debate on several aspects of embolization, including tumors most appropriate for embolization, embolic agent of choice, ideal size of embolic, and the choice of vessel(s) to embolize. This detailed review of pertinent vascular anatomy, embolization technique, results, and complications should allow practitioners to maximize treatment outcomes in this setting.

Intra-arterial Stroke Management.

Acute ischemic stroke is a leading cause of death and the leading cause of disability in the United States. Cerebral neuronal death begins within minutes after threshold values of blood oxygen saturation are crossed. Prompt restoration of oxygenated blood flow into ischemic tissue remains the common goal of reperfusion strategies. This article provides a brief overview of acute ischemic stroke, a summary of the major intra-arterial stroke therapy trials, and comments on current training requirements for the performance of intra-arterial therapies.

Carotid artery stenting: review of technique and update of recent literature.

Stroke is the fourth leading cause of death and the number one cause of long-term disability in the United States. Carotid stenosis is an important cause of ischemic strokes, accounting for 20 to 25%. Previous studies have established carotid endarterectomy as standard of care of symptomatic patients with > 50% stenosis and asymptomatic patients with > 60% stenosis; recently, carotid artery stenting has emerged as an alternative treatment for carotid stenosis. Several studies have been published comparing carotid artery stenting with endarterectomy with mixed results. In this article, the authors discuss carotid artery stenting technique, the results from the most recent trials, and future directions.