Search

Purpose: The AAST Organ Injury Scale is widely adopted for splenic injury severity but suffers from only moderate inter-rater agreement. This work assesses SpleenPro, a prototype interactive explainable artificial intelligence/machine learning (AI/ML) diagnostic aid to support AAST grading, for effects on radiologist dwell time, agreement, clinical utility, and user acceptance., Methods: Two trauma radiology ad hoc expert panelists independently performed timed AAST grading on 76 admission CT studies with blunt splenic injury, first without AI/ML assistance, and after a 2-month washout period and randomization, with AI/ML assistance. To evaluate user acceptance, three versions of the SpleenPro user interface with increasing explainability were presented to four independent expert panelists with four example cases each. A structured interview consisting of Likert scales and free responses was conducted, with specific questions regarding dimensions of diagnostic utility (DU); mental support (MS); effort, workload, and frustration (EWF); trust and reliability (TR); and likelihood of future use (LFU)., Results: SpleenPro significantly decreased interpretation times for both raters. Weighted Cohen's kappa increased from 0.53 to 0.70 with AI/ML assistance. During user acceptance interviews, increasing explainability was associated with improvement in Likert scores for MS, EWF, TR, and LFU. Expert panelists indicated the need for a combined early notification and grading functionality, PACS integration, and report autopopulation to improve DU., Conclusions: SpleenPro was useful for improving objectivity of AAST grading and increasing mental support. Formative user research identified generalizable concepts including the need for a combined detection and grading pipeline and integration with the clinical workflow., (© 2024. The Author(s), under exclusive licence to American Society of Emergency Radiology (ASER).)

Purpose: There is a growing body of diagnostic performance studies for emergency radiology-related artificial intelligence/machine learning (AI/ML) tools; however, little is known about user preferences, concerns, experiences, expectations, and the degree of penetration of AI tools in emergency radiology. Our aim is to conduct a survey of the current trends, perceptions, and expectations regarding AI among American Society of Emergency Radiology (ASER) members., Methods: An anonymous and voluntary online survey questionnaire was e-mailed to all ASER members, followed by two reminder e-mails. A descriptive analysis of the data was conducted, and results summarized., Results: A total of 113 members responded (response rate 12%). The majority were attending radiologists (90%) with greater than 10 years' experience (80%) and from an academic practice (65%). Most (55%) reported use of commercial AI CAD tools in their practice. Workflow prioritization based on pathology detection, injury or disease severity grading and classification, quantitative visualization, and auto-population of structured reports were identified as high-value tasks. Respondents overwhelmingly indicated a need for explainable and verifiable tools (87%) and the need for transparency in the development process (80%). Most respondents did not feel that AI would reduce the need for emergency radiologists in the next two decades (72%) or diminish interest in fellowship programs (58%). Negative perceptions pertained to potential for automation bias (23%), over-diagnosis (16%), poor generalizability (15%), negative impact on training (11%), and impediments to workflow (10%)., Conclusion: ASER member respondents are in general optimistic about the impact of AI in the practice of emergency radiology and its impact on the popularity of emergency radiology as a subspecialty. The majority expect to see transparent and explainable AI models with the radiologist as the decision-maker., (© 2023. The Author(s), under exclusive licence to American Society of Emergency Radiology (ASER).)

Background: AI/ML CAD tools can potentially improve outcomes in the high-stakes, high-volume model of trauma radiology. No prior scoping review has been undertaken to comprehensively assess tools in this subspecialty., Purpose: To map the evolution and current state of trauma radiology CAD tools along key dimensions of technology readiness., Methods: Following a search of databases, abstract screening, and full-text document review, CAD tool maturity was charted using elements of data curation, performance validation, outcomes research, explainability, user acceptance, and funding patterns. Descriptive statistics were used to illustrate key trends., Results: A total of 4052 records were screened, and 233 full-text articles were selected for content analysis. Twenty-one papers described FDA-approved commercial tools, and 212 reported algorithm prototypes. Works ranged from foundational research to multi-reader multi-case trials with heterogeneous external data. Scalable convolutional neural network-based implementations increased steeply after 2016 and were used in all commercial products; however, options for explainability were narrow. Of FDA-approved tools, 9/10 performed detection tasks. Dataset sizes ranged from < 100 to > 500,000 patients, and commercialization coincided with public dataset availability. Cross-sectional torso datasets were uniformly small. Data curation methods with ground truth labeling by independent readers were uncommon. No papers assessed user acceptance, and no method included human-computer interaction. The USA and China had the highest research output and frequency of research funding., Conclusions: Trauma imaging CAD tools are likely to improve patient care but are currently in an early stage of maturity, with few FDA-approved products for a limited number of uses. The scarcity of high-quality annotated data remains a major barrier., (© 2023. The Author(s), under exclusive licence to American Society of Emergency Radiology (ASER).)

Objectives: Weight percentiles are generally reported without any indication of error. This variation can lead a fetus being mistakenly classified erroneously as having intrauterine growth restriction (IUGR) or macrosomia. The goal of this study was to compare estimated weight percentiles with the actual observed weight percentile for each gestational age in a large cohort of fetuses being scanned in our institution., Methods: After IRB approval the radiology information system data base was retrospectively searched for all obstetrical US reports obtained during the late second and third trimesters from July 1, 2014, until July 1, 2020. Demographic information, fetal weight, and weight percentile information were obtained from these reports. Quantile-quantile plots were created for all gestational ages and all ethnicities., Results: Our study included 6259 ultrasounds in 4060 patients. Mean maternal age of the total group was 31.68 years (ranging 15-53 years). When all subjects were considered, the median values in our QQ plots approximated the line of identity. However, there was considerable variation for a given estimate, implying that estimated fetal weight percentiles are only very rough predictors of the actual percentile., Conclusion: Estimated fetal weight percentiles are only very rough predictors of the actual percentile. We therefore suggest that estimates of the weight percentile should be reported along with an estimate of the expected variation. Recognition of variations in weight percentile should be considered in the greater clinical context, and could potentially prevent misdiagnosis of growth restriction and macrosomia as well as the subsequent overutilization of resources, unnecessary interventions, and maternal stress., (© 2022 American Institute of Ultrasound in Medicine.)

Background: At our level 1 trauma center, we overread cross-sectional scans on all transferred patients, initially reported by community radiologists (CR). We designed a unique peer feedback learning model to address discrepancies encountered with outside overreads with the goal of practice improvement on the part of the CRs. Although there is ample literature on internal peer review and peer learning programs, no publications address errors committed by peers outside institutional boundaries. In this paper we describe our model and report a survey analyzing the perception of the program by the CRs., Materials and Methods: Outside CT and MR exams and reports of patients transferred to our level 1 trauma center were imported into PACS and prospectively overread by specialist trauma radiologists. Our report contained a summary of the outside findings as well as our findings. In the case of a significant discrepancy, a paper copy of our final report was sent by US mail to the originating CR. When the program had been active for 18 months, an invitation to participate in a survey was sent to all radiologists who had been sent reports., Results: Eight thousand three hundred forty patients were transferred, of which 4331 (52%) had 9175 exams with outside reports (8666 CT scans and 509 MRI). One hundred seventy six final report letters containing significant discrepancies were sent to 139 individual radiologists. These 139 radiologists also later received our survey letter. Thirty-eight (27%) responses were received. Thirty-two respondents (84%) recalled receiving the report and reviewed the exam in question. Twenty-eight of them (85%) agreed with the overread and 30 (88%) believed that our feedback program should be continued., Conclusion: We have designed a novel peer feedback learning model to address discrepancies in outside overreads which is administratively simple and well received by the CRs getting feedback. Those CR who responded to the survey rated the experience favorably and wanted the practice continued, although the overall response rate did not allow statistical analysis. Also, institutions trying to design similar or new peer learning models can benefit from our experience., (Copyright © 2022 Elsevier Inc. All rights reserved.)

Purpose: To determine the rate and nature of significant discordances between community and subspecialist emergency radiologists' interpretations of cross-sectional exams performed on patients transferred to our trauma center., Methods: Outside hospital CT and MRI exams performed on transfer trauma patients are routinely overread by subspecialist emergency radiologists, specifying either concordance or discordance with the interpretation by the community radiologist. We evaluated the discordant reports for clinical significance, defined as an additional finding or difference in interpretation which was likely to affect patient management. The total rate of significant discordances, rate by modality, rate by body region, and rate per patient transferred were calculated. The most common errors were identified, and the distribution of errors among individual community radiologists was examined., Results: 9175 exams were reviewed. Significant discordances were encountered in 4.1% of exams: 3.9% for CT and 6.7% for MRI; 5.1% for head and neck exams, 3.3% for spine, 3.8% for torso, and 2.9% for extremities. The discordance rate per patient transferred was 7.7%. The most common discordances involved missing injuries to the cranio-cervical junction, missing or misinterpreting vascular injuries in the neck, and incompletely characterizing facial fractures. Discordances were evenly spread among 220 community radiologists., Conclusion: There is frequent discordance between community and emergency radiologists' interpretations of CT and MRI exams, leading to a change in transferred patient management. Thus, trauma center radiologists provide added value overreading these patients' exams. It is difficult to predict which patients or exams will contain discordances, justifying routine overreading of all such exams., (© 2022. American Society of Emergency Radiology.)

The purpose of this review is to understand the role of imaging in the diagnosis and management of non-traumatic subarachnoid hemorrhage (SAH). SAH is a life-threatening emergency and a relatively common entity, the most common etiology being ruptured aneurysms. Multiple conundrums exist in literature at various steps of its imaging workup: diagnosis, management, and follow-up. We target our review to highlight the most effective practice and suggest efficient workup plans based on literature search, and describe in detail the clinical diagnostic and prognostic scales, role of CT scan, lumbar puncture, and MR, including angiography in the diagnosis and workup of SAH and its complications, and try to simplify the conundrums. Practical knowledge of imaging workup of SAH can help guide correct management of these patients, so as to reduce morbidity and mortality without resource overutilization., (© 2021. American Society of Emergency Radiology.)

Purpose: Occasionally, a patient requiring computed tomography (CT) with poor venous access or in an unstable condition may have a sheath as their only form of IV access. Many institutions prohibit contrast power injection through the side ports, for concern for rupture. We hypothesize that under conditions encountered in CT scanning, the side arm of the sheath introducer is safe for power injection of contrast material., Methods: In this in vitro study, we injected contrast at different flow rates varying from 2 to 8 ml/s, through the side port of the sheath. Time-pressure graphs were obtained for each injection from the injector display. The assembly was observed for any signs of leakage or material failure., Results: There was not a single event of leak, rupture, or displacement. Maximum pressures were within the range of conventional contrast administrations., Conclusion: This study suggests that iodinated contrast can safely be injected through the side port of a venous sheath.

Tuberculosis is a devastating disease and has shown resurgence in recent years with the advent of acquired immunodeficiency syndrome. Central nervous system involvement is the most devastating form of the disease, comprising 10% of all tuberculosis cases. The causative organism, Mycobacterium tuberculosis , incites a granulomatous inflammatory response in the brain, the effects of which can be appreciated on magnetic resonance imaging (MRI), which can thus be used for diagnosis of the same. Neurotuberculosis can present in various patterns, which can be identified on MRI. The meningeal forms include leptomeningitis and pachymeningitis. Parenchymal forms of neurotuberculosis include tuberculoma in its various stages, tubercular cerebritis and abscess, tubercular rhombencephalitis, and tubercular encephalopathy. Each pattern has characteristic MRI appearances and differential diagnoses on imaging. Complications of neurotuberculosis, usually of tubercular meningitis, include hydrocephalus, vasculitis, and infarcts as well as cranial nerve palsies. Various MRI sequences besides the conventional ones can provide additional insight into the disease, help in quantifying the disease load, and help in differentiation of neurotuberculosis from conditions with similar imaging appearances and presentations. These can enable accurate and timely diagnosis by the radiologist and early institution of treatment in order to reduce the likelihood of permanent neurological sequelae., Competing Interests: The authors report no conflict of interest.