Search

Automatic segmentation of angiographic structures can aid in assessing vascular disease. While recent deep learning models promise automation, they lack validation on interventional angiographic data. This study investigates the feasibility of angiographic segmentation using in-context learning with the UniverSeg model, which is a cross-learning segmentation model that lacks inherent angiographic training. A retrospective review, after IRB approval, identified 234 patients who underwent interventional fluoroscopy of the celiac axis with iodinated contrast from January 1, 2019, to December 31, 2022. From 261 acquisitions, 303 maximum contrast images were selected, each generating a 128 × 128 pixel partition for arterial detail analysis and binary mask creation. Image-mask pairs were divided into three classes of 101 pairs each, based on arterial diameter and bifurcation number. UniverSeg was tested class independently in a fivefold nested cross-validation. Performance analysis for in-context learning determined average model convergence for class sizes from 1 to 81 pairs. The model was further validated by repeating the tests on the inverse segmentation task. Dice similarity coefficients for decreasing diameters were 78.7%, 72.5%, and 59.9% (σ = 5.96, 7.99, 14.29). Balanced average Hausdorff distances were 0.86, 0.71, and 1.16 (σ = 0.37, 0.52, 0.68) pixels, respectively. Inverted mask testing aligned with UniverSeg expectations for out-of-context problem sets. Performance improved with support class size, vessel diameter, and reduced bifurcations, plateauing to within ± 1.34 Dice score at N = 51. This study validates UniverSeg for arterial segmentation in interventional fluoroscopic procedures, supporting vascular disease modeling and imaging research., Competing Interests: Declarations. Conflict of Interest: The authors declare no competing interests., (© 2024. The Author(s) under exclusive licence to Society for Imaging Informatics in Medicine.)

Background Recent advancements, including image processing capabilities, present new potential applications of large language models such as ChatGPT (OpenAI), a generative pretrained transformer, in radiology. However, baseline performance of ChatGPT in radiology-related tasks is understudied. Purpose To evaluate the performance of GPT-4 with vision (GPT-4V) on radiology in-training examination questions, including those with images, to gauge the model's baseline knowledge in radiology. Materials and Methods In this prospective study, conducted between September 2023 and March 2024, the September 2023 release of GPT-4V was assessed using 386 retired questions (189 image-based and 197 text-only questions) from the American College of Radiology Diagnostic Radiology In-Training Examinations. Nine question pairs were identified as duplicates; only the first instance of each duplicate was considered in ChatGPT's assessment. A subanalysis assessed the impact of different zero-shot prompts on performance. Statistical analysis included χ 2 tests of independence to ascertain whether the performance of GPT-4V varied between question types or subspecialty. The McNemar test was used to evaluate performance differences between the prompts, with Benjamin-Hochberg adjustment of the P values conducted to control the false discovery rate (FDR). A P value threshold of less than.05 denoted statistical significance. Results GPT-4V correctly answered 246 (65.3%) of the 377 unique questions, with significantly higher accuracy on text-only questions (81.5%, 159 of 195) than on image-based questions (47.8%, 87 of 182) (χ 2 test, P < .001). Subanalysis revealed differences between prompts on text-based questions, where chain-of-thought prompting outperformed long instruction by 6.1% (McNemar, P = .02; FDR = 0.063), basic prompting by 6.8% ( P = .009, FDR = 0.044), and the original prompting style by 8.9% ( P = .001, FDR = 0.014). No differences were observed between prompts on image-based questions with P values of .27 to >.99. Conclusion While GPT-4V demonstrated a level of competence in text-based questions, it showed deficits interpreting radiologic images. © RSNA, 2024 See also the editorial by Deng in this issue.

Radiographic quality control is an integral component of the radiology workflow. In this study, we developed a convolutional neural network model tailored for automated quality control, specifically designed to detect and classify key attributes of wrist radiographs including projection, laterality (based on the right/left marker), and the presence of hardware and/or casts. The model's primary objective was to ensure the congruence of results with image requisition metadata to pass the quality assessment. Using a dataset of 6283 wrist radiographs from 2591 patients, our multitask-capable deep learning model based on DenseNet 121 architecture achieved high accuracy in classifying projections (F1 Score of 97.23%), detecting casts (F1 Score of 97.70%), and identifying surgical hardware (F1 Score of 92.27%). The model's performance in laterality marker detection was lower (F1 Score of 82.52%), particularly for partially visible or cut-off markers. This paper presents a comprehensive evaluation of our model's performance, highlighting its strengths, limitations, and the challenges encountered during its development and implementation. Furthermore, we outline planned future research directions aimed at refining and expanding the model's capabilities for improved clinical utility and patient care in radiographic quality control., (© 2024. The Author(s) under exclusive licence to Society for Imaging Informatics in Medicine.)

Radiology Imaging plays a pivotal role in medical diagnostics, providing clinicians with insights into patient health and guiding the next steps in treatment. The true value of a radiological image lies in the accuracy of its accompanying report. To ensure the reliability of these reports, they are often cross-referenced with operative findings. The conventional method of manually comparing radiology and operative reports is labor-intensive and demands specialized knowledge. This study explores the potential of a Large Language Model (LLM) to simplify the radiology evaluation process by automatically extracting pertinent details from these reports, focusing especially on the shoulder's primary anatomical structures. A fine-tuned LLM identifies mentions of the supraspinatus tendon, infraspinatus tendon, subscapularis tendon, biceps tendon, and glenoid labrum in lengthy radiology and operative documents. Initial findings emphasize the model's capability to pinpoint relevant data, suggesting a transformative approach to the typical evaluation methods in radiology., (©2024 AMIA - All rights reserved.)

Rotator cuff (RC) tears are a common source of pain and decreased shoulder strength. Muscle length is known to affect muscle strength, and therefore evaluating changes in supraspinatus muscle length associated with RC pathology, surgical repair, and post-operative recovery may provide insights into functional deficits. Our objective was to develop a reliable MRI-based approach for assessing supraspinatus muscle length. Using a new semi-automated approach for identifying 3D location of the muscle-tendon junction (MTJ), supraspinatus muscle length was calculated as the sum of MTJ distance (distance between 3D MTJ position and glenoid plane) and supraspinatus fossa length (distance between root of the scapular spine and glenoid plane). Inter- and intra-operator reliability of this technique were assessed with intraclass correlation coefficient (ICC) and found to be excellent (ICCs > 0.96). Muscle lengths of 6 patients were determined before RC repair surgery and at 3- and 12-months post-surgery. Changes in normalized muscle length (muscle length as a percentage of pre-surgical muscle length) at 3 months post-surgery varied considerably across patients (16.1 % increase to 7.0 % decrease) but decreased in all patients from 3- to 12-months post-surgery (0.3 % to 17.2 %). This study developed a novel and reliable approach for quantifying supraspinatus muscle length and provided preliminary demonstration of its utility by assessing muscle length changes associated with RC pathology and surgical repair. Future studies can use this technique to evaluate changes over time in supraspinatus muscle length in response to clinical intervention, and associations between muscle length and shoulder function., Competing Interests: Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2024. Published by Elsevier Ltd.)

Purpose: To determine the relationship between resident imaging volumes and number of subspecialty rotations with Diagnostic Radiology In-Training (DXIT) subspecialty scores., Methods: DXIT-scaled subspecialty scores from a single large diagnostic radiology training program from 2014 to 2020 were obtained. The cumulative number of imaging studies dictated by each resident and specific rotations were mapped to each subspecialty for each year of training. DXIT subspecialty scores were compared against the total subspecialty imaging volume and the total number of rotations in a subspecialty for each resident year. A total of 52 radiology residents were trained during the study period and included in the dataset., Results: There was a positive linear relationship between the number of neuro studies and scaled neuro DXIT scores for R1s (Pearson coefficient: 0.29; p-value: 0.034) and between the number of breast studies and the number of neuro studies with DXIT scores for R2s (Pearson coefficients: 0.50 and 0.45, respectively; p-values: 0.001 and 0.003, respectively). Furthermore, a positive significant linear relationship between the total number of rotations in cardiac, breast, neuro, and thoracic subspecialties and their scaled DXIT scores for R2 residents (Pearson coefficients: 0.34, 0.49, 0.33, and 0.32, respectively; p-value: 0.025, 0.001, 0.03, and 0.036, respectively) and between the total number of nuclear medicine rotations with DXIT scores for R3s (Pearson coefficient: 0.41; p-value: 0.016)., Conclusion: Resident subspecialty imaging volumes and rotations have a variable impact on DXIT scores. Understanding the impact of study volume and the number of subspecialty rotations on resident medical knowledge will help residents and program directors determine how much emphasis to place on these factors during residency., (Copyright © 2023 Elsevier Inc. All rights reserved.)

Imaging examination selection and protocoling are vital parts of the radiology workflow, ensuring that the most suitable exam is done for the clinical question while minimizing the patient's radiation exposure. In this study, we aimed to develop an automated model for the revision of radiology examination requests using natural language processing techniques to improve the efficiency of pre-imaging radiology workflow. We extracted Musculoskeletal (MSK) magnetic resonance imaging (MRI) exam order from the radiology information system at Henry Ford Hospital in Detroit, Michigan. The pretrained transformer, "DistilBERT" was adjusted to create a vector representation of the free text within the orders while maintaining the meaning of the words. Then, a logistic regression-based classifier was trained to identify orders that required additional review. The model achieved 83% accuracy and had an area under the curve of 0.87., (©2023 AMIA - All rights reserved.)

Purpose: As the COVID-19 pandemic continues, efforts by radiology departments to protect patients and healthcare workers and mitigate disease spread have reduced imaging volumes. This study aims to quantify the pandemic's impact on physician productivity across radiology practice areas as measured by physician work Relative Value Units (wRVUs)., Materials and Methods: All signed diagnostic and procedural radiology reports were curated from January 1st to July 1st of 2019 and 2020. Physician work RVUs were assigned to each study type based on the Medicare Physician Fee Schedule. Utilizing divisional assignments, radiologist schedules were mapped to each report to generate a sum of wRVUs credited to that division for each week. Differential impact on divisions were calculated relative to a matched timeframe in 2019 and a same length pre-pandemic time period in 2020., Results: All practice areas saw a substantial decrease in wRVUs from the 2020 pre- to intra-pandemic time period with a mean decrease of 51.5% (range 15.4%-76.9%). The largest declines were in Breast imaging, Musculoskeletal, and Neuroradiology, which had decreases of 76.9%, 75.3%, and 67.5%, respectively. The modalities with the greatest percentage decrease were mammography, MRI, and non-PET nuclear medicine., Conclusion: All radiology practice areas and modalities experienced a substantial decrease in wRVUs. The greatest decline was in Breast imaging, Neuroradiology, and Musculoskeletal radiology. Understanding the differential impact of the pandemic on practice areas will help radiology departments prepare for the potential depth and duration of the pandemic by better understanding staffing needs and the financial effects., (Copyright © 2020 Elsevier Inc. All rights reserved.)

The Coronavirus disease 2019 (COVID-19) presents open questions in how we clinically diagnose and assess disease course. Recently, chest computed tomography (CT) has shown utility for COVID-19 diagnosis. In this study, we developed Deep COVID DeteCT (DCD), a deep learning convolutional neural network (CNN) that uses the entire chest CT volume to automatically predict COVID-19 (COVID+) from non-COVID-19 (COVID-) pneumonia and normal controls. We discuss training strategies and differences in performance across 13 international institutions and 8 countries. The inclusion of non-China sites in training significantly improved classification performance with area under the curve (AUCs) and accuracies above 0.8 on most test sites. Furthermore, using available follow-up scans, we investigate methods to track patient disease course and predict prognosis.

Rationale and Objectives: The COVID-19 pandemic has greatly impacted radiology departments across the country. The pandemic has also disrupted resident education, both due to departmental social distancing efforts and reduced imaging volumes. The purpose of this study was to assess the differential impact the pandemic had on radiology resident imaging volumes by training year and imaging modality., Materials and Methods: All signed radiology resident reports were curated during defined prepandemic and intrapandemic time periods. Imaging case volumes were analyzed on a mean per resident basis to quantify absolute and percent change by training level. Change in total volume by imaging modality was also assessed. The number of resident workdays assigned outside the normal reading room was also calculated., Results: Overall percent decline in resident imaging interpretation volume from the prepandemic to intrapandemic time period was 62.8%. R1s and R2s had the greatest decline at 87.3% and 64.3%, respectively. Mammography, MRI and nuclear medicine had the greatest decline in resident interpretation volume at 92.0%, 73.2%, and 73.0%, respectively. During the intrapandemic time period, a total of 478 resident days (mean of 14.5 days per resident) were reassigned outside of the radiology reading room., Conclusion: The COVID-19 pandemic caused a marked decrease in radiology resident imaging interpretation volume and has had a tremendous impact on resident education. The decrease in case interpretation, as well as in-person teaching has profound implications for resident education. Knowledge of this differential decrease by training level will help residency programs plan for the future., (Copyright © 2020 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.)

Purpose: To assess the interobserver and intraobserver agreement of fellowship trained chest radiologists, nonchest fellowship-trained radiologists, and fifth-year radiology residents for COVID-19-related imaging findings based on the consensus statement released by the Radiological Society of North America (RSNA)., Methods: A survey of 70 chest CTs of polymerase chain reaction (PCR)-confirmed COVID-19 positive and COVID-19 negative patients was distributed to three groups of participating radiologists: five fellowship-trained chest radiologists, five nonchest fellowship-trained radiologists, and five fifth-year radiology residents. The survey asked participants to broadly classify the findings of each chest CT into one of the four RSNA COVID-19 imaging categories, then select which imaging features led to their categorization. A 1-week washout period followed by a second survey comprised of randomly selected exams from the initial survey was given to the participating radiologists., Results: There was moderate overall interobserver agreement in each group (κ coefficient range 0.45-0.52 ± 0.02). There was substantial overall intraobserver agreement across the chest and nonchest groups (κ coefficient range 0.61-0.67 ± 0.06) and moderate overall intraobserver agreement within the resident group (κ coefficient 0.58 ± 0.06). For the image features that led to categorization, there were varied levels of agreement in the interobserver and intraobserver components that ranged from fair to perfect kappa values. When assessing agreement with PCR-confirmed COVID status as the key, we observed moderate overall agreement within each group., Conclusion: Our results support the reliability of the RSNA consensus classification system for COVID-19-related image findings., (Copyright © 2020 The Association of University Radiologists. Published by Elsevier Inc. All rights reserved.)

Minimally invasive implantable cardiac devices used in valve repair and replacement, cardiovascular support, and partial chamber and appendageal occlusion represent a burgeoning area of both bioengineering and clinical innovation. In addition to familiarizing the reader with the radiographic appearance of the most commonly utilized and encountered newer devices, this review will also address the relevant clinical and pathophysiological indications for usage and deployment as well as potentially encountered complications., (Copyright © 2018 Elsevier Inc. All rights reserved.)

Purpose: Order entry protocol selection of advanced imaging studies is labor-intensive, can disrupt workflow, and may displace staff from more valuable tasks. The aim of this study was to explore and compare the behaviors of radiologic technologists and radiologists when determining protocol to identify opportunities for workflow automation., Methods: A data set of over 273,000 cross-sectional examination orders from four hospitals within our health system was created. From this data set, we isolated the 12 most frequently requested examinations, which represent almost 50% of the entirety of advanced imaging volume. Intergroup comparisons were made between behavior of radiologic technologists and radiologists or residents when determining protocol. Frequencies of changes were calculated. Common parameters of changed examinations were identified., Results: The overall change rate for both radiologists and residents (4%) is very low and comparable to the overall change rate of radiologic technologists (1%). The change rates for the 12 most ordered examinations were calculated and compared individually. Most examinations that underwent change involved a patient with a low estimated glomerular filtration rate, a patient with a contrast allergy, or a provider ordering a general examination but in fact wanting an organ-specific protocol or an angiographic study., Conclusion: Order entry protocol selection of the most frequently ordered advanced imaging examinations was rarely a value-added activity because these examinations are rarely changed. Changes follow predictable patterns that make order entry protocol selection of most radiology orders for advanced imaging amenable to workflow automation., (Copyright © 2018 American College of Radiology. Published by Elsevier Inc. All rights reserved.)

Objective: To determine the appropriateness rating (AR) of advanced inpatient imaging requests that were not rated by prospective, point-of-care clinical decision support (CDS) using computerized provider order entry., Materials and Methods: During 30-day baseline and intervention periods, CDS generated an AR for advanced inpatient imaging requests (nuclear medicine, CT, and MRI) using provider-selected structured indications from pull-down menus in the computerized provider order entry portal. The AR was only displayed during the intervention, and providers were required to acknowledge the AR to finalize the request. Subsequently, the unstructured free text information accompanying all requests was reviewed, and the AR was revised when possible. The percentage of unrated requests and the overall AR, before and after radiologist review, were compared between periods and by provider type., Results: CDS software prospectively generated an AR for only 25.4% and 28.4% of baseline and intervention imaging requests, respectively; however, radiologist review generated an AR for 82.4% and 93.6% of the same requests. During the respective periods, the percentage of baseline and intervention imaging requests considered appropriate was 18.7% and 22.9% by prospective CDS software rating and increased to 82.4% and 88.7% with radiologist review., Conclusion: Despite limited effective use of CDS software, the percentage of requests containing additional, relevant clinical information increased, and the majority of requests had overall high appropriateness when reviewed by a radiologist. Additional work is needed to improve the amount and quality of clinical information available to CDS software and to facilitate the entry of this information by appropriate end users., (Copyright © 2017 American College of Radiology. Published by Elsevier Inc. All rights reserved.)

Retrospective data mining has tremendous potential in research but is time and labor intensive. Current data mining software contains many advanced search features but is limited in its ability to identify patients who meet multiple complex independent search criteria. Simple keyword and Boolean search techniques are ineffective when more complex searches are required, or when a search for multiple mutually inclusive variables becomes important. This is particularly true when trying to identify patients with a set of specific radiologic findings or proximity in time across multiple different imaging modalities. Another challenge that arises in retrospective data mining is that much variation still exists in how image findings are described in radiology reports. We present an algorithmic approach to solve this problem and describe a specific use case scenario in which we applied our technique to a real-world data set in order to identify patients who matched several independent variables in our institution's picture archiving and communication systems (PACS) database.

Purpose: To examine the effect of integrating point-of-care clinical decision support (CDS) using the ACR Appropriateness Criteria (AC) into an inpatient computerized provider order entry (CPOE) system for advanced imaging requests., Methods: Over 12 months, inpatient CPOE requests for nuclear medicine, CT, and MRI were processed by CDS to generate an AC score using provider-selected data from pull-down menus. During the second 6-month period, AC scores were displayed to ordering providers, and acknowledgement was required to finalize a request. Request AC scores and percentages of requests not scored by CDS were compared among primary care providers (PCPs) and specialists, and by years in practice of the responsible physician of record., Results: CDS prospectively generated a score for 26.0% and 30.3% of baseline and intervention requests, respectively. The average AC score increased slightly for all requests (7.2 ± 1.6 versus 7.4 ± 1.5; P < .001), for PCPs (6.9 ± 1.9 versus 7.4 ± 1.6; P < .001), and minimally for specialists (7.3 ± 1.6 versus 7.4 ± 1.5; P < .001). The percentage of requests lacking sufficient structured clinical information to generate an AC score decreased for all requests (73.1% versus 68.9%; P < .001), for PCPs (78.0% versus 71.7%; P < .001), and for specialists (72.9% versus 69.1%; P < .001)., Conclusions: Integrating CDS into inpatient CPOE slightly increased the overall AC score of advanced imaging requests as well as the provision of sufficient structured data to automatically generate AC scores. Both effects were more pronounced in PCPs compared with specialists., (Copyright © 2015 American College of Radiology. Published by Elsevier Inc. All rights reserved.)