ISLE: An Intelligent Streaming Framework for High-Throughput AI Inference in Medical Imaging.

As the adoption of artificial intelligence (AI) systems in radiology grows, the increase in demand for greater bandwidth and computational resources can lead to greater infrastructural costs for healthcare providers and AI vendors. To that end, we developed ISLE, an intelligent streaming framework to address inefficiencies in current imaging infrastructures. Our framework draws inspiration from video-on-demand platforms to intelligently stream medical images to AI vendors at an optimal resolution for inference from a single high-resolution copy using progressive encoding. We hypothesize that ISLE can dramatically reduce the bandwidth and computational requirements for AI inference, while increasing throughput (i.e., the number of scans processed by the AI system per second). We evaluate our framework by streaming chest X-rays for classification and abdomen CT scans for liver and spleen segmentation and comparing them with the original versions of each dataset. For classification, our results show that ISLE reduced data transmission and decoding time by at least 92% and 88%, respectively, while increasing throughput by more than 3.72 × . For both segmentation tasks, ISLE reduced data transmission and decoding time by at least 82% and 88%, respectively, while increasing throughput by more than 2.9 × . In all three tasks, the ISLE streamed data had no impact on the AI system's diagnostic performance (all P > 0.05). Therefore, our results indicate that our framework can address inefficiencies in current imaging infrastructures by improving data and computational efficiency of AI deployments in the clinical environment without impacting clinical decision-making using AI systems.

Economic and Environmental Costs of Cloud Technologies for Medical Imaging and Radiology Artificial Intelligence.

Radiology is on the verge of a technological revolution driven by artificial intelligence (including large language models), which requires robust computing and storage capabilities, often beyond the capacity of current non-cloud-based informatics systems. The cloud presents a potential solution for radiology, and we should weigh its economic and environmental implications. Recently, cloud technologies have become a cost-effective strategy by providing necessary infrastructure while reducing expenditures associated with hardware ownership, maintenance, and upgrades. Simultaneously, given the optimized energy consumption in modern cloud data centers, this transition is expected to reduce the environmental footprint of radiologic operations. The path to cloud integration comes with its own challenges, and radiology informatics leaders must consider elements such as cloud architectural choices, pricing, data security, uptime service agreements, user training and support, and broader interoperability. With the increasing importance of data-driven tools in radiology, understanding and navigating the cloud landscape will be essential for the future of radiology and its various stakeholders.

Exploring the Clinical Translation of Generative Models Like ChatGPT: Promise and Pitfalls in Radiology, From Patients to Population Health.

Generative artificial intelligence (AI) tools such as GPT-4, and the chatbot interface ChatGPT, show promise for a variety of applications in radiology and health care. However, like other AI tools, ChatGPT has limitations and potential pitfalls that must be considered before adopting it for teaching, clinical practice, and beyond. We summarize five major emerging use cases for ChatGPT and generative AI in radiology across the levels of increasing data complexity, along with pitfalls associated with each. As the use of AI in health care continues to grow, it is crucial for radiologists (and all physicians) to stay informed and ensure the safe translation of these new technologies.

Clinical Impact and Generalizability of a Computer-Assisted Diagnostic Tool to Risk-Stratify Lung Nodules With CT.

OBJECTIVE: To evaluate whether an imaging classifier for radiology practice can improve lung nodule classification and follow-up. METHODS: A machine learning classifier was developed and trained using imaging data from the National Lung Screening Trial (NSLT) to produce a malignancy risk score (malignancy Similarity Index [mSI]) for individual lung nodules. In addition to NLST cohorts, external cohorts were developed from a tertiary referral lung cancer screening program data set and an external nonscreening data set of all nodules detected on CT. Performance of the mSI combined with Lung-RADS was compared with Lung-RADS alone and the Mayo and Brock risk calculators. RESULTS: We analyzed 963 subjects and 1,331 nodules across these cohorts. The mSI was comparable in accuracy (area under the curve = 0.89) to existing clinical risk models (area under the curve = 0.86-0.88) and independently predictive in the NLST cohort of 704 nodules. When compared with Lung-RADS, the mSI significantly increased sensitivity across all cohorts (25%-117%), with significant increases in specificity in the screening cohorts (17%-33%). When used in conjunction with Lung-RADS, use of mSI would result in earlier diagnoses and reduced follow-up across cohorts, including the potential for early diagnosis in 42% of malignant NLST nodules from prior-year CT scans. CONCLUSION: A computer-assisted diagnosis software improved risk classification from chest CTs of screening and incidentally detected lung nodules compared with Lung-RADS. mSI added predictive value independent of existing radiological and clinical variables. These results suggest the generalizability and potential clinical impact of a tool that is straightforward to implement in practice.

Criteria for the translation of radiomics into clinically useful tests.

Computer-extracted tumour characteristics have been incorporated into medical imaging computer-aided diagnosis (CAD) algorithms for decades. With the advent of radiomics, an extension of CAD involving high-throughput computer-extracted quantitative characterization of healthy or pathological structures and processes as captured by medical imaging, interest in such computer-extracted measurements has increased substantially. However, despite the thousands of radiomic studies, the number of settings in which radiomics has been successfully translated into a clinically useful tool or has obtained FDA clearance is comparatively small. This relative dearth might be attributable to factors such as the varying imaging and radiomic feature extraction protocols used from study to study, the numerous potential pitfalls in the analysis of radiomic data, and the lack of studies showing that acting upon a radiomic-based tool leads to a favourable benefit-risk balance for the patient. Several guidelines on specific aspects of radiomic data acquisition and analysis are already available, although a similar roadmap for the overall process of translating radiomics into tools that can be used in clinical care is needed. Herein, we provide 16 criteria for the effective execution of this process in the hopes that they will guide the development of more clinically useful radiomic tests in the future.

Medical Student Perspectives on the Impact of Artificial Intelligence on the Practice of Medicine.

INTRODUCTION: Concerns about radiologists being replaced by artificial intelligence (AI) from the lay media could have a negative impact on medical students' perceptions of radiology as a viable specialty. The purpose of this study was to evaluate United States of America medical students' perceptions about radiology and other medical specialties in relation to AI. METHODS: An anonymous, web-based survey was sent to 32 radiology interest groups at United States medical schools. The survey was comprised of 6 questions assessing medical student perceptions of AI and its potential impact on radiology and other medical specialties. Responses were voluntary and collected over a 6-month period from November 2017 to April 2018. RESULTS: A total of 156 students responded with representation from each year of medical school. Over 75% agreed that AI would have a significant role in the future of medicine. Most (66%) agreed that diagnostic radiology would be the specialty most greatly affected. Nearly half (44%) reported that AI made them less enthusiastic about radiology. The majority of students (57%) obtained their information about AI from online articles. Thematic analysis of free answer comments revealed mostly neutral comments towards AI, however, the negative responses were the strongest and most detailed. CONCLUSIONS: US medical students believe that AI will play a significant role in medicine, particularly in radiology. However, nearly half are less enthusiastic about the field of radiology due to AI. As the majority receive information about AI from online articles, which may have negative sentiments towards AI's impact on radiology, formal AI education and medical student outreach may help combat misinformation and help prevent the dissuading of medical students who might otherwise consider the specialty.

Interpreting Radiographs with Concurrently Obtained Patient Photographs.

A technology for automatically obtaining patient photographs along with portable radiographs was implemented clinically at a large academic hospital. This article highlights several cases in which image-related clinical context, provided by the patient photographs, provided quality control information regarding patient identification, laterality, or position and assisted the radiologist with the interpretation. The information in the photographs can easily minimize unnecessary calls to the patient's nursing staff for clarifications and can lead to new methods of physically assessing patients. Published under a CC BY 4.0 license.

CT-Based 3D Printing of the Glenoid Prior to Shoulder Arthroplasty: Bony Morphology and Model Evaluation.

To demonstrate the 3D printed appearance of glenoid morphologies relevant to shoulder replacement surgery and to evaluate the benefits of printed models of the glenoid with regard to surgical planning. A retrospective review of patients referred for shoulder CT was performed, leading to a cohort of nine patients without arthroplasty hardware and exhibiting glenoid changes relevant to shoulder arthroplasty planning. Thin slice CT images were used to create both humerus-subtracted volume renderings of the glenoid, as well as 3D surface models of the glenoid, and 11 printed models were created. Volume renderings, surface models, and printed models were reviewed by a musculoskeletal radiologist for accuracy. Four fellowship-trained orthopaedic surgeons specializing in shoulder surgery reviewed each case individually as follows: First, the source CT images were reviewed, and a score for the clarity of the bony morphologies relevant to shoulder arthroplasty surgery was given. The volume rendering was reviewed, and the clarity was again scored. Finally, the printed model was reviewed, and the clarity again scored. Each printed model was also scored for morphologic complexity, expected usefulness of the printed model, and physical properties of the model. Mann-Whitney-Wilcoxon signed rank tests of the clarity scores were calculated, and the Spearman's ρ correlation coefficient between complexity and usefulness scores was computed. Printed models demonstrated a range of glenoid bony changes including osteophytes, glenoid bone loss, retroversion, and biconcavity. Surgeons rated the glenoid morphology as more clear after review of humerus-subtracted volume rendering, compared with review of the source CT images (p = 0.00903). Clarity was also better with 3D printed models compared to CT (p = 0.00903) and better with 3D printed models compared to humerus-subtracted volume rendering (p = 0. 00879). The expected usefulness of printed models demonstrated a positive correlation with morphologic complexity, with Spearman's ρ 0.73 (p = 0.0108). 3D printing of the glenoid based on pre-operative CT provides a physical representation of patient anatomy. Printed models enabled shoulder surgeons to appreciate glenoid bony morphology more clearly compared to review of CT images or humerus-subtracted volume renderings. These models were more useful as glenoid complexity increased.

Will machine learning end the viability of radiology as a thriving medical specialty?

There have been tremendous advances in artificial intelligence (AI) and machine learning (ML) within the past decade, especially in the application of deep learning to various challenges. These include advanced competitive games (such as Chess and Go), self-driving cars, speech recognition, and intelligent personal assistants. Rapid advances in computer vision for recognition of objects in pictures have led some individuals, including computer science experts and health care system experts in machine learning, to make predictions that ML algorithms will soon lead to the replacement of the radiologist. However, there are complex technological, regulatory, and medicolegal obstacles facing the implementation of machine learning in radiology that will definitely preclude replacement of the radiologist by these algorithms within the next two decades and beyond. While not a comprehensive review of machine learning, this article is intended to highlight specific features of machine learning which face significant technological and health care systems challenges. Rather than replacing radiologists, machine learning will provide quantitative tools that will increase the value of diagnostic imaging as a biomarker, increase image quality with decreased acquisition times, and improve workflow, communication, and patient safety. In the foreseeable future, we predict that today's generation of radiologists will be replaced not by ML algorithms, but by a new breed of data science-savvy radiologists who have embraced and harnessed the incredible potential that machine learning has to advance our ability to care for our patients. In this way, radiology will remain a viable medical specialty for years to come.

Machine Meets Biology: a Primer on Artificial Intelligence in Cardiology and Cardiac Imaging.

PURPOSE OF REVIEW: An understanding of the basics concepts of deep learning can be helpful in not only understanding the potential applications of this technique but also in critically reviewing literature in which neural networks are utilized for analysis and modeling. RECENT FINDINGS: The term "deep learning" has been applied to a subset of machine learning that utilizes a "neural network" and is often used interchangeably with "artificial intelligence." It has been increasingly utilized in healthcare for computational "learning", especially for pattern recognition for diagnostic imaging. Another promising application is the potential for these neural networks to improve the accuracy in the identification of patients who are at risk for cardiovascular events and could benefit most from preventive treatment in comparison with more conventional statistical techniques. The importance of such tailored cardiovascular risk assessment and disease management in individual patients is far reaching given that cardiovascular disease is the leading cause of morbidity and mortality in the world. Nearly half of myocardial infarctions and strokes occur in patients who are not predicted to be at risk for cardiovascular events by current guideline-based approaches. Equally important are individuals who are not at risk for cardiovascular events and yet are given expensive and unnecessary preventive treatment with potential untoward side effects. The application of powerful artificial intelligence/deep learning tools in medicine is likely to result in more effective and efficient health care delivery with the potential for significant cost savings by shifting preventative treatment from inappropriate to appropriate patient subgroups.

Reinventing Radiology: Big Data and the Future of Medical Imaging.

PURPOSE: Today, data surrounding most of our lives are collected and stored. Data scientists are beginning to explore applications that could harness this information and make sense of it. MATERIALS AND METHODS: In this review, the topic of Big Data is explored, and applications in modern health care are considered. RESULTS: Big Data is a concept that has evolved from the modern trend of "scientism." One of the primary goals of data scientists is to develop ways to discover new knowledge from the vast quantities of increasingly available information. CONCLUSIONS: Current and future opportunities and challenges with respect to radiology are provided with emphasis on cardiothoracic imaging.

Finding the sweet spot for metformin in 18F-FDG-PET.

PURPOSE: The effect of oral hypoglycemic agents on fluorine-18-flurodeoxyglucose (F-FDG) uptake is less understood than the effect of exogenous insulin. In this study, the effect of withholding metformin on F-FDG distribution in subsequent PET imaging was evaluated. PATIENTS AND METHODS: A retrospective observational study was carried out. A total of 15 patients taking metformin were grouped according to the time interval from the last dose of metformin to F-FDG-PET. Those who received PET after stopping metformin for less than 24 h were compared with those stopping metformin 24-48 h before PET. The F-FDG uptake and PET image fidelity for these groups were compared qualitatively and the associated blood glucose was recorded. The average standardized uptake value of the liver, tongue, and subcutaneous tissues among the groups were also compared. RESULTS: The F-FDG-PET distribution and image quality were found to be the best at time points greater than 24 h following metformin dose. There was significantly increased F-FDG uptake in the liver and tongue and tongue-to-liver ratio in patients who had metformin within 24 h of F-FDG-PET compared with those who last had metformin greater than 24 h before F-FDG-PET; however, the F-FDG uptake in the subcutaneous tissues was unchanged. CONCLUSION: Less than 24 h between metformin dose and F-FDG-PET resulted in increased muscle and fat uptake in addition to increased bowel uptake. Abnormal F-FDG uptake can limit the diagnostic quality of an exam and affect F-FDG uptake in cancer. The emerging role of biguanides in the treatment of cancer calls for more personalized standardization for F-FDG-PET in the presence of oral hypoglycemic agents.

RSNA Diagnosis Live: A Novel Web-based Audience Response Tool to Promote Evidence-based Learning.

Audience response systems have become more commonplace in radiology residency programs in the last 10 years, as a means to engage learners and promote improved learning and retention. A variety of systems are currently in use. RSNA Diagnosis Live™ provides unique features that are innovative, particularly for radiology resident education. One specific example is the ability to annotate questions with subspecialty tags, which allows resident performance to be tracked over time. In addition, deficiencies in learning can be monitored for each trainee and analytics can be provided, allowing documentation of resident performance improvement. Finally, automated feedback is given not only to the instructor, but also to the trainee. Online supplemental material is available for this article. © RSNA, 2017.

Use of Radiology Procedure Codes in Health Care: The Need for Standardization and Structure.

Radiology procedure codes are a fundamental part of most radiology workflows, such as ordering, scheduling, billing, and image interpretation. Nonstandardized unstructured procedure codes have typically been used in radiology departments. Such codes may be sufficient for specific purposes, but they offer limited support for interoperability. As radiology workflows and the various forms of clinical data exchange have become more sophisticated, the need for more advanced interoperability with use of standardized structured codes has increased. For example, structured codes facilitate the automated identification of relevant prior imaging studies and the collection of data for radiation dose tracking. The authors review the role of imaging procedure codes in radiology departments and across the health care enterprise. Standards for radiology procedure coding are described, and the mechanisms of structured coding systems are reviewed. In particular, the structure of the RadLex™ Playbook coding system and examples of the use of this system are described. Harmonization of the RadLex Playbook system with the Logical Observation Identifiers Names and Codes standard, which is currently in progress, also is described. The benefits and challenges of adopting standardized codes-especially the difficulties in mapping local codes to standardized codes-are reviewed. Tools and strategies for mitigating these challenges, including the use of billing codes as an intermediate step in mapping, also are reviewed. In addition, the authors describe how to use the RadLex Playbook Web service application programming interface for partial automation of code mapping. © RSNA, 2017.

Relations of blood pressure and head injury to regional cerebral blood flow.

Hypertension confers increased risk for cognitive decline, dementia, and cerebrovascular disease. These associations have been attributed, in part, to cerebral hypoperfusion. Here we posit that relations of higher blood pressure to lower levels of cerebral perfusion may be potentiated by a prior head injury. Participants were 87 community-dwelling older adults - 69% men, 90% white, mean age=66.9years, 27.6% with a history of mild traumatic brain injury (mTBI) defined as a loss of consciousness ≤30min resulting from an injury to the head, and free of major medical (other than hypertension), neurological or psychiatric comorbidities. All engaged in clinical assessment of systolic and diastolic blood pressure (SBP, DBP) and single photon emission computed tomography (SPECT). Computerized coding of the SPECT images yielded relative ratios of blood flow in left and right cortical and select subcortical regions. Cerebellum served as the denominator. Sex-stratified multiple regression analyses, adjusted for age, education, race, alcohol consumption, smoking status, and depressive symptomatology, revealed significant interactions of blood pressure and head injury to cerebral blood flow in men only. Specifically, among men with a history of head injury, higher systolic blood pressure was associated with lower levels of perfusion in the left orbital (ß=-3.21, p=0.024) and left dorsolateral (ß=-2.61, p=0.042) prefrontal cortex, and left temporal cortex (ß=-3.36, p=0.014); higher diastolic blood pressure was marginally associated with lower levels of perfusion in the left dorsolateral prefrontal cortex (ß=-2.79, p=0.051). Results indicate that men with a history of head injury may be particularly vulnerable to the impact of higher blood pressure on cerebral perfusion in left anterior cortical regions, thus potentially enhancing risk for adverse brain and neurocognitive outcomes.