Journal highlights
The following are highlights from the current issues of RSNA’s peer-reviewed journals.
Protocol Guidance for Pediatric Photon-Counting CT
Performing CT in children comes with unique challenges such as greater degrees of patient motion, smaller and densely packed anatomy, and potential risks of radiation exposure. Since photon-counting detector (PCD) CT results in decreased radiation dose and noise, it is important to review the relevant technical aspects and principles specific to protocol development on the new PCD CT platform to realize the potential benefits for this population.
In an article published in Radiology, Kelly K. Horst, MD, Mayo Clinic, Rochester, MN, and colleagues address clinical challenges of PCD CT in pediatric patients by outlining fundamental terminology, explaining technical aspects and providing protocol guidance for the first commercially available, and only Food and Drug Administration-cleared, PCD CT scanner.
Rather than being evidence-based, the PCD CT protocols provided represent the recommendations of radiologists and medical physicists from four pediatric imaging centers. The authors note that images generated from the PCD CT scanner offer a fundamentally different representation of detected photon energy compared with prior energy-integrating detector (EID) CT technology.
“PCD CT represents a significant technical advance in CT imaging of patients, especially the very young,” the authors conclude.
Read the full article, “Multi-institutional Protocol Guidance for Pediatric Photon-counting CT.” Follow the Radiology editor on X @RadiologyEditor.
Noncontrast images of the chest in a 15-year-old patient using (A) conventional energy-integrating detector (EID) CT (section thickness, 1 mm; scan time, 2.2 sec; volume CT dose index, 5.68 mGy) with fast-pitch scan mode (Flash; Siemens Healthineers) and (B) photon-counting detector CT (section thickness, 0.2 mm; scan time, 1.2 seconds; 0.25-second rotation time; pitch, 3.2; volume CT dose index, 2.38 mGy) performed a few months later with fast-pitch scan mode (Flash + ultrahigh resolution [UHR]).
Complete legend details can be found at https://doi.org/10.1148/radiol.231741 © RSNA 2024.
An Update on the Imaging, Management and Prevention of Human Papillomavirus-Related Neoplasms
Five percent of cancers globally can be attributed to human papillomavirus (HPV), the most common sexually transmitted infection. HPV reproduces in flat cells with irregular boundaries, or squamous epithelium, and is the most common source of viral-related tumors.
There are low-risk and high-risk subtypes of HPV. Cancers of the cervix, vagina, vulva, anal canal, rectum and oropharynx are the common HPV-related malignancies caused by high-risk subtypes.
In an article published in RadioGraphics, Venkata S. Katabathina, MD, University of Texas Health, San Antonio, and colleagues discuss the epidemiologic and pathologic conditions of HPV-related malignancies and emphasize the current updates in the World Health Organization (WHO) classification. The authors also describe the impact of vaccination and screening programs.
“Imaging plays a pivotal role in the diagnosis, staging, assessment of treatment response, and surveillance of HPV-related malignancies. Novel immunotherapies to increase T-cell response are being investigated, and updated screening and vaccination guidelines are being implemented to improve the overall outcomes in HPV-related cancers,” the authors conclude.
Read the full article, “Update on Pathologic Conditions, Imaging Findings, Prevention, and Management of Human Papillomavirus-Related Neoplasms,” at RSNA.org/RadioGraphics. This article is also available for CME on EdCentral or at RSNA.org/Learning-Center. Follow the RadioGraphics editor on X @RadG_Editor.
HPV-positive squamous cell carcinoma of the right tongue base in a 64-year-old man. (A) Axial contrast-enhanced CT image shows a small enhancing mass in the right tongue base (arrow). (B) Axial FDG PET/CT fusion image shows diffusely increased metabolic activity in the right neck lymph nodes (arrows). Pathologic examination confirmed HPV-positive SCC. The TNM staging is T1N1, which portends an excellent prognosis.
https://doi.org/10.1148/rg.230179 © RSNA 2024
Extracting Information from Free-Text Radiology Reports
Large language models (LLMs) are uniquely adaptable to new tasks and show untested potential to harness large, unstructured radiology report databases.
In an article published in Radiology: Artificial Intelligence, Bastien Le Guellec, MS, Centre Hospitalier Universitaire de Lille, France, and colleagues sought to assess the performance of a local open-source LLM in various information extraction tasks from real-life emergency brain MRI reports.
Two radiologists identified MRI scans performed in the emergency department of a French quaternary center for headaches, and four radiologists scored the reports’ conclusions as either normal or abnormal, labeling abnormalities as either headache-causing or incidental.
The LLM reviewed 2,398 emergency brain MRI free-text reports and achieved high performance metrics for detecting the presence of headache in the clinical context, detecting contrast medium injection use in the protocol, and categorizing the studies as either normal or abnormal. The authors note that the LLM also performed causal inference between a radiologic finding and a symptom with 82% accuracy.
“An open-source LLM was able to extract information from free-text radiology reports with excellent accuracy without requiring further training,” the authors conclude.
Read the full article, “Performance of an Open-Source Large Language Model in Extracting Information from Free-Text Radiology Reports,” and the related commentary, “A New Era of Text Mining in Radiology with Privacy-Preserving LLMs.” Follow the Radiology: Artificial Intelligence editor on X @RadiologyAI.
RSNA Journals Impact Factors Announced
RSNA Journals demonstrate impressive influence according to the impact factor numbers in the newly released 2024 Clarivate Analytics Journal Citation Reports. Impact factor measures the relevance and influence of academic journals based on citation data.
Radiology, the most cited journal in its field, ranked first in journal citation indicator and second in impact factor among 204 journals in the Radiology, Nuclear Medicine and Medical Imaging category. It received an impact factor of 12.1 and was cited 59,748 times.
Published regularly since 1923, Radiology, is edited by Linda Moy, MD.
RadioGraphics earned an impact factor of 5.2 with 15,185 citations. The popular journal, which first earned an impact factor in 2015, is edited by Christine Cooky Menias, MD.
This is also the second consecutive year that RSNA’s subspecialty journals, Radiology: Artificial Intelligence, Radiology: Cardiothoracic Imaging and Radiology: Imaging Cancer have achieved impact factors.
Radiology: Imaging Cancer, edited by Gary D. Luker, MD, jumped to a 5.6 impact factor with citations increasing from 342 in 2023 to 491 this year.
Radiology: Artificial Intelligence, edited by Charles E. Kahn, Jr., MD, earned an 8.1, also reflecting a year over year increase in citations from 1,483 to 2,163.
Radiology: Cardiothoracic Imaging, edited by Suhny Abbara, MD, earned a 3.8 impact factor with 1,226 citations.
Find the Radiology suite of journals at RSNA.org/Journals.