The Resident Roentgen File curates multimedia educational resources from our vast portfolio of benefits to help ARRS In-Training Members prepare for exams, fellowship, and beyond.
Interest in artificial intelligence (AI) applications for radiology has grown dramatically in recent years. In particular, one deep learning architecture—the convolutional neural network (CNN)—has shown strong diagnostic potential in medical imaging. However, CNNs and other similar deep learning methods have several intrinsic challenges to their implementation that both medical professionals and developers should be keenly aware of. An article from the latest issue of ARRS’ glossy member magazine and website, InPractice—coauthored by Ian Chan, ARRS Resident Advisory Subcommittee Chair—expertly reviewed the internal structures of CNNs, while detailing several less discussed challenges shared with related AI methodologies and their implementation in clinical practice.
Each month, AJR Editor in Chief Andrew Rosenkrantz handpicks an article from “the yellow journal” specifically for ARRS In-Training Members. This month’s AJR Editor’s Choice article looks at the clinical importance of incidental homogeneous renal masses at portal venous phase CT.
Authors’ rationale—Homogeneous renal masses are commonly encountered on contrast-enhanced CT and may lead to unnecessary recommendations for additional imaging.
Sample—Three hundred patients with 346 homogeneous well-defined renal masses measuring 10–44 mm and 21–39 HU.
Imaging test—Portal venous phase CT.
Data source—Retrospective review at 12 institutions.
Outcome—Presence of a clinically important renal mass, defined as a solid mass without fat, a mass classified as Bosniak IIF or higher, or confirmed malignancy.
Reference standard—Renal mass protocol CT or MRI, ultrasound showing definitive characterization, or at least 5 years of imaging or clinical follow-up.
What the authors found—None of the masses were clinically important based on the reference standard; 72% measured 21–30 HU, and 28% measured 31–39 HU.
Authors’ insight—A threshold of greater than 30 HU, rather than greater than 20 HU, is warranted for recommending follow-up imaging for small homogeneous renal masses seen on portal venous phase CT.
Our Abstract Submission site is now open for Oral Presentations, Scientific Online Posters, and Educational Online Posters to be considered for the ARRS 2022 Annual Meeting, May 1–5, in New Orleans, virtually, and on demand. The deadline to submit your original abstract for our 2022 Annual Meeting is August 31, 2021.
Also specially selected for ARRS In-Training Members, July’s Web Lecture is Nonpulmonary Thoracic Masses. Residents and fellows will review non-lung cancer tumors for which FDG PET/CT is indicated for initial staging; describe the impact of imaging features of pleural and chest wall tumors on treatment and prognosis; and review the most common differential diagnoses of anterior, middle, and posterior pediatric mediastinal masses.
Our brand-new, open-access toolkit for in-training imagers and practicing radiologists alike, RadTeams.org helps hone your skills for establishing, growing, and sustaining high-functioning teams—an essential part of providing effective, safe, and high-level care to patients.
Featuring didactic lectures and interactive case-based reviews presented by accomplished faculty, ARRS In-Training Members enjoy anytime, anywhere access to every session of both these symposia for one full year, until September 9, 2022.
Practical Pediatric Imaging Symposium—September 8–9, 2021
Abdominal MRI: Practical Applications and Advanced Imaging—September 9–10, 2021
We invite residents and fellows to submit to the ARRS Gold Medal Story Series, which will include member narratives from around the world—giving names and faces to those imaging professionals who conquered the day-to-day challenges of battling COVID-19.
How Residents and Fellows Can Join ARRS for FREE:
- Visit the ARRS Membership page, select In-Training Members, and complete the application online.
- Fill out the ARRS In-Training Membership Application, and mail or fax it back to us.