RSNA 2021 Artificial Intelligence Preview

Road to RSNA 2021: Artificial Intelligence Preview

By Erik L. Ridley, staff writer
November 15, 2021

Welcome to the first installment of this year's Road to RSNA preview of the RSNA 2021 meeting, which will return to McCormick Place after a one-year hiatus due to the COVID-19 pandemic.

We're pleased to once again provide a modality-by-modality overview of select scientific presentations to serve as your guide to this year's meeting, whether you're making the trip to Chicago or attending virtually on your computer.

Fittingly, our journey along the Road to RSNA begins with our preview of artificial intelligence (AI). Although much has changed in radiology around the world over the last two years, AI's dominant role at the RSNA meeting remains unchallenged. As usual, AI research will be the focus of a variety of dedicated scientific sessions, as well as sprinkled throughout the scientific program at RSNA 2021.

AI is increasingly being investigated for its potential utility in predicting patient outcomes and guiding treatment. For example, researchers will share how AI can inform steroid treatment in COVID-19 patients, predict survival in patients with lung cancer, forecast all-cause mortality from total-body dual-energy x-ray absorptiometry (DEXA) exams, and evaluate treatment response in patients with hepatocellular carcinoma.

Presenters will also detail progress being made in utilizing deep-learning algorithms to glean more information from imaging studies, including to analyze bone mineral density from hip radiographs, detect health disparities in COVID-19 patients, and predict patient demographic information from chest x-rays. Of course, AI's growing utility in breast imaging applications will be featured in a number of presentations across all relevant modalities, including digital breast tomosynthesis.

On the nuts and bolts side of things, AI continues to demonstrate promise as a tool to speed up imaging exam times -- up to 70% faster for spine MRI studies -- and reduce dose, including 90% less dose needed for gadolinium-based contrast agents. Federated learning is also increasingly demonstrating value as a way to train high-quality, generalizable algorithms without having to share training data between institutions.

See below for previews of these and other AI-related scientific talks at this year's RSNA meeting. Of course, these are just a sample of the content on offer; many other scientific presentations, scientific posters, hot topic sessions, refresher courses, and educational exhibits on AI topics also await attendees. For more information on those presentations and other abstracts, view the RSNA 2021 meeting program.

You'll also want to check out the AI Showcase, which will be located in the South Hall of McCormick Place this year. As of early November, 93 vendors were scheduled to showcase their wares in this dedicated area on the exhibit floor.

Those visiting the AI Showcase in person will also have the opportunity to visit the RSNA's Imaging AI in Practice interactive exhibit, which will feature 22 vendors demonstrating AI technologies and the integration standards needed to embed AI into the diagnostic radiology workflow. Featuring 32 different products, the interactive exhibit will showcase the use of AI and health IT standards throughout the radiology workflow in real-world scenarios such as ischemic stroke and lung nodules, according to the RSNA.

AI differentiates colorectal polyps on CT colonography
Sunday, November 28 | 10:30 a.m.-11:30 a.m. | SSGI01-4 | Room TBA
This proof-of-concept study will demonstrate the potential for artificial intelligence (AI) technology to differentiate colorectal polyps spotted on CT colonography exams.
Combining radiomics, AI with PET/MRI helps assess nodal status
Sunday, November 28 | 10:30 a.m.-11:30 a.m. | SSNMMI01‐3 | Room TBA
Here, researchers will talk about how applying radiomics and machine learning to FDG-PET/MRI can noninvasively assess nodal status and treatment planning for breast cancer patients.
AI captures bone mineral density data from hip x-rays
Sunday, November 28 | 1:00 p.m.-2:00 p.m. | SSMK03-1 | Room TBA
In this talk, researchers will show how bone mineral density data can be obtained from hip radiographs by an artificial intelligence (AI) algorithm.
Models predict mortality risk on total-body DEXA exams
Sunday, November 28 | 1:00 p.m.-2:00 p.m. | SSMK03-3 | Room TBA
Artificial intelligence can spot changes in body composition over time on total-body dual-energy x-ray absorptiometry (DEXA) exams and predict mortality risk, according to this presentation.
AI catches missed limb, pelvic bone fractures on x-rays
Sunday, November 28 | 1:00 p.m.-2:00 p.m. | SSMK03-5 | Room TBA
An artificial intelligence (AI) algorithm can detect limb or pelvic bone fractures that were initially missed on radiographs, according to this study from France.
Algorithm enhances survival predictions in lung cancer
Monday, November 29 | 9:30 a.m.-10:00 a.m. | SSCH03-2 | Room TBA
Artificial intelligence-based assessment of biological chest age on chest radiographs can help to predict survival in lung cancer patients, according to this presentation.
Federated learning facilitates algorithm training
Monday, November 29 | 9:30 a.m.-10:00 a.m. | SSGI05-2 | Room TBA
In this talk, researchers will share their success in using federated learning to train deep-learning algorithms for liver and tumor segmentation on hepatic CT exams.
Algorithm helps monitor HCC treatment response
Monday, November 29 | 9:30 a.m.-10:30 a.m. | SSGI06-5 | Room TBA
A deep learning-based software application can improve the objectivity of treatment response evaluation in patients with hepatocellular carcinoma (HCC), according to this presentation.
Can AI help guide management of COVID-19 patients?
Monday, November 29 | 1:30 p.m.-2:30 p.m. | SSPH05-2 | Room TBA
In this session, researchers will discuss the potential of artificial intelligence (AI) for supporting treatment decisions in COVID-19 patients.
Coronary stenosis algorithm bolsters reproducibility
Monday, November 29 | 3:00 p.m.-4:00 p.m. | SSCA04-2 | Room TBA
Coronary stenosis evaluation on coronary CT angiography can be improved with the assistance of an artificial intelligence algorithm, according to this presentation.
AI reduces false positives on breast ultrasound
Monday, November 29 | 3:00 p.m.-4:00 p.m. | SSBR04-6 | Room TBA
Researchers will present their findings from a clinical trial that used artificial intelligence (AI) in an effort to reduce false positives on breast ultrasound.
Deep-learning algorithm predicts survival for COPD
Tuesday, November 30 | 9:30 a.m.-10:30 a.m. | SSCH04-2 | Room TBA
A deep-learning algorithm can enable significantly more accurate survival predictions in patients with chronic obstructive pulmonary disease (COPD), according to this study.
'Plug-and-play' AI method helps with microcalcifications on DBT
Tuesday, November 30 | 9:30 a.m.-10:30 a.m. | SSPH09-6 | Room TBA
Researchers will present results from a "plug-and-play" artificial intelligence (AI) algorithm that they say has the potential to improve detection of microcalcifications on digital breast tomosynthesis (DBT) images.
AI finds there's more on chest x-rays than meets the eye
Tuesday, November 30 | 3:00 p.m.-4:00 p.m. | SSIN05-6 | Room TBA
A patient's biological age, sex, ethnicity, and insurance status can be predicted by an artificial intelligence (AI) model's analysis of his or her chest radiograph, according to this presentation.
Explainable artificial intelligence elevates prostate MRI reads
Wednesday, December 1 | 8:00 a.m.-8:30 a.m. | SSGU05-4 | Room TBA
An explainable deep-learning software application can be highly accurate for characterizing prostate lesions on biparametric MRI exams, according to this scientific presentation.
AI illuminates health disparities in COVID-19 patients
Wednesday, December 1 | 9:30 a.m.-10:30 a.m. | SSIN06-1 | Room TBA
In this talk, researchers will describe how an artificial intelligence (AI) algorithm can reveal racial or ethnic disparities in COVID-19 patients based on their chest x-rays.
Deep learning identifies CT biomarkers that help diagnose type 2 diabetes
Wednesday, December 1 | 9:30 a.m.-10:30 a.m. | SSGI11-3 | Room TBA
Deep learning can identify CT biomarkers that help detect and predict type 2 diabetes in patients undergoing CT for other indications, according to this presentation.
Deep-learning algorithm triages emergency head CTs
Wednesday, December 1 | 1:30 p.m.-2:30 p.m. | SSMS05-3 | Room TBA
A deep-learning algorithm trained on only normal head CT exams can triage emergency cases for priority review by radiologists, according to this presentation.
Deep-learning tool triages women with decreased breast density
Wednesday, December 1 | 3:00 p.m.-4:00 p.m. | SSBR09-4 | Room TBA
In this Wednesday talk, researchers will present findings from their study of nearly 2,700 women that used a convolutional neural network in predicting mammographic density percentage from MRI scans.
AI model shows promise in detecting lesions from DBT data
Wednesday, December 1 | 3:00 p.m.-4:00 p.m. | SSBR09-3 | Room TBA
In this talk, researchers will discuss an artificial intelligence (AI) algorithm that detects lesions using datasets from digital breast tomosynthesis (DBT) exams.
Can AI enable 90% lower gadolinium dose in brain MRI?
Thursday, December 2 | 9:30 a.m.-10:30 a.m. | SSNR14-2 | Room TBA
An artificial intelligence (AI) algorithm can enable high-quality contrast-enhanced brain MRI exams to be acquired using only 10% of the normal dose of a gadolinium-based contrast agent, according to this multicenter clinical study.
Deep learning meets need for speed in spine MRI
Thursday, December 2 | 9:30 a.m.-10:30 a.m. | SSNR14-5 | Room TBA
In this session, researchers will show how a deep learning-based image reconstruction method can deliver up to 72% faster spine MRI scan times along with perceived improvements in image quality.
AI-based measurements increase utility of cardiac MRI
Thursday, December 2 | 9:30 a.m.-10:30 a.m. | SSCA10-1 | Room TBA
Automated cardiac MRI measurements correlate better with invasive pulmonary hemodynamics than manual measurements and could also help in predicting mortality, according to this study.