Facebook Twitter LinkedIn Email Print
By Erik L. Ridley, AuntMinnie staff writer
October 31, 2016

Welcome to the first installment of this year's Road to RSNA preview of the 2016 RSNA meeting in Chicago. For the eighth year in a row, we're providing a modality-by-modality overview of the most important scientific sessions to serve as your guide to events at McCormick Place.

Our journey along the Road to RSNA begins with our preview of Imaging Informatics, and specifically PACS, deep learning, and teleradiology. Coverage of presentations on related topics such as structured reporting, critical results management, clinical decision support, radiation dose monitoring software, and other analytics topics will be included in our upcoming Healthcare IT section. Meanwhile, 3D and computer-aided detection (CAD) will be the focus of our Advanced Visualization preview, which will also be a later stop on the Road to RSNA.

Deep learning is by far the hottest area in imaging informatics research today, and the dizzying pace of innovation in this field will be fully evident at RSNA 2016. Worried that deep learning will soon make radiologists obsolete? Don't be. Researchers are largely focusing on applying deep-learning technology to improve the practice of radiology and aid radiologists, not replace them.

For example, presenters will describe how deep-learning technology shows potential in applications such as providing automated analysis of image quality, helping to assess risk for breast cancer, differentiating breast tumors, and screening for osteoporosis on CT exams. Other areas of interest include applying deep learning to avoid unnecessary breast biopsies, identify prostate cancer, and uncover biomarkers for disease, as well as detect and label anatomy and disease on imaging studies.

These methods are also being applied to predict whether a patient is likely to miss their radiology appointment and to extract acute findings from radiology reports. In addition to the plethora of scientific presentations on the schedule, the RSNA has also planned a number of other events geared toward machine learning. These include an "Eyes of Watson" demonstration of IBM's technology platform, which will allow attendees to make diagnoses on imaging cases and then watch Watson perform real-time processing on the case. There will also be a hands-on workshop and a controversy session on machine learning.

On the PACS side, attendees will be able to take in presentations on topics such as correlating pathology and radiology results within PACS and also a PACS software tool that improves accuracy for measuring lesions.

See below for previews of these and other imaging informatics-related scientific papers and posters at this year's RSNA meeting. Of course, these are just a sample of the content on offer; a host of refresher courses and educational exhibits on a wide variety of imaging informatics topics also await those who make the trip to Chicago. For more information on those talks and other abstracts in this year's scientific and educational program, click here.

Scientific and Educational Presentations
Machine learning can assess MR image quality
Sunday, November 27 | 11:25 a.m.-11:35 a.m. | SSA22-05 | Room S405AB
German researchers will explain in this scientific presentation how machine-learning technology can handle the challenge of assessing the image quality of MR studies.
Inadequate history may not affect emergency radiology reads
Sunday, November 27 | 11:35 a.m.-11:45 a.m. | SSA06-06 | Room N226
In this presentation, researchers from a teleradiology services provider will share how a lack of adequate clinical history -- surprisingly -- did not lead to more interpretation errors on emergency radiology cases.
It helps to correlate radiology, pathology in PACS
Sunday, November 27 | 12:05 p.m.-12:15 p.m. | SSA12-09 | Room S403A
In this scientific presentation, researchers will describe how correlating radiology and pathology results within their PACS software yields time savings and efficiency gains for radiologists who perform image-guided procedures.
Deep learning can find, label images in PACS
Sunday, November 27 | 12:30 p.m.-1:00 p.m. | IN200-SD-SUA1 | Lakeside, IN Community, Station 1
In this poster presentation, researchers will highlight the potential of deep-learning techniques for categorizing and labeling images stored on PACS archives.
Deep learning may help assess breast cancer risk
Monday, November 28 | 9:00 a.m.-9:10 a.m. | RC215-03 | Arie Crown Theater
Deep learning may be able to help clinicians evaluate mammographic parenchymal patterns for assessing breast cancer risk, according to researchers from the University of Chicago.
Deep learning + CADx combo performs best in breast tumors
Monday, November 28 | 11:30 a.m.-11:40 a.m. | SSC08-07 | Room S402AB
In this talk, researchers will describe how the combination of deep convolutional neural networks and computer-assisted diagnosis (CADx) software yields improved diagnostic performance in differentiating breast tumors on full-field digital mammography and ultrasound.
AI offers value as teaching tool for radiologists
Monday, November 28 | 11:40 a.m.-11:50 a.m. | SSC08-08 | Room S402AB
Researchers in this session will share how artificial intelligence (AI) technology can be used to teach people how to think like a radiologist and overcome common biases during image interpretation.
Deep learning shows promise for reading chest x-rays
Monday, November 28 | 12:15 p.m.-12:45 p.m. | IN211-SD-MOA2 | Lakeside, IN Community, Station 2
A deep-learning method could be used to provide a more "human-like" diagnosis on chest x-rays, according to a group from the U.S. National Institutes of Health Clinical Center.
Deep learning can detect osteoporosis on CT exams
Monday, November 28 | 12:15 p.m.-12:45 p.m. | IN212-SD-MOA3 | Lakeside, IN Community, Station 3
Artificial intelligence based on deep-learning techniques may be able to automatically screen for osteoporosis in routine abdominal CT exams, according to this poster presentation.
Deep learning may be able to avoid some breast biopsies
Monday, November 28 | 3:20 p.m.-3:30 p.m. | SSE02-03 | Room E450A
Deep-learning technology has the potential to help decrease the number of unnecessary biopsies performed on benign breast microcalcifications, according to this scientific presentation.
Natural language processing, machine learning extract acute findings on reports
Tuesday, November 29 | 10:50 a.m.-11:00 a.m. | SSG07-03 | Room S402AB
In this presentation, researchers will share how natural language processing and machine-learning technology can extract acute findings from radiology reports.
Deep learning detects, labels vertebrae on lumbar MRI
Tuesday, November 29 | 12:15 p.m.-12:45 p.m. | IN227-SD-TUA4 | Lakeside, IN Community, Station 4
A new study has shown that deep-learning technology can detect and label vertebrae on lumbar MRI studies.
Deep learning shows promise for spotting prostate cancer
Tuesday, November 29 | 12:15 p.m.-12:45 p.m. | IN229-SD-TUA6 | Lakeside, IN Community, Station 6
In this poster presentation, researchers will highlight the potential of deep learning in detecting clinically significant prostate cancer on multiparametric MRI scans.
Deep learning may help spot interstitial lung disease
Tuesday, November 29 | 12:45 p.m.-1:15 p.m. | IN235-SD-TUB3 | Lakeside, IN Community, Station 3
This poster presentation will describe how deep learning may offer assistance in the challenging task of identifying interstitial lung disease.
Analytics helps spot risk for missed appointments
Tuesday, November 29 | 3:30 p.m.-3:40 p.m. | SSJ13-04 | Room S402AB
In this session, researchers will describe how predictive analytics can help identify patients who may be at higher risk for missing their radiology appointment.
Deep learning can segment, measure lymph node clusters
Wednesday, November 30 | 10:30 a.m.-10:40 a.m. | SSK17-01 | Room S404AB
In this scientific session, researchers will reveal how an automated method based on deep learning shows promise for segmenting and measuring the volume of lymph node clusters.
Deep learning can label partially annotated images
Wednesday, November 30 | 12:15 p.m.-12:45 p.m. | IN242-SD-WEA3 | Lakeside, IN Community, Station 3
In this poster presentation, a study team will show how deep-learning techniques can be used to fully segment and label disease regions on images that have only been partially annotated.
Deep learning could help uncover disease biomarkers
Wednesday, November 30 | 12:45 p.m.-1:15 p.m. | IN247-SD-WEB2 | Lakeside, IN Community, Station 2
This poster presentation will illustrate the potential of deep-learning methods for finding imaging biomarkers in relatively uncommon diseases, such as nasopharyngeal cancer.
PACS display features affect diagnostic performance
Wednesday, November 30 | 3:20 p.m.-3:30 p.m. | SSM12-03 | Room S403A
Italian researchers will report in this session on the display characteristics and settings that most affect the identification of cerebral infarction on brain CT exams.
Can display features improve DR interpretations?
Wednesday, November 30 | 3:50 p.m.-4:00 p.m. | SSM12-06 | Room S403A
In this scientific presentation, an Italian group will show how higher spatial resolution and luminance levels on medical displays may be able to improve radiologist performance in reading digital radiography (DR) images.
PACS software tool boosts lesion measurement accuracy
Thursday, December 1 | 11:40 a.m.-11:50 a.m. | SSQ10-08 | Room S403A
In this talk, researchers will share how a PACS-integrated software tool can help radiologists be more efficient and accurate in documenting lesion measurements.