By Eric Barnes, staff writer
November 8, 2016

If the brave new world of advanced visualization, 3D, and computer-aided detection (CAD) that will be on display at RSNA 2016 has a theme, it's that radiologists aren't working alone anymore.

The evolution of digital visualization and detection tools and the rise of artificial intelligence are approaching the point where machines will work more as partners than assistants in radiology. And this smart help comes not a minute too soon, considering the explosion of images, technologies, protocols, costs, and challenging demographics that define today's healthcare environment.

Today's CAD schemes, increasingly aided by machine learning algorithms and crunched by fast processors, are continually improving their ability to pore over shapes and textures and patterns and inhomogeneities to find signs of pathology, saving time and healthcare dollars.

Biomarkers, a term denoting almost any biological process doctors might want to look at, are a perfect match for smart detection algorithms that don't get tired or bored, even if the attention span of the human operator should start to flag. More investigators are now including them in the search for disease.

Meanwhile, 3D printing of pathology, organs, and other spare parts is jumping in to offer tools and teaching aids to supplement the hands-on training that used to be covered by years of experience, cadavers, and human teachers with time to impart their wisdom. Highly functional 3D printed organs may still be a ways off, but few seem to doubt that they are coming.

Even today, the well-functioning systems that support detection, diagnosis, and decision-making offer a bit of salvation for radiology, freeing practitioners to think about larger questions surrounding patient care.

The challenges of ensuring that these systems help rather than hinder are significant, ranging from integration and workflow problems to managing the false positives and false negatives that put patient care at risk. But the sheer breadth of questions today's systems address, and the scope of the progress they have made, is real.

To take just a few examples from this year's sessions, assessing liver fibrosis is pretty straightforward now, say radiologists from Wisconsin, who built a semiautomated tool to look for surface nodularity at CT. 4D flow MRI can quickly assess ventricular function, noninvasively and without radiation.

In lung nodule detection, a homegrown 3D lung CAD scheme from Japan can pick up the smallest ones at even smaller doses when advanced iterative reconstruction is turned on, with no increase in false-positive detections at ultralow doses. Or maybe you need to know about the genetics of your patient's non-small cell lung cancer. Radiogenomics mapping with CT can tell you.

Parametric MR mapping of prostate cancer is helping radiologists in Italy detect more clinically significant lesions and speed up workflow. If you overlay 3D electroanatomic mapping on a 3D heart, ablation for tachycardia becomes more intuitive.

A little time spent learning about 3D, maximum intensity projections, and cinematic rendering could be well spent. Cinematic rendering is a photorealistic 3D rendering technique that enables a more in-depth 3D evaluation of the underlying anatomy, presenters say.

There's even a session on using today's tools optimally to make your and your patients' lives better while dealing with a deluge of images. And isn't that what it's all about?

Learn more about these sessions in the previews below. To view RSNA's complete listing of abstracts for this year's scientific and educational program, click here.

Scientific and Educational Presentations
CT software noninvasively stages liver fibrosis
Sunday, November 27 | 11:05 a.m.-11:15 a.m. | SSA09-03 | Room E450B
Researchers from the University of Wisconsin-Madison have developed a semiautomated tool to assess for liver surface nodularity corresponding with liver fibrosis. The tool can predict both the presence of liver fibrosis and the disease stage.
Should 4D flow replace standard MRI for LV function assessment?
Sunday, November 27 | 12:05 p.m.-12:15 p.m. | SSA03-09 | Room S502AB
4D flow is a volumetric cardiac MR technique that in 10 short minutes provides both anatomic and flow information over the cardiac cycle in a nonbreath-hold acquisition. Is it ready to replace standard MRI for left ventricular (LV) assessment?
4D CT kinematic assessment boosts radioulnar joint view
Sunday, November 27 | 12:05 p.m.-12:15 p.m. | SSA15-09 | Room S406B
4D CT with a multiplanar reformatted technique is useful for evaluating motion of the distal radioulnar joint, according to this study being presented on Sunday afternoon.
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.
IR boosts lung CAD detection at all dose levels
Monday, November 28, 2016 | 10:50 a.m.-11:00 a.m. | SSC03-03 | Room S404CD
Japanese researchers found huge dose reductions combined with high nodule detection sensitivity when they used advanced iterative reconstruction (IR) with their newly developed computer-aided detection (CAD) scheme for lung nodules.
Radiogenomics mapping joins lung nodule features with genes
Monday, November 28 | 11:20 a.m.-11:30 a.m. | SSC08-06 | Room S402AB
Researchers from Stanford University built a radiogenomics map that linked high-level metagenes detailing lung cancer pathways with CT images, opening a window on predicting survival.
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.
CAD boosts MRI proficiency for prostate cancer
Monday, November 28 | 11:50 a.m.-12:00 p.m. | SSC08-09 | Room S402AB
Adding computer-aided detection (CAD) results in greater sensitivity and interreader agreement for radiologists interpreting multiparametric MRI scans for prostate cancer, as well as reduced reading times, according to researchers from Italy.
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.
How to build a collaborative 3D printing lab
Monday, November 28 | Approximately 1:30 p.m.-2:00 p.m. | RCC23D | Room S501ABC
This course, the last in Monday's 3D Printing: Clinical Applications I session that runs from 12:30 p.m. to 2:00 p.m., is a how-to for creating a collaborative 3D printing lab in today's environment of rapidly accelerating technology development.
3D printing improves understanding of congenital heart disease
Monday, November 28 | 2:30 p.m.-4:00 p.m. | RCB24 | Room S401CD
This hands-on 3D printing course will explore how to recognize congenital heart disease, the most common significant birth defect and one that usually requires surgical treatment. This particular session focuses on the double-outlet right ventricle in all its forms and permutations.
Implementing quality control is essential with 3D printing
Monday, November 28 | 3:10 p.m.-3:20 p.m. | SSE13-02 | Room S404CD
As 3D printing has surged in popularity, quality control (QC) has become an essential partner in the development process. Radiologists at the University of Pittsburgh Medical Center have built a comprehensive and methodical QC plan for the task.
3D heart models take on cardiac ablation
Tuesday, November 29 | 9:50 a.m.-10:00 a.m. | RC303-06 | Room S504AB
Hoping to build a better interventional technique for doctors treating ventricular tachycardia with ablation guided by electroanatomic mapping (EAM), Italian researchers used CT to create a 3D heart model for each patient, overlaying EAM on the CT images to guide ablation. Hint: It works better than EAM alone.
How to turn image files into 3D-printed models
Tuesday, November 29 | 10:30 a.m.-12:00 p.m. | RCA32 | Room S401AB
This hands-on course taught by 3D printing experts will use practical examples to help participants learn how to convert cross-sectional images into 3D-printed models.
4D CT brain segmentation offers subtle clues in stroke
Wednesday, November 30 | 10:30 a.m.-10:40 a.m. | SSK15-01 | Room N229
Dutch researchers have successfully segmented gray and white matter in the brain using 4D CT. The technique provides a subtle assessment of the health of the brain for stroke treatment, and it could be useful in further segmentation and anatomic labeling, according to the group.
3D printing for congenital heart disease: Crisscross heart
Wednesday, November 30 | 2:30 p.m.-4:00 p.m. | RCB44 | Room S401CD
This hands-on 3D printing course will look at congenital heart disease, focusing on crisscross or twisted heart and related conditions.
Quantitative CT analysis boosts spinal metastasis detection
Wednesday, November 30 | 3:20 p.m.-3:30 p.m. | SSM18-03 | Room N229
Epidural spine metastases are defined as a radiologic critical finding because any delay in reporting may result in substantially increased morbidity. But how can imaging be optimized to find more of them?
Advanced visualization and the cinematic 3D rendering prototype
Thursday, December 1 | All day | PH125-ED-X | Lakeside, PH Community
Despite continuous evolution in postprocessing techniques, most 3D visualization and planning still occur on a very basic visual level. This all-day educational event covers the basics of maximum intensity projections and volume rendering techniques, and it goes on to discuss the more visually rich world of cinematic rendering of 3D volumes.
Surviving today's environment while making the most of advanced imaging
Thursday, December 1 | 8:30 a.m.-10:00 a.m. | RC653 | Room S403B
This thoughtful and much-needed session on practical informatics for the practicing radiologist aims to help radiologists use today's tools to deal with a wave of new technologies and rising workloads, while staying sane and happy.