Más información con Karla Gonzalez, +52 1 55 2862 5590
Más información con Karla Gonzalez, +52 1 55 2862 5590
When it comes to imaging the pelvis, radiologists agree that ultrasound (US) is the right tool for diagnosis. However, opinions diverge on when MRI should be used, and the topic was debated during a Wednesday Controversy Session.
“In a skilled hand and using all the techniques that are available to the sonologist, ultrasound can localize sources of pelvic pain and discomfort and performs equally, if not better than MRI,” said Beryl Benacerraf, MD, from Brigham and Women’s Hospital and Harvard Medical School in Boston.
Deborah Levine, MD, from Beth Israel Deaconess Medical Center and Harvard Medical School in Boston, agreed that US is where radiologists should begin and finish with pelvic imaging, but added that in some instances more information is needed. “Pelvic MRI is a problem solving tool that should be used when an US is inconclusive about diagnosis or doesn’t provide enough information to confirm treatment direction,” Dr. Levine said.
For Dr. Benacerraf, technical advances, such as 3-D, color Doppler and real-time transvaginal dynamic US, make US an effective imaging tool for gynecologic patients. “These additional technologies add value, but only when the radiologist is in the room examining the patient simultaneously and uses the advances to tailor the imaging based on what the patient says,” Dr. Benacerraf said.
There are instances though, countered Dr. Levine, when US doesn’t provide a sufficiently complete picture of the nature of a mass or when there are complications during pregnancy.
“MRI can give additional information about the nature of a mass that may change the decision to perform surgery, such as when the mass is an exophytic fibroid,” Dr. Levine said. “When uterine artery embolization is planned, MRI can show the exact number, size and location of fibroids or guide decisions regarding a hysteroscopy or laparoscopic approach.”
In pregnancy, Dr. Levine noted, MRI is helpful when there is a suspicion of fetal abnormality or prior to fetal surgery, when the surgeon wants to be sure that the abnormality being treated is the only one present.
In conclusion, Dr. Levine affirmed that there is a place for MRI when imaging the female pelvis. “When surgical intervention is being considered, MRI can provide additional information that confirms or alters the direction of treatment,” she said. “It can also assist with patient counseling, in situations where more detail is needed, or where reassurance is needed about the conservative management of benign masses.”
According to Dr. Benacerraf, US can provide a diagnostic answer without the need for further testing. “If we continue as a profession to end so many ultrasound reports with ‘MRI is recommended,’ requesting physicians and payers may become disenchanted with ultrasound as a modality thus decreasing its value,” she said. “Consequently we need to provide adequate training and maintenance of competency in order to ensure that all ultrasound procedures are performed with high quality resulting in confident diagnoses.”
By Elizabeth Gardner
A standing-room-only crowd turned out to hear foremost experts discuss how artificial intelligence (AI) and machine learning (ML) are impacting radiology at Tuesday’s RSNA/American Association of Physicists in Medicine (AAPM) symposium.
The discipline now upending radiology as we know it — applying big-data processing power and techniques to digital images — is developing rapidly, but will need the trained brains of radiologists to make it into a daily tool for diagnosis and treatment.
“There is a lot of excitement and also a lot of questions,” said Paul E. Kinahan, PhD, moderator of the symposium, “Machine Learning in Radiology: Why and How?”
Dr. Kinahan, vice chair of radiology research and head of the Imaging Research Laboratory at the University of Washington, Seattle, noted that dozens of exhibitors at this year’s meeting, many of them exhibiting for the first time, are offering AI-based products.
During the session “Harnessing Artificial Intelligence,” presenter Keith Dreyer, DO, gave a high-level explanation of the complexities involved in teaching computers to read images. The audience was split among radiologists, physicists and vendors of AI tools.
“Machines are getting smarter faster than people are,” said Dr. Dreyer, vice chair of radiology and director of the Center for Clinical Data Science at Massachusetts General Hospital, Boston, and chair of the American College of Radiology’s Commission on Informatics.
He said early science-fiction depictions of AI that all but replace the human brain have given way to less complex, but immediately useful ML, which, for example, tells Amazon customers what other products they might be interested in based on past purchases.
“It’s not as sexy as AI, but it is a necessary foundation for AI to take hold,” Dr. Dreyer said.
Getting Involved in AI
A major obstacle to developing AI for imaging is the lack of what Dr. Dreyer called a “healthcare AI ecosystem.” He said radiology needs universally accepted ways to develop and incorporate AI, similar to the DICOM image standard, in order to make it easy for developers to create new applications and integrate them into imaging devices and clinical information systems.
Dr. Dreyer urged audience members to get involved in the first important task of developing medical imaging use cases. One of his slides showed a matrix of thousands of tiny squares, each one representing the intersection of a radiology specialty, an imaging modality, a part of the body and the lab and/or pathology findings that inform the imaging study. Each square is a possible use case for AI that will need clinicians to help develop the rules.
AI for Cancer Treatment
In his presentation, “Assistive AI for Cancer Treatment,” Antonio Criminisi, PhD, a principal researcher at Microsoft in Cambridge, U.K., described the company’s InnerEye project and how it can help pinpoint the location of tumors for targeted radiotherapy.
The technology is based on a principle similar to the one that drives the Microsoft Kinect game system, which senses human movement and “builds” a replica of the body inside the game. Dr. Criminisi’s team is teaching the InnerEye software to analyze pelvic images and identify anatomical structures, particularly the prostate, to speed the now laborious task of locating exactly which spots to radiate.
In another example, the team combined images of a brain tumor from six different imaging modalities in order to quantify changes in tumor volume.
“This technology is not for doing things that you already know how to do well, but to do things that you wish you were able to do,” Dr. Criminisi said.
But perfecting AI isn’t as important as getting it to work well enough for daily use, Dr. Dreyer said, noting the broad spectrum of accuracy among human radiologists.
“Do you need AI to be at the top?” he asked. “Or would it be enough for it to be somewhere in the middle, which is easier to achieve? I would argue there are many places on the globe where it would be adequate to have a good solution.”
Radiology began as a discipline of capturing images, and, like photography, has evolved by emulating and improving on how the eye works. But advancements such as compression and image analysis algorithms and artificial intelligence (AI) are quickly transforming radiology into a discipline that instead mirrors how the brain works.
These new developments may someday render radiology images — as we know them today — irrelevant, according to Daniel K. Sodickson, MD, PhD, who on Monday delivered this year’s New Horizons Lecture, “A New Light: The Birth, and Rebirth, of Imaging.”
“I’m pleased to announce the death of the MR protocol,” said Dr. Sodickson, “Not quite yet, because there is lot of work to do and there will always be need for tailored studies to answer a particular question. But MRI is a little like art photography now: Lie still, don’t move, hold your breath, and do it again. It’s not a very modern paradigm.”
Radiologists, on the other hand, will maintain their value, he added.
“We are more than just our images,” Dr. Sodickson said.
Dr. Sodickson, chair of the National Institutes of Health study section on biomedical imaging technology, is credited with founding the field of parallel imaging, which allows distributed arrays of detectors to gather MR images at previously inaccessible speeds.
In his lecture, he predicted that imaging studies will become less like still photography and more like streaming video. Scanners that acquire information from many different angles at the same time will capture patients’ data continuously and algorithms will select it as needed to reconstruct images for specific purposes.
“You can see a moving abdomen and the flow of contrast,” said Dr. Sodickson, vice chair for research in the Department of Radiology, director of the Bernard and Irene Schwartz Center for Biomedical Imaging, and a professor of radiology, physiology and neuroscience at NYU School of Medicine, in the NYU Langone Health System in New York City. “You can freeze the heart and look at respiratory function, or track the coronary arteries and freeze them at any point you like.”
An Alphabet Soup of Algorithms
This revolution in physical modeling is changing the way radiologists interpret image information and connect organ-level maps to underlying cellular architecture and molecular composition.
At the same time, an alphabet soup of AI algorithms are deriving new information from sometimes very low quality image data, and may someday change the way imaging devices are designed, Dr. Sodickson said. In fact, recent implementations of AI for image reconstruction have begun to resemble the neural processing of complex, continuous sensory data streams.
He compared this change to the difference between looking at a single image, which gives a single stream of visual information, and being at a live concert, which generates multiple streams of information that the brain quickly sorts through, eliminating unnecessary information and pinpointing what to focus on.
Dr. Sodickson predicted that techniques such as MR fingerprinting (which isolates unique information in MR images that might be used to identify specific tissue types, cell types, or diseases) will take image information out of the realm of subjectivity, depending on the skill of the reader and the technician and the features of the scanner, and make it as objective as a blood test.
MR fingerprints will be available in a database and radiologists will look for a match to a specific patient’s information, just as police identify criminals by running their fingerprints against the FBI database.
Where do these changes leave radiologists? In a good place, Dr. Sodickson said.
He suggested that radiologists begin thinking of themselves as “information innovators,” due to their expertise in image data acquisition. The field is already headed in that direction, said Dr. Sodickson, pointing to the vast increase in sessions on machine learning and AI offered at major radiological meetings — including RSNA — in the past year.
“It’s a turbulent age for imaging — there is no doubt,” Dr. Sodickson said, “But I hope I have convinced you it is a golden age for innovation.”
Along with being persistent innovators, tomorrow’s radiologists must work to establish themselves as imaging, information science and image-guided therapeutics experts who will play a vital role on healthcare teams, said Roderic I. Pettigrew, PhD, MD, in an opening session lecture on Sunday in Arie Crown Theater.
In his presentation, “Tomorrow’s Radiology,” Dr. Pettigrew, founding director of the National Institute of Biomedical Imaging and Bioengineering, stressed that the overall goal of today’s healthcare enterprise is to achieve healthy longevity — to be born healthy, acquire no significant disease and to reach the end of life without pain or suffering from disease.
“That bold vision requires technological innovation for earlier precision diagnostics and therapeutics,” Dr. Pettigrew said. “And tomorrow’s radiology will play a critical role in achieving this goal.
“We emphasize innovation because we realize that like imagination, there is no end to innovation,” Dr. Pettigrew said.
For example, he referred to Peter Basser, PhD, of the National Institutes of Health and a team of researchers who invented and developed MR diffusion-tensor imaging, which is now used to map the human connectome.
“His latest innovation actually quantifies neural conduction latency,” Dr. Pettigrew said. “This has the potential to offer new insights into some of our most challenging healthcare conditions in the brain, such as mental health and psychiatric illness.”
Dr. Pettigrew also discussed a 4-D motion compensation MRI technique that allows for the in-utero assessment of cerebral function.
“This technique is able to depict the functional neural network involved in inward thinking,” Dr. Pettigrew said. “This is the so-called ‘default’ mode network believed to be disturbed in autism and related to emotional development.”
A technique like this could allow for the evaluation of a number of factors related to development, ranging from something as simple as a person’s diet to the impact of listening to music, Dr. Pettigrew said.
Other areas in which imaging innovations have had a significant effect on health outcomes include the use of CT to evaluate the risk for coronary artery disease, mixed-reality MRI-guided planning for breast conserving surgery and the potential of MRI-ultrasound (US) in treating prostate cancer via focal laser ablation.
Radiologic Innovations Critical to the Future
Radiologists are also demonstrated innovators. For example, Dr. Pettigrew referred to a program at Massachusetts General Hospital, Boston, called the Radiology Consultation Clinic in which patients meet directly with radiologists to review images with the idea of getting them to take more responsibility for their own health and to follow their physicians’ advice.
“This is the kind of role and impact we can have,” Dr. Pettigrew said, adding that this type of innovation is directly tied to the role radiologists will play going forward — that of imaging and data science experts who are active members of the healthcare team.
This shift is demonstrated by some of the themes highlighted in recent Radiology articles, such as integrated diagnostics, radiomics, machine learning and artificial intelligence.
“The key message here is that modern imaging science is information and data science,” Dr. Pettigrew said. “Extracting the data from these images will improve the value proposition in imaging.”
And to do so, he added, tomorrow’s radiologists must follow a clear path.
“Radiologists must leverage the digital revolution, the state of hyperconnectivity we are in, big data science and artificial intelligence.”
RSNA’s Annual Meeting Preview highlights the presentations, exhibits and attractions planned in 2017.