Despite the current “perfect storm” of political, economic, and social confusion, radiologists can breathe easier, knowing their profession is conscientiously working with other groups to improve the industry and healthcare overall.

“Within the house of radiology, we come from different perspectives, practices, and locations,” said Howard B. Fleishon, M.D., MMM, chair of the American College of Radiology (ACR) in a memo published in the May 28 

Journal of the American College of Radiology.

“But, we are all part of the community of radiology. We are stronger together.”

To further augment the profession’s position as a leader in healthcare, he said, the ACR has forged partnerships with multiple other organizations, bringing together skills and talents for a number of causes that can improve patient care. He outlined several initiatives that highlight these efforts.

ACR and RSNA partnership to advise the U.S. Food & Drug Administration about autonomous artificial intelligence

ACR, RSNA, and American Association of Physicists in Medicine collaboration to create the Medical Imaging Data Resource Center

Radiology Heath Equity Coalition creation by ACR, American Board of Radiology, American Roentgen Ray Society, Association of University Radiologists, RSNA, Society of Chairs of Academic Radiology Departments, and Society of Interventional Radiology

ACR representation at the World Health Organization, International Atomic Energy Association, International Society for Strategic Studies in Radiology, and International Society of Radiology

ACR representation at the G7 for artificial intelligence policy development

Lung Cancer Coalition with the American Cancer Society, the National Lung Cancer Round Table, the American Lung Association, and GO2 Foundation for Lung Cancer

Partnership for joint recommendations for breast cancer patients with the American Society of Breast Surgeons, National Accreditation Program for Breast Centers, the National Comprehensive Cancer Network, the Commission on Cancer, and the American College of Surgeons

Advocacy for CT colonography insurance coverage through collaboration with the Colon Cancer Alliance, Medical Imaging and Technology Alliance, and the American Gastroenterological Association

It is through these initiatives and partnerships, he said, that radiology can help healthcare navigate the broader issues for the benefit of patients.

“Speaking with one voice, we will stand out from all the noise and confusion we find around us,” he concluded. “Standing together, we have the opportunity to promote our profession and serve our patients. Stronger together as the ACR and the radiology community.”

For more coverage based on industry expert insights and research, subscribe to the Diagnostic Imaging e-newsletter

Articles

In the midst of a “perfect storm,” radiology can see the benefits of community solidarity.

Rad Strong: Through Collaborations, “We Are Stronger Together”

Relying on shortcuts for work isn’t always a great idea – and that holds true for artificial intelligence (AI) with chest X-rays and COVID-19, as well.

New research from the University of Washington (UW), shows that AI, like humans, has a tendency to lean on shortcuts for disease detection with these scans. If the tools are deployed clinically, investigators said, the result could be diagnostic errors that impact real patients.

Rather than learning from actual medical pathology and clinically significant indicators, the team, led by Paul G. Allen School of Computer Science & Engineering doctoral students, Alex DeGrave, who is also a medical student in the UW Medical Scientist Training Program, and Joseph Janizek, also a UW medical student, showed that algorithms used during the pandemic relied on text markers and patient positioning specific to each dataset to predict whether someone was COVID-19-positive.

The team published their results May 31 in 

“A physician would generally expect a finding of COVID-19 from an X-ray to be based on specific patterns in the image that reflect disease processes,” DeGrave said. “But, rather than relying on those patterns, a system using shortcut learning might, for example, judge that someone is elderly and, thus, infer that they are more likely to have the disease because it is more common in older patients. The shortcut is not wrong per se, but the association is unexpected and not transparent. And, that could lead to an inappropriate diagnosis.”

By depending on these shortcuts, the team said, the algorithms are far less likely to work outside of their original setting where they were trained and tested. Models that do use shortcuts have a greater chance of failure when used in a different facility, opening the door for potentially incorrect diagnoses and improper treatment, he added.

Chest X-ray Plus Artificial Intelligence Can Predict Worst Outcomes for COVID-19-Positive Patients

It’s the lack of transparency – the “black box” phenomenon associated with AI – that contributes heavily to the problem, the team said. And, they suspected it existed with the models touted as detecting COVID-19, so they conducted their own training and testing to assess the models’ trustworthiness. Specifically, they focused on “worst-case confounding,” a situation, resulting in shortcuts, that arises when an AI tool lacks sufficient training data to learn the underlying pathology of a disease.

“Worst-case confounding is what allows an AI system to just learn to recognize datasets instead of learning any true disease pathology,” Janizek said. “It’s what happens when all of the COVID-19-positive cases come from a single dataset while all of the negative cases are in another. And, while researchers have come up with techniques to mitigate associations like this in cases where those associations are less severe, these techniques don’t work in situations where you have a perfect association between an outcome, such as COVID-19 status, and a factor like the data source.”

For their study, the team trained several deep convolutional neural networks on X-ray images from a dataset that replicated the approach that was used in existing papers. In a two-step process, they tested each model’s performance with an internal image set from the initial dataset that was not part of the training data, and they tested performance on a second, external dataset that was meant to be a new hospital system.

Their analysis showed that once models were removed from their original setting – where they showed high performance – they floundered. In outside environments, accuracy fell by half. DeGrave and Janizek’s team called this the “generalization gap,” asserting that confounding factors were behind the models’ predictive success with the initial dataset.

The team also applied explainable AI techniques in an attempt to pinpoint image features that played the largest role in each model’s predictions. They had similar results when they tried to reduce confounding by training the models with a second dataset that included both positive and negative COVID-19 images pulled from similar sources.

These results, the team said, shed light on the degree to which high-performance medical AI system can depend on undesirable, unreliable shortcuts for disease detection. They also contradict the belief that confounding isn’t as much of a danger when datasets come from similar sources. Fortunately, DeGrave said, it is unlikely that the models tested in this study were deployed clinically, largely because providers still use RT-PCR for COVID-19 diagnosis rather than chest X-rays.

Overall, said senior author Su-In Lee, Ph.D., associate professor in the Allen School, the findings from this study underscore the fundamental role explainable AI will play in ensuring the safety and efficacy of these models in medical decision-making.

“Our findings point to the importance of applying explainable AI techniques to rigorously audit medical AI systems,” he said. “If you look at a handful of X-rays, the AI system might appear to behave well. Problems only become clear once you look at many images. Until we have methods to more efficiently audit these systems using a greater sample size, a more systematic application of explainable AI could help researchers avoid some of the pitfalls we identified with the COVID-9 models.”

For more coverage based on industry expert insights and research, subscribe to the Diagnostic Imaging e-Newsletter 

Research shows artificial intelligence models rely on shortcuts for detecting COVID-19.

Shortcuts Are Bad – Even for AI

Due to advancements in catheter-based therapies in interventional radiology, conditions that once required large open surgeries, months of recovery, and had many associated complications, are now being treated with minimally invasive procedures by radiologists. One such intervention is uterine fibroid embolization (UFE). Although UFE has been performed for almost 25 years, the procedure continues to evolve.

Uterine fibroids are the most common benign tumor of reproductive age women, and they can cause debilitating symptoms. In many instances, conservative medical management fails or is associated with an unfavorable side effect profile. In these cases, the gold standard has been either myomectomy or hysterectomy. However, both are major invasive surgeries, associated with significant costs, long hospitalizations/recovery, and significant morbidity. Given recent concerns regarding long-term effects on cardiovascular and hormonal health, uterine sparing therapies have grown in popularity.

UFE has emerged as a minimally invasive alternative after first being performed in the mid 1990s. The techniques have been borrowed from transcatheter embolization of the uterine arteries in pelvic trauma/bleeding. In this instance, the temporary occlusion of the uterine arteries supplying the uterus results in infarction of uterine fibroids. UFE has been shown to have numerous benefits, including shorter hospital stays, earlier return to normal activities, and the ability to use conscious sedation rather than general anesthesia

Despite the fact that UFE has been shown to be a safe and effective treatment, the procedure remains under-utilized. A reported 65 times more hysterectomies are performed per year in the United States compared to UFE interventions

. This is partly due to patient awareness. A recent Society of Interventional Radiology report shows that only 62 percent of women have even heard of UFE treatment. Of the women who have heard of UFE, approximately three-fourths of those patients had not heard about it from their gynecologist

. Interventional radiologists need to play a larger role in the multidisciplinary team when discussing treatment options for symptomatic fibroids. There also needs to be a greater emphasis on patient awareness and marketing moving for UFE forward.

Treatment options for patients who wish to preserve their uterus and future fertility include myomectomy and UFE. In 2020, the FEMME trial (A Randomized Trial of Treating Fibroids with Either Embolization or Myomectomy to Measure the Effect on Quality of Life Among Women Wishing to Avoid Hysterectomy) was published in the 

 directly comparing the two therapies. This study served as the first robust multicenter, randomized label trial evaluating the topic

. The study demonstrated a four-point difference in the quality-of-life questionnaire score in favor of myomectomy at two years (84.6 vs 80.0), which was statistically significant but likely not a clinically meaningful difference. Furthermore, UFE patients had shorter hospital stay, quicker return to work, fewer bleeding adverse events, and similar pregnancy rates to myomectomy patients. Lastly, they also both had similar results in regard to willingness of women to have their particular procedure again

One of the main barriers to widespread adoption of UFE is post-procedural pain. Initial post-embolization pain can last up to 24 hours and can be severe in nature. Most commonly, it is managed with controlled analgesia or epidural anesthesia, requiring hospital admission. Secondary pain can last days after the initial procedure, and is commonly managed with narcotics. Recently, Superior Hypogastric Nerve Block (SHNB) interventions have been performed in conjunction with UFE and have been shown to be effective post-procedural pain suppressants. This enhanced procedure reduces overall pain scores, need for opioids, and allows for same day discharge

. The widespread use of SHNB may alleviate patient fears regarding the procedure and allow for further adoption of UFE as a first line therapy for symptomatic fibroids.

In summary, UFE has proven to be a safe and highly effective treatment for symptomatic uterine fibroids. Recent studies have proven that it compares favorably with myomectomy. Post-embolization pain has been a barrier to widespread adoption. However, the recent use of combined pain interventions, such as Superior Hypogastric Nerve Blocks in conjunction with UFE, have been shown as highly effective in reducing pain, ultimately allowing for same day outpatient procedures. These advancements, along with continued public exposure, and patient education, point towards the widespread adoption of UFE as a first line therapy for symptomatic uterine fibroids.

1. Dariushnia SR, Nikolic B, Stokes LS, Spies JB; Society of Interventional Radiology Standards of Practice Committee. Quality improvement guidelines for uterine artery embolization for symptomatic leiomyomata. 

. 2014;25(11):1737-1747. doi:10.1016/j.jvir.2014.08.029

2. Poll H. 2017. The fibroid fix: What women need to know. Society of Interventional Radiology; Executive Report.

3. Gupta JK, Sinha A, Lumsden MA, Hickey M. Uterine artery embolization for symptomatic uterine fibroids. 

. 2012;(5):CD005073. Published 2012 May 16. doi:10.1002/14651858.CD005073.pub3

4. Makris GC, Butt S, Sabharwal T. Unnecessary hysterectomies and our role as interventional radiology community. 

. 2020;3(1):46. Published 2020 Jul 14. doi:10.1186/s42155-020-00138-x

5. Manyonda I, Belli AM, Lumsden MA, et al. Uterine-Artery Embolization or Myomectomy for Uterine Fibroids. 

. 2020;383(5):440-451. doi:10.1056/NEJMoa1914735

6. Caridi TM, Spies JB, Kohi MP. Myomectomy versus Uterine Artery Embolization: More Alike than Different. 

. 2020;31(11):1838-1839. doi:10.1016/j.jvir.2020.08.023

7. Yoon J, Valenti D, Muchantef K, et al. Superior Hypogastric Nerve Block as Post-Uterine Artery Embolization Analgesia: A Randomized and Double-Blind Clinical Trial. 

. 2019;292(1):269. doi:10.1148/radiol.2019194008

8. Pereira K, Morel-Ovalle LM, Taghipour M, et al. Superior hypogastric nerve block (SHNB) for pain control after uterine fibroid embolization (UFE): technique and troubleshooting. 

. 2020;3(1):50. Published 2020 Sep 27. doi:10.1186/s42155-020-00141-2

The need for a growing role for interventional radiologists.

Uterine Fibroid Embolization: Past, Present, and Future

A podcaster I like was mini-ranting about some sociopolitical issue of the day. As such talking heads often do, he was speaking as if to someone of the opposite opinion—politician, bigmouth celebrity, whatever. In the process, he shared an interesting quip that, months later, inspired today’s column: “Don’t make me get interested.”

He went on to explain that it’s an implied threat and invited his audience to use it when appropriate. He’s a capable, smart guy, with more than a few resources, including money and connections. If something gets his attention enough to make him want to take action, he has more than a decent chance of making things happen.

He’s got plenty already going on in his life, so not just anything is going to rise to that level. Most stuff is going to fly below his radar…or briefly provoke thought or emotional reaction. But, if something interests/angers/enthuses him sufficiently to take it on as a personal project, he’ll be committed enough to make an impact. Hence, the threatening aspect: Don’t make him get interested in something you’d rather he not pursue, such as the opposite side of a political campaign…because he might just make your guy lose.

There are any number of ways you might thusly provoke someone. Jeopardizing their interests with your actions or statements…or inactions, pertinent things left unsaid…is a common one. Most rads (and, indeed, other professionals across the board) have experienced this in their career paths — if one gets the idea that their future in a current job is not as promising as it might be elsewhere, sooner or later they’ll get interested in the possibility of other jobs.

That can happen if they’re paid less than they think they’re worth, given undesirable hours, or have no hope of upward mobility within their group’s ranks. Also if they come to feel that promises aren’t being kept or things are otherwise not moving in hoped-for directions. A sense of being insufficiently appreciated and/or respected as a professional. None of these dissatisfactions have to be reasonable or realistic to anybody other than the individual, mind — all that matters is that they’re sufficiently motivating to the individual.

This sort of thing isn’t just about changing jobs, of course. Rads can get interested in all sorts of things that others in the group might wish they wouldn’t. Perceived abuse of protocoling by referrers, for instance, or excessive use of imaging at all. Next thing you know, a rad who got interested in there being too many triphasic abdominal scans or middle-of-the-night MR is waging a personal war against the clinicians and making everybody unhappy in the process.

So what do you do to avoid getting someone interested in (and then pursuing) things you’d rather they not? Well, the utopian answer would be to have everyone freely discuss everything of interest to them—encouraging shy folks to come out of their shells and participate so they don’t suffer in silence—and then come up with universally-satisfying, diplomatic solutions. Or at least make sure people understand the whys and wherefores of not getting everything they want.

Obviously, not going to happen 100 percent…but it’s still probably a good idea to make efforts in that direction whenever possible. If people feel like they’re being heard, even solicited for their thoughts, it goes a long way. The opposite—a culture where opinions are not sought out at all, or even discouraged/disparaged when shared—adds insult to injury. Someone who might not have gotten interested in going their own way could be motivated to do so by a sense of “They brushed me aside? I’ll show 

Another approach comes to mind in light of a pup who joined my home two weeks back. Quite a different personality than our previous doggo (who, unfortunately, had developmental renal dysplasia and barely made it to age 4). The current pooch is far more energetic and brighter…thus, capable of getting into far more trouble.

 of things get her interested: Places she shouldn’t go, things she shouldn’t get/destroy, and our poor cats who were accustomed to being mostly left alone by the previous dog. Just as with humans, we’re obviously not going to let the dog do whatever she pleases…but, we can’t exactly reason with her about why she can’t have this, do that, etc.

A big part of the solution, just as with humans, is redirection: Provide something harmless (or, ideally, productive) to get and maintain their interest. One might call it 

direction, since it’s far easier to control the situation if you can prevent the unwanted interest from occurring in the first place.

For a puppy, that means things like playing with them, getting them toys that will hold their interest…even “slow feeders” that make meals take more time and effort so they can’t inhale their kibble (and, then, possibly vomit it back up a few minutes later on the living room rug). An engaged mind is less likely to find mischief than an idle one.

For a human—let’s say a radiologist—it might mean proactively giving them things to do beyond simply plugging away at a never-ending worklist. Committee seats, additional levels of responsibility, such as subspecialty chief or Quality Assurance poobah. Heck, make up a role if you can identify something that suits and would appeal to them.

Bonus: Most rads, approached with an open query about some special role they’d like to take on, will probably feel complimented. Instant morale boost and sense of engagement by the group. They might even respond by telling you what potentially-negative interest they were liable to develop, giving you a chance to put them on a related task in a far more constructive way.

Providing something productive to focus on can help side-step unwanted attention.

Don’t Make Me Get Interested

Large filling defect with irregular margins seen on this upper GI series on fluoroscopy.

On upper GI series, any filling defects projecting from the wall of the stomach may be seen. Ulceration and cavitation may also be present. Margins of the filling defect may be smooth or irregular.

Patient presents for upper GI series with this finding... What is your diagnosis?


Image IQ Quiz: What's On This Upper GI Series?

In recent years, the personalization of medicine has accelerated greatly as a result of tremendous advances in genomics. Mapping of the human genome, including understanding the effect even minor abnormalities in the genetic structure can have, has enabled providers to select one treatment option over another even when two patients share similar factors, such as age, gender, weight, ethnicity, lifestyle, family history, etc.

Yet, despite the availability of all that data, understanding not only how effective a treatment plan might be, as well as how well it’s currently going, has remained a challenge. Because even when there are genetic similarities between patients and their tumors, the reality is they may still react differently to the same therapy. Often, providers won’t know for certain until the course of treatment is well underway – or completed.

At least, that has been the case until now. Fortunately, new advances in radiomics are helping providers obtain those answers earlier in the process. When combined with artificial intelligence (AI), the phenotype data being generated by radiomics is even beginning to yield biomarkers that can predict how a specific tumor or lesion will react to various treatment options, helping to guide the decision process.

The key is radiomics’ ability to see below the surface of a lesion. A typical radiology image is limited to two dimensions: length and width, basically the long and short. Radiologists can see whether a lesion is growing, shrinking, or remaining the same size in reaction to a course of treatment, but that is all they can measure.

Radiomics, however, uses AI-driven analytics to extract meaningful data from traditional imaging modalities, such as CT, MRI, or PET scans. It, then, curates, annotates, and analyzes that quantitative data to deliver a wealth of information that cannot be observed optically in an image.

With radiomics, providers gain accurate readings of a lesion’s depth, volume, density, doubling size, textures, and other features, all the while comparing those measurements to an AI-enabled model based on thousands of healthy organ images. The result is that providers not only gain a view of lesion progression that formerly could only be achieved through invasive surgeries, such as biopsies; they also receive information about lesion changes that have 

 been available by any means. And, with built-in deep learning (DL) capabilities, the radiomics system continues to improve its accuracy and effectiveness as more images are ingested and analyzed, and more outcomes are confirmed, regardless of the type of lesion.

Here’s how that information can be applied to personalize the treatment plan: When a lesion is discovered, providers can use the patient’s genetic information along with phenotype data from radiomics to match the patients to others with similar profiles. They can, then, review the outcomes those previous patients achieved with various treatment plans to determine which yielded the best results on a consistent basis. This, then, becomes the starting point for treatment.

The added advantage of radiomics, however, is that providers can continue to monitor the effect the treatment is having across thousands of data points. If it is not achieving the intended results at certain checkpoints, providers can change the treatment plan immediately rather than waiting until the treatment is completed. They can, then, continue this same process of monitoring and adjusting to ensure they deliver the best possible results for each patient based on their individual response to the treatment plan.

There is also a practical, business-oriented side to using radiomics to further personalize a treatment plan. When health plans are reviewing treatment options, their tendency is to prefer the one (or two) with the lowest up-front cost. Yet, that might not always be the best course of action for the patient.

With the deeper data and history generated through radiomics, providers can demonstrate to health plan authorization reviewers that the preferred treatment, while potentially more expensive on the surface, is not only the one most likely to generate the best health outcomes but will also cost the health plan the least in the long term.

The more healthcare providers can base treatment plans on individual needs and characteristics, the more value they can deliver to their patients. Radiomics offers the next breakthrough in bringing a more personalized approach to cancer patients – before, during, and after treatment.

Rose Higgins serves as Chief Executive Officer of 

, a pioneer in applied radiomics, which is the cutting-edge field of extracting novel data and biomarkers from medical images.

Mimi Huizinga, MD, MPH, FACP, serves on the HealthMyne board of directors. Dr. Huizinga is the Vice President and Head of U.S. Oncology Medical, Novartis.Huizinga is a physician executive with deep experience achieving clinical excellence across the care continuum driven by evidence-based care, data, and analytics.

A look at the insights and financial benefit provided by radiomics.

Radiomics Offers the Next Breakthrough in Treatment Planning for Personalized Medicine

Making advanced PET part of the treatment process for men who have recurrent prostate cancer can greatly improve their outcomes over solely using traditional imaging, according to new research.

Conventional imaging, such as CT or MRI, frequently does not offer enough detail for providers to have complete confidence in suggesting post-prostatectomy radiation, said a team of investigators from the Winship Cancer Institute at Emory University. But, using 

F-fluciclovine-PET/CT makes it easier for providers to see the uptick in amino acid movement that often factors into their decision.

The team published their findings this month in 

“The decision to offer post-prostatectomy radiation is complex, because conventional imaging can leave unanswered questions on the best approach for treatment planning,” said co-principal investigator Ashesh B. Jani, M.D., professor of radiation oncology at Winship. “This research has found that integrating advanced PET imaging using 

F-fluciclovine into the treatment planning process allows us to do a better job of selecting patients for radiation therapy, guiding radiation decision and planning, and ultimately, keeping our patients’ cancer under control.”

The radiopharmaceutical, developed by Blue Earth, a subsidiary of Bracco, has been approved by the U.S. Food & Drug Administration for use in men who are suspected of prostate cancer recurrence based on their prostate-specific antigen (PSA) levels.

Based on their research, Jani’s team determined the men who received 

F-fluciclovine experienced a superior cancer control rate of 75.5 percent after three or four years. Their result outpaced what men who underwent conventional imaging experience – a 63-percent and 51.2-percent control rate at three and four years, respectively.

For this EMPIRE-1 study, from September 2012 to March 2019, the team enrolled 165 patients with prostate cancer who had PSA levels that were detectable post-prostatectomy, but who had no findings on conventional imaging. These men were, then, randomly assigned to one of two study arms – radiotherapy guided solely by conventional imaging or radiotherapy augmented by 

Both groups saw similar toxicity, and the most common adverse events were late urinary frequency or urgency (present in 37 out of 81 patients in the conventional imaging group and 31 out of 76 individuals in the PET group) and acute diarrhea (present in 11 conventional imaging patients and 17 PET patients).

These results validate what the researchers had believed about the effect of using this radiopharmaceutical.

“The question that we wanted to answer in this study was whether the treatment plan effect informed by 

F-fluciclovine-PET imaging had a positive effect in the lives of patients,” said David M. Schuster, M.D., FACR, professor of radiology and imaging sciences and director of the nuclear medicine and molecular imaging division at Emory. “The EMPIRE-1 trial allowed us to determine whether using 

F-fluciclovine-PET imaging influences patient outcomes for the better, and the significant results confirm that it does.”

Using advanced PET imaging instead of just traditional imaging alone can improve outcomes for patients with prostate cancer. 



Prostate Cancer Control Improved by Adding PET Imaging to Treatment Planning

Diagnostic Imaging's Weekly Scan: The Top 5

Slideshows

Take a quick look at Diagnostic Imaging's Top 5 stories of the week.

Diagnostic Imaging's Weekly Scan: May 31, 2021, to June 4, 2021

If a picture is worth a thousand words, at least some of those are good for radiology reports. At least that’s the approach being used to train residents at the University of Massachusetts Medical School.

Roughly 70 percent of interpretation errors with diagnostic imaging studies are linked to visual perception mistakes. These problems can happen a lot more often when providers focus more on the “answer” lurking in the image rather than having a more in-depth discussion of what the scan might reveal. This approach is a habit among millennials, said the team led by Andrew M. Singer, M.D., a radiologist and director of ultrasound at UMMS.

“We have observed that residents often do not discuss findings or form an impression in a methodical way, but rather jump to one answer, as if interpreting the image as a multiple-choice test question,” the team said in their study published in the May 31 

. “This change in approach has likely been compounded by the introduction of PACS, remote reading, lack of clinical rounds in the reading room, and the introduction of [artificial intelligence], all of which has created distance between radiologists and clinicians and has decreased one-to-one interactions.”

To combat this trend, and because millennials respond well to immersive, interactive learning environments, the team landed on an unusual strategy – they turned the art museum into a visual learning lab. A reading room filled with Baroque, Classical, and modern images that helped residents fine-tune the nuanced skills needed for more accurate, deft diagnoses.

Beginning in 2016, UMMS implemented this art museum teaching strategy with radiology residents who had completed their first three months of training, including emergency radiology, chest, abdomen, pelvis, and neuroradiology. The workshop consisted of a one-hour lecture followed by a two-hour museum visit.

“We chose the timing of our workshop to synchronize with the stage of their training where they would have gained a basic understanding of the process of interpreting radiologic images which would inform their understanding of the similarities to art observation techniques,” the team explained.

During the lecture, residents were exposed to the main tenets of Visual Thinking Strategies (VTS) – detail, storytelling, ambiguity, and empathy – as the same elements are fundamental to meaningful radiologic interpretation, they said. The students observed four differently styled paintings, and researchers posed and expounded on three VTS questions so that they correlated well to making radiologic findings, communicating the impression, and ACR Imaging 3.0, as well.

What is the story that the painter is telling?

What is the mood expressed by the artist?

Are your observations biased by your beliefs?

“The purpose was to identify and describe findings without interpretation, and then to give the findings context, in order to clearly analyze and present the ‘story’ that the artist is conveying in the artwork,” the team said. “This parallels the radiologist’s presentation of the findings and impression for an imaging study.”

Based on the feedback from 35 residents who have participated in the workshop, the researchers determined that the art museum is an effective environment for residents to explore their observational, communication, and analytical skills. Residents reported feeling more comfortable with closely analyzing studies and that they had developed a deeper understanding of weaving findings together cohesively in order to see the overall patient picture more clearly.

For the future, the team said they intend to expand their outside-the-box training efforts to include digital simulations, 3D and virtual anatomy interactive touchscreens, escape room competitions, and “Case of the Week” social media posts.

“Moving forward, we suggest that the radiology reading room not be thought of as an impersonal utilitarian ‘Work Station,’ but rather as a ‘collaborative digital imaging studio’ for reviewing images, static and video, with the added capabilities of digital data manipulation, image reconstruction technology and the capabilities to interface remotely with other radiologists and the clinical providers,” they said. “By transferring skills acquired at the museum to the ‘Radiology Digital Imaging Studio,’ we hope that our program helps to accelerate the radiology residents’ learning process and adds to the enjoyment of radiology.”

Teaching radiology residents the skills used to observe and analyze works of art can improve their interpretation and reporting skills.



Want to Teach a Millennial Rad? Try a Night at the Museum

Due, in large part, to a lack of access to chronic disease management services, homeless patients are more likely to undergo a CT scan for a suspected or confirmed seizure in the emergency department.

Across western urban settings, homeless individuals have an eight-fold increased rate of epilepsy and seizures, and they frequently seek care in acute settings where there is little-to-no longitudinal knowledge about their condition. This leads to a much higher incidence of CT scans that can elevate their cumulative radiation dose.

“We found that homeless patients had a significantly increased number of presentations to the emergency department and a significantly higher number of head computed tomography, with a resulting increase in radiation exposure,” said a team from St. James Hospital in Dublin, Ireland, led by neurologist Elisabeth M. Doran. “Only receiving acute care for seizures leads to sub-optimal seizure control, which then leads to higher frequency of acute presentations.”

Doran’s team published their findings in the May 31 issue of 

Neuroimaging in the ED for Seizures Not Required as Often

As part of an overall effort to improve healthcare delivery to homeless patients, the team detected this higher CT utilization among this group. To determine the severity of the problem, they examined anonymized data from 88 homeless patients who experienced seizures between 2015 and 2020 and compared it with data from a control group of 88 patients who were not homeless.

Based on their assessment, they found that homeless patients – who had a higher number of seizures related to substance abuse – averaged six visits to the emergency department for seizures while the control group – which was more likely to have an established epilepsy diagnosis – averaged one visit to the emergency department.

Overall, the cohort of homeless patients underwent 250 brain CTs over five years with a median radiation exposure of 4.3 mSv. Nine patients from the group had 10 or more CT brain scans within the five-year study period. That led to an exposure of at least 14 mSv for those patients, the team said – an equivalent to 700 plain X-rays. In contrast, no patients from the control group had 10 or more CT scans during the study period.

“We found that compared to housed patients, homeless patients had a considerably higher number of presentations to emergency departments and more frequent CT brain investigations, which led to a significantly increased radiation exposure,” the team explained. “As such, we estimate that some of the patients reach cumulative radiation exposure doses which put them at risk of radiation-induced cancer.”

Taking steps to improve access to epilepsy care – particularly specialist services – is one avenue to potentially lowering the number of times these homeless patients visit the emergency department, the team said. Working to integrate specialist services, such as remote consultation, into both the emergency and primary care settings could help augment care.

“If specialist services find ways to tailor their care more to the needs of this vulnerable population, we may be able to improve their seizure control and reduce the frequent [emergency department] admissions for seizures through that,” they said.

Homeless patients suffering from seizures undergo more CT scans, largely due to lack of access to chronic care, leading to a higher radiation dose and cancer risk.

