Leaders in Breast Health: Louise Miller, Mammography EducatorsJul 8, 2021
Louise Miller is the co-founder of Mammography Educators and former Director of the Mammography Practicum at the University of California School of Medicine, where she taught for 25 years and is a Society of Breast Imaging Honorary Fellow. She has taught or trained over 50,000 technologists globally and helped develop, coordinate, and implement mammography screening and breast health education programs internationally.
Dr. Eliot Siegel, our Clinical Advisory Board Member, speaks with Louise on the impact of COVID-19 on mammography positioning and her perspective on the value of A.I. in breast screening.
Positioning techniques are known to vary considerably among radiologic technologists. There are concerns that with COVID-19 restrictions there may be even more of a tendency for positioning errors to occur due to anxiety in both patients and technologists, social distances, and PPE requirements.
Training technologists has always been important in ensuring quality of care in mammography and continuing to do so during COVID-19 is going to require careful oversight of performance to ensure continued quality care delivery. How do you see A.I. enabling the development of evidence-based training protocols and what impact do you think A.I. may have in the next few months and years as a mechanism to provide feedback to technologists?
COVID-19 definitely has added to the challenges that existed before social distancing and PPE requirements were in place. Patients are more apprehensive, and this makes positioning even more difficult at times.
In your own experience, do you believe COVID has had a deleterious impact, or do you think most technologists are getting around that?
I think they’re learning to work around it. We have technologists who are going to go the extra mile and those who are going to use that as an excuse. We also have patients complaining, which is a very natural process during this period. Everybody has anxiety about what’s going on, especially the patients. When they come for a regular mammogram, they have a high level of anxiety anyway, this just heightens it.
I think that when we look at it historically, positioning has long been a problem, certainly during my career of over 35 years. The last ACR manual was updated on the positioning component in 1999, so there have been no updates for standardized positioning techniques for use with full field digital and DBT.
To add to the confusion, feedback regarding image improvement is variable and subjective. So, it’s hard to believe that no standard positioning exists for mammography. In general radiography, every technologist in the world positions a foot, a hand, an ankle, a chest, or a pelvis using the same technique. But in mammography, we’re all over the place.
We did a survey about a year ago and asked 100 mammography technologists, “do you think that the techs in your department position the same way?” 81% said no. No wonder we’re having a problem.
Unfortunately, there has been very little information or training on standardized techniques available for technologists during the past 20 years. Consequently, technologists have adapted by using a huge variation in positioning methods with little or no idea how to improve them. Unfortunately, most of these techniques are not effective and counterproductive to improving and optimizing image quality.
Another factor is that the technologists’ licensure process and MQSA itself do not require updated, hands-on mammography positioning training for technologists. Mammography technologists need an efficient, proficient, consistent, reproducible, and ergonomically sound method of positioning that unfortunately has not been universally taught. Perhaps most important is that the technologists need to be able to clearly identify the deficiencies in their images, hopefully with the use of A.I., and then be able to problem solve using these proven methods and techniques.
Most technologists can tell if they don’t have the IMF or they don’t have the pec muscle down to the PNL, but they don’t know what to do about it. They don’t specifically know how to correct the problem.
Let’s say that I have software that uses A.I. for mammography quality control. How can I design a program around an A.I. system that would provide that feedback to the technologist right away and then in the aggregate on a monthly basis, for example? How could I, essentially, then successfully build a training program that’s evidence-based around the A.I. software?
This is what’s really exciting for me, Dr. Siegel, is being involved with the development of these programs. Fortunately, we know exactly how to get the technologist to this point. Using A.I., we want to be able to get them excited about using this incredible tool to learn to help themselves and their patients. I believe that technologists and radiologists will begin to view A.I. as a way to increase image quality.
We need to be able to objectively and instantly identify areas of deficiencies for the technologists and then provide the technologist with immediate feedback. And this is where it really gets interesting. We want to be able to tell them exactly what to do to correct the deficiency with an objective and precise analysis using A.I.
We’ve developed quick, on the spot, live correctional-action steps that are essential for the technologists. They must also have access to comprehensive learning resources that will assist in the learning process beyond that immediate feedback. Once we’re able to objectively measure image quality using A.I., we’ll be able to clearly see the efficacy of the use of standardized positioning techniques and the appropriate and objective feedback for improvement.
This will undoubtedly result in the improvement of image quality which in turn means the detection of more treatable breast cancers.
How close are we to that point now? I’ve seen some fantastic software with regard to that quality feedback. Are you aware of anything that would be the wizard that would coach somebody through and then potentially track trends as time goes on?
I think we’re very close to doing that. I think the preliminary studies are showing that this is absolutely a possible reality. The most important thing that I think we need to understand is that it’s not just about the analysis of the image. We do all of these MQSA audits and the technologist is told they don’t get IMF enough times or they only get it 10% of the time. We know when we’re not good at something, but the problem is that there’s nobody there to support them and say what they need to do to fix it.
We know that standardized positioning works. Bassett published a study in 1993 saying that after ACR standardized positioning training, image quality improved by 76%. That’s huge.
So, what we can do is work together to make available to the technologists not just the techniques that they should learn initially, but how they can sustain the use of those techniques that are consistent, reproducible, efficient, proficient, and ergonomically sound. A lot of technologists have workplace related injuries and that money is better spent on programs like A.I.
I think giving them solutions to the problem is the key. Not simply identifying the problem, but the objective identification and then an objective recommendation for improvement is absolutely essential.
We’re starting to work on it already and we have some of these quick steps and detailed instructions for the technologists, so we’re very excited.
Discover Technologist Training with Densitas and Mammography Educators
Densitas and Mammography Educators have partnered to tackle technologist training for mammography positioning challenges with embedded Mammography Educators content. Efficiently target the mammography positioning challenges that your technologists face with a unique integration that feeds a cycle of continuous improvement
Let’s Stay Connected, Subscribe for Updates
Join our email list to stay up to date on the latest advancements in breast health technology.