Greg Freiherr, Industry Consultant
Greg Freiherr, Industry Consultant

Greg Freiherr has reported on developments in radiology since 1983. He runs the consulting service, The Freiherr Group.

Blog | Greg Freiherr, Industry Consultant | Radiology Imaging| July 20, 2016

The Case Against Quantitation

imaging

Image courtesy of Pixabay.

Making healthcare more objective and precise promises to increase efficiency and reduce costs — the key ingredients of value medicine. At the University of Pennsylvania, the Penn Center for Innovation has developed an image quantification tool that might help achieve both goals.

Called the Automated Anatomy Recognition (AAR) system, it localizes and delineates cancer in multiple body regions using magnetic resonance imaging (MRI), computed tomography (CT) and positron emission tomography/computed tomography (PET/CT) images. When applied to radiation therapy planning, AAR can reduce the time needed to delineate organs from several hours to less than 5 minutes. Successes like this fuel speculation that quantitation may be broadly applied to make medicine more efficient and effective. But quantitation is a means to an end, typically a highly focused end.

Essentially quantitation in medical imaging defines a range of numerical values associated with a pathology. At best, it adds information that may be used to streamline analyses or increase the certainty of conclusions. At worst, it creates the illusion that diagnosis can be reduced to numbers. And therein lies the potential problem. 

It is all too easy to assign to quantitation greater value than it deserves. It cannot simply be extended to all aspects of medical interpretation. No one has suggested doing so closely. And for good reason.

Human Experience

When interpreting exams, radiologists draw from their experience. They may consider quantitative measures, interpreting them in a broader context.

The very basis of enterprise imaging is to provide this context to referring physicians who can see radiological images and to radiologists who can call up lab results and path reports. If quantitative imaging adds more information, great. But it is human nature to follow the shortest path between two points; to take the path of least resistance; to delegate to machines what might better be accomplished under our own power.

A few years ago I saw my neighbor across the street riding his lawn tractor down the driveway to get his mail. He laughed when he saw me watching, answering, “Never do what a machine can do.”

It turned out not to be causative that he had just finished cutting the backyard and was on lawn tractor; or that the ride to the mailbox might have been done on a whim. A few days later I soon saw him driving his lawn tractor over to his next-door neighbor. Never mind that he could walk faster than his tractor could drive. Never mind that the walk would have done him good. These thoughts were non-starters for my neighbor.

Ride To Nowhere

And so it goes with machines. If they reduce labor or simplify complex tasks, they are sure to be a hit. Unfortunately what might be left out of the equation is whether the end result is superior to what would have been achieved more conventionally.

We are at a time when radiologists are being pushed to take on a bigger role in healthcare. In a word, they are vying for “more.” More opportunity to make more of an impact by becoming more involved in patient care — not just in diagnostic work-ups but therapy planning and treatment assessment. And it is happening.

At the NYU Langone Medical Center, radiologists are making morning rounds — virtual ones, to be precise — communicating with medical staff over the center’s wireless network, sharing images and insights, and improving patient care. As told in a Wall Street Journal article, one such interaction led to the discovery of a defect in the hip of a six-month-old patient operated on for a congenital heart defect.

Making this possible was the human ability to see the whole patient, to see what otherwise might not have been seen, if the focus been on just one aspect of the patient’s care.

Radiologists are more familiar than any other medical specialist with the imaging options available for patient assessment and how those options might be utilized. An MR image might tell enough of the story to draw a conclusion. Or MR results might be fused with those coming from a PET scan to see the functional and anatomical together. Or MR might be used in real-time to guide radiation therapy … or to guide a neuro intervention. 

Once the lead leveraging their skills, radiologists go well beyond simply interpreting images, exemplifying the limits of quantitation.

The phrase “to err is human” has led to the pursuit of increasingly precise and objective means of analysis. And that’s good. Quantitation is another arrow in the quiver of precision medicine. But radiologists must be called upon to consider which arrow to pull from that quiver. Knowing that — and when to fire those selective arrows — has never been more important.

Hanging in the balance are patient welfare and lower healthcare costs. This is happening at a time when American medical institutions are beginning to creak under the healthcare needs of an aging population.

How well these institutions hold up may depend on how well healthcare providers choose from the sources of information available to them — and how well the resulting information is managed. The answers will not to be found worshiping at the digital altar.

In our zeal to be more precise, we must not put at risk the unique human ability to perceive what cannot be exactly defined — or quantified.

Editor's note: This is the third blog in a four-part series on patient centricity. The first blog, “Value Medicine: Radiology’s Big Chance,” can be found here. “What To Do About The Draper Effect” can be found here.

Related Content

Video Plus Brochure Helps Patients Make Lung Cancer Scan Decision

Image courtesy of the American Thoracic Society

News | Lung Cancer | April 19, 2019
A short video describing the potential benefits and risks of low-dose computed tomography (CT) screening for lung...
FDA Clears GE's Deep Learning Image Reconstruction Engine
Technology | Computed Tomography (CT) | April 19, 2019
GE Healthcare has received 510(k) clearance from the U.S. Food and Drug Administration (FDA) of its Deep Learning Image...
Artificial Intelligence Performs As Well As Experienced Radiologists in Detecting Prostate Cancer
News | Artificial Intelligence | April 18, 2019
University of California Los Angeles (UCLA) researchers have developed a new artificial intelligence (AI) system to...
A smart algorithm has been trained on a neural network to recognize the appearance of breast cancer in MR images

A smart algorithm has been trained on a neural network to recognize the appearance of breast cancer in MR images. The algorithm, described at the SBI/ACR Breast Imaging Symposium, used “Deep Learning,“ a form of machine learning, which is a type of artificial intelligence. Graphic courtesy of Sarah Eskreis-Winkler, M.D.

Feature | Artificial Intelligence | April 12, 2019 | By Greg Freiherr
The use of smart algorithms has the potential to make healthcare more efficient.
Videos | RSNA | April 03, 2019
ITN Editor Dave Fornell takes a tour of some of the most interesting new medical imaging technologies displa
Johns Hopkins Medicine First in U.S. to Install Canon Medical's Aquilion Precision
News | Computed Tomography (CT) | March 26, 2019
March 26, 2019 — Johns Hopkins Medicine now has access to the first...
NIH Study of Brain Energy Patterns Provides New Insights into Alcohol Effects

NIH scientists present a new method for combining measures of brain activity (left) and glucose consumption (right) to study regional specialization and to better understand the effects of alcohol on the human brain. Image courtesy of Ehsan Shokri-Kojori, Ph.D., of NIAAA.

News | Neuro Imaging | March 22, 2019
March 22, 2019 — Assessing the patterns of energy use and neuronal activity simultaneously in the human brain improve