Dr. Thomas Ibounig on Incidental Rotator Cuff Abnormalities and Shoulder MRI

Great.

And so putting all this data together, um, what are your takeaways from the study? How did you...

Can you describe a little bit about how you, um, analyzed the data? And there was some discussion of different, um, regression models within the paper, but I'm curious your insight into the choices for the three different regression models that were described and how you tried to make sense of the potential relationships between, as you described, not only the imaging findings, but qui- quite a comprehensive clinical battery of tests, which to your credit, we think about a lot as orthopedic manual physical therapists of utilizing imaging as not a standalone finding, but as part of our information that we analyze in conjunction with quite a thorough clinical exam.

We kind of pride ourselves on that.

So how do you do that from a research standpoint, factoring in all of those different variables to try to make sense of it?

Yeah, that's, that's a good question.

Of course, [sighs] not, not an easy, not an easy question to, to tackle because there are many different ways, and probably no- none of those ways are completely perfect, but...

So I think that what our, what we achieved with our sample was that we, we, we got a very, very good, um, certainty o- on the prevalence of those imaging findings in asymptomatic and symptomatic shoulders.

Uh, but then, of course, those imaging findings and symptoms, um- They could be caused, for example, looking at the shoulder also by other structures, not only the rotator cuff.

So how do you want to make sure that not, uh, osteoarthritis of the shoulder or AC joint problem, uh, biceps problems are the cause of symptoms? So, so the first adjustment that we, that we performed in addition to a, to the adjustment, uh, according to age, sex, uh, region, and education, which would be, um, general risk factors or, or, or factors affecting, uh, prevalence.

So, so we, we looked at other imaging findings and adjusted for those, and we saw that the difference in prevalence between symptomatic and asymptomatic shoulders was reduced a little bit, uh, regarding full thickness rotator cuff tears.

So when the, when the, um, uh, difference in prevalence in our base sample was 8.1%, 6.5% in asymptomatic shoulders and 14.6% in, in symptomatic shoulders, um, then after adjusting for these other, uh, imaging abnormalities, the, the prevalence difference was reduced to 6.7%, so a little bit reduced.

And then we thought, okay, when no one is treating, uh, or at least no, no surgeon admits that he's, he or she is treating only, uh, based on imaging findings alone, so of course we have the clinic, clinical examination playing, playing a very big role into that.

So we also wanted to look, well, how, how does clinical examination, uh, findings, if we use those, we should be able to better identify those with symptoms and those with, with no symptoms, uh, symptomatic rotator cuff tears.

Uh, and we were really surprised, or we were on one hand surprised and then on the other hand not surprised that actually the difference diminished completely, and we didn't get any, uh, improvement in, in our diagnostics using these clinical tests.

Um, and I think that's also-- I think that orthopedic surgeons are not good at, at, uh, or, or very structural orientated, but from what I, from, from my experience, although physiotherapists are much better in, in accepting the fact that, uh, pain or symptoms are not only based on structural problems, I think where, where physical therapists also are a little bit biased probably is, uh, those, uh, clinical maneuvers and, and clinical examination.

We're very fond and proud of all our, our maneuvers that we have, and, and we are really, especially when we are really experienced, very much able to identify with certain maneuvers, um, uh, the cause of pain and symptoms.

But then when we look at the literature, uh, actually, no, first of all, I think that also regarding clinical examination, we have some, uh, lack of good studies, but, but those studies that we have actually show that our, uh, isolated clinical examination maneuvers are, are like the flip of a coin mostly.

So they also don't seem to help us, uh, on their own as much as, as we would think.

And then I think it's putting all those things together, and that's really difficult to, to, uh, at least at the moment, I think that AI might help us in this, but, but putting all those small pieces of the puzzle together, uh, and, and doing research on that is really difficult.

So although single maneuvers or a single image might not be able to, to show us the cause of pain, um, at least with a very high certainty, putting all those stuff together might help us a little bit more.

In our model, it did not work out, and I, I don't think that, that single maneuvers are very good or, or single imaging.

But then there might be some, some things, and maybe it's even more patient, uh, history or patient baseline characteristics that, that play a bigger role, uh, in, in, in showing us what, what's the cause of symptoms in each individual patients, because I think that it probably depends a lot on the, on the person where the pain comes from.