Interstitial Lung Abnormalities (ILAs)

Podcast Transcript

Raed Dweik, MD:

Hello and welcome to the Respiratory Exchange Podcast. I'm Raed Dweik, Chief of the Integrated Hospital Care Institute at Cleveland Clinic. This podcast series of short, digestible episodes is intended for healthcare providers and covers topics related to respiratory health and disease in the areas of lung, critical illness, sleep, infectious disease and related disciplines. We will share with you information that will help you take better care of your patients. I hope you enjoy today's episode.

Dan Culver, DO:

Hello. Thank you for joining this episode of Respiratory Exchange. I'm Dr. Dan Culver. I'm the Chair of Pulmonary Medicine at Cleveland Clinic. Today I have with me two of my colleagues who are experts in interstitial lung disease. I’m joined today by Ruchi Yadav, MD from Thoracic Radiology,

Ruchi Yadav, MD:

Hi everyone.

Dan Culver, DO:

And by Aman Pande, MD from the Division of Pulmonary Medicine,

Aman Pande, MD:

Hello.

Dan Culver, DO:

Today we're going to talk about interstitial lung abnormalities or ILAs. When I first heard about ILAs, I thought this is another trick that the ILD folks are foisting on us with a new acronym, and these have become quite popular. It's hard to pick up a journal nowadays without seeing some kind of paper about ILAs. And so just a level set for everybody. Maybe we can just start off with what are ILAs?

Ruchi Yadav, MD:

You know, before the Fleischner Society position paper in 2020, ILAs, as far as at least the radiologists were concerned, there was no standard definition for it. So, we would call them non-specific markings in the lungs on a CAT scan; we would call them changes which may suggest aging. And, since there was no standard definition, there were some radiologists who wouldn't even put them in the impression.

And so I think the first time was when the position paper came out in 2020, they came up with a standard definition for ILAs and the Fleischner Society called them findings that are detected incidentally in patients who do not have any known ILD or do not have suspected ILDs and these findings are bilateral, independent, and they said that they should involve at least 5% of the lung zones. So that was the first time there was an address made to these very subtle findings on a CAT scan, which are different than ILD, that as a radiologist we started paying more attention to.

Dan Culver, DO:

So, this is like learning a new word. As soon as you put a name on it, suddenly you start seeing it everywhere. And I noticed in radiology reports, suddenly we were seeing a lot more of these than we'd ever noticed.

[00:03:04]

Ruchi Yadav, MD:

That's true. I mean, and as you rightfully said, Dan, there's been a lot of traction in the last, I would say, five years.

 Dan Culver, DO:

No pun intended, with “traction.” [laughs]

Ruchi Yadav, MD:

[Laughs] and I think this is because we now have drugs, which can have an impact. And that's the reason there's a lot more interest in identifying these because they can impact clinically.

 Dan Culver, DO:

So we now are starting to pay attention to this concept of incidentally discovered findings on CT scan.

 And quite often those CT scans are done for other reasons. Somebody comes to the ED for a variety of reasons, or they're having a lung cancer screening test or a coronary calcium scoring and yet you find these things on the imaging, which aren't quite often as remarkable as what you might see with ILD. But they're there and now we're calling them out.

So that brings up the issue since these are so prevalent compared to ILD, why is this a big deal? Why are people talking about this so much? Why did this gain traction, so to speak, Aman?

Aman Pande, MD:

Well, I think that it's since the publication of a few papers, many, around 2017. In JAMA there was a big publication that described four large, well-characterized cohorts of patients.

This included patients from the Framingham Heart study and patients from the COPD gene cohort. These were all folks who had had CAT scans of the chest and who were followed and monitored over a period of time, five to 10 years.

And what the researchers found was quite striking: that when they divided patients into two groups, those who had ILAs and those who did not, the rates of mortality were different and uniformly in all these studies, they found higher mortality among patients who had ILAs. And that, I think, generated a lot of interest.

And around that time, we'd also come up with a couple of new medications to treat these fibrotic lung diseases. And the combination of those, I think, has spurred a lot of interest in the field.

[00:05:17]

Dan Culver, DO:

Yeah, it is really taking off. And I think something the field is struggling with is the fact that these are so prevalent and, by definition, asymptomatic. And so, we'll get a little bit later to what do you do about them?

But I also wonder if you want to comment on some of the other similarities in terms of genetics or in terms of physiology between what's been seen in the ILA population and what we might see in a more typical ILD population.

Aman Pande, MD:

So the ILA population is sort of a broader spectrum of the ILD population I would say. They tend to be older patients, men more often than women, smokers and patients with some degree of impairment of lung function. And, in terms of genetics, there is also a higher proportion of these individuals who have certain abnormalities like the Mach five B promoter mutation, for instance.

[00:06:15]

Ruchi Yadav, MD:

Yeah, I think so. There's a lot of similarities and if you look at these risk factors, be it the clinical presentation, the genetics between patients of ILA versus ILD, they share those similar things. One thing that I can add is short telomeres is also another risk factor. So, as Aman said, I think they're the same spectrum. It's ILA is on the same spectrum. Early stage diseases that we pick up in ILD's are the other end of the spectrum.

Dan Culver, DO:

So, it brings to mind the metaphor of an iceberg where a small portion of it is discernible and it's ILD and it's symptomatic. And a huge proportion of it is below the surface and may never cause a problem. But certainly, it seems like in the populations a lot of the risk factors, the genetics, overlap quite a bit.

And so, I think the concept of a spectrum makes a lot of sense to me. Now, you mentioned, Ruchi, that there's been a statement by the American Thoracic Society recently that revised the definition of ILA and that this played off a 2020 position paper from the Fleischner Society. Those are the two main documents in this space.

So, do either of you care to discuss what the changes are, whether or not the Fleischner document is still relevant, or if the ATS made some significant changes that are relevant for the audience?

Ruchi Yadav, MD:

So, the ATS statement, I think it was much needed, and it still has a lot of criteria in its definition, which it has retained from Fleischner, but it made certain changes.

And the couple of changes which really stand out is in the Fleischner paper patients, the high-risk patient population were not included in ILA. And the new paper by the ATS removes this exclusion of a high-risk population. So, patients who are smokers in lung cancer screening, patients with autoimmune disease, patients with familial history of pulmonary fibrosis. And another exclusion is the need for the findings to be incidental.

At the same time, we have to remember that some of the patients that are identified are going to be incidental. As you rightfully said, some patients undergoing coronary artery CT are going to be incidentally detected. To me, in practice, one of the biggest subset populations, which will be incidental for ILAs, are going to be oncology surveillance patients with thoracic and extra thoracic malignancy undergoing surveillance.

[00:09:05]

The ATS statement paper says, as with Fleischner, the findings have to be bilateral, but gives an exception that in certain high-risk populations, such as patients with familial pulmonary fibrosis, the findings can be unilateral. And I think the biggest distinction is what it did to make a distinction between ILA and ILD, with the presence of symptoms and physiological impairment and CT showing progression or features of fibrosis.

So those to me were the major changes that the ATS made over Fleischner and, I mean, there's certain clinical aspects to it also, which I'm gonna let Aman talk about with recommendations, regard to follow up and stuff.

Aman Pande, MD:

Sure. No, I actually like the new ATS guideline. I think it's simpler. And it's, I think, more directed at the radiologist, so that they remove the clinical component from this.

It's descriptive and it's broad. It catches all these patients, and then it allows the clinician to put those findings in perspective and see if there's a family history or if there's an exposure, or if there is an associated physiologic impairment and symptoms, that's when it becomes a disease. So, I think it's a broader and easier, and simpler definition. I like it.

Dan Culver, DO:

One of the other things I like about it is that it gives a number of exclusions - what's not ILA - and it does beg the question of what are the features, radiographically, that you typically tend to describe as being a securely ILA, confidently?

What are ones that are, you know, perhaps less associated with risks for progression? And then, what are some things that you see, Ruchi, as calls that are made sometimes that are not clearly ILA, and yet people get excited about them and put them in a report and generate that kind of follow-up.

Ruchi Yadav, MD:

So, three parts to your question. I'm going to start with the first one talking about what are the CT features that actually constitute ILA.

So, features such as reticulation, ground glass abnormalities, architectural distortion, traction, bronchiectasis and honeycombing. So these are features that constitute ILA. The Fleischner society came up with subcategories of these morphological findings.

And so, the first category was Nons subpleural Islas, which where you have reticulation or ground glass, which does not have a predominant subpleural predilection. So, these findings are away from the subdural aspects of the-

Dan Culver, DO:

So that would be an unusual pattern if we're talking about ILD. It would make you wonder.

[00:11:59]

Ruchi Yadav, MD:

Right, that's true. Yeah. And again, you know, one of the reasons, and I really like this classification, is because it has some prognostic implications. So those who have non-subpleural ILAs have less clinical significance. They do not have physiological impairment and no risk of progression, or they're not associated with increased mortality.

And then we have subpleural ILAs, where you have features which could be fibrotic and non-fibrotic. So, they categorize, when you have ground glass and reticulation in the subpleural aspects of the lung: that was subpleural non fibrotic ILA.

And, then you have subpleural fibrotic ILA, where you have features of traction bronchiectasis and/or honeycombing within the subpleural aspects of the lung with architectural distortion.

Ruchi Yadav, MD:

And the importance is: when you have subpleural fibrotic ILA, this is the patient who has a risk. These are the imaging features with increased chances of progression, with increased physiological impairment and increased risk of mortality. So those were the radiological features in terms of ILA.

It's important that we do not call some of the other incidental findings that we may see on CT's and term them as an ILA. One of the things that I see sometimes, see my residents do, is when we see this fibrosis adjacent to the osteophytes in the vertebral body, I mean - it is fibrosis - but it is basic. That is not something that you call an ILA.

When your finding is a unilateral, you know, that's when you do not call them ILA. When you have, like smoking-related respiratory bronchiolitis, central lobular nodules, those are not really ILA. When you see findings in a patient who's in heart failure with edema and interlobular septal thickening, that's not ILA, you know, findings, which may suggest aspirations such as central lobular nodules and bronchial wall thickening and airspace. So, kind of refraining, you know, those are not really ILA.

In terms of, as I said, there are certain features which have increased risk of progression on CT scan and a lot of data, which suggests that when the findings are subpleural and in the lower lobes and findings are fibrotic, such as you know, when you see traction, bronchiectasis, honeycombing, and even reticulation these are imaging features that have increased risk of progression over time.

Dan Culver, DO:

Yeah, I definitely want to get into the progression piece a little bit more here 'cause I think that's really clinically relevant. But before we do that, I just want to get your directive perhaps to the listenership about people who have this less than 5% of involved lung and who have non fibrotic findings in dependent areas and, and perhaps this, public service plea around prone imaging. So, do you want to talk about that for a minute?

[00:15:00]

Ruchi Yadav, MD:

I think it's gonna be patient based, you know. So if you have a high-risk patient who has just dependent abnormalities in the lungs, then it's extremely crucial that we do prone imaging to ensure that there is no subtle ILD in that location.

But if this isn't, if it's just in a general population, you see just dependent densities when your risk of having any kind of ILD or ILE is low. You know, those are dependent atelectasis, so you need to be very mindful.

Dan Culver, DO:

Yeah, I'm thinking a lot about those probably get over called, right. A person who might be overweight has only dependent nonfibrotic features. And more likely than not, you're dealing with atelectasis there rather than you're dealing with a true ILA.

Ruchi Yadav, MD:

That's true. And I think just looking at your prior imaging is extremely helpful. I mean, those are your best. I think for radiologists looking at priors is an amazing tool. You don't want to over-image a patient, so just look at a prior.

If you have a prior and there was nothing there, that also helps you make that distinction that this is just gravitational atelectasis. And as I said, again having - if it's a patient with scleroderma, your suspicion for ILD is high. I mean, these patients have a high incidence of interstitial lung disease, then you know I'm gonna take that with a pinch of salt. Even if I feel that it's dependent atelectasis, I think it would be prudent to do and prone imaging in those patients to exclude very early ILD.

Dan Culver, DO:

Sure, sure. Well, you mentioned progression, and I know there's risks of radiographic progression and there are radiographic features that predict progression, but Aman, what do we know about progression across the whole population and who gets it and what the chances are?

Aman Pande, MD:

Well, that's an important question, Dan, that I don't have a very good answer to unfortunately. What we have are research definitions, and in my opinion, these research definitions are somewhat loose and easy to hit.

For instance, if there is radiographic progression to the eye of the radiologist, then that's progression. If there is a decline in pulmonary function without any particular specifications about what decline, that's progression. And then if a patient were to report symptoms, that's progression.

So that's how it's defined in the research world. And based upon those criteria a fair proportion of our patients with ILA will show evidence of progression.

How much of that is clinically significant progression, and how much of that is going to evolve into clinically significant disease? That is the important question. I don't think we quite have an answer to that yet.

Dan Culver, DO:

You alluded a little earlier to high mortality in the ILA population compared to the patients from the same cohorts that did not have ILAs.

[00:18:05]

How closely do you think that the data around progression relate to risk of mortality? We would normally think, of course, that we would want to see that these had translated into overt ILD in order to attribute mortality. Is there a solidity around those data?

Aman Pande, MD:

So, a couple things I would say here. First is that it's all-cause mortality that they're talking about. So, it's not risk respiratory-specific mortality. And the other very big thing is that it's an association. And association is vastly different from causation.

It could very well be that these are a marker of some other process that is driving the mortality, and so we have to bear that in mind. The association is striking, it's important, but we need to learn more about it.

Dan Culver, DO: Okay. So, you know, one of the things that I think is coming down the pike and all of us are waiting for it to be clinically available more than it is now, are more sophisticated and robust ways to measure radiographic progression.

You mentioned the eyeball test before, and I always wonder about the kappa for the eyeball test. It's usually not very high when we look at it. Even for things like honeycomb cystic changes, which should be quite clear, there's a poor kappa, at least at the best moderate agreement for most of those features.

So it begs the question of the hot topic of the day. When is AI going to get us out of all of this? And before we even get to AI, are there at least some quantitative platforms that you think are useful, Ruchi?

Ruchi Yadav, MD:

Yeah. I mean, I think AI is the future and I think it's not ready as yet, but in due course. There will be a time, I'm hoping, in the next five to 10 years where it's going to revolutionize the imaging of ILAs and ILDs.

As you mentioned, I mean all these definitions of ILAs is very logical and it's based on visual assessment. And when you do visual assessment, as you know, there's a lot of subjective variability. There's interobserver variability and it's not reproducible. And I think that's the strength of AI. So, all these AI tools machine learning and quantitative CTs, they do an excellent job. You know, we are trying to pick up very early disease and they're very good at detecting those and I think, more importantly, they do a better job than radiologists in quantifying those changes and diagnosing subtle interval change.

So, if you're looking for disease progression, and especially in ILA, that progression, those who have a progressive phenotype is very slow. And these AI tools do a great job in detecting that interval change, the subtle progression.

[00:20:56]

Unfortunately, all these tools with all the data that we have is in the research room. There's still a huge gap between clinical practice and research. And I think the need is that, you know, we need to do more prospective studies moving forward, to validate all the specificity that we have with all these new AI tools.

And, I think somewhere along the line we need to standardize the CT protocols. There are the CT measurements that we get from all these AI tools, you know, that tends to be a lot of variability because the CT protocols have not been standardized.

Ruchi Yadav, MD:

And so I think that's also a very important thing that we need to be mindful of when we are trying to evaluate ILA and ILDs. We want the slice thickness to be less than 1.5 millimeters. We want to make sure that we are doing high resolution CAT scans. We can get away with low-dose CT for lung cancer screening, but we need to do proper HR CT for these patients to pick up those nuanced changes.

And you know, one thing I will definitely say is that photon counting CT, that we [Cleveland Clinic] have that available. Not very many sim centers have that, but I think it's a great tool, especially in the domain of ILAs. We can go to 0.2 millimeter slice thickness with these photon counting CTs.

And as a result, picking up very subtle reticulation, picking up very subtle traction bronchiectasis and traction bronchiectasis. Again, features which have implications to us for patient survival and mortality is extremely--, it's easier for the radiologist to make those when we go sub-millimeter and slice thickness.

So, I think in the next five to 10 years, I believe we'll be there when all these quantitative imaging tools that we have and the promising results that they show on our research population, we will be able to have them in clinical practice. And one thing that I'm super, super excited about is seeing these tools being used for population-based screening.

I mean, lung cancer screening is exploding throughout US and we already are using CAD for lung cancer screening for nodule detection. We are using it for identifying coronary artery calcification in those patients. So having these tools to evaluate ILA and ILD in these patients, when you're doing those huge numbers, it’s going to be, I think, a game changer.

Dan Culver, DO:

You're already scaring me, Ruchi. We're going to have lots of disease we find, and I hope that we are good enough to know what to do with the things we find. I'm afraid that we might be lagging in the actionable data as compared to what we can find, so we'll see.

That does bring up the issue though, of common clinical scenario is a patient is found to have ILA. Now they come in to see you, Dr. Pande, in your clinic. They may or may not have some risk factors for interstitial lung disease. What are you going to do?

[00:24:15]

Aman Pande, MD:

Oh boy. It depends on whether or not there are certain predisposing factors. Like for instance, Dr. Yadav commented on patients with scleroderma.

Now that's a very important population where there is a high prevalence of interstitial disease and it's actionable. These are patients who are on immunosuppressant therapy and often that finding will lead to a change in treatment. So, there's that extreme. Then there's included in that are also folks who have familial ILD.

So, if there's this patient who has the incidental findings happens to have a couple of first-degree relatives with IPF or any kind of fibrotic disease, well, that changes the equation. In that case, I'm much more likely to make stronger recommendations for closer follow up. On the other hand, if we have someone who is incidentally found and doesn't have any of those risk factors, then I am careful about how I make these suggestions.

I always tell patients that I'm asking you to come back out of an abundance of caution. I don't want you to worry. I don't think that this necessarily means that this is something that is going to develop into a life-changing disease, but out of an abundance of caution, maybe you should come back in three years and have another set of PFTs. And if I see a change on that, or if I were to see some new symptoms, that's when I would consider getting another CT scan of the chest.

Dan Culver, DO:

So as some monitoring but trying not to get the patient to go onto Google right away and look up what happens in IPF and go down that hole.

Aman Pande, MD:

Very true. And actually, that it's all about screening really, and the principles of screening, and this is why we screen for lung cancer and we don't screen for prostate cancer.

The reason is that with lung cancer, when you detect it early, you can do something that will make a huge difference to the outcome. And with prostate cancer, it's very prevalent and even if you detected it early, it's unlikely that it would change the outcome very much. So, I would have to say that ILAs are more like prostate cancer in that regard.

Dan Culver, DO:

I think that's important to bear in mind as we see more and more prevalent use of antifibrotics. And I think the old adage of, well, when you're a hammer, all the world appears to be a nail, is something we have to bear in mind as we navigate through this ILA space.

Aman Pande, MD:

I couldn't agree with you more, Dan.

There's a lot of anxiety that our patients have around this. Anyone who comes to see us with an interstitial lung abnormality with the word fibrosis anywhere in the report usually has visited Google and usually is very anxious and has lost sleep.

[00:27:11]

So, I think we also have to reassure our patients that for those very early changes without risk factors the likelihood that this is going to make a big difference to their life is relatively small, and our recommendations are born out of an abundance of caution.

Dan Culver, DO:

There you have it, ILAs are all over the place, even though we didn't know they existed 15 years ago. Now we're quite aware of them. Seems like we're going to be able to measure them in a more sensitive and more sophisticated way pretty soon. We're going to understand more about their impact on the population level, and hopefully at some point in time how we manage individual patients will catch up to that population level data.

So, I want to thank both of you for joining me today. Thank you Ruchi.

Ruchi Yadav, MD:

 Thank you, Dan, for having me. It was a pleasure talking with you.

 Dan Culver, DO:

And thank you Aman.

Aman Pande, MD:

Oh, thank you very much for having me. This was fun.

Dan Culver, DO:

That concludes another episode of Respiratory Exchange. Thank you again for joining us, and we look forward to seeing you on the next edition.

Raed Dweik, MD:

Thank you for listening to this episode of the Respiratory Exchange Podcast. You can find additional podcast episodes on our website, clevelandclinic.org/podcasts, or wherever you get your podcasts.