Interstitial Lung Disease: An Overview

Interstitial lung disease—or ILD—is a category comprising a massive number of diseases and disorders, which affect the lung through inflammation, fibrosis, or some combination of the two. It is often quoted that there are more than 200 interstitial lung diseases, although this number may be somewhat misleading: exactly what counts as a distinct ILD depends on how finely one chooses to divide the taxonomy; drug-induced interstitial pneumonias alone, for example, account for an impressive number of individual entries. (For an outstanding, and somewhat humbling, catalogue of drugs with known pulmonary toxicity, see Pneumotox.)

Even if the number is somewhat inflated, the underlying point stands: interstitial lung disease is a difficult topic. There are a lot of diseases, a lot of acronyms, and a vocabulary in which findings, patterns, and diagnoses have an unfortunate tendency to sound alike or literally share the same name.

How I got here

My own interest in the field began in 2017, when I was an internal medicine resident. I enrolled in an elective ILD clinic for a year—roughly 20 half-day sessions—mostly because I realized I did not know much about the topic. Somewhere along the way, I grew to love it. ILD was an evolving field where keeping up with current evidence mattered, but so did being deeply interested in the particulars of a patient's life: where they worked, what they did for fun, what medications they took, what they might have inhaled, what illnesses ran in their family, and whether some seemingly incidental detail might turn out not to be incidental at all. It was also a field in which communicating compassionately was an essential skill, since the job sometimes included telling a person for the first time that they were carrying a life-limiting, terminal diagnosis.

I also loved the preservation of certain skills from general medicine, particularly given the overlap between interstitial lung disease and systemic autoimmune rheumatic disease, or SARD (i.e., the artist formerly known as “connective tissue disease”). A good ILD clinic could require you to think like a pulmonologist, radiologist, rheumatologist, pathologist, occupational historian, and general internist, sometimes within the same hour.

This website is my attempt to create the kind of go-to resource I would have wanted when I first walked into that clinic in 2017. I had excellent teachers and mentors, but breaking into the field was still surprisingly difficult. Tracking what had already been considered and become passé, what was known, what was maybe known, and what remained controversial required a lot of wandering through papers, guidelines, reviews, and terminology. There seemed to be an endless supply of acronyms. More confusingly, it was not always obvious what constituted a descriptor, what constituted a pattern, and what constituted an actual diagnosis. (Worse yet, sometimes everyone involved appeared to understand the distinction except me.)

I plan to do my best to sort all of that out here.

A few big ideas about ILD

When I try to break interstitial lung disease down into broad strokes, there are a few ideas that I think make the rest of the field substantially easier to understand.

(1) We try to make the best diagnosis we reasonably can

At centers with expertise in ILD, there is generally an effort to pursue the best possible diagnosis. This matters partly because diagnosis can influence management: whether an exposure needs to be removed, whether specific autoimmune diseases should be tested for, whether immunomodulatory or antifibrotic treatment (or both) is appropriate — the list goes on. It also matters because careful diagnosis supports the ongoing effort to understand this broad class of diseases and, when appropriate, allows interested patients to enroll in clinical trials from which they may genuinely benefit.

But the phrase best possible diagnosis is intentional. Not every patient emerges from an ILD evaluation with a perfectly labeled disease, and forcing greater specificity than the evidence allows does not make the diagnosis better. Sometimes the data support a confident diagnosis; sometimes they support a provisional diagnosis with meaningful alternatives; and sometimes, after a great deal of expert thought, the most accurate diagnosis remains unclassifiable ILD. The point is to get as close to naming the underlying disease as the available information reasonably permits, while being clear about how certain (or uncertain) we actually are.

(2) There are a few major kinds of information we use to get there

The first is clinical information, which in ILD can encompass much more than the usual medical history. People's jobs, hobbies, homes, medications, environmental exposures, smoking histories, families, pets, (when their hair went gray!,) and other illnesses can all matter. So can symptoms that initially seem to belong to another organ entirely, particularly when they suggest an underlying SARD or a telomere biology disorder. ILD therefore rewards a particular kind of nosiness: the socially acceptable medical kind, where discovering that someone restores old furniture, keeps birds, cuts stone, takes an unusual medication, or has had mildly swollen fingers for three years may suddenly reorganize an entire case.

The second major source of information is radiographic. High-resolution chest CT gives us a collection of findings—ground-glass opacity, reticulation, traction bronchiectasis, honeycombing, nodules, cysts, consolidation, mosaic attenuation, and so forth—which can be considered alongside their distribution and assembled into recognizable patterns. Almost none of these patterns are perfectly specific. Instead, they alter probabilities. A particular pattern may make one diagnosis quite likely, another less likely, and several others still entirely possible.

The third broad source of information is lavage or tissue. In contemporary ILD practice, bronchoalveolar lavage is generally most helpful when there is a very specific question one hopes it will answer. Lung tissue can likewise be enormously informative, particularly when obtained through surgical approaches, but tissue is not free information. Biopsies carry risks, and surgical lung biopsy in particular has a calculable risk of serious morbidity and mortality, and even a beautiful piece of pathology still has to be interpreted alongside everything else we know about the patient. It’s possible to undergo the procedure and still not have a diagnosis.

These distinctions took me a while to feel comfortable with when I started learning ILD. A descriptor is not necessarily a pattern, and a pattern is not necessarily a diagnosis (BUT a pattern may become a diagnosis if we cannot find a reason for the pattern - why ILD, why?!). The same pattern can appear in several diseases, and the same disease can produce more than one pattern. Remembering what level of the hierarchy you are talking about prevents a remarkable amount of confusion.

Which brings us to another big idea.

(3) More information is not automatically better information

As in many other areas of medicine, the general direction of ILD practice—at least in the region where I practice—has been toward minimizing invasive testing when the information obtained is unlikely to justify the risk involved. This has been supported by efforts to understand what can be learned from high-resolution chest CT alone, including when a radiographic pattern can tell us, with sufficient confidence, what we would probably see under the microscope if we actually pursued a biopsy.

The question, then, when considering a biopsy is not simply, could we get tissue? It is: What uncertainty are we trying to resolve? How likely is tissue to resolve it? Would resolving it change what we do? And is the value of that information worth the risk of obtaining it? Sometimes the answer is emphatically yes. Sometimes it is emphatically no. Much of expert ILD practice lives in the territory between those two answers.

This is also why multidisciplinary discussion is so central to the field. The clinical history, CT, pathology, serologic evaluation, and other pieces of data are not separate answer keys waiting to be consulted in sequence. They modify one another. A pathology pattern means something different in one clinical context than another; the same CT can imply a different diagnosis after a previously unknown exposure is discovered. The goal is to put all of these pieces on the table and ask what explanation makes the most sense when they are considered together. I try to make this systematic in a four question approach.

(4) Sometimes what the disease is doing matters as much - or more - than exactly what we call it

One important development in ILD has been attention to disease behavior in addition to disease identity. The clearest example I can think of for this is progressive pulmonary fibrosis, or PPF. To be more specific, PPF is not a disease, but rather a clinical descriptor, which refers to progression of fibrosis that can occur across a range of fibrotic interstitial lung diseases. And while the underlying diagnoses may remain quite different, the recognition that many of these patients share a clinically important pattern of progression has enabled researchers to study them together and has supported treatment strategies that cross traditional diagnostic boundaries.

In this, there is a broader idea worth keeping in mind as you move through the rest of this site: we still care deeply about determining which disease a patient has. But there may also be useful information in simply asking what the disease is doing and whether its current activity appears inflammatory, fibrotic, or a mix of both. Knowing what the disease is and what kind of activity it is demonstrating (if any, as some diseases may become quiescent) are related questions, but they are not quite the same one.

Putting it together

All of these ideas become important when we start working through actual diagnostic and management reasoning in ILD. We want the most useful diagnosis the evidence can support, but we also want to be honest about the limits of that evidence. We integrate the patient's story with their CT Chest and, when useful, consider obtaining information from invasive testing, but only pursue that testing when answering any remaining questions is likely to matter. And after deciding what we think the disease is, we continue watching what, if anything, it does.

That sounds tidy when compressed into a paragraph. Actual ILD clinics, predictably, are less tidy. So the next step is to take these broad ideas and turn them into something more practical: an approach to a patient with suspected interstitial lung disease.

A Clinical Approach to ILD

Roads to recognition

There are a number of ways that patients end up referred to ILD centers of excellence for suspected ILD. Based on my own experience (and potentially ignoring studies that could provide specific numbers/percents), here are some examples:

1. During the evaluation for chronic cough or chronic, progressive dyspnea, a patient undergoes a high-resolution CT Chest that reveals an imaging pattern associated with ILD

2. During a low-dose CT Chest for lung cancer screening, a patient is incidentally noted to have an imaging pattern associated with ILD

3. During screening in patients with known risk for ILD, such as patients with SARD, screening PFT and/or CT Chest demonstrate a pattern associated with ILD, and, less commonly,

4. During an episode of acute hypoxemic respiratory failure, when rapidly progressive ILD is suspected, often after failure to improve despite antibiotics, diuretics, and other therapies or because early in the patient’s course a CT Chest demonstrates a pattern highly suggestive of an inflammatory interstitial lung disease

With the exception of the lattermost, each of these scenarios should reasonably result in referral to an outpatient ILD practice. And, as you can see, at least in high resource settings, the discussion of an interstitial lung disease typically flows out from imaging that has already been obtained. Therefore, in discussing an approach, I will focus on approaching suspected ILD, rather than trying to design an approach for when to suspect it. (Does that make sense? I think it makes sense.) However, where possible, I will try to note the sensitivity and specificity of various findings, a la the style of rational physical exam, which could also presumably be useful for suspecting ILD in the first place.

History

***

Exam

***

Labs

***

Imaging

***

Other Data

***

Reasoning

***[4QM-ILD]