Expert Perspectives on the Identification and Management of Drug-Induced Interstitial Lung Disease

Axel Grothey, MD

Clinical Experience With Interstitial Lung Disease

OncLive^®: What has been your experience with drug-induced interstitial lung disease (ILD)?

Axel Grothey, MD: My personal experience with drug-induced ILD is that the clinical symptoms can actually vary quite significantly, depending on how much of the lung is affected. Some- times, patients are completely asymptomatic, and we just see a hazy opacity bilaterally on routine CT scans as a first indicator that some ILD could be involved.

While I’ve had patients who were on antibody-drug conjugates [ADCs] and presented with ILD, fortunately none had life-threatening complications. I have seen patients who develop patchy bilateral interstitial infiltrates on scans, and it’s definitely something I’m very alert to. The next step is withholding the treatment, and depending on the symptoms’ severity, starting glucocorticoids, although the value of steroid treatment is not quite validated—it’s more anecdotal experience that we have. Withholding the drug is clearly important, and then [possibly] starting some steroids.

Jarushka Naidoo, MBBCh: I have a research interest, as well as a clinical interest, in patients who develop immune-related adverse events [AEs] from immunotherapy for cancer and specifically for lung cancer, which is the tumor type and the patients I look after. We know that the pneumonitis that occurs in patients with lung cancer tends to be more severe and that this is a more common immune-related AE in this cancer type compared to other cancer types. [We have had] extensive experience of diagnosing patients who have developed pneumonitis from anti–PD-1 or anti-PD-L1 agents, either as standard of care or in the context of clinical trials.

I think some of the observations we have made, in terms of looking after these patients, is that unlike [with] bleomycin where there is a characteristic involvement of the bases, there is no characteristic radiographic feature that indicates pneumonitis. If there is a radiographic abnormality, that may span a number of appearances.

The second issue I think is discerning between pneumonitis and other competing differential diagnoses, such as pneumonia, progression of cancer within the lungs, and some rare infections such as fungal pneumonias. For many of these complications, particularly lung cancer patients, the images can all look very similar, and the clinical presentations can all look very similar. For that reason, in those who are symptomatic, I usually advocate for trying to do an extensive work up.

Hope S. Rugo, MD, FASCO: The way we first found that CDK 4/6 [cyclin-dependent kinase 4 and 6] inhibitors could cause ILD was in trials that were expanding the use of these drugs and treating patients in a later-line setting. The bottom line is that it was discovered that all of the CDK 4/6 inhibitors can cause ILD. The rates are extremely low, and the severity is generally low as well. I actually haven’t seen anything clinically significant in terms of ILD with CDK 4/6 inhibitors. If you see pulmonary changes on scans or a patient has symptoms, you have to keep this in mind, but it is uncommon.

Tanios Bekaii-Saab, MD: We have seen ILD—or its precursor, which is pulmonary damage and pneumonitis that ultimately leads to ILD—with a number of agents across the last 2 to 3 decades. These were rare occurrences, amounting to about 1% to 2% of all patients treated with some forms of chemotherapy or immunotherapy agents, specifically PD-1 and PD-L1 inhibitors.

Although the incidence is extremely low, at least with the agents that we commonly use in clinic, when it happens it is can be quite devastating for patients; they end up with long- term toxicities to the lung. Many can be disabled and on lifetime oxygen because of these complications. These are very serious complications when they do occur.

Tiffany Traina, MD: ILD and pneumonitis have become
AEs that, as oncologists, we need to be increasingly aware of. There have been many drugs that we’ve used over the years that we know have increased the potential risk for pneumonitis and often we have been reactive to symptoms or imaging findings. More recently, some very powerful drugs that have tremendous efficacy have also been associated with ILD or pneumonitis.

As a practicing oncologist, we likely all have seen some degree of drug-related ILD or pneumonitis from the very active agents that we used over time. I have seen this occur rarely with traditional cytotoxic agents. I’ve seen this occur with CDK-4/6 inhibitors and having participated in some of the antibody-drug conjugate trials, particularly with trastuzumab deruxtecan, I did experience ILD or pneumonitis in the context of the clinical trials. I would say that following guidelines within a protocol is helpful, but now that we have these drugs available to us off-study I think it is increasingly important for oncologists to be aware of the potential risks, be aware of some of the risk factors that are associated with higher likelihood of ILD, and to understand the management of ILD.

Monitoring for ILD

OncLive^®: How do clinicians monitor for ILD?

Grothey: At this point in time, GI oncologists do not specifically monitor patients for ILD. We are aware of ILD as a potential AE, and we would of course be more alert to patient symptoms if they come with a dry cough and shortness of breath, and radiographic signs. My threshold to obtain a CT scan, if a patient is on a potential ILD-inducing agent, is very low: If they got a cough or other clinical symptoms, I’d obtain a high-resolution chest CT scan. I would be aware of that and not necessarily monitor outpatients, but have a lower threshold to really start a diagnostic process to work up [any of] their symptoms.

ILD is a rare AE, but I think we need to start being more aware of it, and then having the referral pattern either to our pulmonologists or to people who know how to handle ILD internally. I don’t think we proactively monitor patients as much as we will react to unexpected findings or symptoms that patients develop. This might change with the emerging agents and the inclusion of ADCs and immunotherapy in more of our treatment approaches.

Naidoo: In terms of monitoring for ILD, most patients with cancers, including lung cancer, will have a baseline CT scan; that includes a CT scan of their chest before they start an anticancer therapy. At that time, it may be appreciated that patients may have a background ILD or a background process within their lung that may predispose them to the development of a drug-induced ILD. If something like that occurs or a risk factor for a patient is flagged and known, then these patients may be directed towards a pulmonologist for a risk assessment before the anticancer therapy is begun.

If we perceive that this patient has an increased risk of ILD, or they have a known history of ILD, either de novo or for some other reason, we will evaluate this patient with our pulmonologists, obtain a full pulmonary function evaluation, rule out other complications, perhaps right heart failure, and do an echo, all of this sort of work up before we start treatment in order to sort of provide a risk assessment to the patient of how likely or unlikely we think it is for an ILD to happen.

Rugo: We have several important monitoring parameters. First, we monitor patients who are on agents that are associated with ILD with regular scans. That’s important because if
a patient has been stable for a little while or doesn’t have pulmonary findings in terms of their site of cancer, they might not have scans that include their lungs. If you know that a drug is associated with a risk of ILD, it’s really important that you do get some kind of regular imaging in order to evaluate for ILD. The second is a much earlier assessment of the lungs in patients who have relatively mild symptoms—say, an upper respiratory infection and a cough. [Even] for very nonspecific symptoms, I always get a chest x-ray. Third, both patients and staff need education about really jumping on a symptom if a patient calls in, rather than waiting and saying, “Call me if it gets worse.” I have reinforced the need to report symptoms with patients and with our nurses.

Traina: It’s important to recognize that pneumonitis
or ILD is a potential risk for patients on a drug like trastuzumab deruxtecan. Having that mindset to recognize when a patient might have an incidental finding of ground glass opacities on imaging when they are asymptomatic, that sort of grade 1, mild ground glass opacities, could be an early indication of ILD. Other symptoms to monitor for on a regular basis are things like a new onset of a cough, particularly a dry cough, or dyspnea, some shortness of breath, or even fever. These can all be early indications of an interstitial lung disease or pneumonitis. I find that if this potential is in your differential diagnosis, you are more likely to respond and react quickly to this.

Is ILD a Class Effect Among ADCs?

OncLive^®: Is ILD a class effect among ADCs?

Grothey: It’s not fully clear whether ADCs in general have ILD as a class effect or whether ILD is related to specific agents. I do believe we see some ADCs with a higher prevalence and incidence of ILD, but we need larger studies and further investigation to really figure out whether it’s a class effect or some agents stand out.

Rugo: It’s a great question: Is there a class effect among ADCs that we’re going to be seeing with ILD? I don’t necessarily believe there is, because we haven’t seen it with all of the ADCs. I would, however, keep in mind that for a rare toxicity, it might take a very long time to really understand that a drug is associated with ILD. I do think that we don’t know the answer to whether this is a class effect among ADCs yet. I will say that the risk—if there is a risk—with all ADCs is very low. For some agents there is a class effect, but the incidence varies a lot among the agents. For example, with oral tyrosine kinase inhibitors, ILD has been reported with many different agents, and there do appear to be some class effects. I think that the jury is generally out, but it’s not clear that all ADCs cause ILD.

Factors that Place Patients at Higher Risk for ILD

OncLive^®: What patient-related factors might place them at higher risk for ILD?

Traina: The literature has shown that there are particular patient factors that might place them at higher risk for develop- ing ILD. Those include having some underlying interstitial lung disease to begin with at baseline. For patients who are smokers, previous smokers, or who have poor a performance status, there is concern for any increased potential risk of interstitial lung disease.

There has been some description that patients over the age of 60, or patients of Japanese or African American descent, may also have a higher risk of interstitial lung disease with the use of trastuzumab deruxtecan. Men have been described as having a slightly higher risk of ILD. This drug has been looked at in other cancers besides HER2-positive breast cancer.

Having multiple prior lines of chemotherapy and being heavily pretreated [is another factor]. Many of the patients who were treated in the DESTINY-Breast01 study seemed predisposed to, or had a higher risk of, ILD.

Signs and Symptoms of ILD and How ILD is Diagnosed?

OncLive^®: What are the signs and symptoms of ILD? How is it diagnosed?

Grothey: Patients’ symptoms and signs that make us suspect ILD come in various forms and shapes. Many patients are asymptomatic when they come in, then we are surprised to see patchy infiltrates on CT scans. We get a call from radiology and they say, “Hey, a patient might have an infection or some ILD related to some toxic or autoimmune or hypersensitivity effect due to medication.” Those patients would be classified as grade 1 because they might be asymptomatic. We discontinue the treatment and follow these patients to see if they develop symptoms, namely cough and some shortness of breath upon exertion.

Naidoo: The most common signs and symptoms that patients with a drug-induced ILD will present with are usually cough and shortness of breath. Less commonly, patients may present with fever or chest pain. The other notable diagnostic factor that should be kept in mind by any provider is some- times patients may not necessarily notice cough or shortness of breath when they are at rest, but if this is suspected, doing a walking oxygen saturation or assessing shortness of breath after a short walk is often a good way to diagnose pneumonitis or drug-induced ILD. I think the most important take-home message is that drug-induced ILD, particularly from the PD-1 or the PD-L1 agents, can be fatal in some situations. Unlike some other AEs, I think drug-induced ILD is more challenging in that patients may become acutely unwell and their ILD may progress quickly, resulting in a sudden clinical deterioration.

Rugo: Identifying ILD can really be a challenge. We have not seen other clinical manifestations other than cough, shortness of breath, and sometimes tightness in the chest as a particular symptom associated with ILD; those are the real symptoms to look for.

What I would recommend and highlight is that if a patient comes in with a respiratory symptom, it should be evaluated right away. I think a really important message is to not to let your guard down. When a patient does have a symptom, it is important to think, “Are they on a drug that could cause ILD, and what should I do to evaluate this, given the symptoms that a patient is expressing?”

Staff and patients must also understand that if there are symptoms and a suspicion that it might be ILD, that it is important to stop the drug first; you can always restart.

Bekaii-Saab: Overall, the presenting symptoms for pneumonitis are mostly nonspecific. A red flag is when a patient, one who doesn’t necessarily have any baseline lung disease or significant lung metastases, comes in with a dry cough, with dyspnea on exertion, and then when checked in clinic their respiration rate is higher than expected and they are oxygenating a little bit lower than usual. Then the patient would get a chest x-ray, typically, where you can see some infiltrates, and that could lead to a CT scan. The chest x-ray has some low level of sensitivity and specificity, I think enough to start with, but if the patient is quite symptomatic and the chest x-ray looks completely fine, you would push for a CT scan at that point. Then, if the patient shows any evidence of pneumonitis progressing into ILD, any level of lung injury, you should stop the agent immediately; the patient should not get the drug any further.

Traina: ILD can really present across the gamut of symptoms. At the extreme, patients can be quite dyspneic with an exercise tolerance reduction. They can develop dry cough and have fevers from the underlying inflammation in the lung disease. Those would be at the extremes. In my experience with the incidental asymptomatic findings, I’ve seen ground glass opacities appear on routine imaging in patients that were entirely asymptomatic, which ultimately turned out to be a drug-related effect of the cytotoxic leading to ILD. We need to be open to and recognize that ILD can present itself in a wide variety of ways.

Management of ILD

OncLive^®: How is ILD managed?

Grothey: When we talk about management of ILD on ADCs, I discontinue the agent. I do not believe that re-exposing with a lower dose would put patients at a lower risk because there’s not a clear dose-effect relationship. This might just be a immune-related reaction, sometimes an inflammatory reaction induced by these ADCs, which might be independent of the actual dose used. I would not just use a dose reduction and then continue therapy.

When we discontinue the causative agent, a lot of patients improve, but it might take time. This is mainly related to the fact that the ADCs’ antibodies have a longer half-life. The biologic half-life of these drugs is anywhere between 3 to 6 weeks, so we cannot expect to have dramatic changes within a very short time when we discontinue these agents. It’s different with small molecules—erlotinib, for instance—or conventional chemotherapy like gemcitabine, where the shorter half-life might actually contribute to a faster resolution.

It’s not just the drug that causes the ILD, but it’s also the induced immunologic reaction, which can actually be delayed. We have some patients whose therapy was already stopped, and they developed delayed ILD. It has been recorded, so the critical issue is to monitor patients very closely over time, and do not expect that everything will revert and regress right away when you discontinue the agent.

The other point is, dependent on the degree of the inflammation, there might be some long-term scarring effects that we see in the lung. The lung function might actually be compromised to some degree on a permanent basis. This really depends on how much of the lung is affected and with how much inflammation.

Naidoo: My greatest experience in this area is with the anti– PD-1 or PD-L1 agents that are used in lung cancers and now many different cancer types for which they are licensed. For those patients, immune checkpoint inhibitors are usually held at the development of pneumonitis because we appreciate that the half-life of the PD-1 or PD-L1 agent persists after dosing is held. Usually, in a patient with symptomatic pneumonitis, a grade 2 or greater, we would hold therapy, and even in some grade 1 cases we may consider to hold therapy. In those who develop a severe pneumonitis that requires them to be admitted to the hospital or the intensive care unit, it may be appropriate to permanently discontinue therapy.

With other agents, like chemotherapies and targeted therapies, this would involve holding therapy and, in some cases, a dose reduction may be instituted. This depends on the clinical severity of the pneumonitis or drug-induced ILD event.

Rugo: Managing ILD is very much dependent on the institution, but in my practice, it depends on symptoms. A patient who develops a mild cough or changes on everolimus who can be easily treated with steroids is something that we are used
to, and we generally don’t involve pulmonology. However, with newer agents where ILD has been identified, if a patient develops symptomatic ILD, I think involving a pulmonary specialist is important; they may be able to help you monitor and understand the significance of both the pulmonary findings and the symptoms for a specific patient.

The other thing is that I think for oncologists, it’s important
to have an early intervention with steroids. That’s important, because you can always take back the steroids, but you can’t [go back and] start them earlier if you didn’t start them. There does appear to be a difference. For example, with the most recent ILD that we have identified with trastuzumab deruxtecan, starting steroids earlier and identifying this ILD earlier will make a big difference in patients not progressing to higher-grade ILD.

Serious events have been reported in patients who progressed from early low-grade to higher-grade ILD, and you may want to start steroids with very early-grade ILD. You’re going to take a differential approach based on the risks for this specific agent and its half-life. Do the ILD symptoms go away when you discontinue therapy? CDK 4/6 inhibitors have a short half-life, and you have to recognize [ILD], not continue to treat. ADCs have a long half-life, so you may want to treat [potential ILD] earlier. You may want to have a much lower threshold for using steroids for these patients and understand that it could take some time for the symptoms to resolve.

Bekaii-Saab: If ILD is suspected, evaluations should include a chest x-ray and then a high-resolution CT scan. In our institution, we do consult with the pulmonologist. They have to be involved. You want to rule out reversible causes, like infections.

If you suspect ILD, then interruption of trastuzumab deruxtecan—or any drug that can cause ILD—needs to be immediate. Depending on the symptoms, you [make] prompt intervention with steroids, with glucocorticoids, when appropriate, especially if the patient is requiring oxygen or there is evidence of significant inflammation. It’s important to keep in mind that, even as you’re [ordering] consultations and tests, intervention with steroids may help reduce the severity of the complication and may even cut down on the risk of developing further damage to the lungs.

For any agent that causes ILD, and specifically these antibody-drug conjugates, we have to keep in mind that about half of the patients will not see any improvement in ILD [once it develops], so this will be the new baseline. A small percentage of patients [will] unfortunately die, but hopefully with catching this early and intervening early, we won’t see any further death from this complication.

Traina: If I suspect that a patient has interstitial lung disease, maybe an incidental finding on their routine imaging, where they are clinically symptomatic with dyspnea, cough, or fevers, I am inclined to involve pulmonary quite early in that disease course.

The first step in managing the ILD symptoms is discontinuing the offending agent, so hold therapy. Then depending on the grade and whether a patient is asymptomatic or not, there are guidelines for instituting systemic steroids to reduce that inflammation. A steroid taper needs to be quite slow and over a period of several weeks, even as far as 4 weeks on steroids, to see improvement in symptoms.

Changes in imaging can take even longer. A grade 1 ILD could present with an asymptomatic patient with only some ground glass opacity seen on imaging and you can take a long period of time for that imaging to resolve. Nevertheless, the recommendations would be to hold drug.

Discussing Benefits and Risks of Treatment With Patients

OncLive^®: How do you discuss the benefits and potential risks of treatment with your patients?

Traina: As we face later and later lines of therapy for patients with advanced HER2-positive metastatic breast cancer, we’ve been looking for drugs that could really change the natural history and improve the outcomes for these patients.

Our institution participated in the very early and most recent studies with trastuzumab deruxtecan, and I have treated patients with this really powerful agent. One of the most impressive takeaways is the incredible responses seen and the efficacy of this drug in patients that were heavily pretreated for HER2-positive metastatic disease. To have a drug like trastuzumab deruxtecan that has clearly shown benefit with tremendous response rates, progression-free survival, and really compelling data, is a wonderful benefit for our patients.

When we take on the potential risk of interstitial lung disease, we need to have that conversation with our patients around risk/benefit balance. Patients have been symptomatically improved with the use of this drug when their cancer is under better control and looking at all of the data in its totality, having trastuzumab deruxtecan is 1 of the drugs I can offer my patients in a later line setting is a huge advantage.

Taking a look at the data in its totality, having a drug that has such clinical benefit for patients in later lines of HER2-postive metastatic breast cancer, that benefit seems to outweigh the small potential risk of interstitial lung disease.

It’s important to have these conversations together with our patients so that they’re aware of the potential risks, but also the tremendous benefit from having an active agent. We are all the better for having this drug available to our patients who have seen a number of prior lines of therapy.