All right, hello everyone. My name is Faisal Al-Ghula. I'm a pulmonologist at Deakins Health Indiana. I'll be chairing the session today. Today we'll be talking about multifocal link cancer, which is a very interesting and complex topic. But luckily today we have a distinguished expert panel who are gonna help us to understand this very complex entity. And since the topic is very big, we're gonna start. So I will start with Dr. Frank Ditterbeck. Dr. Ditterbeck is the Chief of Thoracic Surgery, the Surgical Director of Thoracic Oncology at Yale University. Dr. Ditterbeck is leading expert in the multifocal cancer, multifocal link cancer. He has contributed immensely to the development and the definition, the classification, the treatment of the multifocal link cancer. So without further ado, Dr. Ditterbeck, you have the floor. Well, thank you. I've certainly been thinking about this topic for a while and I've been trying to struggle to understand it and frame it in a way that is useful. And I don't think we've quite gotten there, but I think we're getting there. So I don't think these are my slides though, right? These are your slides. All right, I'm sorry about that. All right, this is your slides. Okay. So I have no disclosures. So I'm gonna really focus on multifocal adenocarcinoma. So I think when you're talking about multiple pulmonary sites of lung cancer, you have to kind of differentiate. Are you talking about, I guess this mouse doesn't show up, does it? Nope, it doesn't. Okay. You know, are you talking about multifocal ground glass lipidic adenocarcinoma? Are you talking about second primary lung cancers? Are you talking about separate tumor nodules? Are you talking about mnemonic type of adenocarcinoma? Those are all different entities. I think they behave differently and you should think about them differently. I'm gonna focus really on multifocal adenocarcinoma. Just say a few words about separate primary that I think most people kind of forget about. So when we looked at this and did the analysis for it, you know, this is looking at lung cancers that have a metastatic site where both the primary and the metastatic site were biopsied and looking at concordance of biomarkers. And you can see that there's about a 30% discordance rate between the two. So, you know, biomarker pattern being the same or different is helpful, but it's not the thing that you can hang your hat on without saying, well, there's a possibility it might be discordant. This is the opposite. This is, you know, primary lung cancers, two different primary lung cancers, but looking at clonality of those, you know, you see listed there PCR studies and various sorts of things that were done, but about a third of these had different clonality. Two thirds of them had different clonality and about one third had similar clonality. And if you looked at the outcomes, whether the outcomes were different, whether it was the same or different, they weren't. So a diagnosis of multiple, you know, primaries, clonality, again, is not as helpful, but it's not something where you can say, well, absolutely, if clonality is the same, it has to be a MET. You know, it doesn't have to be a MET and the behavior doesn't suggest that that patient is going to do worse if you think by other criteria that these are separate primaries. So that's all I'm going to say about separate primary lung cancers. Just be a little bit careful about that. So I'm going to focus on multifocal ground glass lipidic adenocarcinoma. So this is a typical, you know, CT. There's a lesion in the apex of the right epilope and a lesion in the left upper lobe. And the clinical criteria that we came up with were, you know, there are multiple subsolid nodules, pure ground glass, part solid, with at least one that is suspected to be a cancer. So that's really the key thing. And I would say this is based primarily on imaging. And you can apply this, whether or not the other lesions have been biopsied, if you're convinced that these are, you know, malignant. So there's also pathologic criteria. And this is if there are multiple foci of lipidic predominant adenocarcinoma, or AIS, or MIA. It doesn't apply to squamous or something like that. And again, it applies whether there is a, you know, detailed histological examination of different, you know, foci and looking at, you know, what is the percent of acinar and papillary and lipidic. And all of those different things are looking at clonality studies. In fact, the recommendation was, do not do that. You don't need to do that. If these are multiple ground glass lesions and lipidic, you know, have a strong lipidic component, you know that this is multifocal adenocarcinoma. And, you know, a detailed histological examination is not gonna change it. And it also applies if you've resected one, and this has been looked at pathologically, and there are other nodules that the patient has, other ground glass capacities that you haven't biopsied, you've decided to just observe those, you still call it multifocal. So why do we separate this out as a separate entity? Well, I think the behavior is different. So at the time that we had defined this for the staging system, we did the systematic literature review, and look at the five-year survival. For all patients, 85%, for N0 patients, 91%. So distinctly favorable survival, better than, you know, adenocarcinoma, than lung cancer in general. Also, recurrence patterns are different. Distant recurrences, really quite rare. You know, new primary lung cancers arising in other, you know, GGOs is much more common. And, you know, there's some regional recurrences there, lung in N2, 3, but probably most of those are actually new primary lung cancers as well. So low propensity for distant mets. This is a different behavior. High propensity for new ground glass lipidic tumors. So different behavior than what we're used to. How do we classify these by the T and an M classification? Well, the T category is determined by the highest T. So by size of the invasive component or the solid component. And then in parentheses, you put in an M for multiple, or three if you know that there are three lesions that the patient has. And the N and the M is assigned collectively for all of them. So how do we manage these? I think you need to do a little bit less investigation for clinical stage because the propensity for nodal mets or distant mets is so low. Basically, you manage each nodule individually. If it warrants intervention, you intervene. If it doesn't, you leave it alone and you watch it. And this is an example of why you don't necessarily intervene on too many. This is a study of patients that had one lung cancer that was resected and it had additional ground glass opacity. So exactly this multifocal adenocarcinoma that we're talking about. And you can see there in 10-year follow-up that about two-thirds of those additional lesions did not progress. So just because he had one lung cancer didn't mean that the others are inevitably going to be a problem. So don't necessarily jump on everything just because it's there. It's really the size of the solid or invasive component that determines prognosis. Multiple studies have shown this. So this is kind of my trigger, how I look at it. A new solid area of two millimeters or greater on mediastinal windows or a consolidated area of greater than six millimeters on lung windows is when I start thinking about doing something about it. But you have to balance it against the rate of change. If it took five years for that solid component to develop, I'm not sure I care. If that patient's 85 years old, I'm not sure that means I'm gonna pull the trigger. You have to balance it. So just to summarize, we're talking about this multifocal adenocarcinoma. So ground glass capacities, maybe part solid on the CT scan, maybe different subtypes of adenocarcinoma, different proportions, maybe the same. You really don't need to do a detailed histological examination for it. So subsolid nodules, lipidic predominant adenocarcinoma. We use the TNNM classification based on the highest T lesion, one NNM for all, and management, I think, less investigation for clinical stage and manage each one individually. So thank you. I will be fairly briefly discussing the role of PET scan and biopsy in the management of these slow-growing subsolid lung nodules. I have no disclosures related to the contents of this presentation. And really, the two key questions I wanna go over with you are what is the role of PET scanning, and if any, in the evaluation of subsolid lung nodules, and really, what's the role of biopsy? And I think what you'll hear from the different talks is, as Dr. Hossain already mentioned, some of the complexity with these issues. And because of the complexity and imperfect data, how practice may vary. But I'll review some of the data that is available on both of these questions. So in fact, first glance at answers from a review paper by Dr. Diederbeck published in 2020 in Clinics and Chest Medicine. It was really PET has no role in the management of ground glass nodules. Now, of course, not all subsolid nodules are purely ground glass, so I will review some of that data. And also, in terms of biopsy, the problem with bronchoscopic or CT-guided biopsy is that both have a very high false negative rate, although there's much less data available on bronchoscopic biopsy. I will review this. So why would you consider PET scan for subsolid nodules? Well, there's some guidelines. Often, as well, we're left, I find, asking ourselves, should we order a PET scan? Because I find a recommendation for PET scan in the conclusion of my CT report. And not always a recommendation that's consistent with what's in our guidelines. The chest guidelines on the management of lung nodules, they're currently being updated, but these are the guidelines from 2013. For PET, for subsolid nodules, recommend PET scan when the solid component is over eight millimeters. And so that PET should not be used for a part-solid nodule that has a solid component that's less than eight millimeters, because really, at that point, the sensitivity of PET scanning is quite low. The Fleischner guidelines, in fact, don't make a recommend, sorry, the Fleischner guidelines recommend PET CT biopsy or resection for subsolid nodules, but really, when the solid component is over eight millimeters, and I'll be curious to hear after from Dr. Hussain, because our Fleischner guidelines for lesions under six millimeters really don't recommend any follow-up, right? So for AAH, that's one question, but we're really mostly here talking about our slow-growing subsolid nodules, and Fleischner, really, as you know, other than the PET question, addresses more nodules that are stable over time, in terms of the intervals of follow-up. The BTS guidelines of 2015, in fact, recognize that PET CT may be suboptimal at characterizing subsolid nodules as either benign or malignant. So looking at some of the data that supports this, as was already mentioned, a lot of the series that are published come from Asia. The first series is coming from Japan, looking at PET in relation to the Noguchi classification, which is a classification that antedates the 2011 ISLAC ERS and ATS classification of adenocarcinoma that was just discussed, but essentially concluded that PET scan had poor sensitivity for subsolid lesions, particularly for the lesions that are presenting as pure ground glass, so really more the AAH and the AIS pathology. Another series from Korea, looking at 89 patients with 134 ground glass predominant nodules, looked at SUV and found that there was a correlation with mean size and a negative correlation with ground glass, so the larger the ground glass component, the less likely to really have any PET uptake, but really no clear advantage, certainly not just in terms of assessing the lesion, but in terms of staging for the more purely ground glass lesions, and this sort of goes with the threshold of detection we know about for PET scan for solid nodules when once we're below eight millimeters, the usefulness just is not there as much. Another series, this one from Japan. Again, this was a multicenter analysis of CT and PET findings, 500 patients with stage one adenocarcinoma. In this series, they did report and they classified nodules according to percentage ground glass and in this case, they used the tumor disappearance ratio, but a bit the same idea as having consolidation to tumor ratio when you try to assess the proportion of solid component relative to ground glass of the whole lesion, and found that the maximum SUV could be a predictor of surgical outcomes, and so that it may have a role, but again, more this idea of using it in the subsolid, the part solid subsolid nodules where you have a significant solid component, but in fact, some of these patients, even though it's a minority, some of these part solid nodules, even when they had a greater than 50% ground glass component, up to 6% had nodal mets and up to 4% recurrence rate was reported in this series. In terms of the role of PET scanning in screen detected subsolid nodule, this was a study from Italy looking at PET CT over the first six years of the single arm Cosmos study that was done in Italy, and they used PET, PET was in all nodules that were CT detected as part of this study, and of course, reported on the important role of PET scan for solid nodules and nodules that were larger, but in fact, even when they looked at subsolid nodules, the sensitivity of PET scan for detection of malignancy, even though we know that a large proportion of these subsolid nodules on this adenocarcinoma spectrum and are, in fact, on a malignancy spectrum, the sensitivity was only 21% for the subsolid nodules that were less than 15 millimeters, and still a very imperfect 64% for subsolid nodules greater than 15 millimeters, so importantly, a negative PET scan really does not exclude malignancy in subsolid nodules. And so these are older recommendations, but I think they kind of fit with the guidelines I've mentioned already of perhaps having a usefulness for PET, but in the more narrow subgroup of patients with subsolid nodules that have a significant solid component, whether that's eight or a 10 millimeter cutoff, and trying to limit its use maybe to that context. In terms of the role of biopsy for these challenging subsolid lesions, the classification of adenocarcinoma published in 2011 was already mentioned. It's important to note that this classification is based on surgical resection specimens, and so, in fact, while as part of the document, a terminology to be used in small biopsies was proposed, but there was also a very important recommendation that a comment should be added to pathology reports when reporting the results of a small biopsy that really an invasive component cannot be excluded if pure lipidic growth is observed in a small biopsy specimen, and so, in fact, the diagnosis of AIS or minimally invasive adenocarcinoma really can't be firmly established without looking at the histology of the entire tumor. Looking at some of the series that have been published, and again, the data is mainly coming from CT-guided biopsy that I'm presenting, there are series where CT-guided biopsy has been shown to provide a reasonable result in terms of sampling of these subsolid nodules. This was a study, again, from Korea by Kim and colleagues looking at the diagnostic accuracy of CT-guided core biopsies in 50 patients with predominantly ground glass nodule. They reported a pretty high diagnostic accuracy. What's important is that from their report, they removed five non-diagnostic samples, that's 10% of their samples, so already out of 50 patients having TTNA, five had a non-diagnostic result. Of course, if you remove those results and report diagnostic accuracy only in the patients who had a useful and informative TTNA, the diagnostic accuracy starts to look better. So in fact, when they looked at final pathology versus the biopsy, there was a discordance in histology in almost a third of lesions that were established to be malignant following resection. Another series, this one's from Memorial Sloan Kettering of 221 patients who had a CT-guided biopsy and went on to have resected lung adenociae. Similar idea, the concordance between biopsy and surgery for the predominant subtype of adenociae was 77%, and the sensitivity for the more aggressive types of adenociae was only 48%. And the last series I'm showing is from Mass General, 86% who had TTNA or TTNB for subsolid nodules. A high success rate reported, a pneumothorax rate similar to what's reported in a previous series that I showed from Korea of around 20% with a variable chest tube rate, lower than that 20%, but still a non-diagnostic biopsy rate of anywhere from 11% to up to 18% with the more pure ground glass lesions. And so really to remember that biopsying subsolid nodules, first of all, there's the question of whether it's necessary given how likely these lesions are to be malignant when they persist, when they start evolving slowly, but also to recognize that a negative biopsy does not rule out malignancy, just like a negative PET does not rule out malignancy and has a negative, a limited role, and yeah, that you cannot determine with definitively the level of invasiveness from a small biopsy. The literature on bronchoscopy for subsolid nodules is even more limited. We know from earlier reviews of our data on bronchoscopy from earlier or older reviews of our data on bronchoscopy for peripheral lung nodules, certainly conventional bronchoscopy has a limited role for any peripheral nodules in particularly less than two centimeters. In the NAVIGATE trial, which was a large multicenter trial of electromagnetic navigation, a small proportion of patients had subsolid nodules that were targeted by bronchoscopy. The data was not reported specifically in that subgroup. Similarly, in the ACQUIRE registry, which was a prospective multicenter registry of a range, this is pre-robotics, but a range of advanced bronchoscopy techniques, specific data to the subsolid lesion group was not reported. So in fact, looking at more recent robotic bronchoscopy data certainly a couple of the recent systematic reviews that have looked at our data on advanced bronchoscopy, including robotics for peripheral lesions, they'll typically report diagnostic yield when it's pooled according to the more typical predictive factors of diagnosis, nodule size, presence or absence of a bronchus sign, but the data is not reported specifically on subsolid nodules. And I found one recent multicenter perspective series of, retrospective, sorry, series of robotic-assisted bronchoscopy. And again, one issue, this is the topic of another session tomorrow morning on some of the challenges with diagnostic yield in our bronchoscopy studies of the lung periphery. They did not report a strict definition of diagnostic yield. However, even with that non, more liberal definition of diagnostic yield, you could see that in pure ground glass nodule, the yield was lower, closer to 60% for subsolid lesions with robotic-assisted bronchoscopy. So in conclusion, PET has a limited role in the management of subsolid nodules, certainly no role for pure ground glass lesions, a possible role with solid component of eight to 10 millimeters, and a negative PET does not exclude malignancy. Biopsy as well has a limited role. The subtype cannot be determined. A negative TTNA cannot exclude malignancy. And there's a question of whether there could be a role prior to SBRT. I think even that is a question that's up in the air at this point, but that may be one case where a biopsy could be considered maybe point of discussion for the end of our session. Thank you for your attention. So my talk is entitled Immunotherapy, New Paradigms in Lung Cancer Treatment. I have no financial disclosures. So Dr. Jackman gave a couple of very well-delineated timelines. This is my first academic meeting in Boston since I did my internship here at Beth Israel before there was a Beth Israel deaconess over 30 years ago. So I was much more sentimental when I was preparing the slides for this talk. It was only natural to think about how things have changed since I was an intern. That's me over there, just a few miles from here. So back then, if somebody was diagnosed with lung cancer, the standard was to stage them via TNM. Sometimes we did a procedure called a media stenoscopy that you may have heard of. We didn't know where the PET scan was. And patients were referred for surgery, radiation, or chemotherapy. These are the ACCP treatment guidelines from 2003, and you can see that the main driver of referrals were stage, and pretty much stage alone. There was no individual characteristics that went into the consideration of referral, except for maybe some medical comorbidities such as COPD that might preclude surgery. This all began to change a few years later with reports that people with adenocarcinoma that had the EGFR mutation had improved overall and progression-free survival if they had an EGFR tyrosine kinase inhibitor. Since then, it's become standard at my institution and many other institutions throughout the country to genotype all new patients with lung cancer to look for markers that delineate so-called driver mutations that might be actionable with medications. And this has resulted, as we've just seen, with a slew of new medications that can take advantage of these driver mutations that didn't exist when I was an intern. This slide here just shows 19 medications, and there are many more. So I was tasked with talking about immunotherapy. First of all, what is immunotherapy? Immunotherapy takes advantage of the fact that the fate of all T-cells are not just contingent on binding to the T-cell receptor. It's contingent on the binding to other receptors concurrently. So for instance, if there's binding of the CD28 receptor, T-cells are activated. But in order to keep this system from running amok with extensive autoimmune disease and inflammation, there are so-called checkpoints that keep the T-cells from all becoming active. So in the lymph nodes, we have CTLA-4, and in the peripheral tissues, we have PD-1 that binds to the PD-L1 ligand, which is present in most somatic cells. And these inhibit the T-cells and promote energy. Most of us would recognize the medications that I've listed at the top. These inhibit the so-called checkpoints that I mentioned. So when you inhibit an inhibitory pathway, just like a double negative in grammar, you wind up with a net positive, or in this case, T-cell activation. Some of the earliest reports that this may be helpful in cancer was a study published in 2010 on ipilimumab and metastatic melanoma. Melanoma is a tumor that doesn't respond well to conventional chemotherapy. So for those in the patients received either ipilimumab or placebo, those in the placebo had a mean overall survival of six months. Those in the treatment group had a median overall survival of 10 months, nearly double. And that's just an improvement that's almost unheard of in the oncology literature, which is notoriously poor. These results were so exciting that the Nobel Prize in 2018 for medicine went to James Allison, who was at MSK and later MD Anderson, and to Suko Honjo, who was in Japan, for their discovery of cancer therapy by the inhibition of negative immune regulation. Dr. Allison was able to leverage his fame by joining a band called The Checkpoints, where he plays harmonica and backup vocals. So I would encourage anyone to fact check me on this, but I believe that this is the only example of a Nobel Prize winner playing in a classic rock cover band. So since then, there are really innumerable articles that talk about how immunotherapy has helped lung cancer and other malignancies, and we can easily talk about these for hours. But this talk is only supposed to be 10 to 15 minutes, so what I thought I would do is talk about three specific situations. Immunotherapy for perioperative, in the perioperative setting. Immunotherapy for unresectable disease. And immunotherapy for small cell lung cancer. And for each of these situations, I'd have a representative article that supports or refutes the use of immunotherapy. So the Checkmate 77T trial was a phase three study, looked at 461 patients with stage 2A to 3B disease, and these went on to surgical resection. So this is the perioperative setting. Patients received conventional chemotherapy, plus nivolumab or placebo. They went on to surgery, and then they had more nivolumab or placebo postoperatively. So at 18 months, the ven free survival was 70% in the treatment group versus 50% in the placebo group. Again, these are statistics that you just don't hear about in the oncology literature at 20% improvement. The Pacific trial looked at nivolumab after chemotherapy in stage 3 lung cancer. So these are people with unresectable disease. 709 patients with stage 3 unresectable non-small cell lung cancer underwent chemoradiation. 473 got nivolumab. 236 received placebo. There was an extended observation period which resulted in several publications, and there were at least three published in high-impact journals. So the 12-month progression-free survival in the treatment group was 55% versus 35% in the placebo group. Again, 20% improvement, almost unheard of. The observation period extended out for about seven years, and at the end of seven years, the median overall survival was 48 months in the treatment group versus three months in placebo. So those who received nivolumab with unresectable non-small cell lung cancer had essentially an 18% survival advantage, again, unheard of. Finally, there's a trial, the Adriatic trial of nivolumab for limited-stage small cell lung cancer. This is a brand-new study that was published less than a month ago in the New England Journal of Medicine. It was the September 13th edition. So as you may know, chemotherapy for small cell lung cancer, typically atopocyte and platinum, has remained largely unchanged for about 50 years. I spoke before about my internship. If you want to speak about the last time that there was a meaningful change in small cell lung cancer, you'd have to go back to when I was in D-care. Seven hundred and thirty patients underwent traditional chemoradiation and were then randomized to divalumab, placebo, or divalumab plus a medication, I apologize, I have trouble pronouncing tremolumab. Now the last group remains blinded, and we'll have to stay tuned to see the results. These are patients who received two checkpoint inhibitor medications. The primary endpoint for this study was overall and progression-free survival, looking at divalumab versus placebo. The median overall survival in the treatment group was 55 months versus 35 months in the placebo group. Again, a significant advantage. As you can see, the Kaplan-Meier graph changes almost immediately. So I think what we're going to see for the first time in 50 years is there's going to be a longstanding change in how we manage small cell lung cancer for the first time since the Nixon administration. Not everyone that's been studied for these immunotherapy drugs has to have cancer. I thought this was a very creative study. It's just an abstract published in the Journal of Thoracic Oncology last year, looking at divalumab as immunoprevention in patients who are thought to be at high risk for developing cancer in the future. So these are just 16 patients with endobronchial dysplasia seen on bronchoscopy who were given divalumab. Nine of these had an improved histologic score with treatment, and the activated T cells were more common in the responder group than those who did not respond. It's worthwhile pointing out that there are immunotherapy treatments besides checkpoint inhibitors that may be coming down the pike. Tumor infiltrating lymphocytes has had some success with melanoma, and it's now being studied at least in our institution with non-small cell lung cancer. So this occurs when T cells are isolated and expanded from the tumor tissue. It's thought that if activated T cells are in the tumor tissue, they may have something to do with actually fighting the tumor. Patients then undergo non-myeloablative lymphodepletion and get their T cells back along with high-dose IL-2. This is a preliminary study of 19 patients, and it's a waterfall plot, and you can see that most of the patients had some sort of improvement in the tumor size and extent. However, only six had a partial response as defined by the investigators before the study, and only one of 19 had a complete response. So this is clearly going to be a work in progress. So let me just mention again that when I was an intern here just a few miles from here, we really had surgery, chemotherapy, and radiation as the only treatments for lung cancer, and patients were referred for treatment based on stage alone. So now we have molecular treatments, and now we have immune treatments, and these are all operating on different planes. So how someone is treated by a conventional treatment with surgery, radiation, et cetera, might be different from how they were treated along with molecular, based on their molecular profile or their immune profile. And so you can get all sorts of different combinations of permutations of treatment based on individual characteristics of each patient. So lung cancer management now is just much more individualized than it was the last time I was in Boston. Most of us are pulmonologists. I think it's worthwhile mentioning toxicity. We're going to talk about this in some greater detail later, but this is what we're seeing now with Sloan Kettering, and it's going to be a major reason for consults, I think, for the rest of us in the future. This is a patient of mine with small cell lung cancer and interstitial lung disease, who the good news is he responded to immunotherapy. The bad news is he became so short of breath that he couldn't walk across the room. Eventually responded to prednisone and infliximab. The way I look at it is the way I look at superheroes, having seen a lot of Marvel movies with my son. If you make enough normal people into superheroes, some of those superheroes are going to turn rogue and start to destroy the village. And so it is with T-cells. So not only do you get pneumonitis, but you can get autoimmunity and numerous other medications in the heart with myocarditis, the pancreas. You can get diabetes. You can get thyroiditis and colitis, just to name some of the more common complications. If you do get diabetes from the pancreas, it's worthwhile pointing out that the changes are permanent in your insulin for life. So to summarize what I've said, we're entering an era when lung cancer patients may benefit from individual lung cancer care based on specific molecular and immunological profiles. It's not just surgery, radiation, and cytotoxic chemotherapy anymore. We can now offer individual therapy on several planes. Immunotherapy seems to be helpful under a variety of situations, and its uses may continue expanding. How it expands is limited only by our creativity and to watch for toxicities. So I welcome any questions, or we can save them until the end. So the next talk is called Immunotherapy from the Pulmonologist's Point of View, Panacea or Poison. So I'm going to tell you a little bit about the dark side of immunotherapy. So you can tell that this talk might be talking a little more about the poison part. So first of all, I do have a couple of ARS questions if you guys want to participate. So if you want to take a second to just open up your QR code here, give you guys a minute here. And then we'll go ahead and start with a little case presentation here. So while you guys are doing that, my name is Ajay Chaudhary. I'm at the MD Anderson Cancer Center. I'm a pulmonologist there. And we see a fair bit of checkpoint inhibitor toxicity. The first two talks talk about how these new therapies have really transformed the care of patients who get lung cancers, but it comes at a price. And that's a price that we pulmonologists need to be prepared to help deal with. So the lesson objectives here are to recognize the association between treatment intensity and the risk for pulmonary toxicity with immune checkpoint inhibitors. We'll talk about some high-risk scenarios. To understand the guidelines to how to treat checkpoint inhibitor pneumonitis and then talk a little bit about steroid refractory pneumonitis. There's been a couple of talks about this during the meeting, so we'll talk briefly about that as well. So here's a case presentation. A 68-year-old male with stage 4 non-small cell lung cancer. He undergoes treatment with carbaplatin, pemetrexid, and pembrolizumab. After the fourth dose of pembrolizumab, he developed severe shortness of breath and presents the emergency room. And immediately, he's put on high-flow nasal cannula, and he started on 2 milligrams per kilogram of methylprednisolone, about 160 milligrams per day. But he has not improved in 48 hours. So what would you do? So those are your options there. A is to increase methylprednisolone to 1,000 milligrams per day, so a pulse dose. B is to give him a dose of infliximab, 5 milligrams per kilogram, one dose. C is to give tocilizumab, 4 milligrams per kilogram, one dose. D is to start intravenous immunoglobulin, IVIG. And E is to start CELCEPT at 1,000 milligrams PID. So we'll go ahead and move to the next. We'll see the options here. Great. There's not really a right answer here. I just want to see what you guys are thinking. We'll talk about the level of data here. But it is curious to see where people start off. And I'll ask this question at the end of the talk as well. All right. So we'll talk a little bit about some of the new therapies. So we've already heard now about how dismal the treatment of lung cancer was just 20 years ago. This is a seminal study by Schiller and others looking at various platinum doublet therapies. And you can see, for patients with stage 4 non-small cell lung cancer, the mortality at one year was about 80%. It was pretty grim. We've already talked about how immunotherapy has transformed these outcomes. So let's talk about some high-risk scenarios for toxicity. The first trial I'll talk about has already been alluded to is the PACIFIC trial. So this is a trial where patients with locally advanced non-small cell lung cancer, so think stage 3A, stage 3B, could not undergo surgery. These patients underwent chemoradiation. And they were treated with either drivalumab, an immune checkpoint inhibitor, or placebo as adjuvant therapy after the chemoradiation. And you can see there's a pretty significant improvement with drivalumab therapy compared to just placebo. And so, I mean, the outcomes in general are better than they were 20 years ago. But despite that, there's an improvement with the use of drivalumab. Unfortunately, this comes at a cost. The use of drivalumab increased the risk of pneumonitis significantly. 34% of the patients developed some sort of pneumonitis with the immune checkpoint inhibitor adjuvant therapy, compared to 25% who got chemoradiation alone. And notably, about 3.5% developed grade 3 or 4 toxicities. And I believe there was one death as well, too. Let's talk about another study here. This is Keynote 799. This did not make it to clinical practice, but this is a phase 2 non-randomized study. And I just include it here because it really illustrates this concept. So there was two cohorts. Cohort A was squamous and non-squamous, and cohort B was non-squamous only. Both cohorts got pembrolizumab up front with chemo and then thoracic radiation. And then it was followed by pembrolizumab as adjuvant therapy. And the AE rate was quite high. And I'm just going to go ahead and jump to pneumonitis. So when you look at grade 3 pneumonitis, so these are patients that are extremely sick. They're requiring oxygen, usually in the hospital. The rate of grade 3 or higher pneumonitis was 8% in cohort A and about 8% in cohort B. And there was 4% deaths due to pembrolizumab or due to this treatment strategy in cohort A and one death in cohort B. So a highly toxic strategy. This was not something that made it to clinical practice. The third study just came out. And this is another really practice-changing study. So one of the things that's interesting about EGFR-positive tumors, they usually don't respond to immune checkpoint inhibitors. They're kind of immune-cold. And this has been known for a long time. So this was a study looking at locally-advanced lung cancers. But in this case, they were EGFR mutant lung cancers. And so they didn't get immune checkpoint inhibitor therapy, and so they got osomertinib, which is the EGFR inhibitor. And the study randomized osomertinib versus placebo in patients who had locally-advanced disease and got chemo radiation. So they got this adjuvant therapy. And there's a huge improvement in survival with osomertinib compared to placebo. But once again, there's an increased risk of pneumonitis. Now, this is not an immune checkpoint inhibitor. This is a tyrosine kinase inhibitor. But there is an increase of about 10% in the risk for pneumonitis. In this case, they lumped everybody as radiation pneumonitis. It's a little bit hard to tell what's due to the targeted therapy and what's due to radiation when you see this particular subset of patients. But once again, you can see that by putting these up front and adding therapies, you are adding to the toxicity. So what I want to highlight is that as we start to find better and better ways to treat lung cancers, as we start to combine these therapies, we do see better outcomes in some of these instances. But we're also seeing a higher rate of toxicities. So I'll talk a little bit about the guidelines for how do we treat pneumonitis. So in blue, I have some guidelines that I would personally treat as a little bit of a suggestion. And in the black are the more clinical guidelines. And I'll talk a little bit about why that is. But grade one patients are generally asymptomatic. They have limited involvement of the lung. Grade two patients, you start to see some symptoms. This is a pretty wide bin. And that is one problem with the clinical grade. You can be minimally symptomatic and be grade two. Or you can be so short of breath that you can barely walk, but not be hypoxemic. You would still be a grade two. And the same for grade three. You could be minimally symptomatic and require a little bit of oxygen. Or you could be on the verge of requiring a non-rebreather. Grade four is a little more homogenous. These are patients that are extremely sick. They require high flow or non-invasive ventilation or mechanical ventilation. And the grade five is fatal pneumonitis. So we'll talk a little bit about how to treat it. The reason I put the radiologic criteria in blue is that it's a little bit confusing. There's some discordance between the clinical and radiological criteria. And we have a little bit of data that's not quite ready to show you. But it does seem like the clinical criteria are the most pertinent, which is good for us, because that's a lot easier to go by. All right. How do we treat pneumonitis? So grade one pneumonitis is relatively easy. But you still got to keep your eye on these patients. These are patients who have some evidence of toxicity. You can consider holding the drug. But honestly, we often treat through these patients. We continue the ICIs. The important thing is to make sure that you're checking the CT to make sure it's stable or improving before you give the next dose of ICI. So the timing is crucial. It's helpful for us to be involved in these cases. I prefer to be involved in these cases. And if they get better, then you can watch them a little less closely. Grade two cases, this is where we absolutely should be involved. So I prefer to get involved as early as possible. In these cases, we hold the ICI treatment until the pneumonitis resolves. I advocate for bronchoscopy. We do see some opportunistic infections and pneumonias and other things. The radiology is not particularly specific. So you see certain patterns that can look like infection or pneumonitis. And so we tend to do the bronchoscopy up front. Can I say that it definitely changes management? I can't say that. But I'd rather err on the side of caution. These patients, we start prednisone at one to two milligrams per kilogram per day once we make the diagnosis of pneumonitis. And we try to taper over four to six weeks. It's a little faster than a sporadic organizing pneumonia. And the reason is we want to get these guys back on their immune checkpoint inhibitor. We don't want to, we want to minimize the gap that they're off. And that is associated with better outcomes. As the symptoms get worse, we treat it as a higher grade. And in blue, this is sort of a conditional grade, I think. If there's no improvement in 48 to 72 hours, there's a suggestion to treat it as grade three. But there's not great evidence to support that. These patients probably should be re-challenged with ICI. If they had a good response to ICI and you get, the pneumonitis resolves, you could try to get them back on their ICI if the oncologist agrees. Grade three cases, there's a little bit of discordance here. But generally speaking, these patients don't go back on their ICI. These patients require oxygen. They're usually in the hospital. You could monitor carefully as an outpatient, but most of these patients end up in the hospital. We still like to do bronchoscopy. We still like to give steroids. And we taper over one to two months. If they don't get better in 48 hours, you might think about infliximab or IVIG or another agent. We'll go through the choice of the agents towards the end of the talk. And I would say ICI re-challenge is generally not recommended. If you do it, it should be in consultation with the oncologist. There needs to be a really good case that they responded well and it's worth the risk of recurrent pneumonitis. There is emerging data on re-challenge. Many patients do tolerate the re-challenge. So it's something that I think hopefully we'll have more clarity on in the next couple years. Grade four cases, I think it's pretty much unanimous. These patients shouldn't be on ICI. Again, the risk is too high. These patients should be on two milligrams per kilogram methylprednisolone to start out with. If there's no improvement, you should start a second-line agent. And once again, we'll talk about what that second-line agent might be. A lot of times, we can't do the BAL here unless they're already on the ventilator. So how do we treat steroid refractory pneumonitis? So let's talk a little bit about what that is. So steroid refractory pneumonitis is usually in the first 48 hours, there's not a significant improvement on steroids in somebody who has a high-grade pneumonitis. There's also some other definitions like steroid-resistant pneumonitis where they require steroids for a longer period of time before they have improvement. There's things like steroid-dependent pneumonitis where you can't wean them off the steroids. That's not what I'm really talking about here. I'm talking about the folks who are really sick, not getting better on high-dose steroids. So what do we do in these cases? So this was a cohort from Memorial Sloan-Kettering. And in this case, there was 26 patients who had steroid-unresponsive pneumonitis. 19 of the patients were treated with infliximab. Six patients were treated with CellCept. Of those, four of them were four failures of TNF inhibition were also treated with MMF or cyclophosphamide. Sorry, that should be CellCept. And then if you look at the mortality, those patients who did not respond to steroids initially, 75% of them died. There's another study from Johns Hopkins against steroid refractory pneumonitis. They defined it as failure to improve after two to 14 days of corticosteroid therapy. So they had 65 patients with pneumonitis, and about 20% fell under the steroid refractory category. In this case, seven of them were treated with IVIG. And of them, four improved, two partially and two fully, and three died. Two were treated with infliximab and both died, and two were treated, three were treated with IVIG and sequential infliximab, and all three died. So unfortunately, there's a lot of confounding my indication here, but it doesn't look that great for infliximab. And this is a third study that just came out from Ohio State from Kevin Ho's group. And in this case, that's actually, the slide didn't update, unfortunately. That's from the previous slide. But in this case, they had about 50 patients that were steroid unresponsive. And of them, 10 were treated with either IVIG or infliximab. And three of the five who were treated with IVIG died, and five out of the five who were treated with infliximab died. So the IVIG group did do a little bit better, but you can see from the right that the number of patients that were under observation was quite small. So the evidence for how we treat steroid refractory pneumonitis is really lacking. And part of this is that we don't understand exactly what drives the biology of steroid refractory pneumonitis, or pneumonitis in general. We focus a lot on the T cells. There are some very small single-cell sequencing studies that talk a little bit about T cell biology. What we know is that, in general, there's a type one inflammation or perhaps a type one polarized Th17 cell-mediated inflammation. And there's a variety of cytokines, including tumor necrosis factor and IL-6. None of these are particularly specific for pneumonitis. And there's really not a great biomarker panel or single biomarker at this point. There are a few studies that are underway looking at this that distinguish pneumonitis from other conditions. The current practice patterns are all evidence-based, or expert guideline-based and not evidence-based. And really, the available evidence is insufficient at this point. Most of what we have to go off of is case series and case reports and that sort of thing. And I think we need to think a little bit bigger. A lot of this has been driven by single-center small studies. This is something that we need to look at multiple centers, different practice patterns, and maybe even think about other conditions, like ARDS or IPF, and see if we can apply some of the knowledge from those diseases to steroid refractory pneumonitis. So I'm going to go back to this case one more time, and I'm going to see how things change a little bit. So we have the thing open. I think it's going to pull up the QR code here. Let's see if anyone changes their opinion on what we should do here. So I'll give you guys just 30 seconds here, and then we'll go ahead and start to see how the responses go. Okay, not much change. You guys are convinced. That's good. I do think this is something that we need to do a little bit better with the data. These are patients that we're going to see a lot more of over the years, and I appreciate everybody's time. Thank you again. Thank you.

multifocal lung cancer

Dr. Faisal Al-Ghula

Dr. Frank Ditterbeck

thoracic surgery

adenocarcinoma

biomarker discordance

imaging strategies

clinical staging

subsolid lung nodules