® focused on advanced lung cancer. In today’s presentations my colleagues and I will review three clinical cases. We will discuss an individualized approach to treatment for each patient, and we’ll review key clinical trial data that impact our decisions. I’m Dr. Mark Socinski from the AdventHealth cancer institute in Orlando, Florida. Today I’m joined by Dr Ed Kim, a medical oncologist from the Levine Cancer Institute in Charlotte, North Carolina; Dr Brendon Stiles, who is a thoracic surgeon from the Weill Cornell Medical Center in New York ; and Dr Tim Kruser, radiation oncologist from Northwestern Medicine Feinberg School of Medicine in Chicago. Thank you all for joining me today. We’re going to move to the first case, which is a case of small cell lung cancer. I’m going to ask Dr Kim to do the presentation.

 Thanks, Mark. It’s my pleasure to walk us through the first case, which is small cell lung cancer. This is a case with a 72-year-old woman who presents with shortness of breath, a productive cough, chest pain, some fatigue, anorexia, a recent 18-pound weight loss, and a history of hypertension. She is a schoolteacher and has a 45-pack-a-year smoking history; she is currently a smoker. She is married, has 2 kids, and has a grandchild on the way. On physical exam she had some dullness to percussion with some decreased-breath sounds, and the chest x-ray shows a left hilar mass and a 5.4-cm left upper-lobe mass. CT scan reveals a hilar mass with a bilateral mediastinal extension. Negative for distant metastatic disease. PET scan shows activity in the left upper-lobe mass with supraclavicular nodal areas and liver lesions, and there are no metastases in the brain on MRI. The interventional radiographic test biopsy for liver reveals small cell, and her PS is 1. Right now we do have a patient who has extensive-stage small cell lung cancer. Unfortunately, it’s what we found. It’s very common to see this with liver metastases.

Case 1: 72-Year-Old Woman With Small Cell Lung Cancer

 Ed, is any additional work-up needed to help you inform any treatment recommendations here, or is she pretty well worked up at this point?

 When you see these folks in clinic, sometimes with small cell cancer patients, they can be very sick, but her labs I assume are OK. She had a biopsy. It looks like it’s sensitive stage to the liver. Brain is clear, so I think I’ve got all the staging info I need with path.

 We’ve seen really a change in the standard of care. What would be your approach in treating this patient?

 As you know, you and I have been doing this for long time. We kept complaining every year about targeted therapy for small cell lung cancer. Platinum-based doublets with cisplatin-etoposide, or carboplatin-etoposide, have really been the standard, and that’s what many have used. There was a period of time where we considered some drugs, but certainly it looks like cisplatin-etoposide. Now, we’ve had the addition of immunotherapy for the frontline setting. I would begin talking to her about a combination of chemotherapy and immunotherapy.

 How does that discussion go in your clinic with regard to extensive-stage small cell lung cancer? Historically we’ve considered this a treatable disease, but a disease that although most patients respond, you know the response is not very long lasting. If you look at progression-free landmarks at a year, it’s like 5%, 6% or so. How does that conversation go, and how has immunotherapy changed that conversation?

 People are excited about immunotherapy. There aren’t as many commercials on as there used to be, but they’ve hit everybody. Small cell is tough; they don’t like hearing that diagnosis. They start doing a search about small cell and they see that it’s aggressive, that it’s harder to treat, that it can be very debilitating. I always find it a little tougher when patients are diagnosed with small cell, especially extensive-stage small cell. There is at least hope that we have immunotherapy treatment that has improved outcomes compared with historical treatments. That’s the mixed bag that I experienced.

 I’m going to ask Dr Stiles a question about the role that surgery may play in this disease. You know my training has always been, and we actually showed in the case presentation the TNM staging system. I was trained many, many years ago that you know the management of small cell may not necessarily be different from that of non–small cell if you just do it by a TNM staging. For instance, if you knew there was a T1N0 small cell, you know outcomes with surgical resection and postoperative treatment seemed to be better than we would anticipate from chemotherapy or radiotherapy.

Brendon, your perspective on this from a surgeon’s point of view? How often do you see it? What are your thoughts around it? I’d be interested in your comments.

 It’s pretty rare. I’ve definitely taken up a handful of T1A small cells, which I never quite felt good about and never really knew what to do afterward. In cases like this, I would shy away, and even in cases with mediastinal adenopathy I tend to shy away. It’s rare to find that single lesion without any lymph nodes. In that case I do feel like local therapy is appropriate and follows NCCN Guidelines. Certainly in this case I would stay far away.

Case 1: Extensive-Stage Small Cell Lung Cancer Background

 These are the NCCN Guidelines. You can see here that we’ve got several category 1 recommendations. You see that we’ve got combinations of chemotherapy with atezolizumab-durvalumab, with both carboplatin and cisplatin. Then the others are more of the older regimens, the backbone of platinum plus etoposide and then platinum plus irinotecan, which got a lot of momentum in the Japan but not as much in the United States, and probably deservedly so. This is IMpower133. This was a carboplatin-etoposide with or without atezolizumab for extensive-stage small cell lung cancer. Atezolizumab is an immunotherapy that is used in several other areas. The trial design was here with a primary end point of overall survival and PFS. It was the triplet versus the doublet with a placebo-controlled aspect. You receive 4 cycles of therapy, and then either had placebo or atezolizumab as maintenance therapy until follow-up progression. There was a PCI that was allowed but it was based on the standard of care that were done. You can see there was no biomarker really assessed as part of the eligibility in the treatment here. These are the survival curves, PFS and OS. As you can see, the median survival was improved by 2 months, 12.3 vs 10.3, with a 1-year survival of over 50%. Those are better than historical controls and were statistically significant. Then there was an updated overall survival close to 23 months, and you can see that effect has continued with a hazard ratio, a 

 value, and an 18-month survival of 34%. Here, we can see the IMpower133 response rates with atezolizumab. You can see that although we do have usually high responses in small cell with chemotherapy, both of these performed well in the 60% range. There was also stable disease that was very similar, but again, we did see an effect with the atezolizumab group with the added immunotherapy.

Chemotherapy-related adverse effects and immunotherapy-related adverse effects are things that we’re always cognizant of. Again, when you look at atezolizumab you’re going to expect to see more of these I/O related events. The overall grade 3/4 toxicities were very low, and that’s been our experience: We do see some severe toxicities but not as many. For instance, my nurse practitioner worked in melanoma for a couple of years and came over to work with me. She was really sacred when we started treating patients with I/O and weren’t doing UAs every week and getting thyroid function tests every week. I said, “No, that’s the way we did them in lung.” She said, “Are you sure?” You know, because she was coming from that melanoma area. Sure enough, that’s how it was, and we see the adverse effects can vary among tumor types with these I/Os.

Case 1: IMpower133 Trial in Small Cell Lung Cancer

 Another study that’s been reported is CASPIAN. This is a platinum-etoposide doublet, and what was unique about this study is it was 3 different arms. You had a combination of durvalumab and tremelimumab—so it’s CTLA4 with the PD-L1, with chemotherapy. That was given for 4 cycles, so the 4-drug regimen, followed by a maintenance durvalumab schedule every 4 weeks. That’s different from every 3 weeks that was given with the concurrent. Then you had the triplet; again, every 3 weeks for 4 cycles, and then the every-4-week maintenance. Then you had the control arm, which was the chemotherapy given up to 6 cycles. You could give more chemotherapy therapy there and again get optional PCI. The primary end point here was again overall. It was a randomized trial with very interesting results. No PD-L1 or anything that you can see was done ahead of time. The survival curve was 12.9 versus 10.5 months. That’s a little more than 2 months that was shown; very similar numbers. As you can see, the control arm did very well, and that’s the thing that’s been highlighted across all these trials. The control arm has behaved as we would expect to see in our daily practice. The PFS also was a little—you could see the curves overlap and then separate, so there was some link separation there. That raised some eyebrows from some folks. But it was still beneficial and clearly small cell is a tough disease early on. Maybe there’s an effective maintenance here or maybe it was the fact that you were able to take 6 cycles of chemotherapy in a control arm. There are many reasons, but the comparison here is that the durvalumab with chemotherapy is what showed the benefit. There was not the same benefit by adding the additional CTLA4 with tremelimumab, that 4-drug regimen, so it was a 3-drug regimen in the comparison that was the best. As you see, updated data, this duration of response here. The response rate is very high. There was certainly an additive effect with durvalumab with the chemotherapy regimen. The brain metastases are very interesting. We know this is a tough area with small patients with small cell, and there was a benefit throughout. They allowed patients on the study with asymptomatic brain metastases, and they showed a benefit in OS as you can see in the Figure 3, Panel A and B, with the addition of durvalumab. There was an effect that they were able to demonstrate.

The adverse effects—it was very well tolerated. You’ll see this as a recurring theme across grade 3 and 4. We’re very good about looking for these adverse effects that can occur both early and later in the course of therapy.

Then we have everyone’s popular drug, pembrolizumab. This is the KEYNOTE-604 study: platinum-etoposide with the combination. Everybody expected this to be something that was going to hold out as positive. Here was the early event, but it missed its cutoff. It was just barely, not significant. It was close but not statistically significant. Although the PFS early on for this study was pretty impressive, the overall survival did not hit. So pembrolizumab, as you saw the NCCN Guidelines is not FDA approved for small cell.

 What about the issues of patient quality of life? There was some patient-reported outcome [PRO] data from the CASPIAN trial. Can you comment on that?

 Yeah. PROs are a very important area, and so is quality of life. We’re seeing that immunotherapy is very well tolerated and that, in fact, we see the benefits of adding this drug. What we see is not necessarily an improvement, but we see a lack of deterioration in quality of life in PROs. If we can hold the line and not make people feel sick but also add another drug on top of things that are going to improve efficacy, then that’s a win.

 The older literature you can go way back to in the early days of treatment, and there were actually a number of studies that suggested that you do palliate symptoms when you shrink a tumor. I would prefer for quality of life and PROs to tell us how many times they got better vs worse later. It’s interesting that that’s the path that they took in the CASPIAN trial.

 Yeah, and the age increase can always be a little clearer. It was 1 of those mandates where they said, “OK, we want to see quality-of-life data, so you should measure it.” Then they’ve been very slow to approve drugs based on that indication.

Case 1: CASPIAN Trial in Extensive-Stage Small Cell Lung Cancer

 What about biomarkers in small cell lung cancer in terms of PD-L1? We hear a lot about TMB. Is it relevant in small cell? You did mention that it was certainly not a selection criterion for any of the small cell trials. But is it helpful?

 If we were all in the room 5 years ago discussing this, we’d probably all vote for TMB knowing that in small cell, there’s a lot of noise and you would expect those patients who smoked to have more dirty tumors. However, if you believe either of those, we still struggle even in the non–small cell lung cancer realm with these biomarkers for immunotherapy. Maybe that will be different when we start looking at things like STK11 or keep with resistance certainly KRAS and others have shown some prediction. In small cell, as we pointed out, you don’t have to worry about testing right now. We don’t have a good marker, and neither of these markers have really shown any promise in helping to direct which patient population should be treated based on that.

Here is the new therapy biomarker. We can see on top that there’s a big question mark. IMpower133 didn’t really reveal anything. As you can see, there are some favorable aspects but nothing really predictive to see whether you’re looking at the blood-based tumor mutational burden or PD-L1 expression. Nothing really stands out as a true marker. Then there is overall survival based on the PD-L1 expression. It’s a little all over the place—nothing really pointed out that would lead us in a direction.

Certainly, from maintenance therapy there was really nothing that pointed out to us that by adding the nivolumab-ipilimumab in CheckMate 451. We think we know that maintenance can help, but not all the drugs necessarily translate to that. Even giving first-line chemotherapy and then having a maintenance strategy really didn’t reveal any predictive aspects to this treatment.

 Ed, let me ask you a little about that maintenance question. Were you surprised by that result that maintenance did not show something here? To be honest with you, I was surprised.

 I was, part of what we believe in some of these currently approved trials is the fact that the maintenance approach afterward is what held the stability of disease or kept it at bay. It’s interesting. The more we study this I/O, even though we classified them as very similar from the beginning, they do behave differently in these trials. Some of it you think may be is random, but some of it also seems to be biology. Ipilimumab-nivolumab worked well in several tumor types, whereas tremelimumab did not work as well even in melanoma. Even from a setting of maintenance therapy, gosh, it would be interesting to see if it was given concurrently, for instance, with nivolumab in this trial, up front, and then had a doublet maintenance. That would be an interesting question to see. But maybe it was just bad luck, or maybe it really doesn’t work.

Case 1: Biomarkers in Small Cell Lung Cancer

 At ASCO you showed KEYNOTE-604. The other presentation was an ECOG phase 2 trial grafting nivolumab on to the same platform of platinum-etoposide. I walked away from ASCO saying, “We really do have a new standard of care.” I know technically KEYNOTE-604 was a negative trial, but it was almost the same effect adding pembrolizumab. Obviously, it won’t gain FDA approval, nor does it have to gain FDA approval in this setting. We have 2 good choices here. In your mind, is durvalumab-atezolizumab like the Coke-Pepsi argument? Are they the same? Do you view them any differently? Which do you use in your practice or your pathway? Comment on that.

 I agree with you first about the pembrolizumab data. It was close, and many of us thought it was going to be positive. And it’s surprising—that’s why you play the game, right? You finish strong. We’ve had both of these regimens on our small cell extensive-stage pathway. We consider them very similar: The profiles are very similar as well. I’ve leaned more toward durvalumab. If you’re splitting hairs, a little more response and a little more OS, but very similar. So, we allow both regimens on our pathways.

 It’s hard for me to ask this question, but you know the atezolizumab-durvalumab is a little more interesting than the carboplatin-versus-cisplatin argument. I think we have plenty of phase 3 data, at least in extensive-stage disease. Let’s say they’re equivalent in that setting. Your thoughts on that in the era of immunotherapy? Most of us have always felt that maybe cisplatin is a little more active, but it’s hard to prove it in this setting. Maybe with immunotherapy, do you have any sense of that or is it just Coke-Pepsi again for the cytotoxic?

 I try to lean toward cisplatin when I can. In the extensive-stage standpoint, sometimes you see patients who are a little sicker. I don’t know if it works any better, especially in an incurable situation, and immunotherapy may be an equalizer for them. So yeah, I agree. It’s less relevant for me in the extensive-stage setting. If we get data down the road in the early stage setting or the local setting, then certainly we lean toward the cisplatin-based therapies.

 What’s your sense? I was going to ask you this early on when you were discussing the NCCN Guidelines, because they still list platinum- etoposide as a standard or an option. Small cell lung cancer is a population of patients who tend to have a high incidence of paraneoplastic syndromes. Many of these, particularly the neurological ones, may be immune related. I guess the question is, what are your thoughts there? How often do you see a patient where you don’t want to use immunotherapy for 1 reason or another in this population? Give us your thoughts about that.

 It’s rare that I find a patient I can’t or don’t want to treat with immunotherapy. Early on we were very conservative in who we approached. If they had even a hint of history of rheumatoid arthritis or anything like that, we got a little skittish. We’ve seen that these are very well-tolerated drugs, so there must be a pretty high bar to really not give the effective therapy. Like COVID-19, we’re still learning. Everybody said, “Don’t use steroids in COVID-19.” It was interesting. Melanoma doctors were really advocating for it early on because of cytokine storm and other things. ID docs kept saying “No, no, no.” Now it’s changed. Everybody is learning. We’re learning about these immune responses, and certainly with paraneoplastic syndromes we’ll get more experience with that for the treatment of patients.

 Let me bring Dr Kruser into the discussion. We’ve had an up-and-down experience with PCI in the extensive-stage disease setting. Should patients with extensive-stage disease get PCI following therapy? It was an option at least for the control arm patients in the CASPIAN trial. I’m not sure about IMpower133. Tim, what are your thoughts there?

 As you know that was the standard of care for a long time, to use PCI in these patients based on some older data before MRI surveillance became an option. I currently do not use it, and I based that decision on the Japanese data, the randomized study that was performed when they did incorporate MRI surveillance. The survival in patients who did not receive the PCI was superior in the Japanese cohorts. Additionally, we know it’s a toxic therapy to some degree, and fundamentally I don’t like using toxic adjuvant radiation in people who are not going to be curable. It never really made sense to me as a paradigm, and I’m glad we have data that support not using it. Whether that pendulum will shift back or not now that we have hippocampal avoidance and memantine remains to be seen. There’s that ongoing SWOG randomized study that is looking at this question with more up-to-date techniques. It’s still an open area, but I keep leaning away from it.

 To follow up on that, there were some old data suggesting that you could get some mileage out of consolidating the mediastinum with thoracic radiotherapy. What are your thoughts about? Are there any patients you would consider giving thoracic radiation therapy to in the extensive-stage setting?

 Yes. That data I feel much better about. That was a current cohort. It was pre-immunotherapy, but it was the publication was in 2014, and the bang for the buck there was very good. It bumped the 2-year-overall survival from 3% to 13% with essentially no increased toxicities. It’s 10 fractions of outpatient radiation, with no grade 3/4 toxicity bump, so I do like to use that. Subgroup analyses from that randomized study have showed that people who have less than 3 extra thoracic sites of disease tend to have a better prognosis. It makes more sense to use it in people who have good response and a relatively modest burden of extra thoracic disease. I do have advocate engaging in a conversation with my patients.

 Yeah. I guess as we wind down case No. 1, I’d like to get Ed Kim’s opinion and then your opinion. If you look at the new standard of care and extensive-stage disease using these agents in addition to chemotherapy, and then you look at the PACIFIC data in non–small cell lung cancer, it seems to me that limited-stage small cell lung cancer is a perfect target for this approach. Ed, your thoughts on a PACIFIC-type approach? I’m actually blanking on the body of water that we have for the name for the limited-stage small cell trial off the top of my head. But I’m sure you remember it, Ed.

 It’s 1 of those. I agree with Tim’s discussion before. I believe the consolidation to the thoracic disease after chemotherapy and immunotherapy doesn’t change that for me. I try to give any type of radiation, other than whole brain, to someone who’s taking immunotherapy. I still believe in that abscopal effect—or that there’s an enhancement, so I like that. Even in my non–small cell patients, I like to give them a little bit of a radiation here and there to treat maybe a site of disease that’s growing out of control vs others. It makes logical sense, especially in light of the PACIFIC data, which surprised a lot of us. That was 1 that clearly came out of left field. If you ranked that as a trial, that ahead on the radar with all the immunotherapies that were being proposed and the different trials, I don’t think there would be many of us who listed that as our top choice as far as where durvalumab especially was going to make it in light of other studies combining concurrent regimens or consolidative regimens. It makes a lot of sense. Durvalumab is now showing that it has some legs behind it and has some ability to add efficacy, and it’s the logical next step. I hope it’s a positive study that will really help move forward our treatment with limited-stage small cell.

 I do hope there will be data for immunotherapy in limited-stage small cell lung cancer. It’s actually the focus of an ongoing randomized study, NRG-LU005, which is looking at both concurrent and maintenance immunotherapy in the limited-stage small cell lung cancer setting with concurrent chemoradiotherapy. Hopefully that is a paradigm I will use for patients moving forward.

Case 1: Small Cell Lung Cancer in the Era of Immunotherapy

 Let’s move on to case 2, this is a case of a patient with 

 mutation-positive non–small cell lung cancer. It’s a 67-year-old woman who presents to her primary care physician complaining of visual disturbances, fatigue, and sporadic headaches. She does have a history of hypertension, hyperlipidemia, is a former smoker, pretty impressive 55-pack-year history. You can see her vital signs there. She does have some decreased breath sounds at the left lower lung base. Otherwise her examination is negative. Her CBC and chemistries are normal.

Her brain imaging demonstrates a 1-cm right parietal mass at the gray-white junction without significant vasogenic edema. Chest CT, abdomen, and pelvis shows a 3.4-cm mass in the left lower lobe and several small liver nodules there. She has a CT-guided transthoracic needle biopsy of the lung lesion showing a grade 2 adenocarcinoma acinar subtype, so she stages a T2aN0M1c. Her ECOG performance status is said to be 1. The initial laboratory results we get back show that she does have a higher expression of PD-L1, 70%, done by the 22C3 test, and the patient is anxious to start therapy, as most of them are. I guess the first discussion that I would have here, and I’ll let Dr. Kim to address this. What would you recommend for molecular testing? Then, what do you do with this PD-L1 status before you know the results of the molecular test?

: Thanks, Mark. I still remember, just to digress, the first day when I was at Anderson Cancer Center, we converted from paper memos in the boxes to email, and it was a requirement that you had to check your email to get messages, and if you missed a meeting because you didn’t check your email, you were at fault. That’s the way you have to initiate change. As a medical oncologist out there you have to do these biomarkers and you must do all of them, and not just part of them. You can’t check your email once a week, it’s just not like that. The problem is out there, there’s no accountability to check on that, and we hear all kinds of excuses. For good or for worse, immunotherapy has been a great treatment initiative for many patients, but it doesn’t allow you to skirt the biomarkers.

The fact that we, even since May, we’ve had 7 different indications for lung cancer including 2 brand new biomarkers, 

 exon 14. You have to do these markers, you have to absolutely do them. You wait for the results, and then you make a decision on treatment. That’s like starting treatment without the biopsy results coming back, however you want to relate it. It’s one of the big no-no’s right now in treatment of non–small cell lung cancer to start therapy without having the molecular results. I ask people, I see this all the time when people are starting chemotherapy. When did chemotherapy get so good that that was going to be the therapy that we need to start? I don’t know what changed.

Case 2: 67-Year-Old Woman With EGFR+ Non–Small Cell Lung Cancer

 There's obviously a growing list of actionable biomarkers with FDA-approved therapies. I’ve been saying for quite some time now that I think one of the greatest things we do as medical oncologists is diagnosing an oncogenic driver where we have an FDA-approved agent. These agents work so much better than chemotherapy. We know that from the 

 space, and this is not a group that historically we think gets a great benefit from immunotherapy. Most of the oncogenic drivers were essentially excluded from the vast majority of trials of chemoimmunotherapy, except for a few like IMpower150. I actually am pretty aggressive about testing. I don’t know what Ed does, but in my practice, I do both tissue and plasma at the time of diagnosis. We had the NILE trial, which suggested that you can find things in plasma you don’t find in tissue. If you live by the credo of never miss an oncogenic driver, I think plasma plays an important part of your initial comprehensive molecular evaluation.

 I will say, Mark, that is progressive and I agree with you. I think that’s the general direction it’s going for sure. We collect both, we don’t run them simultaneously. With the Tempus testing platform you collect both, and then we run the reflex if the tissue doesn’t come back. I agree, I think doing both of them is definitely progressive and helpful.

 Yes, and many years ago, addressing the second issue here about the smoking history, I think Vince Miller, when he was still at Memorial Sloan-Kettering Cancer Center actually presented some data at ASCO. He said that if you restricted your 

 mutation testing to never-smokers, you would miss 60% of the 

 mutations. So there really are no clinical features, it should be the histologic diagnosis that should force you to do the comprehensive testing. You can’t say in an adenocarcinoma, this case here, this woman has a 55-pack-year history of smoking. You’re still going to test her, right?

 Yes. We do the same, we base it more on the histology. You can understand the way old school sort of held on because the way we even got to these biomarkers was because of clinical history, the Asian, never-smokers with adenocarcinoma, and that was always the last part. So I get it, people like to hold on to the oldies but goodies, but no, you need to just test.

 What’s the specific message you would give to the audience now, knowing that she’s strongly PD-L1–positive, but we don’t know any of her molecular genomics yet? What would your message be?

 I’m happy we got something positive. Let’s hope we get something else off the other panel because that will make me even happier. We always go with a driver mutation, and we have to be cautious when we have 

 mutations present, for instance, giving immunotherapy because some of the data have not been favorable and in fact been harmful.

 Yes, and this summarizes the data here. Again, to Ed’s point, Eddie Garon really led the charge in the initial observation of taking patients who were 

 mutation-positive and had high PD-L1 expression, giving them pembrolizumab in the first-line setting, kind of went 0 for 10. There was 1 patient who had a response, but that patient was later found to be 

 mutation negative. When they all progressed, they rolled over to osimertinib, and there were 2 deaths within 6 months of enrollment, including 1 attributed to pneumonitis. So the take-home message is you should have concern. There have also been some data in the 

 space with the same concerns; in that case, a couple of issues with hepatitis.

Really, the take-home message is make sure you have all the information you need to make the decision. Obviously, molecular testing may take a little bit longer, but it’s important to wait for those results. This slide again brings home the point that I just made. Particularly with osimertinib, you can see that that sequence of PD-1, PD-L1 blockade followed by osimertinib does carry with it the risk of severe immune-related adverse events. You don’t want to put your patient in that position by pulling the trigger on immunotherapy too quickly before you know the results of the genomics in this testing. It’s been incorporated into the NCCN recommendations, as well as the SITC recommendations shown here. So it’s an important point to make with regard to managing this population.

Case 2: Biomarker Testing for Non–Small Cell Lung Cancer

 I haven’t actually been in the situation where a patient has come to me for an opinion already on immunotherapy and subsequently was found to have an 

 mutation. I don’t know if that’s happened in your practice.

 Fortunately, it’s more of the situation we discussed where someone started chemotherapy or even I/O and they haven’t been tested yet. We have found actually some other alterations molecularly, but luckily not 

 I did have a discussion earlier this week with one of our colleagues who had actually 2 cases that she was struggling with regard to this thing. Patient was on pembrolizumab, they subsequently found an 

 mutation, and she was struggling with what to do and how to do it and was having some issues with that. I hope this is going away, but how often do you get insufficient reports in your practice?

 It’s rare now. We do a check to make sure. The worst is when you’ve already sent it and you get the QNS back, and that just wasted 2 weeks of everyone’s time. I think more practices are being diligent about looking at the amount of tumor that’s within the specimen, so that it’s adequate. I think it’s happening less and less.

 I would agree with that. I had my first patient in a while this week. I don’t want to offend anybody, but one of the most useless diagnostic procedures, in my opinion, often in the work-up of patients with lung cancer is a bronchoscopy. I had one this week that had the diagnosis of adenocarcinoma but QNS to do either PD-L1 or molecular testing. I saw the patient, it was interesting. He said, “What do you think this bump on my head is?” and on MRI it looks like a classic subcutaneous metastasis, and so we’re going to biopsy that. We had accessible tissue there, but we’ve got to make sure that we’re getting things up front efficiently for this population. How often do you have to do a second biopsy? I assume if you’re not getting much QNS, you’re getting stuff up front?

 Yes. We rarely get second biopsies. I think many of us would think about, if we have a diagnosis, then we would think about doing liquid. If we didn’t have a diagnosis, then another repeat biopsy would be done, and I’d have to scold one of my interventional folks.

 I think it’s important to let the audience know that we throw this term around, 

 mutation, and what we’re mostly talking about are the exon 19 deletions and the exon 20 L858R, but there are others, and I think it’s important to realize that some of these are not sensitivity mutations. The biggest suspect there is the exon 20 insertions, which we really don’t have a good drug yet, at least not an FDA-approved drug at this point. It’s important to realize that all 

 mutations are not created equal, and we do have some uncommon mutations in which you may think about a different drug in this setting.

So this patient has a molecular testing panel, and importantly she is found to have an 

 exon 19 deletion mutation, but we don’t know the time course. So here we have exon 19 deletion in the setting of a high PD-L1 expression. What are your thoughts in terms of therapeutic options for this patient, Ed?

 These days, it’s osimertinib. Get it started. Hopefully there’s not an issue with getting the drug or copay, it’s less of that these days, and get them started.

 When osimertinib was trying to break out of the second-line T790M space, there seemed to be an appetite for this issue of sequencing—give a first-generation or second-generation drug first, and then when T790M rears its ugly head, use osimertinib at that time. There were all these arguments about PFS1, PFS2, would that get you further than using osimertinib up front? I don’t think we’re having that argument anymore.

 Yes, I’m glad. Although baseball is limited right now, the starting pitchers, you use them as your starters on the frontline, and when they are not being so effective, you move them to the bullpen. Eckersley is a classic example of that, you don’t want to put your best pitchers in the bullpen. You want to put good pitchers because they have to finish strong, but you’ve got to start off as best as you can.

 What are your thoughts, we had a recent approval of an anti-VEGF drug in this space with erlotinib/ramucirumab based on the RELAY data. We’ll talk about the FLAURA data in a moment, but has that changed your view at all?

 As you know, there were some inklings of these data and other combinations that had come out about a year ago at some national trials. I never understood, again, that concept of adding an extra drug to get a similar result. There are adverse effects that occur with these drugs. Ramucirumab is much better than bevacizumab, but I don’t know how that helps the patient, especially taking an oral drug, without an IV, that has great effectiveness, has great CNS effectiveness, a great adverse effect profile. It never made a lot of sense to me, but kudos for people doing research and finding a benefit there.

 mutated disease is a very VEGF-driven disease, I like the description that Jared Weiss from the University of North Carolina School of Medicine always says: it really medicalizes your treatment. You’re coming for an infusion every 2, 3 weeks, whatever the schedule is, and it doesn’t necessarily get you any further mileage from the data that we have in the absence of a head-to-head trial.

Case 2: EGFR-Positive Non–Small Cell Lung Cancer

 Let’s ask Tim to comment on the 1-cm lesion we found in the CNS in this patient, and we’ve all agreed that we’re going to start osimertinib here. What are your thoughts as a radiation oncologist about that finding, and what would your recommendation be knowing she has an exon 19 deletion mutation?

 We run into this scenario a lot. In this particular case presentation, fortunately it seems the patient was not symptomatic and got the test results pretty quickly. If we knew that the patient was neurologically intact and this was not in an eloquent area, I would be on board for starting drug therapy and watching this. There was a nice secondary presentation of the FLAURA results for brain metastases specifically, and 90% response rate, and maybe more importantly, no actual progression events. So you don’t have to worry about the 1-cm lesion blossoming into a neurologic deficit or requiring a craniotomy. That said, it’s more challenging as we framed the case originally when people have this finding at diagnosis and it takes a couple of weeks. Very commonly, if someone has something that’s radiosurgery-friendly and we’re waiting for those mutation results, we might use radiosurgery up front while we wait to decide on the systemic therapy, especially if it’s relatively sizeable, and 1 cm is getting toward being relatively sizeable. So it’s a case-by-case decision.

 Yes, I like to say you have little room for error when you’re dealing with CNS disease. If you start people on drug therapy with known CNS disease, how soon do you repeat the MRI to make sure they’re responding?

 I like to see these people and get them in my clinic and tell them along with the medical oncologist here, I treat a lot of lung as well as CNS disease. I think having a neurosurgeon or a radiation oncologist who specializes in brain management is good in terms of localizing these and their eloquence. Typically to your question, 6 weeks would be a reasonable first MRI timeframe, and then you can space them out pretty aggressively. We know that time to progression for these post osimertinib is 18 months or so. Every 3, 6 months I think is reasonable for surveillance.

 Great. Let’s talk a little bit about some data we’ve been alluding to. Here’s the design of the FLAURA trial. This specifically took exon 19 and 21 

 mutations, good performance status patients, stable; CNS metastases were permitted in the trial. The randomization was to dealer’s choice on the control arm of one of the first-generation drugs, erlotinib or gefitinib, and the investigational arm was osimertinib, 80 mg daily. The primary end point was progression-free survival. You can see the secondary end points are included on the bottom of the slide. The stratification factors were nature of the mutation as well as race, Asian versus non-Asian. Any way you want to see, the investigational arm wins right out of the gate, these curves separate early. The control arm of erlotinib/gefitinib, median PFS of 10 months is exactly what we’ve seen in multiple other trials involving these agents, so the control arm performed as we expected. We see almost a doubling, up to 18.9 months, for osimertinib, hazard ratio of 0.46. So a slightly more than 50% reduction in the risk of death or progression over time. Not only statistically, but I think clinically significant in this population.

Ed, I’ll ask you to comment on this slide. I think we’re all waiting for these data, thinking that many of the control arm patients may cross over with T790M to second-line osimertinib. I don’t know that we have the final answer on that yet, but there was a clear survival advantage seen in this trial here with a hazard ratio of 0.8, which is certainly in the range of what we’ve accepted to change therapy in the survival setting. But your thoughts, were you surprised at the survival advantage or not?

 I have to say more relieved, Mark, because I’m a big believer in response and PFS. I think in the metastatic setting, those add value. I think we as scientists and researchers get very hung up on overall survival, especially again in a metastatic setting. I think it’s a little different there, especially with crossover, and certainly I’m glad we’re not going through the exercise of talking about these curves, saying, “Oh, it was probably crossover that confounded the data, etc.”

 I’m just relieved that now this chapter is closed, everybody can see this is the best therapy, and we can move on.

 We alluded to the ramucirumab data, and I completely agree with you, I think the FLAURA data established osimertinib as the undisputed world champion of TKIs [tyrosine kinase inhibitors] in the 

Case 2: FLAURA Study for EGFR+ Metastatic NSCLC

 We alluded to the RELAY trial before, which took a first-generation drug, erlotinib, again in the same population that we saw in FLAURA, added ramucirumab, in this case on an every-2-week treatment.

Another key feature of this is that brain metastases were excluded, were a key exclusion criteria. Again, the primary end point was progression-free survival. You see here the progression-free survival with a hazard ratio of 0.59, so not quite as robust as we saw with the FLAURA data. I don’t think the toxicity shown in the table is surprising to anybody given the use of an anti-VEGF agent. We saw similar results with bevacizumab from one of the Japanese trials, and you see the same effect on PFS as well as the safety in this particular setting, but no convincing benefit in overall survival. These are the data that we’ve seen from the Japanese trial and, again, from an overall survival point of view, certainly not what we saw in the FLAURA trial with the survival advantage with osimertinib here.

Another interesting path is we’ve had a couple of trials that have reintroduced chemotherapy into the combination with a TKI in the 

 mutated space. Again, this is the Japanese trial with gefitinib plus or minus chemotherapy. We did see in this trial not only a PFS, but an overall survival advantage in this population. So very interesting. Obviously, there were no surprises with regard to the adverse events adding chemotherapy back in, pretty much what we would expect. Of course, if you go back, way back, I can’t even remember—Ed, maybe you’ll remember—but all the trials in unselected populations where we added gefitinib to chemotherapy or erlotinib to chemotherapy, all of which were negative in unselected populations, there was no concern about giving the TKI with chemotherapy from a toxicity point of view, and I think these data bear it out.

Case 2: EGFR+ NSCLC Combination Therapies

 I think to Tim’s point, sometimes the molecular testing can take 2, 3 weeks, and the patient has time to see the radiation oncologist, have an opinion, and in this case SRS was done while she was awaiting molecular studies. I don’t think that that’s necessarily a bad thing, and Tim made all of the great points with regard to that. She was started on 80 mg of osimertinib daily, and she had a good partial response. Now unfortunately, 20 months later—remember the median PFS was about 18.9 months, so she’s spot on in terms of when we think she might develop some issues—she has some new liver lesions, she has a new brain lesion; she still retains an ECOG performance status of 1. What do you do now, Ed? You’ve got a lady who’s kind of wearing out her welcome on osimertinib, she’s got new disease in the brain and liver. Do you repeat testing?

 Whenever there is progression on an existing therapy, we like to rebiopsy if possible. If there’s new or existing disease and it’s growing, that’s the first step. Liquid biopsy is also an option that we do, that’s less invasive. As was reported out of Massachusetts General Hospital years ago by Jeff Engelman and others, transformation to small cell that can occur in patients. We always try to get tissue first, because that can be up to 14%, 15%. Then that’s where we get artistic, right, Mark? If somebody has done really well, then we start dabbling, if there can be something added. I know some people are tempted to even add a VEGF sometimes, depending on the type of progression that occurs. I don’t think that’s something we characterize very well in the metastatic setting, is that there are different types of progression. Some make us very nervous and some make us less nervous. I’ve followed a patient who has progressed on a TKI for 3-plus years, and she was still hiking in the mountains and stuff. We don’t quite understand it, and there are different scenarios, but the academic answer would be that you could consider something like IMpower150 because it did include these patients who had an 

 translocation and gave them the 4-drug regimen with the chemotherapy, carboplatin/paclitaxel, and then gave them the VEGF plus the I/O. I think that’s the academic answer. I can’t say that I’ve done it that often, but it is out there.

 Yes, I also retest. Obviously, you have to be aware of histologic transformation. It’s a different clinical situation typically with rapid progression, you should be suspicious of histologic transformation to small cell. In the typical slow-progressing sorts of patients, which also is fairly common, I do test, but I often don’t feel good about testing because I don’t know that it’s informative most of the time. We got so used to retesting when we used the first-generation drugs because everyone was excited about finding T790M. That was a great story because we had osimertinib. Now that we use osimertinib first line, at least my understanding of the resistance mechanism is they’re all over the map with regard to things, and this C797S resistance mutation is not like T790M was, you know, 60% of the patients, it’s more like 8% to 10% of the patients, and I still don’t quite know what to do with that outside of a clinical trial.

You mentioned IMpower150, obviously a trial near and dear to my heart. One of the things that I struggle with is physicians who treat with the KEYNOTE-189 regimen, and these patients were excluded. It seems kind of intellectually dishonest to me. At least the scant amount of data we have, either with monotherapy or with chemoimmunotherapy, does not suggest that this population gets a great benefit from immunotherapy. Your comments on that, Ed?

 Yes, I feel the same way, Mark. I’m not a big fan of doing KEYNOTE-189. I think if you don’t go with a VEGF-type regimen afterward, then it’s a chemotherapy-based regimen. You hope and pray that you find something on the repeat biomarker testing, but that’s generally the route that we will go as well.

 Let me ask Tim. I guess you could make the argument that this lady has oligoprogression, she has a couple of lesions in her liver and a single lesion in her brain. Back in the old days, we got some traction when patients were on erlotinib/gefitinib, and they didn’t want to go to chemotherapy. If they had 1 or 2 spots that were progressing, we’d typically zap them or resect them or something like this. What are your thoughts in a patient like this? Would you ever offer the patient focal radiotherapy? She seems to have a limited number of sites.

 It’s a great question. Of course, as you said earlier, there’s not a lot of room to wait in the brain. Normally we’re going to offer radiosurgery to a patient like this. I think the question more so is, is there value to trying to ablate other extracranial sites of progressive disease? There are some emerging data in that. Interestingly, at ASCO last year there was this SINDAS study out of China looking at up-front SBRT [stereotactic body radiation therapy] for molecularly driven 

 patients, and it showed a survival benefit. Now, this was not the oligoprogression, this was up front, but it does speak to some value to debulking patients with a relatively modest volume of metastatic molecular-driven tumors. I think it’s an area that we’ll know more about in the coming years. I do think it’s entirely reasonable for oligoprogressive disease to keep them on the drug that seems to be working at most foci and add radiation in or surgical resection. I’m sure Dr Stiles has experience with doing some focal surgeries on these types of patients as well.

 Alright, this has been a great discussion. I think we’re going to end case 2 and then transition to case 3.

Case 2: Treatment After Progression of EGFR+ NSCLC

 This is case 3. This is a 63-year-old gentleman who presented to his internist with intermittent cough and some dyspnea on exertion. He has hyperlipidemia, hypothyroidism. His main comorbidity is COPD, which is managed with inhalers. He recently quit smoking, although he has a 40-pack-year history. He presented with intermittent wheezing and the shortness of breath as I mentioned, but otherwise negative. His performance status is a 1 based largely upon the pulmonary disease. The patient’s laboratory test results were normal. A chest CT scan was obtained, which demonstrated a suspicious 3.1-cm spiculated right upper lobe mass. There were also 2 enlarged right mediastinal lymph nodes, measuring 2.5 cm and 1.7 cm, the suggestion of relatively bulky disease, as well as moderate emphysema. The PET scan confirmed the lung lesion is PET avid and also the mediastinal lymphadenopathy, but was without evidence of distant metastases or other lymph node disease. A brain MRI was obtained, which was negative. The patient’s pulmonary function was fairly poor with an FEV1 of 1.2 liters and a DLCO of just 35%.

 Brendon, give us your thoughts in terms of an initial staging approach in this patient. You reviewed the radiographic findings. Let’s assume you’re seeing this patient in your clinic for the first time, how do you go about getting a diagnosis and combine that with the right staging? Would you favor one procedure over another, or how do you think about this?

 Good question. We probably, like most of you, like to go after what we think is the highest stage disease first. Rather than doing a lung biopsy, I would start with staging of the mediastinal nodes. I think it’s dealer’s choice in terms of EBUS versus mediastinoscopy, but I think most have moved away from mediastinoscopy and would prefer EBUS, particularly here in this setting of bulky mediastinal nodal disease. Mediastinoscopy may have an advantage of being better at checking the N3 disease, and I do think it’s an important point for somebody with bulky N2 disease, if you’re considering resection, to rule out N3 disease. I try to make a good effort to get not only the N2 stations, whether there’s 1 N2 station or multiple stations, but also to check some of the N3 nodes.

 In the setting with very bulky nodes, I assume the false negative rate of EBUS is fairly low, correct?

 Right, it should be less than 10%. Typically we can get enough for diagnosis and for molecular from there if people thought that was important as well.

 In knowing that EBUS does have some operator dependency aspects to it, if you came up negative on an EBUS, is this a patient you would do a mediastinoscopy on following that?

 Yes, that’s a great point. This is the classic case where I would not accept a negative EBUS, nor should anyone else in the community. He’s got pretty clear-cut evidence of regional lymph node disease or mediastinal lymph node disease, so a negative EBUS here should be followed with a mediastinoscopy.

 Let me ask, what I think, and I’m not a surgeon, but what I think is difficult to answer is what are the criteria for resectability? Then, more importantly, this term unresectable. Does unresectable mean they’re technically not resectable, or there is not sufficient evidence that they should be resected?

 I think the latter is true, and I think you and I have talked about this before. Probably anything can be resected, and the real question to surgeons is should you or should you not. It’s a tough question. I think bulky N2 disease should give people pause. Yes, I’ve resected the occasional patient with bulky N2 disease, and particularly if we think we can do a lobectomy rather than a pneumonectomy. But that’s often the equivalent, as you all know, of systemic disease. So bulky N2 disease, multistation N2 disease, disease requiring a pneumonectomy, for me often fall into the I-probably-shouldn’t-resect side, even if I thought it was resectable disease.

 My training was in this group of patients, and again to your point, there’s a great deal of heterogeneity in this population of stage III patients. The surgeon’s stage IIIA may be different than a radiation or a medical oncologist’s stage IIIA. I’ve always been taught that the decision about surgery should be made up front, and that this strategy of saying, “Well, let’s give them some induction therapy, see how they do, reassess for surgery down the road,” is kind of a treacherous path to take because we really don’t have necessarily a good way to assess people that way. Is that your thinking, or educate me on that.

 I totally agree with you. I think it’s a hard decision to make after the fact, and this should be made up front based on the discussion with the multidisciplinary team, what your radiation oncologist thinks, what your oncologist thinks, what the patient thinks, because there’s multiple reasonable ways to approach this. It’s a little bit of a fallacy to think that chemotherapy can take an unresectable patient and make him or her resectable. When you can make this decision up front, I always think it’s an easier course for the patient.

 I think that’s a good point. I have seen a number of patients seeking second opinions who have started on therapy, and the assumption is that some sort of induction therapy with standard chemotherapy is going to take a patient who clearly unresectable disease and they’re going to make them all of a sudden resectable. Or if you really think you need a pneumonectomy to start with, you’re going to downstage them so much that you can do a lobectomy, and I don’t know that we have any evidence to support that that would be the right thing to do.

 Yes, I agree, it’s pretty unlikely that you would have that kind of response with chemotherapy.

 Yes. And then, we’ll talk a little bit about this when we get further along in the case, but obviously we’ve had a change in the standard of care with the incorporation of immunotherapy as we alluded to before. Anything in that data set that changes, say, what your thinking was 5 years ago? Has it changed given the PACIFIC data in your mind in any way?

 It’s a great question. We’ve got great options for stage IIIA that we’re not going to resect now, and in the past that might not have looked as good as the PACIFIC arms. I think if you’re in a place where the surgery, where you’re worried about it, I think it’s perfectly good for III and IIIA patients to get chemoradiation and then durvalumab. For me, as immunotherapy keeps creeping earlier and earlier, I think that might be competitive. Certainly to roll the dice for the IIIB or IIIC days, I think should be largely behind us.

 Institutions tend to be either more aggressive or less aggressive in stage III disease. I want to ask Ed and Tim, what’s it like at Levine Cancer Institute? Are the surgeons particularly aggressive in this stage III setting or less aggressive? And then the same question to Tim.

 I’ve actually found our surgical colleagues to be extremely reasonable, even when I was at The University of Texas MD Anderson Cancer Center with Steve Swisher, and that group. They liked induction, but just as Brendon said, I don’t believe in downstaging in lung cancer. It just doesn’t happen. You’re either resectable and you’re given the chemotherapy up front, or you’re unresectable.

 I think that’s an important message to our audience. Tim, your situation there at your institution?

 I would echo a lot of Brendon’s points, and I think my surgeons follow a lot of the paradigms he outlined in that it’s not whether they can be resected, it’s whether they should. When you’re getting to the point where you’re trying to convince yourself that downstaging with chemotherapy is going to turn someone from unresectable into resectable, you’re probably talking about the patient who has such a high risk of distant metastatic disease, that you probably shouldn’t be operating. And our surgeons are very thoughtful and reasonable men in that same way.

Case 3: 63-Year-Old Man With Unresectable Stage IIIA NSCLC 

 A bronchoscopy was performed, and a transbronchial lung biopsy of the right upper lobe mass was performed. This showed adenocarcinoma, and the bronchoscopist was able to biopsy 4R and 7 and confirmed that those were also metastatic nodal disease. They did make an effort to sample the 4L, which was negative. The final clinical stage here was T2aN2M0, stage IIIA. Then based on the bulk and the extent of the mediastinal disease and the active emphysema, the patient’s cancer was deemed to be unresectable, or shouldn’t resect, and he was referred for consideration of concurrent chemotherapy and radiation.

 Ed, your thoughts about molecular testing in this stage III disease? Do you routinely do it?

 I would say that prior to maybe about a month and a half ago, we were just testing in stage IV patients, although we had reflex testing set up with any lung cancer diagnosis. I would say that there’s been some change in the paradigm, and we are now doing molecular testing on everyone. However, if we knew that they were specifically a stage IIIB, unresectable, undergoing concurrent chemoradiation, there are no standard biomarkers we would utilize to make any treatment decisions here.

 Patients with advanced stage III disease end up being stage IV, and it would be pertinent to know at that point. Again, I don’t know that we have any data to date that suggest that we should act on whatever we find from molecular testing, at least in this stage III disease. Obviously, ongoing trials may change that, and of course we all saw the ADAURA trial, so we may all be testing early stage disease, but that’s a subject for a different day. Options for treatment, we’ve already decided this patient is not a surgical candidate. I think we spent the 1990s and early 2000s showing that concurrent chemoradiotherapy is the standard of care and not sequential. I think we can move on and let Dr Kruser address the next step and considerations from a radiation oncology perspective.

: Yes, I think that’s a great introduction. Thank you. As you pointed out, concurrent chemoradiotherapy was established by a trial that was launched in 1994. We’re going to talk a little bit more about how we treat now, both in the setting of RTOG 0617, which is a dose-finding trial, and then touch upon the PACIFIC trial, which has really changed practice. There are a few questions here about how we keep patients optimized, especially now in light of the PACIFIC data, and wanting to make sure they’re in good condition to get on to durvalumab. I would highlight the role for multidisciplinary care in this setting. We as radiation oncologists really rely on our nurses, who are really vital in helping keeping these patients fit through chemoradiotherapy, as well as our nutrition colleagues. When we run into esophagitis, keeping good hydration and caloric intake is very important. Furthermore, social work is actually underrecognized, an important player for people receiving chemoradiotherapy. Treatment breaks for transportation issues related to 6 weeks of chemoradiotherapy can be an unrecognized problem for these patients. So a lot of times our social workers are very important for helping to facilitate that. We’ll talk a little bit more about some of the radiation data, per se.

Case 3: Molecular Testing in Stage III NSCLC 

 Here’s the schema for RTOG 9410. It is an old study, so I won’t spend too much time on this. It compared sequential chemoradiotherapy with 2 forms of concurrent radiotherapy—1 with twice-daily radiation, which we don’t really use and was not significantly different—and the outcomes for the concurrent chemotherapy with once-daily radiation was significantly better. If you look at the study results, here are the overall survival results. At 5 years, the overall survival is 16% for concurrent chemoradiotherapy vs 10% for sequential. This really set the stage for subsequent examinations. Sadly, this was again launched in 1994, and we really didn’t have any breakthroughs until PACIFIC, which we’ll touch upon in a few slides. Here’s RTOG 0617. I kind of previewed this. This was a dose-finding study, standard dose being 60 Gy vs an experimental dose of 74 Gy. It was a randomized study, 460 patients, and overall survival was the primary outcome. As radiation oncologists, we all like to think that more dose is better, and we were a little disheartened to see these results. To many, it was a surprise that the 60-Gy arm is numerically and statistically significantly better than the 74-Gy arm for overall survival. This is the progression-free survival curves. They’re relatively close, but when you flash ahead and look at the overall survival curves, there’s actually more distance between those curves and more significance, suggesting that the excessive dose with 74 Gy is driving some noncancer deaths. These were not all because of progressive events from cancer but seemed to be toxicity related.

Looking at the adverse effects, here’s the 60-Gy arm vs the 74-Gy arm. Numerically, these were not different. That 

 value is actually a little misleading. This was a 4-arm study, which included 2 cetuximab arms. We haven’t discussed those because that therapy was not beneficial. But looking at the standard dose vs the high dose, the adverse effects were not statistically significantly different.

That said, if we look at the treatment-related deaths, there was a concerning number of treatment-related deaths and also noncancer deaths that were maybe not attributed to radiation in the 74-Gy arm. If we look at the ones that were probably or likely attributed to radiation, we see events for radiation pneumonitis, respiratory failure, tracheo-esophageal fistulas. Numerically, these were all more common in the 74-Gy arm. Here are the patterns of failure data for RTOG 0617. It describes how the high dose did not achieve better disease-related outcomes in terms of local or regional control in the 60-Gy arm. I think that was disheartening to many people who ran the study.

Looking at factors related to survival, the high dose on multivariate analysis comes out to be a negative prognostic factor. Additionally, heart dose was a negative predictor, speaking to high dose being associated with noncancer deaths. Likewise, tumor location, the left lower lobe, again probably related to heart dose, was a problematic prognostic factor.

If we summarize RTOG 0617, the standard-dose arm of 60 Gy remains the standard of care, despite efforts to intensify a dose selection or dose escalation in these patients. The median survival was 28 months. The 5-year overall survival was 32% here vs 16% that we highlighted back in the 1994 study. Undoubtedly some were from stage migration but there was also improvement of radiation technology with IMRT utilization and better systemic therapy.

 Tim, thanks for that. I wanted to ask you a couple of questions, because a lot of people were surprised with RTOG 0617. I spent the first part of my career at the University of North Carolina doing dose-escalation trials for 74 Gy into what used to be the CALGB [Cancer and Leukemia Group B trial] there, and obviously we were thinking this would be the winner. I have 2 questions. Why do you think this happened, that 74 Gy was worse? You were alluding a little to normal tissue toxicity, particularly the heart. But also the recurrence patterns here. I just wonder if many radiation oncologists were very conservative in their fields and might have had marginal misses because they had to go to 74 Gy and they didn’t want to over radiate the lung or this or that or everything else. That’s 1 question. The second question is that every radiation oncologist I talk to doesn’t think 60 Gy is a great dose, and they actually go to 66, 70 Gy in select patients. Even though these data as presented are pretty convincing, the nuances of the data are what we all continue to ruminate about and try to figure out how can we explain this. It shouldn’t make sense. I just want to get your perspective on that.

 Those are definitely some great questions. We radiation oncologists all need to be judicious with our radiation. We know from earlier studies, where we gave radiation for post-op RT, for early stage disease, that thoracic radiation—when it’s not adequately addressing disease—is morbid and contributes to mortality. We’re seeing a single of that in the RTOG 0617 data. I do believe it. As to why the disease-related outcomes or local control were not different, there were some data suggesting there were some less-than-rigorous radiation quality in the higher-dose arm, perhaps related to what you alluded to in terms of people scaling back on margins, for example. Did that contribute to marginal misses? It’s hard to know, but potentially.

What dose do I use now, and do I think 60 Gy is the right dose? We are talking about using this with dual-agent concurrent chemotherapy, and that needs to be highlighted. Sixty Gy may be enough, especially in the setting of the PACIFIC data, which we’re going to talk about. I’m rarely going above 60 Gy because I really want to make sure these patients are not harmed by my therapy, because we know we have something that comes after that has an overall survival benefit. I’m very cautious about going beyond 60 Gy. I do it occasionally for the low-volume patient for whom the disease is relatively close in proximity between the nodal disease and the primary disease, and lung and heart parameters are easily met. But we have to take our medicine with these data and believe that more is not always better.

 That’s an excellent perspective. Ed, chemotherapy regimens? That might be 1 of the least interesting things to talk about in terms of what we use with radiotherapy. I think our national platforms are either carboplatin-Taxol or cisplatin-etoposide, and there are a lot of chemotherapy drugs you can use with radiation. I took to heart the PROCLAIM trial. We all thought pemetrexed would be a great radiation drug with cisplatin, and it turned out to be no different from cisplatin-etoposide. Your thoughts?

 I’ve always been a cisplatin fan because in the curative setting, it seems to lean a bit that way. Of course, when I was at MD Anderson, they liked to talk about how radiation is the best in the world at MD Anderson. It’s better than anywhere else in the country. That would negate the cisplatin effect, and you could use carboplatin. I’m kidding, but it was kind of true. As someone who also treated head and neck, we’re not scared of giving high-dose cisplatin. We can give 100 per m2. So I don’t understand when people get really skittish when we’re giving lower doses in lung. I tend to be a cisplatin-etoposide person. We have carboplatin-Taxol weekly on our regimens. I don’t know if it’s necessarily easier to take, but it’s what people are comfortable giving.

Case 3: Chemoradiation for Stage III NSCLC 

 As we discussed, this patient underwent cisplatin-pemetrexed with concurrent thoracic radiotherapy. He had a partial response with a nice shrinkage, both in the primary and nodal disease. Upon restaging imaging, there was no distant spread observed.

 This sets up the patient for the PACIFIC data. He’s had a nice response. One question, both for Tim and Ed: Given that there’s some suggestion that earlier immunotherapy is better, how do you handle the restaging of these patients following completion of chemo-RT? I’ll ask Tim first.

 I do put some credence in those data. The hazard ratio for the patients who got the durvalumab within 14 days was marginally better than most who got it later on. Why you’d like to think that maybe that’s because those were fitter patients, there were also placebo patients who were getting the placebo within 14 days. So the hazard ratio should reflect the activity of the drug in that scenario. It’s something to strive for, and I like to have that discussion with my medical oncologists as they’re getting midway through their radiotherapy: Can we get the scan set up pretty quickly and have this be an expedited transition to the next phase of care?

 Yeah, that data subset was really profound, and the design of the study that allowed it was really great because it certainly has always been ingrained in us to let the patient cool off. Let’s give them a couple of months off and let them get recovered. It reminds me of orthopedic surgery, right when you’re done with a knee surgery, you’re putting them in mobility and getting them moving. It’s the same thing here. Don’t stop, hit them while you’re hitting the tumor hard. We like to start within 2 weeks. We like to have a visit 1 week after the completion of chemoradiation, and then starting the durvalumab quickly, shortly afterward. So we’re very aggressive with that.

 I agree. Occasionally you have a patient who has a great deal of esophagitis or fatigue. They need a little more time to recover. All things being equal, we try to start it. To Tim’s point, I schedule the CT scan within a week after completing radiotherapy, get them authorized, and get them going as quickly as we can. Tim, let’s move on and talk about the PACIFIC trial.

 Here’s the schema for the PACIFIC trial, which we previewed. Randomized study at 2-to-1 randomization of people who have completed chemoradiotherapy. It’s important to note that these patients, having completed chemoradiotherapy, had a scan that did not show progression. They were then randomized, again 2-to-1, to either durvalumab every 2 weeks or placebo for 1 year. There were coprimary end points of progression-free survival and overall survival, which resulted in sister publications in the 

. Here is the overall survival curve. This was updated at ASCO, about 3 years, we’re seeing that 36-month time point. A nice separation of curves early has persisted and numerically actually gotten larger over time, which is encouraging in that it suggests that durvalumab may be curing some patients who would have been otherwise destined to blossom with metastatic disease. It has really catapulted this paradigm into standard of care.

When we look at how these patients progress, the patients on the placebo arm, about a third developed progression disease. That’s cut by about one-third by the use of adjuvant durvalumab. Site by site, those patterns relatively hold in place. A lot of interest on the brain. If you look numerically, it’s about 50% reduction in the brain metastases. Regardless of whether you put a weight on that, it’s probably not statistically valid to say it’s different from the other sites. But I do think that reducing hematogenous disease is going to reduce brain metastases. But we also know that immunotherapy has independent action on brain metastases and might be trying to set up shop. It does work all the extra-thoracic spaces.

Here are some subgroup analyses looking at PD-L1. We can see that there was a nice hazard ratio for those patients with more robust PD-L1 status. All patients were eligible for this study regardless of their PD-L1 status. You notice a fair segment of people did not have PD-L1 testing, unknown category. But some thought-provoking data on whether people with PD-L1–negative tumors may not benefit from this approach. That was not a prespecified or powered analysis. Post hoc analysis—here are the safety summary data. It’s a relatively well-tolerated treatment. If you look at leading to discontinuation, about 15% of patients receiving durvalumab had to discontinue it vs 10% of people receiving placebo. There was nothing particularly surprising here about the safety paradigm for medical oncologists who use these immunotherapy drugs. Commonly, the expected adverse effects in immune-mediated adverse reactions were relatively commonplace. Importantly, the big concern was pneumonitis. How would radiation, thoracic radiation, high-dose radiation play with immunotherapy? While any grade pneumonitis was 33% in the durvalumab arm vs 24% or 25% in the placebo arm, when you look at the high-grade, grade 3/4, they weren’t different, hovering about 3%. This has led people to be quite comfortable with dealing with durvalumab and any pneumonitis that comes thereafter.

Case 3: PACIFIC Trial in Unresectable Stage III NSCLC

 There’s probably little debate between Ed and me about this being the new standard of care. Tim, in the radiation oncology community, I assume this has been adopted as a standard of care.

 Yeah, absolutely. We’re all very heartened by the potential to see these people do well. We’ve always used 60 Gy of radiation in this setting, and we’ve been relatively underwhelmed by the outcome in that we get a lot of esophagitis and cause a lot of toxicity, and the vast majority of these people were developing metastatic disease. We’re all pretty enthusiastic to update this approach.

 We talked about the issue of starting it earlier. We had consensus that earlier is better to the extent that you can get it started within 2 weeks, so we favor doing this. What about the duration of therapy? Is it important to get a year of treatment? I’m asking that question because I did see some post hoc analysis suggesting that if you got close to the defined number of cycles, there seemed to be better outcomes. One has to wonder, is that the treatment or is that just more robust patients, they’re able to hang in there longer. It’s hard to know. Certainly, my practice has been that if this is the way they did it in the phase 3 trial that showed an advantage, that’s the way I’m going to do it in my practice once it becomes standard of care.

Ed, I want to ask you a couple of questions, because this comes up quite often. Does PD-L1 status impact your enthusiasm for this in the consolidation setting?

 Yeah, you showed some of the data about PD-L1 status. We don’t measure it; we don’t use it. I know the Europeans feel a little different, and I guess I’ve never seen them approve on a post hoc analysis in another trial that went the other way. They would use a post hoc analysis to restrict utilization of some of this therapy in patients, but I didn’t feel it was right. You know, it is what it is, but in the United States, there should be no biomarker restriction based on the trial and its design.

 Yeah, let’s stick with that. What about driver biomarkers, like if you happen to know that the patient’s 

 positive or something like this? There were very few patients in the trial, so I don’t know how to make this decision in this particular setting.

 This is clear, Ignorance is your friend. Until we have more mature data that we’re actually harming folks, it’s dangerous to make extrapolations. You use the pemetrexed example earlier. None of us would have thought that there would be any problem substituting in pemetrexed, but it just didn’t pan out. We’ve seen the same thing in head and neck. Right now I follow the trial the way it is. Once we think we have medicine figured out is when it gives us an opposite result, just like the radiation data earlier.

 Right. Tim, a lot of our efforts during the concurrent chemoradiotherapy part, treating esophagitis, dehydration, nutrition, all these sorts of things, there is a risk that certain patients will decline in their performance status. Of course, for PACIFIC, you had to be ECOG 0 to 1. Sometimes at the end of radiotherapy, that last week is pretty tough, and they might be like 1.8 and not exactly a 2, certainly not a 3. But do you extrapolate the PACIFIC data to those kinds of borderline performance status patients, or do you exclude them from it? What’s your thinking there?

 I try to fit my patients into that status. What I mean by that is, when you see that borderline patient, we know that the durvalumab has an overall survival benefit and we really want to strive to get that. The reason I bring that up is there will be times where someone is not doing super great at 35, 40 Gy, and I’m having a conversation with their medical oncologist. Boy, do we need that fifth or sixth cycle of concurrent carboplatin-Taxol. Can we drop those weeks that have a very uncertain benefit, so the patient is not a PS 2 at the end of this? They’re at 1 or 1.5, and we can get the durvalumab. That’s an important foresight to try to have with these patients with this new approach.

Case 3: Standard of Care in Unresectable Stage III NSCLC 

 Let’s move on and talk about what happened to the patient.

 As discussed, the patient went on planned durvalumab. A few cycles in, he developed findings concerning for pneumonitis vs an infectious cause, shortness of breath, mild cough, no clear infectious contacts.

 You reviewed the data suggesting that there was really no increase in grade 3/4 pneumonitis. There was a mild, slight increase in grade 1/2 pneumonitis in this setting. How do you manage these patients? Ed, what’s your strategy for these patients such as this?

 It depends on the severity. This patient is symptomatic, and we’re seeing something on scan, so we would stop the drug, give steroids if we need to. Obviously, this is an evolving process. Sometimes patients, it’s very mild and you can re-challenge them. Most of the time, though, we don’t do that because it can get worse. The good news is, I always tell patients that if I can’t give them the full therapy—and in PACIFIC, most of the patients were able to complete the prescribed therapy—we’re not really sure how long or how much immunotherapy is required for patients to get that lasting effect, to develop that memory in their T cells. If the tumor doesn’t mutate quickly, you might get long-lasting effect even with a few doses. People are happy they’re getting an immune response but upset they can’t get the prescribed therapy. But you never know with these.

 Let’s say you get aggressive with steroids and give them a good course of steroids, and they taper slowly. Do you ever go back? I guess it depends on the severity of the pneumonitis, whether you take the chance of reinstituting immunotherapy. I don’t know that there are any specific rules here.

 It also depends on how much treatment we’ve given them. If we’re in the second half, it’s less inclination from my standpoint. If it started early on, then we have tried to re-challenge if there was a very straightforward response to stopping and giving steroids.

 I’ll chime in here. It’s an important discussion to have with the radiation oncologist as to whether this is seemingly radiation pneumonitis or drug pneumonitis. We’ve pooled some data with Washington University, looking at our experience in PACIFIC patients, for patients deemed to have radiation pneumonitis. Obviously that can be a challenging determination, but you’re somewhat going to use the location of the radiation dosed tumor to help guide that determination as well. It’s a little bit of a characteristic time course. Among our patients who we deem to have radiation pneumonitis, we were actually able to restart the drug without flaring up the pneumonitis again in about 85% of patients in small numbers. I think stopping drug early for a fairly expected radiation toxicity, which can sometimes be managed quite easily, might be a bit premature. A multidisciplinary conversation should be had.

 Yeah, I was going to reinforce that point. It’s incredibly important for the medical oncologist and the radiation oncologist to work together. The right hand has to know what the left hand is doing, and you have to gain consensus in assessment of the patient in these sorts of things. This has been a wonderful discussion of these 3 cases. I want to thank our panel for your thoughtful case presentations. Certainly, it’s been a lively and informative discussion. To our viewing audience, thank you for joining us for this 

® presentation. We hope today’s discussion was a valuable use of your time and that you acquired some practical knowledge that you can take back to your clinic. Thank you so much for joining us.

Clinical Cases in Lung Cancer