What about monitoring patients for recurrence? Typically, we have our surveillance strategies that are ctDNA [circulating tumor DNA] based for the most part. Is any potential for earlier detection with plasma-based approaches?

 At ASCO [American Society of Clinical Oncology Annual Meeting] 2021, there was a good abstract presentation talking about the feasibility of doing ctDNA even in patients with early stage lung cancer. They were able to get samples from the LUCID trial patients. Though the representation of stage I was 50% of those samples, the detection rate in that stage was 24% if you look at it stage by stage. We all know that as the stage goes higher, so does the detection rate. They collected data before the diagnosis, and then they did it at landmark time at 3 months and then at 4 months, 6 months, 9 months, and 17 months. They were able to show with a super strong hazard ratio of 0.20 that if you detected it, then it did relate to an early relapse. Almost all the patients who were detected were ctDNA positive. And the lead time was anywhere from 120 to 200 days. There was another interesting abstract…that also showed they were more successful in picking up these MRD [minimal residual disease] patients and predicting for relapse. We all understand that this may be a predictive marker. The bigger question remains, how can we up the game in this sense? Can we detect it more reliably and not just in factional patients? The other question would be, we picked it up, now what do you do with that? Do you start them on therapy? If you do, based on what? There are no data to use. They’re more prognostic, but I would like them just to become more predictive.

Yeah, we have a lot of work in front of us in terms of proving the clinical utility. At the end of the day, we have to prove clinical utility and that it does make a difference. It certainly is an exciting time.

Videos

The potential clinical utility of blood-based biomarkers in non–small cell lung cancer based on recent data presented at ASCO 2021. 

Blood-Based Biomarkers in NSCLC 

I’m going to go to our Southern California colleagues and ask them about the best approach in the preoperative time period. The question is about neoadjuvant alone, adjuvant, and the arguments for both. What’s the best strategy? Is it the chemotherapy–I/O [immuno-oncology] strategy vs I/O alone? What are your thoughts? I’ll start with Sandip.

 It’s a great question, and tumor boards are about to get even more interesting. But a neoadjuvant strategy necessitates a multidisciplinary evaluation of a patient. Because if you’re in a practice environment, in which you don’t even know a patient is a candidate for neoadjuvant therapy because you see them only after the surgery, then you have 1 choice. It’s adjuvant therapy. A lot of surgery is done in that setting. For those of us who’ve had the opportunity to participate in multidisciplinary discussions, to use neoadjuvant chemoimmunotherapy, my general sense—at least with the current time—is a combination of chemotherapy plus immunotherapy in the neoadjuvant setting would be preferred, even if you’re going to take the IMpower010 approach. You’re going to be given 4 cycles of chemotherapy followed by atezolizumab for a year. We’re not at a juncture in the locally advanced localized setting to do away chemotherapy.

For patients who are stage III, or would require pneumonectomy, or would have a very difficult surgery, at tumor boards, those patients will be considered for neoadjuvant chemotherapy–I/O. The ability to make the surgery easier for the surgeon: convert pneumonectomy to lobectomy, sterilize some of the occult-positive lymph nodes that may be in the mediastinum. For the higher-stage patients—stage IIIA, those questionable stage IIBs that may be IIIAs once you go to the OR [operating room]—a neoadjuvant chemotherapy–I/O approach may be preferred. But there may be patients for whom the mediastinum is completely negative—stage IB, stage IIA—and it’s going to be a lobectomy. It’s going to be relatively easy for the surgeon based on the tumor size due to resection. Those are patients who have the opportunity for neoadjuvant vs adjuvant therapy. That’s where the discussions will likely lie.

One of the things about neoadjuvant therapy, at least the way it’s built, is that we don’t even have survival data. It’s 3 cycles in your adjuvant therapy, and then you’re done. We don’t know what to do if the patient doesn’t get a complete response. Are we going to be doing a year of adjuvant therapy if they have only 20% tumor eradication? These are questions breast oncologists deal with all the time because they’ve been doing neoadjuvant therapy for a long time. In short, things are going to get more complicated before they get less complicated. The general premise is that a patient who requires a more aggressive surgery pneumonectomy and is likely to have positive nodes may have a preferred benefit for neoadjuvant chemoimmunotherapy. It’s probably 1 of the approaches we’ll take at our institutions, seeing how we handle the clinical trials.

To complicate that argument a little, those exact patients may be the best candidates for chemotherapy radiotherapy strategies with I/O. You’re exactly right: It’s going to get more complicated. Many of the questions in this setting we may have to answer all over again because the historical data may not be pertinent. Chaitali, your thoughts? Do you agree with your Southern California colleague?

 I do. One more important factor is your PD-1 status. What are your molecular drivers? If your PD-1 is deeper, is that patient best on neoadjuvant or adjuvant? That needs to be factored in.

If they can work this into an adaptive strategythat could really push the neoadjuvant approach more.

Drs Chaitali Nangia and Sandip P. Patel debate between best treatment approaches for patients with non–small cell lung cancer in the perioperative setting. 

Perioperative Approaches to NSCLC 

Roy, your thoughts on how you see the future. Many, historically, these lung cancer patients don’t necessarily get a biopsy before going to the OR [operating room]. We know the PD-L1 [programmed death-ligand 1] typically is tested on core biopsies, which can sometimes be a challenge in these early patients. Your thoughts on that situation?

 Certainly, like everything else now, we’re going to need more information early in the patient’s course. So post-op, we’re going to need to know PD-L1 status, just like we want to know mutational status. EGFR, of course. But probably, someday all the different actual mutations. So, PD-L1 is an easy test, it takes 1 section. So, I suspect that a tumor board post-op will probably have a PD-L1 at the same time we know the histology. And then the molecular test takes about another week. But I think this is going to be important because we are going to have to use that, at least in the early round, for the actual immunotherapy.

How do you, if the neoadjuvant approach catches on, how do you see the whole bio…testing for biomarker thing playing out in that situation where obviously with the ADAURA data, you might be in a position where that is the standard of care. You would want to get the patient resected and then onto adjuvant Osimertinib [Tagrisso] versus if they were not EGFR-positive, perhaps neoadjuvant IO [immunotherapy] approach would be best. Any thoughts in that regard?

 Again, the earlier testing, the better. So, we probably have to test and do a full panel on every core biopsy, every sampling that we get. And that’s going to be the education. Getting that done around the country, especially for some of these. They’re all quite rare still—10% of patients in the U.S. for EGFR. I think if we’re doing ALK or RET, something like that, so we’re going to have to do widespread testing, and that’s going to have to become something that’s promoted—NGS testing for all. And then the PD-L1, yeah, they get in on every patient now. You don’t even have to ask for it.

Dr Roy S. Herbst comments on the utilization of biomarkers and information needed to help aid in treatment decisions about perioperative therapy for patients with resectable non–small cell lung cancer. 

Biomarkers for Immunotherapy in Resectable NSCLC 

Let’s transition to the data we saw from IMpower010. The phase III data that Heather Wakelee recently presented at ASCO [American Society of Clinical Oncology]. Martin, give us your thoughts on the findings in this trial.

 I think this was a surgical ASCO, interesting data in [CheckMate] 816 and IMpower010. With enlarged randomized biomarker space with IO [immunotherapy] looking at over 1000 patients after complete resection and completion of adjuvant chemotherapy of 4 cycles of a cisplatin-based chemotherapy randomized to complete a year of atezolizumab [Tecentriq]. Looking at outcome data across the PD-L1 [programmed death-ligand 1] spectrum, we’ve seen PD-L1 negatives greater than 1% and greater than 50%. There was clearly a staggered benefit similar like we’ve seen and discuss early in the metastatic setting and then the PACIFIC trial as well. This could be a staggered benefit with a hazard ratio of 0.4 for PD-L1 greater than 50% with a hazard ratio of around 0.63, if I’m not mistaken, for greater than 1%. But buried in there is the 1% to 49% that probably turn out to be somewhere between 0.8 to 0.9, and then really no benefit for the PD-L1 negative. So, my take on this is biomarkers here do matter enough. I think this tie into the discussion about what is better, neoadjuvant or adjuvant chemotherapy with the discussions previously. We’re looking at probably a need for a comprehensive biomarker analysis upfront. We probably need PD-L1 levels in the beginning including EGFR, ALK, and other molecular drivers to really map out the optimal treatment plan afterwards. I really think that this is a regimen where I really would like to see the 1 to 49 subset data. Biologically, I’m impressed that even in the absence of macroscopic disease, and we actually still see immunotherapy responses. I would always think that the immune system needs a tumor microenvironment to actually train on a PD-1/PD-L1 access to activate lymphocytes. So, I think that was very reassuring to see, but it will need to be biomarker driven in the same way as we’ve had in the earlier stage disease. There was clearly an increase in benefit as we move into higher stage disease, higher nodule disease. Some of the data in the lower stages that were enrolled here was probably not mature yet. There was a trend for benefit but not necessarily positive. PD-L1 levels greater than 50% without drivers. And probably other options including neoadjuvant or adjuvant chemoimmunotherapy or combination immunotherapy strategies and to drive for those that are less than 50%. So, no new lessons from stage III and stage IV disease, but affirmed in basically in the surgical setting here. Biomarkers really drive the decision making here for optimal therapy. Otherwise, the data looks very impressive.

Dr Martin Dietrich comments on the IMpower010 trial of atezolizumab versus best supportive care after adjuvant chemotherapy in patients with resected non–small cell lung cancer. 

Atezolizumab After Adjuvant Chemotherapy in Resectable NSCLC 

Roy, you’re the king of ADAURA in the disease-free survival [DFS] end point here. We have in CheckMate 816 a pathologic CR [complete response]. We also have major pathologic response rate end points in these sorts of things. Obviously, many of these are co-primary with event-free survival. What are your thoughts in this setting for a true overall survival versus some of these, hopefully, surrogate end points but we don’t quite know that yet?

 It’s tough. I think DFS is, in my opinion, an approvable and valuable end point. Remember with DFS, we also know patterns of recurrence, that we have recurrence in vital organs, like brain, liver, bone, and we’ve seen, for example, in ADAURA, that those have been decreased. Interestingly, breast cancer targeted therapies have been approved based on DFS for 15-plus years now. I still remember the ASCO [American Society of Clinical Oncology] where that happened with Herceptin [trastuzumab]. That said, I think the results of [CheckMate] 816 are just amazing. Large number of patients as we heard and the difference in pathologic response, extraordinary. But we’re still one step to make this a standard of care. We also don’t know what to do next. Do we continue therapy afterwards? Do we give more chemotherapy? Do we give more immunotherapy? So, I would say, I’d want to see a DFS end point of some sort, at least. And then, if that correlated then in the future, we could start using the pathologic rates if that correlates. But right now, the link has not been made yet. But certainly, the neoadjuvant is here to stay. It’s incredibly exciting. And I can’t help but think that that is going to be the future, but the future still has a little bit of time before it’s ready to take hold.

Do you favor the neoadjuvant approach versus adjuvant? We have several adjuvant trials ongoing at this point, but thoughts?

 I think adjuvant is still going to be much more favorite, and then we can do a poll here. But surgery, a very select patients who are getting surgery even for stage IIIA lung cancer these days. So, I think it’s going to be more adjuvant and the adjuvant will be in certainly in a stage IIs that we don’t even know about until surgery. I think adjuvant is going to be a bigger deal than neoadjuvant. Though I think that the neoadjuvant is nice because it’s allowing us to find the right regimens. And we’re getting some science out of it. And they can actually trick a tumor without any operative complications and allow smaller procedures to be done, for example, fewer right pneumonectomies, these are all good things. But I think adjuvant, we saw the first adjuvant trial now at ASCO and I think there’ll be 2, 3, or 4 more behind them in the next year.

Before we get to that adjuvant trial, I’m going to ask Martin to comment on that in a moment here. But I wanted to get Neal Ready’s perspective on this issue of differences in tumors that may affect resection rates. You’re at an institution with the history of a strong thoracic surgery. We all know the guys and gals at the Duke thoracic surgery department. What are your thoughts on the neoadjuvant and does it raise any issues from your point of view regarding more traditional surgeries followed by adjuvant treatment?

I agree with Roy that a lot of treatment will be adjuvant. I think the way it’ll play out is in patients we would’ve given neoadjuvant treatment to anyway—III A disease, multi-stage II, we’ll end up giving chemo IO [immunotherapy], and patients who are good for initial resection without any problems are then we’ll resect them and then get adjuvant IO. And I think that there are…we did a neoadjuvant pembrolizumab [Keytruda] trial, investigator initiated. And similar to these other studies where it’s clear that giving pre-operative IO is safe and feasible. And we published that data. But when Francis talked to Tommy {D'Amico} about it, with IO alone, with pembrolizumab alone, there’s a degree of inflammation and early fibrosis that you don’t see with chemo alone. So, the way Tommy put it was that it’s not as easy as after chemo alone, but it’s not as bad as after chemo radiotherapy alone.

It’s somewhere between. With IO alone, it probably takes a little bit longer and they end up converting a few of the video assisted surgeries to open. Now, with [CheckMate] 816 it may be different. Maybe when you give 3 cycles of chemoimmunotherapy, you actually shrink the tumor.

The chemo and steroids that may prevent inflammation, so you may actually make it easier. But I think 2 important points when we give preoperative IO is that one, the radiographic responses often underestimate the pathologic responses by a lot. And secondly, when you set up an IO response in the primary tumor, sometimes you see enlargement of regional lymph nodes. That is not cancer progression. That’s inflammation. And interestingly, often it’s like granulomatous disease. So don’t be put off if the preoperative CAT scan doesn’t show a response and don’t be put off if you see enlargement of regional lymph nodes. That’s probably not a true progressive response, progressive…that’s not probably actually progression of cancer. It’s probably just an inflammatory response to the immune system attacking the cancer. I agree, it’s, I think that neoadjuvant chemo IO and IO is going to be as Roy Herbst was saying, an opportunity to get IO into the mix earlier on. When probably there’s less immune evasion in the cancer and the immune system has a better chance to be effective against the cancer.

The impact of endpoints used in clinical trials of neoadjuvant therapy for resectable non–small cell lung cancer as they relate to newer drug approvals and real-world practice.  

Clinical Endpoints and Neoadjuvant Therapy for Resectable NSCLC 

Martin, you set the stage mentioning the issue in resectable disease. Not only did we see that phase II trial that you just reviewed nicely, but we also had a number of recent updates or data sets from some new adjuvant trials in non–small [cell lung cancer]. So, I’m going to ask Sandip to give us his thoughts starting with the CheckMate 816 [clinical trial], which we first saw at AACR [American Association for Cancer Research] this year. But some nice surgical outcome data at ASCO [American Society of Clinical Oncology] this year. So, your thoughts, Sandip?

 I really thought this was one of the most exciting areas as Dr. Herbst mentioned earlier. The use of these agents in early and earlier lines of therapy allows for more patients to potentially benefit. One of the theoretical concerns about the use of neoadjuvant therapy, whether it be chemotherapy, targeted therapy, or immunotherapy, or some combination therein have been the fear of creating a toxicity that would prevent the patient going to surgery and occurred of intense setting. That’s obviously something we want to be close to [a] never event. There are multiple studies that have looked at this over time. There’s the NADIM study from Spain, which [was] with chemotherapy plus nivolumab [Opdivo]. There’s the NEOSTAR study which looked at a pure IO [immunotherapy] combination of nivolumab and ipilimumab [Yervoy]. There’s the LCMC3 study which [was] with neoadjuvant atezolizumab [Tecentriq] and it [is] probably the most, at least, robust data set. The one that’s done the most presentations recently because of the excitement around the data. CheckMate 816, this was a phase III study with neoadjuvant nivolumab plus chemotherapy versus chemotherapy alone and [in] the neoadjuvant setting, it was 3 cycles neoadjuvant chemoimmunotherapy. We found a pathological complete response rate of 24% regardless of PD-L1 [programmed death-ligand 1] status with the integration immunotherapy. Without immunotherapy, just chemo alone, it’s about 2% pathologic complete response rate, and so very exciting. We don’t have yet mature survival data, though that this difference in pathologic complete responses is very striking. And one of the more striking differences that we’ve seen. And I think more recently at ASCO, we saw some of the presentations around surgical outcomes. Dr. Forb presented some of the work with pathologic complete response. Dr. Spice represents some of the work related to surgical outcomes, and not only did neoadjuvant chemoimmunotherapy not create problems in terms of going to surgery, but it also actually improves surgical outcomes, which basically went from it being a theoretical weakness, it's actually being a strength of the utilization of chemoimmunotherapy. And not only were there less delays for patients getting the surgery, but more patients got a lobectomy versus a pneumonectomy. The pneumonectomy’s a much more morbid surgery. Patients have less complications with neoadjuvant chemoimmunotherapy presumably because the tumor shrunk. The surgery was made easier. The operating time was reduced in the OR [operating room]. Many virtuous surgical outcomes that really took what was a theoretical concern around the drugs in neo/adju [neoadjuvant/adjuvant] chemoimmunotherapy. It actually demonstrated it’s a strength, this approach. So, I think we’re still waiting for survival data, but this is extremely provocative in terms of the pathologic complete response rate. And it appears feasible and safe based on the surgical outcomes we’ve seen to date.

I agree with you. That was an impressive presentation with regard to the surgical aspects of this. I don’t think we would anticipate. We’re hoping it wouldn’t be worse, and it looked a little better. And that’s not anything we’ve ever seen, to my knowledge, with chemotherapy. We’ve done a number of neoadjuvant chemotherapy trials, and I don’t think we’ve seen that same effect that was seen in that trial. So that’s quite impressive. And kudos to the CheckMate 816 surgical guys/gals for doing that presentation.

Martin Dietrich, MD, PhD, describes his interpretations of the KEYNOTE-799 study of pembrolizumab plus chemoradiation therapy for unresectable stage III non–small cell lung cancer. 

Roy Herbst, MD, PhD, Yale Cancer Center

Articles

Blood-Based Biomarkers in NSCLC

Perioperative Approaches to NSCLC

Biomarkers for Immunotherapy in Resectable NSCLC

Atezolizumab After Adjuvant Chemotherapy in Resectable NSCLC

Clinical Endpoints and Neoadjuvant Therapy for Resectable NSCLC

Neoadjuvant Immunotherapy for Resectable NSCLC