Adjuvant Treatment Strategies in Resectable Non–Small Cell Lung Cancer

Transcript:

Balazs Halmos, MD, MS: The improvement doesn’t just come in the neoadjuvant space. In the past, we all got used to adjuvant therapy. The patient who went to Dr [Nasser] Altorki’s clinical trial got surgery. After surgery, he was found to have occult nodal disease. There are still a large group of patients we need to consider for adjuvant therapy. Can I ask you to comment on developments in that adjuvant space, where we’ve seen tombstones as opposed to milestones in the past, such as radiation and VEGF targeting? So many things have failed, but recently, there’s some excitement, right?

Martin Dietrich, MD, PhD: We have 2 new studies that have helped us improve outcomes in the adjuvant setting. They are easier to interpret because they latch onto the standard of care that we previously used, which was predominantly surgery followed by adjuvant chemotherapy and then an immunotherapy approach.

The first one is the IMpower010 study [NCT02486718] with adjuvant atezolizumab vs best supportive care. It was a large study that Dr Heather Wakelee presented about 2 years ago. We saw interesting progress for tumors greater than 4 cm with PD-L1 positivity independent of mutational setup, especially in PD-L1–high tumors greater than 50%. We have seen some very significant improvement in disease-free survival and, more recently, in overall survival. The question still is what to do with those that we traditionally think of as not immunotherapy responses: the 1% to 49%’s, the EGFR, and the ALKs?

We do not see the same strength of signal for improvement for those subgroups. A lot of the biomarker lessons from the metastatic space have transferred into IMpower010. In those, an approval for PD-L1 greater than 1%. There has to be much more nuance placed on the biomarker setup to guide a potential benefit. Is it safe to hold? Will the outcomes really change with overall survival? Especially for the PD-L1 lows from 1% to 49%, this hasn’t been fully answered yet.

The [KEYNOTE-091/PEARLS] study [NCT02504372] was a little bit different. It was a large, randomized study looking at a similar patient population with a co–primary end point of analysis for PD-L1 greater than 50%, and then in the PD-L1 agnostic patient population for the adjuvant setting. To my surprise, some of the biomarker lessons we had before were turned upside down. We saw a PD-L1 agnostic benefit for PD-L1–negative lows and highs, and staggered and certainly improved for the PD-L1 highs. This [then ended with] an approval for all-comers’ PD-L1 levels.

This is one where I have difficulties and would love to hear my colleagues’ thoughts on it. Where should we group in the PEARLS study for treatment? Both of them have chemotherapy requirements, so at least 1 cycle of chemotherapy is mandated by the label. It was a smaller subset [that was] not felt to be sufficient for full analysis. I think that patients who are eligible in this setting derive an overall survival benefit from adjuvant chemotherapy. So if an adjuvant therapy approach is chosen, I think it should be offered and evaluated for adjuvant chemotherapy eligibility. I do not want to take any shortcuts here for treatment of patients who have the option to skip to immunotherapy directly. I offer this option to all of my patients in PD-L1 levels greater than 50% if the biomarker setup excludes EFGR or similar oncogenic drivers. For those with levels 1% to 49%, I am still on the fence and would have to have a very detailed discussion with the patients. I certainly have not thought about this for PD-L1 negatives. We have a real tumor sample, so we get a pretty representative idea of how PD-L1 looks. If you have tumor sample [that is] PD-L1 negative, in my practice, I would not consider adjuvant immunotherapy.

Balazs Halmos, MD, MS: A terrific summary. You brought up so many great points on the controversy between the 2 study results. It’s very hard to read. Looking at the discrepancies between the 2 studies, are you sticking with the known biology and logic, or are you willing to yield to the broader approval of pembrolizumab for the TPS [tumor proportion score] agnostic, including zero-patient population?

Joshua Reuss, MD: There’s no right answer to this question. I think patient-informed discussions are important here. The data is the data at the end of the day. However, in the IMpower010 study, it does appear that the PD-L1–high subgroup is really deriving that benefit, so I would agree with Dr Dietrich [about] atezolizumab’s clear benefit in the PD-L1–high population. I am inclined to recommend pembrolizumab for those with 1% to 49% levels in the adjuvant setting, and particularly when you’re talking about lymph-node positivity, occult N2 lymph nodes, where your likelihood of recurrence is very high. I do not know if we should be looking at this as we do with immunotherapy monotherapy in the metastatic setting where the IO [immunotherapy] is in a vacuum.

If you are looking at chemo then immunotherapy, what is the best comparison there? Is it like a KEYNOTE-189 [trial regimen]? I know that’s combination [therapy]. It is not sequential. But maybe we can’t create an analogy to the IO monotherapy studies in the metastatic setting. The PD-L1–negative population is definitely perplexing. I am inclined to have the discussion and offer the therapy to someone who has high risk for recurrence, a large primary tumor, or occult N2 disease involvement.

People may ask what I am going to do when patients recur. Will I be utilizing these chemotherapy chairs for patients who may not need it? My argument there is whatever we can do to prevent a recurrence and to prevent someone from having metastatic disease, whatever we can do to promote cure, trumps what we’re going to do in the metastatic setting.

Kristen A. Marrone, MD: My clinical practice is similar in that we use PD-L1 [grading]. I do not consider atezolizumab unless the PD-L1 is above 50%. If we are going toward using neoadjuvant immunotherapy, how can we actually interpret this data? Could you use pathCR [pathologic complete response] to determine if adjuvant therapy is indicated? I have used neoadjuvant and adjuvant equivalently. It might a be bias of my institution, but in general, that is what we have clinically seen play out in terms of how we have used these FDA approvals in our clinic so far.

Balazs Halmos, MD, MS: Both approaches are here to stay, so all of us need to stay comfortable in terms of learning by literature and adapting it to the patient in front of us. Martin made a very good point in terms of the surgical specimen providing a great sample for TPS testing. In fact, at our tumor board, we have decided never to accept the presurgical cytology for biopsy when making adjuvant decisions. But we are doing it under another surgical specimen. In reality, what’s the difference between 0% vs 1%? If you ask a pathologist, they can make anything 1%, if you ask them todo that. There is a lot of relianceon a technology that’s not perfect for this context, but it is the best we have. That’s just the reality.

Transcript edited for clarity.