Over the past year, there has been a multitude of exciting developments in the treatment of patients with non-small cell lung cancer
Corey Langer, MD
Over the past year, there has been a multitude of exciting developments in the treatment of patients with non-small cell lung cancer (NSCLC): Anti-PD-1 and PDL1 immunotherapies in lung and other tumors made a big splash at ASCO 2013 and continue to generate tremendous excitement in the field; guidelines for molecular testing in lung cancer were published in the Journal of Thoracic Oncology; second-generation ALK inhibitors have shown promise in patients who become resistant to crizotinib (Xalkori); and afatinib (Gilotrif) and erlotinib (Tarceva) have been approved in the first-line metastatic NSCLC setting. For expert analysis of these critical issues, we sat down with Corey Langer, MD, at the 14th Annual International Lung Cancer Congress. Langer is the director of Thoracic Oncology at the Abramson Cancer Center and a professor of Medicine in the Hematology/Oncology Division at the University of Pennsylvania.
OBTN: Can you discuss the breakthroughs with immunotherapy in lung cancer?
Langer: I scarcely consider myself an expert in immunotherapy but this is actually a very exciting field, even though, historically, I’ve categorized myself as an “immuno-skeptic” based in part on my past experience over 20 years ago with highly toxic immunotherapy agents. We conducted trials in that era looking at those agents, including IL-2 and beta-interferon, not just in melanoma and kidney cancer, but in non-small cell lung cancer; and by and large, very few patients responded and toxicity was horrible.
Now we are in a new era. Specifically, there are emerging data in advanced non-small cell lung cancer, particularly in squamous histology, where a phased approach—essentially grafting ipilimumab onto the third cycle and beyond of standard chemotherapy— looks better than standard chemotherapy alone. A randomized phase II study was reported last year and suggested about a 3- or 3.5-month improvement in median survival for patients receiving combination chemotherapy with ipilimumab (during the 3rd and subsequent cycles) versus chemotherapy alone and that has set the stage for a major phase III trial that will isolate the role of ipilimumab in patients with advanced NSCLC and squamous histology.
In the last 2 years, we have seen data for monoclonal antibodies targeting PD-1 and PD-L1. These are single-agent studies in heavily pretreated patients, many of whom have exhausted standard cytotoxic therapy—basically patients out of standard therapeutic options. Patient response rates have ranged from 18% to as high as 25% to 30%, with much less toxicity than we saw historically with interleukin-2 and beta interferon. The schedules of these agents are still being worked out. Whether these agents can combine easily with standard treatment either in the first line or second line is still being worked out as well, but you can bet there are and will be multiple randomized phase II, and more importantly, phase III registration trials that will establish whether these agents truly have a role. So far, at face value, they appear to have a major role.
There is some speculation that patient tumors need to harbor PD-L1 for these agents to be active. Certainly, response rates appear to be considerably higher in those individuals whose tumors express PD-L1, but responses are also seen in PD-L1—negative tumors, so I think the notion that we have a very dichotomous marker—if you have PD-L1 you’re eligible and if you don’t you can’t get these agents—that notion falls by the wayside.
Which patients with non-small cell lung cancer should receive molecular testing?
As far as I am concerned, any patient with advanced NSCLC with nonsquamous histology is a reasonable subject for molecular testing. Many groups, particularly academic centers, will test every patient regardless of histology. I think when it is tough to get adequate specimens, we have to be a little bit careful about applying this strategy generally or uniformly. Certainly, a patient with squamous histology, an older male patient, a heavy smoker, a non-Asian—that patient’s likelihood of having an actionable mutation or translocation is probably less than 1% to 2%. On the other hand, if they do have it, the presence of this molecular abnormality completely alters the nature of the patient’s therapy. We go from a conventional cytotoxic regimen to a potentially more active, less toxic, molecularly directed treatment. There is this ongoing debate now between uniform testing—global testing—versus more targeted testing. I would say at this point in 2013, the vast majority of patients who have advanced non-small cell lung cancer will be tested for actionable molecular markers if adequate tissue is available.
What about molecular testing for early-stage patients?
In 2013, at least in patients with N1 or N2 involvement or stage 1 patients with tumors larger than 4 cm, adjuvant platinum-based chemotherapy for four cycles constitutes the standard postoperative treatment. There is no proven role for a targeted agent. Certainly, it is a research question. We clearly hope that one or more agents will emerge. Should we test these folks routinely? Frankly, I think we should because unfortunately half or more will relapse, and I think it is very important to know their marker status should that inevitable day arrive for many of these patients, so we know whether they harbor an EGFR mutation, if the cancer recurs, particularly if it comes back in a site that is not readily able to be biopsied, it puts you ahead of the game. You know information that would otherwise take 2, 3, or 4 weeks to process. Plus there’s the difficulty of trying to retrieve an older specimen, which might have been obtained at another hospital or which otherwise might not be readily accessible. From a practical standpoint, although it is not immediately actionable, I have asked our thoracic pathologists to perform reflex testing on surgical specimens with the knowledge that it may not be used in the adjuvant setting, but it will often be necessary to know subsequently.
Which mutations are you testing for in your practice?
In our institution we are testing for EGFR, ALK, KRAS—not that the latter is yet actionable, but we have clinical trials that are targeting KRAS-mutant patients. We are also adding ROS 1 (which is a substantiated target of crizotinib), and we are looking for BRAF, since the melanoma drug dabrafenib has clearly shown activity in NSCLC patients whose tumors harbor the BRAF 600 mutation. We are also looking at RET and a variety of other potentially actionable markers. HER2 is relatively rare, but 1% to 2% of NSCLC patients will have that mutation, and again we have established agents that work in breast cancer, which may work in this group. It really opens up the therapeutic portfolio when we know the marker status of individual patients. (See Figure for a breakdown of actionable mutations in the lung cancer population.)
Can you talk about next-generation ALK inhibitors?
There has been quite a bit of intriguing progress made in this whole therapeutic realm of ALK inhibition, which is really quite amazing when you consider that only 4% to 8% of the population are candidates for this class of compounds; and yet we have four drugs now that clearly have activity. Crizotinib is the first in class; however, it may not necessarily be the best in class. We know response rates in those who harbor EML4/ALK translocation are in the 60% to 70% range. A major randomized phase III trial compared crizotinib to standard second- line chemotherapy with either pemetrexed or docetaxel, and for those with the ALK translocation, response rates were threefold higher in the range of 60% versus 20%. Progression-free survival was more than double: 7.7 months versus 3 months. Survival is still premature. If we extrapolate from the EGFR experience, we probably won’t see a survival advantage because ALK-inhibitor-naïve patients who get “standard” second-line chemotherapy and whose disease progresses are going to cross over to crizotinib at the time of disease progression.
Probably one of the most promising of the newer second-generation inhibitors is the Novartis drug LDK378, where much of the original testing was in the second-line setting in patients who had already been exposed to crizotinib. The response rates, at least initially reported, were about 80%. They have dropped down a little bit to about 70%, but we are taking about a heavily pretreated population, many of whom have already received multiple chemotherapy regimens and who have already received crizotinib. They have done well on crizotinib, often for quite awhile, but their disease progressed and they are being treated with a second-line ALK inhibitor.
We have no analagous situation with an EGFR inhibitor. Afatinib is the best of the group, and its response rate in this second-line setting in EGFR mutation- positive patients is 9% or 10%, and yet agents like the LDK drug in the ALK fusion, second-generation setting are generating response rates that are in the high double digits. This is really quite amazing. In addition, the newer agents can cross the bloodbrain barrier. We have seen individuals who have relapsed in the brain and who previously received radiation with subsequent intracranial progression, and some of these lesions shrink or even disappear. But these new agents are not without toxicity.
I have a patient who was actually rescued from death by crizotinib. She first presented to me in 2004-2005. She had what we called then “wet stage IIIb NSCLC” based on a malignant pleural effusion (today, we’d call this stage IVA or M1A). She received systemic chemotherapy with paclitaxel/ carboplatin/bevacizumab and did really well. She was maintained on bevacizumab for some time and ultimately progressed in the hilar area where her original primary was located. We gave her radiation at that point and also incorporated pemetrexed and she did well again. She was on second-line maintenance pemetrexed for 2 to 3 years and her pleural effusion never came back. Unfortunately, over time, about 3 or 4 years into her diagnosis, she developed progressive disease in the hilar area and she went through a whole series of therapeutic agents.
In 2010, she was really starting to get quite ill— short of breath, oxygen-dependent, huge mass in the hilum, multiple new pulmonary nodules in the contralateral lung. We knew she never smoked, we knew that she was EGFR wild type, but we did not have sufficient tissue at that time for additional testing. At around the same time, the ALK translocation story had just emerged and we knew it was far more likely for nonsmokers to have that translocation than the typical advanced NSCLC patient who smoked. The first bronchoscopy was nondiagnostic and she was referred to our institution for a repeat bronchoscopy for EBUS. We obtained sufficient tissue and lo and behold, she was ALK-positive. We placed her on the phase II trial of crizotinib and literally within weeks—less than a month—she went from performance status 2 to performance status 0. This was 2010. It is now July 2013 and this patient, who would have clearly died 3 years ago, has been rescued by a sequence of ALK inhibitor therapies.
Her initial response on crizotinib lasted a year; she then developed progression in the liver, which we treated with radiofrequency ablation [RFA] and ultimately resection, all the while continuing her on crizotinib. Eventually, she developed multifocal diffuse progression including the brain, the other lung, the pericardium and the pleura, and the retroperitoneum. At that point, she was enrolled in the LDK trial and had a fair amount of toxicity but did well. She re-responded; her liver mets actually shrank. Unfortunately, the LDK trial was not designed to allow continuation of therapy in the face of oligo-progression, even if we could remedy that with RFA or local stereotactic radiation; so at that point she was obligated to go off the LDK trial, and I was faced with a dilemma. The “new” ALK inhibitor trial which we had opened at our institution explicitly excluded patients who were exposed to both crizotinib and the LDK trial. So I went back to both crizotinib and pemetrexed. She had responded to each of these agents in the past, and had not been exposed to either agent for quite awhile. We have done this now for 6 months and her disease, thankfully, has been stable for that entire period. Her performance status, while not optimal, is reasonable. She is enjoying life. She has new grandchildren. This is what the treatment of advanced NSCLC is all about in 2013. This woman has lived far longer than she would have lived even 5 or 6 years ago, and it is largely thanks to molecular genetics and the new agents that target these abnormalities that we now have available. I would never have witnessed such a scenario when I started out as a thoracic medical oncologist in 1987. These approaches simply did not exist. Back then, we never had patients living many, many years with advanced metastatic disease.
In 2013, the FDA has approved both afatinib and erlotinib for the first-line treatment of patients with metastatic non-small cell lung cancer whose tumors have EGFR exon 19 deletions or exon 21 (L858R) substitution mutations. Which agent would be your treatment choice for a patient in this setting?
I think either one is a right choice. My choice in this situation, particularly with the very recent approval of afatinib in patients who harbor an actionable EGFR mutation, would be to put them on a clinical trial comparing the two agents. That trial does not exist. So, until I can easily get afatinib, my default option remains erlotinib.
As time goes by, afatinib may actually emerge as a “standard” second-line agent. There is a major randomized phase III trial recently reported by Vince Miller and colleagues that evaluated individuals, most of whom harbored a mutation, who had done well on gefitinib and erlotinib, and at the time of disease progression, they were randomized to afatinib or to a placebo. The afatinib group had a much higher response rate. Overall, they had a threefold improvement in progression-free survival, and if they harbored a mutation, they had a fourfold improvement in PFS. They did not have a survival advantage, but there was no mandate or directive regarding subsequent therapy in the control group. I think on that basis alone, afatinib should at least be offered as a standard second-line in patients who have already received erlotinib, and let’s face it, in 2013, by the very nature of the agents available, the vast majority of folks with an EGFR mutation who were diagnosed more than 3 or 4 months ago have already been exposed to erlotinib. So, for the time being at least, afatinib may become the de facto standard second-line option.
There are intriguing data from Memorial-Sloan Kettering and elsewhere combining afatinib with cetuximab, so essentially enacting global EGFR blockade. Response rates as high as 40% have been observed with progression-free survival in the 4- to 6-month range. And those are really the best data that I have seen in patients with an EGFR mutation who have developed acquired resistance. We have no similar data for erlotinib or gefitinib combined with cetuximab, so in that regard, afatinib, which is an irreversible EGFR inhibitor targeting both HER1 and HER2, may ultimately have an advantage over erlotinib and other first-generation EGFR TKIs. Whether or not it actually displaces erlotinib upfront is not going to be my decision. It may ultimately be an insurer’s decision, and it may be based on which agent is priced appropriately. I have total equipoise when it comes to this issue. Based on how these agents are given, based on their toxicity profiles, and based on my own experience with both agents, I could give one or the other equally to a newly diagnosed patient with an EGFR mutation.
These first-line approvals for afatinib and erlotinib were accompanied by the approval of a companion diagnostic. Some oncologists suggest that the FDA’s trend toward approving agents concurrently with diagnostics is not a good thing. What are your thoughts?
There is major debate about the optimal platform for genetic testing. Unfortunately, a lot of confusion exists, as many different companies test different agents that all target the same molecular abnormality, and they often have completely different assays. It is a little bit like the era of industrialization in the United States when railroad companies all had separate railroad lines and they used different gauge tracks with different widths. A railcar could make it from Boston to Baltimore, but could not go from Baltimore to Richmond, Virginia, because the track size had changed. We need uniform, mutually acceptable testing, but we also need some degree of flexibility. We need various companies who are in competition to at least accept a uniform approach that can be applied globally to our patients so that we are not dealing with different or mandatory companion diagnostics, particularly as the testing technology evolves and improves and we are able to test multiple mutations and translocations simultaneously. In this regard, I think both the FDA and NCI really need to join forces and come up with an acceptable approach that the industry can comply with, which does not handcuff us to specific “companion diagnostics.”
Over the past decade, the identification of driver mutations has drastically enhanced the treatment options for patients with lung cancer. Diagram A illustrates the projected histological profile per every 100 patients with stage IV lung cancer diagnosed today. Diagram B identifies patients in this population with mutations for which targeted therapies have been developed or are being investigated.
Scagliotti GV. Individualized therapy in lung cancer: where are we in 2013? Presented at: 14th International Lung Cancer Congress; July 25 -27, 2013; Huntington Beach, CA.