Corey J. Langer, MD, detailed the best courses of treatment for a patient with non–small cell lung cancer to a group of physicians during a recent Targeted Oncology live case-based peer perspectives discussion.
Corey J. Langer, MD
Corey J. Langer, MD, detailed the best courses of treatment for a patient with nonsmall cell lung cancer (NSCLC) to a group of physicians during a recentTargeted Oncologylive case-based peer perspectives discussion. Langer, director of thoracic oncology, Abramson Cancer Center, and professor of medicine, Hematology/Oncology Division, University of Pennsylvania, reviewed the treatment options based on a case scenario of one patient with stage III NSCLC and another study of a patient withEGFR-mutant stage IV NSCLC.
A 63-year-old man presented to his primary care physician with intermittent cough and difficulty breathing on exertion. He had a medical history notable for hyperlipidemia, which was well managed on simvastatin; hypothyroidism, which was managed on levothyroxine; and chronic obstructive pulmonary disease (COPD), which was managed on inhalers. He had recently quit smoking but had a 40-pack-year history. He demonstrated intermittent wheezing on physical exam and had an ECOG performance status of 1.
His blood work showed that his complete blood count, chemistries, and creatinine were all within normal limits. A chest CT revealed a 3.1-cm spiculated right upper lung (RUL) mass, 2 enlarged right mediastinal lymph nodes measuring 2.5 cm and 1.7 cm, and moderate emphysema. A PET scan confirmed the lung lesion and mediastinal lymphadenopathy without evidence of distant metastases. A brain MRI was negative. Pulmonary functions test showed a forced expiratory volume in 1 minute (FEV1) of 1.2 L and a diffusing capacity of carbon monoxide (DLCO) of 35%. A bronchoscopy with transbronchial lung biopsy and lymph node sampling revealed adenocarcinoma at primary site with positive nodes in stations 4R and 7; level 4L was negative. He was diagnosed with NSCLC T2aN2M0, stage IIIA.
Based on the bulk and extent of mediastinal disease and active emphysema, the patient’s cancer was deemed unresectable, and he was referred for consideration of concurrent chemotherapy and radiation.
TARGETED ONCOLOGY: Please describe the patient’s case, and detail the chosen course of therapy.
Langer:A 63-year-old man presented to his primary care physician with a cough and difficulty breathing on exertion. His medical history was notable for COPD, which was managed with inhalers. He has hypothyroidism, which was managed with levothyroxine, and hyperlipidemia, which was managed with simvastatin. He had a good ECOG performance status of 1, and his labs were good.
His chest CT showed a 3.3-cm spiculated RUL mass with 2 enlarged right mediastinal lymph nodes and moderate emphysema. The PET scan confirmed a lung lesion and mediastinal lymphadenopathy without evidence of distant metastases. Brain MRI was negative. The FEV1 is 1.2 L, with DLCO on the marginal side at 35%.
The patient underwent therapy with cisplatin and etoposide and concurrent thoracic radiotherapy.
TARGETED ONCOLOGY: Do you typically order genetic testing for locally advanced lung cancer to help select the best course of chemotherapy?
Langer:The answer is that is does not matter. The NCCN [National Comprehensive Caner Network] guidelines cover every single thing.1There are data for carboplatin and pemetrexed [Alimta] or cisplatin and pemetrexed. It is no worse than cisplatin and etoposide, but it is probably no better. It is more expensive but maybe a little less toxic. And then, of course, there is the paclitaxel plus carboplatin weekly regimen.
I had to do a talk on this recently and the question was, “Which is the best regimen?” I said that if you go to the VA [Veteran’s Affair] database or the SEER [Surveillance, Epidemiology, and End Results Program] database, the curves are all superimposable. Obviously, we can give carboplatin to a lot of [patients] who cannot tolerate cisplatin.
CASE 1 (continued)
The patient underwent therapy with cisplatin and etoposide and concurrent thoracic radiotherapy.
Follow-up imaging showed a partial response with shrinkage of the primary and nodal lesions. No evidence of distant spread was observed.
TARGETED ONCOLOGY:Could you describe the efficacy benefit observed in the PACIFIC trial and its applicability to this patient?
Langer:[The PACIFIC trial] gets updated about every 6 to 12 months. It was a 2:1 randomization of durvalumab [Imfinzi] versus placebo following chemoradiation. The problem is they were not registered at the time they started. It was after they finished chemoradiation. All of us know that there is a lot of attrition that goes on [in that time]. Patients have toxicity, or they do not make it. Some patients progress; they have closet stage IV [disease], with brain metastases and bone metastases suddenly appearing toward the end of chemoradiation.
They tried to get tissue and did not stipulate that it had to be PD-L1 positive. Durvalumab was given every 2 weeks for up to 12 months. We have all seen the PFS [progression-free survival] curves; this was astounding and unprecedented [with a median PFS of 17.2 vs 5.6 months with placebo (HR, 0.51; 95% CI, 0.41-0.63)].2We have never seen that sort of separation before. But again, it is not the same population [as was seen in other trials with a similar patient population]. That 20% or 30% of patients who would have not made it to this point are not included in that PFS analysis.
TARGETED ONCOLOGY: What are some aspects about the PACIFIC trial that stand out to you?
Langer:The other problem is the placebo is underperforming. I think in this population, we expect it to be 9 to 12 months, not 5 to 6 months. In the context of this trial, there was a 3-fold improvement observed. The hazard ratio is just incredible at 0.51. It was on this basis that it was approved, even before the OS [overall survival] data came through.3
Across the board, regardless of gender, age, smoking status, disease status, tumor histologic type, prior response to therapy, or PD-L1 status andEGFRstatus, there is a benefit.4
Patients were getting these scans a lot sooner than we normally do. In a traditional practice, we do it at about 6, 8, or 12 weeks. The investigators were doing it within 2 weeks to screen out [patients] who had progressed. Those patients who had their last radiation prior to randomization of less than 14 days [durvalumab, n = 120; placebo, n = 62] had an advantage hazard ratio of 0.39.
I had a lot of trouble with that. I think it is a confounding variable. These are the [patients] who are fit and do not have any pneumonitis. They are over their esophagitis, so of course they are going to do well. These are the marathon runners. With other [patients], you see a lot more started after 14 days [durvalumab, n = 356; placebo, n = 175].
TARGETED ONCOLOGY: Please detail the OS data from PACIFIC.
Langer:The survival data were published about a year ago.2Until I saw these data, I was still on the fence about durvalumab. If you think about it, they were not required to get PET scans at the beginning of treatment. This was a global trial, and my concern was that there were a number of patients who might have had occult metastases. If you just do a [CT] scan, in 15% or 20% of patients, you are going to miss those metastases. And of course, they are going to do better on durvalumab than nothing. But in the context of this trial, it still shows phenomenal survival advantage, with a Pvalue of .0025. We have tried TKIs [tyrosine kinase inhibitors], different chemotherapy, docetaxel added to cetuximab [Erbitux], and pushing the dose of radiation to 74 Gy [in this setting].
But there is always an update. In the original data from theNew England Journal of Medicinepaper, they looked at the incidence of new lesions. It seems to have a protective effect in the brain even though these are big antibodies that do not cross the blood­brain barrier; but the effects are still getting in the brain.5
The update from the 2019 American Society of Clinical Oncology [ASCO] Annual Meeting [shows that the data are] holding up.6Of course, [the Kaplan-Meier curves] have dropped a bit, but we are starting to see plateaus in both arms. At 3 years, 57.0% versus 43.5% are alive. Again, the hazard ratio here is good, at 0.69 [95% CI, 0.55-0.86]. It is the standard now, at least in [patients] who qualify for it. Obviously, if [a patient] has bad autoimmune disease, cannot tolerate it for some reason, or was on steroids, they did not get on this trial.
My big concern is the PD-L1 subgroup analysis. If you look at the nonPD-L1 group [<1% expression], the hazard ratio is on the wrong end of the unity. The hazard ratio is 1.36, [according to Antonia and colleagues].2The FDA has approved this across the board. In Europe, they stipulate PD-L1 testing has to show 1% expression or higher, and that is based on this analysis. It is concerning, but we have to remember that it was a post hoc analysis; it was not a part of the original plans.
My approach now is if they have ≥1% expression, I will recommend it with enthusiasm. If it is 0%, I will recommend it with some degree of ambivalence. If they can understand [the statistical analyses], I will show them the data, and we will have joint decision making.
TARGETED ONCOLOGY: Would you start durvalumab in patients with knownEGFRmutations?
Langer:I have offered it. Some of my colleagues out west see a lot moreEGFRmutations. There is a [large] Asian American population where our practice is, probably 15% to 18%. But they are typically not getting it because they are worrying that if they progress and they need a TKI, there will be a lot of toxicity.
I had a patient in her 70s with a lot of comorbidities and a bad heart. Her tumor was right up against it. As soon as she heard about potential cardiac damage from the radiation, she said, “No. Are you kidding?” She had IIIa disease, but it was not bulky. We went around and around with her treatment. With the recommended chemotherapy, we decided to put her on a trial and limit the volume to the heart. We then realized that what we were doing was completely out of indication in the study.
Then I put her on pembrolizumab [Keytruda], and she did well. She had little toxicity. She was on it for a year and a quarter. She almost had lymphoma, and she had lots of other medical problems. She had 80% PD-L1 expression and a complete response. Then it caught up with her. The thing about toxicity with immuno-oncology is that it may not happen in the first month or so. She started having diarrhea around month 15 and started being unstable. We started delaying the treatment [and] eventually stopped it, and now she feels a lot better. We are hopeful.
The bottom line is, do not do this at home. The standard treatment is still chemoradiation. But in exceptional circumstances, it is better than no treatment.
TARGETED ONCOLOGY: What are some of the toxicities associated with durvalumab treatment?
Langer:The toxicities in the PACIFIC trial are not much different in the durvalumab arm compared with the placebo. You do see some immune-mediated adverse events, at 24.2% versus 8.1% with placebo. Pneumonitis at any grade is 33.9% versus 24.8%. I have trouble wrapping my head around those percentages, but that is what they are. Again, this is a privileged population, and these are not patients who had pneumonitis going into the trial. If they had it, they were on steroids, and they would have been delayed.5
The 2019 NCCN guidelines list durvalumab as consolidation therapy for unresectable stage III NSCLC [nonsmall cell lung cancer], good performance status, no progression, and following at least 2 cycles of platinum-based treatment.1
A 66-year-old Caucasian woman presents to her primary care physician complaining of visual disturbances, nausea, fatigue, and sporadic headaches. Her medical history was notable for hypertension, which was managed on candesartan and hyperlipidemia, which was managed on simvastatin. She did not have a smoking history.
On physical exam her blood pressure was 148/70 mmHg and she had decreased breath sounds in the left lower lobe (LLL). Her blood work showed that her complete blood count and chemistries were within normal limits.
A brain MRI demonstrated a 10-mm right parietal mass at the graywhite junction with vasogenic edema. A CT of the chest, abdomen, and pelvis revealed a 3.4-cm mass in the LLL and several small liver nodules. A CT-guided transthoracic needle biopsy of the lung lesion showed grade 2 adenocarcinoma of acinar subtype.
She was diagnosed with NSCLC stage T2aN0M1c and had an ECOG performance status of 1.
TARGETED ONCOLOGY: What is your approach toward molecular testing for a patient with newly diagnosed metastatic NSCLC?
Langer:Penn Medicine is doing next-generation sequencing [NGS] in house. But if I have a high suspicion, I will order spot testingat least just forEGFR,ALK, andROSand they can turn that around in 2 or 3 days. But if you order the whole NGS, it takes 2 to 4 weeks.
The mandatory testing now [includes]EGFR,ALK,ROS1, andBRAF; we have an FDA-approved drug for each of those. Most physicians would strongly considerMEK; I thinkRETwill have 2 approvals in the next year. WithHER2[ERBB2], there are some interesting data with trastuzumab [Herceptin].
After 30 years, we may be starting to solve theKRASpuzzle. There was a clinical science symposium at ASCO for tumors withKRASG12C mutations, which is only about half of all the patients withKRASmutations. It is a new drug from Amgen; 5 of 10 patients had a [partial] response. The rest were at least stable, and the median time to progression had not been reached.7
InNTRK, there is larotrectinib [Vitrakvi] and entrectinib [Rozlytrek].
The recommendations for frontline testing are to offer a comprehensive panel includingEGFR, ALK, ROS1, BRAF, MET, ERBB2, RET, andKRASas the first option. The second option is to offer targeted testing withEGFR, ALK, andROS1, then checkKRASin a single-gene test and expand the panel toBRAF, MET, ERBB2,andRETif theKRAStest is negative.
The problem with the second approach is that it takes a lot of time, and you are requiring more and more tissue. When you do NGS, you can do it off a much smaller tissue specimen. When you are spot testing, you can eventually run out of tissue to look at.
CASE 2 (continued)
Testing showed 55% PD-L1 expression on her tumor cells by the PD-L1 IHC 22C3 pharmDx assay. Molecular panel testing showed that she had anEGFRexon 19 deletion.
TARGETED ONCOLOGY: Based on this patient’s molecular testing, would you give a checkpoint inhibitor?
Langer:NCCN says that there are contraindications for the checkpoint inhibitors in the presence of an oncogene, particularly in autoimmune disease. So if you know about both, you go with the TKI. The same holds true for second-line treatment.
The Society for Immunotherapy of Cancer says that the sequence for nonsquamous, advanced NSCLC and >1 actionable mutation is targeted therapy, chemotherapy, and a checkpoint inhibitor last.
Last year at ASCO during one of the plenary sessions, a study of patients with TKI-naïve,EGFRmutationpositive NSCLC and at least some degree of PD-L1 positivity was presented. They used pembrolizumab in the front line, and the primary endpoint was response. They were going to roll 25 patients, but they hung the trial at 11 due to the total absence of efficacy with the drug.8
What I think is happening with these mutations is a fake-out. The checkpoint inhibitors get to the cancer cell, but they just shake hands with the cell. They are not doing what they are supposed to do with the cell. There is something about oncogenic-driven tumors that makes them resistant.
[Of the] 11 patients accrued, 73% had ≥50% expression. These are patients whom we would have ordinarily treated with pembrolizumab up front. One of 11 had a response, but when they retested the patient, they did not have a mutation. There were 2 deaths within 6 months of enrollment, including 1 due to pneumonitis. One patient who got pembrolizumab and progressed went on a TKI, developed pneumonitis, and died.8
The take-home message is that toxicities during subsequent TKI use raised concerns about that sequence. Current evidence does support this approach.
TARGETED ONCOLOGY: Are there any studies that compare immunotherapy with other standard regimens among groups with EGFR-mutant and wild-type disease?
Langer:Based on a meta-analysis of the second-line trials, the hazard ratio for checkpoint inhibitors versus docetaxel is 0.69. [In patientswith EGFRwild-type disease], the hazard ratio is 0.67, and in the group with EGFR-mutated disease], it is 1.11. There is no survival benefit [in the latter group]. If anything, there was a slight advantage for docetaxel.9
In the group with wild-type disease, clinical trials OAK, CheckMate 057, KEYNOTE-010, and POPLAR have an overlapping unity [with a hazard ratio] of 1.11.9
TARGETED ONCOLOGY: What EGFR inhibitors are best in this setting?
Langer:The FLAURA trial is the reason we now use osimertinib [Tagrisso]. It was a big trial, with almost 600 patients, testing osimertinib versus either erlotinib [Tarceva] or gefitinib [Iressa], targeting the major actionable mutations, exon 19 and 21, not the more obscure actionable mutations. The primary endpoint was PFS, and osimertinib blew it out of the water. From 3 to 6 months, you see a separation in the Kaplan-Meier curves for PFS. There was an astounding hazard ratio of 0.46 with a Pvalue <.001. Survival looks positive, with a hazard ratio of 0.63 and aPvalue of .007.10
The intracranial responses [are similar to those seen in patients without metastases]. Again, these are TKIs that can cross the bloodbrain barrier; this does not hold for all agents. You are not going to see similar responses with PD-1 and PD-L1 inhibitors.
In patients with ≥1 CNS [central nervous system] lesions who were evaluable for response, the objective response rate was 91% versus 68% [with osimertinib and standard EGFR TKIs, respectively]. For the full analysis of patients with any degree of CNS involvement, corresponding rates of response were 66% versus 43%. It is better than first-generation TKIs.11
Treatment discontinuation was less common than with first-generation agents.
Timing is everything. If [the ARCHER study] had been presented 4 years ago, we would be using dacomitinib [Vizimpro], but it came after FLAURA. This trial looked at dacomitinib versus gefitinib. There was PFS improvement, but it was not as impressive as in FLAURA. It was 14.7 months versus 9.2 months [HR, 0.59; CI, 0.47-0.74;P<.0001].12
In the context of this trial, it is impressive, and they did show a survival benefit. They make a bragging point that it is a TKI that has shown a survival advantage versus the first-generation [agents]. I think that will change, but this exists. The median survival of 34.1 months versus 26.8 months seems to be low for this population.
There is a benefit across the board, but there is a lot more toxicity. This is the most toxic TKI. There is diarrhea, paronychia, rash, and stomatitis. The gefitinib [arm] had a little bit more transaminitis and nausea.
CASE 2 (continued)
The patient received osimertinib 80 mg once daily. She experienced a good partial response.
She complained of headaches and worsening fatigue 10 months after initiation of osimertinib. A CT scan showed 3 new liver lesions and a brain MRI showed 1 new lesion. She had an ECOG performance status of 1.
TARGETED ONCOLOGY: What are your therapeutic options at this point?
Langer: The RELAY trial, which was presented at ASCO, looked at erlotinib up front plus ramucirumab [Cyramza]versus placebo. It was a big trial, with 449 patients. There was a major PFS benefit of 19.4 months [versus 12.4 months]. It has a hazard ratio similar to that of other trials, 0.59 [95% CI, 0.461-0.760]. There are a lot of zeros in the Pvalue [P<.0001]. There is more toxicity, including more bleeding, hypertension, proteinuria, with the addition of ramucirumab, but it is mostly grade 1/2 events.13
The IMpower150 trial was published in the New England Journalof Medicineby Mark Socinski. It had over 1200 patients, comparing what we commonly did in the past: paclitaxel, carboplatin, and bevacizumab [Avastin] either alone or with atezolizumab [Tecentriq] added or atezolizumab substituted for bevacizumab. [In] a key subgroup, the patients with EGFR/ALK-positive tumors, the OS hazard ratio was 0.54 [for the 4-drug regimen versus the bevacizumab triplet]. For the wild-type population, the hazard ratio for OS is 0.78. It is intriguing, but we have to remember that this is not checkpoint inhibition only. This is combined with an angiogenesis inhibitor.14