Based on 2 case scenarios, H. Jack West, MD, discusses treatment options and the data that support these options for the management of squamous and nonsquamous non–small cell lung cancer.
H. Jack West, MD
H. Jack West, MD
During aTargeted Oncologylive case-based peer perspectives program, H. Jack West, MD, discussed treatment options and the data that support these options for the management of squamous and nonsquamous nonsmall cell lung cancer (NSCLC). West, director of the Thoracic Oncology Program at the Swedish Cancer Institute in Seattle, Washington, explained disease management and the factors that go into treatment decisions during the dinner event based on 2 case scenarios of patients with NSCLC.
An 81-year-old man presented to his physician with symptoms of coughing, dyspnea, upper back pain, and fatigue requiring frequent rest. His medical history was significant for hypercholesterolemia, which was controlled on pravastatin; hypertension, which was controlled on verapamil; and psoriatic arthritis, for which he was not on treatment for at least 3 years. He was a former smoker who was physically active and played golf weekly. His ECOG performance status was a 1.
Imaging included a chest CT that revealed a 2.5-cm solid mass in the left upper lobe and lymphadenopathy in the left hilar and bilateral mediastinal nodes and bilateral, small pulmonary nodules, with the largest one 8 mm. PET/CT imaging also showed18F-FDG uptake in the lung mass, left hilar and both mediastinal lymph nodes, as well as in the thoracic spine (T5/T6).
Bronchoscopy and transbronchial lung biopsies were performed. Pathology indicated that he had grade III squamous cell carcinoma. His PD-L1 expression levels were also tested by PD-L1 immunohistochemistry 22C3 and he had a tumor proportion score of 65%.
The patient was diagnosed with stage IV squamous NSCLC.
Targeted Oncology: Is there a potential role for radiation therapy to address his symptoms of dyspnea or back pain?
West:One thing I am thinking about a little more is radiation, maybe as an immune sensitizer anyway. I would not necessarily use it gratuitously, but if I am waiting for 2 weeks for the patient to get their testing back, I might have a low threshold for them to pursue palliative radiation. If that is done and helps them with getting a better immune response later, so much the better, even if we do not know it is true.
Targeted Oncology: How do his older age and good performance status factor into this choice?
West:As a general rule, I would say we are learning that chronologic age is much less important than performance status. We will talk about the questions about chemoimmunotherapy versus immunotherapy/immunotherapy.
Targeted Oncology: Do precautions need to be taken considering his psoriatic arthritis?
West:Yes, we should always be mindful of immune-mediated issues, but there is a difference between Guillain-Barré syndrome and psoriatic arthritis. I have a patient [with psoriatic arthritis] whom I have been treating with durvalumab [Imfinzi]. We monitor it, and it has not changed at all. My early signal based on my huge “n-of-1” outlook is that it is OK.
Targeted Oncology:What are some clinical trials that play into your decision to use pembrolizumab?
West:[KEYNOTE-024] is the trial that kind of started it all for first-line therapy. It seems old and quaint, but it’s still just 3 years old.1This was presented in late 2016, with theNew England Journal of Medicineprinting it at the same time, and a couple of weeks later, the FDA approved pembrolizumab [Keytruda] based on this study. [It was] highly positive. [It enrolled] 305 patients with any histology, but they had to have PD-L1 expression ≥50%, which accounts for about 28% to 30% of the patients. They were randomized to platinum doublet of the clinician’s choice, which is histology based, or pembrolizumab every 3 weeks. It was a 2-year maximum, and patients in the control could cross over.
[The results] showed a huge difference in progression-free survival [PFS] and marked improvement in overall survival [OS], and objective response rate [was] 17% higher. Every endpoint [looked] better. It led to a very rapid FDA approval and became the standard of care.
There was a 2-year follow-up just reported in theJournal of Clinical Oncologyshowing that it holds up very well.2It is still phenomenally superior, and this is clearly a treatment of choice.
A different study, KEYNOTE-042, was at the plenary session at the American Society of Clinical Oncology [ASCO] 2018 Annual Meeting, and it was reported as positive.3This is the same kind of study but much larger, with over 1200 patients. It compared histology-appropriate chemotherapy versus pembrolizumab, but here it allowed patients with PD-L1 expression all the way down to 1%, so just expanding and casting a wider net. Among [the] PD-L1positive [patients], half have high PD-L1 expression and half have low expression, and most of those are below 20%. There are not many people who are 40%; they are usually <10%, just over 10%, or >50%. The [results of the] trial showed that pembrolizumab was better for OS in all the groups. If you include patients with 50%, 20%, or 1% expression, yes, it is positive. But the problem is that this includes these patients. It is like Russian nesting dolls. It is not fair because this is not really looking at the lows. Everyone [in the trial] appears to benefit…and what is actually happening is that this is still positive even when you dilute [the trial arms] with people who do not benefit.
When you look at patients who have 1% to 49% PD-L1 expression, they do not benefit. You could say you get the same overall survival (OS), and you do not have to give chemotherapy. That is not entirely true. They do not do as well. One thing about the trial is that there is no potential to cross over. There are only 20% of patients started here who ended up finding a way to get immunotherapy. In the United States when this trial was being conducted, it was standard of care to give immunotherapy in the second line. This group got substandard treatment in our world, and they ended up doing as well as the other arm. I would say pembrolizumab is basically looking as good as a hobbled chemotherapy.
When you look at PFS, it does not look good. Basically, the hazard ratio [HR] is 1.07 in [patients with PD-L1 expression ≥1%], and that is with the benefit of the highPD-L1 patients baked in. When we look at response rates, the responses last longer, but again, the response rate is higher with pembrolizumab in the high-expression group. By the time you get to patients with low expression, there is no benefit. My conclusion…was that this was not a fair study. This was a study where the hand was on the scale in favor of it, and it barely looked good even cheating.
The challenge wasthat [study] had all the fanfare; that was in the plenary session—this study, KEYNOTE-407, was hidden from view.4It was the dark horse, and the reason was the investigators did not have the results submitted to ASCO until late May, so they put it in some session that nobody knew about during the poster sessions for lung cancer. Nobody saw this, but it was actually the “sleeper” that was the important study. This is very similar to KEYNOTE-189 [NCT02578680]. This tested carboplatin and a taxane. It could have been standard paclitaxel or nab-paclitaxel, and patients received either placebo or pembrolizumab with it. There was crossover allowed. Maintenance therapy included pembrolizumab in the experimental arm and supportive care in the other arm. This was significantly better for OS. This was an interim analysis, so the trial stopped right before ASCO, and these data were released. It kind of slipped under the door. This was very important but just underappreciated. The HR for OS was 0.64, so very impressive there.
All the groups benefited; it did not matter whether they got paclitaxel or nab-paclitaxel. This is also something that really impressed me. When you look at the different subgroups by PFS, the HR for highPD-L1 expressers is 0.64, but for the low expressers, it is 0.61. The [overall] PFS HR is 0.56, a good number. PFS by [PD-L1 expression subgroup from lowest to highest expressers] is 0.68, 0.56, and 0.37; every group is a winner here. Again, you have a marked difference of a 23.4% higher response rate in the recipients of the triplet compared with the doublet. This was the only thing that was reported in the abstract, that there was an improvement in response rate. I actually generally feel response rate is not as important as survival, but when we saw that number of 23%, I thought, Well that is certainly a big difference. It is certainly on my radar, but the OS and PFS were not included in the abstract, but boy, did they impress at the actual meeting.
Targeted Oncology:When would you use atezolizumab?
West:Another complicated study compared carboplatin and nab-paclitaxel in the standard arm versus carboplatin, nab-paclitaxel, and atezolizumab [Tecentriq] or carboplatin, paclitaxel, and atezolizumab in the population [with squamous disease] [IMpower130]. PFS met the endpoint [6.3 vs 5.6 months], and the HR was 0.71. The benefit was there across the different PD-1/PD-L1 subgroups. In OS, it falls short. The HR is 0.96. Some charitable people say you can see the Kaplan-Meier curves separating toward the end, so maybe with more follow-up, [the OS will improve]. To which I would say: Is there any world in which this curve is going to be better than the one we saw for KEYNOTE-407? This is just not that impressive.
Here is an odd situation: You have PD-L1 lows; here you have the chemotherapy arm better than the atezolizumab arm, and that’s different from the negatives or the highs. It’s just bizarre. This is where we get into the differences of staining. It has to do maybe with the assay, but otherwise it’s inexplicable that you could have the negatives and the highs do totally different from the lows.
Targeted Oncology:What are your recommendations for each clinical situation?
West:I have clear conclusions here. I would say today for stage III disease, durvalumab for a year is the standard of care for patients who have not progressed and have the performance status to tolerate it without prohibitive toxicity. For stage IV disease, even with all the data that we have, I would say you check for PD-L1 and you check for driver mutations. If they have a driver, that takes priority; they receive an agent that treats that first. If they do not have a driver and high PD-L1 expression, monotherapy is going to be very compelling.
For everybody else, low or negative expressers, it is going to be chemotherapy plus pembrolizumab, and it is going to be histology based. I would use the KEYNOTE-189 regimencarboplatin and pembrolizumab—for the patients [nonsquamous disease] and carboplatin plus paclitaxel or nab-paclitaxel [for the population with squamous disease]. I personally use nab-paclitaxel, but there are no wrong answers there with pembrolizumab.
The only asterisk I would add is if patients are clearly decliningand we see patients who have lost weight in just the 4 weeks since they initially presented to their primary care physician and they are about to start therapy—[and] if their performance status is declining during their work-up, those are patients whom I might be more inclined to give chemoimmunotherapy, even if they have high PD-L1 expression. You might say if their performance status is declining, they cannot take it. My thought is you cannot hold any stops. You basically have to throw everything at them. They may or may not respond to the immunotherapy, and they may have a better response rate to chemoimmunotherapy. You may not know which the active drug is, but if you can do anything to pull them out of the tailspin, you probably only have 1 shot to do that. If they are progressing like this, they are going to be in a wheelchair before they can get chemotherapy. Those are the folks whom I do chemoimmunotherapy for, even if they have a performance status of 60 or 70.
A 63-year-old man presented to his primary care physician with intermittent cough and difficulty breathing on exertion. His medical history was significant for hyperlipidemia, which was well managed on simvastatin; hypothyroidism, which was managed on levothyroxine; and chronic obstructive pulmonary disease, which was managed with inhalers. He was a former smoker with a 40 pack-year history who recently quit smoking.
On physical exam he demonstrated intermittent wheezing. His ECOG performance status was 1.
Imaging included a chest CT, which revealed a 3.1-cm spiculated mass in the right upper lobe and 2 enlarged right mediastinal lymph nodes measuring 1.5 cm and 1.7 cm. Moderate emphysema was also noted. PET/CT imaging confirmed a lung lesion and mediastinal lymphadenopathy without any evidence of distant metastases. An MRI of the brain was negative for metastases.
His creatinine clearance was within normal limits. His pulmonary function tests showed an FEV1 of 1.2 L and a DLCO of 35%.
Bronchoscopy with transbronchial lung biopsy and lymph node sampling revealed adenocarcinoma with positive nodes in stations 4R and 7; level 4L was negative. The patient was diagnosed with T2aN2M0, stage IIIA NSCLC. Based on the extent of mediastinal disease and emphysema, the patient’s cancer was deemed inoperable.
Targeted Oncology: In a patient with stage III adenocarcinoma, are you routinely testing for molecular biomarkers or PD-L1 expression as you would in patients with stage IV disease?
West:If you were to ask 20 experts on this, you would get a pretty fair split on this subject. It is murky, and there is no guidelines-based evidence to say there is a clear standard of care in this situation. I, myself, am not that inclined to do it because if I do not know what to do with that information, it presents a challenge. That said, it is not rare for someone to order it along the way, and if it is available, you may use it. It may or may not be relevant [depending on the clinical situation]. There is debatable utility for it.
He underwent therapy with cisplatin/etoposide and concurrent thoracic radiotherapy. Follow-up imaging showed a partial response with shrinkage of the primary and nodal lesions.
Targeted Oncology: What are the treatment options for the patient?
West:This patient ends up receiving cisplatin and etoposide with concurrent thoracic radiotherapy. This patient in my practice would be getting weekly carboplatin plus paclitaxel for 7 weeks.
Targeted Oncology: What further treatment would you give past chemoradiotherapy based on results of the PACIFIC trial?
West:In the PACIFIC trial, they started people on durvalumab within 6 weeks of chemoradiotherapy. This trial really changed a lot. It was presented late in 2017 at the European Society of Medical Oncology Congress, and previously, there were 10 to 15 years of nothing happening in stage III disease. This was a global study that was liberal in the [clinician’s] choice of what induction could be rendered before durvalumab. You just had to have some regimen that was platinum based for 2 cycles minimum, and about 20% to 25% of the people had induction before induction chemoradiation. Carboplatin plus paclitaxel was the most commonly used regimen in about one-third of patients. There were many choices. Second-most common was cisplatin plus etoposide.
It was a 2:1 randomization [to durvalumab or placebo] in patients who did not progress after chemoradiotherapy and did not have prohibitive toxicities that dropped their performance status; stratification was done by age, sex, and smoking history. Co-primary endpoints were PFS by blinded independent central review and OS.
The updated results did not change much [from initial presentation]. These results were presented by Scott J. Antonia, MD, PhD, [of the Moffitt Cancer Center in Tampa, Florida], and subsequently published in theNew England Journal of Medicine.5There was almost a 3-fold increase in median PFS [17.2 vs 5.6 months]when you look at a landmark, like 1-year PFS, there is nearly a 20% difference [55.7% vs 34.4%], and there is a doubling of the 18-month PFS [49.5% vs 26.7%]—with a HR of 0.51. These are huge differences that are sustained over time.
When this occurred, we did not have an OS reported, but I think most of us who are thoracic focused thought this was a big [enough] difference…to change our standard of care. The FDA agreed and approved this agent before the OS results were reported.
Targeted Oncology: Which disease types benefit from the addition of durvalumab?
West:Subgroup analysis of PFS shows both squamous and nonsquamous disease benefit from durvalumab. The trial investigators designed PACIFIC to look at patients with PD-L1 expression of <25% versus ≥25%; both groups benefited. In this analysis of PFS, patients with allEGFRmutation status had positive results. It seems as if, regardless of age, sex, smoking status, and prior response, all participants benefit.
The OS data were presented at the International Association for the Study of Lung Cancer 19th World Conference on Lung Cancer in Toronto, Canada, and subsequently published in theNew England Journal of Medicine.6There was a significant improvement in OS. Most of the patients are followed for 2 years, and after that, there is approximately an 11% improvement in OS [66.3% vs 55.6%]. The HR is 0.68. The Kaplan-Meier curves maintain the separation that begins 1 month into treatment and is magnified over the first year.
Targeted Oncology: Do patients with low PD-L1 expression benefit from durvalumab?
West:There was another subset analysis, and I think this is a setting in which there is open question depending on your beliefs. In the subset analysis that was required by the European Medicines Agency, patients with a PD-L1 expression <1% were included. That group seems to not do as well as far as PFS and OS. It is on the unfavorable side. The European regulatory agency has not allowed durvalumab for these patients.
There is some debate about whether that is appropriate or not. This is a group of patients who were included in a trial that was positive. In general, when we see a subset analysis of a negative trial and there is 1 group who comes out favorable, we do not ignore the negative trial and just treat those patients [who saw a benefit]. It is a post hoc analysis in a subset of patients. By the alternative view, you do not necessarily preclude 1 group with an overlapping confidence interval who benefited less. Some clinicians would consider not giving durvalumab to patients who have test results that indicate PD-L1 expression <1%.