Feliciano Determines Which Patients Are Best Served by Either Immunotherapy or Targeted Therapy

During a <em>Targeted Oncology </em>live case-based peer perspectives presentation, Josephine Louella Feliciano, MD, explained the treatment considerations and decisions she makes when treating a patient with non&ndash;small cell lung cancer in the clinic and the data that support these options to a group of physicians.

Josephine Louella Feliciano, MD

During aTargeted Oncologylive case-based peer perspectives presentation, Josephine Louella Feliciano, MD, explained the treatment considerations and decisions she makes when treating a patient with non—small cell lung cancer (NSCLC) in the clinic and the data that support these options to a group of physicians. Feliciano, an assistant professor of oncology, The Johns Hopkins University School of Medicine, Johns Hopkins Kimmel Cancer Center, and medical director of the Thoracic Oncology Program and co-director of the Outpatient Oncology Clinic at Johns Hopkins Bayview Medical Center, discussed these treatment options with the group in relation to 2 case scenarios of patients with NSCLC.

Case 1

A 63-year-old man presented to his physician with intermittent cough and difficulty breathing on exertion. His past medical history included hyperlipidemia, which was well managed on simvastatin; hypothyroidism, which was managed on levothyroxine; and chronic obstructive pulmonary disease, which was managed on inhalers. He had recently quit smoking but had a 40 pack-year history of smoking.

On physical exam he demonstrated intermittent wheezing and had an ECOG performance status of 1. He had a creatinine clearance that was within normal limits.

A chest CT 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. A PET scan confirmed the lung lesion and mediastinal lymphadenopathy without evidence of distant metastases. A brain MRI was completed and was negative.

Pulmonary function test showed a forced expiratory volume in 1 second of 1.2 L/second and diffusing capacity of the lungs for carbon monoxide of 35%. A 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 staged with T2aN2M0, stage IIIA disease.

Based on the extent of mediastinal disease and emphysema, the patient&rsquo;s cancer was deemed inoperable, and he was referred for consideration of concurrent chemotherapy and radiation.

Do you typically order genetic testing for locally advanced lung cancer?

No, this is not a typical practice in the community. If the patient has metastatic disease, genetic testing is usually ordered.

What are the treatment options for this patient? Neoadjuvant chemotherapy? Neoadjuvant chemoradiation? Chemotherapy and concurrent radiation? Chemotherapy and sequential radiation?

The patient is already deemed to be inoperable. From the case, we know he is not a surgical candidate, so we would not be considering neoadjuvant chemotherapy. Because of that, he is also not a candidate for neoadjuvant chemoradiation. In some robust patients who are surgical candidates, we do consider trimodality therapy for stage IIIA disease. But the standard of care for a patient who is inoperable is chemotherapy with concurrent radiation. The survival benefit, regardless of newer data, for concurrent therapy is much better than sequential therapy.

Are you ever recommending surgery in patients with stage IIIA disease?

It depends. We typically see our patients in a multidisciplinary clinic, and if someone is operable in terms of cardiac function or pulmonary function, as well as anatomy, and the surgeon deems them operable, then we consider trimodality therapy with chemotherapy and radiation, followed by surgery. Typically, it is the nonbulky stage IIIA tumors. Incidental IIIA tumors are candidates for mediastinoscopy or endobronchial ultrasound. We are not usually doing this procedure on bulky disease.

We do a fair amount of trimodality therapy. Some institutions are solely neoadjuvant chemotherapy, even for stage IIIA disease. For operable patients, these institutions will consider neoadjuvant chemotherapy in stage IIIA disease.

Case 1 (continued):

The patient underwent therapy with cisplatin/etoposide and concurrent thoracic radiation therapy. Follow-up imaging showed that the patient had achieved a partial response (PR) with shrinkage of the primary and nodal lesions.

Does the patient require further treatment, and how would you follow up with this patient after progressive disease?

Yes, the patient requires further therapy.

Something to consider that was not available a year ago are the data from the PACIFIC trial. These patients had stable disease [SD] or a PR. Your patient might progress after chemoradiation, and so at our institution, it is rather reflexive to order [genetic] testing so that we have it available if or when the patient progresses. But that is not standard, and there is no indication for a tyrosine kinase inhibitor [TKI] after chemoradiation in this patient if they had a response or stable disease.

How do the results of the PACIFIC trial influence your decisions here?

The PACIFIC study has been a groundbreaking study for us in oncology and radiation oncology. This was a randomized, phase III, double-blind study that was reported twice: first with the primary endpoint of progression-free survival [PFS],1and then we subsequently followed up with overall survival [OS] advantage.2

The patients who were included in the study had stage III, locally advanced, unresectable, non—small cell lung cancer [NSCLC] that has not progressed after definitive platinum-based chemoradiation.

These were patients who had SD or a PR. They were then randomized in a 2:1 ratio to receive either durvalumab [Imfinzi], 10 mg/kg every 2 weeks for up to 12 months [n = 473], or placebo, 10 mg/kg every 2 weeks for up to 12 months [n = 236]. They had to have good performance status and a life expectancy of ≥12 weeks; and the researchers used archived tissue that had already been collected. This tissue was not from a new biopsy or a biopsy after the chemoradiation had been completed, which I think is an important consideration.

The majority of patients did receive the durvalumab. The primary endpoints were PFS and OS, and secondary endpoints were response rates, duration of response [DOR], safety, and patient reported outcomes.

The primary endpoint of PFS was met and was significant at 17.2 months versus 5.6 months. This is important because of the question, “Does PFS translate into OS?” Remember that the objective is trying to cure patients, and fewer patients had progressed over the course of this trial. But we also see that OS was statistically significant and quite robust [HR, 0.68;P= .0025].

What the trial designers also conducted was a subgroup analysis in the intention-to-treat population. Remember that this is a subgroup analysis. The factors that were emphasized were that for the majority of patients, there appears to be benefit in favor of durvalumab. Where there may be some vagueness is in patients who haveEGFRmutations. But keep in mind that many patients have an unknownEGFRmutation status and this is a subset analysis. If you had a patient with anEGFRmutation who received chemotherapy and radiation, and now you are trying to decide if durvalumab is the next treatment to try, it is difficult to definitively come out and say [you should not use it in this patient] based on this data, because this is just a subset analysis and many patients [in PACIFIC] did not have the molecular testing results.

What is exciting is that the PFS with durvalumab does appear to translate into an OS. What is important is that the 24-month OS at 66.3% vs 55.6%, that is an absolute benefit of 10%. If you think about all the other cancers that you might treat—adjuvant breast cancer, for example—the absolute benefit is on the order of 3% or 2%. This finding is so important because this particular disease has a high fatality.

When you measure a patient for PR, do you use a CT or PET scan?

We are using a CT scan. In general, we do not tend to follow with PET scans because they can be hard to interpret after radiation. If we do order a PET scan, it is usually because we are concerned that the patient has progressed. We have some sense that the patient has progressed to advanced disease. For follow-up examinations, we are generally ordering a CT scan to confirm that the disease is stable or responding.

When you are treating the patient with durvalumab, how often do you order scans?

We are typically monitoring with CT scans, but with the routine complete blood count, chemistry, and thyroid function test because we are looking for signs of autoimmune toxicity.

What do the guidelines say about length of time to treat with chemoradiation?

Most of the guidelines for definitive chemoradiation are 3 to 6 months. Insurance companies typically allow for 3 to 4 months. Some will say only 6 months.

How does the patient&rsquo;s PD-L1 status affect your decision to use durvalumab?

There is concern that [with] <1% PD-L1 expression, there may not be a benefit. This does not take into account biomarkers like tumor mutational burden [TMB]. There were a lot of data presented at the American Society for Clinical Oncology 2018 Annual Meeting that indicated that TMB and PD-L1 status were independent of each other. I think that is an important point. TMB is not something that is routinely tested. Currently, only Foundation Medicine’s assay [FoundationOne CDx] tests for TMB. It is very challenging. If I had a patient with 0% PD-L1 expression, and I knew that up front, I would not necessarily withhold durvalumab just because of their PD-L1 status.

We have seen this in CheckMate 227 where we are looking at TMB and PD-L1.3The results in this trial indicated you can have low PD-L1 but high TMB and still have a good response to immunotherapy.

If you are considering a patient for clinical trials, we are having the patient undergo repeat biopsies because we are seeing the tumor dynamics. The microenvironment dynamics can fluctuate during treatment.

What about radiation and durvalumab?

There are also some questions about radiation and durvalumab. In subgroup analyses, radiation received <14 days before randomization appeared to have a more robust benefit from durvalumab versus those patients who received it ≥14 days before randomization.

I say with caution that it is possible that radiation can upregulate factors that boost a response from immunotherapy. However, I think it is possible that those who received radiation ≥14 days before were the robust patients to begin with, and so this has to be taken with a grain of salt.

We do try to begin the immunotherapy as soon as the patient has few side effects and toxicities from the chemoradiation, but I would not say the 14 days is a hard-and-fast rule.

One suggestion for this boosted response could be the abscopal effect, but there are a lot of hypotheses trying to explain radiation’s possible boosting effect.

What are some of the adverse events (AEs) associated with durvalumab in the PACIFIC trial?

The other important factors to consider are related to safety data. In general, serious toxicities, grade 3/4 toxicities, were relatively similar between placebo and durvalumab. The real areas of interest are radiation pneumonitis. What was showed was an increase in any-grade pneumonitis with durvalumab [33.9% vs 24.8%]; however, grade 3/4 pneumonitis was relatively acceptable [3.4% vs 2.6%].

From this study, the grade 3/4 toxicity is what you would expect from the chemoradiation. What is seen is that the addition of durvalumab did not significantly increase cases of serious pneumonitis. I think we will also learn more about that as more people who were not included in this study go on to receive durvalumab. I would expect we will see more real-world evidence of who is actually getting treated. I know the FDA has done analyses looking at who is actually getting immunotherapy for metastatic disease. They are finding that older patients and patients who have more comorbidities than the patients who were enrolled in the PACIFIC trial are getting these drugs.

You should remember that this was a healthy population in PACIFIC. Whether or not this translates into actual mortality or actual cases of pneumonitis remains to be seen.

I will say that at our institution—we have a study that is almost like a registry—we are having any patient who is receiving immunotherapy and being followed longitudinally have blood samples taken [and] microbiome examinations of stool samples. We are trying to determine if these toxicities are real and if can we predict who might experience these toxicities. There is a lot of interest in finding out what is actually happening to these patients.

Case 2

A 66-year-old Caucasian woman presented to her physician complaining of visual disturbances, nausea, fatigue, and sporadic headaches. His past medical history included hypertension, which was managed on candesartan (Atacand), and hyperlipidemia, which was managed on simvastatin. He had never smoked.

On physical exam he demonstrated a blood pressure of 148/70 mm Hg and left lower lobe auscultation revealed decreased breath sounds. His complete blood count was within normal limits.

An MRI of the head demonstrated a right parietal mass at the gray—white junction with vasogenic edema. A CT of the chest, abdomen, and pelvis revealed a 3.4-cm mass in the left lower lobe and several small liver nodules. CT-guided transthoracic needle biopsy of the lung lesion showed grade 2 adenocarcinoma, which was considered acinar.

Molecular testing revealed anEGFRexon 19 deletion. The patient was diagnosed with T2aN0M1c NSCLC and had an ECOG performance status of 1.

Is next-generation sequence (NGS) testing reflexive at your institution?

In general at our institution, if a patient has nonsquamous disease, the test is reflexive. But for a patient with squamous carcinoma who is a nonsmoker, I have to ask for the test.

What if PD-L1 testing comes back before NGS testing in a patient suspicious for a driver mutation?

I do see quite a few referrals where the PD-L1 came back first but the suspicion is high. I will not treat with immunotherapy in that setting if I have a high suspicion because there is concern about the safety of giving immunotherapy before a TKI, for example. Is there a risk of increasing the chance the patient develops pneumonitis? I do get nervous about that.

If the patient is symptomatic, though, we sometimes cannot wait 3 weeks to make a clinical decision in most of these patients.

Case 2 (continued):

The patient developed symptoms related to vasogenic edema.

How do you treat her now?

If the patient is symptomatic with vasogenic edema and there is 1 lesion, those are situations where I might consider neurosurgery followed by stereotactic boost.

One thing that we are finding, especially in a case like this with a nonsmoker, is we are trying to move away from whole-brain radiation unless the patient has a lot of disease. This is not only because of the length it takes to get the patient on systemic therapy afterwards but because of the quality-of-life and early dementia issues, especially in a nonsmoker with few comorbidities.

If you have a high suspicion for a mutation, I would argue that it has a high response in the brain, whether it’s anALKor anEGFRmutation. However, that delay in starting treatment when the patient becomes symptomatic can be fatal. We have definitely seen patients with an untreated brain metastasis [that] leads to bleeding before they can get started on treating the cancer.

What if the patient has an oligometastatic lesion? Is removing the primary lesion an option to consider?

I would say that is not standard practice. For example, I have a patient who has anALKmutation and had an area of progression that was symptomatic. I biopsied it, and he had a resistance mutation to alectinib [Alecensa]. We ended up using radiation at the site as local definitive therapy. In that time, his brain was riddled with metastases. We put him back on the alectinib and all the brain metastases went away. There is also heterogeneity in tumors where you may have resistance mutation at 1 site but not in the others.

Case 2 (contined):

NGS testing was positive for anEGFRmutation.

How does the discovery of a driver mutation in this patient influence your treatment decision making?

This patient in our case ends up having a mutation. I would not recommend immunotherapy up front. When considering a TKI, osimertinib [Tagrisso] has demonstrated very good intracranial penetration. However, what are the cutoffs when consulting with the radiation oncologist? Their comfort level is that if the lesion is >1 cm and the patient is symptomatic, the patient should be treated, especially if there is vasogenic edema, such as this case. When we have small metastases, <1 cm, and the patient is asymptomatic, then I would feel more comfortable [treating systemically].

The other factor is the location of tumors. A cerebellar lesion would be in a location that we would want to treat. We have had patients with a cerebellar tumor that after we treated it and it responded, it herniated. Again, we are trying to prevent a more serious neurologic complication. Radiation oncologists tell me that location of the tumor is very important. If it is located in the midbrain, cerebellar, brainstem, a slight swelling will cause something very detrimental. I have had patients herniate in the past when they respond to treatment.

How do you treat a patient such as this who develops pseudoprogression?

Our radiation oncologists have said that it is safe to give TKIs concurrent with stereotactic radiation. When giving immunotherapy, you [can] see the phenomenon of pseudoprogression in an area that has been radiated. It can be very difficult to tell if it is a true progression or just inflammation from the immunotherapy, which seems more common in patients treated with stereotactic radiation. It can be tricky because pseudoprogression, or inflammation, can be just as bad as the metastases in the first place. A lot of these patients are on long-term steroids, and the question of how to taper the steroids down is a challenge. This is going to be an emerging issue that we see as patients live longer. We do have patients on immunotherapy who are living 8 to 10 years now. I think we are going to see more longer-term sequelae that we did not anticipate because we did not anticipate them to live so long. It is a good problem to have, in a way.

In frontline testing, which genes should be routinely tested for alterations in lung cancer?

For biomarker testing, the mandatory must-test minimum areEGFR,ALK,ROS1, andBRAF. I would say these are changing with each FDA-approved drug. ForNTRK, we have entrectinib for any tumor that has theNTRKfusion gene. We have had patients withMETamplification or MET exon 14 skipping that, because it is listed in NCCN [National Comprehensive Cancer Network] guidelines, you can get it approved for crizotinib [Xalkori].

This is really a moving target right now. Not every panel can test them all at once, so it makes sense to test these because they have higher frequencies in patients and quite a few FDA-approved drugs. This is really evolving.

If the mutation is listed in the guidelines, most insurers will cover the treatment, even if there is no FDA-approved drug.

What can medical oncologists do to help ensure sufficient tissue quantity is collected and that tissue is used efficiently? What needs to be communicated to the interventionalist conducting the biopsy? To the pathologist?

The process and protocol when ordering tissue tests can vary by institution, and [they affect] patient care. For example, a patient with metastatic disease comes in with spinal cord compression or brain metastases. These mean the tissue comes in a different path in pathology than if the tissue was collected via bronchoscopy. There are a lot of challenges associated with this.

One thing that is important is that even though all of the mutations are in 1% or 2% of the population, if you have 10 different mutations, that adds up to a lot of patients. The problem is still similar: You do not have enough tissue to make the diagnosis.

Do you prefer to use osimertinib in this patient?

This person has central nervous system [CNS] disease. Things to think about when you are choosing a TKI [are] drug availability, toxicity, and does it affect the brain.

Most people use osimertinib in the frontline. I do have some patients on erlotinib [Tarceva]. It just depends on the scenario. One of my concerns, or why I like osimertinib, is the CNS progression, especially for my young patients. That is such an awful area to progress, and if they have 1 metastasis in the brain, I worry that they will develop more. I do think osimertinib is well-tolerated.

How do the FLAURA data apply to this patient?

The FLAURA data are often cited for first-line osimertinib versus standard of care forEGFR-mutated advanced NSCLC.4[The] standard of care was erlotinib or gefitinib [Iressa] in patients who had an exon 19 mutation. The primary endpoint was PFS, and the secondary endpoints were objective response rate, DOR, OS, and safety. One thing to point out is that many of these patients were Asian. The trial did meet the primary endpoint, with a median PFS of 18.9 versus 10.2 months.

What they found in FLAURA was that the PFS was significantly greater in the osimertinib arm. [These were] the early data, but the median OS have not been reached and appear, at least early on, to be better in the osimertinib arm. The only thing I will say is that the number of patients who crossed over, about half of patients in the standard-of-care arm, did not go on to receive any kind of therapy at the time of progression. Actually, it was high in the osimertinib arm as well.

I have often wondered why half the patients did not receive a second-line therapy, and part of the answer is tied to the location of the patient’s treatment and what is available; but I just find it interesting. The other thing is when you look at subgroups stratified by types of treatment in the second line, many of them do not get platinum doublets or platinum-containing chemotherapy. Nonetheless, the PFS is significantly better than that in the subgroups for which there appears to be a benefit of osimertinib over standard TKI use.

The data that we talk about in terms of using this agent over another TKI have been with the CNS metastases. Regardless of CNS metastases, osimertinib fares better. But I think these CNS responses are encouraging, especially for younger patients for whom CNS progression would be awful for their quality of life. That is something we are thinking about when justifying which TKI to choose. And what we see is that not all patients were required to have an MRI of the brain. It is possible that [a] patient [was] categorized as not having CNS metastases. They could have had small metastases that were not captured.

What were the common toxicities observed with osimertinib?

In terms of grade 3/4 toxicity, 30% of patients given osimertinib had grade 3 AEs and 2% had grade 4 compared with standard-care [therapy in which patients experienced] grade 3 AEs at 37% and 4% for grade 4. Again, osimertinib is pretty well tolerated.

In terms of an AE leading to discontinuation, patients who took osimertinib were less likely to discontinue compared [with] the standard TKI, which again speaks to our anecdotal experience.

The ARCHER 1050 study looked at dacomitinib [Vizimpro] versus gefitinib, which did result in dacomitinib’s approval in the first-line setting.5Patients were randomized to either dacomitinib [or] gefitinib. The primary endpoint was PFS. For OS, we [saw] a slight advantage of dacomitinib [34.1 vs 26.8 months (HR, 0.76; 2-sidedP= .0438)], but this is a second-generation TKI versus a first-generation TKI.

This only included EGFR exon 19, and there did not appear a difference seen in Asian versus non-Asian patients.

Grade 3 toxicities, particularly diarrhea, nail changes, rash, and stomatitis were higher in the dacomitinib arm compared with the gefitinib arm. Many patients required dose modification, 66.5% in the dacomitinib arm versus 8% in the gefitinib arm.

Case 2 (continued):

The patient received osimertinib at 80 mg once daily. She experienced a good PR to treatment.

Ten months later she complained of headaches and worsening fatigue. A CT scan showed 3 new liver lesions and a brain MRI showed 1 new lesion. Her ECOG performance status was 1.

Assume mutation testing was performed and revealed anEGFRC797S resistance mutation. How would this influence your decision making?

This occurs very rarely, and there is no FDA-approved drug. It occurs in <10% of resistance mutations in patients who have received osimertinib. There is a drug in trials that could address this, and my patients are responding well, but I cannot recall the name of the drug. That is why I encourage that if a patient can get on a clinical trial, then they should.

Would you consider a checkpoint inhibitor in a patient who has progressed on a TKI?

If you accept the data from IMpower150, then you have to consider the patient as a candidate for bevacizumab [Avastin].6We have to ask, “Does this person have a history of GI bleeding, diverticulosis, blood clot?” That’s always a challenge in metastatic lung cancer.

[IMpower150] did include patients with anEGFRmutation who did regress on TKI, whereas KEYNOTE-189 [NCT02578680] did not allow patients with mutations. Do we believe that immunotherapy is effective in patients with a mutation versus not? I would say there is some concern that if we are giving immunotherapy alone to patients with a mutation, then the benefit may not be there compared with chemotherapy.


  1. Antonio SJ, Villegas A, Daniel D, et al; Pacific Investigators. Durvalumab after chemoradiotherapy in stage iii non—small-cell lung cancer.N Engl J Med. 2017;377(20):1919-1929. doi: 10.1056/NEJMoa1709937.
  2. Antonio SJ, Villegas A, Daniel D, et al; Pacific Investigators. Overall Survival with Durvalumab after Chemoradiotherapy in Stage III NSCLC.N Engl J Med. 2018;379(24):2342-2350. doi: 10.1056/NEJMoa1809697.
  3. Hellmann MD, Ciuleanu T-E, Pluzanski A, et al. Nivolumab plus ipilimumab in lung cancer with a high tumor mutational burden.N Engl J Med. 2018;379(22):2093-2104. doi: 10.1056/NEJMoa1801946.
  4. Soria J-C, Ohe Y, Vansteenkiste J, et al; FLAURA Investigators. Osimertinib in untreatedEGFR-mutated advanced non—small-cell lung cancer.N Engl J Med. 2018;378(2):113-125. doi: 10.1056/NEJMoa1713137.
  5. Mok TS, Cheng Y, Zhou X, et al. Dacomitinib (daco) versus gefitinib (gef) for first-line treatment of advanced NSCLC (ARCHER 1050): final overall survival (OS) analysis.J Clin Oncol. 2018;36(suppl 15; abstr 9004). doi: 10.1200/JCO.2018.36.15_suppl.9004.
  6. Socinski MA, Jotte RM, Cappuzzo F, et al; the IMpower150 Study Group. Atezolizumab for first-line treatment of metastatic nonsquamous NSCLC.N Engl J Med. 2018;378(24):2288-2301. doi: 10.1056/NEJMoa1716948.