Expert Perspectives: ESMO 2017 Changes Lung Cancer Practice

Floriana Morgillo, MD, PhD; H. Jack West, MD; Stefan Zimmermann, MD


September 11, 2017

Like much of oncology at large at the moment, developments in lung cancer are among the hottest topics at the 2017 European Society for Medical Oncology (ESMO) Congress in Madrid, Spain. Medscape has a team of specialists in thoracic oncology at ESMO 2017 who are providing their perspectives on the clinical impact of the new lung cancer data being presented.

For more real-time coverage across the entire field of oncology, visit Medscape's ESMO 2017 Live Blog, which features specialists in breast, melanoma, lung, central nervous system (CNS), and prostate cancer providing their perspective on practice-changing data from ESMO 2017, or visit our ESMO conference page for all conference coverage.

H. Jack West, MD: This is my first ESMO, and I'd say that today has been arguably one of the biggest days for practice-changing results that I can remember since focusing on thoracic oncology about 15 years ago. It was great to see lung cancer as the focus of the first Presidential Session at ESMO, with the huge room filled with tons of people and energy. It's amazing to think that for most of the first 5 years or so after I entered the field, people asked with disbelief why someone would want to focus on lung cancer. It's difficult to keep up with the major advances and new treatment options, so let me start with the first one, with more to follow later.

The first big result, befitting the first presentation in the Presidential Session by Luis Paz-Ares, MD, PhD, head of the medical oncology service at University Hospital 12 de Octubre in Madrid, was of the PACIFIC trial of consolidation durvalumab or placebo every 2 weeks for a year, after patients with unresectable stage III non–small cell lung cancer (NSCLC) completed concurrent chemotherapy (at least two cycles) and chest radiation.

More than 700 patients were randomly assigned 2:1 to immunotherapy or placebo, and the interim analysis focused on progression-free survival (PFS), not the co-primary endpoint of overall survival (OS). We knew from a press release that the trial was positive, but we didn't know whether the differences would be enough to be very impressive, especially without reporting on OS. The results today were accompanied by—actually, preceded by—the publication of the trial by Scott Antonia, MD, PhD, and colleagues in the New England Journal of Medicine.[1]

It didn't disappoint. There was a striking difference in PFS—median 16.8 versus 5.6 months, HR 0.52 (highly significant)—with the difference in PFS sustained at 12- and 18-month landmarks. All efficacy endpoints were similarly superior for durvalumab, including median time to death or distant metastases (23.2 vs 14.6 months; P < .001). Though there was a slightly higher rate of grade 3/4 toxicity (about 5% more with durvalumab than placebo), this was a very marginal difference compared with the major efficacy difference with durvalumab.

Although there has been some back-and-forth discussion about the significance of results that don't include OS, particularly in a potentially curative setting, the prevailing view has been the one I share, that these results are remarkably impressive and warrant a change in the standard of care. To be completely convincing, we should see a survival benefit. But seeing not just a marginally statistically significant PFS difference but a threefold increase in median PFS, and a difference sustained out to at least 18 months, I'm convinced that this must be associated with an improvement in OS that must be clinically significant. I'm less confident that immunotherapy will lead to an increase in the proportion of patients who are truly "cured" and will never see a relapse of their cancer, but I'm hopeful that that will also be the case.

There really hasn't been a major change in our management of stage III NSCLC in more than a decade. That changed today.

Fewer CT Scans Needed in Early Lung Cancer Surveillance

Floriana Morgillo, MD, PhD: Virginie Westeel, MD, PhD, head of thoracic oncology at the University Hospital of Besançon, France, presented the second of three phase III lung cancer studies featured in the Presidential Symposium: the IFCT-0302 trial results.

The context is early-stage resected NSCLC patients. The majority of medical societies' (ESMO, NCCN, ASCO) clinical practice guidelines recommend follow-up visits every 3-6 months for 2-3 years, less often (eg, annually) thereafter. Visits include history, physical examination, and chest CT as appropriate, although there is a lack of well-designed randomized controlled trials addressing the effect on survival of using chest CT scan instead of chest x-ray. Despite the lack of such evidence, a majority of physicians require CT scan during follow-up visits of their patients.

The IFCT-0302 trial is the first randomized study with the goal of defining the best follow-up method in resected NSCLC, by comparing the impact on overall survival of two follow-up programs: one consisting of clinical examination and chest x-ray, and the other consisting of a more intensive approach with clinical examination, chest x-ray, and thoracoabdominal CT scan plus bronchoscopy (optional for adenocarcinomas). In both arms, procedures were repeated every 6 months after randomization during the first 2 years and yearly until 5 years.

The study, at a median follow-up of 8.7 years, does not demonstrate a significant benefit in OS between arms (HR = 0.92; 95% CI, 0.8-1.07; P = .27) thus demonstrating chest x-ray as a valid option to monitor disease relapse instead of CT scan. A trend in favor of the CT scan approach can be seen (median OS was 8.2 years for arm 1 and 10.3 years for arm 2), indicating that we may need a longer follow-up to appreciate the benefit in OS in this subset of patients followed by CT scan. Of note, it seems that this trend is due to the ability of the CT scan to detect second primary tumors.

The results of this study are very informative. Especially in the first 2 years, monitoring of resected NSCLC patients probably will not exclude CT scan as a follow-up method. The importance of CT scanning is linked to the ability to detect local and distal recurrence with higher sensitivity than chest x-ray in previously resected patients at an asymptomatic stage, therefore allowing earlier interventions to take place. The early detection of local or distant recurrence after curative surgery retains a high potential for improving survival rates, especially if second surgery is possible.

Furthermore, the use of more tailored therapeutic approaches, eventually integrated with locoregional therapeutic approaches, may significantly improve survival time in patients with relapsed disease that is detected earlier.

Last but not least, the CT scan retains its role in the early diagnosis of second primary tumors.

Dr West: Dr Morgillo gave a thorough summary already of the very important IFCT-0302 trial that clearly demonstrated an absence of benefit for intensive postoperative surveillance after surgery for early-stage, resected NSCLC. The comparability of survival was convincing, suggesting that our judgment-based routine of often favoring CT scans every 6 months for the first few years (and sometimes more often than that) serves no clear purpose except to reassure ourselves that we're being vigilant. Dr Westeel was circumspect in her comments; she didn't say that it was pointless and wasteful, instead suggesting that there may still be value in less frequent scans to detect second cancers.

But it leaves us with the question of whether oncologists (or others following these patients) will change their practice and stop doing CT scans, or do them far less frequently. Much of the chatter on Twitter was about the concern that patients will seek the reassurance of more frequent CT scans, so perhaps we should continue with our routines as usual.

I definitely appreciate the value in patient satisfaction and quality of life, but I personally have an issue with doing interventions of proven nonvalue just to appease patients whose expectations are not based on solid judgment or evidence, but instead on a bias and desire. Too many people reflexively presume that more is better—more serum marker testing, more antibiotics, more expensive imaging. Too often, the serial PET scan is an emblem of ridiculous waste, and it's good that ASCO made this one of the elements of their Choosing Wisely campaign (which is far more tactful than my rejected campaign name, "Stop Being Such a Wasteful Moron"). But I still see it overused because the oncologist and/or the patient just want more.

My view is that it is our job to explain to patients that there is no benefit to frequent CT surveillance, and that there is harm in giving an antibiotic to everyone with a transient cough who asks for a prescription. If these interventions are being administered just to palliate patient anxiety, better communication and/or lorazepam is a better choice.

That said, I think it is worth recognizing that patients at known high risk for lung cancer, typically based on smoking history and age, have a proven survival benefit from annual low-dose chest CT screening. I have picked up more than a dozen new second lung cancers in patients I've followed over the past 10 years, and in each case they have been stage I and treated with surgery or SABR (stereotactic ablative radiotherapy), all without relapse over further follow-up. Patients with a history of a lung cancer are, by definition, at high risk for another lung cancer, so I will be inclined to continue annual screening, but I will move away from my judgment favoring more frequent CT scanning over the first 2-3 years.

Every day, we are confronted with clinical situations in which there are no good data. In those settings, we must rely on our best judgment. But we need to respect the data when it comes along and informs us that our judgments are not correct.

Mine isn't the only perspective on this complex topic. What do others think? Please let us know in the comments section of this article.

Osimertinib: A New First-line Standard

Dr West: Another of the likely practice-changing presentations on lung cancer from ESMO 2017 was by Suresh S. Ramalingam, MD, director of medical oncology at Emory University in Atlanta, Georgia, on the FLAURA study, a global phase III trial that randomly assigned (1:1) 556 patients with an activating EGFR mutation to first-line standard-of-care treatment with a first-generation EGFR tyrosine kinase inhibitor (TKI) (which turned out to be gefitinib administered to two thirds of patients, erlotinib to one third) or initial treatment with the third-generation EGFR TKI osimertinib, which is currently approved only as a second-line therapy for the subset of EGFR mutation-positive patients with a T790M mutation associated with acquired resistance to another EGFR TKI, such as gefitinib, erlotinib, or afatinib (approximately 50%-60% of patients).

Approximately two thirds of patients were Asian, nearly two thirds were never-smokers, and about 20% had CNS metastases at the time of treatment. Patients on the standard-of-care arm were permitted to cross over to osimertinib if they were found to have a T790M mutation upon progression on one of the first-generation EGFR TKIs. The primary endpoint was PFS, based on investigator assessment.

We knew from a prior press release that the trial met its primary endpoint, but the key to whether it would be practice-changing was the magnitude of the benefit. With osimertinib available as a second-line therapy for at least the 50%-60% of patients with a T790M mutation first-line, and with a median PFS in that setting of about 9 months, an absolute difference in PFS of only 5 months favoring osimertinib—while statistically significant—wouldn't necessarily lead us to change practice and give up a potentially valuable line of therapy.

But that's not what we saw. The median PFS with osimertinib was 18.9 months, compared with a very reasonable median PFS of 10.2 months with standard of care (gefitinib or erlotinib), and a hazard ratio of 0.46 (P < .001). The benefit with osimertinib was seen across all clinically defined subgroups, in both common activating EGFR mutations, and it was nearly identical in patients whether they had presented with or without brain metastases. The response rates were very comparable (80% with osimertinib, 76% with standard of care), but the duration of those responses was significantly longer with osimertinib.

Because patients with an EGFR mutation typically live several years and commonly receive several lines of subsequent therapy after their first-line treatment, it has been very difficult to detect survival differences based on one first-line therapy versus another. However, the FLAURA trial demonstrated an HR for OS (interim analysis with only 25% maturity) of 0.63 (P = .0068), which was not considered statistically significant because a higher threshold P value of .0015 or lower was required to satisfy the O'Brien-Fleming rule. Nevertheless, a hazard ratio of 0.63 for overall survival is an impressive complement to the highly significant PFS difference.

Of course, toxicity is a significant issue as well, and this was also more favorable for osimertinib. Though treatment was generally well tolerated on both arms, overall rates of grade 3/4 toxicity were lower with osimertinib, particularly in regard to rash and liver function test elevations.

Though I often find concluding statements from pharma-led studies to be overreaching and obviously self-serving, I think this was entirely justified.

Dr Ramalingam concluded with the statement that osimertinib is a new standard of care for first-line treatment in patients with an activating EGFR mutation. Though I often find concluding statements from pharma-led studies to be overreaching and obviously self-serving, I think this was entirely justified. In fact, I think the key question is whether it should be considered "a standard of care" or "the standard of care." As generic EGFR TKIs enter the market soon, I'm sure we'll need to weigh cost versus efficacy, and I would say that the cost of osimertinib relative to other options is a significant concern. But for the clearly superior efficacy, including within the CNS (which wasn't the focus of this presentation but is also a major benefit with osimertinib relative to other EGFR TKI options), as well as the toxicity benefit in a setting of prolonged and ongoing treatment, osimertinib is a very clear winner. I am hopeful that we'll see a broadening of the indication very soon, and I will definitely be inclined to switch as many of my current patients on erlotinib to osimertinib as soon as is feasible, as well as start my newly diagnosed patients with an activating EGFR mutation on osimertinib.

Another game changer.

When Should Checkpoint Inhibitors Be Discontinued?

Stefan Zimmermann, MD: Immune checkpoint inhibitors: How long is long enough? Is more actually more?

At least in advanced NSCLC treated with immune checkpoint inhibitors, the optimal duration seems to be at least until progression or limiting toxicity, and not a fixed duration of 1 year. This is a question of paramount importance: When is the best moment to stop? Is it at a definitive time point after therapy initiation, after first response, after maximal response, after adverse event occurrence, or based on a biological dynamic marker?

In an era of financially stretched healthcare systems, and because of toxicity, de-escalation trials in immunotherapy are of obvious interest yet are extremely hard to set up with the industry. One such trial, CheckMate 153, was just presented today in the metastatic NSCLC session, and the results are enlightening.

In contrast to the limited data available in melanoma, CheckMate 153 provides invaluable insight into what happens after we stop anti-PD1 therapy after 1 year. For the 17% of patients who had not progressed after a year of nivolumab therapy, continuation of a checkpoint inhibitor yielded a significant benefit in PFS (HR for progression or death, 0.42) and a trend for an OS benefit (HR for death, 0.63). The study also clearly demonstrated that most patients in whom treatment was stopped were not successfully salvaged with re-treatment after progression. Benefit was seen across subgroups; whether this holds true for patients in complete response remains unclear.

Finally, and thankfully, few new-onset toxicity events were observed, so toxicity issues should not weigh too much in this clinical decision. More is more in this particular setting.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.