Practice-Changing Results From ASCO 2015: Immunotherapy in Previously Treated Advanced Non-Small Cell Lung Cancer

H. Jack West, MD


July 31, 2015

Immunotherapy wasn't merely the big story of the American Society of Clinical Oncology (ASCO) 2015 meeting; it was the tornado that left a vacuum in its wake, as some of us discussed—the unfortunate side effect of everything that wasn't immunotherapy being swept aside as last year's model of quaint anticancer therapy.

In the world of lung cancer, at least, immunotherapy trials delivered data that merited real excitement. We saw truly practice-changing results that provided a glimpse of a new era in which we might expect lung cancer treatments to be transformed by integrating immunotherapy across many settings. For now, that will start with previously treated advanced non-small cell lung cancer (NSCLC), and we will also see where immunotherapy is heading next.

One critical trial we knew to expect was the CheckMate-017 trial, which had already changed practice even before we saw most of the data. It was presented by Dr David Spigel[1] and published the same day.[2] On the basis of the US Food and Drug Administration's (FDA's) approval of the programmed cell death protein 1 (PD-1) inhibitor nivolumab (Opdivo®) in early March, we already knew that it led to a highly statistically significant improvement in overall survival (OS) (median, 9.2 vs 6.0 months; hazard ratio [HR], 0.59; P <.001) compared with docetaxel as second-line therapy in a group of 272 chemotherapy-pretreated patients with advanced squamous NSCLC.

What we saw at ASCO 2015 corroborated our preliminary enthusiasm: The survival benefit was accompanied by a significantly longer progression-free survival (PFS) (3.5 vs 2.8 months; HR, 0.62; P < .001) and higher objective response rate (20% vs 9%; P = .008). Medians, however, are less impressive than the extent and duration of benefit in responders, as evidenced by the difference in 1-year OS of 42% vs 24%, 1-year PFS of 21% vs 6.4%, and duration of response "not yet reached" vs 8.4 months, all dramatically favoring nivolumab. Of note, the toxicity profiles of these treatments also clearly favored nivolumab, with 57% vs only 7% of patients experiencing grade 3-5 adverse events on docetaxel vs nivolumab, respectively.

The last critical portion of the study in squamous NSCLC was the finding of no hint of difference in efficacy as a function of protein levels of programmed death receptor ligand-1 (PD-L1). Every subgroup, regardless of clinical variable or PD-L1 cutoff level, demonstrated a clear superiority for nivolumab.

The accompanying trial, CheckMate-057, randomly assigned 582 chemotherapy-pretreated patients with advanced nonsquamous NSCLC to receive nivolumab vs docetaxel and was presented by Dr Luis Paz-Arez.[3] Unlike CheckMate-017, CheckMate-057 was only known to be positive by a prior press release[4] with no data.

Here, both the nivolumab and docetaxel groups demonstrated a median OS approximately 3 months longer than the same treatments produced in the squamous NSCLC population, but we again saw another approximately 3-month improvement in median OS: 12.2 vs 9.4 months (HR, 0.73; P = .0015). There was no significant difference in PFS, and a somewhat less pronounced improvement in objective response rate (19% vs 12%; OR, 1.7; P = .0246); as in the population with advanced squamous NSCLC, nivolumab shone brightest with regard to the longer-term endpoints of 1-year OS (51% vs 39%), 1-year PFS (19% vs 8.1%), and duration of response (17.2 vs 5.6 months). The safety profile was also remarkably superior for nivolumab.

Aside from the longer OS relative to patients with squamous NSCLC, patients with nonsquamous NSCLC also differed in demonstrating significant differences in efficacy as a function of PD-L1 tumor expression. Although OS was marginally better with increasing levels of PD-L1 expression from a threshold of >1% of cells to >5% to >10%, the clearest cutoff was simply expression of >1% or not: the benefit with nivolumab was entirely confined to PD-L1–positive patients.

These results suggest that nivolumab should be considered the unequivocal standard of care for all eligible patients with previously treated advanced squamous NSCLC and for many patients (arguably, the vast majority) with previously treated advanced nonsquamous NSCLC.

Along with the debatable value of filtering patients with nonsquamous NSCLC on the basis of PD-L1 expression, a separate presentation[5] revealed that never-smokers and patients with an activating EGFR mutation or ALK rearrangement also seemed to derive no meaningful benefit from nivolumab. This suggests that such patients have scant immunogenic targets compared with patients who have cancers with a greater "mutational load," which has been correlated with greater probability of response to immune checkpoint inhibitors.[6]

Should patients with nonsquamous NSCLC be screened for tumor PD-L1 expression before treatment with nivolumab is recommended? There are several factors to consider. First, we have yet to see whether the FDA will require PD-L1 positivity in this setting. Some may argue that the trial was positive for the broad, unselected population and should be applied in a similarly unselected way. But corroborating evidence of the potential value of PD-L1 expression was demonstrated in the POPLAR trial of atezolizumab (MPDL3280A),[7] a PD-L1 inhibitor, compared with docetaxel in 287 patients with previously treated advanced NSCLC (squamous or nonsquamous) who were divided into groups on the basis of PD-L1 expression levels in tumor cells, immune cells, or both. This study revealed a superior survival for atezolizumab that was most pronounced in the patients with the highest immune cell or tumor cell expression of PD-L1, with a stepwise reduction in benefit of the immunotherapy relative to docetaxel with lower levels of PD-L1 expression.

Taken together, these data lead me to conclude that PD-L1 expression, though still limited by varied testing approaches and definitions of positivity, is a biomarker with real potential value, at least in patients with nonsquamous NSCLC. Why would it be less valuable in patients with squamous NSCLC? The higher mutational load in squamous NSCLC may offer competing drivers of immunogenicity that render PD-L1 less important.

Regardless, if PD-L1 expression testing were to require a repeat biopsy at progression followed by a delay of several weeks to learn PD-L1 expression levels, this will become a practical barrier that will probably lead oncologists to be far less inclined to pursue it unless it is required and thus otherwise favor treatment with nivolumab on an unselected, empirical basis without delay, on the basis of the desire of patients and oncologists to never miss the potentially dramatic benefit of an immune checkpoint inhibitor.


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.