The OAK study—recently reported by Rittmeyer and colleagues and reviewed in this issue of The ASCO Post—is the first study to show patients with previously treated non–small cell lung cancer (NSCLC) treated with a humanized antibody (atezolizumab, Tecentriq) directed against the programmed cell death ligand 1 (PD-L1) have prolonged overall survival compared with those treated with docetaxel.1 This study led to the U.S. Food and Drug Administration (FDA) approval of atezolizumab in October 2016 to treat patients with metastatic NSCLC whose disease progressed during or after platinum-containing chemotherapy. This is in addition to the two immune checkpoint agents directed against the programmed cell death protein 1 (PD-1) already approved for previously treated NSCLC: nivolu-mab (Opdivo) and pembrolizumab (Keytruda).
I would like to comment on the potential differences between targeting PD-L1 vs PD-1, the trial designs and results, and the potential predictive biomarkers employed in the different studies. Also I plan to focus on the hazard ratios between the two different treatments for the purposes of comparing the studies with each other.
Targeting PD-L1 vs PD-1
Tumor cells and tumor-infiltrating cells can make PD-L1 (a protein that interacts with PD-1 or B7-1) programmed cell death receptors on T cells, and PD-L1 can suppress signals to the T cells that would otherwise attack cancer cells. The administration of an antibody directed against PD-L1 or PD-1 can interrupt the signal inhibiting the T cells from attacking the cancer and allow the T cells to assault the cancer cells.
Agents directed against PD-L1 and PD-1 have received FDA approval for treatment of patients with melanoma, kidney cancer, NSCLC, and bladder cancer. Investigators have hypothesized that directly targeting PD-L1 with atezolizumab may leave the PD-L2 and PD-1 interaction intact and cause less autoimmunity than the agents that target PD-1 directly (nivolumab and pembrolizumab).
Trial Designs and Results
There are significant differences in the trials for patients with previously treated NSCLC with atezolizumab, nivolumab, and pembrolizumab. The OAK trial reported on the efficacy of atezolizumab vs docetaxel in 850 patients with previously treated squamous and nonsquamous NSCLC (1–2 regimens) without prospective biomarker selection and yielded a hazard ratio of 0.73 for overall survival (median survival of 13.8 months for those treated with atezolizumab vs 9.6 months for those treated with docetaxel). The hazard ratios in favor of atezolizumab over docetaxel were similar at 0.73 for both the squamous and nonsquamous NSCLC populations, despite survival being longer for patients with nonsquamous NSCLC treated with either docetaxel or atezolizumab than for those with squamous cell lung cancer treated with either agent.1 The longer survival of patients with nonsquamous NSCLC vs those with squamous NSCLC has also been seen in the trials of nivolumab and pembrolizumab.
PD-L1 staining is now being adopted in the initial characterization of NSCLC to identify patients who are candidates for initial checkpoint inhibitor therapy rather than chemotherapy.— Bruce E. Johnson, MD
Similar to the OAK study, two different randomized trials for patients with previously treated squamous and nonsquamous NSCLC compared the survival of patients treated with nivolumab with that of those treated with docetaxel.2,3 Once again, there was no biomarker selection for protocol entry. The hazard ratio in favor of nivolumab over docetaxel for 582 patients with previously treated nonsquamous NSCLC was 0.73—identical to that seen in the OAK study, where 74% of the patients had nonsquamous NSCLC.2 The hazard ratio in favor of nivolumab over docetaxel for the 272 patients with previously treated squamous NSCLC was 0.59—somewhat more favorable for nivolumab than the overall number in the OAK study—but the 95% confidence intervals overlap.3
Cigarette smoking has been linked to improved outcomes with checkpoint inhibition, but the smoking history by histology was not provided in the OAK study to compare the potential impact of different smoking rates between the two studies with nivolumab.
The pembrolizumab KEYNOTE-010 study included both squamous and nonsquamous NSCLC, and the entry criteria for the pembrolizumab trial were different from those of the OAK study and the two nivolumab studies. The KEYNOTE-010 study required patients to have at least 1% of the tumor cells stained with a PD-L1 antibody, which excluded about 34% of the patients who had less than 1% staining. An additional difference in this study was that it incorporated two different doses of pembrolizumab: 2 mg/kg and 10 mg/kg administered every 3 weeks.
The hazard ratios were remarkably similar to the findings in both the OAK and nivolumab studies. The hazard ratios in favor of pembrolizumab were 0.71 with 2 mg/kg (n = 344) and 0.61 with 10 mg/kg (n = 346). The hazard ratio was more favorable in the squamous than the nonsquamous NSCLC patients (0.74 and 0.63, respectively, for the pooled pembrolizumab groups vs docetaxel).1,4 My conclusion from the four different publications is that distinguishing among the different checkpoint inhibitors by their therapeutic efficacy in these populations is thus far difficult.
Potential Predictive Biomarkers and Toxicity
The three assays used propose to assess PD-L1 in the tumor (nivolu-mab and pembrolizumab) or in the tumor plus the tumor-infiltrating cells (atezolizumab). Three different antibodies were used: SP142 PD-L1 immunochemistry (IHC) for the atezolizumab studies; a PD-L1 antibody–automated IHC (clone 28-8) for nivolumab; and a 22C3 antihuman antibody for pembrolizumab. All three checkpoint inhibitors show an association between high PD-L1 staining and greater response, but to differing degrees. Of the three agents, only pembrolizumab requires at least 1% of the tumor cells to stain for PD‑L1 for entry onto the clinical trial.
One of the key findings in the OAK study is that the patients with negative PD-L1 staining for tumor cells but positive staining for tumor-infiltrating cells have a similar survival benefit from atezolizumab over docetaxel as the overall patient group (hazard ratio = 0.73). Although it appears that it is not critical to assess PD-L1 staining for previously treated patients with NSCLC, testing for PD-L1 protein expression has become important for patients with previously untreated advanced NSCLC. Further research has shown that previously untreated patients with high expression of PD-L1 (> 50% of the tumor cells staining positive; about 20%–30% of the patients with NSCLC) treated with pembrolizumab live longer (hazard ratio = 0.60) than those treated with platinum-based chemotherapy.5,6 Therefore, PD-L1 staining is now being adopted in the initial characterization of NSCLC to identify patients who are candidates for initial checkpoint inhibitor therapy rather than chemotherapy.
Comparison of the toxicities across these different studies is difficult, due to the low frequency of grade 3 and 4 treatment-related adverse events in patients treated with atezolizumab, nivolumab, and pembrolizumab (about 7%–16%). We must await further pooled analyses of the different agents across multiple studies to make an accurate comparison.
Three different checkpoint inhibitors are now available for the treatment of previously treated NSCLC: atezolizumab, nivolumab, and pembrolizumab. Thus far, it is difficult to identify which of these three agents has the greatest efficacy based on the results of these large phase III clinical trials within the limitations of comparing across studies. It will be helpful for the different companies to provide information on toxicities across different studies to help compare the side effects of these agents to assess whether we should favor checkpoint inhibitors based on their toxicity profile. ■
Dr. Johnson is Chief Clinical Research Officer, Dana-Farber Cancer Institute, and Professor of Medicine, Harvard Medical School, Brigham and Women’s Hospital, Boston.
Disclosure: Dr. Johnson reported no conflicts of interest.
1. Rittmeyer A, Barlesi F, Waterkamp D, et al: Atezolizumab versus docetaxel in patients with previously treated non-small-cell lung cancer (OAK): A phase 3, open-label, multicentre randomised controlled trial. Lancet 389:255-265, 2017.
2. Borghaei H, Paz-Ares L, Horn L, et al: Nivolumab versus docetaxel in advanced nonsquamous non-small-cell lung cancer. N Engl J Med 373:1627-1639, 2015.
3. Brahmer J, Reckamp KL, Baas P, et al: Nivolumab versus docetaxel in advanced squamous-cell non-small-cell lung cancer. N Engl J Med 373:123-135, 2015.
4. Herbst RS, Baas P, Kim DW, et al: Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): A randomised controlled trial. Lancet 387:1540-1550, 2016.
5. Reck M, Rodríguez-Abreu D, Robinson AG, et al: Pembrolizumab versus chemotherapy for PD-L1-positive non-small-cell lung cancer. N Engl J Med 375:1823-1833, 2016.
6. Johnson BE: Divide and conquer to treat lung cancer. N Engl J Med 375:1892-1893, 2016.