Adjuvant Osimertinib in EGFR-Mutant Early-Stage NSCLC: Does HRQoL Influence Decisions?

In this issue of Clinical Cancer Research, Majem and colleagues (1) are to be congratulated on their health-related quality of life (HRQoL) data in the randomized ADAURA trial (2) as a secondary endpoint of the trial. The data showing no detrimental effect of 3 years of osimertinib on HRQoL are quite complete and potentially important. However, there are issues with the interpretation of the results, especially before the trial is compete and overall survival (OS) data become available. What do we know about the role of EGFR tyrosine kinase inhibitors (TKI) in lung cancer? In advanced metastatic disease, we know from the LCMC trial that patients with EGFR mutation who receive a TKI in any line of therapy live longer than those who never receive an EGFR TKI (3). We also know form multiple randomized trials comparing an EGFR TKI to a platinum doublet that the overall survival is equivalent whether the TKI is given before or after chemotherapy (4). However, the first-line use of the EGFR TKI is preferred due to a higher response rate which improved symptoms more often, less treatment toxicity, improved HRQoL, and longer progression-free survival (PFS). Will earlier use of an EGFR TKI in the adjuvant setting improve survival compared with later use for only those who relapse?

What do we know about the use of EGFR TKIs in early stages? In earlier stages, there have been multiple randomized trials evaluating first-generation EGFR TKIs (5). The first of these to be reported was the “RADIANT” trial which randomized patients to receive 2 years of adjuvant erlotinib or placebo (6). In EGFR-mutant patients, there was a significant improvement in disease-free survival (DFS) but there was no improvement in overall survival, likely because patients who received placebo and crossed over to erlotinib at progression had the same survival benefit as patients who received it earlier as in later stage trials. The other randomized trials comparing adjuvant first-generation EGFR TKIs with adjuvant chemotherapy to placebo with adjuvant chemotherapy all showed a significant improvement in DFS but no improvement in OS (5). A meta-analysis of these trials confirmed that the DFS was significantly longer with the EGFR TKI but there were no significant differences in OS (5). The inability of 2 years of adjuvant EGFR TKIs to improve cure rates and survival is likely due to their inability to produce complete responses in advanced stages indicating that they produce only 1–2 log cell kill. This degree of cell kill was not able to improve cure rates with erlotinib or gefitinib. The major question is whether a longer treatment with a superior EGFR TKI, osimertinib, will produce sufficient cell kill in early-stage patients to improve cure rates. If so, survival improvement will be evident. If not, it may be better to define which patients are most likely to relapse and determine if early use of osimertinib can cure any of these patients.

The ADAURA trial was designed to determine if a third-generation EGFR TKI, osimertinib, given for a longer period of time (3 years vs. 2 years) would improve PFS and OS after surgery and standard adjuvant therapy (2). The trial had a very complicated design as patients could receive or not receive adjuvant chemotherapy at the choice of the trial sites and patients were randomized to osimertinib or placebo after the decision about adjuvant chemotherapy. Thus, there were 4 treatment groups (no chemo and placebo TKI (n = 136); no chemo and osimertinib (n = 137); chemo and placebo (n = 207), and chemo and osimertinib (n = 202). The primary DFS and OS analyses were based on a comparison only of those who received osimertinib or placebo. The DFS was strongly in favor of osimertinib for DFS with a HR of 0.30. The early survival data were similar with 3-year OS exceeding 90% in both groups. No survival analyses comparing those who received adjuvant chemotherapy with or without osimertinib versus those who received no chemotherapy with or without osimertinib have been reported. If we recall that the HR for meta-analyses of adjuvant chemotherapy's effect on OS was only 0.89, requiring many thousands of patients to show statistical improvement (7). If we analyze the ADAURA groups with and without adjuvant chemotherapy separately, it seems highly unlikely that there will be any statistically significant survival differences in the final analyses.

Patients who are early stage and completely resected are asymptomatic from their disease and thus treatment cannot improve symptoms. There is no question that adjuvant osimertinib will produce increased toxicity and increased costs but what about HRQoL? The results of Majem and colleagues demonstrate that that the toxicity of the osimertinib did not impair HRQoL in a measurable way (1). Another positive aspect of adjuvant osimertinib is a reduction in CNS relapse because osimertinib crosses the blood–brain barrier. Is preventing these CNS relapses in a few patients better than treating those who relapse and not subjecting cured patients to years of adjuvant therapy? Majem and colleagues concluded that “the HRQoL data provide support for the adjuvant use of osimertinib and provide “clinically meaningful improvement in DFS.” We believe that such conclusions cannot be made until the OS data become available. And if there is no survival benefit, where do we go from here?

There is little debate that the goal of therapy for patients with early-stage NSCLC is to cure their disease and thus improve overall survival. Such cures rates and long survival occur in more than 50% of patients with early-stage NSCLC in recent adjuvant therapy trials with surgery and adjuvant or neoadjuvant chemotherapy (7). These cured patients would not benefit from additional adjuvant therapy and bear the expense and toxicity of any such therapy. These cured patients should be excluded from future trials and their exclusion could reduce future trial sample sizes. While there are no established means of identifying patients who are cured, a number of studies are showing that measures to assess minimal residual disease (MRD) may define patients with high and low risk of recurrence (8). These MRD assessments such as circulating DNA should be incorporated into future trial designs to exclude patients least likely to recur such as in the ongoing MERMAID trial.

It is reassuring to know that adjuvant osimertinib does not impair HRQoL but this will be most relevant if the 3 years of therapy improve cure rates and overall survival.

T. Patil reports personal fees from AstraZeneca, Pfizer, Genentech/Roche, Takeda, Mirati Therapeutics, Janssen, EMD Serono, Sanofi, and Elevation Oncology outside the submitted work. P.A. Bunn Jr reports personal fees from AstraZeneca, Eli Lilly, Verastem Oncology, CStone, Ascentage, and Viecure outside the submitted work. No other disclosures were reported.