Adjuvant TKI Therapy for Early-Stage EGFR Mutation-Positive NSCLC: Is Disease-Free Survival Enough?

H. Jack West, MD


July 07, 2017

The headline in ASCO Post seemed promising enough: "Adjuvant Gefitinib Delays Recurrence in EGFR-Positive NSCLC."[1] It describes the provocative result of a trial reported as highly positive at ASCO 2017,[2] in which 222 Chinese patients with N1 or N2 node-positive, resected, activating EGFR mutation-positive non-small cell lung cancer (NSCLC) were randomly assigned to receive either standard cisplatin/vinorelbine chemotherapy for four cycles or 2 years of adjuvant daily gefitinib at the standard dose of 250 mg daily. The highlight was a more than 10-month difference in median disease-free survival (DFS) (28.7 vs 18.0 months; hazard ratio, 0.60; P = .005); lower toxicities across many parameters with gefitinib; and more patients reporting improved quality of life using various measures if they received gefitinib rather than conventional chemotherapy.

Despite such promise at first blush, this study is mired in glaring shortcomings that undermine any argument that it should change practice, and it raises major questions about what our goal should be in running trials of novel therapies in a curative setting for patients with cancer.

This study is mired in glaring shortcomings that undermine any argument that it should change practice.

First, many people favor shaping the question as whether epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI) therapy should be added sequentially to adjuvant platinum-based chemotherapy, with its well-established survival benefit in higher-risk resected NSCLC, rather than replacing a proven therapy with an unproven one. This is especially true in a trial in which two thirds of the patients had stage IIIA disease, with a risk for recurrence in the range of 80% or more, and a setting in which adjuvant chemotherapy can significantly reduce that risk, far more than the difference achieved in lower-stage resected NSCLC.[3]

Moreover, the staging procedures in this trial were not clearly articulated but did not include uniform PET or brain imaging during preoperative staging, making it likely that many patients had more advanced disease than was reported. This concern is underscored by the finding that the DFS curves for both arms decline precipitously beyond 2 years out, nearly converging by 3 years (34% vs 27%, favoring gefitinib) and continuing to drop steadily after that, with no plateau for either arm. There is no suggestion of any plateau of prolonged DFS for either arm, which is a far worse result than we would expect to see in an adjuvant therapy trial—strongly suggesting that most of these patients had metastatic disease that was not detected in a potentially suboptimal preoperative staging workup.

With these severe limitations, it is unlikely that these results will or should lead to a marked change in recommendations for adjuvant therapy and don't suggest a clear mandate to do molecular marker testing for patients with resected NSCLC. I'm afraid that what we most clearly learned from this study was how even a phase 3 randomized trial can undermine its utility with poor rigor when enrolling appropriate patients.

The bigger question I see is whether an improvement in an endpoint such as DFS is an appropriate goal when we're treating for cure. The general premise of adjuvant therapy is that we administer it to eradicate micrometastatic disease that may persist after surgery, thereby converting a subgroup of patients with invisible metastatic disease from being destined to experience relapse and die of their disease to truly cured through additional systemic therapy. Although we might hope that DFS serves as a reliable early surrogate for overall survival (OS) here, the evidence in this setting suggests otherwise, particularly in studies of adjuvant targeted therapy, such as an EGFR TKI for patients with EGFR mutation-positive NSCLC. Single-arm[4] and randomized trials[5] have tested adjuvant erlotinib after standard adjuvant chemotherapy and reported improvement in DFS for patients with EGFR mutation-positive disease that was not associated with any evidence of an improvement in OS from what would otherwise be expected.

These data suggest that postoperative EGFR TKI therapy may not eradicate micrometastatic disease, but instead may only postpone progression after acquired resistance develops. If we consider giving 1-2 years of therapy and then reliably see a convergence of the DFS curves with no improvement in OS, I think this is an insufficient benefit to subject a large population of patients, some of whom are already cured with no further therapy, to prolonged adjuvant therapy. This is especially true if this added intervention entails skin, gastrointestinal, and sometimes other toxicities, as well as the financial toxicity of extremely expensive therapy that will be conferred on some patients as well as broader society.

We may need to debate these issues further as we consider whether the emerging results of the PACIFIC trial of consolidation durvalumab versus placebo for a year after concurrent chemotherapy and radiation for stage III, unresectable NSCLC warrant a change in practice. We've recently learned from a press release[6] that this trial is positive for a significant improvement in progression-free survival (PFS), which was the primary endpoint of the trial. However, PFS can be especially challenging to assess following the marked treatment-related changes in and around the primary tumor after chemotherapy/radiation. If the difference in PFS is profound, it will probably serve as an effective surrogate for improved OS, particularly because the cure rate in this locally advanced setting is only about 20% at best. But if it is only a modest improvement in PFS that won't necessarily translate to an OS benefit, a significant PFS benefit may not be sufficient to warrant an extra year of intravenous therapy for every patient, at a cost approaching $200,000 for the full course of treatment. For that price, both in patient time and financial burden, we should expect a profound survival benefit.

This question may become even more challenging if the preponderance of data from our adjuvant and consolidation therapy trials suggest that the benefits of these therapies dissipate after discontinuation of the treatment. If a fixed course of treatment is not conferring true cures, we will need to be very judicious in considering whether patients who may already be cured should receive longitudinal therapies that could possibly be just as effective when administered upon progression, but only to the patients with recurrent/progressing disease, who are then proven to actually need subsequent therapy.


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.