[W]e believe that the survival results and possibly the disease response results are false negative. However, the trial clearly demonstrates feasibility of treatment assignment for patients with advanced NSCLC across countries and academic, nonacademic, and private practice settings.
—Gerold Bepler, MD, PhD, and colleagues
A trial reported by Gerold Bepler, MD, PhD, of Karmanos Cancer Institute, Detroit, and colleagues in Journal of Clinical Oncology assessed whether chemotherapy selected on the basis of in situ ERCC1 and RRM1 protein levels could improve outcomes in patients with advanced non–small cell lung cancer (NSCLC).1
In this international phase III trial, 275 patients were randomly assigned 2:1 to an experimental or control arm. The experimental arm (n = 183) consisted of four groups according to ERCC1 and RRM1 status: patients with low-ERCC1 (≤ 66.0) and low-RRM1 (≤ 40.5) received gemcitabine at 1,250 mg/m2 on days 1 and 8 and carboplatin AUC 5 on day 1 every 21 days (n = 56); those with low-RRM1/high-ERCC1 received gemcitabine at 1,250 mg/m2 on days 1 and 8 and docetaxel at 40 mg/m2 on days 1 and 8 every 21 days (n = 37); those with high-RRM1/low-ERCC1 received docetaxel at 75 mg/m2 on day 1 and carboplatin AUC 6 on day 1 every 21 days (n = 26); and those with high-ERCC1/high-RRM1 received vinorelbine at 35 mg/m2 on days 1 and 15 and docetaxel at 50 mg/m2 on days 1 and 15 every 28 days (n = 64).
All patients in the control arm (n = 92) were treated with gemcitabine plus carboplatin regardless of protein levels. Among the 92 patients, profiles were low-RRM1/low-ERCC1 in 33 patents, high-RRM1/low-ERCC1 in 15, low-RRM1/high-ERCC1 in 19, and high-RRM1/high-ERCC1 in 25.
There were no significant differences between the experimental arm and the control arm with regard to baseline demographic or disease characteristics. The trial was powered to detect a 32% improvement in 6-month progression-free survival. Patients in both arms received a median of four treatment cycles, with there being no difference among the different treatment groups in the experimental arm in number of cycles received. Median follow-up times were 32.9 months in the experimental arm and 26.3 months in the control arm.
There were no significant differences in progression-free survival between the experimental arm and the control arm or among the different treatment groups in the experimental arm. The 6-month progression-free survival rate and median progression-free survival were 52.0% and 6.1 months in the experimental arm and 56.5% and 6.9 months in the control arm.
Median progression-free survival durations in the experimental arm were 5.0 in the gemcitabine/carboplatin group, 6.5 in the docetaxel/carboplatin group, 6.1 in the gemcitabine/docetaxel group, and 6.7 months in the docetaxel/vinorelbine group. By comparison, median progression-free survival durations in the four counterpart protein-expression subgroups in the control arm were 8.1, 8.1, 6.0, and 6.3 months for the low-RRM1/low-ERCC1, hihg-RRM1/lowERCC1, low-RRM1/high-ERCC1, and high-RRM1/high-ERCC1 subgroups, respectively.
The only significant difference in the subgroup comparisons was a prolonged progression-free survival in the low-RRM1/low-ERCC1 control subgroup vs the gemcitabine/carboplatin group in the experimental arm. These two subgroups served as an internal control in the trial, since patients had the same protein expression profile and received the same treatment. The finding of a significant difference between them was thus unexpected, and, along with other considerations, raises the possibility the survival results of the trial are false-negative.
There were no differences in overall survival between the experimental arm and the control arm, among treatment groups in the experimental arm, or between the experimental arm treatment groups and their counterpart protein expression subgroups in the control arm. Median overall survival was 11.0 months on the experimental arm and 11.3 months in the control arm. Objective response rates (all partial responses) were 38.8% in the experimental arm and 36.5% in the control arm.
There were no significant differences in total adverse events, serious adverse events, or treatment discontinuation due to toxicity between the experimental and control arms or among the treatment groups in the experimental arm.
Approach Is Feasible
In assessing the feasibility of protein analysis to guide chemotherapy selection, it was determined that the trial would meet its feasibility endpoint if more than two-thirds of registered patients had successful protein analysis. Of the 331 registered patients, protein levels were successfully determined in 301 (91%).
The median time from informed consent to completed gene analysis was 11 days (range, 1–47 days). Rebiospsy for the specific purpose of gene expression analysis was required in 17% of patients. Of specimens used for protein analysis, 202 were histologic and 103 were cytologic, with no significant difference between procedures in successful vs unsuccessful initial protein analysis.
Reasons for False-negative Results
As noted, the investigators believe that the survival results, and perhaps the disease response results, are false-negative, as suggested by the significantly better progression-free survival observed in the control subgroup of low/low protein expression compared with the experimental group with low/low expression, both of which received gemcitabine/carboplatin. No differences in parameters known to affect survival were found between the two groups.
A number of factors may have resulted in inaccurate protein expression measurements and inaccurate assignment of patients to protein expression groups and thus contributed to generation of false-negative results. No significant correlations between best disease response and protein or mRNA ERCC1 or RRM1 levels were observed in patients receiving gemcitabine/carboplatin, as has been previously reported. This finding may be related to interinstitutional variations in specimen collection and processing observed in the current trial and in other studies.
There was also no correlation between ERCC1 protein and mRNA levels. This may be related to the fact that the reagent used for in situ measurement (monoclonal antibody 8F1) may detect other molecules of similar immunoreactivity and cellular localization as ERCC1. In contrast to earlier findings, data obtained from fresh-frozen tumor specimens showed no correlation of RRM1 protein and mRNA levels in the current trial.
The investigators commented, “[We]believe that the survival results and possibly the disease response results are false negative. However, the trial clearly demonstrates feasibility of treatment assignment for patients with advanced NSCLC across countries and academic, nonacademic, and private practice settings.”
They continued, “We conclude that further assay development with special attention to reagent specificity, day-to-day assay conditions, and site-specific specimen processing is desirable before another trial is launched.” ■
Disclosure: For full disclosures of the study authors, visit jco.ascopubs.org.
1. Bepler G, Williams C, Schell MJ, et al: Randomized international phase III trial of ERCC1 and RRM1 expression-based chemotherapy versus gemcitabine/carboplatin in advanced non-small-cell lung cancer. J Clin Oncol. May 20, 2013 (early release online).