Updated survival analysis of JBR.10: A randomized phase III trial of vinorelbine/cisplatin versus observation in completely resected stage IB and II non-small cell lung cancer (NSCLC)

7501

Background: JBR.10 was one of a number of phase III trials that established adjuvant cisplatin based chemotherapy as a recommended treatment in completely resected NSCLC . Long-term follow-up of these trials is important to document persistent benefit and potential late toxicities of adjuvant therapy. We report the updated survival data for JBR.10 with more than 9 years median follow up. Methods: Patients with completely resected stage IB (T2N0) or II (T1–2N1) NSCLC were randomized to receive 4 cycles of vinorelbine/cisplatin or observation.. Kaplan-Meier curves were generated for overall (OS) and disease specific survival (DSS). Log-rank test was used to compare survival distribution and to test cause specific hazard. For the competing risk analysis, the Gray test was used to test the difference in cause specific incidences. All efficacy analyses were done on an ITT basis. Results: 482 patients were randomized. Data cut-off for this update was July 2008. Median follow-up is 9.3 years (3.2–13.8 y). 12 patients were lost to follow up, a median 4.9 years from randomization (1.5–12 years). 271 deaths have occurred, 73% due to lung cancer or its treatment. Survival analysis continues to show a benefit for chemotherapy: HR .78 (CI .61-.99, p=.04). The benefit appears to be confined to N1 patients: median OS 6.8 y versus 3.6 y, HR .68 (CI .5-.92, p=.01). N0 patients did not appear to benefit: HR 1.03 (CI .7–1.52, p=.87). Chemotherapy significantly prolonged DSS, HR.73 (CI .55-.97, p=.03) Competing risk analysis showed observation to be associated with significantly higher risk of death from lung cancer (p=.02) with no difference in incidences of death from other causes between arms (p=.62). Conclusions: Prolonged follow-up of patients in the JBR.10 trial continues to show a benefit in survival for adjuvant chemotherapy. This benefit appears to be confined to N1 patients. There was no increase in death from other causes in the chemotherapy arm.

[Table: see text]

TGF-α and amphiregulin levels in non-small cell lung cancer (NSCLC) patients (pts) treated with erlotinib/placebo in the National Cancer Institute of Canada Clinical Trials Group (NCIC CTG) BR.21

11023

Background: In BR.21, erlotinib prolonged survival of previously treated NSCLC pts. We examined the predictive/ prognostic effects of the EGFR ligands, TGFa and amphiregulin(Am). Methods: Plasma was collected prior to treatment in consenting pts. TGFa and Am were analyzed by ELISA (R&D Systems). Samples were blinded, measured in duplicate and ligand concentrations determined following interpolation of a standard curve generated from known quantities of recombinant proteins. Sensitivity was ∼3pg/ml. Cutoff points for TGFa were <10 (low) and ≥10 (high); for Am <1.5 (low) vs 1.5–10 (intermediate) vs >10 (high). Standard statistical methods were used to correlate biomarker data with baseline characteristics and outcomes. Results: In 731 randomized pts, 539 were evaluable for both markers; there was a significant interaction (p=0.03) between the evaluable and inevaluable cohorts (erlotinib benefit favoring the evaluable cohort); evaluable pts were more likely to be male, > 60 and to have EGFR amplification. Baseline high TGFa was associated with poorer ECOG PS (p=0.03) while high Am was associated with worse PS (p<0.0001), anemia (p<0.0001), non response to prior therapies (p=0.01) and no-prior platinum therapy (p=0.02). Although TGFa was not prognostic for OS (HR=1.0, p=0.99), high Am was (HR: Middle vs low: 0.94, 95% C.I 0.49 -1.81; high vs.low: 2.04, 95% C.I 1.43 - 2.91; p=0.0002). High TGFa was a poor prognostic factor for PFS (HR 1.4, p=0.07) as was elevated Am (HR: Middle vs low: 1.02, 95% C.I 0.56 -1.85; high vs. low: 1.67, 95% C.I 1.19 - 2.35, p=0.01). In multivariate analyses including both ligands, high Am remained prognostic, but neither ligand was predictive. Conclusions: High levels of baseline Am appears to be a poor prognostic factor, while low levels of TGFa showed a trend to predict benefit from erlotinib in univariate analyses. [Table: see text]

[Table: see text]