1271P - Analysis of PFS and OS from the SATURN study according to EGFR IHC status using the H-score reading method

Date 29 September 2012
Event ESMO Congress 2012
Session Poster presentation I
Topics Biomarkers
Non-Small-Cell Lung Cancer, Metastatic
Presenter Julien Mazières
Authors J. Mazières1, W. Brugger2, F. Cappuzzo3, P. Middel4, A. Frosch5, I. Bara6, G. Klingelschmitt7, B. Klughammer8
  • 1Thoracic Oncology Unit, Larrey Hospital, 31400 - Toulouse/FR
  • 2Hematology/oncology, Schwarzwald-Baar Klinikum, University of Freiburg, Villingen-Schwenningen/DE
  • 3U.o. Oncologia Medica, Istituto Toscano Tumori, Livorno/IT
  • 4Department Of Pathology, Targos Advance AG, Kassel/DE
  • 5Department Of Pathology, Göttingen University Hospital, Gottingen/DE
  • 6Global Medical Affairs Oncology, F. Hoffmann-La Roche Ltd, Basel/CH
  • 7Mdbb Biometrics, F. Hoffmann-La Roche Ltd, Basel/CH
  • 8Department Of Biomarkers, F. Hoffmann-La Roche Ltd, Basel/CH

Abstract

Background

In the phase III SATURN study, maintenance erlotinib significantly prolonged progression-free survival (PFS) vs placebo in patients (pts) with advanced non-small-cell lung cancer (NSCLC) and non-progressive disease after first-line chemotherapy (Cappuzzo et al, Lancet Oncol 2010). Epidermal growth factor receptor (EGFR) expression analysis by immunohistochemistry (IHC) found no significant difference in PFS (p = 0.63) or overall survival (OS; p = 0.52) with erlotinib by EGFR IHC status (Brugger et al, J Clin Oncol 2011). Recently, Pirker at al. (Lancet Oncol 2012) presented data on EGFR expression as a predictor of OS for first-line chemotherapy plus cetuximab in the phase III FLEX study, using a novel method (H-score) to assign EGFR IHC status. We used this method to reassess samples from the SATURN study.

Methods

The H-score method assigns an IHC score to each pt on a continuous scale of 0–300, based on the percentage of cells at different staining intensities (Pirker et al, Lancet Oncol 2012). As per this method, the outcome-based discriminatory threshold IHC H-score for our analysis was set at 200 and existing samples were re-read and classed as low (H-score <200) or high (≥200) EGFR protein expression. An exploratory cut-off was also assessed, at <10% or ≥10% membrane staining. PFS and OS (in terms of hazard ratios [HRs] and 95% confidence intervals [CIs] by Cox model, and log-rank p-values), were reanalysed based on these new classifications.

Results

The original EGFR IHC analysis used samples from 370 and 372 pts in the erlotinib and placebo arms, respectively. This analysis examined existing samples from 351 and 361 pts in the erlotinib and placebo arms, respectively. PFS and OS according to EGFR IHC by different methods are shown in the Table.

Conclusions

Re-scoring of EGFR IHC status in SATURN by H-score found similar benefit of erlotinib treatment in PFS or OS for subsets with high or low EGFR expression, in overall or EGFR WT populations.

Table: 1271P

SATURN protocol-defined EGFR IHC+ SATURN protocol-defined EGFR IHC– EGFR IHC byH-score ≥ 200 (high) EGFR IHC byH-score < 200 (low) EGFR IHC byH-score ≥ 10% EGFR IHC byH-score < 10%
n erlotinib: 308 placebo: 313 erlotinib: 62 placebo: 59 erlotinib: 146 placebo: 157 erlotinib: 205 placebo: 204 erlotinib: 267 placebo: 287 erlotinib: 84 placebo: 74
PFS 0.69 (0.58–0.82) log-rank p < 0.0001 0.77 (0.51–1.14) log-rank p = 0.1768 0.68 (0.53–0.86) log-rank p = 0.0010 0.76 (0.62–0.93) log-rank p = 0.0076 0.71 (0.59–0.84) log-rank p < 0.0001 0.79 (0.57–1.10) log-rank p = 0.1627
OS 0.77 (0.64–0.93) log-rank p = 0.0063 0.91 (0.59–1.38) log-rank p = 0.6482 0.80 (0.62–1.05) log-rank p = 0.1099 0.80 (0.64–1.01) log-rank p = 0.0639 0.78 (0.64–0.96) log-rank p = 0.0161 0.85 (0.60–1.22) log-rank p = 0.3901
n (EGFR WT) erlotinib: 170 placebo: 157 erlotinib: 28 placebo: 27 erlotinib: 86 placebo: 84 erlotinib: 103 placebo: 97 erlotinib: 147 placebo: 144 erlotinib: 42 placebo: 37
PFS in WT 0.79 (0.63–0.99) log-rank p = 0.0336 0.65 (0.37–1.13) log-rank p = 0.1146 0.69 (0.51–0.95) log-rank p = 0.0188 0.84 (0.63–1.12) log-rank p = 0.2166 0.78 (0.62–0.99) log-rank p = 0.0400 0.69 (0.44–1.10) log-rank p = 0.1126
OS in WT 0.77 (0.60–0.99) log-rank p = 0.0402 0.64 (0.35–1.20) log-rank p = 0.1608 0.78 (0.55–1.10) log-rank p = 0.1563 0.76 (0.55–1.05) log-rank p = 0.0964 0.78 (0.60–1.02) log-rank p = 0.0698 0.72 (0.43–1.19) log-rank p = 0.2004

Note: HR < 1 is in favour of erlotinib

Disclosure

W. Brugger: Financial Interest: Advisory Board.

F. Cappuzzo: Financial Interest: Advisory board, F. Hoffmann-La Roche Ltd.

I. Bara: Employed by F. Hoffmann-La Roche Ltd.

G. Klingelschmitt: Employed by F. Hoffmann-La Roche Ltd.

B. Klughammer: Stock Ownership: F. Hoffmann-La Roche Ltd Employed by F. Hoffmann-La Roche Ltd.

All other authors have declared no conflicts of interest.