Table 1.

Evidence for Biomarker Use in Advanced NSCLC.

Trials	Number of patients	Findings
Histology
7 trials – E4599 plus randomized phase II trial, BR21, SWOG database analysis (S9806, S0003, S9308), JMEI, JMBD, JMEN	5408	Adenocarcinoma histology prognostic for ↑OS. Adenocarcinoma predictive of ↑ PFS and OS for bevacizumab. Squamous ca predictive of toxicity from bevacizumab. Adenocarcinoma predictive of ↑ ORR, but not OS for erlotinib. Non-squamous histology predictive of ↑ORR, PFS and OS for pemetrexed
EGFR protein expression (IHC)
6 trials – BR21, ISEL, IPASS, INTEREST, SATURN, TALENT	2691	EGFR protein expression predictive of ↑ORR and OS in BR21 and ISEL, but not predictive of TTP or OS in other studies
EGFR copy number (FISH or qPCR)
9 trials – BR21, ISEL, IPASS, INTEREST, SATURN, TALENT, TRIBUTE, INTACT I/II	2994	Some evidence that EGFR gene copy number prognostic for ↓ OS EGFR gene copy number predictive of ↑OS in BR21 and ISEL, but not predictive of PFS or OS in other trials
EGFR mutation
14 trials – BR21, ISEL, SATURN, IPASS, First Signal, WJTOG3405, NEJ002, INTEREST, IDEAL I/II, INTACT I/II, TALENT, TRIBUTE	3259	EGFR gene mutations prognostic for ↑ OS. EGFR mutations predictive of ↑ ORR, but not OS in 2nd/3rd line therapy. In studies comparing EGFR TKI to chemotherapy, EGFR mutations predictive of ↑ ORR and PFS. OS data are immature
K-Ras mutations
8 trials	2442	Mixed information re prognostic value of K-Ras mutations. Likely weakly prognostic for ↓ OS. BR10 suggests K-Ras mutations predict lack of OS benefit from adjuvant chemotherapy. Two studies suggestive K-Ras predictive of ↓ OS for patients on EGFR TKI, while 4 trials show K-Ras not predictive of ORR, PFS or OS differences