Erlotinib (TARCEVA) in pancreatic carcinoma? Two years ago, erlotinib (TARCEVA) was licensed for the treatment of patients with locally advanced or metastatic non-small cell lung cancer, who had at least one previous failed chemotherapy. For a year, the tyrosine kinase inhibitor has also been licensed in combination with gemcitabine (GEMZAR) for patients with metastatic pancreatic carcinoma. As recently as summer 2006, Committee for Medicinal Products for Human Use (CHMP) of EMEA, the European Medicines Agency, had refused an extension of the indication to patients with locally advanced or metastatic pancreatic cancer because of a negative benefit-risk balance: in the only phase III study submitted for this indication, which included 596 patients, the addition of erlotinib to gemcitabine prolonged overall survival (primary endpoint) compared to placebo by a median of only twelve days (6.34 versus 5.93 months), which was statistically significant. As regards the progression-free survival, one of the secondary endpoints, the difference was only six days (3.75 vs. 3.55 months). A positive influence on quality of life was not demonstrated. On the contrary: significantly more patients on erlotinib complained of diarrhoea. Adverse events classified as serious occurred more frequently with the tyrosine kinase inhibitor (51% vs. 39%) The death of five patients (2%), all in the erlotinib arm, was attributed to the toxicity of the treatment regime (EMEA: European public assessment report [EPAR] TARCEVA, Oct. 2007; available at: http://www.emea.europa.eu/htms/human/epar/t.htm). A few months later, however, the CHMP recommends extension of the licence for patients with metastatic pancreatic carcinoma. However, there are no new data - the reasons given for the change of mind are a subsequent subgroup analysis and the advice of leading European cancer specialists (EMEA: Questions and answers on the change to the marketing authorisation for TARCEVA of 14th Dec. 2006). Whether all experts made a free and independent assessment is, however, doubtful: one of the four experts stated to the European Agency that he had received lecture fees from Roche, the supplier of erlotinib. Another, who had not listed any connections with Roche in a declaration of conflicts of interest signed in June 2006, declared in an article submitted in November 2006 and published in May 2007 that he had worked as an adviser for the company and had received payments both from Roche and from its subsidiary Genentech (EMEA: letter of 29 June 2007; SANOFF, H.K. et al.: J. Clin. Oncol. 2007; 25: 1891-7). The CHMP finally approved extension of erlotinib's licence although the experts differed very much in their assessments of the validity of the study, the clinical significance of the result and the observed toxicity of the tyrosine kinase inhibitor. We advise against the use of erlotinib in pancreatic carcinoma, -ed.