Abstract 4113: Kinase activity based biomarkers: Identification of prognostic and erlotinib response prediction markers in NSCLC patients

Abstract

Abstract Background: Reliable diagnostic tests are needed to identify early stage non-small cell lung carcinoma (NSCLC) patients with poor prognosis. Concomitantly there is a clear need for tests that enable the selection of patients who will benefit from targeted therapy with kinase inhibitors. We evaluated kinase activity profiles in two groups of early stage NSCLC patients, either for prognosis of long- or short-term survival, or for predicting erlotinib drug response. Method: Retrospective studies were performed on fresh frozen resection material of two groups of early stage NSCLC patients. The first group consisted of 48 short- and long-term survivors who underwent a complete surgical resection (5+ years follow-up). The second group consisted of 14 NSCLC patients who received 3 weeks of neo-adjuvant treatment with erlotinib prior to complete surgical resection. Response evaluation to neo-adjuvant treatment was based on histopathological examination of the surgical specimens. For both studies, kinase activity profiles of lysed cryosections of tumour tissues were generated in the presence and absence of protein tyrosine kinase inhibitors on PamChip® peptide micro-arrays, comprising 144 tyrosine containing peptides, derived from known phosphorylation sites of human proteins. Partial least square discriminant analysis was used to construct prediction models. ClustalW alignment algorithms were used to investigate the most informative phosphorylation sites. Results: Kinase activity profiles obtained in the absence of inhibitor did not distinguish between subgroups (long- versus short-term survival, responder or non-responder to TKI), whereas ratios of inhibited versus non-inhibited signals resulted in distinct classifiers predicting survival for the first group, and response for the second group. Multivariate unsupervised analysis with leave-one-out cross-validation resulted in an error rate for survival prediction of 29%. In the drug response prediction 13 out of 14 patients were correctly classified. Conclusion: This is the first study to show that kinase activity profiles of tumour tissue exposed to a kinase inhibitor can be used to identify NSCLC patients likely to respond to erlotinib treatment. Furthermore, based on kinase activity profiles of early stage NSCLC tumours, a prognostic classifier, for a set of 48 patients, was obtained. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 4113. doi:10.1158/1538-7445.AM2011-4113