作者: Nicola Personeni , Steffen Fieuws , Hubert Piessevaux , Gert De Hertogh , Jef De Schutter
DOI: 10.1158/1078-0432.CCR-08-0449
关键词:
摘要: Purpose: To evaluate the usefulness and pitfalls inherent to assessment of epidermal growth factor receptor (EGFR) gene copy number (GCN) by fluorescence in situ hybridization (FISH) for outcome prediction cetuximab metastatic colorectal cancer. The value testing KRAS mutation status, addition EGFR GCN, was also explored. Experimental Design: FISH analysis 87 cancer patients treated with done, recording individual GCN per cell using different samples tumor. Performances published cutoff points summaries distribution were assessed response prediction. Results: In our data set, two performed less well than their training yielding positive predictive values negative between 40.0% 48.3% 81.0% 86.5%, respectively. Among explored, mean right-tailed yielded highest performances. A >= 2.83 provided an area under curve 0.71. Important heterogeneity repeated measures observed within tumors (intraclass correlation, 0.61; within-class SD, 0.40), leading potential misclassifications status 7 18 (38,8%) if a point used. multivariable analysis, significant information independent predict (P = 0.016) overall survival 0.005). Conclusions: We confirm association increased after cetuximab. However, because reproducibility concerns, any decision making based on is not warranted.