Sir,

We wish to thank Sesboüé et al (2012) for their interest in our study on the prediction of anti-EGFR therapy in colorectal cancer (CRC) (Ålgars et al, 2011). While, as they discuss, the mechanism of gene copy number (GCN) alterations may affect the response to targeted treatments, this question was beyond the focus of our work. Our aim was to test a simple hypothesis: does EGFR immunohistochemistry (IHC)-guided silver in situ hybridisation analysis predict treatment response better than previously used methods. Our hypothesis was based on the fact that EGFR expression in CRC, as examined by IHC, is heterogenous within tumours (Moroni et al, 2005; Ålgars et al, 2011). Thus, EGFR GCN (or Chr-7 polysomy) analysis from areas with highest IHC intensity might reflect the tumour's sensitivity to anti-EGFR Abs better than unguided fluorescence in situ hybridisation analysis. We correlated the results with three different clinical parameters; clinical benefit, as evaluated by RECIST criteria, progression-free survival (PFS), and overall survival (OS). The results showed that in unselected or KRAS wild-type (WT) tumours, EGFR GCN increase (cutoff 4.0) significantly predicts outcome by all three parameters. The mean PFS in the KRAS WT/EGFR GCN high group was three times longer than in the KRAS WT/EGFR GCN low group, and the OS of the KRAS WT/EGFR GCN high patients was four times longer than OS of the KRAS WT/EGFR GCN low patients.

The EGFR/Chr-7 ratio did not in our analysis have any predictive value. While 51 out of 78 tumours had an EGFR GCN over the cutoff value 4.0, only two of them had an EGFR/Chr-7 ratio >2 indicating that pure EGFR amplification in CRC is a rare event (at least when evaluated by our methods). Thus, EGFR GCN increase in association with Chr-7 polysomy appears to be the prevalent pattern, which is associated with responsiveness to anti-EGFR treatment. Sesboüé et al (2012) site four studies to support their claim that only ‘true’ EGFR amplification would be meaningful for the treatment response. Of these studies, Cascinu et al. correlated EGFR GCN changes to response with EGFR small molecular inhibitor, gefitinib, which is not used in CRC (Cascinu et al, 2008). One of the other three studies lacks information of the KRAS status of the tumours (Razis et al, 2008), and only one of these three studies correlated EGFR alterations with time to progression (Razis et al, 2008), whereas the other two studies assessed merely the response rates to anti-EGFR therapy (Moroni et al, 2005; Frattini et al, 2007). In our understanding, none of the studies therefore support the conclusion of Sesboüé et al (2012). As a final note, the dogma of the importance of the gene/chromosome ratio seems to be falling apart also in the case of HER2. A recent study of 1888 breast cancer patients treated with or without trastuzumab demonstrated that trastuzumab benefit was independent of HER2/centromere 17 ratio and Chr-17 copy number (Perez et al, 2010).