Next generation sequencing revealed RET or RAS mutation in medullary thyroid cancer that were negative at sanger sequencing

We sequenced 114 medullary thyroid cancer samples for RET and RAS mutations by Sanger sequencing (SS). Forty three /114 (37.7%) resulted to be RET/RAS negative. These samples were reanalyzed with a next generation sequencing system (ION S5) using a custom panel. Twenty-nine cases were confirmed to be RET/RAS negative and also negative for any other analysed oncogene. Fourteen cases became either RET or RAS positive. Three cases were positive for H-RAS (2 G12R, 1 Q61R) with a mean allelic frequency (AF) of 34.91%. These mutations were reanalyzed by SS and still considered as negative. Eight cases were positive for RET (4 M918T, 1 C634W, 1 C620R, 1 p.Asp898_Glu901del, 1 S1024F, 1 Q762K) with a mean AF of 13.66%. These mutations were reanalysed by SS that showed the presence of a very low peak for the mutated allele, that would not be considered as positive without the NGS information. The 3 other cases were also positive for RET but the discrepancy was due to technical problems: 1 case had a germline M918T RET mutation not revealed by SS for the presence of a single nucleotide polymorphism (SNP) at the 5’ end of one primer in the mutated allele that caused the amplification of the wild type allele only; 1 case had a mutation L56M in exon 2 that is not usually investigated with the routine SS; 1 case had a somatic M918T mutations revealed by SS only following an increase in the MgCl2 in the amplification mix. We also compared the AF of mutations found by NGS with the ratio between the height of the SS peak of the mutated and wild type allele. With this analysis we demonstrated that the ratio of the SS peaks corresponds to the AF at NGS. In conclusion, our data indicate that NGS identifies mutations in cases negative at Sanger due to the low frequency of the mutated allele and/or to the presence of conditions that allow the amplification of the wild type allele only (i.e. SNPs or other modifications) thus it should be preferred for diagnostic purposes. Furthermore, we show that although SS is not a quantitative method the ratio of the peak heights of the two alleles likely indicates the relative amount of the mutated allele.