A single-nucleotide polymorphism (SNP) is an inheritable single nucleotide variation between members of species or paired chromosomes. SNP makes the difference in the susceptibility to the development of cancer when driver mutations occur. In addition to the genetic change, accumulated profiles of active transcriptional epigenetic markers (such as H3K4me1, H3K4me2, H3K4me3, H3K9ac and H3K27ac) and transcriptional repressive markers (such as H3K9me3 and H3K27me3) at SNPs greatly influence the variable possibility to oncogenesis.   

You can read these articles

"Emerging patterns of somatic mutations in cancer." by Watson 2013.

"Patterns of somatic mutation in human cancer genomes" by Greenman 2007.

"Statistical Analysis of Pathogenicity of Somatic Mutations in Cancer" by Greenman 2006.

A single-nucleotide polymorphism (SNP) is an inheritable single nucleotide variation between members of species or paired chromosomes. SNP makes the difference in the susceptibility to the development of cancer when driver mutations occur. In addition to the genetic change, accumulated profiles of active transcriptional epigenetic markers (such as H3K4me1, H3K4me2, H3K4me3, H3K9ac and H3K27ac) and transcriptional repressive markers (such as H3K9me3 and H3K27me3) at SNPs greatly influence the variable possibility to oncogenesis.   

It would be interesting to search COSMIC with a list of sites from dbSNP (or 1000 genomes project) to see whether any known somatic mutations are co-localised with SNPs.

(although obviously a germline polymorphism in one individual could occur as a somatic mutation in a different individual - really a baseline sample from the individual is needed to identify their own set of germline polymorphisms before somatic mutations occur)

Surely, we know that some people are more prone to the cancer than the others and this fact has roots in SNP variations in important molecules like p53. DNA repair proteins and the cell cycle check point molecules each have several SNP variants with a their specific degree of function. It is proportional that the existence of specific SNPs or de novo mutations in vulnerable sites can lead to the elevated risk of error and eventually the cancer.

I believe that a polymorphism, in the purest sense is meant to mean that the genetic change is present in a significant proportion of the population and is not pathogenic.  Tumors may harbour many somatic mutations, some of which are driving the malignancy to proliferate and spread, whereas others are just along for the ride.  Its possible, by chance, that such a passenger mutation would be identical to a known polymorphism and thus a SNP would be present a somatic change in a tumor.  However, if that change was driving tumor growth, then it probably shouldnt be classified as a polymorphism as it is in fact pathogenic.