Carcinogen exposure from tobacco smoking is the major cause of upper aerodigestive tract cancer, yet heavy smokers only have about a 10% life-time risk of developing one of these cancers. Current technologies allow only limited prediction of cancer risk and there are no approved screening methods applicable to the general population. We developed a method to assess somatic mutational load using small-pool PCR (SP-PCR) and analysed mutations in DNA isolated from cells obtained by mouth rinse. Mutation levels in the hypermutable tetranucleotide marker D7S1482 were analysed in specimens from 25 head and neck squamous carcinoma (HNSCC) cases and 31 controls and tested for associations with age, smoking history and cancer status. We found a significant association between mutation frequency and age (P=0.021, Generalized Linear Model (GLM), N=56), but no influence of smoking history. Cases had higher mutation frequencies than controls when corrected for the effects of age, a difference that was statistically significant in the subgroup of 10 HNSCC patients who were treated with surgery only (P=0.017, GLM, N=41). We also present evidence that cancer status is linked to levels of nonunique, and presumably clonally derived, mutations in D7S1482. Insertion mutations were observed in 833 (79%) of 1058 alleles, of which 457 (43%) could be explained by insertion of a single repeat unit; deletion mutations were found in 225 (21%) of tested alleles. In conclusion, we demonstrate that the sensitive detection of single molecule mutations in clinical specimens is feasible by SP-PCR. Our study confirms an earlier report that microsatellite mutations increase with age and is the first to provide evidence that these mutations may be associated with cancer status in individual subjects.