Counting alleles to predict recurrence of early-stage colorectal cancers

Background Chromosome imbalances occur in many cancers and represent important biological properties of tumours. However, measurements of such imbalances are difficult. We used a new, quantitative approach to investigate the prognostic value of chromosome imbalances in early-stage colorectal cancers. Methods We studied 180 patients with no evidence of lymph-node or distant metastases at the time of surgery. DNA from paraffin-embedded tumours was tested for imbalances of chromosome 8p and 18q by digital SNP (single-nucleotide polymorphism)—a technique in which each allele in a sample is directly counted. Surviving patients had median follow-up of 68 months, and disease recurrence was used as the clinical endpoint. Findings Tumours were divided into three groups: “L” tumours (n=93) had allelic imbalances of chromosomes 8p and 18q, “L/R” tumours (n=60) had allelic imbalances of either chromosome 8p or 18q but not both, and “R” tumours (n=27) retained allelic balance for both chromosomes. 5-year disease-free survival was 100% (95% CI 80–100) for patients with R tumours, 74% (61–87) for patients with L/R tumours, and 58% (47–69) for those with L tumours. These differences were significant (p<0•0001) and were independent of other variables—eg, Dukeʹs stage A tumours of class L were much more likely to recur than Dukeʹs stage B tumours of class R (p=0•002). Interpretation In patients without metastasis, allelic imbalance is a better predictor of prognosis than histopathological stage.