A prognostic multigene classifier for squamous cell carcinomas of the larynx

Survival after diagnosis of laryngeal cancer has not improved over the last 20 years. Selection of patients for radio- and chemotherapy or surgery or follow-up strategies based on a prognostic classifier could improve survival without unduly extending radical surgery. We performed microarray gene expression analysis and developed a four-gene classifier for laryngeal cancer using Prediction Analysis of Microarray and leave-one-out cross validation. A four-gene classifier containing the non-coding gene H19, the histone HIST1H3F and the two small nucleolar RNAs, SNORA16A and SNORD14C was developed that assigns cases to low and high risk classes. The high risk class has a relative risk of 6.5 (CI=1.817-23.258, Fisher exact test p<0.0001). The maternally imprinted gene H19 is the top classifier gene.