Recently we identified a molecular basis for differentiating benign and malignant follicular thyroid tumors. The purpose of these studies was to determine whether molecular analysis can be used to differentiate papillary thyroid carcinomas from benign thyroid nodules. Gene expression patterns of 14 papillary thyroid carcinomas and 21 benign tumors were analyzed by oligonucleotide array analysis. The carcinomas included seven classical papillary thyroid carcinomas (PTC) and seven follicular variant of PTC (FVPTC), and the benign tumors included 14 follicular adenomas and seven hyperplastic nodules. A hierarchical clustering analysis was performed to examine the groups for potential differences. The combined PTC and FVPTC groups had a distinct gene expression profile compared with the benign lesions. The sensitivity for a diagnosis of carcinoma was 93%, with a 100% specificity (one FVPTC clustered with the benign nodules). Cancer gene profiles contained both known (Met and galectin-3) and previously unidentified genes. Gene profiling is a reliable means of distinguishing PTC, FVPTC, and benign tumors of the thyroid. These gene profiles may provide insight into the pathogenesis of papillary thyroid carcinoma and may ultimately enhance the preoperative diagnosis of thyroid nodules on a molecular basis.