Lymphoma of germinal center cell (GC) origin generally is an indolent malignancy that transforms progressively into a more aggressive disease. According to the World Health Organization classification, lymphomas of follicular center cell origin are classified as either large B-cell lymphoma (LBCL) or follicular lymphoma (FL). The authors tested the feasibility of performing gene expression profiling using amplified RNA from fine-needle aspirates (FNA) obtained from lymph nodes. Twenty-four samples from patients with a diagnosis of FL or LBCL were obtained after Institutional Review Board-approved informed consent was obtained. The diagnoses were confirmed by 2 pathologists and were classified into 2 groups (10 LBCL samples and 14 FL samples) by using conventional morphology and immunophenotyping. One hundred nanograms of total RNA were subjected to 2 cycles of standard, double-stranded complementary DNA synthesis and in vitro transcription for target amplification using a small-sample target-labeling protocol. The biotinylated cRNA from each sample was hybridized to gene chips. Gene expression profiling results were analyzed first by principal-component analysis (PCA) by using a list of 146 probe sets that represented 62 genes that are characteristic of an activated B-cell (ABC) signature or a GC signature. The analysis identified 5 LBCL samples with an ABC cell signature. Using a list of 207 probe sets that represented 113 genes involved in FL transformation, PCA analysis identified 2 overlapping clusters corresponding to FL and GC-diffuse LBCL. To improve this classification further, the authors generated a list of 72 genes that were expressed differentially between FL and GC-LBCL. Using this list of genes, PCA analysis demonstrated a clear separation between FL and GC-LBCL. However, five FL samples clustered as an intermediate group between FL and GC-DLBCL. These samples were characterized morphologically by a mixed cell pattern with relatively fewer large, noncleaved lymphocytes and more small, cleaved lymphocytes. The results support the feasibility of FNA-based transcription profiles in patients with FL or LBCL, which, in combination with morphology and immunophenotyping, can help in the subtyping of these entities.