Written informed consent for participation in this study was obtained from all patients prior to surgery. A small portion of the tissue removed during the procedure was processed for the study after routine pathological analysis per the routine standard care for the patients. After review by a pathologist, a piece of each tumor specimen was de-identified from the patient and given a unique identification number and was immediately taken, without freezing, to be implanted in nude mice for xenograft generation. A separate piece of the same tumor was microdissected by a pathologist to assure that greater than 80% of tissue contained HNSCC prior to being snap frozen in liquid nitrogen. Three primary tumor specimens are included in this study. Seven normal tissue specimens were obtained from patients who underwent uvulopalatopharyngo- plasty for sleep apnea. After review by a pathologist, a section of dissected mucosal layer from discarded UPPP specimens was immediately frozen in liquid nitrogen. This is the first study to directly compare the whole-genome locus specific promoter methylation profiles of primary HNSCC, HNSCC tumor xenografts, and HNSCC cell lines. Although we had a relatively small sample size, our findings are consistent with several previous studies have compared methylation levels in tumors and cell lines from multiple different tissue types and have reported increased levels of methylation in tumor-derived cell lines when BEZ235 compared to primary tumors., including one study that compared tumor, xenograft, cell line and normal tissue whole-genome methylation profiles. One recent study by Houshdaran and colleagues compared methylation profiles of ovarian epithelial cell carcinomas and cell lines at 1,505 GpG sites, representing 808 genes. This study demonstrated a distinct difference in methhylation profiles of tumors and cell lines. 8% of the genes that we found to be differentially methylated in our study were also found to be differentially methylated BIBW2992 in the smilar comparison made by Houshadaran and colleagues with ovarian cancers and ovarian cancer cell lines. Another recent study by Milne and colleagues compared methylation profiles of a gastric cancer specimen, a xenograft from that cancer and 2 gastric cancer cell lines. Similarly to our study, Milne and colleagues found that the methylation pattern in cell lines is different from primary tissues. This study, however, only looked at 38 genes. Paz and colleagues compared methylation of 15 genes in primary tumors and cell lines for 12 different tumor types, including HNSCC, and found that the methylation profile for these 15 genes is similar in primary tumors and corresponding cell lines. This study, however, looked at methylation only for a subset of genes and did not evaluate the whole-genome promoter methylation profile. Of the 15 genes evaluated by Paz and colleagues, only DCC is seen in common with the 98 genes differentially methylated between normal tissue and primary tumors. The high degree of similarity between primary tumors and xenografts derived from those tumors, and the substantial discordance in methylation between primary tumors and cancer cell lines demonstrates that xenografts are more accurate models for promoter methylation in cancer than cancer cell lines. Although it is expected that the methylation of each individual xenograft would be similar to its corresponding primary tumor, all of the tumor-xenograft pairs had overall similar methylation patters as seen on hierarchical clustering while all of the cell lines had similar methylation patterns.