It is becoming increasingly apparent that there are subgroups of patients with oral leukoplakia that are at extremely high risk for cancer development despite complete excision. Histology alone is a relatively poor indicator of such lesions; however, several biomarkers show promise. Although we are beginning to be able to detect such lesions, as yet, we have little knowledge of the mechanism(s) underlying such aggressive behavior or of an effective treatment for such cases. Objective: The goal of this study was to develop a micro-tissue array for in-situ analysis of DNA, RNA and protein alterations in high-risk OPLs with an aim to identifying critical information on gene targets and altered pathways to guide future intervention studies. The data presented here describes initial studies using this array of DNA alterations in the EGFR pathway. Method: The tissue array consisted of 35 oral dysplasias. Twenty were refractory to treatment (D-RTT) showing recurrence after resection. These cases were selected from an ongoing longitudinal study. The remaining 15 cases were randomly chosen from the population-based archives of the BC Oral Biopsy Service, selected to represent the general population of dysplasias (D-GP). A control array was constructed with 20 normal oral mucosae and 10 squamous cell carcinomas (SCCs). Copy number alterations to the EGFR locus (7p12) and to the centromere region of chromosome 7 (CEP7) were determined in these samples by fluorescent in situ hybridization (FISH) analysis. Results: There was no significant difference in demographics and degree of dysplasia for D-RTT and D-GP groups. However, several FISH profiles showed altered frequencies in D-RTT compared to D-GP groups. There was an increase in the proportion of cases with cells showing >2 EGFR signals: seen in 1/20 (5%) normal, 6/15 (40%) D-GP cases (P = 0.03, for comparison to normal), all 20/20 (100%) D-RTT cases (P < 0.0001 for comparison to D-GP), and 8/10 (80%) SCC. The presence of cells with either EGFR > 2 &/or CEP7 > 2 was also elevated, seen in 100% D-RTT lesions and 47% D-GP (P = 0.0003). In the latter case, when a cutoff of at least 10% of cells with the profile was used, these proportions changed to none of the normal, 27% D-GP (P < 0.0001, for comparison to normal), 75% D-RTT (P = 0.007, for comparison to D-GP), and 80% of SCC. To characterize the types of FISH patterns that could be more prevalent in D-RTT and SCC, a dendrogram (using Cluster and TreeView software) was created and identified 17 cases with the numerical abnormalities clustered together, including 3/15 (20%) D-GP (examination of history revealed 2 cases were recurrences), 9/20 (45%) D-RTT, and 5/10 (50%) SCC. Conclusion: The data suggest that FISH profiles can facilitate the identification of OPLs that could prove refractory to surgical removal. Future studies will increase sample number and focus on other members of this pathway (sponsored by R01 DE13124, NIDCR)

[Proc Amer Assoc Cancer Res, Volume 46, 2005]