5550

Although the likelihood of cure for low stage NSCLC is much higher than for advanced NSCLC, many stage 1b and 2 tumors recur and cause death after surgery. Furthermore, patients destined to recur cannot be distinguished by current methods. Better methods to define those at high risk might permit these patients to receive intensive adjuvant or neoadjuvant therapy, or forego surgery altogether. Genomic alterations detected by fluorescence in situ hybridization (FISH) aid diagnosis for NSCLC and are associated with response of late stage NSCLC to certain therapeutics. Here we evaluate FISH as a prognostic tool in the management of early stage disease.
 Formalin-fixed paraffin embedded tumors from 65 patients with stage 1b and 2 NSCLC were analyzed by FISH using a commercial probe set specific for the 8q24 (MYC), centromere 6, 7p12 (EGFR), and 5p15 loci (Abbott Molecular). Of these patients, 22 progressed within 1 year, 5 progressed between 1 and 3 years, 15 progressed after 3 years and 22 never progressed (≥34.5 months follow-up). FISH signals were enumerated in 40 to 120 cells (ave=82 cells) per specimen to obtain copy numbers for each of the five loci. Various classifiers derived from the genomic copy numbers, and a range of cutoff values for each classifier were used to categorize patients into two groups, and times to progression were compared by contingency tables (2-sided Fischer's exact test p-values) and Kaplan Meyer analysis (log rank p-values).
 Several FISH classifiers were found to associate strongly with prognosis when patients that progressed within 3 years of surgery (unfavorable prognosis) were compared to patients that either did not progress or progressed more than 3 years post surgery (favorable prognosis). A single FISH classifier was capable of dividing patients into a group of 36 in which 26 (72.2%) had the favorable prognosis, and a group of 28 in which only 11 (39.3%) had the favorable prognosis (p=0.011). Median times to progression in the former and latter groups were 88.0 and 13.0 months, respectively (p=0.022). A second FISH classifier was capable of dividing patients into a group of 20 in which 16 (80%) had the favorable prognosis, and a group of 44 in which 21 (47.7%) had the favorable prognosis (p=0.028). Median times to progression in the former and latter groups were 113 and 17.3 months, respectively (p=0.0098). Two classifiers could be combined to divide patients into a group of 54 in which 37 (68.5%) had the favorable prognosis, and a group of 10 in which none had the favorable prognosis (p<0.0001). Median times to progression in the former and latter groups were 64.7 and 10.0 months, respectively (p<0.0001). These data indicate that FISH classifiers based on genomic copy numbers of the EGFR, MYC, 5p15, and centromere 6 loci may be of prognostic importance in early stage NSCLC, and FISH performed on biopsy specimens or resected tumors may prove valuable in decisions related to surgery and adjuvant therapy.

99th AACR Annual Meeting-- Apr 12-16, 2008; San Diego, CA