Postoperative prognosis of node-negative breast cancers predicted by gene-expression profiling on a cDNA microarray of 25,344 genes.

Background: In Japan, postoperative relapse occurs within five years in 9.2 to 16% of patients whose breast cancers have not metastasized to lymph nodes at the time of initial surgery(node-negative, n0). Attempts to find molecular markers able to classify n0 breast cancers in terms of postoperative prognosis have not been successful.
Methods: To identify molecular indicators of prognosis for this type of cancer, we used a cDNA microarray consisting of 25,344 human genes to investigate expression profiles of 12 primary breast cancers from patients whose tumors recurred within five years after surgery(5Y-R) and 12 from patients who survived disease-free for more than five years (5Y-F).
Results: Sets of genes characterizing each group in terms of expression patterns in the tumors were selected by Mann-Whitney and random-permutation tests: these panels included 21 genes expressed highly in 5Y-R tumors than in 5Y-F tumors, and 37 with higher expression in the 5Y-F group than in the 5Y-R group.
Conclusions: We established a scoring system to prediction of postoperative prognosis which was 100% accurate as to the actual clinical outcomes of the 24 cases and therefore might be useful for predicting prognosis of n0 breast cancers in a clinical setting. The prognostic score system clearly separated the two groups without any overlap, and accurately predicted prognosis in 6 additional cases. Moreover, the extensive list of tumor-related genes identified in these experiments provides valuable information about progression of breast cancer and suggests potential target molecules for therapy of n0 breast cancers.