Evolution in endometrial cancer: evidence from an immunohistochemical study.

BACKGROUND: It is poorly described how endometrial cancer biology changes during tumor evolution. We hypothesize that characterization of molecular targets in recurrent lesions is more relevant for targeting treatment. METHODS: Paired biopsies from primary and recurrent endometrial cancer tumors (n = 85) were stained immunohistochemically for the following proteins: estrogen receptor (ER), progesterone receptor (PR), stathmin (correlating with phosphatidylinositol 3-kinase activity), HER-2/neu, WT1 (Wilms tumor gene 1), phospho-mammalian target of rapamycin (p-mTOR), and p53. Each tumor was scored, using a semiquantitative and subjective grading system. Discordance, a change in expression between primary and recurrent tumor, was defined as ≥ 2 step change; concordance was ≤ 1 step change. The frequency of change was correlated with established prognostic markers in endometrial cancer. RESULTS: Fifty-six patients (67%) were diagnosed with endometrioid carcinoma, 23 (27%) with serous/clear cell carcinoma. A change in expression between primary and recurrent tumor was noted in 7% to 31% of patients for ER, PR, stathmin, HER-2/neu, WT1, p-mTOR, and p53. Concordant-positive cases for PR were significantly correlated with stage, tumor grade, and histological subtype. Expression of ER, p53, and p-mTOR in cytoplasm in the recurrent tumor correlated significantly with survival. CONCLUSIONS: Endometrial cancer biology changes over time. The decision on targeted treatment should preferably be based on recurrent tumor characteristics.