Predicting everolimus treatment efficacy in patients with advanced endometrial carcinoma: a GINECO group study.

This study aimed to determine whether the expression of various tumor biomarkers of the mTOR pathway predicts tumor response to everolimus in metastatic recurrent endometrial cancer. Tumor blocks from 44 patients of a phase II clinical trial receiving everolimus until progression or toxicity were collected and evaluated at 3 and 6 months for response. Thirty-six blocks were available for analysis of ER, PR, HER2, LKB1, PI3K, PTEN, pAKT, 4E-BP1, p4E-BP1, and S6RP expression by immunohistochemistry, PTEN deletion by FISH, and mutational status of K-RAS, PIK3CA, PTEN, and AKT1 genes. Twelve of 34 evaluable patients had partial response or stable disease (PR, SD) and 22 had progressive disease (PD). Immunohistochemistry showed that no protein expression could predict response to everolimus. Neither could loss of PTEN expression or PTEN deletion or PTEN mutation predict patient outcome. Thirty-one samples were assessable for K-RAS mutations (ten for PR+SD and 21 for PD). There are only four patients with K-RAS mutations and none of them responded to treatment. Median progression-free survival (PFS) and overall survival (OS) were longer in patients without K-RAS mutations (PFS 3.12 ± 1.7 months versus 1.05 ± 0.4 months, p < 0.001; OS 9.28 ± 2.0 months versus 2.30 ± 1.4 months, p = 0.034). In conclusion, the level of expression of proteins of the PI3K/mTOR pathway tested in this study cannot predict response to everolimus. However, endometrial cancer patients with K-RAS mutations do not seem to derive benefit from everolimus treatment.