In a study reported in Clinical Cancer Research, Fujimura and colleagues developed and tested models to predict prostate-specific antigen (PSA) recurrence and cancer-specific survival in patients with treatment-naive prostate cancer and bone metastases using mRNA expression of genes involved in androgen and estrogen signaling and genes in stem cell-like cells.
Cancer and stromal tissues were collected via laser-captured microdissection. Markers analyzed for association with clinical outcome consisted of: androgen receptor and related genes APP, FOX family, TRIM36, Oct1, and ACSL3; stem cell-like cell genes Klf4, c-Myc, Oct3/4, and Sox2; estrogen receptor (ER) and Her2 expression; and PSA and C-reactive protein (CRP).
Analysis of the ability to predict PSA recurrence showed an area under the curve value of 1.0 in both training (n = 46) and validation (n = 30) cohorts for Sox2, Her2, and CRP expression in cancer cells, androgen receptor and ERα expression in stromal cells, and clinical parameters. A model using Oct1, TRIM36, Sox2, and c-Myc expression in cancer cells, androgen receptor, Klf4, and ERα expression in stromal cells, PSA, Gleason score, and extent of disease identified favorable-, intermediate-, and poor-risk groups, with 5-year cancer-specific survival rates of 90%, 32%, and 12% in the training set and 75%, 48%, and 0% in the validation set.
The investigators concluded, “Expression levels of androgen and estrogen signaling components and [stem cell] markers are powerful prognostic tools.” ■
Fujimura T, et al: Clin Cancer Res. July 1, 2014 (early release online).