Overexpression of Periostin in Tumor Biopsy Samples Is Associated With Prostate Cancer Phenotype and Clinical Outcome

Carlo Cattrini, Alessandra Rubagotti, Pier Vitale Nuzzo, Linda Zinoli, Sandra Salvi, Simona Boccardo, Marta Perachino, Luigi Cerbone, Giacomo Vallome, Maria Maddalena Latocca, Elisa Zanardi, Francesco Boccardo

Research output: Contribution to journalArticlepeer-review


We retrospectively investigated the prognostic significance of periostin, an extracellular matrix protein, in tumor biopsy samples of 215 patients with prostate cancer. We found that periostin expression can predict the outcome of specific subgroups of patients. In addition to the prostate-specific antigen level and/or Gleason score, the immunohistochemical assessment of periostin expression could be useful in clinical practice to predict the prognosis. Background: Overexpression of periostin (POSTN) is associated with prostate cancer (PCa) aggressiveness. We investigated the prognostic significance of POSTN expression in tumor biopsy samples of patients with PCa. Methods: We scored POSTN expression by immunohistochemistry analysis on 215 PCa biopsy samples using an anti–POSTN-specific antibody. A total immunoreactive score (T-IRS) was calculated by adding the POSTN staining scores of stromal and epithelial tumor cells. Prostate-specific antigen (PSA) progression/recurrence-free survival (PFS), radiographic progression/recurrence-free survival (rPFS), and overall survival (OS) were the study end points. Results: A total of 143 patients received therapy with radical attempt, whereas 72 had locally advanced or metastatic disease and received hormone therapy alone. Median T-IRS was 9 and 12 (range, 0-20), respectively (P = .001). Overall, we found a weak positive correlation of T-IRS with prebiopsy PSA levels (r = 0.166, P = .016) and Gleason score (r = 0.266, P < .000). T-IRS ≥ 8 independently predicted for shorter PSA-PFS and OS (hazard ratio [HR] [95% confidence interval (CI)] ≥ 8 versus < 8: 1.50 [1.06-2.14], P = .024 and 1.92 [1.20-3.07], P = .007, respectively). In the subgroup analysis, the association between T-IRS and patient outcome was retained in patients who received therapy with radical attempt (HR [95% CI] ≥ 8 vs. < 8: rPFS: 2.06 [1.18-3.58], P = .01; OS: 2.36 [1.24-4.50], P = .009) and in those with low to intermediate Gleason scores (HR [95% CI] ≥ 8 vs. < 8: PSA-PFS: 1.65 [1.06-2.59], P = .028; rPFS: 2.09 [1.14-3.87], P = .018; OS: 2.57 [1.31-5.04], P = .006). Conclusion: POSTN T-IRS on PCa biopsy samples independently predicted the risk of recurrence, progression, and death in patients with localized disease and in those with low to intermediate Gleason scores.

Original languageEnglish
Pages (from-to)e1257-e1265
JournalClinical Genitourinary Cancer
Issue number6
Publication statusPublished - Dec 1 2018


  • Extracellular matrix proteins
  • Periostin
  • Prognostic biomarkers
  • Prostate biopsy
  • Prostate cancer

ASJC Scopus subject areas

  • Oncology
  • Urology


Dive into the research topics of 'Overexpression of Periostin in Tumor Biopsy Samples Is Associated With Prostate Cancer Phenotype and Clinical Outcome'. Together they form a unique fingerprint.

Cite this