TY - JOUR
T1 - Assessment of 54 biomarkers for biopsy-detectable prostate cancer
AU - Parekh, Dipen J.
AU - Ankerst, Donna Pauler
AU - Baillargeon, Jacques
AU - Higgins, Betsy
AU - Platz, Elizabeth A.
AU - Troyer, Dean
AU - Hernandez, Javier
AU - Leach, Robin J.
AU - Lokshin, Anna
AU - Thompson, Ian M.
PY - 2007/10/1
Y1 - 2007/10/1
N2 - Objective: We analyzed the association of 54 biomarkers from seven classes including adipokines, immune response metalloproteinases, adhesion molecules, and growth factors with prostate cancer risk adjusting for the Prostate Cancer Prevention Trial (PCPT) risk score. Methods: A total of 123 incident prostate cancer cases and 127 age-matched controls were selected from subjects in the San Antonio Center for Biomarkers of Risk of Prostate Cancer cohort study. Prediagnostic serum concentrations were measured in the sample collected at baseline using LabMAP technology. The odds ratios (OR) of prostate cancer risk associated with serum concentrations of 54 markers were estimated using univariate conditional logistic regression before and after adjustment for the PCPT risk score. Two-way hierarchical unsupervised clustering techniques were used to evaluate whether the 54-marker panel distinguished cases from controls. Results: Vascular endothelial growth factor, resistin, interleukin 1Ra (IL-1Ra), granulocyte colony-stimulating factor, matrix metalloproteinase-3, plasminogen activator inhibitor, and kallikrein-8 were statistically significantly (P < 0.05) underexpressed in prostate cancer cases, and α-fetoprotein was statistically significantly overexpressed in prostate cancer cases, but all had area underneath the receiver-operating characteristic curve <60%; none were statistically significant adjusting for multiple comparisons (P < 0.0008) or after adjustment for the PCPTrisk score. Statistical clustering of patients by the marker panel did not distinguish a separate group of cases from controls. Conclusions: This age-matched case-control study did not support findings of increased diagnostic potential from a 54-marker panel when compared with the conventional risk factors incorporated in the PCPT risk calculator. Future discovery of new biomarkers should always be tested and compared against conventional risk factors before applying them in clinical practice.
AB - Objective: We analyzed the association of 54 biomarkers from seven classes including adipokines, immune response metalloproteinases, adhesion molecules, and growth factors with prostate cancer risk adjusting for the Prostate Cancer Prevention Trial (PCPT) risk score. Methods: A total of 123 incident prostate cancer cases and 127 age-matched controls were selected from subjects in the San Antonio Center for Biomarkers of Risk of Prostate Cancer cohort study. Prediagnostic serum concentrations were measured in the sample collected at baseline using LabMAP technology. The odds ratios (OR) of prostate cancer risk associated with serum concentrations of 54 markers were estimated using univariate conditional logistic regression before and after adjustment for the PCPT risk score. Two-way hierarchical unsupervised clustering techniques were used to evaluate whether the 54-marker panel distinguished cases from controls. Results: Vascular endothelial growth factor, resistin, interleukin 1Ra (IL-1Ra), granulocyte colony-stimulating factor, matrix metalloproteinase-3, plasminogen activator inhibitor, and kallikrein-8 were statistically significantly (P < 0.05) underexpressed in prostate cancer cases, and α-fetoprotein was statistically significantly overexpressed in prostate cancer cases, but all had area underneath the receiver-operating characteristic curve <60%; none were statistically significant adjusting for multiple comparisons (P < 0.0008) or after adjustment for the PCPTrisk score. Statistical clustering of patients by the marker panel did not distinguish a separate group of cases from controls. Conclusions: This age-matched case-control study did not support findings of increased diagnostic potential from a 54-marker panel when compared with the conventional risk factors incorporated in the PCPT risk calculator. Future discovery of new biomarkers should always be tested and compared against conventional risk factors before applying them in clinical practice.
UR - http://www.scopus.com/inward/record.url?scp=35448972867&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=35448972867&partnerID=8YFLogxK
U2 - 10.1158/1055-9965.EPI-07-0302
DO - 10.1158/1055-9965.EPI-07-0302
M3 - Article
C2 - 17932343
AN - SCOPUS:35448972867
SN - 1055-9965
VL - 16
SP - 1966
EP - 1972
JO - Cancer Epidemiology Biomarkers and Prevention
JF - Cancer Epidemiology Biomarkers and Prevention
IS - 10
ER -