A clinicogenetic model to predict lymph node invasion by use of genome-based biomarkers from exome arrays in prostate cancer patients.
10.4111/kju.2015.56.2.109
- Author:
Jong Jin OH
1
;
Seunghyun PARK
;
Sang Eun LEE
;
Sung Kyu HONG
;
Sangchul LEE
;
Hak Min LEE
;
Jeung Keun LEE
;
Jin Nyoung HO
;
Sungroh YOON
;
Seok Soo BYUN
Author Information
1. Department of Urology, Seoul National University Bundang Hospital, Seongnam, Korea. ssbyun@snubh.org
- Publication Type:Original Article ; Research Support, Non-U.S. Gov't
- Keywords:
Exome;
Genotype;
Lymph nodes;
Predictive value of tests;
Prostate neoplasms
- MeSH:
Aged;
Biomarkers, Tumor/*genetics;
Biopsy;
DNA, Neoplasm/genetics;
Exome;
Gene Frequency;
Genome;
Genotype;
Humans;
Lymph Node Excision;
Lymph Nodes/pathology;
Lymphatic Metastasis;
Male;
Middle Aged;
*Models, Genetic;
Neoplasm Invasiveness;
Polymorphism, Single Nucleotide;
Predictive Value of Tests;
Prospective Studies;
Prostatectomy;
Prostatic Neoplasms/*genetics/pathology/surgery
- From:Korean Journal of Urology
2015;56(2):109-116
- CountryRepublic of Korea
- Language:English
-
Abstract:
PURPOSE: Genetic variations among prostate cancer (PCa) patients who underwent radical prostatectomy (RP) and pelvic lymph node dissection were evaluated to predict lymph node invasion (LNI). Exome arrays were used to develop a clinicogenetic model that combined clinical data related to PCa and individual genetic variations. MATERIALS AND METHODS: We genotyped 242,186 single-nucleotide polymorphisms (SNPs) by using a custom HumanExome BeadChip v1.0 (Illumina Inc.) from the blood DNA of 341 patients with PCa. The genetic data were analyzed to calculate an odds ratio as an estimate of the relative risk of LNI. We compared the accuracies of the multivariate logistic model incorporating clinical factors between the included and excluded selected SNPs. The Cox proportional hazard models with or without genetic factors for predicting biochemical recurrence (BCR) were analyzed. RESULTS: The genetic analysis indicated that five SNPs (rs75444444, rs8055236, rs2301277, rs9300039, and rs6908581) were significant for predicting LNI in patients with PCa. When a multivariate model incorporating clinical factors was devised to predict LNI, the predictive accuracy of the multivariate model was 80.7%. By adding genetic factors in the aforementioned multivariate model, the predictive accuracy increased to 93.2% (p=0.006). These genetic variations were significant factors for predicting BCR after adjustment for other variables and after adding the predictive gain to BCR. CONCLUSIONS: Based on the results of the exome array, the selected SNPs were predictors for LNI. The addition of individualized genetic information effectively enhanced the predictive accuracy of LNI and BCR among patients with PCa who underwent RP.
