Some of previous immunohistochemical studies have compared

Some of previous immunohistochemical studies have compared the levels of AR expression in low-grade versus high-grade and/or non-muscle-invasive versus muscle-invasive tumors. There was a significant or marginal decrease (Boorjian et al., 2004, Miyamoto et al., 2012b, Rau et al., 2011, Shyr et al., 2013, Tuygun et al., 2011) or no significant difference (Kashiwagi et al., 2016a, Mir et al., 2011, Wirth et al., 2017) in the rates of AR positivity between low-grade and high-grade tumors. However, no studies have shown considerably elevated expression of AR in high-grade tumors, compared with low-grade tumors. Similarly, significant or marginal down-regulation (Boorjian et al., 2004, Boorjian et al., 2009, Miyamoto et al., 2012b, Shyr et al., 2013, Tuygun et al., 2011, Williams et al., 2015) or no significant change (Jing et al., 2014, Kashiwagi et al., 2016a, Mir et al., 2011, Wirth et al., 2017) in AR expression was seen in muscle-invasive tumors, compared with non-muscle-invasive tumors. However, 2 other studies showed a significant increase in AR positivity in muscle-invasive tumors (Rau et al., 2011, Zhuang et al., 1997). Prognostic significance of AR expression in urothelial cancer specimens has been assessed, but the findings remain controversial. While 3 studies have suggested a correlation of AR positivity with the risk of tumor progression (Mashhadi et al., 2014, Miyamoto et al., 2012b, Zheng et al., 2011), other 2 have showed that patients with AR-positive tumor have a lower risk of disease recurrence (Nam et al., 2014, Tuygun et al., 2011). Other studies in 5-Iodotubercidin specimens (Mir et al., 2011, Kauffman et al., 2011, Kim et al., 2015) as well as all studies assessed in upper urinary tract specimens (Kashiwagi et al., 2016a, Rau et al., 2011, Wirth et al., 2017) failed to reveal a significant difference in the prognosis of patients between AR-positive versus AR-negative tumors. It has recently been suggested that muscle-invasive bladder cancers are initially androgen-sensitive for their growth but the sensitivity is then lost due to activation of certain genes possessing an ARE in their promoter region in an androgen-independent manner, as seen in prostate cancer, which ultimately induces metastatic potential of tumor cells (Gakis and Stenzl, 2013). Therefore, AR expression may not necessarily serve as a prognosticator in patients with bladder cancer. Nonetheless, a recent study involving 296 patients with pT1/non-muscle-invasive bladder cancer, overexpression of AR mRNA was associated with significantly lower risks of tumor recurrence/progression and cancer-specific mortality (Sikic et al., 2017). Thus, conflicting results as to the status of AR expression in urothelial tumors and its associations with histopathological characteristics or patient outcomes have been reported. To further assess these, we recently conducted a meta-analysis of 9 retrospective studies involving 1309 patients in which AR expression was immunohistochemically detected in bladder specimens (Ide et al., 2017). There were no statistically significant differences in AR expression between non-tumor versus tumor (odds ratio = 1.138; P = 0.336) as well as between non-muscle-invasive versus muscle-invasive tumors (odds ratio = 0.666; P = 0.356). However, AR expression was significantly down-regulated in female tumors, compared with male tumors (odds ratio = 0.658; P = 0.027), as well as in high-grade tumors, compared with low-grade tumors (odds ratio = 0.575; P < 0.001). Moreover, in patients with non-muscle-invasive tumor AR expression was associated with a significantly lower recurrence-free survival rate (hazard ratio = 0.593; P = 0.006), but not with a lower progression-free survival rate (hazard ratio = 0.533; P = 0.223). Meta-analysis data for the association between AR expression and prognosis in patients with muscle-invasive tumor were not available.