[1074] Androgen Receptor Expression in Genitourinary Neoplasms

Elizabeth Williams, John Higgins, Jesse McKenney, Ankur Sangoi, Megan Troxell. Oregon Health & Science University, Portland, OR; Stanford, Stanford, CA; Cleveland Clinic Foundation, Cleveland, OH; El Camino Hospital, Mountain View, CA

Background: Androgen receptor (AR) is strongly expressed in the vast majority of prostatic adenocarcinomas. Historically, AR expression was considered to be relatively specific to prostate carcinoma, maintained in poorly differentiated carcinomas, and amenable to immunohistochemical analysis. However, there remains little published data regarding the specificity of AR for prostate cancer, even amongst genitourinary cancers. Few studies attempting to correlate AR status with outcome in bladder carcinoma report variable rates of AR positivity, while we are aware of only a rare study of AR in renal carcinomas.
Design: Tissue microarrays containing prostate, kidney, and bladder tumors were stained with an antibody to AR. The microarrays were constructed as 0.6 mm tissue cores, most cases were represented at least in duplicate. Unstained tissue microarray sections were stained with antibody clone AR27 (Vector Laboratories, Burlingame CA) at a dilution of 1:10 after cc1 standard retrieval on a Ventana XT instrument, with Ultraview detection (Ventana, Tucson AZ). Cores were scored by 2 observers as negative, weak positive, strong positive, or absent/unscoreable.
Results: As expected, 95% of prostate carcinomas were positive for AR with 73% of tumors strongly positive (N=230). Interestingly, 19% of invasive urothelial carcinomas of the bladder and 16% of renal pelvis were also positive, though many stained weakly (N=190 and 43, respectively); furthermore, 33% of non-invasive urothelial carcinomas were AR positive (N=107). A substantial percentage of primary renal cell carcinomas were AR positive, with 19% of 307 cases staining. From a metastatic renal cell carcinoma cohort, 28% of metastases were AR positive (N=126). Incidentally, 16% of angiomyolipomas and 6% of oncocytomas were also AR positive (N=12 and 15, respectively).
Conclusions: Our data show that the sensitivity of AR immunohistochemistry for prostate cancer is 94.8%, among our cohort of invasive genitourinary tumors. However, the specificity of AR is only 79.5%, since 15-30% of invasive urothelial and primary or metastatic renal cell carcinomas are also AR positive. Thus, the specificity of AR is not ideal for diagnostic confirmation of a poorly differentiated carcinoma of suspected genitourinary origin. AR may remain useful as a component of an immunostain panel.
Category: Genitourinary (including renal tumors)

Wednesday, March 6, 2013 1:00 PM

Poster Session VI # 180, Wednesday Afternoon

 

Close Window