[753] Seminal Vesicle Invasion at Radical Prostatectomy: Correlation with Magnetic Reasonance Images

R Elliott, NB Bloch, W Dewolf, Y Fu, M Sanda, J Tomaszewski, A Wagner, N Rofsky, EM Genega. Beth Israel Deaconess Medical Center, Boston, MA; University of Pennsylvania, Philadelphia, PA

Background: Seminal vesicle invasion (SVI) in newly diagnosed prostate cancer (PC) is associated with a poor prognosis and treatment decisions may rely on preoperative radiologic detection. Endorectal coil magnetic reasonance imaging (MR) is reported to be accurate in demonstrating SVI prior to prostatectomy. In this study, we characterized SVI histologically with whole mount prepared radical prostatectomy specimens and correlated the findings with MR.
Design: Prostate glands (PG) from 100 consecutive patients with biopsy proven PC who underwent MR were selected for histologic review. The PG slices were kept intact and sections prepared on 2x3 slides and the seminal vesicles (SV) were sliced on cross section. Features of SVI evaluated were number of tumor foci, location and tumor size (measured microscopically). MR evaluation of SVI was obtained from MR reports; a reading was considered positive if SVI was considered present, probable, suspicious, possible or cannot be ruled out..
Results: Of the 100 patients, 8 had histologically identified SVI (3 bilateral, 5 unilateral); MR agreed in 3 cases (1 present, 1 probable, 1 possible). Of these 3 cases, 45-80% of the PG was involved by tumor and the Gleason score (GS) in 2 cases was 9(4+5) and in one case was 7(4+3) tertiary 5. SVI was only present (by direct extension) at the PG-SV junction and the foci measured at least 2cm. In 5 cases, SVI was not identified by MR. The PG involvement in 4 of these 5 cases ranged from 5-15% and one PG had 70% tumor involvement; tumor was at the PG-SV junction only in 1 case, in the free SV wall only in 3 cases, and in both the PG-SV junction and free SV wall in 1 case. In 4 cases the GS was 7 (3 cases 3+4; 1 case 4+3 (+5)) and in one case it was 9(4+5). The foci of tumor in the free SV wall and PG-SV junction ranged from 1mm-4mm. The SV wall was not distorted in any case. MR reported SVI in an additional 4 cases (3 possible, 1 cannot rule out); however, no SVI was identified at pathologic review.
Conclusions: MR detection of SVI is more likely to be accurate when described as present or probable and tumor volume and GS are high. MR detects tumor at the PG-SV junction when the tumor nodule measures >1cm. Small foci of carcinoma 4mm in the SV wall or at the PG-SV junction, and without microscopic distortion are unlikely to be detected by MR. These smaller foci may not be detected when seminal vesicles are randomly sampled.
Category: Genitourinary (including renal tumors)

Tuesday, March 10, 2009 9:30 AM

Poster Session III # 88, Tuesday Morning