Perineural invasion and seminal vesicle involvement predict pelvic lymph node metastasis in men with localized carcinoma of the prostate

Purpose: We evaluate the contribution of perineural invasion and seminal vesicle biopsy results in predicting pelvic lymph node metastases in men with T1 or T2 adenocarcinoma of the prostate. Materials and Methods: A total of 212 men with localized prostate cancer were evaluated for serum prostate specific antigen (PSA), clinical stage, Gleason score and the presence of perineural invasion. Each patient had undergone seminal vesicle biopsies and a laparoscopic pelvic lymph node dissection before definitive therapy. The pretreatment prognostic values, presence of perineural invasion and seminal vesicle involvement were compared to the results of the laparoscopic pelvic lymph node dissection. Differences in proportions were tested using the Pearson chi-square test. The effect of multiple variables was tested using a stepwise logistic regression analysis. Results: PSA ranged from 1.6 to 190 ng./ml. (median 11), and 52% of patients had Gleason score 7 or greater and 67.5% had clinical stage T2b or greater disease. Of the 212 patients 37 (17.5%) had perineural invasion, 43 (20.3%) seminal vesicle involvement and 21 (10%) positive node dissections. A PSA greater than 20 ng./ml. (20 versus 6.8%, p = 0.006), Gleason score 7 or greater (15.5 versus 3.9%, p = 0.005), clinical stage T2b or greater (14 versus 0.6%, p = 0.004), presence of perineural invasion (27 versus 6%, p = 0.0001) and seminal vesicle involvement (32.6 versus 4.1%, p <0.0001) influenced nodal findings. However, in the logistic regression model only the positive seminal vesicle biopsy (p = 0.0006), presence of perineural invasion (p = 0.04) and PSA greater than 20 ng./ml. (p = 0.044) were significant variables. Of the 21 men with positive node dissections 18 (85.7%) had a positive seminal vesicle biopsy or perineural invasion. Separation of patients into a high risk group defined by a positive seminal vesicle biopsy or perineural invasion, or a low risk group defined as the absence of these features yielded a significant association with nodal involvement; (28 versus 2%, p <0.0001). A separate analysis of the patients with a negative seminal vesicle biopsy demonstrated that only perineural invasion (19 versus 2%, p = 0.0002) and PSA greater than 20 ng./ml. (12 versus 2%, p = 0.01) conferred a greater risk of nodal metastases. A logistic regression analysis in the negative seminal vesicle biopsy group discarded all of the variables other than perineural invasion as significant. Conclusions: A positive seminal vesicle biopsy is the most significant predictor of pelvic lymph node metastases in men with T1 or T2 prostate cancer. Perineural invasion is also an independent predictor of nodal disease. Patients with either of these features should undergo pelvic lymph node dissection before receiving definitive therapy.