Validation of a biopsy-based pathologic algorithm for predicting lymph node metastases in patients with clinically localized prostate carcinoma.

Abstract

BACKGROUND: The authors validated an algorithm for the preoperative prediction of lymph node (LN) metastases in patients with clinically localized prostate carcinoma. The algorithm was applied to sextant biopsy material and radical retropubic prostatectomy (RRP) stage obtained from a cohort of men who were treated at the authors' institution. METHODS: Four hundred forty-three patients underwent systematic sextant biopsy and RRP with staging lymphadenectomy. The original algorithm was based on systematic sextant biopsy data and classified patients into three risk groups for LN metastases based on the biopsy result. If > or = 4 of 6 biopsies contained any Gleason Pattern 4 disease, then the patient was at high risk for LN metastases (45%). Patients with > or = 1 of 6 biopsies with dominant Gleason Pattern 4 disease (excluding high-risk patients) had an intermediate predicted risk (19%) of LN metastases. All other patients had a low predicted risk of LN metastases (2.2%). The authors assed the percentage of patients who were positive and negative for LN metastases and calculated the specificity and negative predictive value in the series when patients were classified according to the original algorithm. RESULTS: Twenty of 443 patients had intraoperative LN metastases. When applied to the current data, the Hamburg algorithm classified 404 patients in the low-risk group, 30 patients in the intermediate-risk group, and 9 patients in the high risk group. The incidence of LN metastases was 2.47% in the low-risk group, 20% in the intermediate-risk group, and 44.4% in the high-risk group. The negative predictive value for the low-risk group was 97.52%, and the specificity was 94.14%. CONCLUSIONS: The Hamburg algorithm proved a valid tool for the prediction of lymphatic spread in this validation study on data from the authors' institution. The algorithm may serve as a tool to select patients who do not need to undergo pelvic lymphadenectomy at the time they undergo RRP, hence reducing morbidity and expense. More importantly, with the increasing numbers of men undergoing treatment options in whom LN dissection is not performed, this validated algorithm provides an important selection basis regarding the appropriateness of a therapy that does not routinely include LN staging.