L. Cheng et al., Preoperative prediction of surgical margin status in patients with prostate cancer treated by radical prostatectomy, J CL ONCOL, 18(15), 2000, pp. 2862-2868
Purpose: We sought to determine the preoperative factors associated with su
rgical margin status in patients who underwent radical prostatectomy for pr
ostate cancer.
Patients and Methods: The study group consisted of 339 patients who were tr
eated by radical retropubic prostatectomy and bilateral pelvic lymphadenect
omy at the Mayo Clinic. None received preoperative adjuvant therapy. The me
an age at the time of surgery was 66 years (range, 45 to 79 years). All spe
cimens were totally embedded and whole-mounted. Positive surgical margin wa
s defined as the presence of cancer cells at the inked margins. Numerous pa
thologic characteristics in needle biopsies and preoperative clinical findi
ngs were analyzed.
Results: The overall margin positivity rate was 24%. In univariate analysis
, preoperative serum prostate-specific antigen (PSA) level, Gleason score,
perineural invasion, percentage of cancer in the biopsy specimens, and numb
er and percentage of biopsy cores involved by cancer were all associated wi
th positive surgical margins. In multivariate analysis, preoperative serum
PSA level (odds ratio for a doubling of PSA levels, 1.9; 95% confidence int
erval, 1.5 to 2.4; P < .001) and percentage of cancer in the biopsy specime
ns (odds ratio for a 10% increase, 1.3; 95% confidence interval, 1.2 to 1.4
; P < .001) were predictive of margin status in radical prostatectomy. With
use of preoperative serum PSA level and percentage of cancer in the biopsy
as predictors of surgical margins, the overall accuracy as measured by the
area under the receiver operating characteristic curve was 0.74.
Conclusion: Preoperative serum PSA level and percentage of cancer in the bi
opsy specimens were independently associated with surgical margin status in
patients who underwent radical prostatectomy for prostate cancer. The comb
ination of these two factors provider a high level of predictive accuracy f
or margin status. J Clin Oncol 18:2862-2888. (C) 2000 by American Society o
f Clinical Oncology.