Radical prostatectomy and brachytherapy are two commonly employed methods of treating localized prostate cancer. Due to a lack of randomized trials, the selection of optimal treatment for these patients remains controversial. A wealth of single-institution data is available, but comparisons are difficult because of differences in patient selection and outcome end points. This has resulted in a frustrating situation for patients and physicians alike. A number of articles have attempted to interpret these data, and two different conclusions have emerged. Several authors have concluded that the available evidence indicates both treatments achieve approximately equal results and should be offered to patients with low-risk disease.[1,2] Other physicians have concluded that in the absence of randomized trials, radical prostatectomy should be considered superior to brachytherapy and remain the standard of care.[3,4] The primary reasons cited by these authors for continuing to regard radical prostatectomy as superior to brachytherapy are listed in Table 1.[3-7] In this review, we will address the latest results of treatment with each modality and attempt to draw conclusions regarding their relative merits. Results of Treatment Selection of Series
A Pubmed search was utilized in an attempt to identify all available series reporting results for transrectal ultrasound (TRUS)-guided interstitial low-dose-rate ("seed") brachytherapy and radical prostatectomy in the prostate-specific antigen (PSA) era. Updates published in abstract form were used when available. For the sake of comparison, only series that reported PSA-based outcomes were considered. Series were selected for analysis if they included data for at least 100 total patients in order to ensure adequate experience with the procedure. Authors generally agree that an accurate comparison of treatment results can only be achieved if the patients compared have similar prognostic factors.[1,3,8] Therefore, series are presented here only if they reported results according to standard pretreatment prognostic factors. The definitions of risk groups used are shown in Table 2. Finally, the importance of adequate length of follow-up has been demonstrated.[ 3,9,10] Therefore, only series with a median follow-up greater than 3 years are presented. We have attempted to include the latest update of each series. In the case of groups that have published results numerous times, we have used their latest report that segregates the patients based on prognostic factors. Biochemical Disease-Free Survival in Brachytherapy Series
Tables 3, 4, and 5 list the results of available brachytherapy series for patients with low-, intermediate-, and high-risk disease, respectively.[ 1,2,11-22] Reports by Merrick,[ 23,24] Singh,[25] and Blank[26] are not included because they had either fewer than 100 patients or less than 3 years median follow-up. Results published by Koutrouvelis are not shown because the implants were performed via a nonstandard technique.[ 27] The results of all of the above series are consistent with, or superior to, the results presented in these tables. Finally, the 32-patient series from Dr. Holm[28]-the first transperineal implants ever reported, last published in 1991-which reported poorer results, was also excluded. The majority of the series did not include any patients who underwent hormonal manipulation in order to rule out a possible confounding effect. Zelefsky et al[11] did include 13% of patients who underwent androgen deprivation for a mean duration of 2 months. Patients in the Dattoli,[22] Lederman,[12] and Kollmeier[13] papers received hormonal ablation in 30%, 16%, and 60% of the cases, respectively. D'Amico[29] reported patients who received androgen deprivation separately, and they are not shown here. In all cases, hormonal ablation was of short duration (3 to 6 months). Of note, the series from Ragde et al[14] clearly presents the worst published results for low-risk patients treated with brachytherapy. These were among the first transperineal implants performed in the United States (treated from 1987 to 1989), thus representing the learning curve for this group, as well as a "discovery curve" for this procedure, during which time important refinements in technique and dosimetry were made. Improvement in technique, and therefore in results, should be expected with increasing experience, and this is demonstrated in a paper by Grimm et al.[15] These investigators compared their data from 1986 to 1987 with their data from 1988 to 1989, and demonstrated an improvement in long-term biochemical no evidence of disease (bNED) rate from 66% in the 1986 to 1987 cohort to 87% in the 1988 to 1989 cohort, despite very similar patient characteristics. Once each table was completed, a weighted average of bNED rates for each prognostic group was calculated using data from all of the included series. The percentage bNED at 5 years in each series was multiplied by the number of patients in that study, and these results were added together and then divided by the total number of patients in the risk group. The data from Sylvester et al[16] were not used in these calculations, as the number of patients in each risk group was not available. Biochemical Disease-Free Survival in Radical Prostatectomy Series
Tables 6, 7, and 8 present the available data from major series of radical prostatectomy for patients with low-, intermediate-, and high-risk features.[ 1,2,4,17,30-34] The series from the University of California, Los Angeles,[ 35] is not presented because it did not stratify results by pretreatment prognostic factors. Barry et al[36] did not report biochemical failure as an end point. Because several of these series do not report results stratified by risk group, but rather, by individual prognostic factors, the data for those series are presented for patients who had features characteristic of that group; ie, patients with a Gleason score of 7 are presented in the intermediate-risk table. It should be noted that this may have a slight negative effect on the outcome of patients in the low-risk table, because some of the subset of patients who had a PSA < 10 ng/mL would be expected to have a Gleason score of 7 or higher, and vice versa. Data for patients stratified according to their multifactor risk group is presented where available. The results from the University of Pennsylvania,[1] Baylor,[30] Cleveland Clinic,[32] William Beaumont Hospital,[30] and Urology Health Center[2] are derived from data published in papers that compared radical prostatectomy to radiation therapy. The Johns Hopkins group has published bNED rates for 5 and 10 years[31]; both are reported here, the 5-year rate to allow comparison to other series, and the 10-year rate to show long-term results. Because many of the surgical series do not provide the number of patients in each prognostic group, a useful weighted average could not be calculated.