incision into an otherwise organ-confined cancer [1,5].

28 The Authors. Journal compilation 28 BJU International Original Article IMPACT ON PROGRESSION OF POSITIVE SURGICAL MARGINS AFTER RP PFITZENMAIER et al. BJUI BJU INTERNATIONAL Positive surgical margins after radical prostatectomy: do they have an impact on biochemical or clinical progression? Jesco Pfitzenmaier, Sascha Pahernik, Tina Tremmel, Axel Haferkamp, Stephan Buse and Markus Hohenfellner Department of Urology, Medical Center, University of Heidelberg, Germany Accepted for publication 27 March 28 Study Type Prognosis (case series) Level of Evidence 4 OBJECTIVE To prospectively examine the effects of the margin status after radical prostatectomy (RP), the location, and the number of positive surgical margins (PSMs) on biochemical and clinical outcome, as even if there seems to be little debate that there is a higher risk of both local and distant recurrence of prostate cancer in the face of a PSM the significance of a PSM after RP is only followed for biochemical progression in most studies. PATIENTS AND METHODS From our prospective database, 46 consecutive well-described patients without neoadjuvant and direct postoperative adjuvant therapy who underwent RP were included. The median age was 64.7 years, the median preoperative PSA level was 7.9 ng/ ml, and the median follow-up was 5.2 years. We analysed pathological tumour stage, grading, number and location of PSMs, PSAfree survival, local recurrence-free survival, metastasis-free survival, prostate cancerspecific and, overall survival prospectively. RESULTS The overall rate of PSMs was 17.2%. The number was higher in higher stage (P <.1) and higher grade tumours (P =.41). For a PSM the PSA recurrence rate was 64.3%, the local recurrence rate was 18.6%, the development of distant metastasis was 15.7% and therefore much higher than in patients with negative margins (2.5%, 2.7%, and 1.5%). A PSM was an adverse predictor for PSA-free survival (P <.1), local recurrence-free survival (P =.2), and development of metastasis (P =.3) on multivariate analysis. The number and location of PSMs was of no additional prognostic value. CONCLUSIONS A PSM increases the risk of biochemical and clinical e.g. local, disease progression after RP. The number and location of PSMs is of minor importance. Although only 2% of patients with a PSM will develop local recurrence, surgeons should continue to strive to reduce the rate of PSMs to improve cancer control. KEYWORDS prostate cancer, PSA, positive surgical margin, recurrence, survival INTRODUCTION After radical prostatectomy (RP) PSA relapse and clinical recurrence of prostate cancer are reported to be associated with various factors, such as pretreatment serum PSA levels, pathological tumour stage, grading, and s (PSMs) [1,2]. PSMs have been reported in 5 46% of specimens [2,3], whereas various factors have an affect on those margins, e.g. cancer volume, surgical technique, and pathological evaluation. A PSM is defined as cancer at the inked margin of the specimen [1,4]. This can follow from incising into the extraprostatic cancer in patients with extracapsular extension or by incision into an otherwise organ-confined cancer [1,5]. There are conflicting studies regarding the risk of biochemical progression with a PSM. Several reports found that a PSM adversely affects prognosis after RP [4,6 8], whereas other investigators were not able to show this association [9,1]. Various studies indicated that the apex is the most common site of PSMs with the apical margin involving of as many as 64% of all PSMs [11]. But this specific location of a PSM was not predictive for disease recurrence in a recent study by Pettus et al. [2], and the same results were reported by Fesseha et al. [9] and van den Ouden et al. [12]. Therefore, opposed to other solid tumours, a PSM might not imply an adverse prognostic outcome, or a need for adjuvant treatment, in all patients. The bulk of studies evaluating the prognostic significance of a PSM have used PSA recurrence as a surrogate marker for outcome after RP [1,7]. Only one recent study added the risk of local recurrence to test the outcome [8]. To investigate more thoroughly the prognostic value of a PSM, we extended our analysis to variables such as local recurrence rate, development of metastasis, prostate cancer-specific survival, and overall survival, using a well described prospective cohort of 595 consecutive prospectively gathered patients with prostate cancer who underwent RP. In all, 46 patients had no neoadjuvant or direct postoperative adjuvant 28 THE AUTHORS JOURNAL COMPILATION 28 BJU INTERNATIONAL 12, 1413 1418 doi:1.1111/j.1464-41x.28.7791.x 1413

PFITZENMAIER ET AL. therapy and these patients were evaluated. Furthermore, the site and the number of PSMs were correlated to the outcome variables. PATIENTS AND METHODS We report on a well-described prospective cohort of 595 consecutive men who underwent RP for prostate cancer between 199 and 26 at our institution. Of these 11 patients (17%) received neoadjuvant hormonal treatment and 11 (18.5%) received direct postoperative treatment for various reasons. Apparently, some patients received both. In all, 46 patients did not receive neoadjuvant or direct postoperative adjuvant treatment (neither radiation nor hormonal therapy), which means no adjuvant treatment only dependent on the pathological report. These 46 patients were evaluated for our present study. Adjuvant therapy was certainly initiated after biochemical or clinical progression. Patient consent was received in all cases. The median (range) age of this cohort was 64.7 (41.1 8.1) years, the preoperative serum PSA level was 7.9 (.3 12.3) ng/ml, and the median follow-up was 5.2 (.4 15.8) years. We analysed preoperative serum PSA levels, pathological tumour stage, WHO grading, number of PSMs, location of PSMs, time of PSA recurrence, time of local recurrence, and time of distant metastatic recurrence, as well as prostate cancer-specific and overall survival in a prospective manner. We used WHO grading as Gleason score was established at our Department of Pathology in the late 199s; therefore WHO grading was available for all patients from 199 onwards. We are convinced that using the WHO grading is acceptable, since, here too, we distinguish between well, moderately, and poorly differentiated tumours. Follow-up intervals with PSA measurements were performed every 3 months during the first year, semi-annually for the second year, and yearly thereafter. The time for biochemical progression was defined as the time from RP to the time of recurrence of serum PSA levels of.2 ng/ml, or until last follow-up, if the patient did not have PSA relapse. At follow-up, every patient had a DRE and TRUS to assess any clinical evidence of local recurrence. In the case of PSA recurrence Variable NSM, n (%) PSM, n (%) Total number P* TABLE 1 Pathological stage Tumour characteristics and pt 2 246 (91.8) 22 (8.2) 268 margin status of the 46 pt 3a 73 (69.5) 32 (3.5) 15 <.1 men after RP pt 3b 4 17 (51.5) 16 (48.5) 33 WHO grading I 9 (9.9) 9 (9.1) 99 II 28 (8.6) 5 (19.4) 258.41 III 38 (77.6) 11 (22.4) 49 Total 336 (82.8) 7 (17.2) 46 *Chi-square test. or suspicion on DRE or TRUS, CT or MRI as well as bone scintigraphy for suspicion of distant metastasis were performed to detect the site of recurrence. Local disease recurrence was defined as recurrence of disease as confirmed by positive cancer histology near the vesicourethral anastomosis (biopsy or transurethral resection of the anastomotic region) or clear evidence of local recurrence on CT or MRI. All prostate specimens were fixed in formalin for 24 h and inked subsequently. The cone technique was used for assessing margin status at the apex of the prostate [13,14]. Therefore, the distal part of the prostate is amputated and the tissue is sectioned in a sagittal plane. The remainder of the prostate was cut at 4-mm intervals. An overall RP WHO grading was determined, and the tumour was staged according to the fifth Union Internationale Contre le Cancer/TNM 1997 classification. Because at our department of pathology Gleason score was established in the late 199s, we used WHO grading which was available for all patients with prostate cancer since 199 when the prospective database was started. We used the chi-square test for associations between tumour characteristics and the presence of a PSM or (NSM). Kaplan Meier analysis was used for the estimation of freedom from biochemical recurrence, local recurrence, metastatic progression, and disease-specific and overall survival. The log-rank test was used to calculate the differences in outcome between patients with PSMs and NSMs. A Cox proportional hazards model was constructed to test for the association of variables with biochemical relapse, local recurrence, and metastatic progression from the day of surgery. Several parameters were evaluated for predicting the probability of biochemical recurrence, local recurrence, metastatic progression, including tumour grade (I + II vs III), tumour stage (organ confined [pt2] vs extraprostatic disease [pt3a/b and pt4]), status of lymph nodes (positive vs negative), preoperative serum PSA levels, and surgical margin status (PSM vs NSM). Statistical significance in this study was set at P <.5. All reported P values are two-sided. RESULTS During the median follow-up of 5.2 years 47 patients (11.6%) of our cohort died from any cause, whereas only eight patients (2.%) died from prostate cancer after a median (range) follow-up of 3.9 (1.8 9.3) years. In all, 114 patients (28.1%) developed biochemical relapse after a median of.9 (.2 11.9) years, 22 patients (5.4%) had local recurrence after a median of 3.2 (.5 8.5) years, and 16 patients (3.9%) developed distant metastatic disease after a median of 3.7 (.9 9.7) years. Table 1 shows the distribution of the pathological tumour stage, WHO grading and their association to the margin status. In all, 66.% of the tumours were organ-confined; five patients (1.2%) had nodal involvement. Table 2 shows the number and location of the PSMs. The Kaplan Meier curves show the PSA-free survival (Fig. 1a), the local recurrence-free survival (Fig. 1b), and the distant metastasisfree survival (Fig. 1c) with regard to the margin status for all patients (n = 46). The prostate cancer-specific survival was better for patients with a NSM (P <.1) whereas the overall survival was not different (P =.33) after 5.2 years. Figures 2 4 show the PSA-free survival for subcategories, i.e. nonorgan-confined tumours (Fig. 2), grade I and II tumours (Fig. 3), and grade III tumours 1414 JOURNAL COMPILATION 28 THE AUTHORS 28 BJU INTERNATIONAL

IMPACT ON PROGRESSION OF POSITIVE SURGICAL MARGINS AFTER RP FIG. 1. Kaplan Meier curves for all 46 patients showing PSA-free survival (a), local recurrence-free survival (b) and metastasis-free survival (c) after RP as a function of surgical margin status. The log-rank test was used for statistical analysis. a c Metastasis free survival 1,,8,6,4,2 p <.1 5 1 1,,8,6,4,2 p <.1 5 1 15 FIG. 2. PSA-free survival for patients with nonorgan-confined tumours. Kaplan Meier curves show the PSA-free survival after RP as a function of surgical margin status for nonorgan-confined tumours. The log-rank test was used for statistical analysis. There were similar curves for local recurrence-free survival (P =.11) and metastasisfree survival (P =.2) in nonorgan-confined tumours (not shown). 1,,8,6,4,2 p <.1 5 1 15 (Fig. 4) with regard to the margin status. The PSA-free survival for organ-confined tumours was also better for patients with a NSM (P <.1). b,8,6,4,2 p <.1 15 5 1 15 Local recurrence free survival 1, FIG. 3. PSA-free survival for patients with grade I and II tumours. Kaplan Meier curves show the PSAfree survival after RP as a function of surgical margin status for grade I and II tumours. There were similar curves for local recurrence-free survival (P <.1) and metastasis-free survival (P <.1) in grade I and II tumours (not shown). The log-rank test was used for statistical analysis. 1,,8,6,4,2 p <.1 5 1 15 On univariate analysis, classical risk factors such as grade, pathological tumour stage, and preoperative serum PSA levels were also risk factors in our patient cohort for PSA recurrence, local recurrence, and the development of distant metastasis as shown TABLE 2 The number and location of PSMs in our prospective cohort of 46 men with prostate cancer Variable N (%) Number of PSMs, i.e. NSM 336 (82.8) 1 46 (11.3) 2 19 (4.7) 3 5 (1.2) Number and location of PSMs 1 apical 24 (5.9) 1 nonapical 22 (5.4) Several 24 (5.9) FIG. 4. PSA-free survival for patients with grade III tumours. Kaplan Meier curves show the PSA-free survival after RP as a function of surgical margin status for grade III tumours. The log-rank test was used for statistical analysis. 1,,8,6,4,2 p <.4 5 in Table 3. The grade did not influence prostate cancer-specific survival (P =.135) and overall survival (P =.588) after a followup of 5.2 years. Organ-confined tumour stage and preoperative PSA levels were responsible for a better prostate cancer-specific survival (P <.1 and P =.31) but not for a better overall survival (P =.215 and P =.776) during the follow-up. Using Cox regression for multivariate analysis these results were mainly confirmed for PSA recurrence but less for local recurrence and the development of distant metastasis (Table 3). Only a PSM was of prognostic significance for PSA recurrence, local recurrence, or development of distant metastasis (Table 3). 1 The risk for patients with a PSM of developing PSA recurrence was 3.213 (2.126 4.855) times higher, of developing local recurrence was 4.643 (1.785 12.79) times higher, or of developing distant metastasis was 6.649 28 THE AUTHORS JOURNAL COMPILATION 28 BJU INTERNATIONAL 1415

PFITZENMAIER ET AL. (1.915 23.88) times higher compared with patients with a NSM. In this patient cohort a apical PSM did not have an adverse outcome for PSA recurrence (P =.917) or local recurrence (P =.529) when compared with any other nonapical PSM or several PSMs (P =.218 and P =.28). In addition, in this patient cohort one PSM did not have an adverse outcome for PSA recurrence (P =.57) or local recurrence (P =.25) when compared with two PSMs or when compared with three or more PSMs (P =.463 and P =.519). DISCUSSION TABLE 3 P values for univariate and multivariate Cox regression analyses of classical risk factors for prostate cancer in our patient cohort of 46 men Variable PSA recurrence Local recurrence Metastasis Univariate Preop. PSA <.1.15.25 Pathologic T-stage <.1.5.13 Grade.1.1.58 Multivariate Preop. PSA <.1.779.356 Pathological T-stage.3.329.368 Pathological N-stage.57.994.989 Grade.71.15.217 PSM <.1.2.3 Preop., preoperative; T, tumour; N, lymph node. The present prospective study indicates that a PSM after RP for prostate cancer is associated with an increased risk of PSA recurrence, local recurrence, and the development of distant metastasis. This was also shown for subcategories, such as nonorgan-confined tumours or grade I, II, and III tumours. After a median follow-up of 5.2 years, patients with a NSM had a better disease-specific but not overall survival. This was confirmed on multivariate analysis when adjusting for other known risk factors such as preoperative serum PSA levels, tumour stage, and grading. The rate of PSMs significantly increased with pathological tumour stage and grade. Similar analysis for PSA-free survival has been previously performed and produced analogous results [1,15]. For example, Swindle et al. [1] reported on a significant adverse affect on PSA-free survival for patients with PSMs using multiple statistical methods in a study of 1389 consecutive patients with prostate cancer. In their models, they tried different statistical methods to calculate the biochemical outcome with and without adjuvant therapy modalities but as in most studies there is no statement on local recurrence. In contrast, we excluded all patients with neoadjuvant and/or direct postoperative adjuvant therapy from the present study. This is because reports have shown lower rates of PSMs with the duration of neoadjuvant hormonal therapy prior to RP [16 18]. Additionally, postoperative adjuvant therapy either radiation therapy [19,2], or hormonal treatment [21] may alter the outcome. In the present series, we confirmed classical risk factors of prostate cancer outcome, such as preoperative serum PSA levels, tumour stage, and grading. Beside these prognostic factors only the surgical margin status can be influenced by the surgeon and this is why it is of major importance that surgeons must try to further reduce the rate of PSMs. Even if there is little debate that there is a higher risk of both local and distant recurrence in the face of PSMs, there are almost no prospective studies to prove this estimation. In addition to many other reports of an adverse outcome for PSA-free survival, we focused on clinical consequences of a PSM such as local recurrence and development of metastasis. There is only one recent report with a well-described patient cohort of 281 patients, which reported a significant association between the margin status and the local recurrence rate on univariate analysis but there was no significance on multivariate analysis [8]. By contrast, in the present study a PSM was of prognostic significance even on multivariate analysis. This might be due to the inclusion of tumour volume and the percentage of Gleason pattern 4 and 5 in the multivariate analysis by Vis et al. [8], which we did not do. In contrast to other malignancies, such as breast or rectum cancer, local disease progression in prostate cancer will mostly not lead to disease-specific death within the first 1 years [8] after RP, but it can be associated with a psychological burden for the patient as well as with local or systemic therapy due to PSA progression. Most prostate cancer studies use PSA recurrence as a surrogate marker for the outcome in patients. This is necessary due to the otherwise mandatory long follow-up in these patients. Nevertheless, in the present study we showed that a PSM will not only lead to a higher rate of biochemical failures (64.3% PSMs vs 2.5% NSMs) but also to a higher rate of local recurrences (18.6% vs 2.7%) and distant metastasis (15.7% vs 1.5%), as well as to a higher prostate cancer-specific death rate (8.6% vs.6%), and overall death rate (17.1% vs 1.4%). Therefore, in the present study we demonstrate that a biochemical failure will lead to clinical consequences during further follow-up. In the present study, the local relapse rate in patients with a PSM was the same as in the study of Vis et al. [8], which showed that 8% of patients with a PSM would not have local recurrence. Possible explanations might be: (i) tumour cells were destroyed at the site of surgery by electrocoagulation, (ii) nutrition of the tumour cells is disrupted by ligation of the vasculature, (iii) tumour cells can be destroyed by inflammatory effects, and (iv) tumour cells cannot grow without connection to the primary tumour (caused by changes of the tumour microenvironment). The number as well as the location of PSMs was not of prognostic value in the present study. This confirms a report of 498 patients with prostate cancer by Pettus et al. [2], where a PSM at the apex conveyed a similar risk of recurrence compared with a nonapical PSM or multiple PSMs. In their study, the apical PSM did not independently predict biochemical recurrence on multivariate analysis. These results were also confirmed by Fesseha et al. [9], who reported on 59 patients that the prognosis for a patient with an apical PSM, in 1416 JOURNAL COMPILATION 28 THE AUTHORS 28 BJU INTERNATIONAL

IMPACT ON PROGRESSION OF POSITIVE SURGICAL MARGINS AFTER RP an otherwise organ-confined tumour, is the same as for a patient with a NSM. The present results are supported by a report of van den Ouden et al. [12] who also found a PSM to be predictive for the outcome in 273 patients. This also is reported by Epstein et al. [22] with a solitary apical PSM being associated with higher recurrences rates and shorter times to progression than NSMs. The differences in recurrence rates maybe because at the apex the prostate capsule cannot be identified. For this reason such artificial margins cannot be reliably separated from biologically significant margins [2,22]. Some limits of the present prospective study must be mentioned. The follow-up of 5.2 years is relatively short for the investigation of local recurrence, development of metastasis, disease-specific survival, and overall survival in prostate cancer. We know that the median time from PSA recurrence after RP to metastatic progression is 8 years [23]. Nevertheless, there is only one recent prospective study having a longer median follow-up of 7 years and the local recurrence rate was the same in patients with PSMs as in the present study. In addition, PSA recurrence usually develops during the first 2 years of follow-up and >9% of patients relapse within 5 years after RP [24,25]. In addition, the back-flow of our prospective follow-up database is between 9 and 95% each year, which accounts for a very meticulous follow-up and very few patients that are lost to follow-up. The second limitation of the present study might be the underestimation of local recurrences. None of the patients received positron-emission tomography-ct for further evaluation. This tool might be helpful in the future to increase the sensitivity of detecting local recurrences. This detection problem of local recurrences in prostate cancer is well known in the few studies that have evaluated the local recurrence rate after a PSM. Nonetheless, TRUS and DRE were performed on every patient to detect local recurrence and if necessary CT or MRI were performed additionally. In conclusion, the present data indicate that a PSM is of prognostic significance for PSA recurrence, local recurrence, and the development of distant metastasis on multivariate analysis after RP. The PSM rate increased with tumour stage and grade. Furthermore, a solitary apical PSM shows no differences to a solitary nonapical PSM for biochemical recurrence, local recurrence, or the development of distant metastasis. Because it is the only parameter that can be influenced by the surgeon, urologists must continue to strive to reduce PSMs. ACKNOWLEDGEMENTS We would like to thank Maria Pritsch, Ph.D., Institute of Medical Biometry and Informatics, University of Heidelberg, Germany and Hildegard Jakobi for their assistance with the statistics. CONFLICT OF INTEREST None declared. REFERENCES 1 Swindle P, Eastham JA, Ohori M et al. Do margins matter? The prognostic significance of s in radical prostatectomy specimens. J Urol 25; 174: 93 7 2 Pettus JA, Weight CJ, Thompson CJ, Middleton RG, Stephenson RA. Biochemical failure in men following radical retropubic prostatectomy: impact of surgical margin status and location. J Urol 24; 172: 129 32 3 Graefen M. The after radical prostatectomy why do we still not really know what it means? Eur Urol 26; 5: 199 21 4 Ohori M, Wheeler TM, Kattan MW, Goto Y, Scardino PT. Prognostic significance of s in radical prostatectomy specimens. J Urol 1995; 154: 1818 24 5 Stamey TA, Villers AA, McNeal JE, Link PC, Freiha FS. Positive surgical margins at radical prostatectomy: importance of the apical dissection. J Urol 199; 143: 1166 73 6 Blute ML, Bostwick DG, Seay TM et al. Pathologic classification of prostate carcinoma: the impact of margin status. Cancer 1998; 82: 92 8 7 Karakiewicz PI, Eastham JA, Graefen M et al. Prognostic impact of positive surgical margins in surgically treated prostate cancer: multi-institutional assessment of 5831 patients. Urology 25; 66: 1245 5 8 Vis AN, Schroder FH, van der Kwast TH. The actual value of the surgical margin status as a predictor of disease progression in men with early prostate cancer. Eur Urol 26; 5: 258 65 9 Fesseha T, Sakr W, Grignon D, Banerjee M, Wood DPJr, Pontes JE. Prognostic implications of a positive apical margin in radical prostatectomy specimens. J Urol 1997; 158: 2176 9 1 Stamey TA, McNeal JE, Yemoto CM, Sigal BM, Johnstone IM. Biological determinants of cancer progression in men with prostate cancer. JAMA 1999; 281: 1395 4 11 Shah O, Melamed J, Lepor H. Analysis of apical soft tissue margins during radical retropubic prostatectomy. J Urol 21; 165: 1943 9 12 van den Ouden D, Hop WC, Kranse R, Schroder FH. Tumour control according to pathological variables in patients treated by radical prostatectomy for clinically localized carcinoma of the prostate. Br J Urol 1997; 79: 23 11 13 Bott SR, Kirby RS. Avoidance and management of s before, during and after radical prostatectomy. Prostate Cancer Prostatic Dis 22; 5: 252 63 14 Ayala AG, Ro JY, Babaian R, Troncoso P, Grignon DJ. The prostatic capsule: does it exist? Its importance in the staging and treatment of prostatic carcinoma. Am J Surg Pathol 1989; 13: 21 7 15 Blute ML, Bergstralh EJ, Iocca A, Scherer B, Zincke H. Use of Gleason score, prostate specific antigen, seminal vesicle and margin status to predict biochemical failure after radical prostatectomy. J Urol 21; 165: 119 25 16 Gleave ME, Goldenberg SL, Chin JL et al. Randomized comparative study of 3 versus 8-month neoadjuvant hormonal therapy before radical prostatectomy: biochemical and pathological effects. J Urol 21; 166: 5 7 17 Soloway MS, Sharifi R, Wajsman Z, McLeod D, Wood DP Jr, Puras-Baez A. Randomized prospective study comparing radical prostatectomy alone versus radical prostatectomy preceded by androgen blockade in clinical stage B2 (T2bNxM) prostate cancer. The Lupron Depot Neoadjuvant Prostate Cancer Study Group. J Urol 1995; 154: 424 8 18 Aus G, Abrahamsson PA, Ahlgren G et al. Three-month neoadjuvant hormonal therapy before radical prostatectomy: 28 THE AUTHORS JOURNAL COMPILATION 28 BJU INTERNATIONAL 1417

PFITZENMAIER ET AL. a 7-year follow-up of a randomized controlled trial. BJU Int 22; 9: 561 6 19 Stephenson AJ, Scardino PT, Kattan MW et al. Predicting the outcome of salvage radiation therapy for recurrent prostate cancer after radical prostatectomy. J Clin Oncol 27; 25: 235 41 2 Bolla M, van Poppel H, Collette L et al. Postoperative radiotherapy after radical prostatectomy: a randomised controlled trial (EORTC trial 22911). Lancet 25; 366: 572 8 21 Messing EM, Manola J, Sarosdy M, Wilding G, Crawford ED, Trump D. Immediate hormonal therapy compared with observation after radical prostatectomy and pelvic lymphadenectomy in men with nodepositive prostate cancer. N Engl J Med 1999; 341: 1781 8 22 Epstein JI, Pizov G, Walsh PC. Correlation of pathologic findings with progression after radical retropubic prostatectomy. Cancer 1993; 71: 3582 93 23 Pound CR, Partin AW, Eisenberger MA, Chan DW, Pearson JD, Walsh PC. Natural history of progression after PSA elevation following radical prostatectomy. JAMA 1999; 281: 1591 7 24 Johansson JE, Andren O, Andersson SO et al. Natural history of early, localized prostate cancer. JAMA 24; 291: 2713 9 25 Kattan MW, Eastham JA, Stapleton AM, Wheeler TM, Scardino PT. A preoperative nomogram for disease recurrence following radical prostatectomy for prostate cancer. J Natl Cancer Inst 1998; 9: 766 71 Correspondence: Jesco Pfitzenmaier, Department of Urology, University of Heidelberg, INF 11, 6912 Heidelberg, Germany. e-mail: Jesco.Pfitzenmaier@med.uniheidelberg.de Abbreviations: RP, radical prostatectomy; (P)(N)SM, (positive) (negative) surgical margin. 1418 JOURNAL COMPILATION 28 THE AUTHORS 28 BJU INTERNATIONAL