Journal of Surgical Oncology 2014;109:818–822

Non-Apical Positive Surgical Margins After Radical Prostatectomy for pT2 Prostate Cancer is Associated With the Highest Risk of Recurrence MARTIN ANDREAS RØDER, MD, PhD,1,2* SANDRA KAWA, MD,1,2 THOMAS SCHEIKE,2,3 BIRGITTE GRØNKÆR TOFT, MD,2,4 JACOB BJERG HANSEN, MD,2,4 KLAUS BRASSO, MD, PhD,1,2,5 BEN VAINER, MD, PhD, DMSc,2,4 AND PETER IVERSEN, MD1,2,5 1 Copenhagen Prostate Cancer Center, Rigshospitalet, Copenhagen, Denmark Department of Clinical Medicine, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark 3 Department of Biostatistics, Department of Public Health, University of Copenhagen, Copenhagen, Denmark 4 Department of Pathology, Rigshospitalet, Copenhagen, Denmark 5 Department of Urology, Rigshospitalet, Copenhagen, Denmark

2

Background and Objective: To investigate how location of positive surgical margins (PSM) in pT2 tumors affect the risk of biochemical recurrence (BR). Methods: The study includes 1,133 consecutive patients from 1995 until end of 2011, who had organ‐confined disease (pT2) following RP. The location of PSM was stratified into apical and non‐apical. BR was defined as the first PSA
0.2 ng/ml after RP. Risk of BR was analyzed with Kaplan–Meier and Cox regression analysis. Results: Median follow‐up was 3.6 years (range: 0.5–15.5 years). The overall pT2 PSM rate was 26.3%. Overall, a pT2 with PSM had a 3.1‐fold increased risk of BR compared to margin negative patients. Patients with pT2 apical and non‐apical PSM had a 5‐year biochemical recurrence‐free survival of 84.9% (95% CI: 77.6–92.2%) and 78.6% (95% CI: 71.3–85.9%), respectively. In multivariate analysis, pT2 apical and non‐apical PSM was individually associated with a 2.2‐ and 3.8‐fold increased risk of BR compared to margin negative patients. Conclusion: In our cohort the location of pT2 PSM was associated with time to BR, that is, patients with non‐apical pT2 PSM endured the highest risk of BR compared to apical PSM. This may indicate that not all patients with pT2 PSM should be offered adjuvant therapy.

J. Surg. Oncol. 2014;109:818–822 ß 2014 Wiley Periodicals, Inc.

KEY WORDS: radical prostatectomy; pT2 tumors; biochemical recurrence; positive surgical margins

INTRODUCTION

MATERIALS AND METHODS

The presence of tumor cells at the resection margin after radical prostatectomy (RP) define a positive surgical margin (PSM) [1]. Risk of biochemical recurrence (BR) after RP is associated with tumor characteristics, including margin status [2]. Studies have shown that PSM increases the risk of BR while it is unclear whether the anatomical location of the PSM is associated with the risk of BR [3–6]. A number of studies have demonstrated that non‐apical PSM may be associated with the greatest risk of BR although the exact causal relationship remains unclear. Radiotherapy following RP has been shown to primarily improve biochemical recurrence‐free survival (BRFS) but also metastasis‐free and overall survival in one study in patients considered at risk for recurrence [7–9]. However, optimal timing of radiotherapy following RP has not been established, and large studies investigating adjuvant versus salvage therapy are underway [10]. Also, the selection of patients for additional treatment after RP is complicated by the lack of clear consensus for defining patients at risk of BR, metastasis and death of PCa after RP. Attempts to risk‐stratify have been presented using tumor characteristics and PSA dynamics [11]. Although the prognostic significance of PSM is generally accepted, the exact role of PSM per se in the assessment of risk for recurrence is unclear, especially in pT2 tumors [12]. In an attempt to refine the indication for adjuvant therapy, the objective of this study was to investigate whether the location of positive surgical margins (PSM) affect the risk of biochemical recurrence (BR) in a consecutive cohort of patients who underwent RP for pT2 PCa in a single institution during a 16‐year period who did not receive any adjuvant therapy.

Between 1995 and 2011, 1,649 patients underwent RP at the Department of Urology Rigshospitalet, Copenhagen, Denmark. Patient data have prospectively been collected in a database approved by the Danish Data Protection Agency (file#2006‐41‐6256). According to our treatment protocol, patients with clinically localized PCa and a life expectancy of 10 years or more were offered RP. Patients with PSA > 10 ng/ml and/or biopsy GS
7 underwent a bone scan prior to RP to rule out metastatic (M1) disease. Staging evaluation included bone scan, abdominal computed tomography and in selected cases magnetic resonance imaging. Patients were staged according to UICC’s TNM classification 2002 (6th edition) [13]. Open retropubic RP was performed according to Walsh (performed by six surgeons). Robotic assisted laparoscopic surgery (RALP) (DaVinci1) was introduced in 2009 and was performed by two of the six surgeons. The present study includes 1,133 consecutive patients who in the period from August 1995 until end of 2011 were

ß 2014 Wiley Periodicals, Inc.

Conflicts of interest: none. *Correspondence to: Martin Andreas Røder, MD, PhD, Copenhagen Prostate Cancer Center, Rigshospitalet, Tagensvej 20, Afsnit 7521, 2200 Copenhagen, Denmark. Fax: þ4535452726. E‐mail: [email protected] Received 3 December 2013; Accepted 15 January 2014 DOI 10.1002/jso.23573 Published online 13 February 2014 in Wiley Online Library (wileyonlinelibrary.com).

PSM After Radical Prostatectomy found to have pT2 tumors at histopathological examination after RP. All patients operated before 2006 were pathologically reviewed to adopt the current consensus guidelines of The International Society of Urological Pathology 2005 [1]. Postoperatively, patients were followed with PSA measurements every 3 months for the first year, thereafter twice a year for 2 years and then once annually. Nerve‐sparing (NS) surgery was offered to selected patients with preoperative self‐evaluated erectile function sufficient for coitus. Unilateral NS was performed in patients with cT1‐cT2a/b, no tumor in apical biopsies, PSA < 10 mg/ml and biopsy GS 3 þ 4 with a maximum of three positive biopsies on the nerve‐spared side. Bilateral NS was performed only in patients with non‐palpable disease (cT1). After fixation in 4% buffered formalin, RP specimens were inked for optimal orientation during microscopic examination (anterior surface in blue, posterior surface in black). Slicing was performed as previously described [14], ensuring laterality, orientation in apical‐basal direction and representation of the resection margin of the total prostate gland. The tissue was paraffin embedded and cut in 3–4 mm tissue sections, and the slides were stained with haematoxylin and eosin. When necessary, immunohistochemical staining for p63, high molecular weight cytokeratin (CK34bE12) and racemase P504S (after 2008) was performed. The surgical margin was considered positive if invasive prostatic glands were located at the inked margin. The location of PSM was reported as specific as possible. The location of PSM was defined as apical when including the most apical, sagitally sectioned slide and the first horizontal section and basal when including the most basal, sagitally sectioned slide. For statistical purposes, the apical group included PSMs found exclusively at the apex. The primary objective of this study was to analyse risk of BR in patients with pT2 PSMs focusing on location of PSM. BR was defined as the first PSA
0.2 ng/ml after RP. No patients received adjuvant therapy until BR was confirmed. Time to BR was calculated from the date of surgery. Kaplan–Meier estimation was used for univariate analysis of biochemical recurrence‐free survival (BRFS). Multivariate analysis was done using Cox proportional hazard model, including age, pT‐category, specimen Gleason score (GS), PSA, surgeon, NS technique and type of surgery (robotic vs. open). The three most inexperienced surgeons were excluded to account for learning curve. Results are presented as hazard ratios (HR) with 95% confidence limits (CI). Baseline PSA was entered on the log scale base 2, therefore HR is for twofold difference in PSA. HR for age represents an increase per 10 years. For categorical covariates, pT2a/b, specimen GS6, surgeon A (most procedures), open surgery, non‐NS surgery, and negative margin status was entered as references. Trend test for distributions between covariates was done using chi‐square (x2) statistics. Tests for interaction between covariates were done. Test for proportionality was done using cumulative residuals (using the timereg package of R) [15]. No patients were lost to follow‐ up. P < 0.05 was considered statistically significant. Statistical analysis was done using R software, version 3.0 (General public license).

RESULTS Patient characteristics are listed in Table Ia. Median follow‐up was 3.6 years (range: 0.5–15.5 years). No patients were lost to follow‐up. No statistical significant differences in distribution of preoperative or specimen pathology parameters between pT2 apical and non‐apical PSM were found, Table Ib. There was a trend toward an association between pT2 non‐apical PSM and nerve‐sparing surgery (P ¼ 0.09) in univariate analysis. This trend was not significant in multivariate logistic regression, where the risk of non‐apical PSM was not associated with nerve‐sparing surgery (OR ¼ 1.2, 95% CI ¼ 0.6–2.6; P ¼ 0.52). Overall, patients with pT2 disease had an estimated 5‐ and 10‐year BRFS of 88.6% (95%CI: 86.2–91%) and 76% (95%CI: 69.7–82.2%), Journal of Surgical Oncology

819

TABLE I. Baseline Characteristics (pT2, N ¼ 1,133) Median Pre‐ and peri‐operative characteristics Age, years 63 PSA, ng/ml 7.9 PPB 30%

Pre‐ and peri‐operative characteristics cT‐category T1 T2a/b T2c T3 Biopsy Gleason score 6 3þ4 4þ3
8 N/A Surgeon A B C D E F Type Open Robot Procedure Nerve‐spared Non nerve‐spared Specimen histopathology pT‐category pT2a/b pT2c Specimen Gleason score 6 3þ4 4þ3
8 N/A Margin status Positive Negative Margin location Apical Non‐apical

IQR

Min–max

59–66 5.5–11 17–50

40–75 0.2–60 0.02–100 N

%

647 311 160 15

57.1 27.5 14.1 1.3

602 426 64 26 15

53.1 37.6 5.6 2.3 1.4

459 320 133 106 88 27

40.5 28.2 11.7 9.4 7.8 2.4

987 146

87.1 12.9

375 758

33.1 69.1

163 970

14.4 85.6

388 577 136 27 5

34.2 50.9 12.0 2.4 0.5

298 835

26.3 73.7

148 150

49.7 50.3

respectively (figure not shown). Patients with negative margin status had a 5‐year BRFS of 91.1% (95%CI: 88.6–93.6%) compared to 81.8% (95%CI: 76.6–86.9%) for patients with PSM, Figure 1 (log‐rank P < 0.001). In multivariate analysis a pT2 PSM was significantly associated with the risk of BR with a HR of 3.1 compared to margin negative, Table II. In Kaplan–Meier analysis a trend for significant impact on 5‐year BRFS survival according to location of PSM was demonstrated. The 5‐ year BFRS was 84.9% (95% CI: 77.6–92.2%) for apical PSM and 78.6% (95% CI: 71.3–85.9%) for non‐apical PSM, respectively (log‐rank test P ¼ 0.09). In multivariate analysis, pT2 apical PSM was associated with a 2.2‐fold significant risk of BR compared to margin negative. Correspondingly, patients with pT2 non‐apical PSM had a 3.8‐fold significant risk of BR compared to margin negative patients, Table III. There was a trend for significant difference in the risk of BR between

820

Røder et al. TABLE III. Risk of Biochemical Recurrence in pT2 Tumors, Stratified on Margin Location

Fig. 1. Biochemical recurrence‐free survival in pT2 tumors, stratified on margin status.

apical and non‐apical PSM (P ¼ 0.08). In Cox adjusted cumulative hazard curve, modeling PSA, age, pT‐category and specimen GS, there was an absolute 10% difference in risk BR between apical and non‐ apical PSM patients at 5 years (12% vs. 22%), Figure 2. During follow‐up, a total of five patients developed metastatic disease, evaluated by bone scan. Also, 46 patients died, whereof four patients died of PCa.

DISCUSSION We present a series of 1,133 consecutive patients with pT2 tumors after RP operated during the past 16 years. An important strength to this study is that none of the patients received any additional oncological treatment before BR was confirmed, and no patients were lost to follow‐up.

Age, per increasing 10 years PSA, for every doubling pT‐category pT2a/b (reference) pT2c Specimen Gleason score 6 (reference) 3þ4 4þ3
8 Type of surgery Open (reference) Robot Surgeon A (reference) B C Procedure Non nerve‐sparing (reference) Nerve‐sparing Margin location Margin negative (reference) Apical Non‐apical

HR

95% CI

P‐value

0.9 1.5

0.6–1.1 1.2–2.0

0.53