Local recurrence after retropubic radical prostatectomy for prostate cancer does not exclusively occur at the anastomotic site

Authors


Correspondence: Urs E. Studer, Department of Urology, University of Berne, Anna-Seiler-Haus, Inselspital, CH-3010 Berne, Switzerland.

e-mail: urology.berne@insel.ch

Abstract

What's known on the subject? and What does the study add?

  • Local recurrence after radical prostatectomy (RP) for clinically organ-confined prostate cancer is largely assumed to occur at the anastomotic site, as reflected in European and North American guidelines for adjuvant and salvage radiotherapy after RP. However, the exact site of local recurrence often remains undetermined.
  • The present study shows that roughly one out of five patients with local recurrence after RP has histologically confirmed tumour deposits at the resection site of the vas deferens, clearly above the anastomotic site. This should be considered when offering ‘blind’ radiotherapy to the anastomotic site in patients with biochemical recurrence alone.

Objective

  • To determine the anatomical pattern of local recurrence and the corresponding clinical and pathological variables of patients treated with retropubic radical prostatectomy (RRP).

Patients and Methods

  • In all, 41 patients with biopsy confirmed local recurrence after extended pelvic lymph node dissection and RRP performed between January 1992 and December 2009 at a single tertiary referral academic centre were retrospectively studied.
  • The site of local recurrence as assessed on computed tomography or magnetic resonance imaging was reviewed. Two sites were identified: the vesicourethral anastomotic site and the cranial resection margin of the surgical bed, where the vas deferens was transected and clipped.
  • Age and serum prostate-specific antigen (PSA) level at RRP, pathological tumour and nodal stage, Gleason score, tumour location, surgical margin status, age and serum PSA level at the time of local recurrence, and time to diagnosis of local recurrence were assessed for the two sites and compared with the chi-square or Wilcoxon rank sum tests as appropriate.

Results

  • Local recurrence occurred at the anastomotic site in 31/41 (76%) patients and at the resection site of the vas deferens in nine of 41 (22%) patients. One patient had distinct lesions at both sites.
  • There was no significant difference in any of the clinical and pathological variables between patients with local recurrence in the former and latter site.

Conclusion

  • Most local recurrences after RRP occur exclusively at the anastomotic site. However, 22% of locally recurrent cases had tumour at the resection site of the vas deferens. This should be taken into account when considering adjuvant or salvage radiation therapy.

Introduction

Retropubic radical prostatectomy is a standard treatment for clinically organ-confined prostate cancer [1, 2]. Although surgery provides excellent cancer control, 17–40% of all patients will have biochemical recurrence by 10 years after RRP [3-6]. In these patients, the distinction between local recurrence and metastatic disease, which has therapeutic consequences, is usually based on clinical and pathological parameters as well as on PSA kinetics.

Apparently isolated local recurrence is reported in 2–21.5% of all patients with organ-confined prostate cancer 3–15 years after surgery [2, 4, 7, 8]. These patients can be offered salvage radiotherapy. If the disease is not yet detectable but local recurrence is highly suspected, it is traditionally assumed that the anastomotic site is the site of residual disease. This is reflected in the guidelines of the Radiation Oncology Group of the European Organisation for Research and Treatment of Cancer (EORTC), which define the clinical target volume for radiotherapy after RRP as encompassing the urethrovesical anastomosis centrally, the bladder neck cranially and the apex caudally [9].

However, the anatomical site of local recurrence after RRP has so far received little attention [10-13]. Herein we retrospectively assessed the anatomical pattern of local recurrence in patients treated by extended pelvic lymph node dissection and RRP for clinically organ-confined prostate cancer.

Patients and Methods

Between January 1992 and December 2009, 875 consecutive patients with clinically organ-confined prostate cancer underwent extended pelvic lymph node dissection and RRP according to the previously described technique [14]. Of note, the vas deferens was clipped and transected cephalad or at the tip of the seminal vesicle on both sides, and no clips were used elsewhere. The lymphatics were ligated in the inguinal area and bipolar coagulation was used elsewhere. All operations were performed by staff urologists. Regardless of the final pT stage, no adjuvant radiotherapy or androgen-deprivation therapy was given. Our institutional follow-up scheme includes DRE and serum PSA level evaluation after 3 months and, if the PSA level is <0.1 ng/mL, after 24 months and at the discretion of the treating physician thereafter. In case of biochemical recurrence (defined as a PSA level >0.2 ng/mL and rising), patients were re-evaluated by DRE, CT or, alternatively, by MRI of the abdomen and pelvis as well as by bone scan. In every case a biopsy via the transrectal route was taken from the suspected lesions, which were visualised on imaging and/or palpable.

In all, 44 patients (5%) were diagnosed with histologically confirmed local recurrence. All clinical and pathological data were evaluated retrospectively from a prospectively maintained database. Pathological tumour and nodal stage were classified according to the 2002 TNM classification. Patients for whom CT or MRI findings were not available (three patients) were excluded from further analysis. The remaining 41 patients were evaluated, including seven patients with concomitant bone metastases.

The anatomical pattern of local recurrence was determined by reviewing the imaging studies in collaboration with an experienced uro-radiologist who was ‘blinded’ to the clinical and pathological data. Two locations were defined: (i) the vesicourethral anastomotic site, which comprises the bladder neck and the area distal to it, and (ii) the resection site of the vas deferens, defined as the retrotrigonal area above the former seminal vesicle tip where the vas had been clipped and transected. Of note, the resection site of the vas deferens lies 3–4 cm cephalad to the vesicourethral anastomotic site. Figures 1 and 2 show the two sites of local recurrence schematically and on CT images.

Figure 1.

Diagrammatic representation of local recurrence at the anastomotic site (red line) and at the cranial aspect of the surgical bed, where the vas deferens had been clipped (grey line) and transected on both sides (blue line).

Figure 2.

Local recurrence at the cranial aspect of the surgical bed. Axial (A) and sagittal reconstructed (B) CT images show contrast-enhancing soft tissue on the right at the level where the vas deferens had been clipped and transected.

All RP specimens were reviewed by a uropathologist who was ‘blinded’ to the clinical data. Three prostate tumour locations were defined: (i) the prostatic base, i.e. the area in contact with the inferior part of the bladder, (ii) the prostatic apex, i.e. the area adjacent to the urogenital diaphragm and (iii) the middle and lateral lobes, i.e. between the prostatic base and apex. Similarly, location of positive surgical margins was documented.

Differences in clinical and pathological variables between patients with local recurrence in the two locations were assessed with the chi-square test for categorical variables and the Wilcoxon rank-sum test for continuous variables using commercially available software. All reported P values are two-sided, statistical significance was set at 0.05.

Results

Histologically confirmed local recurrence was diagnosed after a median (interquartile range, IQR) time of 59 (27–84) months from RRP; the median (IQR) serum PSA level at biopsy was 9.2 (4.8–29.4) ng/mL. Local recurrence was found at the anastomotic site in 31/41 (76%) patients and at the resection site of the vas deferens in nine of 41 (22%). One patient had distinct lesions at both locations and was excluded from the analysis.

There was no statistically significant difference between patients with local recurrence at the anastomotic site and those with local recurrence at the resection site of the vas deferens for age, PSA level at the time of RRP, pathological tumour and nodal stage, Gleason score and time to diagnosis of local recurrence. Prostate tumour location in the RP specimen did not correlate with the site of local recurrence. The primary tumour was at the prostatic base in 21/31 (68%) patients with recurrence at the anastomotic site and in five of nine patients with recurrence at the resection site of the vas deferens (P = 0.8). The primary tumour was also detected at the prostatic apex in 29/31 (94%) with recurrence at the anastomotic site and in eight of nine patients with recurrence at the resection site of the vas deferens (P = 1.0). Positive surgical margins and seminal vesicle invasion were present at a similar rate in both groups (P = 0.7 and P = 0.5, respectively; Table 1). Furthermore, location of positive surgical margins did not correlate with the site of local recurrence. In patients with local recurrence at the anastomotic site, positive surgical margins were found at the prostatic apex in 11/15 (73%) patients, at the prostatic base in four of 15 (27%) patients and at the posterior/lateral aspect of the prostate in seven of 15 (47%) patients. In patients with local recurrence at the resection site of the vas deferens, positive surgical margins were found at the prostatic apex in four of five (P = 0.9) patients, at the prostatic base in one of five (P = 0.8) patients and at the posterior/lateral aspect of the prostate in two of five (P = 0.8) patients.

Table 1. Clinical and pathological characteristics of patients with local recurrence at the anastomotic site and patients with local recurrence at the cranial aspect of the surgical bed.
 Patients with biopsy confirmed local recurrence (n = 41)Recurrence at the anastomotic site (n = 31)Recurrence at the resection site of the vas deferens (n = 9)P
  1. *chi-square test; Wilcoxon rank-sum test; 1 patient had distinct lesions at both locations and was excluded from further analysis; percentages may not sum to 100% due to rounding.
Median (IQR):    
Age at RRP, years63 (59–68)63 (58.5–68)63 (62–68)0.5
Serum PSA level at RRP, ng/mL15.3 (11.9–25.0)14.5 (11.4–25.5)19.5 (15.3–23.4)0.8
N (%) or n/N    
Pathological tumour stage:   1.0*
pT214 (34)11 (35)3/9 
pT316 (39)12 (39)4/9 
pT411 (27)8 (26)2/9 
Pathological lymph node status:   0.3*
pN019 (46)16 (52)3/9 
pN+22 (54)15 (48)6/9 
Median (IQR) Gleason score7 (7–8)7 (7–7)7 (7–8)0.5*
N (%) or n/N:    
Gleason score, distribution,   0.8*
2–69 (22)7 (23)2/9 
721 (51)17 (55)4/9 
8–1011 (27)7 (23)3/9 
Tumour location:    
Prostatic base21 (68)5/90.8*
Prostatic apex29 (94)8/91.0*
Middle and lateral lobes31 (100)9/9
Positive surgical margin:21 (51)15 (48)5/90.7*
Prostatic base4 (27)1/90.8*
Prostatic apex11 (73)4/90.9*
Posterior/lateral7 (47)2/90.8*
Seminal vesicle invasion18 (44)13 (42)5/90.5*
Median (IQR):    
Age at local recurrence, years69 (65–72)69 (61.5–72)69 (68–74)0.5
Serum PSA level at local recurrence, ng/mL9.2 (4.8–29.4)9.7 (4.6–41.5)9.2 (6.7–17.4)0.8
Time from RRP to biopsy confirmed local recurrence, months59 (27–84)59 (27.5–81.5)67 (27–104)0.8

Discussion

In the present series, 22% of patients with locally recurrent cancer after RRP had tumour exclusively at the resection site of the vas deferens (clip), respectively along the remaining vas deferens. These findings diverge from the usual assumption that local recurrence after RRP occurs exclusively at the anastomotic site.

Most earlier information on the anatomical pattern of local recurrence after RRP was based on TRUS imaging. Connolly et al. [10] noted that in 61 patients with biopsy confirmed local recurrence after RP, 66% of all lesions occurred at the anastomotic site, 13% in the area posterior to the trigone, and 16% posterior to the bladder neck. Leventis et al. [11] reported that most (56%) biopsy confirmed local recurrences after RRP were detected at the vesicourethral anastomosis, while only two of 41 (5%) men had a lesion identified in the retrovesical space. Other sites of local recurrence were the areas posterior and anterior to the bladder neck and lesions at multiple sites. Scattoni et al. [12] found that 10% of local recurrences documented by TRUS-guided biopsy were located in the retrovesical space. The discrepancy in local recurrence pattern between the present and other studies may be attributed to our use of CT and MRI as diagnostic tools, which might allow for tumour visualisation beyond the classic locations depicted by TRUS. Indeed, in an endorectal coil MRI study evaluating 39 patients with clinically documented local recurrence after RP and a mean PSA level of 2.18 ng/mL 40% of all local recurrence sites were also detected in the retrovesical space [13]. Moreover, the definition of local recurrence sites has not been uniform. In the aforementioned studies, the retrovesical space may not coincide with the resection site of the vas deferens, as defined in the present study. We used the level of the surgical clip marking the resection site of the vas deferens to define the former tip of the seminal vesicle, clearly above the anastomotic site.

Local recurrence at the anastomotic site represents residual disease left at either the apical margin or bladder neck, and is most likely due to less than optimal surgery with incomplete resection of the primary tumour. On the other hand, the pathophysiology of lesions at the resection site of the vas deferens remains unknown. Anatomopathological studies analysing the pattern of tumour spread in patients with prostate cancer with seminal vesicle invasion have shown that prostate tumour can be seen in continuity along the ejaculatory duct wall [15-17]. At the beginning of last century, two French anatomists, Cunéo and Marcille [18], were probably the first to describe one pathway of lymphatic drainage of the prostate along the vas deferens to the external iliac lymph nodes (Fig. 3) [18]. One may speculate therefore that lesions at the resection site of the vas deferens may be caused by lymphatic spread along the vas, not by direct infiltration. It is noteworthy in this context that in the present patients, tumour location in the RP specimens did not correlate with the site of local recurrence, i.e. distribution of the primary tumour location (prostatic base vs lobes vs apex) in patients with local recurrence at the resection site of the vas deferens was similar to that in patients with local recurrence at the anastomotic site. Likewise, we could not find a correlation between location of positive surgical margins and location of local recurrence.

Figure 3.

Lymphatic drainage of the pelvis, according to Cunéo and Marcille (1901) [18]. Of note, the lymphatic drainage from the base of the prostate along the vas deferens to the external iliac lymph nodes was already emphasised at that time.

In light of the hypothetical lymphogenic spread of prostate cancer along the vas deferens, the fact that 22% of local recurrences after RRP were not found at the anastomotic site has potential implications in two particular contexts: postoperative radiotherapy and seminal vesicle-sparing prostatectomy.

In patients undergoing either adjuvant or salvage radiotherapy, current practice involves local application of radiation to the former prostatic area. The EORTC published a report recommending that post-prostatectomy radiation be delivered to the urethrovesical anastomosis anteriorly and posteriorly, the bladder neck cranially and the apex caudally, with safety margins of 5 mm in all directions [9]. However, the suitability of the former prostatic area as the radiation target remains a major area of uncertainty. While superior local control was seen in patients receiving adjuvant radiation therapy after RRP vs RP alone in the EORTC trial 22911, ≈40% of patients showed a biochemical failure and there was no difference in cancer-specific or overall survival [19]. It could be postulated that the heterogeneous pattern of local recurrence might explain certain cases of radiotherapy failure. If the anastomotic site is erroneously assumed to be the recurrence site, the true recurrence site may not be adequately targeted. Interestingly, the Radiation Therapy Oncology Group consensus guidelines delineate a larger clinical target volume that includes the cut end of the vas deferens cranially [20]. Whether this would be sufficient to effectively treat recurrences along the remaining vas deferens, i.e. cephalad to the clip, remains an open question. Moreover, as locally advanced, aggressive tumours may often represent a systemic disease and not only a local problem [21], the role of postoperative radiotherapy is further challenged, especially in view of the recent EORTC trial data cited above [19].

The rationale of seminal vesicle-sparing prostatectomy has been frequently discussed recently, with the expectation that the incidence of postoperative urinary incontinence and erectile dysfunction may be reduced [22-24]. Based on the present findings, local recurrence at the resection site of the vas deferens indicates lymphogenic dissemination along the vas. This raises concerns about seminal vesicle-sparing procedures, at least in high-risk patients.

The present results were retrospectively obtained in a limited series of patients and require prospective validation in a larger dataset. The true incidence of local recurrences and their location in the present patients is not known, as we did not systematically search for them. Furthermore, many patients were operated on at our institution and then followed up by their referring practitioner. Some patients could have been referred directly to the radiation oncologist for management of biochemical recurrence, or given androgen-deprivation therapy. However, it remains indisputable that a certain number of RRP patients develop local recurrence outside the anastomotic site.

In conclusion, local recurrence after RRP does not occur exclusively at the anastomotic site, as previously assumed. In the present study, an appreciable number of patients had local recurrence where the vas deferens was clipped and transected. We hypothesise that these lesions may be caused by lymphangiosis along the vas and suggest that this is one pathway of prostate cancer dissemination. In view of the reported lack of a survival advantage with adjuvant radiotherapy [19], this is one more reason to refrain from ‘blind’ radiotherapy to the anastomotic site in patients with biochemical recurrence alone.

Acknowledgements

We would like to thank Dr Pasquale Cirillo from the Department of Mathematical Statistics, University of Berne, Switzerland, for his kind assistance in the statistical analyses.

Conflict of Interest

None declared.

Details of Ethics Approval

The study was exempt of full review by the Ethics Committee of the Canton of Berne. Because obtaining patient consent to use medical data in large retrospective studies is not realistic, the institutional expert commission for professional secret in medical research issued a limited authorisation to disclose confidentiality for research purposes on 30 June 2003. This authorisation is valid for retrospectives studies including patients that have been treated at the institution (University of Berne, Inselspital, Berne, Switzerland) exclusively.

Funding

None.

Abbreviations
EORTC

European Organisation for Research and Treatment of Cancer

IQR

interquartile range

(R)RP

(retropubic) radical prostatectomy

Ancillary