. A positive surgical margin (PSM) has been proven to be associated with biochemical recurrence (BCR) rates of 50% at 10 yr after radical prostatectomy (RP) [1]. In addition, the risk of BCR after RP is two to four times greater if the margin is positive, even after excluding patients who received adjuvant therapy [2].

Serum prostate-specific antigen (PSA), clinical stage, and biopsy Gleason can predict PSMs preoperatively [3], [4], [5], [6]. Surgical technique and individual surgeon's expertise also play an important role in the incidence of PSM in open RP after controlling for other clinical and pathologic variables [2]. Of these prognostic factors, only surgical technique can be modified and tailored to the prostate side(s) where clinical and biopsy findings suggest a higher burden of disease.

Nonetheless, prior studies on the association between degree of neurovascular bundle (NVB) dissection and PSM rates failed to show if the PSM was on the prostate side where the NVB was excised or on the contralateral side, particularly in patients who had unilateral NVB resection or in those with unilateral prostate cancer (pT2a) [6], [7], [8]. That is why the present study seeks to evaluate the association of both preoperative clinical factors and degree of NVB dissection with PSM by prostate side (side-specific analysis) among patients undergoing laparoscopic radical prostatectomy (LRP) for clinically localized prostate cancer.

Over the last 5 yr, LRP, with or without robotic assistance, has quickly become popular among urologists, and it is therefore believed that there is a need for accurate identification of variables predictive of PSMs in patients undergoing LRP.

2. Material and methods

Between October 2002 and April 2005, 411 patients were treated with transperitoneal LRP by a single surgeon at our institution [15]. Four patients who received neoadjuvant therapy were excluded, leaving 407 patients with 814 evaluable prostate sides (right and left). Of these, 728 prostate sides that harboured prostate cancer in the final pathology analysis, making these patients at risk for PSM, composed the study population. Prostate sides without prostate cancer in the final report were excluded because by definition they will always have a negative SM irrespective of the degree of NVB dissection performed (Fig. 1).

Fig. 1. Midglandular axial section of a prostate mapping inked with hematoxylin-eosin demonstrates that it does not matter how inaccurate the neurovascular bundle dissection may have been on one side and that it will always be called a “negative margin” as long as there are no malignant cells in contact with ink. Absence of cancer cells on that prostate half works as a surrogate for this, explaining why prostate sides without cancer cells were excluded from the analysis. Cancer cell are present in the right peripheral zone (D). The red dot shows an area of extracapsular extension. The arrow demonstrates an area of prostate amputation on the left lateral face. On the basis of the definition of positive margin (see text), this zone is considered a “negative margin” despite the capsular violation.

For this analysis, clinical stage was categorized into side-specific palpable and nonpalpable nodule on preoperative physical examination. Serum PSA level was routinely measured with the use of the Hybritech Tandem-R PSA assay (Hybritech, San Diego, CA, USA).

All outside biopsy specimens were reviewed by specialized uropathologists at our institution. We recorded side-specific Gleason score (sum of the primary and secondary patterns) and the greatest percentage of cancer among all positive needle cores for each side.

Endorectal coil magnetic resonance imaging (MRI) was performed on 379 (93%) patients at least 4 wk after the prostate biopsy. The MRI technique and extracapsular extension (ECE) criteria have already been described [8]. Prostate volume on MRI was calculated with the ellipsoid formula: 0.52×width×length×height.

The degree of NVB dissection was determined by the operating surgeon on the basis of the plane where the dissection took place [9] and registered immediately after completion of the procedure (Fig. 2).

Fig. 2. T2-weighted magnetic resonance imaging axial section at the midprostate showing the different degrees of neurovascular bundle (NVB) dissection. The intrafascial (Intra) dissection follows the plane between the prostatic capsule and the prostatic fascia at the posterolateral angle of the gland. The interfascial (Inter) dissection is performed lateral to the prostatic fascia into the thickness of the NVB. The extrafascial (Extra) dissection is carried out lateral to the entire NVB.

Intrafascial dissection (complete NVB preservation) was defined as a NVB dissection that followed the plane between the prostatic capsule and the prostatic fascia at the posterolateral angle of the gland only. In this case, the NVB was completely “preserved,” and the medial aspect of the NVB (side in contact with the prostate) was covered by the prostatic fascia. In this type of dissection, as the prostate is excised, the Dennonvilliers fascia will remain on the posterior surface of the prostate, but the posterolateral surfaces of the prostate where the NVBs resided will be bare of both prostatic and Dennonvilliers fascias. Prostatic fascia also remains covering the anterolateral aspect of the prostate (Fig. 2, Fig. 3). Interfascial dissection (partial NVB preservation) was defined as a NVB dissection lateral to the prostatic fascia into the thickness of the NVB. In this case the NVB was partially “damaged,” and the prostatic fascia stayed on the prostate side rather than covering the medial aspect of the NVB (Fig. 2, Fig. 3). We considered extrafascial dissection (complete/almost complete NVB damage) when the dissection was carried out lateral to the entire NVB. In this case the NVB was completely or almost completely resected (Fig. 2, Fig. 3).

Fig. 3. Midglandular axial section of a prostate mapping inked with hematoxylin-eosin. Note in A the difference between intrafascial and interfascial techniques. B shows the difference between extrafascial technique and a “wider” interfascial dissection of the neurovascular bundles.

All radical prostatectomy specimens were sectioned with whole-mount technique by specialized uropathologists. A PSM was defined as tumour cells in contact with ink on the specimen. The Gleason system was used for histologic characterization and scored by the sum of the primary and secondary patterns.

A side-specific intrafascial dissection of the neurovascular bundle was usually performed in prostate sides with no palpable nodules, biopsy Gleason score 3+3 or 3+4, maximum percentage of positive biopsy ≤10% (depending on the location of such biopsy), one of three, or one or two of six positive cores, particularly when the positive cores were located medially, or in the absence of suspected ECE on MRI. However, a prostate side with a single positive biopsy with Gleason score 4+3 or 4+4, or a prostate side with a maximum percentage of cancer >10% compromising a medial core without signs of ECE may be suitable for an intrafascial dissection. These criteria should not be taken as a strict rule but as a general guideline.

2.1. Statistical methods

Multivariable associations between preoperative variables and SM status were analyzed with the use of logistic regression. Generalized estimating equations were used to account for patients contributing more than one prostate side to the analysis. We prespecified a single multivariable analysis that included preoperative serum PSA levels, clinical stage (palpable vs. nonpalpable nodule), biopsy Gleason grade (≤6 vs. 7 vs. 8–10), prostate volume on MRI, degree of NVB dissection (intrafascial vs. interfascial vs. extrafascial), and suspicion of ECE on MRI. Because of incomplete data, the side-specific greatest percentage of cancer among all positive needle cores was analyzed by an exploratory univariate analysis. Analysis of variance and comparison of proportions were used when indicated. All tests were two-sided, with p≤0.05 considered statistically significant. Statistical analyses were conducted with the use of Stata, version 8.2 (Stata Corp, College Station,TX, USA).

3. Results

One or more PSMs were identified in 59 of the 728 prostate sides. Eight prostate sides had more than one PSM in the ipsilateral lobe, making a total of 51 distinct PSMs (Table 1). Table 2 details the locations of the PSM. The largest difference between left and right sides was observed at the apex, where left-sided tumours had twice as many PSMs. Table 3, Table 4show the preoperative clinical characteristics of patients (age, body mass index [BMI], serum PSA, and prostate volume) and of prostate sides (maximum percentage of tumour in cores, biopsy Gleason sum, suspected ECE on MRI, degree of NVB dissection, and clinical stage), subdivided by SM status.

The specimen pathologic evaluation showed that 492 (70%) of the 700 prostate sides that underwent MRI had the cancer confined within the prostate capsule and 208 (30%) had ECE. The side-specific PSM rate was 7% for the organ-confined tumours and 9% for those with ECE.

On multivariable analysis, the variables associated with a higher probability of PSM were higher preoperative PSA (p=0.02), Gleason grade 7 compared with ≤6 (p=0.001), interfascial NVB dissection compared with intrafascial dissection (p<0.0005), and smaller prostate volume (p=0.01) (Table 5). Although interfascial dissection of the NVB was independently associated with an increased risk of PSM compared with intrafascial dissection (p=0.01), we found no significant difference between extrafascial and intrafascial dissection (p=0.7). In addition, biopsy Gleason sum 7 (p=0.04), but not Gleason ≥8 (p=0.6), was significantly associated with increased risk of PSM compared with Gleason ≤6.

Table 6 shows in detail the occurrence of PSM by degree of NVB dissection, PSM location, and pathologic stage. The patients undergoing interfascial NVB dissection had higher proportions of PSM located at the apex and posterior aspects of the prostate in comparison with the intrafascial group, whereas the incidence of PSM on the anterior aspect did not vary. The proportion of PSM among prostate sides that underwent interfascial NVB dissections increased from 11% in pT2 to 30% in pT3 prostate sides. Because of incomplete data, the effect of greatest percentage of cancer in any positive needle core from a given prostate side was explored in a univariate way. A higher maximum percentage of cancer in any positive needle core was associated with a greater chance of a side-specific PSM (odds ratio [OR]: 1.02; 95% confidence interval [CI], 1.005, 1.034; p=0.01). BMI was not included in our prespecified multivariate model because of the low number of events. In univariate analysis, we found no statistically significant association between BMI and side-specific PSM (OR: 0.98; 95%CI, 0.88, 1.08; p=0.7). To analyze the effect of prostate side on PSM, we looked only at the 321 patients who had prostate cancer on both prostate lobes (n=642 prostate sides). In univariate analysis, we found no significant difference in risk of PSM between left and right sides (OR: 0.71; 95%CI, 0.36, 1.4; p=0.3).

4. Discussion

Several investigations have correlated preoperative findings with SM [10], [11], [12]; however, this is the first side-specific analysis of PSMs that also admits three degrees of NVB dissection: intrafascial (complete NVB sparing), interfascial (partial NVB sparing), and extrafascial (complete/almost complete NVB resection). This study contrasts with prior reports, which admitted two options: nerve sparing and non-nerve sparing [5], [6], [7]. This approach may be much more reliable because the current standards for prostate cancer surgery include tailoring the surgical technique to the side-specific preoperative clinical and biopsy data, trying to preserve the best possible quality of life [11]. In addition, few studies have evaluated the relation of MRI in conjunction with other preoperative parameters to predict PSM [3], [12].

Since the advent of conventional and robotically assisted LRP, PSM rates have been closely scrutinized, calling for a rigorous evaluation of factors affecting SM in a contemporary series [13].

It is reasonable to believe that the wider the dissection around the gland, the lower the chances of having a PSM; however, there is controversy in the literature [14]. In an open RP series, Villers and colleagues [9] decreased their PSMs from 21% to 6% by performing extrafascial dissection of the NVB among pT2 patients with tumour volume <2 cc. In opposition, Ward and colleagues [6] reported 34% and 42% PSM rates among patients undergoing nerve-sparing and wide NVB dissections, respectively. Similarly, Sofer and colleagues [5] described PSM rates of 24% and 31% among patients whose open RP was nerve sparing and non-nerve sparing, respectively. LRP reports are no less contradictory [7], [13], [15].

We found correlation between surgical techniques of nerve sparing and PSM. Although one would have predicted a lower PSM rate for those dissections in which an additional layer of fascia was presumably preserved (interfascial dissection), the opposite was observed. This observation is explained in part by differences in the patients selected among the three techniques as depicted in Table 7, in which prostate sides undergoing interfascial and extrafascial techniques had significantly higher proportions of biopsy Gleason scores 7–10 and more advanced clinical stages.

Interfascial dissection of the NVB is performed in two settings: one, when clinical data raises the suspicion of ECE, and second, when adhesions are identified intraoperatively, obliging the surgeon to make a wider margin of resection to avoid a PSM. In our study, prostate sides subjected to interfascial NVB dissection had higher proportions of PSM at the expense of an increased number located at the apex and posterior aspects of the prostate (Table 6). This finding led us to believe that a technical error (particularly during interfascial dissections) could have caused many of the PSMs. We subsequently reviewed the surgery video tapes of these six pT2 prostate sides and found that the most common reasons why a PSM occurred appeared to be prostatic adhesions, excessive “peeling” of the prostatic fascia (rather than straight cutting) in the presence of those adhesions, and narrow deep pelvimetry or bleeding vessels interfering with adequate visualization. Interestingly, eight of the nine PSMs occurring in pT3 prostate sides with interfascial dissection were located at the posterior aspect of the prostate, whereas five of the six PSMs occurring in pT2 prostate sides with interfascial dissection were located at the apex (four on the left and one not specified). This finding suggests that the interfascial dissection was wide enough to minimize the risk of PSMs posterolaterally in pT2 prostate sides but was not as wide as it should have been in pT3 prostate sides.

While we concede that the interfascial technique could have not been as precise as intended in some patients, contributing to their higher PSM, we confirm that adequately selected patients can undergo laparoscopic preservation of the NVB (intrafascial technique) with acceptable PSM rates.

Despite the common belief that smaller glands are easier to remove, PSM rates seem to be higher for smaller glands [16], [17], [18]. Foley and colleagues [16] found a 12% increase in the PSM rates (43% vs. 55%) when they compared 44 patients whose prostates weighed over 75g with 396 patients with prostates below that weight. Hsu and colleagues [17] found a PSM rate of 27.1% in men with prostates <26cm3 and 15.4% in men with prostates >50cm3. Similarly, Chang and colleagues [18] reported a PSM rate of 10.5% for prostates >75g and 22.9% for prostates weighing <75g (p=0.01) among 400 patients who underwent LRP (111 transperitoneal and 289 extraperitoneal approaches).

Our data concur with those findings because 67% of the 51 PSMs occurred in specimens at or below the median weight of 50g (interquartile range [IQR]: 42, 61). Of note, the median prostate volume on MRI was 32g (IQR: 24, 42). Thus, MRI underestimated the volume of the prostate by an average of 18g (SD: 9.9g) in our population.

Whether this finding is the result of diagnosis lead time bias [19] and/or worse disease present in smaller glands remains to be confirmed [20]. However, this may not fully explain the association in our series because prostate volume was independent of serum PSA level and Gleason sum as a predictor of PSM in multivariable analysis. It may simply be a pure technical problem of LRP, but this warrants further investigation.

Similar to previous studies [3], [7], [12] we found higher biopsy Gleason sum and PSA to be predictors of a PSM. Prostate sides harbouring a biopsy Gleason sum of 7 had a significantly increased risk of PSM; however, those with Gleason sum of 8 did not reach statistical significance, most likely because of the low number of cases (Table 4). In opposition to other reports [7], [12], [18], clinical stage was not associated to PSM in multivariable analysis.

Suspected ECE on endorectal coil MRI had neither protective effect nor increased risk for PSMs, which means that, whereas this variable may help with surgical planning, its capacity to decrease PSM rates remains to be prospectively investigated. Although increasing BMI may impact PSMs [17], we did not find such an association, possibly because individuals with high BMI were underrepresented. Of the 407 patients, 79 had a BMI between 30 to 35, and only 10 had a BMI >35.

It is well established that tumour volume is associated with higher PSM rates [1], [12] and prostate biopsy data may help to predict tumour volume and risk of a PSM. Although overall percentage of cancer in the biopsy specimen (total millimeters of cancer in the biopsy specimen divided by total millimeters of biopsy tissue) is a more accurate predictor of cancer volume, we used the maximum percentage of cancer in any core of the specific side because not all prostate biopsies were done at our institution. Of interest, 201 (28%) of the 728 prostate sides that harboured cancer had a negative ipsilateral preoperative biopsy. A PSM was identified in 4% of them, compared with an 8% PSM rate among the 527 prostate sides with a positive biopsy. In other words, surgeons should not become overconfident about dissecting the NVB extremely close to the prostate side where the biopsy was negative.

Our results should be interpreted with care because the analysis was limited to a single surgeon at a single tertiary institution. The limited number of events did not allow us to include all potentially relevant variables in the multivariable analysis. Pelvimetric parameters also play a crucial role in the degree of surgical difficulty, a territory that warrants further research.

The data shown in Table 6 should not mislead the reader to believe that the risk of PSM among pT3 prostate sides is lower with the intrafascial dissection than with the inter- or extrafascial approaches, on the logical grounds that, by definition, the prostate is bare of periprostatic tissue during the intrafascial dissection, leading to a PSM wherever there is ECE. During the intrafascial dissection, the prostate capsule is bare of tissue only on the posterolateral aspect; therefore, the lower PSM rate of pT3 prostate sides treated with intrafascial dissection is related not only to patient selection but also to the fact that the location of ECE of prostate sides with negative surgical margins was away from the posterolateral aspect of the gland.

NVB-sparing technique is not a risk factor for PSMs in adequately selected patients undergoing LRP. Our not-yet-validated criteria for intrafascial technique include, but are not limited to, the variables included in the ECE probability nomogram [21] and preoperative MRI data. Elevated serum PSA (>10ng/ml), small glands (<30g), biopsy Gleason scores of 7, degree of NVB dissection, and presence of bulky disease as indicated by a high percentage of tumour on biopsy cores should be considered by laparoscopic surgeons when planning the operation to decrease the incidence of PSMs.

References

1. Cheng L, Darson MF, Bergstralh EJ, Slezak J, Myers RP, Bostwick DG. Correlation of margin status and extraprostatic extension with progression of prostate carcinoma. Cancer. 1999;86:1775–1782.

2. Swindle P, Eastham JA, Ohori M. Do margins matter? The prognostic significance of positive surgical margins in radical prostatectomy specimens. J Urol. 2005;174:903–907.

3. Cheng L, Slezak J, Bergstralh EJ, Myers RP, Zincke H, Bostwick DG. Preoperative prediction of surgical margin status in patients with prostate cancer treated by radical prostatectomy. J Clin Oncol. 2000;18:2862–2868.

4. D’Amico AV, Whittington R, Malkowicz SB. Critical analysis of the ability of the endorectal coil magnetic resonance imaging scan to predict pathologic stage, margin status, and postoperative prostate-specific antigen failure in patients with clinically organ-confined prostate cancer. J Clin Oncol. 1996;14:1770–1777.

5. Sofer M, Hamilton-Nelson KL, Schlesselman JJ, Soloway MS. Risk of positive margins and biochemical recurrence in relation to nerve-sparing radical prostatectomy. J Clin Oncol. 2002;20:1853–1858.

6. Ward JF, Zincke H, Bergstralh EJ, Slezak JM, Myers RP, Blute ML. The impact of surgical approach (nerve bundle preservation versus wide local excision) on surgical margins and biochemical recurrence following radical prostatectomy. J Urol. 2004;172:1328–1332.

7. Guillonneau B, Vallancien G. Laparoscopic radical prostatectomy: the Montsouris technique. J Urol. 2000;163:1643–1649.

8. Wang L, Mullerad M, Chen HN. Prostate cancer: incremental value of endorectal MR imaging findings for prediction of extracapsular extension. Radiology. 2004;232:133–139.

9. Villers A, Stamey TA, Yemoto C, Rischmann P, McNeal JE. Modified extrafascial radical retropubic prostatectomy technique decreases frequency of positive surgical margins in T2 cancers <2cm (3). Eur Urol. 2000;38:64–73.

10. Tigrani VS, Bhargava V, Shinohara K, Presti JC. Number of positive systematic sextant biopsies predicts surgical margin status at radical prostatectomy. Urology. 1999;54:689–693.

11. Ohori M, Kattan MW, Koh H. Predicting the presence and side of extracapsular extension: a nomogram for staging prostate cancer. J Urol. 2004;171:1844–1849.

12. Wieder JA, Soloway MS. Incidence, etiology, location, prevention and treatment of positive surgical margins after radical prostatectomy for prostate cancer. J Urol. 1998;160:299–315.

13. Katz R, Salomon L, Hoznek A, de la Taille A, Antiphon P, Abbou CC. Positive surgical margins in laparoscopic radical prostatectomy: the impact of apical dissection, bladder neck remodeling and nerve preservation. J Urol. 2003;169:2049–2052.

14. Holmes GF, Walsh PC, Pound CR, Epstein JI. Excision of the neurovascular bundle at radical prostatectomy in cases with perineural invasion on needle biopsy. Urology. 1999;53:752–756.

15. Katz R, Salomon L, Hoznek A. Patient reported sexual function following laparoscopic radical prostatectomy. J Urol. 2002;168:2078–2082.

16. Foley CL, Bott SR, Thomas K, Parkinson MC, Kirby RS. A large prostate at radical retropubic prostatectomy does not adversely affect cancer control, continence or potency rates. BJU Int. 2003;92:370–374.

17. Hsu EI, Hong EK, Lepor H. Influence of body weight and prostate volume on intraoperative, perioperative, and postoperative outcomes after radical retropubic prostatectomy. Urology. 2003;61:601–606.

18. Chang CM, Moon D, Gianduzzo TR, Eden CG. The impact of prostate size in laparoscopic radical prostatectomy. Eur Urol. 2005;48:285–290.

19. D’Amico AV, Whittington R, Malkowicz SB, Schultz D, Tomaszewski JE, Wein A. A prostate gland volume of more than 75cm3 predicts for a favorable outcome after radical prostatectomy for localized prostate cancer. Urology. 1998;52:631–636.

20. Freedland SJ, Isaacs WB, Platz EA, et al.. Prostate size and risk of high-grade, advanced prostate cancer and biochemical progression after radical prostatectomy: a search database study. J Clin Oncol. 2005;23:7546–7554.

21. Ohori M, Kattan MW, Koh H, et al.. Predicting the presence and side of extracapsular extension: a nomogram for staging prostate cancer. J Urol. 2004;171:1844–1849.

Fernando P. Secin, Angel Serio, Fernando J. Bianco Jr., Nicholas T. Karanikolas, Kentaro Kuroiwa, Andrew Vickers, Karim Touijer, Bertrand Guillonneau

Department of Urology, Sidney Kimmel Center for Prostate and Urologic Cancers, Memorial Sloan-Kettering Cancer Center, New York, NY, United States

Accepted 23 October 2006 published online 10 November 2006.