Prognostic significance of positive surgical margins in patients with extraprostatic carcinoma after radical prostatectomy
Version of Record online: 5 SEP 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 6, pages 1215–1219, 15 September 2002
How to Cite
Kausik, S. J., Blute, M. L., Sebo, T. J., Leibovich, B. C., Bergstralh, E. J., Slezak, J. and Zincke, H. (2002), Prognostic significance of positive surgical margins in patients with extraprostatic carcinoma after radical prostatectomy. Cancer, 95: 1215–1219. doi: 10.1002/cncr.10871
- Issue online: 5 SEP 2002
- Version of Record online: 5 SEP 2002
- Manuscript Accepted: 24 JUN 2002
- Manuscript Revised: 21 SEP 2001
- Manuscript Received: 7 SEP 2000
- prostatic neoplasms;
- extraprostatic extension;
A significant number of prostate adenocarcinoma patients undergoing radical prostatectomy are found to have microscopic extraprostatic disease extension. A majority of these patients have focal extraprostatic extension limited to one or both sides of the prostate. In addition, positive surgical margins are a common pathologic finding in this patient subgroup. In the current study, the authors evaluated the impact of positive surgical margins as an independent predictive factor for prostate specific antigen (PSA) progression in patients with pT3a/b N0M0 carcinoma.
The Mayo Clinic prostate cancer registry list provided 1202 patients with pT3a/b NO prostate carcinoma (no seminal vesicle or regional lymph node involvement) who underwent a radical prostatectomy between 1987–1995. To reduce confounding variables, patients who received preoperative therapy or adjuvant therapy were excluded, resulting in 842 patients who were eligible for analysis.
A total of 354 patients (42%) had ≥ 1 positive surgical margins whereas 488 patients (58%) demonstrated no margin involvement. The sites of margin positivity were as follows: apex (n = 163), base (n = 47), posterior prostate (n = 227), and anterior prostate (n = 11). A total of 111 patients had ≥ 2 positive surgical margins. The 5-year survival free of clinical recurrence and/or biochemical failure (postoperative PSA level > 0.2 ng/mL) for patients with no positive surgical margins was 76% and was 65% for patients with 1 positive surgical margin (P = 0.0001). There was no significant difference in biochemical disease progression between patients with 1 versus those with ≥ 2 surgical margins (65% vs. 62%). Multivariate analysis revealed that positive surgical margins were a significant predictor (P = 0.0017) of clinical disease recurrence and biochemical failure (relative risk, 1.55; 95% confidence interval, 1.18–2.04) after controlling for preoperative PSA, Gleason score, and DNA ploidy.
In the current study, positive surgical margins were found to be a significant predictor of disease recurrence in patients with pT3a/b NO prostate carcinoma, a finding that is independent of PSA, Gleason score, and DNA ploidy. The benefit of adjuvant therapy in optimizing recurrence-free survival remains to be tested. Cancer 2002;95:1215–9. © 2002 American Cancer Society.
Long-term progression-free survival analyses reveal that radical prostatectomy provides excellent control when prostate carcinoma is organ confined.1, 2 It has been held that prostate carcinoma extending beyond the capsule is associated with increased risk of disease progression and poorer prognosis. Widespread use of total serum prostate specific antigen (PSA) has resulted in an increase in the pathologic organ-confined rate; however, a significant number of patients (20%) undergoing radical prostatectomy for presumed localized disease are found to have unsuspected extraprostatic disease extension.3 These patients may have either focal extraprostatic extension (i.e., no more than one high-power field outside the prostate on no more than two separate sections) or extensive extraprostatic extension.4 Previous studies have demonstrated that such extraprostatic extension is related to variables associated with poor prognosis such as high Gleason score, nondiploid tumors, and increased tumor volume.5 Epstein et al. have argued for tumor stratification based on the extent of capsular penetration, surgical margins, and Gleason score.6 Others have debated that focal extraprostatic extension in the absence of other prognostic risk factors has a favorable prognosis.7 Extraprostatic extension, particularly in the presence of positive surgical margins, has been considered to be a more adverse prognostic finding.7, 8 In the current study, we present a multivariate analyses of the impact of surgical margin positivity as an independent predictor of PSA failure and disease progression in patients with extraprostatic extension (pT3a/b disease). It is our opinion that a better understanding of the natural history of pathologic T3 disease will lead to the improved identification of patients with clinical and pathologic parameters that predict a high probability of tumor recurrence and thus allow for the improved selection of patients for adjuvant therapy trials or the determination of surveillance intervals.
MATERIALS AND METHODS
Between 1987 and 1995, 5568 patients underwent bilateral pelvic lymphadenectomy and radical retropubic prostatectomy for clinically confined adenocarcinoma of the prostate. Of these, 1202 patients were found to have extraprostatic extension of the tumor (pT3a/b N0 disease). Patients who received preoperative or postoperative therapy were excluded from the study, leaving 842 patients for retrospective analysis. The mean age of the patients was 66 years (range, 45–82 years).
Patients were followed at 3–4-month intervals for 2 years postoperatively, 6 months for the next 3 years, and annually thereafter. The mean follow-up in these patients was 4.9 years (range, 6 months–10 years). Digital rectal examination and serum PSA were performed at each visit, whereas radionuclide bone scan, plain radiographs, and computed tomography scans were performed when indicated. In patients who underwent follow-up at another institution, the PSA concentration was determined at the study institution by means of a mailed-in blood specimen, or the patients were contacted annually and additional medical information was obtained from local physicians.
A total serum PSA level of ≥ 0.2 ng/mL was considered evidence of disease progression or recurrence. Palpable local recurrence was confirmed by transrectal needle biopsy. Systemic disease progression was diagnosed on the basis of positive results on bone scan or plain radiography. A combined endpoint in disease progression was defined as an increase in the postoperative PSA level (≥ 0.2 ng/mL), local recurrence, or distant metastatic disease.
Biopsy and radical prostatectomy specimens were graded histologically on the basis of the Gleason scoring system. The pathologist examined the radical prostatectomy specimens immediately after resection using macroscopic inspection; multiple frozen sections; and routine fixed, paraffin-embedded sections as previously described.9 All surgical margins were evaluated routinely, including the prostatic base and apex, bladder neck, capsule, periprostatic soft tissue, and seminal vesicles. Inked surgical margins were defined as positive or negative based on direct contact of malignant cells with the margins. All tumors were analyzed for deoxyribonucleic acid (DNA) ploidy by flow cytometry of the paraffin-embedded tissues. The revised TNM staging system (based on tumor, lymph node, and metastatic involvement) was used for staging.9, 10
Survival from the time of surgery to overall death, clinical disease recurrence (local/systemic), and the combined endpoint of clinical disease recurrence and PSA failure (> 0.2 ng/mL) were estimated using the Kaplan–Meier method. Survival curves were compared by univariate analysis using the log-rank test. Multivariate survival analysis evaluated the effect of positive surgical margins on disease progression, adjusted for preoperative PSA, Gleason score, and DNA ploidy by the Cox proportional hazards model. All tests were two-sided with an α level of 0.05.
Table 1 shows the anatomic sites of surgical margin positivity of the 842 patients with pT3a/b prostate carcinoma who received no adjuvant therapy. A total of 488 patients (58%) did not have any evidence of surgical margin involvement whereas 354 patients (42%) had ≥ 1 anatomic site involved. In addition, Table 1 displays the distribution of positive anatomic sites stratified by the number of sites involved. One hundred eleven patients (13.2%) had ≥ 2 surgical margins involved. The most common sites of surgical margin involvement were the posterior and apical margins. The prostatic base was involved in only 47 patients (5.6% of total).
|No. of sites per patient with positive margins||Specific positive anatomic sites||No. of patients||Overall (%) (n = 842)||Patients with positive margins (%) (n = 354)|
Table 2 shows the relative risk of positive surgical margins (adjusted for PSA, Gleason score, and DNA ploidy using the Cox proportional hazards model) on the endpoints of overall death, clinical disease recurrence, and the combined endpoint of clinical disease recurrence and PSA failure. Positive surgical margins were significant only for the combined endpoint (relative risk, 1.55; 95% confidence interval [95% CI], 1.18–2.04).
|Endpoint||No. with endpoint||RR (95% CI)||P value|
|Overall death||45||1.43 (0.78–2.62)||0.25|
|Systemic or local clinical recurrence||57||1.00 (0.58–1.72)||0.99|
|PSA failure||217||1.55 (1.18–2.04)||0.001|
The 5-year survival rate free of clinical disease recurrence and PSA failure was 76% for patients with pT3a/b disease without surgical margin involvement. By comparison, patients with evidence of involvement of the surgical margin had a statistically significantly lower 5-year survival rate at PSA failure of 64% (P = 0.0001). There was no significant difference in PSA disease progression for those patients with 1 vs. those patients with ≥ 2 involved surgical margins (65% vs. 62%) (Table 3).
|Margins involved||No. of patients||Events||5-yr (%)||SE||7-yr (%)||SE|
|None||488||120||76||± 2.2||68||± 2.9|
|1||243||80||65||± 3.4||55||± 5.3|
|2||111||40||61||± 5.0||56||± 6.1|
Tables 4A–C demonstrate the impact of surgical margin status in patients subclassified by Gleason score as having low (Gleason score of 2–4), moderate (Gleason score of 5–6), and aggressive tumors (Gleason score of 7–10). Surgical margin status is reported to be highly predictive of biochemical disease recurrence in patients with moderate (P = 0.001) and aggressive (P = 0.05) tumors, but to our knowledge, no impact has been reported in patients with low-risk (Gleason score of 2–4) tumors.
|Margin involved||No. of patients||Events||5-yr event-free||7-yr event-free|
|Negative||16||3||81||± 9.8||81||± 9.8|
|Margin involved||No. of patients||Events||5-yr event-free||7-yr event-free|
|Negative||270||52||81||± 2.7||75||± 3.5|
|Positive||204||61||67||± 3.6||51||± 11.9|
|Margin involved||No. of patients||Events||5-yr event-free||7-yr event-free|
|Negative||193||59||72||± 3.6||58||± 5.5|
|Positive||132||55||56||± 5.0||44||± 6.2|
After adjusting for total serum PSA, Gleason score, and DNA ploidy, multivariate analysis (Table 2) revealed that a positive surgical margin retained a significant association with a combined clinical and PSA failure endpoint (P = 0.002).
The goal of radical prostatectomy is complete removal of the tumor while it still is confined to the prostate; however, approximately 30–40% of patients with clinically localized prostate carcinoma have pathologic evidence of extraprostatic disease.11, 12 Tumor extending into the periprostatic fat has been associated with a higher rate of disease progression. Paulson reported a 70% 10-year tumor progression rate with a positive surgical margin for patients with extraprostatic disease.13 Lowe and Lieberman demonstrated an increased rate of disease progression for patients with pT3 disease compared with those with organ-confined disease at 4 years (36% vs. 3.5%).11 The practice of examining the embedded radical prostatectomy specimen has made the discovery of extraprostatic disease in association with or without a positive surgical margin a common histologic finding after surgery.3, 6, 7 In this series, 354 of 852 patients with pT3 (pT3a/b) disease (42%) were found to have ≥ 1 positive surgical margin in association with histologic evidence of extraprostatic disease extension. Patients with seminal vesical involvement (pT3c disease) or positive lymph nodes have been shown to have a much poorer outcome with regard to PSA-free survival and in our opinion clearly are candidates for adjuvant therapy trials.1, 8 Therefore, patients with seminal vesicle involvement or lymph node metastasis, or those who received adjuvant therapy, were excluded from the current study cohort. To define the outcome of surgically treated pT3 (pT3a/b) disease further, an analysis of a large number of patients with extraprostatic disease extension was performed to measure the impact of surgical margin status on outcome.
The lack of uniform reporting of the pathologic disease stage has led to confusion regarding the natural history of pT3 disease. To our knowledge, until recently many studies defined positive surgical margins in organ-confined carcinoma as pT3 disease,10 whereas in other series such cases were classified as pT2 disease. The staging system agreed on at the 1996 International Conference on Prostatic Intraepithelial Neoplasia and Pathological Staging (which was sponsored by the World Health Organization, the American Urological Association, the International Society of Urologic Pathologists, and the Mayo Clinic) states that only patients demonstrating extraprostatic disease extension or seminal vesicle extension are considered to have pT3 disease regardless of surgical margin status.14, 15 A classification of pT3a disease signifies focal extraprostatic disease extension (i.e., no more than one high-power field outside the prostate on no more than two separate sections). In excess of these criteria, the extraprostatic involvement would be considered extensive, and thus the patient would be considered to have pT3b disease.15 Prior studies have demonstrated the negative impact of positive surgical margins in patients with organ-confined carcinoma,2 but to our knowledge the finding in patients with extraprostatic disease remains unclear.
Overall, the 5-year biochemical progression-free survival for 842 patients with pT3a/b disease was 71%. Epstein et al. found elevated PSA in 18% of patients with limited extraprostatic disease extension versus 35% in patients with extensive extraprostatic disease extension at 5 years.6 This finding is similar to the 64% progression-free survival rate at 4 years reported by Lowe and Lieberman in their series of patients with pT3a/b disease.11 In the series reported by Paulson, the disease-free probability was much less in a cohort of patients with advanced, surgical margin-positive disease (only 30% at 10 years).13 However, this analysis did not control for extraprostatic involvement or seminal vesicle involvement. In a study of 507 patients with clinically confined prostate carcinoma, Epstein et al. reported that a multivariate analysis failed to demonstrate statistical significance for extraprostatic disease extension after adjusting for Gleason score and surgical margin status.15 The data from the current study suggest that a majority of patients with limited extraprostatic disease extension are potentially cured with radical prostatectomy alone. The 5-year progression-free survival rate for patients with negative surgical margins in the face of extraprostatic disease extension was 76% versus 64% in patients in whom ≥ 1 surgical margins is involved. We did not find any significant difference with regard to PSA progression in those patients with 1 versus those patients with ≥ 2 involved surgical margins (65% vs. 62%).
To assess the impact of surgical margin status independent of the Gleason score of the tumor, we subclassified patients into three separate Gleason scores. Patients with limited extraprostatic extension in the Gleason score 2–4 group had minimal rates of disease progression and the surgical margin status was not found to be a significant predictor of disease recurrence. In patients with tumors with Gleason scores of 5–6 tumors, surgical margin status was a reported to be a significant predictor of recurrence (P = 0.0008), as was anticipated. In those patients with high-grade tumors (a Gleason score of ≥ 7), the surgical margin status remained a significant predictor of disease recurrence. This finding is in agreement with Epstein et al., who reported that surgical margin status was predictive of recurrence in men with tumors with a Gleason score of 7 and limited extraprostatic extension.15 Somewhat different results were reported by Ohori et al., who noted that surgical margin status did not appear to be correlated with disease progression in 97 men with extraprostatic disease extension and Gleason scores of ≥ 7.7 D'Amico et al. reported a multivariate analysis that found positive surgical margin status to be predictive of failure with extracapsular extension in the absence of high-grade disease or seminal vesicle involvement.16
As in other series, the majority of failures reported herein were biochemical only (85%) whereas 40 patients (15%) had local or systemic disease with a median of 4.7 years to disease progression. These results should be considered preliminary because our findings are based on the interval to disease progression assessed by PSA levels in patients followed for a limited time. Longer follow-up is needed to determine the true clinical significance of the impact of surgical margin status on disease-free survival. However, when controlling for other variables such as preoperative PSA level, Gleason score, and DNA ploidy, our multivariate analysis revealed that a positive surgical margin is a significant predictor (P = 0.0017) of clinical recurrence or PSA progression.
The results of the current study indicate that surgical margin status is important in those patients with limited extraprostatic carcinoma. The overall prognosis of patients with extraprostatic disease (pT3a/b disease) in the absence of surgical margin involvement is excellent and comparable to that of patients with organ-confined disease.2 Conversely, patients with positive surgical margins are reported to have a PSA failure-free survival rate that is significantly lower and is independent of preoperative PSA level, Gleason score, and DNA ploidy. The benefit of adjuvant therapy or the adjustment of surgical techniques to optimize recurrence-free survival in both these groups remains to be evaluated.