the Departments of Pathology and Laboratory Medicine, Urology, and Division of Biostatistics, Indiana University School of Medicine, Indianapolis, IN

    ABSTRACT

    PURPOSE: Clinical outcome is variable in prostate cancer patients treated with radical prostatectomy. The Gleason histologic grade of prostatic adenocarcinoma is one of the strongest predictors of biologic aggressiveness of prostate cancer. We evaluated the significance of the relative proportion of high-grade cancer (Gleason patterns 4 and/or 5) in predicting cancer progression in prostate cancer patients treated with radical prostatectomy.

    PATIENTS AND METHODS: Radical prostatectomy specimens from 364 consecutive prostate cancer patients were totally embedded and whole mounted. Various clinical and pathologic characteristics were analyzed. All pathologic data, including Gleason grading variables, were collected prospectively.

    RESULTS: A multiple-factor analysis was performed that included the combined percentage of Gleason patterns 4 and 5, Gleason score, tumor stage, surgical margin status, preoperative prostate-specific antigen (PSA), extraprostatic extension, and total tumor volume. Using Cox regression analysis with bootstrap resampling for predictor selection, we identified the combined percentage of Gleason patterns 4 and 5 (P < .0001) and total tumor volume (P = .009) as significant predictors of PSA recurrence.

    CONCLUSION: The combined percentage of Gleason patterns 4 and 5 is one of the most powerful predictors of patient outcome, and appears superior to conventional Gleason score in identifying patients at increased risk of disease progression. On the basis of our results, we recommend that the combined percentage of Gleason patterns 4 and 5 be evaluated in radical prostatectomy specimens. The amount of high-grade cancer in a prostatectomy specimen should be taken into account in therapeutic decision making and assessment of patient prognosis.

    INTRODUCTION

    The Gleason grading system has been adopted recently by the WHO as the international standard for histologic grading of prostate cancer.1 It is based on evaluation of the glandular architecture of the tumor at low magnification; the Gleason score is derived by summation of the predominant pattern (primary Gleason pattern) with the next most prevalent pattern (secondary Gleason pattern). The original Gleason grading scheme does not address the issue of tertiary (third most prevalent) patterns.2

    Prostate cancer is multifocal.3,4 Molecular studies5,6 suggest that multiple tumors can arise independently within a single prostate, probably due to field effect. In a detailed mapping study of 115 whole-mount prostatectomy specimens, Arora et al7 found that two or more separate adenocarcinoma foci were present in 87% of radical prostatectomy specimens, and there often was extensive histologic heterogeneity among tumors within the same specimen. Gleason scores of individual tumors often did not correlate with overall Gleason scores. Tertiary, minor, high-grade components (Gleason pattern 4 or 5) were often identified in specimens with overall Gleason scores of 5 or 6.7 Cancer heterogeneity is well recognized in a variety of tumor types. In many cancers, the presence of any high-grade elements dictates the biologic behavior of tumors.8–11 In this study, we evaluated the significance of the relative proportion of high-grade cancer (Gleason patterns 4 and/or 5) in predicting cancer progression in patients treated with radical prostatectomy.

    PATIENTS AND METHODS

    Patients

    The study group consisted of 364 consecutive patients treated with radical retropubic prostatectomy with or without bilateral pelvic lymphadenectomy between 1999 and 2003 at the Indiana University Hospital (Indianapolis, IN). Patients who received preoperative androgen-deprivation therapy were excluded. Serum prostate-specific antigen (PSA) was measured using the Immulite PSA assay (Diagnostics Products Corporation, Los Angeles, CA). Biochemical recurrence was defined as a PSA value of at least 0.1 ng/mL after surgery.12–14 The patients were followed at 1, 3, 6, 12, 18, and 24 months during the first two years. Thereafter, follow-up was adjusted according to the clinical situation but occurred at least annually. The mean follow-up was 14 months (median, 12 months; range, 1.5 month to 48 months). This research was approved by the Indiana University institutional review board.

    Radical Prostatectomy Specimens

    Each radical prostatectomy specimen was totally embedded and processed by the whole-mount method.6,15–22 Each prostate was weighed, measured, inked, and fixed in 10% neutral formalin. Following fixation, the apex and base were coned and serially sectioned. The seminal vesicles were sectioned parallel to the junction with the prostate and submitted entirely for examination. The remaining prostate was serially sectioned perpendicular to the long axis of the gland from the apex to the tips of the seminal vesicles, and whole mount sections were prepared.

    Histopatholologic evaluation of the prostatectomy specimens were performed by a single pathologist (L.C.), and the data were collected prospectively. The 1997 TNM (tumor, nodes, metastasis) system was used for pathologic staging.23 Six patients had lymph node metastasis at the time of surgery. Surgical margins were considered positive when carcinoma cells were in contact with the inked margins.24 All cancers were assigned Gleason scores according to the Gleason grading system.2 The Gleason score is obtained from the summation of primary Gleason pattern (the most prevalent pattern) and secondary Gleason pattern (the second most prevalent), based on assessment of the entire specimen. The percentage of Gleason pattern 4 or 5 cancer in each case (ie, the proportion of the tumor comprising Gleason pattern 4 or 5 cancer) was evaluated separately and semiquantitatively on a 5%-increment scale ranging from 0% to 100%. The combined percentage of Gleason patterns 4 and 5 was derived by the summation of the percentages of Gleason patterns 4 and 5. Only malignant epithelial elements were considered in the calculation of the percentage of Gleason patterns 4 and/or 5. The volume of carcinoma in the entire prostate was calculated using the grid method15,25 and was the sum of the volumes of individual foci of tumor.

    Statistical Analysis

    SAS version 8.2 (SAS Institute, Cary, NC) was used to perform the analysis. Table 1 lists patient characteristics categorized according to whether their cancers did or did not include components of Gleason patterns 4 or 5. The Wilcoxon rank sum test was used to test for associations between presence of Gleason patterns 4 or 5 and continuous variables. Fisher's exact test was used to test for associations between presence of Gleason patterns 4 or 5 and categoric variables.

    Time to recurrence of PSA was tested for association with pathologic stage, surgical margins, Gleason score, combined percentage of Gleason patterns 4 and 5, preoperative PSA, and extraprostatic extention using Cox univariate and multiple regression analyses. Ties were handled by the Breslow method. There were 45 patients with recurrence and 303 patients (87%) were excluded. For Cox multiple regression, bootstrap resampling was used in order to implement variable selection and choose the model with the most important predictors.26 Robust parameter estimates and SEs were calculated from the final model. A total of 500 bootstrap samples, generated using resampling with replacement, were analyzed using Cox regression. Backward elimination was used to reduce the number of factors, and only models with one to four variables were considered due to the small number of PSA recurrence.27 Convergence was confirmed at each step of the elimination procedure. The criterion for each factor remaining in the last step for each replication was P < .1, marginal significance level. From the results of the 500 replications, a process was used separately to select predictors based on stringent and relaxed threshold conditions.

    To further evaluate the importance of the combined percentage of Gleason patterns 4 and 5, we assessed the correlation between the combined percentage of Gleason patterns 4 and 5 and other patient characteristics using either Spearman's correlation coefficient or Fisher's exact test.

    RESULTS

    Table 1 lists patient characteristics categorized according to the presence or absence of Gleason pattern 4 or 5 cancer in the radical prostatectomy specimen. Among 243 patients with Gleason pattern 4, the median percentage of Gleason pattern 4 tumor was 20% (range, 1% to 80%). Among 81 patients with Gleason pattern 5, the median percentage of Gleason pattern 5 tumor was 10% (range, 1% to 80%). Among the 244 patients with either Gleason pattern 4 or 5, the median percentage of combined Gleason patterns 4 and 5 was 20% (range, 1% to 100%); 163 specimens had Gleason pattern 4 only, one specimen had Gleason pattern 5 only, and 80 specimens had both Gleason patterns 4 and 5. The overall Gleason scores were 4 (one case), 5 (63 cases), 6 (114 cases), 7 (151 cases), 8 (9 cases), and 9 (26 cases). Among 63 patients with overall Gleason scores 5, Gleason patterns 4 and 5 were present in 20 patients; among 114 patients with overall Gleason scores 6, Gleason patterns 4 and 5 were present in 38 patients.

    In univariate Cox regression analyses using the original sample, the presence of extraprostatic extention, higher tumor volume, Gleason score, the combined percentage of Gleason patterns 4 and 5, preoperative PSA and surgical margins all were associated with PSA recurrence. Stage T2 was associated with less recurrence compared with stage T3. Results from univariate regression analyses are presented in Table 2 (P < .0001).

    Time to PSA recurrence was also analyzed using Cox multiple regression analysis including predictors pathologic stage, surgical margins, tumor volume, preoperative PSA, extraprostatic extension, Gleason score, and the combined percentage of Gleason patterns 4 and 5 (Table 3). In order to avoid overfitting, a bootstrap resampling procedure with 500 replications was used to perform Cox regression analyses using backward elimination. A method, described by Sauerbrei,26 was used to simplify the regression model as well as to obtain robust parameter estimation. The goal of this method is to identify the most important predictors. Multiple Cox regression analyses were performed with 500 replicated samples. Each factor in the replicated models was counted conditional on two threshold values, P < .01 for the stringent version and P < .05 for the relaxed version.

    Model selection and robust estimation of parameters were based on 415 replications due to factor correlation and model stability. Although many of the factors were weakly to moderately correlated, Gleason score and the combined percentage of Gleason patterns 4 and 5 were highly correlated (r = 0.85; P < .0001). Although all models of replicated data included one of the two predictors, very few included both. These results suggest that only one of these factors should be included in the model. In addition, models with five or six factors were excluded due to limitations of total number of PSA recurrence. Results for the remaining replications are presented in Table 4 for the stringent and relaxed conditions.

    Using both conditions and the criterion of at least 30% occurrence, predictors Gleason patterns 4 and 5 and tumor volume were selected. Robust parameter estimates were calculated from bootstrapped replications with models including predictors, Gleason patterns 4 and 5, and tumor volume. Estimates were comparable to original data results. Higher Gleason patterns 4 and 5 and tumor volume were associated with increased risk of PSA recurrence (hazard ratio [HR] = 1.028, P < .0001; and HR = 1.104, P = .011, respectively). For a 1-unit increase in Gleason patterns 4 and 5, there was a 2.8% increase in hazard of PSA recurrence. For a 1-unit increase in tumor volume, there was a 9.9% increase in hazard of PSA recurrence. Figure 1 shows the relationship between the combined percentage of Gleason patterns 4 and 5 and incidence of PSA recurrence.

    To further explore the importance of the combined percentage of Gleason patterns 4 and 5, the association between the combined percentage of Gleason patterns 4 and 5 and other patient characteristics was evaluated. We found that the combined percentage of Gleason patterns 4 and 5 correlated with age (r = 0.11; P = .03), preoperative PSA (r = 0.18; P = .001), pathologic stage (P < .0001), total tumor volume (r = 0.47; P < .0001), extraprostatic extension (P < .0001), seminal vesicle invasion (P < .0001), surgical margin status (P < .0001), and perineural invasion (P < .0001). Correlations with age and extraprostatic extension were weak. The combined percentage of Gleason patterns 4 and 5 was not associated with multifocality.

    DISCUSSION

    The biologic aggressiveness of prostate cancer is directly related to the degree of tumor differentiation. The Gleason grading system is widely used in characterizing prostate cancer, and has been linked to clinical outcome in numerous reports, including Gleason's original study.2 In this report, we evaluated 364 totally embedded, serially sectioned, whole-mount prostatectomy specimens that were processed uniformly. We found that Gleason score and the combined percentage of Gleason patterns 4 and 5 were predictive of PSA recurrence. However, based on Cox regression analysis with bootstrap resampling for predictor selection, the combined percentage of Gleason patterns 4 and 5 was the stronger predictor of PSA recurrence. The prognostic utility of the Gleason grading system may be improved by a simple quantification of high-grade (Gleason pattern 4 or 5) tumor. On the basis of our results, we recommend that the combined percentage of Gleason patterns 4 and 5 be reported in the examination of radical prostatectomy specimens.

    The Gleason score is derived from the summation of primary Gleason pattern (the most prevalent pattern) and secondary Gleason pattern (the second most prevalent), and has been well established as one of the most powerful predictors of patient outcome.10,25,28–30 One of the limitations of the Gleason scoring system is that the majority of prostate carcinomas have Gleason scores of 6 or 7. Prostate cancer is a multifocal and heterogenous disease with diverse histopathologic patterns.5,6,31 In evaluating a series of whole-mounted radical prostatectomies, Aihara et al found that more than half contain three or more different Gleason patterns.31 The influence of tertiary Gleason pattern on Gleason grading in a given tumor was not assessed in the original Gleason grading scheme.2,32 McGowan et al concluded that simple identification of primary and secondary Gleason pattern was inadequate to assess prognosis, and that the worst histologic pattern was most important in predicting patient outcome.33 Recently, Pan et al found that a high-grade component (Gleason pattern 4 or 5 tumor), though comprising only a small percentage of the neoplasm, portended a worse prognosis after radical prostatectomy.34

    Evidence has accumulated supporting the concept that the percentage of Gleason pattern 4 or 5 tumor is an important prognostic factor.35–38 With the same Gleason score 7, tumors with more than 50% of Gleason pattern 4 (ie, Gleason score 4 + 3) portend a significantly worse prognosis than tumors with less than 50% Gleason pattern 4 tumor (ie, Gleason score 3 + 4).39–42 McNeal et al35 analyzed 209 entirely embedded radical prostatectomy specimens and found that the percentage of Gleason patterns 4 and 5 was predictive of lymph node metastasis. Stamey et al37 found that the combined percentage of Gleason patterns 4 and 5 provided additional prognostic information beyond Gleason score and proposed that pathologists should move away from the traditional Gleason scoring system and simply estimate the combined percentage of Gleason patterns 4 and 5. In the current study, our multiple factor analyses indicated that the combined percentage of Gleason pattern 4 or 5 tumors was superior to Gleason score in predicting prognosis.

    Accurate histopathologic analysis of prostate cancer is an important aspect of clinical management. One of the strengths of our study is that all pathologic data were collected prospectively and evaluated by a single genitourinary pathologist, ensuring accurate assessment of Gleason grading in each case and eliminating interobserver variability. Our use of the whole-mount technique for uniform specimen processing minimizes the required number of paraffin blocks and slides per cross-section, ensures assessment of the entire specimen, allows accurate histopathologic analysis of various patterns of Gleason grade in the same prostate, and facilitates accurate estimation of cancer volume. Nonetheless, several limitations of our study should be acknowledged. Patient follow-up was short, necessitating the use of a surrogate end point for analysis. Small sample size in terms of outcome events limited the number of predictors that could be assessed in multiple regression analysis. Therefore, our results should be considered preliminary. Nonetheless, identification of patients at risk of early treatment failure is clinically important. We found that the presence of Gleason pattern 4 or 5 tumor was correlated with other adverse features commonly associated with worsening prognosis—higher preoperative PSA levels, higher pathologic stage, greater total tumor volume, and higher incidences of extraprostatic extension, seminal vesicle invasion, and positive surgical margins. Quantification of high-grade prostate cancer provides additional prognostic information beyond conventional Gleason score and can be readily performed in routine pathology practice.

    Gleason grading provides an estimate of biologic aggressiveness of prostate cancer. Our data suggest that the inherent biologic aggressiveness of prostate cancer is closely linked to the presence of high-grade (Gleason pattern 4 or 5) tumor. Quantification of the combined percentage of Gleason patterns 4 and 5 prostate cancer in a radical prostatectomy specimen provides the best estimate of risk for cancer progression, and serves as one of the most powerful predictors of patient outcome. Our findings suggest that this strategy may be superior to conventional Gleason scoring in identifying patients at increased risk of disease progression. Therefore, we recommend that the combined percentage of Gleason patterns 4 and 5 be evaluated and reported in radical prostatectomy specimens. The combined percentage of Gleason patterns 4 and 5 should be considered in stratification variables in designing future clinical trials.

    Authors' Disclosures of Potential Conflicts of Interest

    The authors indicated no potential conflicts of interest.

    NOTES

    Authors' disclosures of potential conflicts of interest are found at the end of this article.

    REFERENCES

    Eble JN, Epstein JI, Sauter G, et al: WHO Classification of Tumours: Pathology and Genetics. Tumours of the Urinary and Male Reproductive System. Lyon, France, IARC Press, 2004, pp 159–215

    Gleason DF, Mellinger GT, the Veterans Administration Cooperative Urologic Research Group: Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol 111:58–64, 1974

    Greene DR, Egawa S, Neerhut G, et al: The distribution of residual cancer in radical prostatectomy specimens in stage A prostate cancer. J Urol 146:1069–1076, 1991

    Eichelberger LE, Cheng L: Does pT2b cancer exist Critical appraisal of the 2002 TNM classification of prostate cancer. Cancer 100:2573–2576, 2004

    Bostwick DG, Shan A, Qian J, et al: Independent origin of multiple foci of prostatic intraepithelial neoplasia: Comparison with matched foci of prostate carcinoma. Cancer 83:1995–2002, 1998

    Cheng L, Song SY, Pretlow TG, et al: Evidence of independent origin of multiple tumors from patients with prostate cancer. J Natl Cancer Inst 90:233–237, 1998

    Arora R, Koch MO, Eble JN, et al: Heterogeneity of Gleason grade in multifocal adenocarcinoma of the prostate. Cancer 100:2362–2366, 2004

    Broders AC: The grading of carcinoma. Minn Med 8:726–730, 1925

    Broders AC: Carcinoma grading and practical application. Arch Pathol Lab Med 2:376–381, 1926

    Cheng L, Bergstralh EJ, Slezak J, et al: Dedifferentiation in metastatic progression of prostate cancer. Cancer 86:657–663, 1999

    Cheng L, Neumann RM, Nehra A, et al: Cancer heterogeneity and its biologic implications in the grading of urothelial carcinoma. Cancer 88:1663–1670, 2000

    Babaian RJ, Troncoso P, Bhadkamkar VA, et al: Analysis of clinicopathologic factors predicting outcome after radical prostatectomy. Cancer 91:1414–1422, 2001

    Koch MO, Foster RS, Bell B, et al: Characterization and predictors of prostate specific antigen progression rates after radical retropubic prostatectomy. J Urol 164:749–753, 2000

    van den Ouden D, Bentvelsen FM, Boeve ER, et al: Positive margins after radical prostatectomy: Correlation with local recurrence and distant progression. Br J Urol 72:489–494, 1993

    Humphrey PA, Vollmer RT: Intraglandular tumor extent and prognosis in prostatic carcinoma: Application of a grid method to prostatectomy specimens. Hum Pathol 21:799–804, 1990

    Hong H, Koch MO, Foster RS, et al: Anatomic distribution of periprostatic adipose tissue: A mapping study of 100 radical prostatectomy specimens. Cancer 97:1639–1643, 2003

    Cheng L, Darson MF, Bergatrahl EJ, et al: Correlation of margin status and extraprostatic extension with progression of prostate carcinoma. Cancer 86:1775–1781, 1999

    Ng JC, Koch MO, Daggy JK, et al: Perineural invasion in radical prostatectomy specimens. J Urol 172:2249–2251, 2004

    Poulos CK, Daggy JK, Cheng L: Prostate needle biopsies: Multiple variables are predictive of final tumor volume in radical prostatectomy specimens. Cancer 101:527–532, 2004

    Poulos CK, Koch MO, Eble JN, et al: Bladder neck invasion is an independent predictor of prostate-specific antigen recurrence. Cancer 101:1563–1568, 2004

    Emerson RE, Koch MO, Daggy J, et al: Closest distance between tumor and resection margin in radical prostatectomy specimens: Lack of prognostic significance. Am J Surg Pathol 29:224–229, 2005

    Poulos CK, Daggy JK, Cheng L: Preoperative prediction of Gleason grade in radical prostatectomy specimens: The influence of different Gleason grades from multiple positive biopsy sites. Mod Pathol 18:228–234, 2005

    Fleming ID, Cooper JS, Henson DE, et al: AJCC Cancer Staging Manual. Philadelphia, Lippincott Raven, 1997, pp 203–207

    Cheng L, Slezak J, Bergstralh EJ, et al: Preoperative prediction of surgical margin status in prostate cancer patients treated by radical prostatectomy. J Clin Oncol 18:2862–2868, 2000

    Cheng L, Bergstralh EJ, Cheville JC, et al: Cancer volume of lymph node metastasis predicts progression in prostate cancer. Am J Surg Pathol 22:1491–1500, 1998

    Sauerbrei W: The use of resampling methods to simplify regression models in medical statistics. Applied Statistics 48:313–329, 1999

    Harrell FE, Lee KL: The practical value of logistic regression: Proceedings of the Tenth Annual SAS Users Groups International Conference, SAS Institute Inc, Cary, North Carolina:1031–1036, 1985

    Bostwick DG: Grading prostate cancer. Am J Clin Pathol 102:S38–S56, 1994

    Cheng L, Zincke H, Blute ML, et al: Risk of prostate cancer death in patients with lymph node metastasis. Cancer 91:66–73, 2001

    Bostwick DG, Grignon DJ, Hammond EH, et al: Prognostic factors in prostate cancer: College of American Pathologists consensus statement 1999. Arch Pathol Lab Med 124:995–1000, 2000

    Aihara M, Wheeler TM, Ohori M, et al: Heterogeneity of prostate cancer in radical prostatectomy specimens. Urology 43:60–77, 1993

    Mellinger GT, Gleason DF, Bailar J: The histology and prognosis of prostate cancer. J Urol 97:331–337, 1967

    McGowan DG, Bain GO, Hanson J: Evaluation of histological grading (Gleason) in carcinoma of the prostate: Adverse influence of highest grade. Prostate 4:111–118, 1983

    Pan C, Potter SR, Partin AW, et al: The prognostic significance of tertiary Gleason patterns of higher grade in radical prostatectomy specimens: A proposal to modify the Gleason grading system. Am J Surg Pathol 24:563–569, 2000

    McNeal JE, Villers AA, Redwine EA, et al: Histologic differentiation, cancer volume, and pelvic lymph node metastasis in adenocarcinoma of the prostate. Cancer 66:1225–1233, 1990

    McNeal JE, Kindrachuk RA, Freiha FS, et al: Patterns of progression in prostate cancer. Lancet 1:60–63, 1986

    Stamey TA, McNeal JE, Yemoto CM, et al: Biological determinants of cancer progression in men with prostate cancer. JAMA 281:1395–1400, 1999

    Blackwell KL, Bostwick DG, Myers RP, et al: Combining prostate specific antigen with cancer and gland volume to predict more reliably pathological stage: The influence of prostate specific antigen cancer density. J Urol 151:1565–1570, 1994

    Herman CM, Kattan MW, Ohori M, et al: Primary Gleason pattern as a predictor of disease progression in Gleason score 7 prostate cancer: A multivariate analysis of 823 men treated with radical prostatectomy. Am J Surg Pathol 25:657–660, 2001

    Lau WK, Blute ML, Bostwick DG, et al: Prognostic factors for survival of patients with pathological Gleason score 7 prostate cancer: Differences in outcome between primary Gleason grades 3 and 4. J Urol 166:1692–1697, 2001

    Sakr WA, Tefilli MV, Grignon DJ, et al: Gleason score 7 prostate cancer: A heterogeneous entity Correlation with pathologic parameters and disease-free survival. Urology 56:730–734, 2000

    Chan TY, Partin AW, Walsh PC, et al: Prognostic significance of Gleason score 3 + 4 versus Gleason score 4 + 3 tumor at radical prostatectomy. Urology 56:823–827, 2000

查询更多Predictor相关信息在本站>>

《临床肿瘤学医学期刊》2005年5月第23卷第5期