Prognostic value of the Gleason score in prostate cancer

Transcription

1 BJU International (22), 89, Prognostic value of the Gleason score in prostate cancer L. EGEVAD, T. GRANFORS*, L. KARLBERG*, A. BERGH and P. STATTIN Department of Pathology and Cytology, Karolinska Hospital, Stockholm, *Department of Urology, Central Hospital, Västerås, and Departments of Pathology and Urology and Andrology, Umeå University Hospital, Umeå, Sweden Objective To investigate the prognostic value of the Gleason score in prostate cancer. Patients and methods A consecutive series of 35 men with prostate cancer diagnosed at transurethral resection (975 99) and with no curative treatment was analysed. There was no assessment of prostate-specific antigen level during this period. The mean (range) age at diagnosis was 73.7 (52 95) years and the mean follow-up was 6.4 ( 22) years. The influence of Gleason score and the percentage of the specimen area with tumour (% cancer) on diseasespecific survival were assessed using Kaplan Meier analyses. Results Of 35 cancers, 22% had a Gleason score of 4 5, 29% of 6, 8% of 7 and 32% of 8. At the follow-up, 89% of the men had died, of whom 42% had died from prostate cancer. The disease-specific -year survival was 56%. The disease-specific mean survival (DSMS) for Gleason score 4 5, 6, 7 and 8 was 2, 6, and 5 years, respectively (P<.). The DSMS did not differ significantly between Gleason 4 and 5 or between 8. There was a trend towards shorter survival for Gleason 4+3=7 (DSMS 9 years) than GS 3+4=7 (DSMS 3 years; P=.6). Gleason score and % cancer were independent predictors of DSMS (P<.). Conclusion The long-term prognosis of prostate cancer on deferred treatment is predicted well by the Gleason score. Four prognostic categories of prostate cancer are suggested, i.e. Gleason score 4 5, 6, 7 and 8. Keywords prostatic neoplasms, pathology, classification, prostatectomy, male, human Introduction Since the first articles by Gleason [ 3] many studies have been published on the prognostic value of the Gleason grading system [4 8]. Most of them studied the Gleason score as a predictor of recurrence after total prostatectomy [4 6] or radiotherapy [7,8]. However, some of these studies have a short follow-up and as the patients in these studies received treatment intended to cure, the prognostic effect of histopathology will not be apparent if a patient with an unfavourable tumour was cured by the treatment. Thus, studies on series of men with prostate cancer with no curative treatment, with a long follow-up and where a substantial proportion of the men died from their cancer remains of interest when the prognostic value of a marker is being evaluated. Many of the studies that have investigated the natural history of prostate cancer have classified the tumours according to a three-tier grading system [9 5]. Hence, the results are difficult to compare with more recent data on the Gleason score as a predictor of recurrence after, e.g. total prostatectomy. The aim of the Accepted for publication 7 December 2 present study was to investigate the prognostic value of the Gleason score in prostate cancer. Patients and methods The study included 35 consecutive men with prostate cancer diagnosed at TURP for obstructive symptoms from April 975 to December 99 at Västerås Central Hospital, Sweden. During the study period the approach to the diagnosis and treatment of prostate cancer was very conservative. If the patient had obstructive symptoms and a palpable tumour, histopathological confirmation of the diagnosis was obtained at TURP, with no previous fine-needle aspiration cytology or coreneedle biopsy, and serum levels of PSA were not measured. Of all men, 44 were classified as clinical stage Ta, 3 stage Tb, 5 stage T2, 37 stage T3 and eight stage T4; in eight men, the clinical stage was unknown. None of the patients had received hormonal treatment or radiotherapy before TURP. The patients were managed by deferred treatment. Hormonal treatment was used in patients with symptomatic, locally progressive or metastatic disease. The patient records and the death certificates were reviewed. The patients were classified as alive at follow-up on 5 September 2, dead from prostate 538 # 22 BJU International

2 GLEASON SCORE IN PROSTATE CANCER 539 cancer or dead from unrelated causes. The histological slides were reviewed by one of the authors (L.E.) and cancer outlined with India ink. The Gleason score was assessed and the percentage of the specimen area involved with tumour (% cancer) estimated as focal (<5%) and at % intervals (, 5, 6, 2 etc.). Survival was analysed using Kaplan Meier plots with log-rank comparison between groups. The relative survival was calculated as the ratio between the observed survival of the patients and the expected survival of age-matched controls according to Swedish population tables [6]. A Cox proportional hazards model was used to compare Gleason score, % cancer, T stage, distant metastases and patient age at diagnosis as prognostic variables. The chi-squared test was used for comparing proportions between groups; in all tests P<.5 was considered significant. Results The mean (range) age of the men at diagnosis was 73.7 (52 95) years with 32% (97/35) of the patients aged <7 years. The relative and cause-specific survival rates correlated closely, indicating a correct assessment of the cause of death (Fig. ). The mean age at diagnosis of men who later died from prostate cancer was 72.4 years and of men who died from unrelated disease was 75.8 years. The mean (range) time from diagnosis to follow-up was 6.4 ( 22) years. At follow-up, 89% (27/35) of the patients were dead; of these, 4.7% had died from prostate cancer (Table ). Among patients aged <7 years at diagnosis, over half had died from prostate cancer, compared with over a third of those aged >7 years (Table ). The disease-specific mean (95% CI) survival (DSMS) was 2.4 (.2 3.6) years. The mean (95% CI) disease-specific survival rate after 5 and years was 74.4 ( )% and 55.6 ( )%, respectively (Fig. ). Of all prostate cancer-related deaths, 4.6% (47/3) occurred >5 years after diagnosis and.6% (2/3) occurred after years. The distribution of the Gleason scores is shown in Fig. 2; Gleason scores 5 7 included 65% (98/35) of all patients. The CIs of the survival of the men with Gleason scores 4 5 and 8 overlapped, while there was no overlap between those with Gleason scores 5 and 6, 6 and 7, and 7 and 8 (Table 2). Therefore, the tumours were grouped into four prognostic categories, i.e. scores 4 5, 6, 7 and 8, respectively. The DSMS was 9.5, 5.6, 9.9 and 4.7 years for the four categories, respectively (P<.; Fig. 3); of these patients, 4.5%, 23.6%, 4.8% and 7.2% respectively, died from prostate cancer (Table 3). The respective relative risks for death from prostate cancer in these categories were. (score 4 5, reference category), 5.9, 2.4 and A larger proportion of the cancer-related fatalities occurred late among patients with Gleason 4 6 tumours than in patients with Gleason 7 tumours. Of the patients who died from prostate cancer, 58% (4/24) with Table The number (%) of patients dead from disease, from unrelated causes and alive at the follow-up Outcome Age at diagnosis, years <7 o7 All.8 Dead from: prostate cancer % of all % of dead other causes Alive at follow-up Survival % Fig.. Disease-specific survival of all 35 patients (green solid line) compared with the relative survival (red dotted line; the ratio between the observed survival and the expected survival of agematched controls). The numbers of patients at risk at diagnosis and at 5, and 5 years were 35, 6, 72 and GS Fig. 2. The distribution of Gleason scores; score 5 7 included 65% of the tumours # 22 BJU International 89,

3 54 L. EGEVAD et al. Disease-specific survival GS 4 5 GS 6 GS 7 Disease-specific survival GS Fig. 3. The tumours were grouped into four categories according to the Gleason score, i.e. 4 5, 6, 7 and 8. The disease-specific survival of these categories differed significantly (P<.); the numbers of patients at risk in the respective categories at 5//5 years were 46/3/2, 53/25/6, 3//2 and 32/7/. Fig. 4. The disease-specific survival of patients with tumours involving f5% of the specimen area (green solid line) was significantly better than in patients with larger tumours (>5% cancer, red dotted line; P<.). The numbers of patients at risk at 5//5 years in the respective categories was 35/9/4 and 26/53/7. Table 2 DSMS according to the Gleason scores Gleason score DSMS (95% CI), years ( ) (7. 8.8) (2.4 6.) 7.4 ( ) ( ) 9 5. ( ) 4.7 (.8 7.6) Table 4 The distribution of Gleason scores within each T stage Stage Gleason score, n (%) Total Ta 23 (52) 8 (4) 2 (4.5) (2) 44 Tb 23 (22) 4 (4) 25 (24) 4 (4) 3 T2 7 (6) 26 (25) 9 (8) 43 (4) 5 T3 3 (8) 2 (5) 6 (6) 26 (7) 37 T Unknown Table 3 The number (%) of patients dead from prostate cancer according to Gleason score category Gleason score Dead from prostate cancer Total (4.5) (23.6) (4.8) (7.2) 94 Total 3 35 Gleason score, % cancer, T stage and patient s age were entered in a Cox proportional hazards model, Gleason score (P<.) and % cancer (P=.), but not age (P=.3) were independent predictors of DSMS. When T stage was substituted with distant metastases in this model, age was still not independent (P=.), while the other variables remained independent (P<. and P=., respectively). Gleason 4 6 and 33% (33/99) with Gleason 7 died after >5 years (P=.24). Men with Gleason 4+3 tumours had a DSMS of 9. years compared with 2.5 years for men with Gleason 3+4 tumours (P=.6). In 2% (6/35) of the men, the specimens contained <5% cancer; the DSMS rate of this group was significantly higher than for those with a larger tumour (P<.; Fig. 4). The distribution of Gleason scores according to T stage is shown in Table 4. When the Discussion Knowledge of the natural history of prostate cancer, i.e. the prognosis with no curative treatment, remains of interest. Many studies have reported the outcome of prostate cancer with deferred treatment [9 4,7] but in most the tumours were graded using a three-tier system that is poorly standardized and now largely abandoned. In the present study, all specimens were reviewed by the same pathologist and graded according # 22 BJU International 89,

4 GLEASON SCORE IN PROSTATE CANCER 54 to the Gleason system. The Gleason score had a very strong prognostic value as a predictor of death from cancer. Only three patients (4%) with prostate cancer of Gleason score 4 5 died from cancer, as opposed to 7% of men with cancers of Gleason score 8. The Gleason scores have previously been grouped as Gleason scores 2 5, 6 7 and 8 [8], or 2 4, 5 6 and 7 [9], or 2 4, 5 7 and 8 [2], or 2 4, 5 6, 7 and 8 [2]. However, in the present study the DSMS of men with Gleason scores 4 5 and 8 overlapped considerably, while the prognosis of those with Gleason scores 6 and 7 was clearly differed. Thus we suggest that the Gleason scores of prostate cancer are categorized in four prognostic groups, i.e. 4 5, 6, 7 and 8. No tumours were assigned Gleason scores 2 or 3. The grade pattern is extremely rare; when grade is found, there is usually enough cancer of other grades present to exclude the grade pattern from the Gleason score. Some pathologists are also reluctant to use Gleason scores 4 5, but such tumours are usually transition-zone cancers and, as shown here, they have an excellent prognosis and should be distinguished from the more common Gleason score 6 tumours. Only 4% of men with Gleason score 4 5 tumours died from cancer, as opposed to 24% of men with Gleason score 6 tumours. Conventional histopathological variables that have been correlated with prognosis of prostate cancer are Gleason score, extraprostatic extension, margin status of the prostatectomy specimen, tumour volume and seminal vesicle invasion [4,6,22]. The practical value of some of these variables, e.g. extraprostatic extension, margin status and tumour volume, is limited because they are unknown until the patient has undergone total prostatectomy. The value of the tumour volume as an independent prognostic variable has been debated. In some studies tumour volume has had independent prognostic value [23], while in others it has added no prognostic information when other variables, e.g. extraprostatic extension, margin status and Gleason score, were included in the analysis [24]. In the present study, % cancer in a TURP specimen predicted outcome in a Cox model including T stage, Gleason score and age. For studies of prognostic markers in prostate cancer there are some advantages with the present TURP series over series using total prostatectomy or core biopsy specimens. First, TURP series are as close to the natural history of the disease as possible. Some patients treated by total prostatectomy are cured by their surgery, and in these cases the effect of the prognostic marker is abolished. Second, a long follow-up is very important for an accurate estimate of the prognosis of prostate cancer and this is still not feasible for most studies on cancer detected by core biopsies. In the present study, 89% of the patients were dead at the follow-up and the mean time from diagnosis to follow-up was 6.5 years. Hence, final outcome data could be collected from the vast majority of patients. Despite the relatively advanced age of this series, 42% of the patients who were dead at follow-up had died from prostate cancer, which is also an advantage when a prognostic marker is being evaluated. Third, although fewer prostate cancers are detected at TURP, with the increasing use of PSA testing, prostate cancer is still detected in <% of all TURP specimens, even when there is a normal serum PSA level before surgery [25]. Thus at present it remains of interest to know the prognosis of these TURP-detected cancers, to optimize subsequent treatment. There are several pitfalls in the interpretation of a series of prostate cancers detected at TURP. Transitionzone tumours are most likely over-represented compared with total prostatectomy specimens. The distribution of the Gleason scores was somewhat wider than usually seen in total prostatectomy specimens, with a larger proportion of the tumours assigned the lowest or the highest Gleason scores. Some small tumours may be resected in toto at TURP and in those patients the clinical course evidently does not describe the natural history. Epstein et al. [26] studied TURP-detected cancers with <5% cancer in the specimen and a Gleason score of <7 (pta). The patients underwent total prostatectomy and in three (4% of the cases), no residual tumour was found. Furthermore, a TURP specimen only allows the examination of a proportion of the tumour and the features of the tumour may be different in the periphery of the prostate. However, sampling artefacts are also a pitfall in core-biopsy studies and the strong prognostic value of the Gleason score in the present study indicates that the histopathology in most specimens is representative of the tumours. In the present decade the increased use of PSA testing and TRUS-guided needle biopsies to diagnose prostate cancer have changed the criteria for TURP. The characteristics of the patients and their tumours in a contemporary TURP series will be very different from that in The purpose of this paper was to analyse the value of the Gleason score in material obtained by TURP and the overall outcome in these patients should not be seen as an indicator of the prognosis of patients currently undergoing TURP for prostate cancer. In conclusion, the Gleason score was a strong prognostic factor in prostate cancer when evaluated by cancer-specific survival in patients with obstructive voiding symptoms, and prostate cancer treated with TURP and subsequent deferred treatment. Based on these findings, we suggest that the Gleason score should be categorized in four groups, i.e. 4 5, 6, 7 and 8 for optimal prognostic information. # 22 BJU International 89,

5 542 L. EGEVAD et al. Acknowledgements We thank Björn Tavelin, Oncological Center, Umeå University Hospital for calculating the relative survival. References Gleason DF. Classification of prostatic carcinomas. Cancer Chemotherapy Reports 966; 5: Gleason DF, Mellinger GT. Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol 974; : Gleason DF. Histologic grading and clinical staging of prostatic carcinoma. In Tannenbaum M ed. Urologic Pathology: the Prostate. Philadelphia: Lea and Feibiger, 977: Catalona WJ, Smith DS. 5-year tumor recurrence rates after anatomical radical retropubic prostatectomy for prostate cancer. J Urol 994; 52: Lerner SE, Blute ML, Bergstralh EJ, Bostwick DG, Eickholt JT, Zincke H. Analysis of risk factors for progression in patients with pathologically confined prostate cancers after radical retropubic prostatectomy. J Urol 996; 56: Epstein JI, Partin AW, Sauvageot J, Walsh PC. Prediction of progression following radical prostatectomy. A multivariate analysis of 72 men with long-term follow-up. Am J Surg Pathol 996; 2: Green GA, Hanlon AL, Al-Saleem T, Hanks GE. A Gleason score of 7 predicts a worse outcome for prostate carcinoma patients treated with radiotherapy. Cancer 998; 83: Zagars GK, Ayala AG, von Eschenbach AC, Pollack A. The prognostic importance of Gleason grade in prostatic adenocarcinoma: a long-term follow-up study of 648 patients treated with radiation therapy. Int J Radiat Oncol Biol Phys 995; 3: Adolfsson J, Carstensen J, Lowhagen T. Deferred treatment in clinically localised prostatic carcinoma. Br J Urol 992; 69: 83 7 Adolfsson J. Deferred treatment for clinically localized prostate cancer. Eur J Surg Oncol 995; 2: Adolfsson J, Steineck G, Hedlund PO. Deferred treatment of clinically localized low-grade prostate cancer: actual -year and projected 5-year follow-up of the Karolinska series. Urology 997; 5: Johansson JE, Adami HO, Andersson SO, Bergstrom R, Krusemo UB, Kraaz W. Natural history of localised prostatic cancer. A population-based study in 223 untreated patients. Lancet 989; : Johansson JE, Adami HO, Andersson SO, Bergstrom R, Holmberg L, Krusemo UB. High -year survival rate in patients with early, untreated prostatic cancer. JAMA 992; 267: Johansson JE, Holmberg L, Johansson S, Bergstrom R, Adami HO. Fifteen-year survival in prostate cancer. A prospective, population-based study in Sweden. JAMA 997; 277: Stattin P, Bergh A, Karlberg L, Tavelin B, Damber JE. Long-term outcome of conservative therapy in men presenting with voiding symptoms and prostate cancer. Eur Urol 997; 32: Hakulinen T. On long-term relative survival rates. J Chronic Dis 977; 3: Chodak GW, Thisted RA, Gerber GS et al. Results of conservative management of clinically localized prostate cancer. N Engl J Med 994; 33: de las Morenas A, Siroky MB, Merriam J, Stilmant MM. Prostatic adenocarcinoma, reproducibility and correlation with clinical stages of four grading systems. Human Pathol 988; 9: Lessells AM, Burnett RA, Howatson SR et al. Observer variability in the histopathological reporting of needle biopsy specimens of the prostate. Human Pathol 997; 28: Albertsen PC, Fryback DG, Storer BE, Kolon TF, Fine J. Long-term survival among men with conservatively treated localized prostate cancer. JAMA 995; 274: Epstein JI, Pizov G, Walsh PC. Correlation of pathologic findings with progression after radical retropubic prostatectomy. Cancer 993; 7: Gerber GS, Thisted RA, Scardino PT et al. Results of radical prostatectomy in men with clinically localized prostate cancer. JAMA 996; 276: Humphrey PA, Walther PJ, Currin SM, Vollmer RT. Histologic grade, DNA ploidy, and intraglandular tumor extent as indicators of tumor progression of clinical stage B prostatic carcinoma. A direct comparison. Am J Surg Pathol 99; 5: Epstein JI, Carmichael M, Partin AW, Walsh PC. Is tumor volume an independent predictor of progression following radical prostatectomy? A multivariate analysis of 85 clinical stage B adenocarcinomas of the prostate with 5 years of followup. J Urol 993; 49: Monda JM, Barry MJ, Oesterling JE. Prostate specific antigen cannot distinguish stage Ta (A) prostate cancer from benign prostatic hyperplasia. J Urol 994; 5: Epstein JI, Oesterling JE, Walsh PC. The volume and anatomical location of residual tumor in radical prostatectomy specimens removed for stage A prostate cancer. J Urol 988; 39: Authors L. Egevad, MD, PhD, Associate Professor, Consultant. T. Granfors, MD, PhD, Consultant. L. Karlberg, MD, PhD, Consultant. A. Bergh, MD, PhD, Professor, Consultant. P. Stattin, MD, PhD, Associate Professor, Consultant. Correspondence: L. Egevad, Department of Pathology and Cytology, Karolinska Hospital, SE-7 76 Stockholm, Sweden. lars.egevad@onkpat.ki.se Abbreviation: DSMS, disease-specific mean survival. # 22 BJU International 89,