Practice patterns for prostate cancer in nine central and northern Italy radiation oncology centers: a survey including 1759 patients treated during two decades (1980–1998)

Int. J. Radiation Oncology Biol. Phys., Vol. 52, No. 5, pp. 1310 –1319, 2002 Copyright © 2002 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016/02/$–see front matter

PII S0360-3016(01)02783-3

CLINICAL INVESTIGATION

Prostate

PRACTICE PATTERNS FOR PROSTATE CANCER IN NINE CENTRAL AND NORTHERN ITALY RADIATION ONCOLOGY CENTERS: A SURVEY INCLUDING 1759 PATIENTS TREATED DURING TWO DECADES (1980 –1998) STEFANO MARIA MAGRINI, M.D.,* FILIPPO BERTONI, M.D.,* VITTORIO VAVASSORI, M.D.,† SERGIO VILLA, M.D.,‡ EMANUELA CAGNA, M.D.,§ ERNESTO MARANZANO, M.D.,㛳 MAURIZIO PERTICI, M.D.,¶ RENATO PRADELLA, M.D.,# MASSIMO ALCIDE SPEDIACCI, M.D.,** ANDREA CHIAVACCI, M.D.,†† ENRICA AMBROSI, M.D.,* LORENZO LIVI, M.D.,¶ ALESSANDRO MAGLI, M.D.,† RITA BELLAVITA, M.D.,㛳 ALBERTO BOSSI, M.D.,§ AND GIANPAOLO BITI, M.D.¶ *Department of Radiation Oncology, Istituto del Radio, “O. Alberti” Brescia University, Brescia, Italy; †Department of Radiation Oncology, Varese Hospital, Varese, Italy; ‡Department of Radiation Oncology, National Cancer Institute, Milan, Italy; § Department of Radiation Oncology, Como Hospital, Como, Italy; 㛳Department of Radiation Oncology, Perugia University, Perugia, Italy; ¶Department of Radiation Oncology, Florence University, Florence, Italy; #Department of Radiation Oncology, Mantova Hospital, Mantova, Italy; **Department of Radiation Oncology, Arezzo Hospital, Arezzo, Italy; †† Department of Radiation Oncology, Pistoia Hospital, Pistoia, Italy Purpose: Prostate cancer patients in Italy are offered the choice of the full spectrum of possible treatment options for their disease, but the diffusion of the more recent technological refinements among the Radiation Oncology centers is not homogeneous and there is a need to establish a reference “historical” data source. This retrospective study describes the changing patterns in prostate cancer patient practice and the therapeutic results obtained in nine Radiation Oncology centers of Northern and Central Italy (five in Northern Italy and four in Central Italy). Methods and Materials: A total of 1759 prostate cancer patients, radically treated in the nine radiotherapy (RT) centers between 1980 and 1998, made up the study population. Data collected for each patient included clinical, pathologic, therapeutic features, and toxicity. The overall survival, disease-specific survival (DSS), and clinical relapse-free survival (RFS) were calculated for the whole series and for the subsets of patients defined by different clinical, pathologic, and therapeutic features, according to three accrual periods (A, 1980 –1990; B, 1991–1994; and C, 1995–1998). Univariate and multivariate analyses were performed to identify prognostic factors related to survival and late adverse effects (cystitis and proctitis) probability. Results: Patient accrual increased markedly during the 2 decades considered, and the percentage of cases with Stage C or D disease dropped from 49% (period A) to 43% (period B) to 37% (period C) (p < 0.0001, chi-square). The baseline prostate-specific antigen value was available for 10%, 76%, and 95% of the cases treated in the three different periods. The major changes in the therapeutic options were an increase in dose to the prostate (>66 Gy in 44%, 84%, and 93% of the patients treated in period A, B, and C, respectively); a reduction in treated volumes, including pelvic lymphatic drainage (56 –39% before 1995, 22% thereafter); and an increase in cases treated in association with hormonal therapy (50% before 1991, 80% thereafter). Lower energy (<10 MV) photon beams were progressively abandoned (12% before 1990 vs. 6 –7% thereafter), along with an increase in the use of blocks (60% in the last 4 years of the study vs. about 30 – 40% before 1995) and “conformal” RT (applied in 41% of cases treated after 1994). The actuarial RFS, DSS, and overall survival rate at 5 years was, respectively, 60% ⴞ 2%, 75% ⴞ 2%, 66% ⴞ 2% for period A; 74% ⴞ 2%, 90% ⴞ 1%, 83% ⴞ 2%, for period B; and 67% ⴞ 5%, 90% ⴞ 2%, 79% ⴞ 5% for period C. The actuarial overall survival, DSS, and RFS rate for the whole series of 1759 patients was 77% ⴞ 1%, 86% ⴞ 1%, and 68% ⴞ 1% at 5 years, respectively. Multivariate analysis showed that only American Urologic Association stage, grade, dose to the prostate, accrual period, association with hormonal treatment after (or both after and before) RT (only in terms of DSS and RFS), and baseline prostate-specific antigen value (only for RFS) retained prognostic significance in the final Cox model. Conclusion: The increase in the accrual of prostate cancer patients radically treated with RT has been accompanied by considerable changes in the clinical features at presentation, as well as in the staging and treatment procedures. Patients treated more recently had better survival results. An earlier stage and more favorable grade were linked with better overall, DSS, and RFS at multivariate analysis. Lower prostate-specific Reprint requests to: Stefano M. Magrini, M.D., Department of Radiation Oncology, Istituto del Radio, “O. Alberti” Piazzale Spedali Civili 1 CAP Brescia 25123 Italy. Tel: 0039030/3995271;

Fax: 030/396700; E-mail: [email protected] Received Jul 16, 2001, and in revised form Nov 15, 2001. Accepted for publication Nov 20, 2001. 1310

Practice patterns for prostate cancer in Italy

● S. M. MAGRINI et al.

1311

antigen baseline values were also related to better RFS. Better results were obtained with higher radiation doses, and the dose to tumor seemed the most important treatment-related prognostic factor. The toxicity (cystitis and proctitis, every Radiation Therapy Oncology Group grade) was substantially the same in the different accrual periods, but larger treated volumes and higher doses appeared to increase the incidence of late effects. © 2002 Elsevier Science Inc. Prostate cancer, Radiotherapy, Patterns of care.

INTRODUCTION

METHODS AND MATERIALS

External beam radiotherapy (RT) did not become a popular treatment modality for prostate cancer in Italy until the end of the 1980s. A strong surgical tradition and an inadequate number of radiation oncology centers, especially in some regions of the country, prevented an earlier diffusion of this type of therapy. However, since then, an increasing number of prostate cancer patients have undergone radical RT, and long-term data on many of these patients are now available. Many studies, directed to define the best clinical and technical choices for prostate cancer radical RT (volumes to be treated, usable doses, usefulness of the association with hormonal treatment), have been published (1–5). This phase of clinical research activity has not ended, because some large international randomized trials have only recently been closed (e.g., European Organization for Research and Treatment of Cancer 22961 and Radiation Therapy Oncology Group 9408 and 9413). Moreover, new (and revised) RT techniques are in different stages of development and are increasingly tested for prostate cancer (e.g., “conformal” techniques, intensity-modulated RT, brachytherapy) (6 – 8). Prostate cancer patients in Italy are offered the full spectrum of possible treatment options for their disease, but the diffusion of the more recent technologic refinements among the radiation oncology centers is heterogeneous; therefore, it is necessary to establish a reference “historical” data source. This may be useful in evaluating the changes in the referral pattern for RT, the staging procedures adopted (and their clinical effects), and the diffusion and impact of the new RT techniques (9 –13). The data obtained could also allow rough comparisons with other treatment modalities (i.e., surgery, hormonal treatment). The main clinical, pathologic, and therapeutic factors linked with prognosis can be identified retrospectively. Finally, the analysis of a sufficiently large, unbiased database may help to put in the right perspective the information given to the patients. To reach these objectives, nine radiation oncology centers (five in Northern Italy and four in Central Italy) decided to combine their archival data of radically treated prostate cancer patients. The survey covered the years 1980 through 1998 and included 1759 cases. We describe in detail the changing patterns in prostate cancer patient practice in the nine centers participating in the study, as well as the therapeutic results (disease control, survival, toxicity) obtained during three subsequent accrual periods and in the different clinical, pathologic, and therapeutic subgroups of the entire series.

Features of participating centers Three large academic centers, a National Cancer Institute, and five regional hospital centers participated in the study; they represent 50% of the university-based, 35% of the non– university research hospitals, and 38% of the regional hospital-based centers located in three of the Central and Northern Regions of Italy (Lombardy, Tuscany, and Umbria). None of the six private centers of the same geographic area participated in the study. New patient recruitment per year ranged from 500 to 2000 among the nine participating centers. The number of megavoltage treatment units per institution ranged from 1 to 5. At least one treatment planning system was available in all nine centers (range 1– 4, with most of them equipped with at least one fully threedimensional treatment planning system). The equipment available in the institutions varied during the two decades considered for the study, and the information given refers to the equipment available in 1998. Three centers began their full clinical activity in the 1990s; two had had archival data for prostate cancer patients treated radically only from the second half of the 1980s. This was taken into account when considering the variation in the overall accrual rate for radical prostate cancer RT in the different calendar years. The protocols adopted for the treatment of prostate cancer patients were developed independently within each center. (National guidelines for the treatment of prostate cancer were produced in 1999 by a panel of pathologists, radiologists, urologists, and radiation oncologists, under the aegis of the Italian National Research Council; however, these guidelines present only the viable therapeutic options by stage and are not intended to dictate a specific treatment protocol.) Each center provided the available data on the cases of prostate cancer treated radically. All the participating institutions were able to provide the data shown in Tables 1–3 for every case, as well as the follow-up information. To ensure that the data regarding each item of the database would have been homogeneously interpreted and collected in the different centers, “ad hoc” meetings were organized with the responsible researchers of each institution. The contribution of each institution to the whole series of 1759 cases ranged from 5% to 22%. The distribution of the cases according to the different characteristics listed in Tables 1–3 varied largely with time during the period considered; however, no significant intercenter differences in the clinical features of the cases observed were found within the same period.

1312

I. J. Radiation Oncology

● Biology ● Physics

Volume 52, Number 5, 2002

Table 1. Main clinical features of entire series of 1759 prostate cancer patients, according to three different accrual periods Accrual period Clinical feature Mean age (y) Stage A B C D Unknown Grade I II III Unknown Gleason sum 2–4 5–7 8–10 Unknown/not done Baseline PSA (ng/mL) 1–5 6–10 10–20 ⬎20 Unknown/not done Average baseline PSA (ng/mL)

1980–1990 (n ⫽ 373) 66

1991–1994 (n ⫽ 381)

1995–1998 (n ⫽ 1005)

68

69

All (entire series) (n ⫽ 1759) 68

38 (10) 153 (41) 111 (30) 70 (19) 1

22 (6) 191 (51) 136 (36) 27 (7) 5

73 (7) 602 (61) 277 (28) 28 (9) 25

133 (8) 946 (55) 524 (30) 125 (7) 31

108 (33) 160 (49) 57 (18) 48

96 (29) 153 (46) 80 (25) 52

226 (24) 525 (55) 203 (21) 51

430 (27) 838 (52) 340 (21) 151

10 (34) 15 (52) 4 (14) 334

50 (26) 104 (55) 36 (19) 191

207 (24) 541 (63) 114 (13) 143

267 (25) 660 (60) 154 (15) 678

8 (22) 7 (20) 6 (16) 15 (42) 337 30.6

53 (19) 61 (21) 67 (23) 107 (37) 93 39.3

116 (12) 229 (24) 277 (30) 326 (34) 57 26.8

177 (14) 297 (23) 350 (28) 448 (35) 487 29.8

Abbreviation: PSA ⫽ prostate-specific antigen. Percentages (in parentheses) refer to total number of patients of each group with data available (cases with “unknown” parameter not considered).

Table 2. Staging procedures adopted for entire series of 1759 prostate cancer patients according to three different accrual periods Accrual period Procedure Abdominal/pelvic CT No Yes Abdominal/pelvic MRI No Yes Transrectal US No Yes Transrectal MRI No Yes Bone scan No Yes Staging lymphadenectomy No Yes

1980–1990 (n ⫽ 373)

1991–1994 (n ⫽ 381)

1995–1998 (n ⫽ 1005)

All (entire series) (n ⫽ 1759)

224 (60) 149 (40)

45 (12) 336 (88)

260 (26) 745 (74)

529 (30) 1230 (70)

371 (99) 2 (1)

357 (94) 24 (6)

887 (88) 118 (12)

1615 (92) 144 (8)

148 (40) 225 (60)

40 (11) 341 (90)

122 (12) 883 (88)

310 (18) 1449 (82)

373 (100) —

373 (98) 8 (2)

934 (93) 71 (7)

1680 (95) 79 (5)

171 (46) 202 (54)

52 (14) 329 (86)

86 (9) 919 (91)

309 (18) 1450 (82)

242 (65) 131 (35)

348 (91) 33 (9)

928 (92) 77 (8)

1518 (86) 241 (14)

Abbreviation: US ⫽ ultrasonography. Values in parentheses expressed as percentages of entire number of patients of each group.

Practice patterns for prostate cancer in Italy

● S. M. MAGRINI et al.

1313

Table 3. Main therapeutic features of entire series of 1759 prostate cancer patients according to three different accrual periods Accrual period Therapeutic feature Dose to prostate (Gy) ⬍60 60–65 66–69 70 ⬎70 Volume treated Prostate ⫾ SV Prostate ⫹ pelvis RT technique AP–PA ⬎2 fields Arc Blocks No Yes Beam energy (MV) 60 Co ⬍10 10–18 ⬎18 Conformal techniques Yes No Association with HT RT only HT before RT* HT after RT HT before ⫹ after RT

1980–1990 (n ⫽ 373)

1991–1994 (n ⫽ 381)

1995–1998 (n ⫽ 1005)

All (entire series) (n ⫽ 1759)

20 (5) 110 (29) 103 (28) 77 (21) 63 (17)

5 (1) 58 (15) 117 (31) 156 (42) 45 (11)

8 (1) 64 (6) 243 (25) 457 (45) 233 (23)

33 (2) 232 (14) 463 (26) 690 (39) 341 (19)

165 (44) 298 (56)

232 (61) 149 (39)

782 (78) 223 (22)

1179 (67) 580 (33)

21 (6) 308 (83) 44 (11)

5 (1) 298 (78) 78 (21)

2 (—) 928 (92) 75 (8)

28 (1) 1534 (87) 197 (12)

266 (71) 107 (29)

241 (64) 140 (36)

412 (40) 593 (60)

919 (52) 840 (48)

29 (8) 14 (4) 147 (39) 183 (49)

2 (1) 24 (6) 141 (37) 214 (56)

— 62 (6) 421 (42) 522 (52)

31 (2) 100 (6) 709 (40) 919 (52)

— 373 (100)

34 (9) 347 (91)

410 (41) 595 (59)

444 (25) 1315 (75)

187 (50) 42 (11) 64 (17) 80 (22)

72 (19) 75 (19) 72 (19) 162 (43)

199 (20) 199 (20) 181 (18) 426 (42)

458 (26) 316 (18) 317 (18) 668 (38)

Abbreviations: SV ⫽ seminal vesicles; RT ⫽ radiotherapy; HT ⫽ hormonal therapy. * Or before and during RT. Values expressed as percentages (in parentheses) of entire number of patients in each group.

Accrual periods Three periods were identified during the interval selected for analysis (1980 –1998): 1. The “pre–prostate-specific antigen (PSA) era,” between 1980 and 1990: a limited accrual period, with only a very few cases staged and followed with the help of the PSA test 2. The 1991–1994 period, when the Italian RT centers experienced a rapid rise in prostate cancer patient accrual, new techniques began to be used, and hormonal manipulation became increasingly popular, to the point that most patients were already receiving hormonal treatment when referred for RT 3. The 1995–1998 period, when radical RT reached its largest diffusion and the number of patients increased because of the large-scale use of the PSA test in the general population. These 4 years also witnessed the adoption of conformal RT techniques by most of the centers

Features of the series and follow-up methods The cumulative series included 2008 cases treated with RT from 1978 through 1999, but the patients treated before 1980 were excluded from analysis (even though included in the cumulative database). Similarly, patients accrued during 1999 were excluded, so that the minimum follow-up at the time of analysis amounted to 18 months. This left 1759 cases available for analysis. Tables 1–3 show the main clinical, pathologic, and therapeutic features of the whole series of 1759 cases. Updated follow-up information was obtained from patient records at every single center; for all those not followed until death or until the closing date of the study (December 31, 2000), supplementary information was gained by inviting patients to present for clinical controls. The minimal follow-up was at least 18 months; the effective follow-up was, on average, 48 months. Less than 5% of the cases were lost to follow-up. Data on crude and cause-specific survival only were obtained for the cases lost to follow-up through the vital statistics offices of their cities, the general practitioner, or relatives. Moreover, follow-up information rel-

1314

I. J. Radiation Oncology

● Biology ● Physics

Volume 52, Number 5, 2002

Fig. 1. Yearly accrual of patients treated radically for prostate cancer in the centers with archives active since 1980 (number and trend, total n ⫽ 1057).

ative to a single institution (265 cases included in the cumulative database) were considered only for crude and causespecific survival, because most such patients underwent active clinical follow-up outside that center. The follow-up procedures varied slightly among the institutions considered; however, some general statements should be made because of the long study period. All the patients were followed for an interval of 3– 6 months between the follow-up visits during the first 2–3 years; the visits were every 6 months or annually thereafter. The follow-up included at each visit a physical examination (general and local). This prompted, when needed and according to the clinical suspicion of persisting/relapsing disease, the use of the same specific diagnostic tests performed during the baseline workup (e.g., ultrasonography, CT, bone scan, and MRI). Some diagnostic examinations were introduced into clinical practice and used only years after the starting date of the study. For example, MRI and MRI with an endorectal coil were available in some of the participating centers after 1990; serial PSA determinations were also routinely available only after 1990. After 1990, most patients were followed with the help of PSA serial determinations. An abnormal PSA value, per se, did not imply the use of other diagnostic procedures. However, the prevailing attitude at the participating centers was to perform other diagnostic tests (apart from the physical examination) when a rising PSA value was recorded. All the cases followed with PSA were reviewed, biochemical relapse was classified according to the American Society for Therapeutic Radiology and Oncology criteria (14), and scored as such only in the absence of any clinical evidence of disease. Study end points and statistical analysis The differences in the distribution of the different clinical, pathologic, and therapeutic features of the series (Tables 1–3), according to the accrual period (1980 –1990, 1991–1994, and 1995–1998) were registered, and their significance evaluated with the chi-square test.

The actuarial overall (uncorrected) survival (OS), disease-specific survival (DSS), corrected for causes of death different from prostate cancer, and clinical relapse-free survival (RFS) were calculated with the Kaplan–Meier method for the whole series and for the subsets of patients defined by the different clinical, pathologic, and therapeutic features. Only clinical relapses were then taken into account to calculate the actuarial RFS. In other words, if any evidence of disease was found at physical examination or according to the other diagnostic tools, at the time of the PSA value rise or thereafter, the relapse was scored as a clinical relapse, and its date registered as the time of the first clinical evidence of disease. This is particularly relevant in interpreting our data on RFS, because the reported values should be interpreted as clinical RFS values and because the more recent cases underwent more intensive follow-up (and staging) procedures. The cumulative probability of late adverse effects (cystitis and proctitis) from RT was calculated for the whole series and its subsets. To quantitate toxicity, the Radiation Therapy Oncology Group–European Organization for Research and Treatment of Cancer (RTOG-EORTC) scoring system was used (14). Comparisons between the actuarial survival values and between toxicity probabilities were made with the log–rank test (univariate analysis) and with the Cox method (multivariate analysis) when appropriate and feasible. Actuarial survival curves, cumulative toxicity probability, and the statistical comparisons between the results obtained in the different subsets of the series were performed with the Statistical Package for the Social Sciences for Windows (SPSS Inc., 1989 –1994). RESULTS Overall, the number of recruited patients (considering only the centers with prostate cancer archives active since 1980, a total of 1057 patients) increased markedly during the period considered (Fig. 1).

Practice patterns for prostate cancer in Italy

The changes in the single main clinical, pathologic, diagnostic, and therapeutic features of the series during the accrual periods 1980 –1990, 1991–1994, and 1995–1998 are depicted in Tables 1–3. Patients treated more recently had more favorable clinical characteristics (Table 1). For example, the percentage of cases with Stage C or D disease dropped from 49% in 1980 –1990 to 43% in the subsequent 4 years and to 37% in 1995–1998 (p ⬍ 0.0001, chi-square). The average baseline PSA value was 39 ng/mL for patients treated during 1991– 1994 and 27 ng/mL for those treated during 1995–1998. The distribution of cases by PSA level was also significantly different in the three accrual periods (p ⫽ 0.02, chi-square); patients with a PSA value ⬎20 ng/mL were less frequent in the last accrual period. The staging procedures also evolved during the long accrual period (Table 2). Although the proportion of patients who underwent bone scanning remained almost the same throughout the study (about 90%), in the past decade, the use of abdominal and pelvic CT for staging was very common (87% and 75% in 1991–1994 and 1995–1998, respectively) compared with 1980 through 1990 (29%). The slight decrease in the use of CT in the last 4 years of the study was paralleled by an increase in MRI of the same anatomic regions (1980 –1990, 0%; 1991–1994, 5%; and 1995–1998, 11%). Accordingly, an increasing number of patients underwent MRI with an endorectal coil in the last few years of the study (3% in 1991–1994 and 11% thereafter). The baseline PSA value was available for 10%, 76%, and 95% of the patients treated in the three subsequent (already defined) accrual periods. Staging lymphadenectomy was progressively abandoned (having been used in 35%, 9%, and 8% of the cases recruited in the three subsequent accrual periods). All the reported differences were statistically significant (p ⬍ 0.0001, chi-square). The therapeutic options underwent three major changes (Table 3). First, the dose to the prostate increased (⬎66 Gy in 44%, 84%, and 93% of the patients treated before 1991, 1991–1994, and 1995–1998, respectively). Second, the number of patients with treated volumes, including pelvic lymphatic drainage reduced significantly (from 56% down to 39% before 1995, and 22% thereafter). Finally, the number of patients receiving different types of hormonal therapy in association with RT increased (50% before 1991 and 80% thereafter). Remarkable changes in RT techniques were also registered. For the patients treated more recently, lower energy (⬍10 MV) photon beams were progressively abandoned (12% before 1990 vs. 6 –7% thereafter), corresponding with an increase in the use of blocks (60% in the last 4 years of the study vs. about 30 – 40% before 1995). The parallel opposed AP–PA field technique, already used in only 6% of those treated before 1990, virtually disappeared thereafter. Finally, conformal RT techniques only recently gained popularity (used in 41% of patients treated after 1994 vs. 0 –9% in the previous accrual periods). All these differences were statistically significant (p ⬍ 0.0001, chi-square). The in-

● S. M. MAGRINI et al.

1315

crease in the use of conformal techniques in the last 4 years of the study was not uniform among the nine centers considered; the percentage of the patients who underwent conformal treatments in the whole series equaled 41%, but the corresponding values for each of the nine centers showed a large variability, ranging from 0% to 100% (specifically, 0%, 0%, 0%, 33.5%, 47.7%, 55%, 59.5%, 60.7%, and 100%). The survival results of the patients most recently treated also seemed better. The actuarial RFS, DSS, and OS (uncorrected) at 5 years was, respectively, 60% ⫾ 2%, 75% ⫾ 2%, and 66% ⫾ 2% for the 1980 –1990 period; 74% ⫾ 2%, 90% ⫾ 1%, and 83% ⫾ 2%, for the 1991–1994 period; and 67% ⫾ 5%, 90% ⫾ 2%, and 79% ⫾ 5% for the last study period (1995–1998). The actuarial OS, DSS, and RFS for the whole series of 1759 patients was, respectively, 77% ⫾ 1%, 86% ⫾ 1%, and 68% ⫾ 1% at 5 years and 50% ⫾ 1%, 71% ⫾ 1%, and 50% ⫾ 2% at 10 years. A subgroup analysis was also performed for the various subsets of patients identified by the different features listed in Tables 1–3. Factors significantly linked with better survival results at univariate analysis were earlier American Urologic Association (AUA) stage, grade ⬍ 3, higher dose to the prostate, lower baseline PSA value (only in terms of RFS); and association of adjuvant hormonal treatment with RT (only in patients with Stage C and D disease). Other factors were significantly linked to a better prognosis at univariate analysis (e.g., absence of pelvic adenopathy either at pelvic CT or histologically; lower Gleason score, for the patients for whom this parameter was available; and some technical parameters, including conformal RT techniques); however, most of these factors did not constitute independent variables (e.g., Gleason score linked with grade; pelvic adenopathy with AUS stage). Table 4 shows the survival results obtained in the main subgroups of the series. The results of multivariate analysis for RFS, DSS, and OS are reported in Tables 5–7, respectively. The parameters entered in the initial model for multivariate analysis of survival and toxicity were those identifying significant differences between subgroups at univariate analysis and, in addition, those judged clinically to be relevant: age, AUA stage, grade, baseline PSA (only for RFS), dose to prostate, treated volume, other RT technique parameters, association with hormonal treatment, and accrual period. Multivariate analysis of survival revealed that only AUA stage, grade, dose to prostate, accrual period, association with hormonal treatment after (or both after and before) RT (for DSS and RFS only), and baseline PSA value (only for RFS) retained prognostic significance in the final Cox model. The 5-year cumulative incidence of late toxicity effects (cystitis and proctitis, every RTOG grade) was substantially the same in the first study period (1980 –1990) and the subsequent accrual periods (1991–1994 and 1995–1998), equaling, respectively, 38% ⫾ 2%, 31% ⫾ 3%, and 37% ⫾ 2%. The same was true for severe late effects (Grade 3– 4

1316

I. J. Radiation Oncology

● Biology ● Physics

Volume 52, Number 5, 2002

Table 4. Five-year overall, disease-specific, and relapse-free survival in different clinical and therapeutic subgroups of the entire series Subgroup Dose to prostate (Gy) ⬍60 60–65 66–69 70 ⬎70 P AUA stage A B C D1 P Grade 1 2 3 x P Baseline PSA (ng/mL) ⬍5 6–10 11–20 ⬎20 Not done/unknown P Association with HT RT only HT before RT* HT after RT HT before ⫹ after RT P Accrual period 1980–1990 1991–1994 1995–1998 P

OS (%)

DSS (%)

RFS (%)

38 ⫾ 10 63 ⫾ 3 85 ⫾ 2 84 ⫾ 3 71 ⫾ 3 0.003

50 ⫾ 11 73 ⫾ 3 90 ⫾ 1 90 ⫾ 1 82 ⫾ 3 0.0001

28 ⫾ 9 57 ⫾ 3 72 ⫾ 2 70 ⫾ 3 73 ⫾ 3 0.0001

89 ⫾ 3 82 ⫾ 1 70 ⫾ 2 64 ⫾ 4 0.0001

95 ⫾ 2 92 ⫾ 1 75 ⫾ 2 69 ⫾ 4 0.0001

89 ⫾ 4 74 ⫾ 2 58 ⫾ 3 52 ⫾ 4 0.0001

84 ⫾ 2 79 ⫾ 2 67 ⫾ 3 69 ⫾ 4 0.0001

91 ⫾ 1 86 ⫾ 1 76 ⫾ 3 75 ⫾ 4 0.0001

77 ⫾ 2 67 ⫾ 2 58 ⫾ 4 61 ⫾ 4 0.0001

84 ⫾ 4 84 ⫾ 4 87 ⫾ 3 77 ⫾ 3 70 ⫾ 2 NS

88 ⫾ 4 94 ⫾ 2 92 ⫾ 2 86 ⫾ 2 78 ⫾ 2 NS

77 ⫾ 5 79 ⫾ 4 72 ⫾ 5 61 ⫾ 4 65 ⫾ 2 0.0004

74 ⫾ 2 75 ⫾ 3 80 ⫾ 3 79 ⫾ 2 NS

84 ⫾ 2 80 ⫾ 3 89 ⫾ 2 85 ⫾ 2 NS

67 ⫾ 2 61 ⫾ 4 80 ⫾ 3 68 ⫾ 2 NS

66 ⫾ 2 83 ⫾ 2 79 ⫾ 5 0.0001

75 ⫾ 2 90 ⫾ 1 90 ⫾ 2 0.0001

60 ⫾ 2 74 ⫾ 2 67 ⫾ 5 0.001

* Or before and during RT. Abbreviations: OS ⫽ overall survival; DSS ⫽ disease-specific survival; RFS ⫽ relapsefree survival; PSA ⫽ prostate-specific antigen; HT ⫽ hormonal therapy; RT ⫽ radiotherapy.

RTOG), with their cumulative probability in the same accrual periods 4.2%, 2.3%, and 5.1%. Among the different clinical and therapeutic features listed in Tables 2 and 3, only a larger treated volume and higher dose appeared to be significantly related to an increased incidence of late effects. The cumulative probability of proctitis and cystitis (all grades) was 20 –31% for doses ⬍70 Gy vs. 40 –53% for higher doses; similarly, patients not treated on pelvic drainage areas showed a reduced incidence of late effects (30% vs. 48%). The same observations held true when limiting the analysis to the probability of Grade 3– 4 late effects (Table 8). Multivariate analysis was possible only for the cumulative probability of late effects for every grade (because of the relatively small number of events scored as Grade 3– 4 RTOG). Higher doses and larger treated volumes were the

only factors that were significantly linked (p ⬍ 0.0001) with an increased probability of proctitis and cystitis. DISCUSSION The data presented show that an increasing number of prostate cancer patients have been treated radically with RT in the past two decades among the participating centers (representing a non-negligible fraction of the Central and Northern Italy radiation oncology departments). The increase in the accrual of these patients was accompanied by considerable changes in the clinical features at presentation, as well as in the staging and treatment procedures. The results of this study allowed certain key points to be defined. First, in a country with a strong surgical tradition, prostate cancer patients are now offered a full range of

Practice patterns for prostate cancer in Italy

Table 5. Multivariate analysis—final Cox model—factors significantly linked with OS Factor Dose to prostate (Gy) ⬍60 60–65 66–69 70 ⬎70 AUA stage A B C D1 Grade 1 2 3 Accrual period 1980–1990 1991–1994 1995–1998

RR

p

Overall p

1 0.6 0.4 0.3 0.5

⬍0.0001

⬍0.0001

1 1.5 2.1 2.9

⬍0.0001

1 1.3 2

0.001

1 0.5 0.5

⬍0.0001

Abbreviations: OS ⫽ overall survival; RR ⫽ relative risk of failure; AUA ⫽ American Urologic Association.

options for the treatment of their disease. The survival results in the present series, obtained in a large cohort of unbiased, consecutive patients, treated in institutions with Table 6. Multivariate analysis—final Cox model—factors significantly linked with DSS Factor Dose to the prostate (Gy) ⬍60 60–65 66–69 70 ⬎70 AUA stage A B C D1 Grade 1 2 3 Association with HT RT only HT before RT* HT after RT HT before ⫹ after RT Accrual period 1980–1990 1991–1994 1995–1998

RR

p

1 0.6 0.3 0.3 0.4

0.0003

1 2.1 6.5 7

⬍0.0001

1 1.4 2

0.01

1 1.5 0.8 0.8

0.03

1 0.4 0.4

⬍0.0001

Overall p

⬍0.0001

● S. M. MAGRINI et al.

1317

Table 7. Multivariate analysis—final Cox model—factors significantly linked with RFS Factor Dose to prostate (Gy) ⬍60 60–65 66–69 70 ⬎70 AUA stage A B C D1 Grade 1 2 3 Baseline PSA value (ng/mL) ⬍5 6–10 11–20 ⬎20 Association with HT RT only HT before RT* HT after RT HT before ⫹ after RT Accrual period 1980–1990 1991–1994 1995–1998

RR

p

Overall p

1 0.5 0.3 0.3 0.3

⬍0.0001

⬍0.0001

1 2.8 4.8 5.4

⬍0.0001

1 1.2 1.8

0.004

1 1 1.4 1.6

0.03

1 1.1 0.6 0.8

0.01

1 0.6 0.7

0.007

Abbreviations: RFS ⫽ relapse-free survival; other abbreviations as in Table 5. * Or before and during RT.

different workloads, both academic and hospital-based, with recruitment areas different under many aspects (e.g., “density” of population and radiation oncology facilities, availability of urologic surgical skills), are in line with those of the literature and therefore also with the surgical results, stage by clinical stage (9, 10, 15–21). Not surprisingly, therefore, accrual in Italian radiation oncology centers grew

Table 8. Cumulative 5-year probability of late effects (proctitis and cystitis) according to dose to prostate and volume treated (prostate vs. prostate and pelvis) 5-y Probability of late effects

* Or before and during RT. Abbreviations: DSS ⫽ disease-specific survival; other abbreviations as in Table 5.

Dose to prostate (Gy) ⬍60 60–65 66–69 70 ⬎70 Volume treated Prostate only Prostate and pelvis

Every grade (%)

p

Grade 3–4 (%) p

20 29 31 40 53

⬍0.0001

— — 3 4.5 6

0.01

30 48

⬍0.0001

2.9 4.4

NS

1318

I. J. Radiation Oncology

● Biology ● Physics

rapidly, even if partly as a consequence of the concurrent increase in disease incidence. Second, some determinants of survival and toxicity already described in the literature were confirmed in the present series (22–26). Some disease features are linked with prognosis. An earlier stage and a more favorable grade are linked with better OS, DSS, and RFS at multivariate analysis. Lower PSA baseline values also correlated with better RFS rates. Among the treatment-related factors of possible prognostic significance, the dose to tumor seemed the most important, with better results obtained with higher radiation doses. That the cumulative probability of toxicity also increased with the radiation dose and with the use of larger treated volumes underscores the need for adequate RT techniques and seems to favor the choice of avoiding pelvic irradiation. Thus, it is not surprising that the recourse to conformal methods of radiation delivery increased rapidly in the last study period (1994 –1998), compared with the previous ones. The association with hormonal therapy seems to be linked with an increase in the DSS and RFS, mainly for advanced cases and for those treated after RT. Third, patients treated more recently had better survival results. This occurred for different reasons, including the more accurate selection of patients for radical RT, of whom the fraction of those with less advanced disease has been larger in more recent years. Some basic treatment parameters are also relevant; a larger fraction of the more recently treated patients were treated with higher doses. However, the accrual period remained in the final Cox model of multivariate analysis as an independent predictor of survival, together with stage and histologic grade and dose to the prostate. This may be explained by the effect of many other factors that changed to a large extent during the long study period. This change almost certainly contributed to the increasing survival rates registered. These factors, when considered individually, did not exert statistically significant effects; however, taken together, they may well affect survival. For example, the staging procedures became progressively more accurate (through the more frequent use of CT and MRI, with or without an endorectal coil), thus allowing for more adequate “tailoring” of the treatment to the individual patient, particularly for those diagnosed with more advanced T-stage disease. The more frequent association of radical RT with hormonal treatment might also have played a role. Finally, in the more recently treated patients, the RT techniques have been substantially modified and refined.

Volume 52, Number 5, 2002

The main issue to emerge from this study is related to the changing “patterns of practice” for prostate cancer RT. The retrospective nature of the analysis obviously leaves many questions unanswered and poses new ones for the future. Some of them appear most relevant to daily radiation oncology practice in our country, taking into account the workload represented by radical prostate cancer treatment, which constitute (1998 figures) between 5% and 10% of the new patient workload of the nine centers taking part in the study. The most relevant question relates to the cost-effectiveness of these treatments. It has been shown that the results seemed better in more recent years. However, in an era of reduced health system expenditure and competition for the allocation of resources, it is necessary to quantify the impact of the different components of RT “practice patterns” on survival. This is true for all the different steps of the treatment workup and delivery. For example, the very high proportion of cases staged with CT or MRI of the abdomen and pelvis in this series (1994 –1998 data) may not be justified in the context of “routine” good clinical practice. For many patients (those with Stage A or B; with Gleason sum ⬍6; with baseline PSA value ⬍10 ng/mL, more frequently observed in the more recent years of the survey and representing about 8% of the cases of the present series treated after 1994), these examinations do not seem to be always worthwhile, when taking into account the available evidence on the risk of lymph node metastases for these subsets of patients. The same is true, for the same patients, for the use of bone scans to exclude bone metastases (27). In addition, delivering radical RT with highly sophisticated techniques involves a non-negligible expenditure of technological and human resources, which may not be always be justified by the clinical results compared with those of “standard” RT techniques and by the current Italian reimbursement policy for such treatments. However, a definitive analysis on the cost/benefit ratio for these techniques on a large, unbiased RT practice case series has not been conducted in our country. Such an analysis might help to standardize the indications for conformal RT, currently not uniformly applied (e.g., among the nine centers participating in this study). Finally, for more advanced cases, the best combination of RT and hormonal treatment remains to be established. Again, the present study evidenced a large-scale recourse to adjuvant and neoadjuvant hormonal therapy, mainly with total androgen blockade. The attending costs are very high, and a clear demonstration of a beneficial effect of such an unselective strategy is currently unavailable.

REFERENCES 1. Asbell SO, Krall JM, Pilepich MV, et al. Elective pelvic irradiation in stage A2, B carcinoma of the prostate, analysis of RTOG 77-06. Int J Radiat Oncol Biol Phys 1988;15:1307– 1316. 2. Bolla M, Gonzalez D, Warde P, et al. Improved survival in

patients with locally advanced prostate cancer treated with radiotherapy and goserelin. N Engl J Med 1997;337:295–300. 3. Hanks GE, Lu J, Marthai M, et al. RTOG protocol 92-02: A phase III trial of the use of long term androgen suppression following neoadjuvant hormonal cytoreduction and radiother-

Practice patterns for prostate cancer in Italy

4.

5.

6.

7. 8.

9. 10. 11. 12. 13. 14.

15.

apy in locally advanced carcinoma of the prostate. Proc ASCO 2000;19:327. Laverdiere J, Gomez JL, Cusan L, et al. Beneficial effect of combination hormonal therapy administrated prior and following external beam irradiation therapy in localized prostatic cancer. Int J Radiat Oncol Biol Phys 1997;37:247– 252. Zelefsky MJ, Leibel SA, Gaudin PB, et al. Dose escalation with three dimensional conformal radiation therapy affects the outcome in prostate cancer. Int J Radiat Oncol Biol Phys 1998;41:491–500. Blasko JC, Wallner K, Grimm PD, et al. Prostate specific antigen based disease control following ultrasound guided 125 iodine implantation for stage T1/T2 prostatic carcinoma. J Urol 1995;154:1096 –1099. Zelefsky MI, Fuks Z, Happersett L, et al. Clinical experience with intensity modulated radiation therapy (IMRT) in prostate cancer. Radiother Oncol 2000;55:241–249. Dearnaley DP, Seddon B, Bidmead M, et al. Target volume definition in conformal radiotherapy for prostate cancer: Quality assurance in the MRC-RT-01 trial. Radiother Oncol 2000; 56:73– 83. Aristizabal SA, Steinbronn D, Heusinkveld RS. External beam radiotherapy in cancer of the prostate: The University of Arizona experience. Radiother Oncol 1984;1:309 –315. Debuyrne F, Beerlage H. The place of radical prostatectomy in the treatment of early localized prostate cancer. Radiother Oncol 2000;57:259 –262. Huland H. Radical prostatectomy: Options and issues. Eur Urol 2001;39(Suppl. 1):3–9. Morphis JG II, Hornback NB. Carcinoma of the prostate: Is radiation therapy an alternative of radical prostatectomy? Indiana Med 1986;79:1939 –1943. Walsh PC, Lepor H. The role of radical prostatectomy in the management of prostatic cancer. Cancer 1987;60(Suppl. 3): 526 –537. American Society for Therapeutic Radiology and Oncology Consensus Panel. Consensus statement: Guidelines for PSA following radiation therapy. Int J Radiat Oncol Biol Phys 1994;17:1155. Bagshaw MA, Kaplan ID, Cox RC. Radiation therapy for localized disease. Cancer 1993;1971:939 –952.

● S. M. MAGRINI et al.

1319

16. Hanks GE, Lee WE, Schultheiss TE. Clinical and biochemical evidence of control of prostate cancer at 5 years after external beam radiation. J Urol 1995;154:456 – 459. 17. Perez AA, Pilepich MV, Zivnuska F. Tumor control in definitive irradiation of localized carcinoma of the prostate. Int J Radiat Oncol Biol Phys 1986;12:523–531. 18. Perez CA, Hanks GE, Leibel SA, et al. Localized carcinoma of the prostate (stages T1b T1c T2 andT3): Review of management with external beam radiation therapy. Cancer 1993; 72:3156 –3173. 19. Ray GR, Cassady JR, Bagshaw MA. Definitive radiation therapy of carcinoma of prostate: A report of 15 years experience. Radiology 1973;106:407– 418. 20. Roach M III, Lu J, Pilepich MV, et al. Long-term survival after radiotherapy: A Radiation Therapy Oncology Group prostate cancer trial. J Urol 1999;161:864 – 868. 21. Shipley WU, Thames HD, Sandler HM, et al. Radiation therapy for clinically localized prostate cancer: A multi-institutional pooled analysis. JAMA 1999;281:1598 –1604. 22. Bagshaw MA, Cox RC, Ray GR. Status of radiation treatment of prostate cancer at Stanford University. Natl Cancer Inst Monogr 1988;7:47– 60. 23. Dearnaley DP, Khoo VS, Normann AR, et al. Comparison of radiation side-effects of conformal and conventional radiotherapy in prostatic cancer: A randomised trial. Lancet 1999; 353:267–272. 24. Lawton CA, Wong M, Pilepich MV, et al. Long term treatment sequelae following external beam irradiation for adenocarcinoma of the prostate: Analysis of RTOG studies 7506 and 7706. Int J Radiat Oncol Biol Phys 1991;21:935–939. 25. Michalsky JM, Pundy JA, Winter K, et al. Preliminary report of toxicity following 3D radiation therapy for prostate cancer on 3 BOG-RTOG 9406. Int J Radiat Oncol Biol Phys 2000; 46:391– 402. 26. Soffen EM, Hanks GE, Hunt MA, et al. Conformal static field radiation therapy treatment of early prostatic cancer versus non conformal techniques: A reduction in acute morbidity. Int J Radiat Oncol Biol Phys 1992;24:485– 488. 27. Carroll PR, Lee KL, Fuks ZY, et al. Cancer of the prostate. In: De Vita VT, Hellman S, Rosenberg SA, editors. Cancer: Principles and practice of oncology. 6th ed. Lippincott; Philadelphia: 2001. p. 1418 –1479.