Recurrence-free survival rates after external ... - Wiley Online Library

1283

Recurrence-Free Survival Rates after External-Beam Radiotherapy for Patients with Clinical T1–T3 Prostate Carcinoma in the Prostate-Specific Antigen Era What Should We Expect?

Deepak Khuntia, M.D.1 Chandana A. Reddy, M.S.1 Arul Mahadevan, M.D.1 Eric A. Klein, M.D.2 Patrick A. Kupelian, M.D.3 1

Department of Radiation Oncology, Cleveland Clinic Foundation, Cleveland, Ohio.

2

Department of Urology, Cleveland Clinic Foundation, Cleveland, Ohio.

3

Department of Radiation Oncology, M. D. Anderson Cancer Center Orlando, Orlando, Florida.

BACKGROUND. The objective of the current study was to report biochemical recurrence–free survival (bRFS) rates among men with T1–T3 prostate carcinoma who were treated with external-beam radiotherapy (RT) at the Cleveland Clinic Foundation (Cleveland, OH).

METHODS. In total, 1352 patients were identified between 1987 and 2000 with a minimum follow-up of 1 year (median follow-up, 55 months; range, 12–189 months). The median radiation dose was 74.0 grays (Gy) (range, 63.0 – 83.0 Gy). The median radiation doses for patients who received ⬍ 68.0 Gy (n ⫽ 201), 68.0 –72.0 Gy (n ⫽ 373), and ⱖ 72.0 Gy (n ⫽ 778) were 66.6 Gy, 70.0 Gy, and 78.0 Gy, respectively. The RT techniques used were standard RT in 41% of patients, 3-dimensional conformal RT in 34% of patients, and intensity-modulated RT in 25% of patients. Androgen-deprivation (AD) therapy lasting ⱕ 6 months was administered to 34% of patients. RESULTS. The 5-year and 7-year bRFS rates were 63% and 59%, respectively. On multivariate analysis, T classification (P ⬍ 0.001), pretreatment prostate-specific antigen level (P ⬍ 0.001), biopsy Gleason score (P ⫽ 0.001), radiation dose (P ⬍ 0.001), and year of therapy (P ⬍ 0.001) were independent predictors of biochemical failure. Age, race, AD therapy, and RT technique did not predict for biochemical failure. For patients with low-risk tumors, the 5-year bRFS rates for those who received RT doses of ⱕ 68.0 Gy, 68.0 –72.0 Gy, and ⱖ 72.0 Gy were 52%, 82%, and 93%, respectively (P ⬍ 0.001); for patients with intermediate-risk tumors, the respective 5-year bRFS rates were 27%, 51%, and 83% (P ⬍ 0.001); and for patients with high-risk tumors, the respective 5-year bRFS rates were 21%, 29%, and 71%, respectively (P ⬍ 0.001). CONCLUSIONS. The most significant therapeutic factor affecting bRFS rates after RT was radiation dose, rather than AD therapy use or radiation technique. Cancer 2004;100:1283–92. © 2004 American Cancer Society. KEYWORDS: conformal radiotherapy, intensity-modulated radiotherapy, prostate carcinoma, prostate-specific antigen, risk group.

Address for reprints: Patrick A. Kupelian, M.D., Department of Radiation Oncology, M. D. Anderson Cancer Center Orlando, 1400 South Orange Avenue, Orlando, FL 32806; Fax: (407) 649-6895; E-mail: [email protected] Received October 3, 2003; revision received December 22, 2003; accepted December 30, 2003. © 2004 American Cancer Society DOI 10.1002/cncr.20093

T

he use of external-beam radiotherapy (RT) in the treatment of patients with localized prostate carcinoma has evolved significantly in the prostate-specific antigen (PSA) era. T classification migration and higher radiation doses, among other factors, have improved outcomes dramatically. The outcomes of patients with localized prostate carcinoma treated at the Cleveland Clinic Foundation (Cleveland, OH) in the PSA era have been reported previously.1 The current study is an update of the Cleveland Clinic Foundation

1284

CANCER March 15, 2004 / Volume 100 / Number 6

experience, with emphasis on long-term outcomes in different prognostic subgroups by radiation dose, radiation technique, and use of androgen-deprivation (AD) therapy.

MATERIALS AND METHODS Between 1987 and 2000, 1352 patients with localized prostate adenocarcinoma were treated with RT at the Cleveland Clinic Foundation. All patients had the following characteristics: a pretreatment PSA level (iPSA), a biopsy Gleason score (bGS), tumor(s) classified as clinical T1–T3, no clinical lymph node involvement, no hormone therapy for ⬎ 6 months, and a minimum follow-up interval of 12 months. The initial evaluation for all 1352 patients included determination of the 1992 American Joint Committee on Cancer clinical T classification. Further work-up with transrectal ultrasound, bone scan, chest X-ray, and computerized tomography (CT) scans of the abdomen and pelvis were obtained according to the individual physician preference. None of these evaluations revealed definite bony or lymphatic metastases. On the bases of iPSA, bGS, and clinical T classification, risk groups were defined as low (T1–T2, bGS ⱕ 6, and iPSA ⱕ 10.0 ng/mL), intermediate (T1–T2, and either bGS ⫽ 7 or iPSA 10.0 –20.0 ng/mL), and high (any T3, any bGS ⱖ 8, any iPSA ⬎ 20.0 ng/mL, or both bGS ⫽ 7 and iPSA 10.0 –20.0 ng/mL). Megavoltage equipment was used to deliver radiation. A fraction size of 1.8 Grays (Gy) was used in 412 patients, and a fraction size of 2.0 Gy was used in 609 patients. The remaining 331 patients received hypofractionated RT up to a total dose of 70.0 Gy at 2.5 Gy per fraction; the equivalent dose at 2.0 Gy per fraction was considered 83.0 Gy. With a total dose of 83.0 Gy for patients who received hypofractionation, the median total prescription dose for all 1352 patients was 74.0 Gy.2 The average total dose for all patients was 74.7 Gy (range, 60.0 – 83.0 Gy). The distribution of dose groups was as follows: ⬍ 68.0 Gy, 201 patients (15%); 68.0 –72.0 Gy, 373 patients (28%); and ⱖ 78.0 Gy, 778 patients (57%). The median dose for patients who received ⬍ 68.0 Gy was 66.6 Gy, compared with a median of 70.0 Gy for patients who received 68.0 –72.0 Gy and a median of 78.0 Gy for patients who received ⱖ 72.0 Gy. Although the radiation delivery techniques varied over the 12 years, there was reasonable consistency. A conformal technique was used for 465 patients (34%), a standard technique was used for 556 patients (41%), and intensity-modulated RT (IMRT) was used for 331 patients (24%). Ninety-five percent of all patients who received standard technique RT (n ⫽ 526) were treated with a 4-field arrangement with standard shielding.

The remaining 30 patients who were treated with standard-technique RT received an arc rotation boost. The conformal technique consisted of a 10-field arrangement (M. D. Anderson technique; initial 4-field followed by a 6-field boost3) in 456 patients (98% of all conformally treated patients). The remaining 9 patients were treated using a 4-field, noncoplanar technique. All 331 patients who were treated with IMRT received a total dose of 70.0 Gy delivered at 2.5 Gy per fraction through 5 fields using a multileaf collimator setup. Descriptions of the IMRT technique have been published previously.2,4 The pelvic lymph nodes were irradiated in 276 patients (20%). The volume of pelvic coverage differed among treating physicians and different periods. Since 1994, patients have had pelvic lymph nodes excluded from the treatment fields. The radiation was prescribed to isocenter in patients who were treated with standard techniques. For patients who received conformal RT, the dose was prescribed to an isodose line covering the target (mean prescription isodose line, 97%; range, 88 –100%). For patients who received IMRT, the dose was prescribed to an isodose line covering the target (mean prescription isodose line, 87%; range, 82–90%). For patients who were treated with conformal or intensity-modulated techniques who had T1–T2 lesions, iPSA levels ⱕ 10.0 ng/mL, and bGS ⱕ 6, the target was the prostate only. The seminal vesicles were treated to full dose as part of the target for patients who had a T3 lesion, an iPSA level ⬎ 10.0 ng/mL, or a bGS ⱕ 7.5 AD therapy for ⱕ 6 months was delivered neoadjuvantly or adjuvantly to 459 patients (34%). None of the patients included in this analysis received AD therapy for a period that exceeded 6 months. The median duration of AD therapy was 6 months (range, 3– 6 months). AD therapy was administered neoadjuvantly and concomitantly with radiation in 65% of patients and was administered adjuvantly in 35% of patients. The minimum follow-up for the entire series was 12 months. Follow-up information was obtained from medical records or from communications with the patient, outside physicians, or hospitals. Follow-up information always included PSA levels; 12,443 follow-up PSA levels were available for analysis, an average of 9 levels per patient. The median follow-up was 55 months (mean, 60 months; range, 12–189 months). Actuarial curves were calculated using the Kaplan–Meier method, and tests of significance of differences between curves were based on the log-rank statistic. Cox proportional hazards were used for multivariate analysis. The first analysis endpoint was biochemical recur-

External RT for Localized PC in the PSA Era/Khuntia et al.

rence–free survival (bRFS). The American Society for Therapeutic Radiology and Oncology consensus definition for biochemical failure was used.6 bRFS was the most reliable endpoint, because different therapies were instituted at the time of biochemical failure in a significant number of patients, rendering clinical failures (local or distant) more difficult to use as endpoints. The second endpoint was clinical recurrence– free survival (local or distant recurrence–free survival), as determined by physical examination, biopsy, or radiologic studies. The time to failure for local and distant recurrences was calculated from the starting date of radiation to the date of clinical failure. The third endpoint was overall survival. The time to death was calculated from the starting date radiation to the date of death.

RESULTS Pretreatment and Treatment Characteristics Table 1 summarizes the pretreatment clinical characteristics and treatment parameters of the 1352 patients by treatment modality. Patients who received RT doses ⱖ 72.0 Gy were slightly younger. The median iPSA level was 9.6 ng/mL, and the mean iPSA level was 15.6 ng/mL (range, 0.4 – 692.9 ng/mL). bGS was higher in patients who received RT doses ⱖ 72.0 Gy (50% had a bGS ⱖ 7). The delivered radiation doses gradually increased over the study period. The median dose in the group that received ⱕ 68.0 Gy was 66.6 Gy, compared with 78.0 Gy in the group that received ⱖ 72.0 Gy, for a difference in median doses of 11.4 Gy. The median dose in the group that received 68.0 –72.0 Gy was 70.0 Gy. Understandably, the median follow-up periods were different for the different dose groups: 40 months for the group that received ⬎ 72.0 Gy (range, 12–148 Gy), 77 months for the group that received 68.0 –72.0 Gy (range, 12–189 Gy), and 92 months for the group that received ⱖ 72.0 Gy (range, 12–148 Gy). RT techniques also changed gradually over the years in the PSA era. Before 1995, the large majority of patients (99%) received RT via standard techniques. Between 1995 and 1998, the majority of patients (71%) received 3-dimensional conformal RT. After 1999, the majority of patients received IMRT (88%). With respect to AD therapy, only patients who received ⱕ 6 months of hormone therapy were included in the study sample, all with a minimum 1 year follow-up after RT. One-third of all patients received hormone therapy for ⱕ 6 months. The use of AD therapy was more prevalent in the more recent years and in patients with more advanced tumors. Consequently, 55% of patients who were treated with radiation doses ⱖ 72.0 Gy received hormone therapy for

1285

TABLE 1 Distribution of Pretreatment and Treatment Parameters by Radiation Dose Groupa

Factor

All

68 to 72 Gy

No. (%)

No. (%)

No. (%)

No. (%)

373 (28)

778 (57)

70 (—) 106 (28) 267 (72)

68 (—) 250 (32) 528 (68)

89 (24) 284 (76)

181 (23) 597 (77)

221 (59) 118 (32) 34 (9)

577 (74) 135 (17) 66 (9)

37 (10) 158 (42) 95 (26) 83 (22)

32 (4) 390 (50) 233 (30) 123 (16)

238 (64) 88 (23) 47 (13)

392 (50) 297 (38) 89 (12)

126 (34) 82 (22) 165 (44)

217 (28) 228 (29) 333 (43)

70.0 (—) 68.0–70.4 (—)

78.0 (—) 72.0–83.0 (—)

307 (82) 66 (18) 0 (0)

49 (6) 398 (51) 331 (43)

345 (92) 28 (8)

350 (45) 428 (55)

Total 1352 (100) 201 (15) Age Median (yrs) 69 (—) 70 (—) ⬍ 65 408 (30) 52 (26) ⬎ 65 944 (70) 149 (74) Race Black 314 (23) 44 (22) White 1038 (77) 157 (78) Clinical tumor classification (AJCC 1992) T1/T2a 930 (69) 132 (66) T2b–c 311 (23) 58 (29) T3 111 (8) 11 (5) iPSA level (ng/mL) ⱕ4 86 (6) 17 (9) ⬎4 to ⱕ10 623 (46) 75 (37) ⬎10 to ⱕ20 391 (29) 63 (31) ⬎20 252 (19) 46 (23) Biopsy Gleason score ⱕ6 753 (56) 123 (61) 7 437 (32) 52 (26) ⱖ8 162 (12) 26 (13) Risk group Low 395 (29) 52 (26) Intermediate 373 (28) 63 (31) High 584 (43) 86 (43) Radiation dose (Gy) Median 74.0 (—) 66.6 (—) Range 63.0–83.0 (—) 63.0–67.0 (—) Radiation technique Standard 556 (41) 200 (99) Conformal 465 (4) 1 (1) IMRT 331 (25) 0 (0) Androgen deprivation (ⱕ 6 mos) No 893 (66) 197 (9) Yes 59 (34) 3 (1)

Gy: Grays; AJCC: American Joint Committee on Cancer; iPSA: pretreatment prostate-specific antigen; IMRT: intensity-modulated radiotherapy. a The chi-square test for significance was performed for ⬍72 Gy versus ⱖ72 Gy.

ⱕ 6 months. Because there was a minimum follow-up of 12 months for all patients, we expect that the effect of hormone therapy will be minimal and that followup PSA levels will be reliable.

Treatment Results Biochemical recurrence A total of 410 patients experienced biochemical failure: 129 patients (64%) in the group that received RT doses ⬍ 68.0 Gy, 172 patients (46%) in the group that received 68.0 –72.0 Gy, and 109 patients (14%) in the group that received ⱖ 72.0 Gy. The 5-year and 10-year

1286


FIGURE 1. Biochemical recurrence–free survival (bRFS) by radiation dose. Symbols represent censored events. Gy: grays.

FIGURE 2.

Biochemical recurrence–free survival (bRFS) by (A) clinical T

classification, (B) pretreatment prostate-specific antigen level (ng/mL), (C) biopsy Gleason score, and (D) risk group. Crosses represent censored events.

bRFS rates were 63% (95% confidence interval [95% CI], 60 – 66%) and 55% (95% CI, 49 – 60%), respectively. The 5-year bRFS rates for patients who received RT doses of ⬍ 68.0 Gy, 68.0 –72.0 Gy, and ⱖ 72.0 Gy were 33% (95% CI, 26 – 40%), 53% (95% CI, 48 –59%), and 81% (95% CI, 77– 85%), respectively (P ⬍ 0.001) (Fig. 1). The 10-year bRFS rates for patients who received RT doses ⬍ 68.0 Gy, 68.0 –72.0 Gy, and ⱖ 72.0 Gy were 28% (95% CI, 20 –35%), 45% (95% CI, 38 –52%), and 80% (95% CI, 76 – 84%), respectively (P ⬍ 0.001). The overall outcomes by T classification, iPSA level, and bGS are displayed in Figure 2A–C. There is a significant difference in bRFS rates between patients with disease classified as clinical T1–2a and patients with disease classified as clinical T2b– c. Patients with

clinical T3 tumors had a 5-year bRFS rate of 42%. Patients with a bGS ⱕ 6 or ⫽ 7 had similar bRFS rates, whereas patients with a bGS ⱖ 8 had significantly worse outcomes. Patients with a bGS ⱖ 8 had a 5-year bRFS rate of 44%. Figure 2D shows bRFS rates by risk group. Throughout the PSA era, radiation doses were increased gradually. The earliest patients in the PSA era received the lowest doses and had the longest followup. For patients who received ⬍ 68.0 Gy, the median RT dose was 66.6 Gy, and the median follow-up was 92 months, compared with 70.0 Gy and 77 months for patients who received 68.0 –72.0 Gy and 78.0 Gy and 40 months for patients who received ⱖ 72.0 Gy. Figure 3A–C shows outcome data for patients in each of the three dose groups according to risk category. Although the lengths of follow-up obviously are different, the outcomes are dramatically different. In each dose group, risk category was a significant predictor of outcome. It is noteworthy that, increasing from a median dose of 66.6 Gy to a median dose of 70.0 Gy (from the ⬍ 68.0 Gy group to the 68.0 –72.0 Gy group), the major improvement in bRFS was observed in the low-risk group. However, increasing from a median dose of 70.0 Gy to a median dose of 78.0 Gy (from the 68.0 – 72.0 Gy group to the ⱖ 72.0 Gy group), the improvement in bRFS was most evident in the intermediaterisk and high-risk groups. The use of AD therapy in combination with RT did increase during the PSA era. Analysis of the outcomes of patients who were treated with RT alone versus patients who received RT and AD therapy demonstrated that the median follow-up for patients who received AD therapy was significantly shorter: 70 months versus 39 months. Analysis of the difference in Kaplan–Meier survival curves between patients who received AD therapy and patients who did not receive AD therapy revealed a statistically significant improvement in bRFS for patients who received AD therapy (Fig. 4A). Similarly, the more modern radiation techniques (conformal RT and IMRT) yielded higher bRFS rates compared with conventional RT techniques (Fig. 4B). Due to the expected effect of hormone therapy on follow-up PSA levels, the analysis of outcomes in the three different dose groups by risk group was repeated excluding patients who received AD therapy. Figure 5A–C displays those results. The observed trends in outcomes were similar to the trends observed for the entire cohort. A prior analysis of a similar cohort of patients treated at the Cleveland Clinic Foundation showed an improvement in outcomes for intermediate-risk and high-risk patients with the use of short-term AD therapy.7 That analysis included all patients at different


1287

FIGURE 4. (A) Biochemical recurrence–free survival according to the use of androgen-deprivation (AD) therapy. No AD: n ⫽ 893; median follow-up, 70 months. AD: n ⫽ 459; median follow-up, 39 months. (B) Biochemical recurrence–free survival according to radiotherapy (RT) technique. Intensity-modulated RT (IMRT): n ⫽ 331; median follow-up, 27 months. Conformal RT: n ⫽ 465; median follow-up, 52 months. Standard RT: n ⫽ 556; median follow-up, 85 months. Crosses represent censored events.

FIGURE 3. Biochemical recurrence–free survival (bRFS) by risk group in the different radiation dose groups. (A) ⬍ 68.0 Grays (Gy) (n ⫽ 201). (B) 68.0 –72.0 Gy (n ⫽ 373). (C) ⱖ 72.0 Gy (n ⫽ 778). Crosses represent censored events.

dose levels, not taking into account the important impact of year of therapy on recurrence-free survival rates. To more accurately reflect outcomes for patients currently treated with radiation, an analysis was performed that included patients who received only doses ⬎ 72.0 Gy (median dose, 78.0 Gy). For the three risk categories, bRFS rates were analyzed by the use of short-term AD therapy. Figure 6A shows that there was no improvement in outcomes with the use of AD therapy in the low-risk group. We observed statistically nonsignificant trends toward outcome improvement associated with the use of AD therapy in the

1288


intermediate-risk and high-risk groups. These findings indicate the presence of only a small additional benefit (if any exists) associated with the use of short-term AD therapy in patients who received high radiation doses (ⱖ 72.0 Gy; median dose, 78.0 Gy). It also is noteworthy that overall outcomes for patients who receive higher radiation doses are excellent: recurrence-free survival rates are expected to be 90 –95% for low-risk patients, 80 – 85% for intermediate-risk patients, and 65–70% for high-risk patients. A multivariate analysis to predict for biochemical failure was performed using the following factors: age (continuous variable), race (African American vs. Caucasian), T classification (T1–2a vs. T2b– c vs. T3), iPSA (continuous variable), bGS (continuous variable), use of AD therapy (yes vs. no), radiation technique (standard vs. conformal/IMRT), radiation dose (continuous variable), and year of therapy (continuous variable). Table 2 summarizes the results of the multivariate analysis. Clinical T classification (P ⬍ 0.001), iPSA (P ⬍ 0.001), bGS (P ⫽ 0.001), radiation dose (P ⬍ 0.001), and year of therapy (P ⬍ 0.001) were independent predictors of biochemical failure; whereas age (P ⫽ 0.31), race (P ⫽ 0.10), use of AD therapy (P ⫽ 0.34), and radiation technique (P ⫽ 0.10) were not. It is noteworthy that although they were found to be significant on univariate analysis, AD therapy and radiation technique were not independent predictors of bRFS.

Clinical recurrence

FIGURE 5.

Biochemical recurrence–free survival (bRFS) by risk group for patients in each radiation dose group who received radiotherapy alone, without any androgen-deprivation therapy. (A) ⬍ 68.0 Grays (Gy) (n ⫽ 197; median follow-up, 92 months). (B) 68.0 –72.0 Gy (n ⫽ 345; median follow-up, 79 months). (C) ⱖ 72.0 Gy (n ⫽ 350; median follow-up, 44 months). Crosses represent censored events.

At the time of last follow-up, of the 410 patients who experienced biochemical failure, 100 had documented clinical failures (local [n ⫽ 26], distant [n ⫽ 62], or both [n ⫽ 12]). For the entire cohort, the actuarial incidence of clinical failures was 7% at 5 years and 14% at 10 years from the time of RT. The actuarial incidence of clinically detectable local failures was 3% at 5 years and 6% at 10 years for the entire cohort. The actuarial incidence of clinically detectable distant failures was 5% at 5 years and 11% at 10 years for the entire cohort. The actuarial incidence of clinical failures was 21% at 5 years after biochemical failure and 28% at 10 years after biochemical failure. There was a statistically significant decrease in the rate of clinically documented local failures with higher radiation doses: the 10-year local recurrence rates for patients who received radiation doses ⱖ 72.0 Gy versus ⬍ 72.0 Gy were 3% and 7%, respectively (P ⫽ 0.001). A significant proportion of patients who experienced biochemical failure received salvage therapy before the development of a clinically detectable recurrence. Consequently, the reliance on clinical failure to assess outcome is questionable.


1289

TABLE 2 Multivariate Cox Proportional Hazards Regression Analysis of Factors that Affected Biochemical Recurrence–Free Survival Factor

P value

Age (continuous variable) Race (African American vs. Caucasian) Clinical T classification (T1–2a vs. T2b–c vs. T3) PSA level (continuous variable) Biopsy Gleason score (continuous variable) Use of androgen deprivation (no vs. yes) Radiation technique (standard vs. conformal/IMRT) Radiation dose (continuous variable) Year of therapy (continuous variable)

0.31 0.10 ⬍0.001 ⬍0.001 0.001 0.34 0.10 ⬍0.001 ⬍0.001

PSA: prostate-specific antigen; IMRT: intensity-modulated radiotherapy.

Overall survival A total of 205 patients died during the study period. Prostate carcinoma was the documented cause of death or the probable cause of death in 78 patients. The 5-year and 10-year overall survival rates were 89% (95% CI, 87–91%) and 61% (95% CI, 55– 68%), respectively. A multivariate analysis was performed to predict for death using the following factors: age (continuous variable), race (African American vs. Caucasian), T classification (T1–2a vs. T2b– c vs. T3), iPSA (continuous variable), bGS (continuous variable), use of AD therapy (yes vs. no), and radiation dose (continuous variable). The independent predictors of death were age (P ⫽ 0.008), bGS (P ⫽ 0.002), T classification (P ⫽ 0.002), and radiation dose (P ⫽ 0.042).

DISCUSSION In the PSA era, the reported outcomes after externalbeam RT for localized prostate carcinoma have varied significantly.1,8 –10 Patient selection, use of hormone therapy, differing radiation doses, and time-related T classification migration are some of the reasons for the observed variation among different external-beam radiation series. The overall outcomes for patients who were treated at the Cleveland Clinic Foundation have been reported previously. In the current report, those

Š FIGURE 6. Biochemical recurrence–free survival (bRFS) according to the use of androgen-deprivation (AD) therapy for patients in each risk group who were treated with a radiation dose of ⱖ 72.0 Grays (Gy). (A) Low risk. No AD: n ⫽ 184; median follow-up, 41 months. AD: n ⫽ 33; median follow-up, 31 months. (B) Intermediate risk. No AD: n ⫽ 95; median follow-up, 41 months. AD: n ⫽ 133; median follow-up, 32 months. (C) High risk. No AD: n ⫽ 71; median follow-up, 68 months. AD: n ⫽ 262; median follow-up, 42 months. Crosses represent censored events.

1290


outcomes are updated. Subset analyses were performed to better evaluate the factors that affected recurrence rates. For the current cohort of patients, who were treated in the period from 1987 to 2000, the 5-year bRFS rate was 63% in 1352 patients with T1–T3 tumors. This is comparable to the 66% bRFS rate reported by Shipley et al. from a pooled series of 1765 patients from 6 separate institutions with T1–T2 tumors. Therefore, the overall results from our series were consistent with what would be expected for patients treated in the PSA era. In our series, there was no evidence that age was a predictor of biochemical failure or clinical failure. Neither biochemical nor clinical recurrence rates were affected by age. Similarly, race was not a predictor of any of the studied endpoints. Therefore, treatment decisions should not be made on the basis of perceived differences in outcomes by age or race. Treatment recommendations should be made independent of both race and age. Overall, the long-term biochemical failure rates in patients with clinical T2b– c tumors was similar to the rates in patients with clinical T3 tumors (Fig. 2A). This emphasizes the importance of documenting the clinical T classification in patients with T2 lesions. Patients with a bGS ⫽ 7 had long-term biochemical failure rates similar to the rates for patients with a bGS ⱕ 6. Patients with a bGS ⱖ 8 had significantly worse outcomes, as seen in Figure 2C. A crucial parameter affecting biochemical recurrence has been the radiation dose delivered to the target areas. The M. D. Anderson Cancer Center randomized trial reported by Pollack et al. demonstrated superior outcomes in patients who received an RT dose of 78.0 Gy compared with patients who received 70.0 Gy.11 Multiple retrospective series from different institutions have confirmed the improvement in cure rates with higher radiation doses.10,12–16 At institutions where the delivered radiation doses were relatively low and within a narrow range, investigators have had difficulty demonstrating an association between higher radiation doses and lower recurrence rates.17 In the current series, the median RT dose for patients who received ⬍ 68.0 Gy was 66.6 Gy, compared with 70.0 Gy for patients who received 68.0 –72.0 Gy and with 78.0 Gy for patients who received ⱖ 72.0 Gy. Therefore, patients who received ⱖ 72.0 Gy had substantially higher doses compared with the other groups. The largest difference in outcome would be expected between patients who received ⬍ 68.0 Gy and patients who received ⱖ 72.0 Gy, as seen in Figures 1 and 3. The difficulty in assessing the impact of radiation dose is related to the fact that patients who receive higher radiation doses typically have been

treated with better radiation techniques, are more likely to have received hormone therapy, and have shorter follow-up periods. Figure 4B shows that a large difference in outcome was observed between patients who were treated with the different radiation techniques: standard RT, conformal RT, and IMRT. However, multivariate analyses (e.g., see Table 2) failed to show that RT technique is an independent predictor of biochemical recurrence rates. This clearly suggests that higher doses, rather than the design of the radiation fields, result in better control of local disease. Therefore, modern radiation techniques do not necessarily improve cure rates; rather, it is the higher radiation doses delivered. With respect to IMRT, although outcomes were promising in the current series, it is too early to make any valid recurrence rate comparisons between IMRT and more traditional conformal techniques. Modern radiation techniques are expected to impact toxicity rates rather than recurrence rates.18 –20 The other complicating factor in assessing the outcomes of patients with prostate carcinoma treated in the PSA era is the use of hormone therapy.21 In the current series, the use of AD therapy was limited to durations of 6 months or less. A minimum of 1 year of follow-up ensures that follow-up PSA levels are reliable for establishing biochemical failure. Similar to radiation technique, a large difference was seen in Kaplan–Meier survival estimates between patients who received 6 months of AD therapy versus patients who received no AD therapy. When patients who received AD therapy were excluded from the analysis, the strong impact of higher radiation doses on recurrence rates still was seen (Fig. 5). In addition, in our series, AD therapy was not an independent predictor of biochemical recurrence, clinical recurrence, or death in multivariate analysis. Therefore, the use of short-term hormone therapy does not explain the observed improved outcomes with increasing radiation doses. Because modern RT typically involves doses exceeding 72.0 Gy, it is important to assess the potential role of short-term AD therapy in improving recurrence rates for patients who receive high radiation doses. There have been suggestions that the impact of short-term hormone therapy decreases with the use of higher radiation doses.1,7 Figure 6A demonstrates that there was no improvement in outcomes for low-risk patients who received RT doses ⱖ 72.0 Gy with the use of AD therapy. In intermediate-risk and high-risk patients, there was a small trend toward decreased recurrence rates with the use of short-term AD therapy (Fig. 6B,C). It is noteworthy that with or without AD therapy, the use of ⱖ 72.0 Gy yielded adequate bRFS rates in patients with high-risk disease. bRFS rates are


expected to be 90 –95% for low-risk patients, 80 – 85% for intermediate-risk patients, and 65–70% for highrisk patients. In addition to radiation dose and use of hormonal therapy, other patient-, tumor-, and treatment-related factors have varied over the years. Known and unknown factors changing over time may affect the outcomes of patients who are treated over long periods. One such change has been Gleason scoring. It is a well known phenomenon that Gleason score determination has shifted over the years: the same tumor is more likely to have a higher score rather than a lower score in more recent years.22 Other factors, such as tumor bulk, also probably have changed in more recent years, with the bulk of tumor decreasing with time. The introduction of year of therapy as a confounding factor in multivariate analyses can control for such changes that have occurred over time.23,24 With the inclusion of year of therapy in the multivariate analyses, the radiation dose, rather than the radiation technique or the use of AD therapy, was an independent predictor of biochemical failure. This finding clearly indicates that radiation dose is the single most important treatment factor influencing cure after patients undergo external-beam RT for localized prostate carcinoma. Clinical outcomes (local and/or distant failures) are difficult to interpret in the PSA era, in which therapeutic intervention frequently takes place at the time of biochemical failure. In the current series, local control was good, with an overall local failure rate of 6% at 10 years after RT. The local failure rate was only 3% among patients who received RT doses ⱖ 72.0 Gy. This finding indicates that although biochemical failure is relatively common, actual clinical local failures are infrequent after external-beam RT for localized prostate carcinoma. The distant failure rate for the entire cohort was 11% at 10 years. The most remarkable observation in the current series was that, among the patients who actually experienced biochemical failure after RT, only 28% developed a clinical failure (either local or distant) at 10 years after the failure date. This is an important observation that can be used in counseling patients when they are faced with rising PSA levels that indicate disease recurrence. Finally, the overall survival rates were comparable what would be expected in an age-matched, normal population. The majority of deaths in our patient cohort were not related to prostate carcinoma. However, the radiation dose still was an independent predictor of overall survival, along with Gleason score, T classification, and patient age. In conclusion, in the PSA era, the most significant therapeutic factor impacting bRFS rates for patients

1291

with localized prostate carcinoma after external-beam RT has been radiation dose, rather than AD therapy use or radiation technique. In addition to bRFS rates, both local control rates and overall survival rates have been improved by higher-than-standard radiation doses.

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

Kupelian PA, Mohan DS, Lyons J, Klein EA, Reddy CA. Higher than standard radiation doses (72 Gy or greater) with or without androgen deprivation in the treatment of localized prostate cancer. Int J Radiat Oncol Biol Phys. 2000;46: 567–574. Mohan DS, Kupelian PA, Willoughby TR. Short-course intensity-modulated radiotherapy for localized prostate cancer with daily transabdominal ultrasound localization of the prostate gland. Int J Radiat Oncol Biol Phys. 2000;46:575– 580. Pollack A, Zagars G, Starkshall G, et al. Conventional vs. conformal radiotherapy for prostate cancer: preliminary results of dosimetry and acute toxicity. Int J Radiat Oncol Biol Phys. 1996;34:555–564. Kupelian PA, Willoughby TR. Short-course, intensity-modulated radiotherapy for localized prostate cancer. Cancer J. 2001;7:421– 426. Katcher J, Kupelian PA, Zippe C, Klein EA, Sohn JW. Indications for excluding the seminal vesicles when treating clinically localized prostatic adenocarcinoma with radiotherapy alone. Int J Radiat Oncol Biol Phys. 1997;37:871– 876. American Society for Therapeutic Radiology and Oncology Consensus Panel. Consensus statement: guidelines for PSA following radiation therapy. Int J Radiat Oncol Biol Phys. 1997;37:1035–1041. Davies-Johns T, Reddy C, Kupelian P. Localized prostate cancer treated with radiotherapy: who benefits from hormonal therapy? Int J Radiat Oncol Biol Phys. 2000;48(Suppl 1):170. 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. Zelefsky MJ, Fuks Z, Leibel SA. Intensity-modulated radiation therapy for prostate cancer. Semin Radiat Oncol. 2002; 12:229 –237. Pollack A, Smith LG, von Eschenbach AC. External beam radiotherapy dose response characteristics of 1127 men with prostate cancer treated in the PSA era. Int J Radiat Oncol Biol Phys. 2000;48:507–512. Pollack A, Zagars GK, Starkschall G, et al. Prostate cancer radiation dose response: results of the M. D. Anderson Phase III randomized trial. Int J Radiat Oncol Biol Phys. 2002;53:1097–1105. Pinover WH, Hanlon AL, Horwitz EM, Hanks GE. Defining the appropriate radiation dose for pretreatment PSA ⱖ 10 ng/mL prostate cancer. Int J Radiat Oncol Biol Phys. 2000; 47:649 – 654. Lyons JA, Kupelian PA, Mohan DS, Reddy CA, Klein EA. Importance of high radiation doses (72 Gy or greater) in the treatment of stage T1-T3 adenocarcinoma of the prostate. Urology. 2000;55:85–90. Zelefsky MJ, Leibel SA, Kutcher GJ, Fuks Z. Three-dimensional conformal radiotherapy and dose escalation: where do we stand? Semin Radiat Oncol. 1998;8:107–114.

1292


15. Zelefsky MJ, Fuks Z, Hunt M, et al. High-dose intensity modulated radiation therapy for prostate cancer: early toxicity and biochemical outcome in 772 patients. Int J Radiat Oncol Biol Phys. 2002;53:1111–1116. 16. Hanks GE, Hanlon AL, Epstein B, Horwitz EM. Dose response in prostate cancer with 8 –12 years’ follow-up. Int J Radiat Oncol Biol Phys. 2002;54:427– 435. 17. Hurwitz MD, Schnieder L, Manola J, Beard CJ, Kaplan ID, D’Amico AV. Lack of radiation dose response for patients with low-risk clinically localized prostate cancer: a retrospective analysis. Int J Radiat Oncol Biol Phys. 2002;53:1106 – 1110. 18. Kupelian PA, Reddy CA, Carlson TP, Altsman KA, Willoughby TR. Preliminary observations on biochemical relapse-free survival rates after short-course intensity-modulated radiotherapy (70 Gy at 2.5 Gy/fraction) for localized prostate cancer. Int J Radiat Oncol Biol Phys. 2002;53:904–912. 19. Kupelian PA, Reddy CA, Carlson TP, Willoughby TR. Dose/ volume relationship of late rectal bleeding after external beam radiotherapy for localized prostate cancer: absolute or relative rectal volume? Cancer J. 2002;8:62– 66.

20. Horwitz EM. Conformal radiotherapy with dose escalation appears safe and effective in localised prostate cancer. Cancer Treat Rev. 2003;29:219 –221. 21. D’Amico AV, Schultz D, Loffredo M, et al. Biochemical outcome following external beam radiation therapy with or without androgen suppression therapy for clinically localized prostate cancer. JAMA. 2000;284:1280 –1283. 22. Chism DB, Hanlon AL, Troncoso P, Al-Saleem T, Horwitz EM, Pollack A. The Gleason score shift: score four and seven years ago. Int J Radiat Oncol Biol Phys. 2003;56:1241–1247. 23. Jhaveri FM, Klein EA, Kupelian PA, Zippe C, Levin HS. Declining rates of extracapsular extension after radical prostatectomy: evidence for continued stage migration. J Clin Oncol. 1999;17:3167–3172. 24. Kupelian PA, Buchsbaum JC, Elshaikh MA, Reddy CA, Klein EA. Improvement in relapse-free survival throughout the PSA era in patients with localized prostate cancer treated with definitive radiotherapy: year of treatment an independent predictor of outcome. Int J Radiat Oncol Biol Phys. 2003;57:629 – 634.