|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology Radical Radiation for Localized Prostate Cancer: Local Persistence of Disease Results in a Late Wave of MetastasesByFrom the Department of Radiation Oncology, Massachusetts General Hospital, Harvard Medical School, Boston, MA. Address reprint requests to John Coen, MD, Department of Radiation Oncology, Massachusetts General Hospital, 100 Blossom St, Cox 3, Boston, MA 02114; email: jcoen{at}partners.org
PURPOSE: To assess whether failure to maintain local control (LC) of prostate cancer after radiation therapy results in a higher incidence of distant metastasis (DM). PATIENTS AND METHODS: From 1972 to 1999, 1,469 patients with clinically localized prostate cancer were treated with radical radiation therapy. Disease outcome was retrospectively reviewed for all patients with more than 2 years of follow-up.
RESULTS: The actuarial 10-year LC rate was 79%. Gleason score CONCLUSION: Patients with locally persistent prostate cancer are at greater risk of DM. The higher initial hazard of DM is consistent either with an increased likelihood of subclinical micrometastases before treatment or with posttreatment tumor embolization. The prolonged time to appearance of DM in locally failing patients and the increasing hazard of DM over time is most consistent with a late wave of metastases from a locally persistent tumor.
ERADICATION OF local prostate cancer is the primary aim of radical radiation therapy. Failure to eradicate tumor in the prostate may result in local disease progression, with substantial morbidity; however, the most significant sequelae from uncontrolled prostate cancer are distant metastasis and the hormonal manipulation its management entails. Persistence of local prostate cancer has been associated with a higher incidence of distant metastasis in several previous studies.1-3 Establishing a causal relationship between failure to eradicate disease in the prostate and the subsequent development of metastases has been difficult. Many of the same factors that predict for local failure also predict for distant metastasis (ie, higher T stage, poorly differentiated tumors, and high prostate-specific antigen [PSA]). The suggestion that these end points may be independent, with common pretreatment predictors, has so far only been countered by a single multivariate analysis that found local recurrence to be an independent prognostic feature for the development of distant metastases.1 The two leading hypotheses for the association of local failure and increased risk for metastasis are the inherent pathobiologic aggressiveness theory and the reseeding theory.2 The pathobiologic aggressiveness theory states that certain tumors are inherently more virulent and are thus more difficult to eradicate locally and more inclined to have micrometastases at diagnosis. Local persistence of disease may demonstrate an independent association with distant metastasis simply because it is a marker for more virulent disease. The reseeding theory states that for localized tumors lacking micrometastases at diagnosis, failure to completely eradicate the tumor may lead to subsequent shedding of tumor cells and a late wave of metastases. Although the biologic aggressiveness theory and the reseeding theory are not mutually exclusive, the reseeding theory is the only one that suggests the aggressive pursuit of local control is warranted. Our study, which represents the largest series available in the literature, demonstrates not only an independent association between local persistence of disease and metastasis, but also a temporally increasing hazard rate in patients with locally persistent disease not observed in locally controlled patients. The increasing hazard rate may represent a late wave of metastatic seeding.
Patients and Treatment Between 1972 and 1999, 1,469 patients with biopsy-proven localized prostate cancer were treated with radical radiation therapy at the Massachusetts General Hospital. Patients with evidence of pelvic lymph node disease and patients who received adjuvant androgen deprivation therapy are not included in this series. All of the patients analyzed had greater than 2 years of follow-up. External-beam radiation therapy was delivered to 1,427 patients, and 42 patients received interstitial brachytherapy by means of a retropubic approach. The median dose to the prostate was 68.4 Gy, ranging from 50 to 80 Gy.
End Points
Statistical Analysis
Definition of Locally Persistent Disease as a Prognostic Feature
Patient Characteristics The median age at diagnosis was 71 years (range, 46 to 90 years). All 1,469 patients had at least 2 years of follow-up, with a median follow-up of 70 months (range, 25 to 253 months). Pretreatment PSA data were available in 59% of patients and the median PSA was 8.8 (range, 0.5 to 661). Of the patients with initial PSA data, 73% had a pretreatment value of less than 15. Overall, 69% of patients had a favorable Gleason score of  7 and 63% had T1 to T2 tumors.
Local Control
Overall local control at 5, 10, and 15 years was 89%, 79%, and 69%, respectively. High Gleason score, T3 to T4 tumors, and high PSA all predicted for an increased incidence of local failure in a univariate analysis (Table 2). Gleason score less than 7 was associated with an 81% 10-year local control rate as opposed to 73% for Gleason score  7 (P = .003). The 10-year local control for patients with an initial PSA 15 was 78% as opposed to 70% for initial PSA more than 15 (P = .002). T1 to T2 tumors had an 81% 10-year local control rate as opposed to 74% for T3 to T4 tumors (P = .02) Pretreatment PSA level was the only independent prognostic factor in a multivariate model that included PSA, Gleason score, and T stage. This model only included 845 patients because many patients did not have pretreatment PSA data. When only T stage and Gleason score were included, the model contained 1,439 patients. T stage was the only independent prognostic feature in this model; however, Gleason score was of borderline significance.
Distant Metastasis and the Effect of Local Control on Distant Metastasis Clinically evident distant metastases were documented in 259 patients in our series. Local control was maintained in 191 of these patients, and 68 were ultimately found to have local disease. Overall DMFS at 5, 10, and 15 years was 86%, 74%, and 64%, respectively. The incidence of distant metastasis was increased in patients with higher Gleason score, higher T stage, and higher PSA level. Gleason score 7 demonstrated a 58% 10-year DMFS versus 82% for Gleason score less than 7 (P < .0001). T3 to T4 tumors demonstrated a 61% 10-year DMFS versus 82% for T1 to T2 tumors (P < .0001). For initial PSA 15, 10-year DMFS was 89% versus 70% for initial PSA more than 15 (P < .0001). DMFS was better in patients maintaining local control. For patients maintaining local control, DMFS at 5, 10, and 15 years was 87%, 77%, and 72%, respectively. For patients with locally persistent disease, DMFS at 5, 10, and 15 years was 80%, 61%, and 37%, respectively. The difference was statistically significant, with P < .0001 (Fig 1 and Table 3).
All four prognostic factors maintained statistical significance when tested in a multivariate model. Local disease status was tested as both a time-dependent and time-independent variable in separate multivariate models. In the time-dependent model, patients’ disease was categorized as locally controlled if there was no evidence of locally persistent prostate cancer at the time of metastatic failure. In the time-independent model, patients were categorized as having locally persistent disease if they had local disease documented at any time over the entire course of follow-up. Local disease status was the strongest predictor for metastatic disease in both models (Table 4). The median time to the appearance of metastasis was longer in patients who failed to maintain local control, 54 months compared with 34 months in patients with locally controlled disease (P = .0003).
There were 68 patients who had evidence of both local and distant failure over the course of their follow-up. Local failure was documented before distant metastasis in 39 patients, distant failure was documented first in 11, and 18 patients had both events documented at the same follow-up date. For patients with local failure as the first event, the median time from local failure to distant metastasis was 13 months (range, 1 to 89 months). For patients with distant failure as the first event, the median time to local failure was 14 months (range, 1 to 36 months).
Hazard of Distant Metastasis and Local Disease Status
Salvage Androgen-Deprivation Therapy Salvage hormonal therapy was initiated in 205 patients for biochemical, local, or distant recurrence in 89, 57, and 59 patients, respectively. Among patients who ultimately developed a local recurrence, 87 (44%) of 198 received hormonal therapy compared with 118 (9%) of 1,271 patients who maintained local control (  2 P < .0001). Among the patients who received salvage hormonal therapy, the proportion of patients who manifested clinical evidence of distant metastasis was higher in patients maintaining local control compared with patients with locally recurrent disease, 47% versus 32%, respectively (P = .04). The median time to the initiation of androgen-deprivation therapy was longer in patients with locally persistent disease than in patients maintaining local control, 54 versus 30 months, respectively (P = .0004). When patients who received salvage hormones for either local failure or biochemical failure were excluded from analysis, local failure continued to be an independent predictor for distant metastasis, and the hazard rate of distant metastasis still increased over time in patients with locally persistent disease.
The association of local control with increases in both overall survival and freedom from distant metastasis has been demonstrated in a variety of disease sites.7 Although several authors have demonstrated an independent association between local control and distant metastasis in patients receiving radical radiation for prostate cancer, only Fuks et al1 were able to provide evidence of a causal relationship1-3 (Table 6). They demonstrated that the median time to distant metastasis was of shorter duration in patients whose disease was locally controlled (37 v 54 months, respectively) and the risk of distant metastasis increased over time in patients whose disease was locally relapsing. Our analysis confirmed both of these observations in a larger patient population predominantly treated with external-beam radiation therapy. It is presumed that the shorter median time to metastasis in patients with locally controlled disease represents the progression of micrometastases that may have been present before local treatment. Median time to metastasis may be prolonged in patients with locally persistent disease, as posttreatment tumor embolization may contribute to the metastatic rate. These posttreatment metastases would manifest clinical symptoms at a later time. The model our data suggests is the development of secondary metastases from local persistence of disease superimposed on a background of occult micrometastases that present themselves over time.
Our data are potentially confounded, as a higher proportion of patients with local persistence of disease received hormonal therapy. Although this intervention could potentially prolong the time to metastasis, the median time to androgen blockade was prolonged to a similar degree as the time to metastasis in patients with locally persistent disease. Furthermore, the median time to distant metastasis in patients with locally controlled disease is shorter than the median time to androgen deprivation in patients with locally persistent disease. Thus, it is unlikely that the higher incidence of androgen-deprivation therapy is the sole reason for the delay in the appearance of metastases in patients with locally persistent disease. The longer interval to androgen-deprivation therapy in patients who ultimately manifest with a local failure is consistent with the theory that locally persistent disease can shed tumor cells, resulting in the late onset of systemic disease. The first presentation of this systemic disease may be a rising PSA or clinically detectable metastatic disease. Biochemical failure was the most frequent reason for the initiation of androgen deprivation in this series. The proportion of patients receiving hormonal therapy who ultimately manifested distant metastasis was higher in patients maintaining local control. Although our analysis is confounded by the more frequent use of androgen-deprivation therapy in patients with locally persistent disease, the manner in which hormones were initiated in this patient group would bias the results away from our conclusions. Androgen blockade could hide the association between distant metastasis and local control. Treatment of the first event could delay or prevent the clinical manifestations of the second event. Thus, the link would appear broken. Using hazard rate analysis, we were able to demonstrate an increasing risk of distant metastasis over time in patients who ultimately develop local failure. The risk of metastasis in patients with locally controlled disease did not change over time. The initial rate of metastatic failure in patients with locally persistent disease was higher than that in patients who maintained local control. Although patients who developed local failure had worse prognostic features at diagnosis, local failure was an independent predictor of metastatic failure in both a time-dependent and a time-independent model. As local failure is the result of failure to eradicate disease in the prostate, it is reasonable to regard these patients as having locally persistent disease after completion of radiation therapy. As many patients may have local persistence of prostate cancer but not declare themselves as local failures over the study period, the impact of local disease on distant dissemination is likely underestimated in this analysis. Systematic rebiopsy of all patients would be the only way to accurately assess the association. Rebiopsy series have demonstrated that a sizable proportion of patients without clinical evidence of locally persistent disease have evidence of residual cancer in the prostate after definitive radiation therapy. Positive rebiopsy rates ranging from 14% to 91% have been documented.8 Crook et al9 documented a 30% positive rebiopsy rate at 30 months. Positive rebiopsy was associated with higher stage tumors and higher PSA nadirs. Dugan et al10 demonstrated that, 2 years after radiation for T3 to T4 tumors, the positive rebiopsy rate was 8% for patients with PSA levels less than 1.0 versus 63% if the PSA level was greater than 1 at the time of biopsy. Positive rebiopsy has been associated with both local and distant failure in several series.8 The association of positive rebiopsy with higher PSA nadirs and higher PSA at the time of biopsy support the use of PSA as a surrogate for residual prostate cancer. The analysis performed by Fuks et al1 is strengthened by a 38% rebiopsy rate, which is higher than in most series. Positive biopsies were available in 67% of patients with a local recurrence, and confirmatory negative biopsies were available in 12% of patients maintaining local control. Our series only includes rebiopsy data on 84 of 1,469 patients. Our data lend support to both of the leading hypotheses explaining the association of distant metastasis and local failure in prostate cancer. The independent association of local failure with distant metastasis is necessary to support either theory. The pathobiologic aggressiveness theory is supported by the initial hazard rates, which demonstrate that the risk for distant metastasis is higher for patients with locally persistent disease from the beginning of the interval after local therapy. These initial rates may represent the risk of micrometastases at diagnosis, which subsequently present themselves as clinical metastases. Patients with local disease resistant to radical radiation may have a higher incidence of micrometastasis before treatment and thus a higher hazard rate for distant metastasis in a time-independent fashion. Both of these features are related to the inherent aggressiveness of the disease. Whereas the hazard rate remains constant in locally controlled patients, it slopes upward with time in patients with local persistence of disease. The kinetics of the development of metastasis in this group suggests a posttreatment event associated with a changing risk of metastasis. The inciting event may represent tumor embolization from persistent local disease giving rise to a late wave of metastases.
Our analysis lends support to the contention that local eradication of disease may result in improved DMFS. It suggests that more aggressive approaches toward local disease are warranted, even in patients with high-risk disease. Dose escalation has resulted in improved biochemical disease-free survival and lower positive rebiopsy rates in several series. Pollack et al11 demonstrated a dose-response relationship in a retrospective series of 1,127 patients, where the 4-year biochemical disease-free survival was 54%, 71%, and 77% for doses of Neoadjuvant hormonal therapy offers the advantage of tumor cytoreduction before the initiation of radiation therapy. This approach theoretically increases the efficacy of radiation therapy without requiring dose escalation and its attendant increased morbidity. Zietman et al8 showed that lower doses of radiation were required for local tumor control when an orchiectomy was performed before irradiation in an androgen-dependent murine adenocarcinoma. Clinical data support the use of neoadjuvant hormonal suppression in patients with locally advanced prostate cancer. The Radiation Therapy Oncology Group 8610 study demonstrated improved local control in patients with T2c to T4 prostate cancer randomized to neoadjuvant and adjuvant androgen suppression compared with radiation alone, 71% versus 46% at 5 years (P < .001).15 A randomized trial from Quebec demonstrated lower positive rebiopsy rates for patients treated with neoadjuvant hormonal suppression followed by radiation.16 Thus, neoadjuvant androgen suppression can play a role in improving the local eradication of prostate cancer with radiation therapy. Although improved biochemical control and local control have been the benchmarks by which newer treatment approaches have been evaluated, they represent viable surrogate end points for distant failure. Our hazard analysis, coupled with the experience of other institutions, has at the least proved local persistence of disease to be a harbinger of distant failure and that rising PSA is certainly correlated with subsequent metastasis. It is reasonable to make the assumption that reduced distant failure will lead to improvements in cause-specific survival and better quality of life. The vast majority of prostate cancer deaths are related to disseminated disease, and disease-related morbidity is primarily secondary to bony disease and the hormonal manipulation its management entails. Our analysis suggests that an aggressive approach to localized prostatic cancer may be beneficial in terms of prolonging survival and improving quality of life.
1. Fuks Z, Leibel SA, Wallner KE, et al: The effect of local control on metastatic dissemination in carcinoma of the prostate: Long-term results in patients treated with I-125 implantation. Int J Radiat Oncol Biol Phys 21: 537-547, 1991[Medline] 2. Zagars GK, Von Eschenbach AC, Ayala AG, et al: The influence of local control on metastatic dissemination of prostate cancer treated by external beam megavoltage radiation therapy. Cancer 68: 2370-2377, 1991[CrossRef][Medline] 3. Kuban DA, El-Mahdi AM, Schellhamer PF, et al: Effect of local tumor control on distant metastasis and survival in prostatic adenocarcinoma. Urology 30: 420-426, 1987[CrossRef][Medline] 4. Kaplan E, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53: 457-481, 1958[CrossRef] 5. Mantel N: Evaluation of survival data and two rank order statistics arising in its consideration. Cancer Chemother Rep 50: 163-170, 1966[Medline] 6. Cox DR: Regression models and life tables. J R Stat Soc 34: 187-220, 1972 7. Suit HD, Westgate SJ: Impact of local control on survival. Int J Radiat Oncol Biol Phys 12: 453-458, 1986[Medline] 8. Zietman AL, Westgeest JC, Shipley WU: Radiation-based approaches to the management of T3 prostate cancer. Semin Urol Oncol 15: 230-238, 1997[Medline] 9. Crook JM, Perry GA, Robertson S, et al: Routine prostate biopsies following radiotherapy for prostate cancer: Results for 226 patients. Urology 45: 624-631, 1995[CrossRef][Medline] 10. Dugan TC, Shipley WU, Young RH, et al: Biopsy after external beam radiation therapy for adenocarcinoma of the prostate: Correlation with original histological grade and current prostate specific antigen levels. J Urol 146: 1313-1316, 1991[Medline] 11. Pollack A, Lewis GS, 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 48: 507-512, 2000[CrossRef][Medline]
12. Pollack A, Zagars GK, Lewis GS, et al: Preliminary results of a randomized radiotherapy dose-escalation study comparing 70 Gy with 78 Gy for prostate cancer. J Clin Oncol 18: 3904-3911, 2000 13. 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 41: 491-500, 1998[CrossRef][Medline] 14. Shipley WU, Verhey LJ, Munzenrider JE, et al: Advanced prostate cancer: The results of a randomized comparative trial of high dose irradiation with conventional dose irradiation using photons alone. Int J Radiat Oncol Biol Phys 32: 3-12, 1995[CrossRef][Medline] 15. Pilepich MV, Krall JM, Al-Sarraf M, et al: Androgen deprivation with radiation therapy alone for locally advanced prostatic carcinoma: A randomized comparative trial of the Radiation Therapy Oncology Group. Urology 45: 616-623, 1995[CrossRef][Medline] 16. Laverdiere J, Gomez JL, Cusan L, et al: Beneficial effect of combination therapy administered prior and following external beam radiation in localized prostate cancer. Int J Radiat Oncol Biol Phys 37: 247-252, 1997[CrossRef][Medline] Submitted January 18, 2002; accepted April 23, 2002.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2002 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
