- Email: [email protected]
USE OF SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER: DATA FROM THE caPSURE DATABASE
oozzdslr7M604-13~.~ ’I’m JOVBNALOF U m m Copyright 0 1998 by -CAN Ummrcru. ASS~CUTION,b c .
vol.160. 1398-1404, October 1998 Printed in USA.
USE OF SECOND TREATM~NTFOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER: DATA FROM THE C A P S U R E DATABASE GARY D. GROSSFELD, DAVID M. STIER, SCOlT C. FLANDERS, JAMES M. HENNING, WARREN SCHONFELD, KAREN WAROLIN AND PETER R. CARROLL From the Department of Umlogy and Mount Zion Cancer Center, University of California, San Fmncisco, Technology Assessment Group, San Fmncisco, California and TAP Holdings, Im., Deerfield, Illinois
ABSTRACT
Purpose: We compare secondary cancer treatment use in patients who underwent definitive local treatment for prostate cancer. Materials and Methods: The rate of second cancer treatment was determined in patients who underwent radical prostatectomy (1,254),radiotherapy (499)or cryosurgery (141)using data from the CaPSURE database, a longitudinal disease registry of patients with prostate cancer. Second treatments started within 3 months aRer initial treatment were defined as adjuvant and those started more than 3 months were defined as nonadjuvant. Using a parametric regression model of survival analysis, second treatment rates were adjusted for differences in clinical arid demographic characteristics, and duration of followup among groups. Results: Of the patients 4%received a second adjuvant treatment and 17% received a second nonadjuvant treatment within 3 years of initial therapy. Adjusted rates of nonadjuvant second treatment were lowest after radical prostatectomy, and 34 and 88%higher after radiation and cryosurgery, respectively (p = 0.01).This finding was most evident in patients with pretreatment prostate specific antigen 10.0 ng./ml. or less, clinical stage TBNOMO disease, or Gleason score 6 or less on diagnostic biopsy, and in those classified as low risk for recurrence based on a combination of these parameters (p = 0.004). Conclusions: Approximately 1 in 5 patients receive second cancer treatment within a mean of 3 years following initial local treatment for prostate cancer. Our data suggest that the likelihood of receiving second treatment was lowest in patients initially treated with radical prostatectomy. Key
WORDS:prostatic neoplasms, prostatectomy, radiation, cryosurgery
There is controversy about the best form of treatment for each stage of prostate cancer. There are no prospective randomized trials directly comparing outcomes following the various local treatments and, thus, the relative effectiveness of competing therapies is not well described.’ In addition, comparison of the available retrospective data is complicated by differences in definitions of disease recurrence for various local treatment options. While serum prostate specific antigen (F’SA) thresholds for recurrence are often used to defme relapse after radical prostate~tomy,~~3 appropriate and consistent PSA thresholds to define recurrence after cryosurgerqT4.5 or radiotherapy6-8 have not yet been agreed on or rigidly tested, despite recent recommendations from the American Society for Therapeutic Radiology and Oncology consensus panel.9 Consequently, the use of serum PSA to compare different primary treatments has been problematic, suggesting that alternative measures of outcome may prove useful. The use of second cancer treatment may be an appropriate end point in comparing outcomes following different primary treatments. Additional treatment following definitive local therapy may be given prophylactically to patients at high risk for disease recurrence, that is those with positive surgical margins, high grade tumors or positive lymph nodes, or therapeutically following biochemical or clinical disease recurrence. In either case the frequency of additional treatment may be an indicator of the effectiveness of the primary treatment modality in preventing disease progression or reAccepted for publication April 9, 1998. Supported by a grant from TAP Holdings, Inc.
currence. The use of additional treatment also provides important information on the magnitude of resource use. We describe the use of additional treatment following definitive local therapy for prostate cancer. Insight into this type of resource use should allow us to compare better local treatment options in terms of cost and effectiveness. We used data from the CaF’SURE database, a longitudinal prostate cancer registry, to determine the likelihood, timing and type of additional treatments in patients who underwent radical prostatectomy, radiotherapy or cryosurgery as definitive local therapy for prostate cancer. MATERIALS AND METHODS
CaPSURE project description. CaPSURE is a longitudinal observational database of patients with prostate cancer recruited through a network of urologists at 29 community and academic urology practice sites distributed regionally throughout the United States. At each site patients with biopsy proved adenocarcinoma of the prostate are invited to join the study consecutively as they present for outpatient care. Subjects may be enrolled despite a considerable interval since tumor diagnosis. Median time since diagnosis is 2.0 years. At enrollment the urologist completes an extensive retrospective prostate cancer history based on the existing medical record. Additional data are recorded prospectively during each office visit subsequent to the baseline, including new procedures, treatments and diagnostic tests. Discharge summaries are obtained for all hospitalizations reported, and all
1398
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER
ICD 9 diagnoses and procedures are abstracted. Patients are followed in the database until death or withdrawal from the study. Institutional review board approval is obtained at each clinical site and patient informed consent for participation is required for data collection. Additional details of the project methodology have been published elsewhere.10 Subjects. Between June 1,1995and March 31,1997a total of 3,765 patients were invited to participate in the project and 3,631 agreed (96.4%).Subjects were excluded from our analysis if malignancy was diagnosed before 1990 (393),date of diagnosis was missing (751,no treatment was recorded within 9 months of diagnostic biopsy (411)or initial treatment was hormonal therapy (858)(see figure). Patients diagnosed before 1990 were excluded from study to provide a contemporary cohort for analysis, and those who received neoadjuvant hormonal treatment also were excluded from study. A total of 1,894patients who underwent radical prostatectomy (1,2541,radiotherapy (499)or cryosurgery (141) comprised the population for analysis and were considered to have undergone definitive initial treatment. Of the radiotherapy patients 378 received external beam, 94 brachytherapy, and 6 external beam and brachytherapy. The type of radiation was unspecified for 21 patients. Outcomes measured. The primary end point was second cancer treatment (radiotherapy, cryosurgery, radical prostatectomy, medical hormonal therapy or orchiectomy) anytime following definitive initial treatment. An attempt was made to distinguish secondary treatments in the context of planned adjuvant therapy from those initiated due to physician assessment of treatment failure or disease progression. Study physicians recorded the dates and types of all treatments but not the clinical rationale for such treatments. PSA values were not recorded between diagnosis and enrollment, making serial PSA data incomplete for patients enrolled a considerable time after diagnosis. Thus, it was not possible to determine specifically whether treatments were prescribed as part of an adjuvant therapy program, or in response to the discovery of disease recurrence or progression. Instead, it was hypothesized that planned adjuvant therapies were initiated within 3 months of primary treatment. The risk of second treatment was analyzed in the entire study population and then separately excluding patients (84)who received second treatment within 3 months of primary therapy. Data analysis. Patient demographic and clinical characteristics were analyzed in the 3 initial treatment groups (tables 1 and 2).For each characteristic a chi-square test or analysis of variance was used to determine if there was a statistically
Biopsy dak mining
Di-b
< 1990 F393
during fird 9 mmdhs 62752
Classification of study patients from CaPSuRE database
1399
significant difference among the groups. To determine which group would benefit the most from a particular treatment modality in terms of lower rates of secondary treatment, we stratified subjeds as low, intermediate or high risk for subsequent treatment based on clinical stage, serum PSA at diagnosis and Gleason score on diagnostic biopsy (see Appendix). This stratification parallels those used in other studies reported in the urology and radiation oncology literature.ll-l6 For patients with no second treatment recorded the time at risk for second treatment equaled the time between the first treatment and the earliest of the date lost to followup, study withdrawal, death or March 31, 1997. For patients with second treatment recorded the time at risk for second treatment equaled the time between the first and second treatments. To adjust for differences in duration of followup after initial treatment the rate of second treatment per person-year at risk was used. In each initial treatment group this rate was calculated as the proportion of patients receiving second treatment divided by mean time at risk for second treatment for all patients in the group. These rates were also calculated separately for patients with nonadjuvant second treatments. To control for baseline differences among the treatment groups the crude rates adjusted for duration of followup only were then adjusted for patient demographic and clinical characteristics using computer software.l6 Characteristics showing a statistically significant relationship with the initial treatment group were included in the regression model with the effects due to treatment group. The adjusted rates were tested for statistically significant differences among the initial treatment groups by comparing the estimated treatment effects with the standard errors. To examine differences among patients in the various clinical risk categories and the 3 combined clinical risk strata (see Appendix) the second treatment rate in each definitive initial treatment group was calculated in each risk category. Computer software was used to test the rates in the clinical risk categories for statistically significant differences, to adjust the rates in each combined risk stratum for differences among the initial treatment groups in demographic characteristics and to test these adjusted rates for statistically significant differences. RESULTS
The demographic and clinical characteristics based on initial treatment are summarized in tables 1 and 2, respectively. Overall the 3 treatment groups differed in ethnicity (p = 0.01),type of insurance (p = 0.001)and year of initial diagnosis (p = 0.001).A higher percentage of patients were white (94%)and a lower percentage were black (1.4%)in the cryosurgery group compared to the other 2 treatment groups (table 1).The treatment groups also differed significantly in pretreatment clinical stage (p = 0.001),serum PSA at diagnosis (p = 0.001)and Gleason score on pretreatment biopsy (p = 0.001)(table 2). In general the radiotherapy and cryosurgery groups appeared to have a more advanced clinical stage, higher serum PSA at diagnosis and higher Gleason score on pretreatment biopsy than the radical prostatectomy group. Similarly, a higher percentage of the radiotherapy and cryosurgery groups were classified as high risk compared to the radical prostatectomy group (p = 0.001). Overall second cancer treatments were administered to 411 patients (21.7%).Adjuvant treatments were given to 48 radical prostatectomy (3.8%),29 radiation (5.8%)and 7 cry* surgery (5.0%)patients. Nonadjuvant treatments were given to 169 radical prostatectomy (14.0%). 128 radiation (27.2%) and 30 cryosurgery (22.4%) patients. Crude and adjusted rates of nonadjuvant second treatment per person-year at risk in the 3 treatment groups are s . ed in table 3. Crude rates were adjusted for differences in duration of fol-
1400
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER TAJXX 1. Demographic charnterktics Initial Treatment
Radical Prostatectomy
CharaderiStiC
(1,254 pts.)
p Value (chi-square)
CrYOSurgerY (141 pts.)
71.2 (6.1)
66.0 (6.1)
86.2 (1,078) 10.1 (126) 2.3 (29) 1.4 (17) (4)
90.1 (448) 6.4 (32) 2.2 (11) 1.2 (6)
94.3 (133) 1.4 (2) 2.8 (4) 1.4 (2) (0)
13.9 26.0 19.1 21.2 19.9
(126)
19.3 24.3 19.8 20.1 16.4
30.0
(369)
63.5
Mean Age (SD) 8 Ethnicitp (No. pts.): white Black Hispanic Other Missing data 8 Education (No. pts.): Less than 12 yrs. High school Some college College graduate Graduate school Missing data % Type of insurance (No. pts.): Preferred pmvider organizatiodfee for service Medim Alone With supplement Health maintenance organization Other Missing 8 Yr. initial biopsy/diagnosis (No. pts.): 1990-1991 1992-1993 1994-1995 1996-1997
Irradiation (499 pts.)
(6.8)
(2)
0.0001 (ANOVA) 0.013
0.205 (236) (173) (192) (181) (346)
19.8 (244) 29.9 (368) 18.7 (230) 1.7 (21) 14.1 27.8 39.6 18.4
(73) (92) (75) (76) (62) (121)
14.3 27.6 18.1 26.7 13.3
(15) (29) (19) (28)
(14) (36) 0.001
8.4 (41)
23.4 (33)
(22)
30.2 (148) 45.7 (224) 13.7 (67) 2.0 (10) (9)
31.2 (44) 29.8 (42) 14.9 (21) 0.7 (1) (0)
(177) (349) (497) (231)
21.2 (106) 35.7 (178) 32.5 (162) 10.6 (53)
2.8 (4) 41.1 (58) 44.7 (63) 11.4 (16)
0.001
TABLE3. Crude and adjusted rates of second treatment in patients
TABLE2. Clinical characteristics
with no adjuvant treatment
% InitialTmlbnent (No. pts.)
Initial Treatment
DiaeaMstage: "OM0 TDJOMO T3NOMO T4orN+orMt Missing data InitialPSA at
0.001 18.6 (224) 76.0 (917) 4.2 (50) 1.2 (15) (48)
26.3 (126) 62.6(300) 9.2 (44) 1.9 (9) (20)
31.4 (43) 46.7 (64) 21.2 (29) 0.7 (1) (4) 0.001
Greater than 20.0
Missing data
12.1(147) 57.0 (693) 21.5 (261) 9.5 (115) (38)
8.8 (42) 44.1 (210) 26.7 (127) 20.4 (97) (23)
7 2 (10) 46.8 (65) 31.7 (44) 14.4(20)
6-10
Missing data combinedrish:
Low Moderate
m Missing data
14.0 (169) 27.2 (128) 22.4 (30) Mean p.at risk for second 2.52 2.92 2.12 treatment 0.0557 0.0933 0.1059 Crude rate of second treatment /person-yr. a t risk* (p = 0.0001) Adjusted rate of second 0.0458 0.0615 0.0861 treatment/person-yr. a t riskt (p = 0.011) * Adjusted for duration of followup only. t Adjusted for duration of followup, age, ethnicity, insurance type, year of diagnosis, TNM stage, FSA at diagnosis and Gleason score on biopsy.
0.001
bi0M 5-6 7
Cryosurgery (134 pts.)
(2)
Ghmnsmneon
2-4
Irradiation (470 pts.)
90 Second treatment (No. pts.)
diagnosis: 40orh 4.1-10.0 10.1-20.0
Radical Prostatectomy (1,206 pts,)
17.6 (208) 61.0 (721) 16.4(194) 5.1 (60) (71)
18.6 (84) 49.2 (222) 18.6 (84) 13.5 (61)
(48)
19.6 (27) 47.1 (65) 21.7 (30) 11.6 (16) (3)
10.0 (112) 56.3 (632) 33.8 (379) (131)
7.8 (33) 40.6(172) 51.7 (219) (75)
6.8 (9) 34.9 (46) 58.3 (77) (9)
0.001
lowup only while adjusted rates were adjusted for duration of followup, and clinical and demographic characteristics. The crude rate of nonadjuvant second treatment per person-year at risk was lowest in the radical prostatectomy, intermediate in the radiotherapy and highest in the cryosurgery group. Similar results were obtained when the analysis was performed in patients who underwent any second treatment, adjuvant or nonadjuvant (data not shown). Considering all nonadjuvant second treatments delivered, the crude rates of second treatment per person-year at risk differed significantly among the 3 treatment groups (p = 0.0001, table 3). The crude rate of nonadjuvant second treatment per personyear at risk was significantly - higher - in the radiotheraw -"
(p = 0.0001) and cryosurgery (p = 0.0012) groups compared to the radical prostatectomy group. ARer adjusting for all clinical and demographic variables simultaneously we again noted a statistically significant difference in the rates of second treatment among the 3 treatment groups (p = 0.01, table 3). The radical prostatectomy group had the lowest rate, the radiotherapy group had an intermediate rate and the cryosurgery group had the highest rate of second treatment. The adjusted rate of second treatment was 34 and 88% higher in the radiotherapy (p = 0.06) and cryosurgery (p = 0.004)groups, respectively, compared to the radical prostatectomy group. The time to and type of second treatments in patients who received nonadjuvant second treatment are reported in table 4. Nonadjuvant second treatments following radical prostatectomy were almost evenly divided between radiotherapy (43%) and hormone deprivation treatment (56%).In contrast, nearly all second treatments following radiotherapy (88%) included some form of hormone deprivation. In the cryosurgery group second treatments were divided among radiotherapy (27%), hormone deprivation treatment (40%) and additional cryosurgery (33%). The rates of nonadjuvant second treatment in specific clinical risk categories based on clinical stage, serum PSA or Gleason score alone are reported in table 5. Results were
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER TABLE4. Characteristics Second Treatment
of
1401
second treatment in Datients with nn adiuuant treatment Mean Yrs.Between Initial + Seeond Treatment (40)
Radical Pmstatectomy (169 uts.)
Irradiation (128 D k . )
~
Radical prostatectomy Irradiation Cryosurgery Medical hormonal therapy Orchiectomy
Not available
(0)
1.44 (42.6) 2.58 (1.2) 2.02 (50.3) 2.58 (5.9)
TABLE5. Rate of second treatment among patients with no adjuvant treatment b y initial treatment and clinical characteristics Initial Treatment Clinical characteristic
Radical prostatectomy Irradiation surgery Comparing Valueall 3 (1,206 pts.) (470 pts’) (134 pts.) Treatments
Disease stage: TlNOMO TZNOMO T3NOMO T4 or N+ or M + PSA range: 4.0 or Less
0.0306 0.0512* 0.2087 0.2741
0.0437 0.1024 0.1346 0.6689
0.0881 0.1027 0.1278 0.8613
Not significant
0.0186* 0.0396 0.0708 0.1703
0.0665 0.0425 0.1008 0.1932
0.0894 0.0860 0.1149
10.005 <0.05
4.1-10.0 10.1-20.0
Greater than 20.0 Gleason score: 2 4 5 4
<0.001
Not significant Not significant
Not significant
0.2164 Not significant
C0.005 0.0587 0.0981 <0.005 0.1399 Not significant 7-10 * For radical pmstatectomy versus irradiation p <0.001. t For radical prostatectomy versus irradiation p <0.005. $ For radical prostatectomy versus irradiation p <0.01. 0.0255t 0.0456$ 0.1296
0.0737 0.0746 0,1280
similar when the rates of all second treatments, adjuvant or nonadjuvant, were analyzed in a similar fashion (data not shown). As expected the rate of second treatment increased as disease severity increased, that is with increasing clinical stage, serum PSA or Gleason score. In each clinical stratum second treatment rates were generally lowest in the radical prostatectomy and highest in the cryosurgery group. The rates of second treatment differed significantly among the 3 treatment groups for patients with clinical stage T2NOMO disease, serum PSA 10.0 ngJml. or less, and Gleason score 6 or less. Significant differences were also noted in these strata, except in the serum PSA 4.1 to 10 ngJm1. stratum, between the radical prostatectomy and radiotherapy groups (table 5). Table 6 summarizes the rates of nonadjuvant second treatment in the 3 combined risk categories after adjusting for clinical and demographic factors. As expected these rates were lowest in the low risk and highest in the high risk categories. The initial treatment strategy was associated with significant differences in the rates of additional treatTABLE6 . Adjusted rate of second treatment in patients with no adjuvant treatment by combined risk group Initial Treatment Combined Risk
Radical ~
LOW
.
~
.
0.0034
~
Irradiation (399 o pts.) ~ 0.0318
t
cryosurgery (125 s pts.)~
P value
0.2198*
0.0041
0.0409 0.5259 0.0337 0.0264 Moderate 0.1419 0.1434 0.1380 High 0.1003 Rate is exDressed as number of mtients with second nonadjuvant treatment per person-year at risk. Rates formcderate and high risk & u p are adjusted for patient age, ethnic group, insurance status, year of biopsy and duration of followup. Rates for the low risk group are adjusted for all of these factors except ethnic group because of the small number of nonwhite patients in this mup. * Based on only 9 observations.
1.15 (1.6) 1.79 (3.9) 3.31 (6.3) 2.82 (74.2) 2.81 (14.1)
Not available
(0)
1.78 (26.7) 0.96 (33.3) 1.99 (36.7) 2.46 (3.3)
ment only in the low risk stratum (p = 0.004). In this category second treatment rates were 9-fold higher in the radiotherapy (p = 0.06) and 65-fold higher in the cryosurgery (p = 0.001) groups compared to the radical prostatectomy group. DISCUSSION
To our knowledge few studies to date have addressed the need for second cancer treatment following initial local therapy for prostate cancer. Followup treatment after radical prostatectomy has been examined in 2 previous studies, which reported that 16% of patients received second cancer treatment within 2 years, 22% within 3 years and up to 35% within 5 years of surgery.1.17 Only 1 prior study has addressed the use of second cancer treatment following radiotherapy, demonstrating that 24% of patients received second cancer treatment within an average of 3 years after initial radiotherapy.18 Comparable estimates are not currently available in patients undergoing initial local cryosurgery. We compared secondary treatment following definitive local treatment using data obtained from a multicenter, longitudinal disease registry of patients diagnosed with prostate cancer. Of these patients 21.7% received some form of second cancer treatment within a mean of 3 years after primary definitive local therapy. Rates of secondary treatment were adjusted for differences in followup, demographic characteristics and clinical characteristics among initial treatment groups as these factors may potentially impact the rate of second treatment independent of initial treatment efficacy.1 ARer adjusting for these factors we noted a significant difference in the rate of secondary treatment use among 3 treatment groups. The radical prostatectomy group had the lowest rate of second treatment, particularly in subjects in the low risk category. The type of second cancer treatment also varied among the 3 initial treatment groups. While 2 or 3 types of secondary treatment were commonly used in the radical prostatectomy and cryosurgery groups, nearly all of the radiation group received some form of hormonal deprivation if followup treatment was required. Although followup treatment using surgery, radiation or cryosurgery may often be delivered with curative intent, hormonal deprivation can be considered only as palliative. These data suggest a need to develop alternative, potentially curative treatment options for patients in whom initial treatment with radiation fails. The only previous study that has compared secondary treatment use among patients receiving different initial local treatments for prostate cancer was recently reported by Fowler et a1.18 The rate of second treatment in 621 patients initially treated with radiation was compared to the rate in 373 initially treated with radical prostatectomy from a previously published study.17 The authors noted no significant difference in the rate of secondary treatment use (26% for radical prostatectomy and 24% for radiotherapy patients 3 years after initial therapy). Comparisons between different series such as these, which rely solely on retrospective data and draw subjects from different populations, may not be valid. Unrecognized patient selection factors may have significantly impacted the rates of second treatment use independent of initial treatment effi-
1402
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER
mcy. Patients eligible for inclusion in the 2 study p u p s ering only second treatments beginning at least 3 months reported by Fowler et al were identified using different meth- &,er completion of primary treatment. This decision did reods and treated during different periods. Second treatment duce the rates of second treatment among our 3 treatment rates were based on patient reports only and important clin- groups but did not change the direction of the between group ical characteristics in each treatment group were not speci- differences. An additional bias against radical prostatectomy fied. In contrast, patients in our 3 treatment groups were may emanate from the definition of disease recurrence diagnosed and treated during the same period. We controlled among the 3 local treatment options. PSA recurrence, or PSA for dif€erences in demographic characteristics, clinical char- progression, is more clearly defined for radical prostatectomy acteristics and length of followup in the final analysis, which than radiotherapy or cryosurgery patients. Consequently, a makes comparisons among the 3 groups more reliable. Fi- physician may be more likely to treat on the basis of a nally, second treatment rates in our study were calculated detectable PSA following surgery by recommending second from a comprehensive review of patient reported data, phy- cancer treatment than to react to a slowly rising PSA after radiotherapy or cryosurgery as the PSA threshold for recursician records and hospital discharge summaries. Resource use by patients undergoing definitive local treat- rence is not as well defmed.4-* Conversely, because all parment for prostate cancer has not previously been well stud- ticipating investigators in this study were urologists, initial ied. Reports that have addressed the overall cost of prostate treatment selection decisions may have potentially biased cancer treatment have focused on prostate cancer ~creening’~the results of the study against radiotherapy and cryosuror treatment of advanced disease.20 Secondary treatment use gem. Lastly, our results may be confounded by other variables is a more direct measure of resource use than the rate of biochemical or clinical disease recurrence, and results from not considered in this analysis, such as co-morbidities and our study suggest that it is not an uncommon occurrence. local practice customs. We recognize these limitations and Since the cost of second treatment may be substantial, re- suggest additional research to eliminate these potential efsource use following definitive local treatment may contrib- fects. However, our results did take into account the most ute significantly to the overall cost of care for prostate cancer relevant clinical variables. While these limitations may exist patients. Our data suggest that this effect may be different in the use of observational data, our data provided informain patients initially treated with radical prostatectomy, ra- tion obtained under conditions of usual practice. In this rediotherapy or cryosurgery. However, assessment of total re- gard, secondary treatment use measures a real world outsource use following definitive local treatment was limited come that was similarly defined for all treatment groups and because we analyzed only the rates of second treatment. captures the actual use that was observed as recorded in the Third cancer treatments were administered to 4% of the database and measured by participating physicians. In the absence of prospective randomized trials that diradical prostaktomy, 5% of the radiotherapy and 8%of the cryosurgery groups, while fourth cancer treatments were rectly compare outcomes following local treatment modaliadministered to 2% of each group. Given the increased ex- ties, estimates of treatment effectiveness have been extrappenditures in the last several years,21 we believe that our olated from uncontrolled retrospective reports that examine findings support the need for further investigation into re- a single treatment modality. These retrospective series are source use and cost of prostate cancer treatment among pa- also affected by selection bias and different tumor control end tients undergoing definitive local therapy. points, which make meaningfbl comparisons among different We relied on nonrandomized retrospective outcome data local treatment options difficult.23~ 24 Furthermore, because and, therefore, our results are subject to selection bias, par- many of these studies originate from single center academic ticipation bias and influence of confounding variables.22 Be- institutions, an additional selection bias is imposed which cause patients were invited to participate in the study and may further limit the broad applicability of the re~ults.25~26 could withdraw at any time, there was a potential for selec- Alternative methods of data analysis and study design must tion bias, which may have been particularly important in the be conceived to compare more reliably different treatment radiotherapy group. All patients in the study were registered options. The use of observational data, such as those obtained in the database and followed by urologists. It is possible that from the CaPSURE database, is a viable option for comparsome patients who were successfully treated with radiother- ing local treatment modalities given our current limitations apy and, therefore, did not require a second treatment were in performing randomized trials. Patient data are entered in followed by a radiation oncologist and not a urologist. Con- the database in a consecutive fashion as patients present for versely, patients in whom radiotherapy failed would almost outpatient care with no regard for clinical characteristics and certainly be referred back to a urologist for second cancer no prior knowledge of research hypotheses. Although it was treatment. Thus, the data presented may overrepresent the clear from the current study that clinicians select the initial number of second treatments in the radiotherapy group. treatment modality with reference to clinical and demoHowever, it is also likely that radical prostatectomy and graphic characteristics, we were able to control for known cryosurgery patients were followed exclusively by urologists. confounding elements in our analysis. Thus, second treatment information may be more complete It is also noteworthy that the rates of second treatment for these patients than radiotherapy patients and, therefore, reported in our study apply to patients initially diagnosed bias the results against these 2 forms of initial local treat- and treated for prostate cancer between 1990 and 1995. ment. Although this is a relatively narrow temporal window, treatOur inability to identify the specific indications for addi- ment techniques have continued to evolve with time. Stage tional treatment may have also biased our results against a migrationF7 better patient selection and improvements in particular local treatment modality. For example, secondary technique, that is fewer positive margins after surgery, use of treatments are not only delivered therapeutically in patients 3-dimensional conformal radiation and image guided brachywith clinical or biochemical evidence of disease recurrence therapy,13.28.29may continue to improve treatment efficacy but may also be given prophylactically in those at high risk and lower second treatment rates in all treatment groups. for disease recurrence following initial treatment. The inclu- Furthermore, the use of neoadjuvant treatment before sursion of such adjuvant treatments may bias our results gery or radiotherapy may also have a significant impact on against radical prostatectomy, since early followup treat- re-treatment rates. These issues will need to be addressed ment decisions for these patients are ofken based on adverse with time as patients undergoing more recent initial or neopathological characteristics, which is not the case for radio- adjuvant treatments continue to be followed for disease retherapy or cryosurgery patients as the prostate remains in currence and occurrence of second treatments. situ. We attempted to minimize this potential bias by considBecause the likelihood of receiving additional cancer treat-
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANGER
ment may influence patient treatment decision, re-treatment rates for the various forms of local therapy should be available to patients and physicians. Furthermore, because retreatment rates influence the overall cost of a local treatment modality, this information may assist health care providers in anticipating patterns of resource use. Given the current uncertainty regarding the relative efficacy of local treatment options for prostate cancer and our inability to perform randomized trials directly comparing local treatments, we believe that our data provide important insight into the outcomes of patients initially treated with 3 common local treatment modalities for prostate cancer. CONCLUSIONS
Of the patients in our study 22% received second cancer treatment within a mean of 3 years following initial local treatment for prostate cancer. There were significant differences in the rates of nonadjuvant second treatment per person-year a t risk based on the type of initial treatment. After adjusting for differences in baseline clinical and demographic characteristics, and followup among the 3 initial treatment groups, the radical prostatectomy group had the lowest rates of nonadjuvant second treatment, followed by the radiotherapy (34% higher) and cryosurgery (88%higher) groups. These differences were most pronounced in patients with clinical stage TSNOMO disease, pretreatment serum PSA 10.0 ng./ml. or less and Gleason score 6 or less on diagnostic biopsy, and those classified as low risk for disease recurrence based on a combination of these factors. Given the current impact of prostate cancer treatment on health care resource use, we believe that these data warrant further investigation into the overall cost of definitive local treatment with these commonly used modalities. Jenny Broering collected data, and Saul Kanowitz and David Pasta assisted with data analysis. APPENDIX
Low risk Intermediate risk High risk
PSA less than 5, stage Tl-T2a NOMO and Gleason 6 or less (no Gleason component 4 or greater) All others PSA greater than 15, stage T3 or T4 or N+ or M + and Gleason 7 or greater REFERENCES
1. Lu-Yao, G. L., Potosky, A. L., Albertsen, P. C., Wasson, J. H., Barry, M. J. and Wennberg, J. E.: Follow-up prostate cancer treatments after radical prostatectomy: a population-based study. J. Natl. Cancer Inst., 88: 166, 1996. 2. Catalona, W. J. and Smith, D. S.: 5-Year tumor recurrence rates after anatomical radical retropubic prostatectomy for prostate cancer. J . Urol., 1 5 2 1837, 1994. 3. Walsh, P. C., Partin, A. W. and Epstein, J. I.: Cancer control and quality of life following anatomical radical retropubic prostatectomy: results at 10 years. J . Urol., 152: 1831, 1994. 4. Coogan, C. L. and McKiel, C. F.: Percutaneous cryoablation of the prostate: preliminary results after 95 procedures. J . Urol., 15t: 1813, 1995. 5. Connolly, J. A,, Shinohara, K, Presti, J. C., Jr. and Carroll, P. R.: Prostate-specific antigen after cryosurgical ablation of the prostate. Defining the appropriate response. Urol. Clin. N. h e r . , 24:415, 1997. 6. Critz, F. A., Levinson, K., Williams, W. H., Holladay, D., Holladay, C. and Griffin, V.: Prostate-specific antigen nadir of 0.5 ng./ml. or less defines disease freedom for surgically staged men irradiated for prostate cancer. Urology, 4 9 668, 1997. 7. G i s t , R. W.: Reference range for prostate-specific antigen levels after external beam radiation therapy for adenocarcinoma of the prostate. Urology, 46: 1016, 1995. 8. Lee, W. R., Hanlon, A. L. and Hanks,G. E.: Prostate s m f i c antigen nadir following external beam radiation therapy for
1403
clinically localized prostate cancer: the relationship between nadir level and disease free survival. J. Urol., 156: 450, 1996. 9. Consensus statement: guidelines for FSA following radiation therapy. American Society for Therapeutic Radiology and Oncology Consensus Panel. Int. J. Rad. Oncol. Biol. Phys., 37: 1035, 1997. 10. Lubeck, D. P., Litwin, M. S., Henning, J. M., Stier, D. M., Mazonson, P., Fisk, R. and Carroll, P. R.: The CaPSURE database: a methodology for clinical practice and research in prostate cancer. CaPSURE Research Panel, Cancer of the Prostate Strategic Urologic Research Endeavor. Urology, 48: 773, 1996. 11. Partin, A. W., Kattan, M. W., Subong, E. N., Walsh, P. C., Wojno, K. J., Oesterling, J. E., Scardino, P. T. and Pearson, J. D.: Combination of prostate-specific antigen, clinical stage, and Gleason score to predict pathological stage of localized prostate cancer. A multi-institutional update. J.A.M.A., 277: 1445, 1997. 12. Bauer, J. J., Connelly, R. R., Sesterhenn, I. A., Bettencourt, M. C., McLeod, D. G., Srivastava, S. and Moul, J . W.: Biostatistical modeling using traditional variables and genetic biomarkers for predicting the risk of prostate carcinoma recurrence after radical prostatectomy. Cancer, 7% 952, 1997. 13. Ohori, M., Goad, J . R., Wheeler, T. M., Eastham, J. A,, Thompson, T. C. and Scardino, P. T.: Can radical prostatectomy alter the progression of poorly differentiated prostate cancer? J . Urol., 1 5 2 1843, 1994. 14. Pisansky, T. M., Kahn, M. J., Rasp, G. M., Cha, S. S., Haddock, M. G. and Bostwick, D. G.: A multiple prognostic index predictive of disease outcome after irradiation for clinically localized prostate carcinoma. Cancer, 7 9 337, 1997. 15. Zagars, G. K, Pollack, A., Kavadi, V. S. and von Eschenbach, A. C.: Prostate-specific antigen and radiation therapy for clinically localized prostate cancer. Int. J. Rad. Oncol. Biol. Phys., 3 2 293, 1995. 16. SAS/STAT User's Guide, Version 6, 4th ed. Cary, NC: SAS Institute, Inc., 1990. 17. Fowler, F. J., Jr., Barry, M. J., Lu-Yao, G., Roman, A., Wasson, J. and Wennberg, J. E.: Patient-reported complications and follow-up treatment after radical prostatectomy. The National Medicare Experience: 1988-1990 (updated June 1993). Urology, 4 2 622, 1993. 18. Fowler, F. J., Jr., Barry, M. J., Lu-Yao, G., Wasson, J. H. and Bin, L.: Outcomes of external-beam radiation therapy for prostate cancer: a study of Medicare beneficiaries in three surveillance, epidemiology, and end results areas. J. Clin. Oncol., 1 4 2258, 1996. 19. Albertsen, P. C.: Screening for prostate cancer is neither appropriate nor cost-effective. Urol. Clin. N. h e r . , 23: 521, 1996. 20. Varenhorst, E., Carlsson, P. and Pedersen, K.: Clinical and economic considerations in the treatment of prostate cancer. Pharmacoeconomics, 6: 127, 1994. 21. Litwin, M.: Economic realities of prostate cancer management. Contemp. Urol., 9 43, 1997. 22. DAgostino, R. B. and Kwan, H.: Measuring effectiveness. What to expect without a randomized control group. Med. Care, suppl., 3 3 AS95, 1995. 23. Zietman, A. L., Shipley, W. U. and Coen, J. J.: Radical prostatectomy and radical radiation therapy for clinical stages T1 to 2 adenocarcinoma of the prostate: new insights into outcome from repeat biopsy and prostate specific antigen followup. J. Urol., 1 5 2 1806, 1994. 24. Barry, M. J.: Editorial: Proving early detection and treatment of prostate cancer does more good than harm: the need to support randomized trials actively. J . Urol., 1 5 2 1903, 1994. 25. Gerber, G. S., Thisted, R. A,, Scardino, P. T., Frohmuller, H. G., Schmeder, F. H., Paulson, D. F., Middleton, A. W.,Jr., Rukstalis, D. B., Smith, J . A,, Jr., Schellhammer, P. F., Ohori, M. and Chodak, G. W.: Results of radical prostatectomy in men with clinically localized prostate cancer. J.A.M.A., 276: 615, 1996. 26. Krongrad, A., Lai, H. and Lai, S.: Sunrival aRer radical prostatectomy. J.A.M.A., 278: 44, 1997. 27. Catalona, W. J., Smith, D. S., Ratliff, T. L. and Basler, J. W.: Detection of organ-conlined prostate cancer is increased through prostate-speciiic antigen-based screening. J.A.M.A., 270: 948,1993. 28. Hanks, G. E., Schultheiss, T. E., Hanlon, A. L., Hunt, M.,Lee,
1404
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER
implications of the findings if confirmed by other studies. They also point out a potential major study flaw in patient selection bias. Since escalation study. Int. J. Rad. Oncol. Biol. Phys., 31: 543,1997. all current CaPSURE database physicians are urologists, the study 29. Ragde, H., Blasko, J. C., Grimm, P. D., Kenny, G. M., Sylvester, may overrepresent radiotherapy patients who have recurrence and J. E., Hoak, D. C., Landin, K and Cavanagh, W.: Interstitial are returning to urological care. The authors should consider exiodine-125 radiation without adjuvant therapy in the treat- panding CaPSURE to include radiation oncologists and possibly ment of clinically localized prostate carcinoma. Cancer, 80: medical oncologists to register a broader range of patients. The Department of Defense Congressionally mandated Center for 442. 1997. prostate Disease Research has undertaken a similar database initiative. Unlike CaPSURE the Center for Prostate Disease Research EDITORIAL COMMENT multicenter database is prospective and includes urologists, radiaThe authors developed a longitudinal observational multicenter tion oncologists and medical oncologists a t 10 Army, Navy and & prostate cancer database and provided us with new information Force medical centers. Clinicians and dedicated research data manabout the contemporary care of American prostate cancer patients. agers a t each site use more than a dozen standardized forms and an In 1,894 men treated by multiple providers between 1990 and 1995 Internet linked relational database to collect comprehensive data. To the authors compared the end point of second cancer treatment date data on more than 5,000 patients have been entered and proamong radical prostatectomy (1,254),all types of radiotherapy (499) active followup continues. We hope to confirm soon the work of the and cryotherapy (141) patients with clinically localized prostate can- authors. cer. The most important finding was that even with adjustment for Judd W.Moul age, race, insurance status, year of diagnosis, TNM clinical stage, Center for Prostate Disease Research pretreatment PSA value and biopsy Gleason sum,radical prostatecUniformed Services University tomy patients had a significantly lower rate of second (nonadjuvant) Bethesda, Maryland cancer treatment. The authors point out the huge resource use
W.R, Eptein, B. E. and Coia, L. R.: Optimization of conformal radiation treatment of prostate cancer: report of a dose
vol.160. 1398-1404, October 1998 Printed in USA.
USE OF SECOND TREATM~NTFOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER: DATA FROM THE C A P S U R E DATABASE GARY D. GROSSFELD, DAVID M. STIER, SCOlT C. FLANDERS, JAMES M. HENNING, WARREN SCHONFELD, KAREN WAROLIN AND PETER R. CARROLL From the Department of Umlogy and Mount Zion Cancer Center, University of California, San Fmncisco, Technology Assessment Group, San Fmncisco, California and TAP Holdings, Im., Deerfield, Illinois
ABSTRACT
Purpose: We compare secondary cancer treatment use in patients who underwent definitive local treatment for prostate cancer. Materials and Methods: The rate of second cancer treatment was determined in patients who underwent radical prostatectomy (1,254),radiotherapy (499)or cryosurgery (141)using data from the CaPSURE database, a longitudinal disease registry of patients with prostate cancer. Second treatments started within 3 months aRer initial treatment were defined as adjuvant and those started more than 3 months were defined as nonadjuvant. Using a parametric regression model of survival analysis, second treatment rates were adjusted for differences in clinical arid demographic characteristics, and duration of followup among groups. Results: Of the patients 4%received a second adjuvant treatment and 17% received a second nonadjuvant treatment within 3 years of initial therapy. Adjusted rates of nonadjuvant second treatment were lowest after radical prostatectomy, and 34 and 88%higher after radiation and cryosurgery, respectively (p = 0.01).This finding was most evident in patients with pretreatment prostate specific antigen 10.0 ng./ml. or less, clinical stage TBNOMO disease, or Gleason score 6 or less on diagnostic biopsy, and in those classified as low risk for recurrence based on a combination of these parameters (p = 0.004). Conclusions: Approximately 1 in 5 patients receive second cancer treatment within a mean of 3 years following initial local treatment for prostate cancer. Our data suggest that the likelihood of receiving second treatment was lowest in patients initially treated with radical prostatectomy. Key
WORDS:prostatic neoplasms, prostatectomy, radiation, cryosurgery
There is controversy about the best form of treatment for each stage of prostate cancer. There are no prospective randomized trials directly comparing outcomes following the various local treatments and, thus, the relative effectiveness of competing therapies is not well described.’ In addition, comparison of the available retrospective data is complicated by differences in definitions of disease recurrence for various local treatment options. While serum prostate specific antigen (F’SA) thresholds for recurrence are often used to defme relapse after radical prostate~tomy,~~3 appropriate and consistent PSA thresholds to define recurrence after cryosurgerqT4.5 or radiotherapy6-8 have not yet been agreed on or rigidly tested, despite recent recommendations from the American Society for Therapeutic Radiology and Oncology consensus panel.9 Consequently, the use of serum PSA to compare different primary treatments has been problematic, suggesting that alternative measures of outcome may prove useful. The use of second cancer treatment may be an appropriate end point in comparing outcomes following different primary treatments. Additional treatment following definitive local therapy may be given prophylactically to patients at high risk for disease recurrence, that is those with positive surgical margins, high grade tumors or positive lymph nodes, or therapeutically following biochemical or clinical disease recurrence. In either case the frequency of additional treatment may be an indicator of the effectiveness of the primary treatment modality in preventing disease progression or reAccepted for publication April 9, 1998. Supported by a grant from TAP Holdings, Inc.
currence. The use of additional treatment also provides important information on the magnitude of resource use. We describe the use of additional treatment following definitive local therapy for prostate cancer. Insight into this type of resource use should allow us to compare better local treatment options in terms of cost and effectiveness. We used data from the CaF’SURE database, a longitudinal prostate cancer registry, to determine the likelihood, timing and type of additional treatments in patients who underwent radical prostatectomy, radiotherapy or cryosurgery as definitive local therapy for prostate cancer. MATERIALS AND METHODS
CaPSURE project description. CaPSURE is a longitudinal observational database of patients with prostate cancer recruited through a network of urologists at 29 community and academic urology practice sites distributed regionally throughout the United States. At each site patients with biopsy proved adenocarcinoma of the prostate are invited to join the study consecutively as they present for outpatient care. Subjects may be enrolled despite a considerable interval since tumor diagnosis. Median time since diagnosis is 2.0 years. At enrollment the urologist completes an extensive retrospective prostate cancer history based on the existing medical record. Additional data are recorded prospectively during each office visit subsequent to the baseline, including new procedures, treatments and diagnostic tests. Discharge summaries are obtained for all hospitalizations reported, and all
1398
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER
ICD 9 diagnoses and procedures are abstracted. Patients are followed in the database until death or withdrawal from the study. Institutional review board approval is obtained at each clinical site and patient informed consent for participation is required for data collection. Additional details of the project methodology have been published elsewhere.10 Subjects. Between June 1,1995and March 31,1997a total of 3,765 patients were invited to participate in the project and 3,631 agreed (96.4%).Subjects were excluded from our analysis if malignancy was diagnosed before 1990 (393),date of diagnosis was missing (751,no treatment was recorded within 9 months of diagnostic biopsy (411)or initial treatment was hormonal therapy (858)(see figure). Patients diagnosed before 1990 were excluded from study to provide a contemporary cohort for analysis, and those who received neoadjuvant hormonal treatment also were excluded from study. A total of 1,894patients who underwent radical prostatectomy (1,2541,radiotherapy (499)or cryosurgery (141) comprised the population for analysis and were considered to have undergone definitive initial treatment. Of the radiotherapy patients 378 received external beam, 94 brachytherapy, and 6 external beam and brachytherapy. The type of radiation was unspecified for 21 patients. Outcomes measured. The primary end point was second cancer treatment (radiotherapy, cryosurgery, radical prostatectomy, medical hormonal therapy or orchiectomy) anytime following definitive initial treatment. An attempt was made to distinguish secondary treatments in the context of planned adjuvant therapy from those initiated due to physician assessment of treatment failure or disease progression. Study physicians recorded the dates and types of all treatments but not the clinical rationale for such treatments. PSA values were not recorded between diagnosis and enrollment, making serial PSA data incomplete for patients enrolled a considerable time after diagnosis. Thus, it was not possible to determine specifically whether treatments were prescribed as part of an adjuvant therapy program, or in response to the discovery of disease recurrence or progression. Instead, it was hypothesized that planned adjuvant therapies were initiated within 3 months of primary treatment. The risk of second treatment was analyzed in the entire study population and then separately excluding patients (84)who received second treatment within 3 months of primary therapy. Data analysis. Patient demographic and clinical characteristics were analyzed in the 3 initial treatment groups (tables 1 and 2).For each characteristic a chi-square test or analysis of variance was used to determine if there was a statistically
Biopsy dak mining
Di-b
< 1990 F393
during fird 9 mmdhs 62752
Classification of study patients from CaPSuRE database
1399
significant difference among the groups. To determine which group would benefit the most from a particular treatment modality in terms of lower rates of secondary treatment, we stratified subjeds as low, intermediate or high risk for subsequent treatment based on clinical stage, serum PSA at diagnosis and Gleason score on diagnostic biopsy (see Appendix). This stratification parallels those used in other studies reported in the urology and radiation oncology literature.ll-l6 For patients with no second treatment recorded the time at risk for second treatment equaled the time between the first treatment and the earliest of the date lost to followup, study withdrawal, death or March 31, 1997. For patients with second treatment recorded the time at risk for second treatment equaled the time between the first and second treatments. To adjust for differences in duration of followup after initial treatment the rate of second treatment per person-year at risk was used. In each initial treatment group this rate was calculated as the proportion of patients receiving second treatment divided by mean time at risk for second treatment for all patients in the group. These rates were also calculated separately for patients with nonadjuvant second treatments. To control for baseline differences among the treatment groups the crude rates adjusted for duration of followup only were then adjusted for patient demographic and clinical characteristics using computer software.l6 Characteristics showing a statistically significant relationship with the initial treatment group were included in the regression model with the effects due to treatment group. The adjusted rates were tested for statistically significant differences among the initial treatment groups by comparing the estimated treatment effects with the standard errors. To examine differences among patients in the various clinical risk categories and the 3 combined clinical risk strata (see Appendix) the second treatment rate in each definitive initial treatment group was calculated in each risk category. Computer software was used to test the rates in the clinical risk categories for statistically significant differences, to adjust the rates in each combined risk stratum for differences among the initial treatment groups in demographic characteristics and to test these adjusted rates for statistically significant differences. RESULTS
The demographic and clinical characteristics based on initial treatment are summarized in tables 1 and 2, respectively. Overall the 3 treatment groups differed in ethnicity (p = 0.01),type of insurance (p = 0.001)and year of initial diagnosis (p = 0.001).A higher percentage of patients were white (94%)and a lower percentage were black (1.4%)in the cryosurgery group compared to the other 2 treatment groups (table 1).The treatment groups also differed significantly in pretreatment clinical stage (p = 0.001),serum PSA at diagnosis (p = 0.001)and Gleason score on pretreatment biopsy (p = 0.001)(table 2). In general the radiotherapy and cryosurgery groups appeared to have a more advanced clinical stage, higher serum PSA at diagnosis and higher Gleason score on pretreatment biopsy than the radical prostatectomy group. Similarly, a higher percentage of the radiotherapy and cryosurgery groups were classified as high risk compared to the radical prostatectomy group (p = 0.001). Overall second cancer treatments were administered to 411 patients (21.7%).Adjuvant treatments were given to 48 radical prostatectomy (3.8%),29 radiation (5.8%)and 7 cry* surgery (5.0%)patients. Nonadjuvant treatments were given to 169 radical prostatectomy (14.0%). 128 radiation (27.2%) and 30 cryosurgery (22.4%) patients. Crude and adjusted rates of nonadjuvant second treatment per person-year at risk in the 3 treatment groups are s . ed in table 3. Crude rates were adjusted for differences in duration of fol-
1400
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER TAJXX 1. Demographic charnterktics Initial Treatment
Radical Prostatectomy
CharaderiStiC
(1,254 pts.)
p Value (chi-square)
CrYOSurgerY (141 pts.)
71.2 (6.1)
66.0 (6.1)
86.2 (1,078) 10.1 (126) 2.3 (29) 1.4 (17) (4)
90.1 (448) 6.4 (32) 2.2 (11) 1.2 (6)
94.3 (133) 1.4 (2) 2.8 (4) 1.4 (2) (0)
13.9 26.0 19.1 21.2 19.9
(126)
19.3 24.3 19.8 20.1 16.4
30.0
(369)
63.5
Mean Age (SD) 8 Ethnicitp (No. pts.): white Black Hispanic Other Missing data 8 Education (No. pts.): Less than 12 yrs. High school Some college College graduate Graduate school Missing data % Type of insurance (No. pts.): Preferred pmvider organizatiodfee for service Medim Alone With supplement Health maintenance organization Other Missing 8 Yr. initial biopsy/diagnosis (No. pts.): 1990-1991 1992-1993 1994-1995 1996-1997
Irradiation (499 pts.)
(6.8)
(2)
0.0001 (ANOVA) 0.013
0.205 (236) (173) (192) (181) (346)
19.8 (244) 29.9 (368) 18.7 (230) 1.7 (21) 14.1 27.8 39.6 18.4
(73) (92) (75) (76) (62) (121)
14.3 27.6 18.1 26.7 13.3
(15) (29) (19) (28)
(14) (36) 0.001
8.4 (41)
23.4 (33)
(22)
30.2 (148) 45.7 (224) 13.7 (67) 2.0 (10) (9)
31.2 (44) 29.8 (42) 14.9 (21) 0.7 (1) (0)
(177) (349) (497) (231)
21.2 (106) 35.7 (178) 32.5 (162) 10.6 (53)
2.8 (4) 41.1 (58) 44.7 (63) 11.4 (16)
0.001
TABLE3. Crude and adjusted rates of second treatment in patients
TABLE2. Clinical characteristics
with no adjuvant treatment
% InitialTmlbnent (No. pts.)
Initial Treatment
DiaeaMstage: "OM0 TDJOMO T3NOMO T4orN+orMt Missing data InitialPSA at
0.001 18.6 (224) 76.0 (917) 4.2 (50) 1.2 (15) (48)
26.3 (126) 62.6(300) 9.2 (44) 1.9 (9) (20)
31.4 (43) 46.7 (64) 21.2 (29) 0.7 (1) (4) 0.001
Greater than 20.0
Missing data
12.1(147) 57.0 (693) 21.5 (261) 9.5 (115) (38)
8.8 (42) 44.1 (210) 26.7 (127) 20.4 (97) (23)
7 2 (10) 46.8 (65) 31.7 (44) 14.4(20)
6-10
Missing data combinedrish:
Low Moderate
m Missing data
14.0 (169) 27.2 (128) 22.4 (30) Mean p.at risk for second 2.52 2.92 2.12 treatment 0.0557 0.0933 0.1059 Crude rate of second treatment /person-yr. a t risk* (p = 0.0001) Adjusted rate of second 0.0458 0.0615 0.0861 treatment/person-yr. a t riskt (p = 0.011) * Adjusted for duration of followup only. t Adjusted for duration of followup, age, ethnicity, insurance type, year of diagnosis, TNM stage, FSA at diagnosis and Gleason score on biopsy.
0.001
bi0M 5-6 7
Cryosurgery (134 pts.)
(2)
Ghmnsmneon
2-4
Irradiation (470 pts.)
90 Second treatment (No. pts.)
diagnosis: 40orh 4.1-10.0 10.1-20.0
Radical Prostatectomy (1,206 pts,)
17.6 (208) 61.0 (721) 16.4(194) 5.1 (60) (71)
18.6 (84) 49.2 (222) 18.6 (84) 13.5 (61)
(48)
19.6 (27) 47.1 (65) 21.7 (30) 11.6 (16) (3)
10.0 (112) 56.3 (632) 33.8 (379) (131)
7.8 (33) 40.6(172) 51.7 (219) (75)
6.8 (9) 34.9 (46) 58.3 (77) (9)
0.001
lowup only while adjusted rates were adjusted for duration of followup, and clinical and demographic characteristics. The crude rate of nonadjuvant second treatment per person-year at risk was lowest in the radical prostatectomy, intermediate in the radiotherapy and highest in the cryosurgery group. Similar results were obtained when the analysis was performed in patients who underwent any second treatment, adjuvant or nonadjuvant (data not shown). Considering all nonadjuvant second treatments delivered, the crude rates of second treatment per person-year at risk differed significantly among the 3 treatment groups (p = 0.0001, table 3). The crude rate of nonadjuvant second treatment per personyear at risk was significantly - higher - in the radiotheraw -"
(p = 0.0001) and cryosurgery (p = 0.0012) groups compared to the radical prostatectomy group. ARer adjusting for all clinical and demographic variables simultaneously we again noted a statistically significant difference in the rates of second treatment among the 3 treatment groups (p = 0.01, table 3). The radical prostatectomy group had the lowest rate, the radiotherapy group had an intermediate rate and the cryosurgery group had the highest rate of second treatment. The adjusted rate of second treatment was 34 and 88% higher in the radiotherapy (p = 0.06) and cryosurgery (p = 0.004)groups, respectively, compared to the radical prostatectomy group. The time to and type of second treatments in patients who received nonadjuvant second treatment are reported in table 4. Nonadjuvant second treatments following radical prostatectomy were almost evenly divided between radiotherapy (43%) and hormone deprivation treatment (56%).In contrast, nearly all second treatments following radiotherapy (88%) included some form of hormone deprivation. In the cryosurgery group second treatments were divided among radiotherapy (27%), hormone deprivation treatment (40%) and additional cryosurgery (33%). The rates of nonadjuvant second treatment in specific clinical risk categories based on clinical stage, serum PSA or Gleason score alone are reported in table 5. Results were
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER TABLE4. Characteristics Second Treatment
of
1401
second treatment in Datients with nn adiuuant treatment Mean Yrs.Between Initial + Seeond Treatment (40)
Radical Pmstatectomy (169 uts.)
Irradiation (128 D k . )
~
Radical prostatectomy Irradiation Cryosurgery Medical hormonal therapy Orchiectomy
Not available
(0)
1.44 (42.6) 2.58 (1.2) 2.02 (50.3) 2.58 (5.9)
TABLE5. Rate of second treatment among patients with no adjuvant treatment b y initial treatment and clinical characteristics Initial Treatment Clinical characteristic
Radical prostatectomy Irradiation surgery Comparing Valueall 3 (1,206 pts.) (470 pts’) (134 pts.) Treatments
Disease stage: TlNOMO TZNOMO T3NOMO T4 or N+ or M + PSA range: 4.0 or Less
0.0306 0.0512* 0.2087 0.2741
0.0437 0.1024 0.1346 0.6689
0.0881 0.1027 0.1278 0.8613
Not significant
0.0186* 0.0396 0.0708 0.1703
0.0665 0.0425 0.1008 0.1932
0.0894 0.0860 0.1149
10.005 <0.05
4.1-10.0 10.1-20.0
Greater than 20.0 Gleason score: 2 4 5 4
<0.001
Not significant Not significant
Not significant
0.2164 Not significant
C0.005 0.0587 0.0981 <0.005 0.1399 Not significant 7-10 * For radical pmstatectomy versus irradiation p <0.001. t For radical prostatectomy versus irradiation p <0.005. $ For radical prostatectomy versus irradiation p <0.01. 0.0255t 0.0456$ 0.1296
0.0737 0.0746 0,1280
similar when the rates of all second treatments, adjuvant or nonadjuvant, were analyzed in a similar fashion (data not shown). As expected the rate of second treatment increased as disease severity increased, that is with increasing clinical stage, serum PSA or Gleason score. In each clinical stratum second treatment rates were generally lowest in the radical prostatectomy and highest in the cryosurgery group. The rates of second treatment differed significantly among the 3 treatment groups for patients with clinical stage T2NOMO disease, serum PSA 10.0 ngJml. or less, and Gleason score 6 or less. Significant differences were also noted in these strata, except in the serum PSA 4.1 to 10 ngJm1. stratum, between the radical prostatectomy and radiotherapy groups (table 5). Table 6 summarizes the rates of nonadjuvant second treatment in the 3 combined risk categories after adjusting for clinical and demographic factors. As expected these rates were lowest in the low risk and highest in the high risk categories. The initial treatment strategy was associated with significant differences in the rates of additional treatTABLE6 . Adjusted rate of second treatment in patients with no adjuvant treatment by combined risk group Initial Treatment Combined Risk
Radical ~
LOW
.
~
.
0.0034
~
Irradiation (399 o pts.) ~ 0.0318
t
cryosurgery (125 s pts.)~
P value
0.2198*
0.0041
0.0409 0.5259 0.0337 0.0264 Moderate 0.1419 0.1434 0.1380 High 0.1003 Rate is exDressed as number of mtients with second nonadjuvant treatment per person-year at risk. Rates formcderate and high risk & u p are adjusted for patient age, ethnic group, insurance status, year of biopsy and duration of followup. Rates for the low risk group are adjusted for all of these factors except ethnic group because of the small number of nonwhite patients in this mup. * Based on only 9 observations.
1.15 (1.6) 1.79 (3.9) 3.31 (6.3) 2.82 (74.2) 2.81 (14.1)
Not available
(0)
1.78 (26.7) 0.96 (33.3) 1.99 (36.7) 2.46 (3.3)
ment only in the low risk stratum (p = 0.004). In this category second treatment rates were 9-fold higher in the radiotherapy (p = 0.06) and 65-fold higher in the cryosurgery (p = 0.001) groups compared to the radical prostatectomy group. DISCUSSION
To our knowledge few studies to date have addressed the need for second cancer treatment following initial local therapy for prostate cancer. Followup treatment after radical prostatectomy has been examined in 2 previous studies, which reported that 16% of patients received second cancer treatment within 2 years, 22% within 3 years and up to 35% within 5 years of surgery.1.17 Only 1 prior study has addressed the use of second cancer treatment following radiotherapy, demonstrating that 24% of patients received second cancer treatment within an average of 3 years after initial radiotherapy.18 Comparable estimates are not currently available in patients undergoing initial local cryosurgery. We compared secondary treatment following definitive local treatment using data obtained from a multicenter, longitudinal disease registry of patients diagnosed with prostate cancer. Of these patients 21.7% received some form of second cancer treatment within a mean of 3 years after primary definitive local therapy. Rates of secondary treatment were adjusted for differences in followup, demographic characteristics and clinical characteristics among initial treatment groups as these factors may potentially impact the rate of second treatment independent of initial treatment efficacy.1 ARer adjusting for these factors we noted a significant difference in the rate of secondary treatment use among 3 treatment groups. The radical prostatectomy group had the lowest rate of second treatment, particularly in subjects in the low risk category. The type of second cancer treatment also varied among the 3 initial treatment groups. While 2 or 3 types of secondary treatment were commonly used in the radical prostatectomy and cryosurgery groups, nearly all of the radiation group received some form of hormonal deprivation if followup treatment was required. Although followup treatment using surgery, radiation or cryosurgery may often be delivered with curative intent, hormonal deprivation can be considered only as palliative. These data suggest a need to develop alternative, potentially curative treatment options for patients in whom initial treatment with radiation fails. The only previous study that has compared secondary treatment use among patients receiving different initial local treatments for prostate cancer was recently reported by Fowler et a1.18 The rate of second treatment in 621 patients initially treated with radiation was compared to the rate in 373 initially treated with radical prostatectomy from a previously published study.17 The authors noted no significant difference in the rate of secondary treatment use (26% for radical prostatectomy and 24% for radiotherapy patients 3 years after initial therapy). Comparisons between different series such as these, which rely solely on retrospective data and draw subjects from different populations, may not be valid. Unrecognized patient selection factors may have significantly impacted the rates of second treatment use independent of initial treatment effi-
1402
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER
mcy. Patients eligible for inclusion in the 2 study p u p s ering only second treatments beginning at least 3 months reported by Fowler et al were identified using different meth- &,er completion of primary treatment. This decision did reods and treated during different periods. Second treatment duce the rates of second treatment among our 3 treatment rates were based on patient reports only and important clin- groups but did not change the direction of the between group ical characteristics in each treatment group were not speci- differences. An additional bias against radical prostatectomy fied. In contrast, patients in our 3 treatment groups were may emanate from the definition of disease recurrence diagnosed and treated during the same period. We controlled among the 3 local treatment options. PSA recurrence, or PSA for dif€erences in demographic characteristics, clinical char- progression, is more clearly defined for radical prostatectomy acteristics and length of followup in the final analysis, which than radiotherapy or cryosurgery patients. Consequently, a makes comparisons among the 3 groups more reliable. Fi- physician may be more likely to treat on the basis of a nally, second treatment rates in our study were calculated detectable PSA following surgery by recommending second from a comprehensive review of patient reported data, phy- cancer treatment than to react to a slowly rising PSA after radiotherapy or cryosurgery as the PSA threshold for recursician records and hospital discharge summaries. Resource use by patients undergoing definitive local treat- rence is not as well defmed.4-* Conversely, because all parment for prostate cancer has not previously been well stud- ticipating investigators in this study were urologists, initial ied. Reports that have addressed the overall cost of prostate treatment selection decisions may have potentially biased cancer treatment have focused on prostate cancer ~creening’~the results of the study against radiotherapy and cryosuror treatment of advanced disease.20 Secondary treatment use gem. Lastly, our results may be confounded by other variables is a more direct measure of resource use than the rate of biochemical or clinical disease recurrence, and results from not considered in this analysis, such as co-morbidities and our study suggest that it is not an uncommon occurrence. local practice customs. We recognize these limitations and Since the cost of second treatment may be substantial, re- suggest additional research to eliminate these potential efsource use following definitive local treatment may contrib- fects. However, our results did take into account the most ute significantly to the overall cost of care for prostate cancer relevant clinical variables. While these limitations may exist patients. Our data suggest that this effect may be different in the use of observational data, our data provided informain patients initially treated with radical prostatectomy, ra- tion obtained under conditions of usual practice. In this rediotherapy or cryosurgery. However, assessment of total re- gard, secondary treatment use measures a real world outsource use following definitive local treatment was limited come that was similarly defined for all treatment groups and because we analyzed only the rates of second treatment. captures the actual use that was observed as recorded in the Third cancer treatments were administered to 4% of the database and measured by participating physicians. In the absence of prospective randomized trials that diradical prostaktomy, 5% of the radiotherapy and 8%of the cryosurgery groups, while fourth cancer treatments were rectly compare outcomes following local treatment modaliadministered to 2% of each group. Given the increased ex- ties, estimates of treatment effectiveness have been extrappenditures in the last several years,21 we believe that our olated from uncontrolled retrospective reports that examine findings support the need for further investigation into re- a single treatment modality. These retrospective series are source use and cost of prostate cancer treatment among pa- also affected by selection bias and different tumor control end tients undergoing definitive local therapy. points, which make meaningfbl comparisons among different We relied on nonrandomized retrospective outcome data local treatment options difficult.23~ 24 Furthermore, because and, therefore, our results are subject to selection bias, par- many of these studies originate from single center academic ticipation bias and influence of confounding variables.22 Be- institutions, an additional selection bias is imposed which cause patients were invited to participate in the study and may further limit the broad applicability of the re~ults.25~26 could withdraw at any time, there was a potential for selec- Alternative methods of data analysis and study design must tion bias, which may have been particularly important in the be conceived to compare more reliably different treatment radiotherapy group. All patients in the study were registered options. The use of observational data, such as those obtained in the database and followed by urologists. It is possible that from the CaPSURE database, is a viable option for comparsome patients who were successfully treated with radiother- ing local treatment modalities given our current limitations apy and, therefore, did not require a second treatment were in performing randomized trials. Patient data are entered in followed by a radiation oncologist and not a urologist. Con- the database in a consecutive fashion as patients present for versely, patients in whom radiotherapy failed would almost outpatient care with no regard for clinical characteristics and certainly be referred back to a urologist for second cancer no prior knowledge of research hypotheses. Although it was treatment. Thus, the data presented may overrepresent the clear from the current study that clinicians select the initial number of second treatments in the radiotherapy group. treatment modality with reference to clinical and demoHowever, it is also likely that radical prostatectomy and graphic characteristics, we were able to control for known cryosurgery patients were followed exclusively by urologists. confounding elements in our analysis. Thus, second treatment information may be more complete It is also noteworthy that the rates of second treatment for these patients than radiotherapy patients and, therefore, reported in our study apply to patients initially diagnosed bias the results against these 2 forms of initial local treat- and treated for prostate cancer between 1990 and 1995. ment. Although this is a relatively narrow temporal window, treatOur inability to identify the specific indications for addi- ment techniques have continued to evolve with time. Stage tional treatment may have also biased our results against a migrationF7 better patient selection and improvements in particular local treatment modality. For example, secondary technique, that is fewer positive margins after surgery, use of treatments are not only delivered therapeutically in patients 3-dimensional conformal radiation and image guided brachywith clinical or biochemical evidence of disease recurrence therapy,13.28.29may continue to improve treatment efficacy but may also be given prophylactically in those at high risk and lower second treatment rates in all treatment groups. for disease recurrence following initial treatment. The inclu- Furthermore, the use of neoadjuvant treatment before sursion of such adjuvant treatments may bias our results gery or radiotherapy may also have a significant impact on against radical prostatectomy, since early followup treat- re-treatment rates. These issues will need to be addressed ment decisions for these patients are ofken based on adverse with time as patients undergoing more recent initial or neopathological characteristics, which is not the case for radio- adjuvant treatments continue to be followed for disease retherapy or cryosurgery patients as the prostate remains in currence and occurrence of second treatments. situ. We attempted to minimize this potential bias by considBecause the likelihood of receiving additional cancer treat-
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANGER
ment may influence patient treatment decision, re-treatment rates for the various forms of local therapy should be available to patients and physicians. Furthermore, because retreatment rates influence the overall cost of a local treatment modality, this information may assist health care providers in anticipating patterns of resource use. Given the current uncertainty regarding the relative efficacy of local treatment options for prostate cancer and our inability to perform randomized trials directly comparing local treatments, we believe that our data provide important insight into the outcomes of patients initially treated with 3 common local treatment modalities for prostate cancer. CONCLUSIONS
Of the patients in our study 22% received second cancer treatment within a mean of 3 years following initial local treatment for prostate cancer. There were significant differences in the rates of nonadjuvant second treatment per person-year a t risk based on the type of initial treatment. After adjusting for differences in baseline clinical and demographic characteristics, and followup among the 3 initial treatment groups, the radical prostatectomy group had the lowest rates of nonadjuvant second treatment, followed by the radiotherapy (34% higher) and cryosurgery (88%higher) groups. These differences were most pronounced in patients with clinical stage TSNOMO disease, pretreatment serum PSA 10.0 ng./ml. or less and Gleason score 6 or less on diagnostic biopsy, and those classified as low risk for disease recurrence based on a combination of these factors. Given the current impact of prostate cancer treatment on health care resource use, we believe that these data warrant further investigation into the overall cost of definitive local treatment with these commonly used modalities. Jenny Broering collected data, and Saul Kanowitz and David Pasta assisted with data analysis. APPENDIX
Low risk Intermediate risk High risk
PSA less than 5, stage Tl-T2a NOMO and Gleason 6 or less (no Gleason component 4 or greater) All others PSA greater than 15, stage T3 or T4 or N+ or M + and Gleason 7 or greater REFERENCES
1. Lu-Yao, G. L., Potosky, A. L., Albertsen, P. C., Wasson, J. H., Barry, M. J. and Wennberg, J. E.: Follow-up prostate cancer treatments after radical prostatectomy: a population-based study. J. Natl. Cancer Inst., 88: 166, 1996. 2. Catalona, W. J. and Smith, D. S.: 5-Year tumor recurrence rates after anatomical radical retropubic prostatectomy for prostate cancer. J . Urol., 1 5 2 1837, 1994. 3. Walsh, P. C., Partin, A. W. and Epstein, J. I.: Cancer control and quality of life following anatomical radical retropubic prostatectomy: results at 10 years. J . Urol., 152: 1831, 1994. 4. Coogan, C. L. and McKiel, C. F.: Percutaneous cryoablation of the prostate: preliminary results after 95 procedures. J . Urol., 15t: 1813, 1995. 5. Connolly, J. A,, Shinohara, K, Presti, J. C., Jr. and Carroll, P. R.: Prostate-specific antigen after cryosurgical ablation of the prostate. Defining the appropriate response. Urol. Clin. N. h e r . , 24:415, 1997. 6. Critz, F. A., Levinson, K., Williams, W. H., Holladay, D., Holladay, C. and Griffin, V.: Prostate-specific antigen nadir of 0.5 ng./ml. or less defines disease freedom for surgically staged men irradiated for prostate cancer. Urology, 4 9 668, 1997. 7. G i s t , R. W.: Reference range for prostate-specific antigen levels after external beam radiation therapy for adenocarcinoma of the prostate. Urology, 46: 1016, 1995. 8. Lee, W. R., Hanlon, A. L. and Hanks,G. E.: Prostate s m f i c antigen nadir following external beam radiation therapy for
1403
clinically localized prostate cancer: the relationship between nadir level and disease free survival. J. Urol., 156: 450, 1996. 9. Consensus statement: guidelines for FSA following radiation therapy. American Society for Therapeutic Radiology and Oncology Consensus Panel. Int. J. Rad. Oncol. Biol. Phys., 37: 1035, 1997. 10. Lubeck, D. P., Litwin, M. S., Henning, J. M., Stier, D. M., Mazonson, P., Fisk, R. and Carroll, P. R.: The CaPSURE database: a methodology for clinical practice and research in prostate cancer. CaPSURE Research Panel, Cancer of the Prostate Strategic Urologic Research Endeavor. Urology, 48: 773, 1996. 11. Partin, A. W., Kattan, M. W., Subong, E. N., Walsh, P. C., Wojno, K. J., Oesterling, J. E., Scardino, P. T. and Pearson, J. D.: Combination of prostate-specific antigen, clinical stage, and Gleason score to predict pathological stage of localized prostate cancer. A multi-institutional update. J.A.M.A., 277: 1445, 1997. 12. Bauer, J. J., Connelly, R. R., Sesterhenn, I. A., Bettencourt, M. C., McLeod, D. G., Srivastava, S. and Moul, J . W.: Biostatistical modeling using traditional variables and genetic biomarkers for predicting the risk of prostate carcinoma recurrence after radical prostatectomy. Cancer, 7% 952, 1997. 13. Ohori, M., Goad, J . R., Wheeler, T. M., Eastham, J. A,, Thompson, T. C. and Scardino, P. T.: Can radical prostatectomy alter the progression of poorly differentiated prostate cancer? J . Urol., 1 5 2 1843, 1994. 14. Pisansky, T. M., Kahn, M. J., Rasp, G. M., Cha, S. S., Haddock, M. G. and Bostwick, D. G.: A multiple prognostic index predictive of disease outcome after irradiation for clinically localized prostate carcinoma. Cancer, 7 9 337, 1997. 15. Zagars, G. K, Pollack, A., Kavadi, V. S. and von Eschenbach, A. C.: Prostate-specific antigen and radiation therapy for clinically localized prostate cancer. Int. J. Rad. Oncol. Biol. Phys., 3 2 293, 1995. 16. SAS/STAT User's Guide, Version 6, 4th ed. Cary, NC: SAS Institute, Inc., 1990. 17. Fowler, F. J., Jr., Barry, M. J., Lu-Yao, G., Roman, A., Wasson, J. and Wennberg, J. E.: Patient-reported complications and follow-up treatment after radical prostatectomy. The National Medicare Experience: 1988-1990 (updated June 1993). Urology, 4 2 622, 1993. 18. Fowler, F. J., Jr., Barry, M. J., Lu-Yao, G., Wasson, J. H. and Bin, L.: Outcomes of external-beam radiation therapy for prostate cancer: a study of Medicare beneficiaries in three surveillance, epidemiology, and end results areas. J. Clin. Oncol., 1 4 2258, 1996. 19. Albertsen, P. C.: Screening for prostate cancer is neither appropriate nor cost-effective. Urol. Clin. N. h e r . , 23: 521, 1996. 20. Varenhorst, E., Carlsson, P. and Pedersen, K.: Clinical and economic considerations in the treatment of prostate cancer. Pharmacoeconomics, 6: 127, 1994. 21. Litwin, M.: Economic realities of prostate cancer management. Contemp. Urol., 9 43, 1997. 22. DAgostino, R. B. and Kwan, H.: Measuring effectiveness. What to expect without a randomized control group. Med. Care, suppl., 3 3 AS95, 1995. 23. Zietman, A. L., Shipley, W. U. and Coen, J. J.: Radical prostatectomy and radical radiation therapy for clinical stages T1 to 2 adenocarcinoma of the prostate: new insights into outcome from repeat biopsy and prostate specific antigen followup. J. Urol., 1 5 2 1806, 1994. 24. Barry, M. J.: Editorial: Proving early detection and treatment of prostate cancer does more good than harm: the need to support randomized trials actively. J . Urol., 1 5 2 1903, 1994. 25. Gerber, G. S., Thisted, R. A,, Scardino, P. T., Frohmuller, H. G., Schmeder, F. H., Paulson, D. F., Middleton, A. W.,Jr., Rukstalis, D. B., Smith, J . A,, Jr., Schellhammer, P. F., Ohori, M. and Chodak, G. W.: Results of radical prostatectomy in men with clinically localized prostate cancer. J.A.M.A., 276: 615, 1996. 26. Krongrad, A., Lai, H. and Lai, S.: Sunrival aRer radical prostatectomy. J.A.M.A., 278: 44, 1997. 27. Catalona, W. J., Smith, D. S., Ratliff, T. L. and Basler, J. W.: Detection of organ-conlined prostate cancer is increased through prostate-speciiic antigen-based screening. J.A.M.A., 270: 948,1993. 28. Hanks, G. E., Schultheiss, T. E., Hanlon, A. L., Hunt, M.,Lee,
1404
SECOND TREATMENT FOLLOWING DEFINITIVE LOCAL THERAPY FOR PROSTATE CANCER
implications of the findings if confirmed by other studies. They also point out a potential major study flaw in patient selection bias. Since escalation study. Int. J. Rad. Oncol. Biol. Phys., 31: 543,1997. all current CaPSURE database physicians are urologists, the study 29. Ragde, H., Blasko, J. C., Grimm, P. D., Kenny, G. M., Sylvester, may overrepresent radiotherapy patients who have recurrence and J. E., Hoak, D. C., Landin, K and Cavanagh, W.: Interstitial are returning to urological care. The authors should consider exiodine-125 radiation without adjuvant therapy in the treat- panding CaPSURE to include radiation oncologists and possibly ment of clinically localized prostate carcinoma. Cancer, 80: medical oncologists to register a broader range of patients. The Department of Defense Congressionally mandated Center for 442. 1997. prostate Disease Research has undertaken a similar database initiative. Unlike CaPSURE the Center for Prostate Disease Research EDITORIAL COMMENT multicenter database is prospective and includes urologists, radiaThe authors developed a longitudinal observational multicenter tion oncologists and medical oncologists a t 10 Army, Navy and & prostate cancer database and provided us with new information Force medical centers. Clinicians and dedicated research data manabout the contemporary care of American prostate cancer patients. agers a t each site use more than a dozen standardized forms and an In 1,894 men treated by multiple providers between 1990 and 1995 Internet linked relational database to collect comprehensive data. To the authors compared the end point of second cancer treatment date data on more than 5,000 patients have been entered and proamong radical prostatectomy (1,254),all types of radiotherapy (499) active followup continues. We hope to confirm soon the work of the and cryotherapy (141) patients with clinically localized prostate can- authors. cer. The most important finding was that even with adjustment for Judd W.Moul age, race, insurance status, year of diagnosis, TNM clinical stage, Center for Prostate Disease Research pretreatment PSA value and biopsy Gleason sum,radical prostatecUniformed Services University tomy patients had a significantly lower rate of second (nonadjuvant) Bethesda, Maryland cancer treatment. The authors point out the huge resource use
W.R, Eptein, B. E. and Coia, L. R.: Optimization of conformal radiation treatment of prostate cancer: report of a dose