0022-5347/05/1736-1947/0
`THE JOURNAL OF UROLOGY®
`Copyright © 2005 by AMERICAN UROLOGICAL ASSOCIATION
`
`Vol. 173, 1947–1952, June 2005
`Printed in U.S.A.
`DOI: 10.1097/01.ju.0000158449.83022.40
`
`LONG-TERM OUTCOME FOR MEN WITH ANDROGEN INDEPENDENT
`PROSTATE CANCER TREATED WITH KETOCONAZOLE AND
`HYDROCORTISONE
`
`MARK SCHOLZ,* ROBERT JENNRICH, STEPHEN STRUM, STANLEY BROSMAN, HENRY JOHNSON
`AND RICHARD LAM
`From Prostate Oncology Specialists, Marina del Rey (MS, SB, HJ, RL), and the Department of Statistics, University of California, Los
`Angeles, Los Angeles (RJ), California
`
`ABSTRACT
`Purpose: The combination of high dose ketoconazole and hydrocortisone (HDK) is active
`against androgen independent prostate cancer (AIPC). Median response times with HDK tend to
`be brief but a significant minority of AIPC patients benefit with extended responses. Well
`characterized response and survival information, especially in the cohort of patients who expe-
`rience these longer, more durable, responses has not been previously reported. Characterization
`of this subgroup is of particular interest since men with long-term responses derive the greatest
`benefit from HDK therapy.
`Materials and Methods: The medical records of 78 patients with AIPC treated with HDK
`between March 1991 and February 1999 were retrospectively reviewed. Baseline clinical and
`laboratory factors predictive of prolonged response and survival were identified.
`Results: The median baseline prostate specific antigen (PSA) before the initiation of HDK was
`25.1. The number of patients with zero, 1 to 3, and more than 3 lesions on bone scan were 25, 35
`and 18, respectively. Median and mean time to PSA progression was 6.7 and 14.5 months.
`Median and mean survival time was 38.0 and 42.4 months, respectively. Response time and
`survival were highly correlated (r ⫽ 0.799). A total of 34 (44%) men had a greater than 75%
`decrease in PSA. The median survival times in men with more vs less than a 75% decrease were
`60 vs 24 months, respectively. In a Cox proportional hazard regression, prolonged survival was
`predicted by percent PSA decrease, extent of disease on bone scan and baseline PSA.
`Conclusions: Ketoconazole can induce prolonged responses, occasionally lasting for years. Long
`responses are more likely to occur in men initiating HDK earlier in the course of disease before
`the cancer burden becomes excessive.
`KEY WORDS: ketoconazole, prostatic neoplasms, prostate-specific antigen, survival
`
`High dose ketoconazole in combination with hydrocorti-
`sone (HDK) is an active regimen in the treatment of andro-
`gen independent prostate cancer (AIPC). Studies demon-
`strating HDK efficacy in AIPC date back to the pre-prostate
`specific antigen (PSA) era.1, 2 In those early years, HDK was
`also used in the treatment of patients with hormone sensitive
`disease.3, 4 After the discovery of PSA, trials incorporating
`PSA as a response indicator to assess HDK in AIPC con-
`firmed high PSA response rates, although the reported me-
`dian response times were generally short.5, 6 These studies
`also identified a subset of patients who experienced more
`extended responses. Survival information, especially in the
`cohort of patients demonstrating more durable responses has
`not been reported. We retrospectively reviewed 78 patients
`with AIPC treated with HDK from a medical oncology prac-
`tice specializing in prostate cancer attempting to identify a
`cohort of men who would derive the most benefit from HDK.
`
`MATERIALS AND METHODS
`Patient selection for inclusion. All patient charts at Pros-
`tate Oncology Specialists, a medical oncology practice spe-
`cializing in prostate cancer, were reviewed for study inclu-
`sion. There were 5 criteria used for selection. 1) Hormone
`refractory prostate cancer as defined by 3 consecutive in-
`Submitted for publication August 18, 2004.
`* Correspondence: 4676 Admiralty Way, Suite 101, Marina del
`Rey, California 90292 (telephone: 310-827-7707; FAX: 310-574-4002;
`e-mail: mark@prostateoncology.com).
`
`creases in PSA despite castrate testosterone levels and pre-
`vious therapy with at least 1 antiandrogen. 2) HDK admin-
`istered for a minimum of 2 months. Patients receiving HDK
`for less than 2 months and who were taken off therapy
`because of HDK induced toxicity were included in the study
`and tabulated as nonresponders failing at 2 months. 3) No
`additional anticancer agents other than hydrocortisone ad-
`ministered concomitantly. 4) Adequate clinical and labora-
`tory information available for review. 5) To ensure adequate
`followup only patients started on HDK before February 1999
`were included.
`Treatment schema. High dose ketoconazole was initiated at
`a dose of 200 mg every 8 hours on an empty stomach. If no
`toxicity was encountered during the first 7 days of treatment
`the dose was increased to 400 mg every 8 hours. Patients
`unable to forego antacid medication were not treated with
`HDK. Patients were routinely administered hydrocortisone
`at a starting dose of 20 mg twice a day with food. In most
`patients, the hydrocortisone dose was decreased to 20 mg
`with breakfast and 10 mg with dinner after 1 month of HDK
`therapy. In a few patients the standard dose of HDK of 1,200
`mg a day was not tolerated. In such patients a dose reduction
`to the starting dose of 200 mg 3 times a day permitted them
`to continue on HDK therapy and remain on study.
`Clinical monitoring schema. Since this was a retrospective
`analysis, there were no specific study criteria. However, all
`patients were treated at 1 facility in a uniform manner.
`Patients underwent baseline bone scans that were repeated
`1947
`
`WCK1016
`Page 1
`
`

`
`1948
`
`MANAGEMENT OF ANDROGEN INDEPENDENT PROSTATE CANCER WITH KETOCONAZOLE
`
`Variable
`
`Age start HDK
`HB Rx mos
`Dx to Rx mos
`Baseline PSA
`Baseline ALP
`Baseline Hct
`Response time (mos)
`PSA % decrease
`Survival (mos)
`
`TABLE 1. Descriptive statistics for 78 men treated with HDK
`
`Mean
`
`69.5
`36.6
`62.1
`121.9
`126.5
`36.9
`14.5
`56
`42.4
`
`Median
`
`70.4
`28.5
`57.2
`25.1
`85.5
`37.2
`6.7
`69
`38
`
`Min
`
`50.4
`3
`8.6
`0.06
`29
`21
`1
`0
`4
`
`Max
`
`87.8
`129
`167.1
`1,867
`761
`44.3
`91
`100
`116
`
`Upper
`Quartile
`74.4
`57
`84.2
`97
`127
`39.2
`17.0
`93
`61
`
`every 6 to 12 months and at the time of PSA progression.
`Patients with measurable adenopathy on computerized to-
`mography had repeat scans periodically and at the time of
`disease progression. PSA levels were relied upon heavily as
`the main clinical indicator of response or progression and as
`such were generally checked monthly. Patients with re-
`sponses greater than 12 months had PSA levels drawn at 2 to
`3 months intervals. Hepatic panels and chemistry panels
`were initially checked monthly because of HDK’s known po-
`tential for hepatotoxicity. Physical examination was done at
`baseline and repeated every 6 months. The majority of pa-
`tients had Gleason scores reviewed by 1 of a small group of
`nationally known pathologists. No patient developed clinical
`progression (changes in bone scan or computerized tomogra-
`phy) without first manifesting progression by PSA criteria.
`PSA decreases of 50% and 75% were only considered valid if
`maintained for at least 4 weeks.7
`Study definitions. Baseline is a laboratory or clinical find-
`ing which immediately preceded the initiation of HDK. Re-
`sponse Time is the time from the initiation of HDK until the
`first of 2 consecutive PSA values that are 50% and at least 2
`ng/dl greater than the nadir PSA value.8 Progression, time to
`progression, or time on HDK as used in the paper are all
`synonymous with Response Time. Survival time is the time
`from initiation of HDK to the time of death from prostate
`cancer. Hormone blockade treatment time (HB Rx Time) is
`the time from starting hormone blockade to the time AIPC is
`diagnosed using the criteria previously described. PC diag-
`nosis to HDK Time (Dx to Rx) is the time from PC diagnosis
`to the time of starting HDK. Bone scan stage is defined as
`stage 1 —no bone metastasis, stage 2—1 to 3 bone metasta-
`ses and stage 3—4 or more bone metastases. Percent PSA
`decrease is defined as the percent decrease in PSA from
`baseline to nadir.
`Statistical issues. All statistical calculations were per-
`formed on Statistica™. The survival variable is censored,
`
`TABLE 2. Characteristics of 78 men treated with HDK
`No. Men
`25
`35
`18
`3
`62
`5
`
`Neg bone scan
`1–3 Lesions on bone scan
`Greater than 3 lesions on bone scan
`Still on treatment
`Stopped responding to HDK
`Stopped HDK for toxicity while respond-
`ing
`Lost to followup while responding
`Treatment changed while responding
`Died of unrelated causes while responding
`Died of unrelated causes after progression
`Died of prostate Ca
`
`3
`1
`4
`1
`61
`
`that is, it contains survival times and censored survival
`times. As a consequence its mean, median, and correlations
`are somewhat biased estimates of population values. These
`estimates are used only for rough descriptive and exploratory
`purposes. Kaplan-Meier survival curves and Cox regressions
`are used to deal with the biases resulting from censoring. The
`Kaplan-Meier curves are used primarily to compare survival
`in subpopulations defined by specified ranges of a single
`variable. Cox regression is used to adjust survival compari-
`sons for potential confounders. Step-down Cox regression is
`used to find significant risk factors and identify potential
`confounders.
`For the survival time, patients suffering a nonprostate
`cancer death, patients lost to followup, patients who stopped
`HDK for reasons of toxicity, and patients alive at the end of
`the study were censored. The same criteria were used for
`response time end point except patients alive at the end of
`the study is replaced by patients responding at the end of the
`study.
`
`FIG. 1. Kaplan-Meier representation of response time for all 78 patients
`
`WCK1016
`Page 2
`
`

`
`MANAGEMENT OF ANDROGEN INDEPENDENT PROSTATE CANCER WITH KETOCONAZOLE
`
`1949
`
`FIG. 2. Kaplan-Meier representation of survival for all 78 patients
`
`FIG. 3. Months survival vs months response time
`
`RESULTS
`In a chart review of 1,200 patients, 78 patients met the
`study inclusion criteria previously defined. Most excluded
`patients were deemed ineligible because they had not re-
`ceived HDK. Some patients were excluded due to insufficient
`data for analysis or because of the simultaneous administra-
`tion of additional anticancer agents. Descriptive statistics for
`baseline and response data for the 78 patients that form the
`basis of this study are summarized in tables 1 and 2. As of
`March 2004, 75 of 78 patients had progressed as defined by
`the progression criteria defined in the materials and methods
`section. The median and mean response time for all 78 pa-
`tients was 6.7 and 14.5 months, respectively. The median
`and mean survival times were 38.0 and 42.4 months, respec-
`tively. Kaplan-Meier representation of response time for all
`78 patients is portrayed in figure 1. Kaplan-Meier represen-
`
`tation of overall survival is portrayed in figure 2. Correlation
`between response time and survival is shown in figure 3.
`Nine laboratory and clinical factors, which were possibly pre-
`dictive of response time and survival, were evaluated in univar-
`iate analysis with the Spearman coefficient. Seven factors had
`an r value of greater than 0.23 (p ⬍0.05) for predicting response
`time and survival (table 3). For response time the 7 factors were
`HB RX time, DX to RX, log of baseline PSA, alkaline phospha-
`tase (ALP), bone scan stage, percent PSA decrease and Gleason
`score. For survival the predictors were the same except that
`Gleason score was lost as a predictor and baseline hematocrit
`(Hct) was added. Step-down Cox proportional hazard regression
`analysis starting with these 7 predictors indicated that for
`response time, 3 factors, the log of baseline PSA, HB Rx Time
`and percent PSA decrease were predictive. For survival log of
`baseline PSA, bone scan stage and percent PSA decrease were
`predictive (table 4).
`
`TABLE 3. Univariate correlations of response time and survival
`Response
`Survival
`Spearman R p Value
`Spearman R p Value
`
`Variable
`
`Age start HDK
`HB Rx time
`Dx to Rx time
`Log baseline PSA
`Baseline ALP
`Baseline Hct
`Bone scan stage
`PSA % decrease
`Gleason score
`
`0.21
`0.25
`0.28
`⫺0.35
`⫺0.25
`0.20
`⫺0.29
`0.75
`⫺0.26
`
`0.07
`0.03
`0.01
`0.001
`0.03
`0.07
`0.01
`⬍0.001
`0.02
`
`0.16
`0.29
`0.24
`⫺0.43
`⫺0.32
`0.23
`⫺0.41
`⫺0.49
`⫺0.16
`
`0.16
`0.01
`0.03
`⬍0.001
`0.004
`0.04
`⬍0.001
`⬍0.001
`0.16
`
`TABLE 4. Predictors of response time and survival using Cox
`regression analysis
`Response
`t Value
`p Value
`
`Variable
`
`Survival
`p Value
`
`t Value
`
`Log base PSA
`Bone scan stage
`
`% PSA decrease
`HB Rx time
`
`3.56
`—
`
`⫺6.74
`⫺2.26
`
`⬍0.001
`Not
`significant
`⬍0.001
`0.024
`
`3.48
`2.80
`
`⫺3.68
`—
`
`0.0005
`0.005
`
`0.0002
`Not
`significant
`
`WCK1016
`Page 3
`
`

`
`1950
`
`MANAGEMENT OF ANDROGEN INDEPENDENT PROSTATE CANCER WITH KETOCONAZOLE
`
`FIG. 4. Kaplan-Meier representation of survival time. Group 0 denotes men with less then 50% decrease in PSA. Group 1 denotes men with
`greater than 50% decrease in PSA.
`
`We then evaluated percent decrease in PSA as a categori-
`cal rather than a continuous variable for predicting survival.
`Kaplan-Meier analysis indicated that a greater than 50%
`decrease predicted increased survival (fig. 4). However, when
`patients with greater than a 50% decrease were split into
`groups constituted by those 50% to 74% vs 75% to 100%, men
`with a 50% to 74% PSA decrease had a survival time similar
`to the men with a PSA decrease of less than 50% (fig. 5). This
`suggests that survival may not be linearly related to percent
`decrease. These Kaplan-Meier analyses can be faulted be-
`cause they ignore possible confounding with other important
`predictors of survival such as baseline PSA, and bone scan
`stage. To deal with this and the nonlinearity, percent de-
`crease was categorized into 4 equal width intervals and used
`as a categorical variable in a Cox regression to predict pros-
`tate cancer death risk after adjusting for log baseline PSA,
`and bone scan stage. Figure 6 is a plot of the resulting
`relative risks. This figure suggests that it is only the larger
`values of percent decrease, those in the range of 0.75 to 1,
`that affect risk. For this category the risk reduction is about
`a factor of 3.4. Only this category is significantly different
`(p ⬍0.001) from the reference category which is the first
`category.
`In terms of toxicity, 5 men who were responding to HDK
`stopped treatment early while still benefiting from an ongo-
`ing PSA response after 12, 24, 25, 34 and 39 months of
`
`therapy. The reason for early discontinuation of HDK was
`excess lassitude and tiredness in all 5 cases. No patients died
`or developed severe toxicity from HDK.
`
`DISCUSSION
`This 78 patient series is notable in 2 distinct ways. First
`followup greater than 13 years enables us to use actual, not
`projected, measurements of survival and response time. This
`longer period of observation demonstrates that a significant
`number of men treated with HDK remained on treatment for
`several years. For example, men in the upper quartile of
`response time had responses lasting for a minimum of 1 and
`a half years. The longest response, which is still ongoing, has
`lasted 91 months. Long response durations tracked closely
`with long survival times.
`A second notable aspect of this patient series suggests a
`possible reason for these relatively long response and sur-
`vival times. Our patient cohort contains a significant number
`of men who developed AIPC before the development of a
`
`FIG. 5. Kaplan-Meier representation of survival time. Group 1
`denotes men with less then 50% decrease in PSA. Group 2 denotes
`men with PSA decrease between 50% and 75%. Group 3 denotes men
`with log PSA decrease greater than 75%.
`
`FIG. 6. Relative risk of death adjusted for log base PSA and bone
`scan stage plotted against percent decrease category.
`
`WCK1016
`Page 4
`
`

`
`MANAGEMENT OF ANDROGEN INDEPENDENT PROSTATE CANCER WITH KETOCONAZOLE
`
`1951
`
`positive bone scan, while those men who did have a positive
`bone scan tended to have a limited extent of metastatic bone
`disease, ie less than 4 lesions. This may explain our some-
`what long median response time of 6.7 months compared
`with other patient series.9 Cox regression analysis in table 4
`did indicate that a more favorable bone scan stage, and a
`baseline PSA less than 25 ng/ml (the median for the 78
`patients), predicted longer survival. The median response
`time of the men who started HDK with positive bone scans
`was 4.8 months, which is consistent with other reported
`series using HDK therapy.
`Percent decrease in PSA in response to therapy predicted
`prolonged response and survival. The degree of PSA decrease
`in response to therapy as an early indicator for predicting
`prolonged survival in androgen independent prostate cancer
`has been reported in multiple studies.10, 11 Percent PSA de-
`crease in response to therapy is an elegant solution to the
`historical problems related to judging drug efficacy in AIPC
`where measurable disease only occurs in a minority of pa-
`tients. Unfortunately a consensus for using PSA response as
`a surrogate for survival is still lacking.12
`Our study finding that a 50% or greater decrease in PSA in
`response to therapy predicts prolonged survival was in line
`with other much larger and validated trials of HDK.13, 14
`However, further analysis of our data indicated that only the
`patients with a 75% or greater decrease in PSA truly had a
`longer survival. The reason that a 50% decrease also pre-
`dicted prolonged survival was because the 50% demarcation
`by its very nature also incorporated patients with a 75% or
`greater decrease in PSA. Clearly a 75% decrease is a strong
`predictor of survival when it is able to impute significance to
`a 50% decrease, a degree of decrease that is substantially
`diluted with the nonpredictive information contained in the
`interval between 50% and 74%.
`There seems to be a lack of awareness in the literature that
`using overlapping variables such as 50% and 75% could in
`some cases indicate falsely that decreases in PSA that reach
`the 50% threshold will necessarily correlate with increased
`survival. When we examined the survival of patients falling
`into the quartile of a 50% to 75% decrease, we found no
`increase in survival compared with men who had less than a
`50% decrease.
`The shortcomings of this study are those characteristic of
`all retrospective studies. One confounding factor that may
`have affected our results was our policy regarding the anti-
`androgen withdrawal response (AAWR). In the majority of
`cases, our usual approach was not to delay starting HDK
`after stopping the preceding antiandrogen, although in a
`minority of cases an attempt at AAWR was implemented
`before HDK was started. Proceeding directly to HDK without
`first performing an AAWR was adopted because of the rela-
`tively infrequent occurrence (20% to 25%) of the antiandro-
`gen withdrawal response, and its relatively short mean du-
`ration of about 3 to 5 months that has been previously
`reported.15, 16 A recent randomized prospective trial showed
`only a 13% response rate.17 We performed an analysis of
`response time and survival in our own patient series (data
`not included) and we did not find a significant difference
`between AAWR followed by HDK compared with no AAWR ie
`immediate HDK. This finding, the absence of a survival
`difference between these 2 different approaches, has been
`previously reported.17
`In the process of patient selection for study inclusion, every
`attempt was made to include all potentially eligible candi-
`dates. Surprisingly we found that relatively few patients had
`stopped HDK for reasons of toxicity. The most common rea-
`son for early discontinuance was lassitude, a well-known
`effect of ketoconazole.18 We have had some success counter-
`acting this side effect by reducing the dose from 1,200 to 600
`mg a day as has been previously reported.19 We think we
`have also observed some amelioration of ketoconazole in-
`
`duced malaise by attempting to reverse the hypogonadally
`induced sarcopenia through a muscle building program un-
`der the supervision of qualified trainers. A group in Canada
`has used this approach successfully to counteract lassitude in
`men treated with hormone blockade.20
`
`CONCLUSIONS
`Retrospective review of 78 patients with androgen inde-
`pendent prostate cancer treated with HDK showed that the
`men in the upper quartile of PSA response had durable
`responses lasting from 17 months to more than 7 years. The
`men most likely to experience extended responses and longer
`survival were those who started HDK with minimal or no
`bone scan involvement and a low baseline PSA. PSA decrease
`of 75% or greater in response to HDK therapy predicted
`better survival compared with men with less than a 75%
`decrease.
`
`REFERENCES
`1. Johnson, D. E., Babaian, R. J., von Eschenbach, A. C., Wishnow,
`K. I. and Tenney, D.: Ketoconazole therapy for hormonally
`refractive metastatic prostate cancer. Urology, 31: 132, 1988
`2. Tapazoglou, E., Subramanian, M. G., Al-Sarraf, M., Kresge, C.
`and Decker, D. A.: High-dose ketoconazole therapy in patients
`with metastatic prostate cancer. Am J Clin Oncol, 9: 369, 1986
`3. Trachtenberg, J. and Pont, A.: Ketoconazole therapy for ad-
`vanced prostate cancer. Lancet, 2: 433, 1984
`4. Aabo, K., Kjaer, M. and Hansen, H. H.: High-dose ketoconazole
`to untreated stage D prostate cancer. Eur J Cancer Clin Oncol,
`24: 431, 1988
`5. Gerber, G. S. and Chodak, G. W.: Prostate specific antigen for
`assessing response to ketoconazole and prednisone in patients
`with hormone refractory metastatic prostate cancer. J Urol,
`144: 1177, 1990
`6. Trump, D. L., Havlin, K. H., Messing, E. M., Cummings, K. B.,
`Lange, P. H. and Jordan, V. C.: High-dose ketoconazole in
`advanced hormone-refractory prostate cancer: endocrinologic
`and clinical effects. J Clin Oncol, 7: 1093, 1989
`7. Bubley, G. J., Carducci, M., Dahut, W., Dawson, N., Daliani, D.,
`Eisenberger, M. et al: Eligibility and response guidelines for
`phase II clinical trials in androgen-independent prostate can-
`cer: recommendations from the Prostate-Specific Antigen
`Working Group. J Clin Oncol, 17: 3461, 1999
`8. Small, E. J., Baron, A. and Bok, R.: Simultaneous antiandrogen
`withdrawal and treatment with ketoconazole and hydrocorti-
`sone in patients with advanced prostate carcinoma. Cancer,
`80: 1755, 1997
`9. Small, E. J. and Vogelzang, N. J.: Second-line hormonal therapy
`for advanced prostate cancer: a shifting paradigm. J Clin On-
`col, 15: 382, 1997
`10. Kelly, W. K., Scher, H. I., Mazumdar, M., Vlamis, V., Schwartz,
`M. and Fossa, S. D.: Prostate-specific antigen as a measure of
`disease outcome in metastatic hormone-refractory prostate
`cancer. J Clin Oncol, 11: 607, 1993
`11. Smith, D. C., Dunn, R. L., Strawderman, M. S. and Pienta, K. J.:
`Change in serum prostate-specific antigen as a marker of
`response to cytotoxic therapy for hormone-refractory prostate
`cancer. J Clin Oncol, 16: 1835, 1998
`12. Scher, H. I. and Kelly, W. K.: Editorial: States and state trans-
`itions are all that really matter. J Urol, 168: 2451, 2002
`13. Small, E. J., McMillan, A., Meyer, M., Chen, L., Slichenmyer,
`W. J., Lenehan, P. F. et al: Serum prostate-specific antigen
`decline as a marker of clinical outcome in hormone-refractory
`prostate cancer patients: association with progression-free
`survival, pain end points, and survival. J Clin Oncol, 19: 1304,
`2001
`14. Scher, H. I., Kelly, W. M., Zhang, Z. F., Ouyang, P., Sun, M.,
`Schwartz, M. et al: Post-therapy serum prostate-specific anti-
`gen level and survival in patients with androgen-independent
`prostate cancer. J Natl Cancer Inst, 91: 244, 1999
`15. Scher, H. I. and Kelly, W. K.: Flutamide withdrawal syndrome:
`its impact on clinical trials in hormone-refractory prostate
`cancer. J Clin Oncol, 11: 1566, 1993
`16. Small, E. J. and Srinivas, S.: The antiandrogen withdrawal
`
`WCK1016
`Page 5
`
`

`
`1952
`
`MANAGEMENT OF ANDROGEN INDEPENDENT PROSTATE CANCER WITH KETOCONAZOLE
`
`syndrome. Experience in a large cohort of unselected patients
`with advanced prostate cancer. Cancer, 76: 1428, 1995
`17. Small, E. J., Halabi, S., Picus, L., Dawson, N., Chen, Y. and
`Vogelzang, N. J.: A prospective trial of antiandrogen with-
`drawal alone or antiandrogen withdrawal in combination with
`high-dose ketoconazole in androgen independent prostate can-
`cer patients: result of CALBG 9583. Proc Am Soc Clin Oncol,
`20: 695, 2001
`18. Small, E. J., Baron, A. D., Fippin, L. and Apodaca, D.: Ketocon-
`azole retains activity in advanced prostate cancer patients
`
`with progression despite flutamide withdrawal. J Urol, 157:
`1204, 1997
`19. Harris, K. A., Small, E. J., Frohlich, M. W., Bok, R., Randall, M.
`and Kakefuda, M.: Prospective trial of low dose ketoconazole
`therapy in patients with androgen-independent prostate can-
`cer. Proc Am Soc Clin Oncol, 20: 2419, 2001
`20. Segal, R. J., Reid, R. D., Courneya, K. S., Malone, S. C.,
`Parliament, M. B., Scott, C. G. et al: Resistance exercise in
`men receiving androgen deprivation therapy for prostate can-
`cer. J Clin Oncol, 21: 1653, 2003
`
`WCK1016
`Page 6