Prostata-shg-bretten.de

P r o s t a t e C a n c e r D N A P l o i d y a n d R e s p o n s e t o S a l v a g e H o r m o n e
T h e r a p y A f t e r R a d i o t h e r a p y W i t h o r W i t h o u t S h o r t - T e r m
T o t a l A n d r o g e n B l o c k a d e : A n A n a l y s i s o f R T O G 8 6 1 0
By A. Pollack, D.J. Grignon, K.H. Heydon, E.H. Hammond, C.A. Lawton, J.B. Mesic, K.K. Fu, A.T. Porter, Purpose: DNA ploidy has consistently been found to be a
Results: DNA nondiploidy was not associated with any
correlate of prostate cancer patient outcome. However, a
of the other prognostic factors in univariate analyses. In
minority of studies have used pretreatment diagnostic ma-
Kaplan-Meier analyses, 5-year overall survival was 70%
terial and have involved radiotherapy (RT)-treated patients.
for those with diploid tumors and 42% for nondiploid tu-
In this retrospective study, the predictive value of DNA
mors. Cox proportional hazards regression revealed that
ploidy was evaluated in patients entered into Radiation
nondiploidy was independently associated with reduced
Therapy Oncology Group protocol 8610. The protocol treat-
overall survival. No correlation was observed between
ment randomization was RT alone versus RT plus short-
DNA ploidy and distant metastasis. The diminished survival
course (ϳ4 months) neoadjuvant and concurrent total an-
in the absence of an increase in distant metastasis was
drogen blockade (RT؉TAB).
related to a reduction in the effect of salvage androgen
Patients and Methods: The study population consisted of
ablation; patients treated initially with RT؉TAB and who
149 patients, of whom 74 received RT alone and 75 re-
had nondiploid tumors had reduced survival after salvage
ceived RT؉TAB. DNA content was determined by image
androgen ablation.
analysis of Feulgen stained tissue sections; 94 patients were
Conclusions: Nondiploidy was associated with shorter
diploid and 55 patients were nondiploid. Kaplan-Meier
survival, which seemed to be related to reduced response to
univariate survival, the cumulative incidence method, and
salvage hormone therapy for those previously exposed to
Cox proportional hazards multivariate analyses were used
short-term TAB.
to evaluate the relationship of DNA ploidy to distant metas-
J Clin Oncol 21:1238-1248. 2003 by American
tasis and overall survival.
Society of Clinical Oncology.
DNA PLOIDY has been investigated as a potential prognos- characterized in the diagnostic material from patients participating tic factor for prostate cancer for many years, and in the in Radiation Therapy Oncology Group (RTOG) protocol 8610.27 vast majority of reports, it has been found to be predictive of RTOG protocol 8610 was a phase III randomized clinical trial patient outcome.1-20 However, most of these studies were done designed to assess the effect of RT plus short-term neoadjuvant using tissue from prostatectomy specimens. Far fewer have and concurrent total androgen blockade (RTϩTAB) as com- examined DNA ploidy as a pretreatment correlate of patient pared with RT alone. The patients enrolled had locally advanced outcome using diagnostic material. Moreover, there are few disease; tumors had a palpable surface area of 25 cm2 or greater.
reports wherein the predictive value of DNA ploidy was inves- In addition, nearly one third of the patients had Gleason score 8 tigated in prostate cancer patients treated with definitive radio- to 10 disease and 8% had documented lymph node involve-ment.27 The purpose of this analysis was to assess the prognostic therapy (RT). Conclusions about the association of DNA ploidy significance of DNA ploidy, as determined by image analysis, with outcome after RT are unclear because the findings have for prostate cancer patients with high-risk features and to been somewhat divergent.21-26 In this analysis, DNA ploidy was determine whether the addition of androgen ablation to RTaffected the prognostic value of these measurements. There were456 evaluable patients entered into the trial, and of these, 149 From the Department of Radiation Oncology, Fox Chase Cancer Center, (33%) had tissue available for DNA ploidy analysis. These and Radiation Therapy Oncology Group, Philadelphia, PA; Department of patients are the subjects of this report.
Pathology, Karmanos CA Institute, and Department of Radiation Oncology,Wayne State University, Detroit, MI; Radiation Therapy Oncology Group, Philadelphia, PA; Department of Pathology, LDS Hospital, Salt Lake City,Utah; Department of Radiation Oncology, Medical College of Wisconsin, Milwaukee, WI; Radiation Oncology Center, Sacramento; and Departmentof Radiation Oncology, University of California, San Francisco, CA; The study population included patients entered into RTOG protocol 8610, Division of Radiation Oncology, Johns Hopkins Hospital, Baltimore, MD; entitled "A phase III trial of Zoladex and flutamide used as cytoreductive and Department of Radiation Oncology, Massachusetts General Hospital, agents in locally advanced carcinoma of the prostate treated with definitive radiotherapy."27 This phase III randomized clinical trial for locally advanced Submitted February 8, 2002; accepted December 18, 2002. prostate cancer was closed in 1991 and accrued a total of 471 patients, 456 Address reprint requests to Alan Pollack, MD, PhD, Department of of whom were evaluable. TAB with flutamide and goserelin acetate (Zola- Radiation Oncology, Fox Chase Cancer Center, 7701 Burholme Ave, dex) was given for a total of 4 months, starting 2 months before radiotherapy Philadelphia, PA 19111; email: A_Pollack@FCCC.edu. 2003 by American Society of Clinical Oncology. Tissue blocks were obtained from 261 (57%) of the 456 evaluable patients.
After hematoxylin- and eosin-stained samples were sectioned and reviewed, Journal of Clinical Oncology, Vol 21, No 7 (April 1), 2003: pp 1238-1248 Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved. DNA NONDIPLOIDY AND SALVAGE HORMONE THERAPY Table 1. Pretreatment Characteristics
Table 2. Potential Pretreatment Predictors of 5-Year* Distant Metastasis, Any
Failure, and Overall Survival (n ؍ 456†)
Abbreviations: DM, distant metastasis; AF, any failure; OS, overall survival; NA, not analyzed; RT, radiotherapy; TAB, total androgen blockade.
*Kaplan-Meier analysis and log-rank test for AF (including death) and OS.
Cumulative incidence and Gray's test for DM.
†There are 429 patients with Gleason score, 149 with DNA-ploidy, and 129 with Abbreviations: KPS, Karnofsky performance status; RT, radiotherapy; TAB, total androgen blockade; NS, not significant.
(Sunnyvale, CA) mounted on an Olympus BH-2 microscope (Lake Success, NY) and a Sony monitor (San Jose, CA) were used to scan the specimens.
†Those with unknown p53 were not included in the analysis.
For each sample, tumor cell and control cell nuclei were taken from the same slide. Each area of interest in the tissue was designated on the slideusing a marking pen. For each slide, a black level and incident light levelwere set for calibration. Each designated area was scanned from left to right sufficient tumor for DNA ploidy analysis was present in 149 patient samples.
covering each field only once. Nuclei were chosen because they appeared not The diagnostic material, which consisted of 113 samples from needle to be overlapped by other nuclei. For each sample, 100 control nuclei biopsies and 36 samples from transurethral resectates, was requested from (endogenous fibroblasts) and 200 tumor nuclei were measured. The DNA participating institutions (Ͼ 100), reviewed centrally by the study pathologist content was plotted as Feulgen-stained DNA versus cell number and (D.J. Grignon) in 98% of cases, and graded according to Gleason.28 A global displayed in histograms. The DNA content mean, SD, and coefficient of Gleason score was assigned. The distribution of patients by Gleason score for variation (CV) were calculated for the control cells (2C control). The mean the study group was 22 in Gleason score 2 to 5, 16 in Gleason score 6, 60 in was used to calculate the DNA index (DI), which was the ratio of the mean Gleason score 7, and 50 in Gleason score 8 to 10; one patient case was not nuclear cell DNA content of tumor population divided by the mean of the 2C graded. The distribution of patients by clinical T category was 38 in T2 and 111 in T3. At the time the trial was initiated, pretreatment prostate-specific Tumor nuclei populations were considered diploid if the main peak DI was antigen (PSA) was not routinely used in the clinic. Pretreatment serum PSA 0.80 to 1.20 with less than 35% of other DNA measurements outside of 2C values were available for only 19 (15%) patients of the study cohort and, as ϩ 2SD (on the basis of the control nuclei population). Populations were a consequence, are not included in the statistical analyses. A prior immuno- considered aneuploid if the main peak DI was less than 0.80 (hypodiploid) or histochemical analysis of p53 status was done in 129 patients29 who greater than 1.20 (hyperdiploid) and were not considered tetraploid. Tet- participated in RTOG protocol 8610. In that report, abnormal p53 expression raploid populations had a DNA index of 1.80 to 2.20. Patients who had (p53-positive by immunohistochemistry) was found to be significantly multiple peaks were considered aneuploid if more than 35% of the tumor correlated with reduced survival. For this reason, p53 status is included in the nuclei population formed peaks in the range greater than 2C Ϯ 2SD.
analysis here. p53 status and DNA ploidy were determined in 113 patients.
DNA Content Measurements by Image Analysis The end points used in the analysis were distant metastasis, any failure, For inclusion in the study, the stained section had to contain identifiable and overall survival. The parameters of distant metastasis and overall carcinoma. Sections were evaluated without knowledge of patient outcome.
survival are as described in the initial report.27 PSA, or biochemical, failure Sections cut 6 ␮m thick on poly-L slides from paraffin-embedded formalin- was included in the definition of any failure. The original treatment protocol fixed tissues were deparaffinized in xylene and rehydrated in a series of was designed before the increasing PSA profile was established as an end ethanol washes (100%, 95%) to a final distilled water step. Slides were then point. Consequently, a PSA of more than 1.5 ng/mL 1 year after random- placed in 5 N HCl for 60 minutes, stained with Schiff's reagent for 60 ization was used as an approximation of biochemical failure. Five patients minutes, rinsed in a sodium metabisulfite rinse (10% Na S O in 1 N HC1), died before 1 year, and 11 other patients did not have posttreatment PSA data dehydrated in reagent alcohol, and then cleared in xylene. Coverslips were for determining biochemical failure; these patients were excluded from the added to slides using synthetic mounting media.
analysis of this end point. Local and regional failures were also included in Measurements were obtained at ϫ200 magnification using 560-nm mono- the definition of any failure. Local failure was defined as an increase in tumor chromatic light. DNA quantification was performed using the Image Measure size of more than 50% for patients in whom complete tumor regression did software program (Phoenix Technology, Inc, Seattle, WA), with a PCVision not occur or as recurrence of a palpable nodule when there was complete Plus digitizing frame-grabber board (Imaging Technology, Inc, Woburn, MA) regression or a positive biopsy of the prostate after 2 or more years of and a Logitech mouse (Fremont, CA). Both a Pulnix TM-745 camera follow-up. Regional metastasis included clinical or radiologic evidence of Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved. Fig 1. Kaplan-Meier survival analysis of overall sur-
vival (upper) and distant failure (lower) for patients with
diploid (--) and nondiploid tumors (- - - -). The tic marks
represent the times at which patients were censored, and
the numbers of patients at risk are displayed above the
x-axis.
disease in the pelvis other than in the prostate. Distant metastasis was defined The distributions of patient characteristics and treatment assignments were as clinical or radiologic evidence of disease outside the pelvis. Any failure compared by the Pearson ␹2 test with the Yates correction factor. Overall was defined as first reported failure, local failure (n ϭ 9), regional failure survival and any failure estimates were derived using the Kaplan-Meier (n ϭ 0), distant failure (n ϭ 11), local plus distant failure (n ϭ 11), biochemical method.30 Gelman et al31 and Gaynor et al32 indicated in their respective failure (n ϭ 112), or death (n ϭ 0). All end points, with the exception of papers that the Kaplan-Meier method generally overestimates distant metas- biochemical failure, were measured from the date of randomization to the first tasis. The cumulative incidence approach was used instead to estimate distant reported failure date or last follow-up date in the absence of failure. The metastasis because it specifically adjusts for competing risk such as dying biochemical failure end point started 1 year after randomization.
without recurrence of prostate cancer.33 Univariate comparisons of overallsurvival and any failure were calculated with the log-rank test.34 Univariate comparisons of local failure, distant failure, and biochemical failure were The published analysis of the evaluable patients on the trial was performed on 456 patients. The current analysis was done on 149 patients, with the Multivariate Cox proportional hazard models were applied to each of the potential for an additional 6 years of follow-up, as compared with the initial three end points. The initial multivariate analyses were restricted to only treatment report.27 As of June 30, 2000, the median follow-up of the alive patients who had DNA ploidy determination. The analyses determined patients in the study cohort was 9 years (range, 1.2 to 11.8 years).
whether DNA content was of prognostic value after adjusting for treatment Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved. DNA NONDIPLOIDY AND SALVAGE HORMONE THERAPY Fig 2. Kaplan-Meier analysis of overall survival for
patients randomly assigned to RT alone (upper) or RT؉TAB
(lower) when subdivided by DNA ploidy status of diploid
(solid line) and nondiploid (- - - -). The tic marks represent
the times at which patients were censored, and the num-
bers of patients at risk are displayed above the x-axis.
assignment and Gleason score as fixed covariates.36 All factors were ploidy determinations, and 129 (28%) had p53 determinations. In only 113 considered as dichotomous variables and coded as follows: treatment (25%) patients were both ploidy and p53 determinations available. There are assignment (0 for RT alone v 1 for RT ϩ hormones), grouped Gleason sums potential problems caused by the missing values. Selection bias may occur, (0 for sums 2 to 6 v 1 for sums 7 to 10), p53 (0 for negative v 1 for positive), wherein the patients in whom the assays were done do not constitute a and DNA content (0 for diploid v 1 for nondiploid). The fitted parameter random sample from the whole study. Consequently, the study cohort may from the Cox model was used to estimate the relative risk associated with have a better or worse outcome than the parent cohort. Moreover, when each prognostic variable and the corresponding 95% confidence interval. A patients with missing values are excluded in the analysis, the number of ratio of 1 would indicate no difference between the two subgroups. The patients to be analyzed may be relatively small, compromising the statistical larger the difference from 1, the greater the difference in the failure rates power needed to detect clinically meaningful differences.
between the two subgroups. The treatment effect was modeled in such a way To adjust for the problem of missing values in the second multivariate that a value less than 1 favored the addition of hormones. DNA ploidy was analysis, two variables instead of one were used to evaluate each marker. For modeled in a way that a value greater than 1 indicates a greater risk of failure DNA ploidy, patients were divided into three categories: determination not for DNA nondiploidy. All of the statistical comparisons were made with done, diploid, and nondiploid. For p53, patients were divided into three categories: determination not done, negative, and positive. The first variable Another multivariate analysis adjusted for two additional factors: p53 and for DNA ploidy would then be 0 for diploid/not done versus 1 for nondiploid, missing tumor determinations. Of the 456 evaluable patients, 149 (33%) had and the second variable would be 0 for nondiploid/not done versus 1 for Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved. Table 3. Cox Proportional Hazards Regression Analyses (n ؍ 149)*
Abbreviations: RR, relative risk; CI, confidence interval; DM, distant metastasis; AF, any failure (including death); OS, overall survival; NA, not analyzed; RT, radiotherapy; *The analysis was performed on 149 patients, in whom all factors were present. p53 status was not included.
diploid. The estimated relative risk of DNA ploidy was figuratively obtained by differences in the distribution of patients between the diploid and subtracting out the two variables. The 27 patients without centrally reviewed nondiploid groups, although a borderline significant relationship Gleason scores were excluded, leaving 429 patients for the analysis.
was seen with Gleason score. Of those that were nondiploid, 83% had a Gleason score of 7 to 10, as opposed to 69% for thosethat were diploid (P ϭ.057). In a prior analysis of this patient On the basis of DNA content measurement, 94 patients were cohort,29 abnormal p53 expression was reported to be a signif- classified as diploid, nine patients were classified as tetraploid, icant correlate of decreased overall survival and so is included and 46 patients were classified as aneuploid. Because the number here. There was no association between the distribution of patients of tetraploid patients was small and not amenable to separate by p53 status and DNA ploidy (correlation coefficient ϭ 0.05).
analysis, the nondiploid patients (tetraploid plus aneuploid; Five-year Kaplan-Meier estimates of overall survival and any n ϭ 55) were pooled, as has been described previously.22 Table failure rates for all patients with the listed variable are shown in 1 shows the distribution of patients by pretreatment characteris- Table 2. Five-year estimates of distant failure rates derived using tics and DNA ploidy. There were no statistically significant the cumulative incidence method are also shown in Table 2. Allfactors listed in the table, including DNA ploidy, affected overall Table 4. Characteristics of Patients Who were Entered in RTOG 8610 by
survival (assigned treatment and p53 status were borderline).
Presence or Absence of Ploidy Data
Estimated 5-year overall survival was only 42% when nondip- loidy was present, versus 70% in diploid patients (Fig 1). Figure 2 shows the breakdown of overall survival rates by protocol treatment assignment and DNA ploidy. Reduced survival rates were observed with nondiploidy in both treatment groups, but the difference between diploid and nondiploid only reached signif- RTOG protocol 8610 had advanced disease, and this is reflected in the high failure of any type and low overall survival rates. Although the main end point of the study was overall survival, the other end points shed light into how survival was affected.
Gleason score, assigned treatment on protocol, and p53 status were associated with all of the end points shown in Table 2.
These associations demonstrate a logical relationship between progression, distant metastasis, and overall survival. Such a pattern, however, was not discerned for DNA ploidy in univariate or Table 5. Univariate Analysis of Outcome by Whether DNA Ploidy
Determination Was or Was Not Done
Abbreviations: RTOG, Radiation Therapy Oncology Group; KPS, Karnofsky performance status; RT, radiotherapy; TAB, total androgen blockade; NS, not †Those with unknown p53 were not included in this analysis.
Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved. DNA NONDIPLOIDY AND SALVAGE HORMONE THERAPY Table 6. Adjusted Cox Proportional Hazards Regression Analyses (n ؍ 429)
Abbreviations: RR, relative risk; CI, confidence interval; DM, distant metastasis; AF, any failure (including death); OS, overall survival; NA, not analyzed; RT, radiotherapy; Cox proportional hazards regression analyses. Survival was worse DNA ploidy was related to overall survival in multivariate when DNA nondiploidy was found, whereas any failure and distant analysis, yet no association with distant metastasis was observed.
metastasis rates were not related to ploidy status.
From these data, it is not intuitive how DNA ploidy affected The initial multivariate analyses were performed using the 149 survival. The lack of a significant correlation between nondip- patients with a DNA ploidy determination. DNA ploidy was loidy and distant metastasis, although reduced survival was associated with overall survival after controlling for assigned evidenced, led us to examine survival after the institution of treatment and Gleason score (Table 3). When this subset was salvage hormone therapy. Figure 3 shows that overall survival at compared with the parent cohort, the differences in baseline 5 years after salvage hormone therapy was significantly lower in characteristics were marginal (Table 4), whereas the differences the presence of nondiploidy (45% v 23%; P ϭ.018). Because the in patient outcome were highly significant (Table 5). For difference in overall survival could be related to an imbalance of example, patients with a DNA ploidy determination had an intercurrent deaths, disease-specific survival results were increased risk of death and any failure. In a previously reported compared. Even though there were fewer patients available analysis, p53 status was found to be a significant prognostic for the analysis of disease-specific survival (n ϭ 40), and variable for survival but was only available for 113 patients with therefore power was reduced, a borderline significant trend a DNA ploidy determination. As a consequence, a multivariate favoring the diploid population was observed (63% v 25% at analysis that adjusted for this population selection effect was 3 years; P ϭ.06). The disease-specific survival results mirror done using 429 patients (see Methods). Table 6 shows that after adjusting for population effects and p53 status, in addition to Figure 4 indicates that the reduced survival of nondiploidy Gleason score and assigned treatment, DNA ploidy was an patients after salvage hormone therapy was the result of salvage independent prognostic factor for overall survival.
hormone therapy resistance of these patients when they were Fig 3. Kaplan-Meier analysis of overall survival after
salvage hormone therapy was started for patients with
diploid (--) and nondiploid tumors (- - - -). The tic marks
represent the times at which patients were censored, and
the numbers of patients at risk are displayed above the
x-axis.
Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved. Fig 4. Kaplan-Meier analysis of overall survival after
salvage hormone therapy (HT) by the treatment random-
ization of RT alone (upper) and RT؉TAB (lower) when
subdivided by DNA ploidy status of diploid (--) and
nondiploid (- - - -). The tic marks represent the times at
which patients were censored, and the numbers of pa-
tients at risk are displayed above the x-axis.
assigned randomly to RTϩTAB. This apparent resistance might absolute rate of distant metastasis that translated into reduced be explained by an unequal distribution of patients with distant survival, but probably more rapid progression to death once metastasis at the time salvage hormone therapy was initiated.
distant metastasis was discernible. Prior treatment with TAB in Figure 5 shows that the same pattern was observed for nondiploid patients may promote resistance to salvage hormone RTϩTAB-treated patients who did not have evidence of distant therapy, thereby shortening survival.
metastasis at the start of salvage hormone therapy. Overall Other potential DNA ploidy-associated differences in treat- survival at 5 years for patients who had distant metastasis at the ment outcome based on the protocol treatment assignments of time of salvage hormone therapy was 15% (n ϭ 14) for patients RT alone versus RTϩTAB are explored in Table 7. In general, randomly assigned to RT alone and 11% (n ϭ 10) for those failure rates were more significantly reduced by RTϩTAB randomly assigned to RTϩTAB. Subdivision by DNA ploidy compared with RT alone for those with diploid tumors than for had no effect on these relationships.
those with nondiploid tumors. These results combined with the Figure 6 is an analysis of time to distant metastasis after findings that survival after salvage hormone therapy is shortened salvage hormone therapy for patients who were free of distant indicate that short-term adjuvant TAB is not advisable in the metastasis at that time. The results indicate that it is not the presence of nondiploidy. Such subgroup analyses are fraught Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved. DNA NONDIPLOIDY AND SALVAGE HORMONE THERAPY Fig 5. Kaplan-Meier analysis of overall survival after
salvage hormone therapy (HT) when distant metastasis
(DM) is absent. The patients are categorized by the treat-
ment randomization of RT alone (upper) and RT؉TAB
(lower) as subdivided by DNA ploidy status of diploid
(- - - -) and nondiploid (--). The tic marks represent the
times at which patients were censored, and the numbers of
patients at risk are displayed above the x-axis.
with potential errors related to the influence of unevenly distrib- loidy.37 However, DNA ploidy analysis has not established a uted prognostic factors. Nonetheless, the results support consid- foothold in clinical practice. One explanation is that the majority eration of DNA ploidy analysis in future trials of short-term of prior studies have not analyzed diagnostic material.
For patients managed by RT, only a handful of reports have examined the association of DNA ploidy with outcome.21-26 Although the results have not been entirely consistent, nondip- Pretreatment prognostic factors have proven valuable in de- loidy has been related to poor patient outcome in the majority of termining prostate cancer patient treatment strategies, especially studies. These reports have involved relatively small numbers of in defining patients who should receive androgen ablation in patients, and additional characterizations of DNA ploidy as a combination with RT. The core factors used in clinical practice prognostic factor are needed. To our knowledge, no reports have are serum PSA, Gleason score, and clinical stage.37 Apart from described the predictive value of DNA ploidy for patients these, and possibly the proportion of cancer in the biopsy specimens,38 the most widely investigated and promising marker The data presented revealed a number of relationships be- of disease progression and reduced survival is DNA nondip- tween DNA ploidy and patient outcome. For the multivariate Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved. Fig 6. Kaplan-Meier analysis of the relationship of
DNA ploidy to distant failure after salvage hormone ther-
apy (HT) when distant metastasis (DM) was absent. This
analysis was restricted to patients randomly assigned to
RT؉TAB. The tic marks represent the times at which pa-
tients were censored, and the numbers of patients at risk
are displayed above the x-axis.
analyses using the 149 patients with DNA ploidy determinations distant metastasis. DNA nondiploidy, in this locally advanced (Table 3) or the adjusted analyses using the 429 patients high-risk cohort, was not a predictor of distant metastasis.
available from the entire cohort (Table 6), nondiploidy was DNA nondiploidy has been correlated with reduced prostate associated with reduced overall survival without any increase in cancer patient survival in numerous reports.37 This correlation distant metastasis. In other studies, nondiploidy consistently has was also observed in the data presented here, despite the lack of been a robust correlate of clinical disease progression (local, a relationship between nondiploidy and distant metastasis. The regional, and distant). Few examples exist in which distant results in Fig 3 indicate that DNA ploidy was associated with metastasis has been analyzed separately,5,22 but there is every survival because the nondiploid cases progressed to death more indication that nondiploidy is related to more rapid progression rapidly after the initiation of salvage hormone therapy. On to distant metastasis and that survival is, therefore, reduced.
further examination, the nondiploid patients who were initially The prostate cancer patient population studied in RTOG treated with RTϩTAB were dramatically more resistant to protocol 8610 was clearly locally advanced and not typical of androgen ablation salvage (Fig 4). Short-term adjuvant TAB, those treated with RT today. Overall, biochemical failure was combined with RT, could have predisposed patients to resistance 78% at 5 years, indicating that the treatments used were to salvage hormone therapy. Although others have shown that inadequate. This biochemical failure rate translated into a 36% response to androgen ablation is less pronounced when DNA distant metastasis rate. The rates of distant metastasis in the nondiploidy is identified,39-43 this is the first description that presence of diploidy and nondiploidy were 35% and 39%, adjuvant androgen ablation in such cases may predispose pa- respectively, which were not different statistically in univariate tients to resistance to androgen ablation salvage.
or multivariate analyses. Because the number of patients exhib- The data presented, however, are not conclusive for a number iting distant failures was 46 of 94 and 22 of 55 for the diploid of reasons. First, there was no evidence of an increased rate of and nondiploid cases, respectively, there were sufficient events distant metastasis in those with nondiploidy. The compromised to measure the effect of DNA ploidy. In multivariate analysis, survival rate of nondiploid patients treated with RTϩTAB after p53 status and Gleason score were independent correlates of salvage hormone therapy seemed to be independent of distantmetastasis rates (Fig 6). Second, Shipley et al44 recently pre-sented an analysis of this type for the entire group of patients in Table 7. Relationship of DNA Ploidy Stratified by Study Randomization to
RTOG protocol 8610, and they did not find a statistically Patient Outcome
significant difference in survival after salvage hormone therapy based on assigned treatment (RT alone v RTϩTAB). Likewise, for the DNA-ploidy cohort studied here, there was no difference in survival after salvage hormone therapy based on assigned treatment (data not shown). However, as shown in Fig 4, survival was reduced for patients who had nondiploidy and who were ϩTAB. Therefore, patients with diploidy who were assigned to RTϩTAB should have had a better survival Abbreviations: RT, radiotherapy; TAB, total androgen blockade; DM, distant than those assigned to RT alone. A nonsignificant trend in this metastasis; AF, any failure; OS, overall survival.
*Log-rank test for AF and OS; Gray's test DM.
direction was observed wherein the survival of diploid patients Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved. DNA NONDIPLOIDY AND SALVAGE HORMONE THERAPY treated with RTϩTAB after salvage hormone therapy was higher prostate DNA ploidy status,12,45,46 which may be measured on than for those treated with RT alone (55% v 39%; P ϭ.54). This thin sections using image analysis. The image analysis technique may have contributed to the difference in survival between for quantifying DNA content used in this report is the preferred diploidy and nondiploidy for those treated with RTϩTAB.
method for the analysis of DNA ploidy status in diagnostic Third, although the cause of death is not always clear for prostate needle biopsy specimens because it requires less tissue than flow cancer patients receiving salvage therapy, there was a slightly cytometry and allows for the histologic separation of normal greater proportion of intercurrent deaths in those with nondip- epithelial and stromal cells from diploid tumor cells (a problem loidy (46%) compared with those with diploidy (39%). Despite that plagues flow cytometric analyses). The data presented show these potentially complicating factors, the overriding concern that locally advanced patients with nondiploidy have reduced raised by the findings is that short-term neoadjuvant or adjuvant survival. The most striking association was that RT plus short- TAB may promote resistance to salvage hormone therapy and more term neoadjuvant and concurrent TAB may predispose patients rapid progression to death once distant metastasis has occurred. This with nondiploidy to reduced survival after salvage hormone hypothesis needs to be confirmed in another group of patients.
therapy. As with all retrospective tumor marker studies, this is a In conclusion, DNA ploidy shows promise in predicting the subset analysis, and although we attempted to correct for outcome of prostate cancer patients. Few reports exist on the selection bias, it is possible that the findings are not representa- relationship of DNA ploidy to outcome after RT, although there tive. One should consider, however, that these data are consistent is ample evidence that nondiploidy is an independent adverse with prior reports that affirm the independent merit of DNA factor that should be considered in treatment planning. Pretreat- ploidy. DNA ploidy will likely become an important factor for ment diagnostic material provides a reasonable representation of the stratification of patients in future trials.
1. Fordham MVP, Burdget AH, Matthews J, et al: Prostatic carcinoma 15. Hawkins CA, Bergstralh EJ, Lieber MM, et al: Influence of DNA cell DNA content measured by flow cytometry and its relation to clinical ploidy and adjuvant treatment on progression and survival in patients with pathologic stage T3 (PT3) prostate cancer after radical prostatectomy.
2. Lee SE, Currin SM, Paulson DF, et al: Flow cytometric determination of ploidy in prostatic adenocarcinoma: A comparison with seminal vesicle 16. Di Silverio F, D'Eramo G, Buscarini M, et al: DNA ploidy, Gleason involvement and histopathological grading as a predictor of clinical recur- score, pathological stage and serum PSA levels as predictors of disease-free survival in C-D1 prostatic cancer patients submitted to radical prostatectomy.
3. Winkler HZ, Rainwater LM, Myers RP, et al: Stage D1 prostatic adenocarcinoma: Significance of nuclear DNA ploidy patterns studied by 17. Ahlgren G, Lindholm K, Falkmer U, et al: A DNA cytometric flow cytometry. Mayo Clin Proc 63:103-112, 1988 proliferation index improves the value of the DNA ploidy pattern as a 4. Nativ O, Winkler HZ, Raz Y, et al: Stage C prostatic adenocarcinoma: prognosticating tool in patients with carcinoma of the prostate. Urology Flow cytometric nuclear DNA ploidy analysis. Mayo Clin Proc 64:911-919, 18. Amling CL, Lerner SE, Martin SK, et al: Deoxyribonucleic acid 5. Adolfsson J, Ro¨nstro¨m L, Hedlund P-O, et al: The prognostic value of ploidy and serum prostate specific antigen predict outcome following salvage modal deoxyribonucleic acid in low grade, low state untreated prostate prostatectomy for radiation refractory prostate cancer. J Urol 161:857-862, 6. Haugen OA, Mjølnerød O: DNA-ploidy as prognostic factor in 19. Diaz JI, Mora LB, Austin PF, et al: Predictability of PSA failure in prostatic carcinoma. Int J Cancer 45:224-228, 1990 prostate cancer by computerized cytometric assessment of tumoral cell 7. Montgomery BT, Nativ O, Blute ML, et al: Stage B prostate adeno- carcinoma: Flow cytometric nuclear DNA ploidy analysis. Arch Surg 20. Cheng L, Darson MF, Slezak J, et al: Correlation of margin status and extraprostatic extension with progression of prostate carcinoma. Cancer 8. Visakorpi T, Kallioniemi O-P, Paronen IYI, et al: Flow cytometric analysis of DNA ploidy and S-phase fraction from prostatic carcinomas: 21. Song J, Cheng WS, Cupps RE, et al: Nuclear deoxyribonucleic acid Implications for prognosis and response to endocrine therapy. Br J Cancer content measured by static cytometry: Important prognostic association for patients with clinically localized prostate carcinoma treated by external beam 9. Wirth MP, Mu¨ller HA, Mansek A, et al: Value of nuclear DNA ploidy patterns in patients with prostate cancer after radical prostatectomy. Eur Urol 22. Gauwitz MD, Pollack A, El-Naggar AK, et al: The prognostic significance of DNA ploidy in clinically localized prostate cancer treated 10. Tinari N, Natoli C, Angelucci D, et al: DNA and S-Phase fraction with radiation therapy. Int J Radiat Oncol Biol Phys 28:821-828, 1994 analysis by flow cytometry in prostate cancer: Clinicopathologic implica- 23. Centeno BA, Zietman AL, Shipley WU, et al: Flow cytometric analysis of DNA ploidy, percent S-phase fraction, and total proliferative 11. Blatstein LM, Ginsberg PC, Daskal I, et al: Flow cytometric deter- fraction as prognostic indicators of local control and survival following mination of ploidy in prostatic adenocarcinoma and its relation to clinical radiation therapy for prostate carcinoma. Int J Radiat Oncol Biol Phys outcome. J Am Osteopath Assoc 93:463-472, 1993 12. Ross JS, Figge H, Bui HX, et al: Prediction of pathologic stage and 24. Pollack A, Zagars GK, El-Naggar A, et al: Near-diploidy: A new postprostatectomy disease recurrence by DNA ploidy analysis of initial prognostic factor for clinically localized prostate cancer treated with external needle biopsy specimens of prostate cancer. Cancer 74:2811-2818, 1994 beam radiotherapy. Cancer 73:1896-1903, 1994 13. Vesalainen S, Nordling S, Lipponen P, et al: Progression and survival 25. Pollack A, Zagars GK, El-Naggar A, et al: Relationship of tumor in prostatic adenocarcinoma: A comparison of clinical stage, Gleason grade, DNA-ploidy to serum prostate-specific antigen doubling time after radio- S-phase fraction and DNA ploidy. Br J Cancer 70:309-314, 1994 therapy for prostate cancer. Urology 44:711-718, 1994 14. Carmichael MJ, Veltri RW, Partin AW, et al: Deoxyribonucleic acid 26. Khoo VS, Pollack A, Cowen D, et al: Relationship of Ki-67 labeling ploidy analysis as a predictor of recurrence following radical prostatectomy index to DNA-ploidy, S-phase fraction, and outcome in prostate cancer for Stage T2 disease. J Urol 153:1015-1019, 1995 treated with radiotherapy. Prostate 41:166-172, 1999 Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved. 27. Pilepich MD, Krall JM, Al-Sarraf M, et al: Androgen deprivation with 38. Iczkowski KA, Bostwick DG: Prostate biopsy interpretation. Current radiation therapy compared with radiation therapy alone for locally advanced concepts, 1999. Urol Clin North Am 26:435-452, 1999 prostatic carcinoma: A randomized comparative trial of the Radiation 39. Forsslund G, Zetterberg A: Ploidy level determinations in high grade and low Therapy Oncology Group. Urology 45:616-623, 1995 grade malignant variants of prostatic carcinoma. Cancer Res 50:4281-4285, 1990 28. Gleason DF: Classification of prostatic carcinomas. Cancer Che- 40. Stege R, Tribukait B, Lundh B, et al: Quantitative estimation of tissue prostate specific antigen, deoxyribonucleic acid ploidy and cytological grade 29. Grignon D, Caplan R, Sarkar FH, et al: P53 status and prognosis of in fine needle aspiration biopsies for prognosis of hormonally treated locally advanced prostatic adenocarcinoma: A study based on RTOG 8610.
prostatic carcinoma. J Urol 148:833-837, 1992 41. Forsslund G, Nilsson B, Zetterberg A: Near tetraploid prostate 30. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 42. Pollack A, Troncoso P, Zagars GK, et al: The significance of 31. Gelman R, Gelber R, Henderson IC, et al: Improved methodology for DNA-ploidy and S-phase fraction in node-positive (Stage D1) prostate analyzing local and distant recurrence. J Clin Oncol 8:548-555, 1990 cancer treated with androgen ablation. Prostate 31:21-28, 1997 32. Gaynor JJ, Feuer EJ, Tan CC, et al: On the use of cause-specific 43. Seay TM, Blute ML, Zincke H: Long term outcome in patients with failure and conditional failure probabilities: Example from clinical oncology pTxNϩ adenocarcinoma of the prostate treated with radical prostatectomy and early androgen ablation. J Urol 159:357-364, 1998 33. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time 44. Shipley WU, Lu JD, Pilepich MV, et al: Effect of a short course of 34. Mantel N: Evaluation of survival data and two new rank order neoadjuvant hormonal therapy on the response to subsequent androgen statistics arising in its consideration. Cancer Chemother Rep 5:163-170, 1966 suppression in prostate cancer patients with relapse after radiotherapy: A 35. Gray RL: A class of K-sample tests for comparing the cumulative secondary analysis of the randomized protocol RTOG 86-10. Int J Radiat incidence of a competing risk. Ann Stat 16:1141-1154, 1988 36. Cox DR: Regression models and life tables. J R Stat Soc Series B 45. Brinker DA, Ross JS, Tran T-A, et al: Can ploidy of prostate carcinoma diagnosed on needle biopsy predict radical prostatectomy stage 37. Pollack A, Zagars GK: Pre-treatment prognostic factors for prostate cancer patients treated with external beam radiotherapy, in Greco C, Zelefsky 46. Ahlgren G, Falmer U, Gadaleanu V, et al: Evaluation of DNA ploidy MJ (eds.): Radiotherapy of Prostate Cancer, Harwood Academic, Singapore, combined with a proliferation index of imprints from core needle biopsies in prostate cancer. Eur Urol 36:314-319, 1999 Information downloaded from jco.ascopubs.org and provided by UNIVERSITAETS UND LANDESBIBLIO on May 9, 2008 Copyright 2003 by the American Society of Clinical Oncology. All rights reserved.

Source: http://www.prostata-shg-bretten.de/Email_Anhang/Pollack2003.pdf