- Email: [email protected]
Prognostic significance of dna ploidy in carcinoma of prostate
PROGNOSTIC
SIGNIFICANCE
IN CARCINOMA STEPHEN ROBERT PHILIP ROBERT
OF PROSTATE
W. DEJTER,
JR., M.D.
E. CUNNINGHAM D. NOGUCHI, V. JONES,
OF DNA PLOIDY
M.D.
JUDD W. MOUL,
M.D.
DAVID G. McLEOD,
M.D.
JOHN H. LYNCH, M.D.
M.D.
From the Division of Urology, Department of Surgery, Georgetown University Hospital, Department of Pathology and Urology Service, Department of Surgery, Walter Reed Army Medical Center, Washington, D.C.
ABST,RACT-Flow cytometry was used to measure the DNA content in archived paraffinembedded human prostatic cancer tissue for 69 patients with known outcomes that presented between 1975 and 1982. Of these, 51 patients had clinically localized lesions and were surgically staged prior to radical prostatectomy, while 18 patients presented with advanced Stage 02 disease. Thirty-six of37 (97.3 %) pathologic Stage B lesions were diploid. In contrast, the majority (72.2 %o) of patients with metastatic disease had aneuploid tumors. The average Gleason grade for aneuploid tumors was 8.2 + 1.98 versus 5.5 + 1.89 for diploid tumors (p < 0.01). For 51 patients with clinically localized tumors, 13.9 percent of diploid tumors with a low Gleason sum (2 to 6) had extracapsular spread of tumor or regional lymph node involvement compared with 83.3 percent of aneuploid tumors with high Gleason scores (7 to 10). The addition of DNA ploidy to degree of glandular differentiation may enhance the prognostic evaluation of prostatic tumors and eventually improve our ability to select patients who are likely to benefit from radical prostatectomy.
The ability to accurately predict the prognosis of individuals with newly diagnosed prostatic cancer continues to elude the practicing urologist. For the most part, clinicians have relied on clinical stage and histologic grade to assess tumor aggressiveness and prognosis. lm4Although valuable, these parameters are imperfect, especially in intermediate grade lesions. In addition, the grading of prostatic carcinoma is a subjective evaluation and therefore may not be uniformly consistent, precise, and reproducible. Orre of the challenges in carcinoma of the prostate is to find the morphologic and biologic features in the tumor which will not only identify the potential for disease progression but also enhance the selection of the optimal treatment for individual patients. Of particular interest is the search for measurable parameters that
UROLOGY
I
MAY
1989 /
VOLUME
XxX111, NUMBER 5
would allow accurate preoperative prediction of the pathologic stage, i.e., localized versus nonlocalized disease, of clinically localized prostatic tumors thereby facilitating the selection of those patients who would benefit from radical excisional therapy. Deoxyribonucleic acid (DNA) ploidy analyzed by flow cytometry (FCM) may be used as an objective and measurable pathologic parameter for the evaluation of prostatic tumors. The change in cell cycle kinetics associated with malignancy may be measured rapidly and reproducibly in large numbers of tumor cells with this technique. The successful application of FCM for the measurement of cellular DNA in archived paraffin-embedded tissue blocks has also made possible the detailed retrospective study of large numbers of surgically staged patients in whom long-term outcome is already known.
361
Relative Representative
DNA content DNA histograms:
We report a study on 69 patients, investigating the relationship of DNA ploidy (measured on paraffin-embedded prostatic tissue) to tumor stage, Gleason grade, and survival in carcinoma of the prostate. Material and Methods Sixty-nine patients with newly discovered, untreated prostatic cancer between 1975 and 1982 had adequate tissue blocks for study and known clinical outcomes. Of these, 51 patients presented with clinical Stage B tumors (all had normal preoperative serum acid phosphatase and negative metastatic surveys by bone scan or skeletal radiographs) and underwent radical retropubic prostatectomy and staging lymphadenectomy (48 patients) or staging lymphadenectomy alone (3 patients found to have metastatic lymph node involvement on frozen section prior to performing radical prostatectomy). Eighteen patients that presented with Stage D2 disease underwent transurethral resection of the prostate for outlet obstruction. Thirty-seven of the 51 patients with clinical Stage B disease were found to have tumor confined to the prostatic capsule and received no further therapy. However, 14 of the 51 clinically localized tumors had either extracapsular extension (including seminal vesicle penetration in 6) of tumor or regional lymph node involvement (Stage C or Dl). Of these, 6 received adjuvant external beam radiation therapy with 4,000 rad, 3 received no therapy (as part of a National Prostatic Cancer Protocol trial for patients with Stages C and Dl disease), and 5 patients were treated with various types of chemo-
362
Relative
DNA content
(A) diploid tumor and (B) aneuploid
tumor.
therapy depending on protocol selection. Time to progression and/or death was analyzed for all 51 patients. Mean and median follow-up for this surgically staged group was 86.5 and 85 months, respectively (range 33 to 135 months). All 18 patients with Stage D2 disease were treated with hormonal manipulation (orchiectomy or DES 1 mg po tid). Patients in this group were followed up for a minimum of sixty months or until death (mean 34.1 months, range 13 to 72 months). Adjacent 5-pm and 50-pm sections were cut with a standard tissue microtome from individual archived paraffin-embedded tumor blocks as well as from age-matched benign prostatic hyperplasia (BPH) specimens obtained by suprapubic prostatectomy. The 5-pm sections were stained with hematoxylin and eosin (H&E) and gra d e d using the Gleason system by one pathologist (RVJ) without knowledge of the clinical stage or patient. The tumor-containing region on the H&E slide was then etched on the corresponding paraffin block. Three 50-pm sections were cut5 and placed in individual finemesh nylon tissue bags* for processing. Single cell suspensions suitable for FCM were prepared using an adaptation of the method described by Hedley et al. 6 The specimens were dewaxed by immersing the tissue bags in 3 changes of 500 mL of xylene for ten minutes each at room temperature, followed by sequential rehydration in 2 changes of 500 mL for ten minutes each of 100, 95, 70, and 50 percent ethanol solutions, also at room temperature. *Cat. #67740010, Shandon Southern Broad Street, Sewickly, PA 15143.
UROLOGY
/
MAY 1989
/ VOLUME
Instruments,
Inc.,
XxX111, NUMBER
515
5
‘TABLE I. Frequency of aneuploidy for prostate carcinomas based on Gleason grade
Gleason Score
-No. of TumorsDiploid Aneuploid
2-4 5-7 8-10
16 25 9
1 3 15
Frequency of Aneuploidy ( % ) 5.8 10.7 62.5
Average Gleason grade for aneuploid and diploid tumors based on pathologic stage
TABLE II.
Pathologic Stage
Diploid
B
5.0 (n = 36) 6.7 (n = 9) D2 7.4 (n = 5) Average Gleason score 5.5 + 1.89 ( f standard deviation) C and Dl
Aneuploid 8.0 (n = 1) 8.8 (n = 5) 7.9 (n = 13) 8.2 f 1.98
KEY: n = number of tumors.
The tissue was rinsed twice in 2 changes of distilled water and then transferred from the tissue bags to 12 x 75 glass test tubes containing 1 mL of 0.5 % pepsin (Sigma) in 0.9 % NaCl, adjusted to pH 1.5 with 2N HCl. The tubes were placed in a 37°C water bath for sixty minutes, with interrnittent vortex mixing. After diluting each sample with 4 volumes of Hanks’ buffered saline solution (HBSS) and filtering through 37pm pore nylon mesh, the cells were centrifuged at 1,000 rpm for five minutes at 25°C. The supernatant was discarded, and the pellets were resuspended in 1 mL of 0.05% Propidium 1odid.e (Sigma) and 100 pg/mL RNase A (Sigma) in HBSS. Cellular DNA content was measured using an Ortho Cytofluorograf model 50 flow cytometer using a 488 nm argon laser excitation beam, counting 10,000 to 20,000 cells per sample. Nuclear size was simultaneously measured using axial extinction (632 nm Helium Neon laser). Human blood lymphocytes were used as diploid standards and agematched BPH specimens served as controls. Examples of typical aneuploid and diploid histograms are shown in Figure 1. Benign prostatic tissue is characterized by diploid patterns whereas prostatic tumors display either diploid or aneuploid histograms. We agree with the criteria set forth by Frankfurt for classifying DNA histograms.7 Briefly, diploid tumors (normal 2C number of chromosomes) exhibit a prominent single GO-G1 peak. Aneuploid tumors (numerical chromosomal abnormality) are characterized by the presence of an additional separate, distinct GO-G1 peak. Histograms were classified as tetraploid only if an apparent subpopulation of GO-G1 cells in the 4C region had a corresponding visible S and G2-M population of cells. This is in contrast to the criteria used by Tribukait, Ronstrom, and Espostis who considered a tumor as tetraploid if the percentage of G2-M ceils in an apparently diploid population exceeded 7 percent. Tetraploid DNA patterns were observed in 21 percent of aneuploid tumors.
UROLOGY
!
MAY 1989
1 VOLUME XxX111. NUMBER 5
Statistical comparison of flow cytometric data was performed with Chi Square test, Wilcoxon test, and linear logistic regression. Actuarial survival curves were prepared with the Kaplan-Meier product limit method and compared using the log rank test. Results The mean age for 51 patients with clinically localized tumors and 18 patients with metastatic Stage D2 disease was 60.5 f 5.7 and 69.9 + 9.6 (age + standard deviation), respectively. The difference in mean ages for these groups was statistically significant (p < 0.05). However, the difference in mean ages for patients with aneuploid tumors compared with diploid tumors in both groups of patients was not significant (p > 0.2). The distribution of aneuploid and diploid tumors based on degree of glandular differentiation for 69 prostate tumors of all stages is depicted in Table I. Tumors were categorized into three groups based on Gleason scores corresponding to well-differentiated (2 to 4), moderately well-differentiated (5 to 7), and poorly differentiated (8 to 10). For 37 patients with surgical Stage B disease, the majority of tumors (45.9%) had an intermediate Gleason score (5 to 7); however many had lesions with low Gleason scores (40.5%) while others (13.5%) had higher scores (highest score was Gleason 8). Among 14 patients with nonlocalized Stages C and Dl tumors and 18 patients with Stage D2 disease, most had poorly or moderately differentiated tumors. The average Gleason grade for aneuploid and diploid tumors broken down by pathologic stage is shown in Table II. Regardless of stage, aneuploid tumors were more poorly differentiated than diploid tmors, though to a lesser extent in Stage D2 lesions. The average Gleason grade for aneuploid lesions was 8.2 f 1.98 compared with 5.5 + 1.89 for diploid tumors (p < 0.01). It is apparent, however, that not all
363
III. Relationship of DNA ploidy to tumor stage for 69 prostate carcinomas
TABLE
Pathologic Stage (n = 37) C and Dl (n = 14) D2 (n = 18)
B
KEY: n =
-No. of TumorsFrequency of Diploid Aneuploid Aneuploidy ( % ) 36 9 5
1
2.7
5
35.7
13
72.2
number of tumors.
high-grade tumors are aneuploid and similarly that not all high-grade lesions have extracapsular spread or regional lymph node involvement. In this series, 5 tumors each with a Gleason sum of 8 (4 diploid) were found to be true surgical Stage B lesions; all 5 patients are alive and had no evidence of disease (NED) at a mean followup of 109 months. The relationship of DNA ploidy to tumor stage is shown in Table III. The increasing frequency of aneuploidy seen with advancing stage is statistically significant (p < 0.001). The upper limit of the 95 percent confidence interval for finding an aneuploid tumor in surgical Stage B disease was 0.12 (based on 1 of 37 patients with diploid pathologic Stage B tumors in this series). The presence of aneuploid stem lines indicated that tumor cells had either spread beyond the prostatic capsule or involved regional lymph nodes in 5 of 6 patients (83.3 %). The 1 patient with an aneuploid pathologic Stage B tumor (Gleason 8) had invasion of the capsule by tumor but no extension outside the capsule and had no evidence of disease at eighty-five months follow-up. Only 1 (diploid tumor) of the 37 patients found to have true Stage B disease progressed to metastatic Stage D2 disease. However, 3 of the 14 patients with nonlocalized lesions progressed to Stage D2 at a mean follow-up interval of 28.3 months. Of these, 1 had an aneuploid tumor and 2 had diploid tumors. Thirty-five of 37 patients with pathologic Stage B lesions were alive at a mean follow-up of 96.4 months (range 75 to 135 months). One patient died after a myocardial infarction (no evidence of disease at autopsy), and 1 patient (diploid tumor) progressed to Stage D2 and died fifty-eight months after radical prostatectomy Eleven of the 14 patients with pathologic Stages C and Dl disease were alive without evidence of metastatic disease at a mean follow-up of 62.5 months. In contrast, 16 of 18 patients with Stage D2 disease were dead of metastases
364
at an average of 30.5 months from presentation Of these, 75 percent had aneuploid stem lines. The influence of DNA ploidy on survival was studied using Kaplan-Meier life table analysis and the log rank test for 69 patients with prostate tumors. After adjusting for stage, a significant difference for survival between aneuploid and diploid tumors was not identified (p > 0.2). The combination of DNA ploidy and Gleason grade for predicting localized versus nonlocalized tumors for 51 patients with clinically localized lesions is shown in Table IV. Tumors were arbitrarily classified into four groups according to degree of glandular differentiation (high Gleason grade = 7 to 10; low Gleason grade = 2 to 6) and ploidy status. A total of 83.8 percent of aneuploid tumors with high Gleason score had spread of tumor cells beyond the prostate compared with only 13.9 percent of diploid low-grade lesions. The relationship of DNA ploidy and Gleason grade to pathologic stage for this group of 51 patients was studied more formally with linear logistic regression treating Gleason score as a continuous predictive variable, ploidy as a binary predictive variable, and pathologic stage (Stage B versus Stages C and Dl) as a binary outcome variable. Both Gleason sum and DNA ploidy considered independently were predictive of pathologic stage (p < 0.01). Although the predictive value of ploidy on pathologic stage when corrected for Gleason grade was not confirmed using this statistical model (p = 0.2) possibly owing to the small total number of aneuploid lesions in this group, the presence of aneuploid stem lines did confer a fourfold increased risk of the tumor having spread beyond the confines of the prostate (Stages C and Dl). Further study on a larger series of patients treated solely by radical retropubic prostatectomy and staging lymphadenectomy is warranted to improve statistical resolution of this finding. Relationship between DNA ploidy, Gleason grade, and pathologic stage for 51 patients with clinically localized tumors
TABLE IV.
DNA Pl. Diploid Diploid Aneuploid Aneuploid
UROLOGY
-No. of TumorsLocalized to Spread Gleason Prostate Beyond Prostate Grade (Stage B) (Stages C, Dl) 2-6 7-10 2-6 7-10
31 (86.1%) 5 (55.6%) : (16.7%)
/ MAY 1989 i
5 (13.9%) 4 (44.4%) 0 5 (83.3%)
Total 36 9 0 6
VOLUME XxX111, NUMBER 5
Comment Flow cytometry for the measurement of cellular DNA content is emerging as an important tool in characterizing the metastatic potential of various malignant human tumorsg~14 To justify the development and application of FCM in the clinical setting, it is necessary not only to show the relationship of DNA ploidy to the currently used pathologic parameters of tumor grade and stage but also to demonstrate that ploidy offers unique or superior information alone or in combination with commonly accepted prognostic parameters. Trib’ukait and associates8 have shown a clear correlation between DNA ploidy and cytologic grade in fine-needle aspiration specimens obtained from 269 untreated prostatic tumors. Our data confirm that in human prostate cancer DNA ploidy is significantly related to the degree of glandular differentiation measured by the Gleason system. Regardless of pathologic stage, the average Gleason sum was higher for aneuploid tumors versus diploid tumors. It is noteworthy that high Gleason score was not necessarily indicative of extraprostatic spread of tumor and poor prognosis; 5 patients with Gleason 8 tumors (4 diploid) were found to have tumor confined to the prostate gland (pathologic Stage B). The frequency of aneuploidy increased significantly with advancing stage of tumor. All but 1 of 37 tumors confined to the prostate (pathologic Stage B) were diploid. This finding corroborates an earlier report by Frankfurt and associates7 in which 11 pathologic Stage B tumors were all diploid (DNA ploidy measurements performed on fresh prostatic tissue). In contrast, 13 of 18 (72.2%) patients with Stage D2 disease had aneuploid tumors. Thus, the presence of aneuploid stemlines appears to be a poor prognostic factor. By combining DNA ploidy and Gleason grade (Table IV), we found that diploid lowgrade (Gleason 2 to 6) tumors were more likely to be localized than a tumor which was only either diploid or Gleason 2 to 6. Five of 6 (83.3 ‘%) aneuploid lesions had tumors which had spread beyond the capsule or involved regional lymph nodes. These data suggest that DNA ploidy in combination with histologic grade may improve our ability to predict the pathologic stage and prognosis of individual patients. A clinical study comparing DNA ploidy and degree of cellular anaplasia (obtained preoperatively by transrectal fine-needle aspira-
UROLOGY
I
MAY 1989
/ VOLUME XXXIII, NUMBER 5
tion in patients with clinically localized tumors) with final pathologic stage would be necessary to substantiate these data and establish the clinical role, if any, of DNA ploidy in the management of this particular group of patients. The influence of DNA ploidy on survival was examined with Kaplan-Meier life-table analysis and the log rank test. Overall, patients with diploid tumors had a survival advantage over patients with aneuploid tumors (p < 0.01); however, when adjusted for stage, the difference in survival curves for aneuploid and diploid tumors was not significant (p > 0.2). Thus, DNA ploidy considered as an independent variable was not a significant predictor of survival in this series. The question of whether or not DNA aneuploidy is present during initial tumor formation and remains stable throughout the course of disease or whether diploid tumors develop into aneuploid lesions during tumor progression has been considered in detail by Frankfurt et al. 7 At this time, neither hypothesis has been proved. Serial biopsy and DNA analysis over many years on large numbers of untreated patients with palpable prostatic nodules would be required to resolve this issue. In summary, our data show that DNA ploidy as an objective pathologic parameter derived from automated measurements correlates well with histologic grade and pathologic stage of human prostate cancers. Of patients who presented with advanced Stage D2 disease, the majority (72.2%) displayed aneuploid DNA patterns. In contrast, 97.3 percent of patients (36 of 37) found to have prostatic malignancy confined to the capsule had diploid tumors. The combination of DNA ploidy and Gleason score may improve the prognostic evaluation of prostatic tumors enhancing the selection of surgically resectable tumor (pathologically confined to the prostate gland) and thus curable lesions. Georgetown University Hospital 3800 Reservoir Road, N.W. Washington, D. C. 20007 (DR. DEJTER) References JT, and Assimos DG: Prognostic significance of tumor grade and stage in the patient with carcinoma of the prostate, Prostate 4: 13 (1983). 2. Gleason DF, Mellinger GT, and Veterans Administration Research Group: Prediction of prognosis for prostatic adenocarcinema by combined histological grading and clinical staging, J Urol 111: 58 (1974). 3. Paulson DF, Piserchia PV, and Gardner IV: Predictors of 1. Grayhack
365
lymphatic spread in prostatic adenocarcinoma. Uro-oncology Research Group study, J Urol 123: 697 (1980). 4. Kraemer SA, et al: Experience with Gleason’s histopathologic grading in prostatic cancer, J Urol 124: 223 (1980). 5. Stephenson RA, Gay H, Fair WR, and Melamed MM: Effect of section thickness on quality of flow cytometric DNA content determinations in paraffin-embedded tissues, Cytometry 7: 41 (1986). 6. Hedley DW, et al: Method for analysis of cellular DNA content of paraffin-embedded pathologic material using flow cytometry, J Histochem Cytochem 31: 1333 (1983). 7. Frankfurt OS, et al: Relationship between DNA ploidy, glandular differentiation, and tumor spread in human prostate cancer, Cancer Res 45: 1418 (1985). 8. Tribukait B, Ronstrom L, and Esposti P: Quantitative and qualitative aspects of flow DNA measurements related to the cytologic grade in prostatic carcinoma, Anal Quant Cytol 5: 107 (1983).
366
9. Balaban G, Merlyn M, Guerry D, and Nowell P: Karyotype stability in the primary tumor and multiple metastases of a human melanoma, Proc Am Assoc Cancer Res 23: 34 (1982). 10. Frankfurt OS, and Huben RP: Clinical applications of DNA flow cytometry for bladder tumors, Urology (Suppl) 23: 29 (1984). 11. Rainwater LM, Hosaka Y, Farrow GM, and Lieber MM: Well-differentiated clear cell renal carcinoma: significance of nuclear deoxyribonucleic acid patterns studied by flow cytometry, J Urol 137: 15 (1987). 12. Frankfurt OS, et al: Flow cytometric analysis of DNA ploidy in primary and metastatic human solid tumors, Cytometry 5: 71 (1984). 13. Fordham MVP et al: Prostatic carcinoma cell DNA content measured by flow cytometry and its relation to clinical outcome, Br J Surg 73: 400 (1986). factors in localized prostatic 14. Pontes JE, et al: Prognostic carcinoma, J Urol 134: 1137 (1985).
UROLOGY
/
MAY 1989
/ VOLUME
XxX111, NUMBER
5
SIGNIFICANCE
IN CARCINOMA STEPHEN ROBERT PHILIP ROBERT
OF PROSTATE
W. DEJTER,
JR., M.D.
E. CUNNINGHAM D. NOGUCHI, V. JONES,
OF DNA PLOIDY
M.D.
JUDD W. MOUL,
M.D.
DAVID G. McLEOD,
M.D.
JOHN H. LYNCH, M.D.
M.D.
From the Division of Urology, Department of Surgery, Georgetown University Hospital, Department of Pathology and Urology Service, Department of Surgery, Walter Reed Army Medical Center, Washington, D.C.
ABST,RACT-Flow cytometry was used to measure the DNA content in archived paraffinembedded human prostatic cancer tissue for 69 patients with known outcomes that presented between 1975 and 1982. Of these, 51 patients had clinically localized lesions and were surgically staged prior to radical prostatectomy, while 18 patients presented with advanced Stage 02 disease. Thirty-six of37 (97.3 %) pathologic Stage B lesions were diploid. In contrast, the majority (72.2 %o) of patients with metastatic disease had aneuploid tumors. The average Gleason grade for aneuploid tumors was 8.2 + 1.98 versus 5.5 + 1.89 for diploid tumors (p < 0.01). For 51 patients with clinically localized tumors, 13.9 percent of diploid tumors with a low Gleason sum (2 to 6) had extracapsular spread of tumor or regional lymph node involvement compared with 83.3 percent of aneuploid tumors with high Gleason scores (7 to 10). The addition of DNA ploidy to degree of glandular differentiation may enhance the prognostic evaluation of prostatic tumors and eventually improve our ability to select patients who are likely to benefit from radical prostatectomy.
The ability to accurately predict the prognosis of individuals with newly diagnosed prostatic cancer continues to elude the practicing urologist. For the most part, clinicians have relied on clinical stage and histologic grade to assess tumor aggressiveness and prognosis. lm4Although valuable, these parameters are imperfect, especially in intermediate grade lesions. In addition, the grading of prostatic carcinoma is a subjective evaluation and therefore may not be uniformly consistent, precise, and reproducible. Orre of the challenges in carcinoma of the prostate is to find the morphologic and biologic features in the tumor which will not only identify the potential for disease progression but also enhance the selection of the optimal treatment for individual patients. Of particular interest is the search for measurable parameters that
UROLOGY
I
MAY
1989 /
VOLUME
XxX111, NUMBER 5
would allow accurate preoperative prediction of the pathologic stage, i.e., localized versus nonlocalized disease, of clinically localized prostatic tumors thereby facilitating the selection of those patients who would benefit from radical excisional therapy. Deoxyribonucleic acid (DNA) ploidy analyzed by flow cytometry (FCM) may be used as an objective and measurable pathologic parameter for the evaluation of prostatic tumors. The change in cell cycle kinetics associated with malignancy may be measured rapidly and reproducibly in large numbers of tumor cells with this technique. The successful application of FCM for the measurement of cellular DNA in archived paraffin-embedded tissue blocks has also made possible the detailed retrospective study of large numbers of surgically staged patients in whom long-term outcome is already known.
361
Relative Representative
DNA content DNA histograms:
We report a study on 69 patients, investigating the relationship of DNA ploidy (measured on paraffin-embedded prostatic tissue) to tumor stage, Gleason grade, and survival in carcinoma of the prostate. Material and Methods Sixty-nine patients with newly discovered, untreated prostatic cancer between 1975 and 1982 had adequate tissue blocks for study and known clinical outcomes. Of these, 51 patients presented with clinical Stage B tumors (all had normal preoperative serum acid phosphatase and negative metastatic surveys by bone scan or skeletal radiographs) and underwent radical retropubic prostatectomy and staging lymphadenectomy (48 patients) or staging lymphadenectomy alone (3 patients found to have metastatic lymph node involvement on frozen section prior to performing radical prostatectomy). Eighteen patients that presented with Stage D2 disease underwent transurethral resection of the prostate for outlet obstruction. Thirty-seven of the 51 patients with clinical Stage B disease were found to have tumor confined to the prostatic capsule and received no further therapy. However, 14 of the 51 clinically localized tumors had either extracapsular extension (including seminal vesicle penetration in 6) of tumor or regional lymph node involvement (Stage C or Dl). Of these, 6 received adjuvant external beam radiation therapy with 4,000 rad, 3 received no therapy (as part of a National Prostatic Cancer Protocol trial for patients with Stages C and Dl disease), and 5 patients were treated with various types of chemo-
362
Relative
DNA content
(A) diploid tumor and (B) aneuploid
tumor.
therapy depending on protocol selection. Time to progression and/or death was analyzed for all 51 patients. Mean and median follow-up for this surgically staged group was 86.5 and 85 months, respectively (range 33 to 135 months). All 18 patients with Stage D2 disease were treated with hormonal manipulation (orchiectomy or DES 1 mg po tid). Patients in this group were followed up for a minimum of sixty months or until death (mean 34.1 months, range 13 to 72 months). Adjacent 5-pm and 50-pm sections were cut with a standard tissue microtome from individual archived paraffin-embedded tumor blocks as well as from age-matched benign prostatic hyperplasia (BPH) specimens obtained by suprapubic prostatectomy. The 5-pm sections were stained with hematoxylin and eosin (H&E) and gra d e d using the Gleason system by one pathologist (RVJ) without knowledge of the clinical stage or patient. The tumor-containing region on the H&E slide was then etched on the corresponding paraffin block. Three 50-pm sections were cut5 and placed in individual finemesh nylon tissue bags* for processing. Single cell suspensions suitable for FCM were prepared using an adaptation of the method described by Hedley et al. 6 The specimens were dewaxed by immersing the tissue bags in 3 changes of 500 mL of xylene for ten minutes each at room temperature, followed by sequential rehydration in 2 changes of 500 mL for ten minutes each of 100, 95, 70, and 50 percent ethanol solutions, also at room temperature. *Cat. #67740010, Shandon Southern Broad Street, Sewickly, PA 15143.
UROLOGY
/
MAY 1989
/ VOLUME
Instruments,
Inc.,
XxX111, NUMBER
515
5
‘TABLE I. Frequency of aneuploidy for prostate carcinomas based on Gleason grade
Gleason Score
-No. of TumorsDiploid Aneuploid
2-4 5-7 8-10
16 25 9
1 3 15
Frequency of Aneuploidy ( % ) 5.8 10.7 62.5
Average Gleason grade for aneuploid and diploid tumors based on pathologic stage
TABLE II.
Pathologic Stage
Diploid
B
5.0 (n = 36) 6.7 (n = 9) D2 7.4 (n = 5) Average Gleason score 5.5 + 1.89 ( f standard deviation) C and Dl
Aneuploid 8.0 (n = 1) 8.8 (n = 5) 7.9 (n = 13) 8.2 f 1.98
KEY: n = number of tumors.
The tissue was rinsed twice in 2 changes of distilled water and then transferred from the tissue bags to 12 x 75 glass test tubes containing 1 mL of 0.5 % pepsin (Sigma) in 0.9 % NaCl, adjusted to pH 1.5 with 2N HCl. The tubes were placed in a 37°C water bath for sixty minutes, with interrnittent vortex mixing. After diluting each sample with 4 volumes of Hanks’ buffered saline solution (HBSS) and filtering through 37pm pore nylon mesh, the cells were centrifuged at 1,000 rpm for five minutes at 25°C. The supernatant was discarded, and the pellets were resuspended in 1 mL of 0.05% Propidium 1odid.e (Sigma) and 100 pg/mL RNase A (Sigma) in HBSS. Cellular DNA content was measured using an Ortho Cytofluorograf model 50 flow cytometer using a 488 nm argon laser excitation beam, counting 10,000 to 20,000 cells per sample. Nuclear size was simultaneously measured using axial extinction (632 nm Helium Neon laser). Human blood lymphocytes were used as diploid standards and agematched BPH specimens served as controls. Examples of typical aneuploid and diploid histograms are shown in Figure 1. Benign prostatic tissue is characterized by diploid patterns whereas prostatic tumors display either diploid or aneuploid histograms. We agree with the criteria set forth by Frankfurt for classifying DNA histograms.7 Briefly, diploid tumors (normal 2C number of chromosomes) exhibit a prominent single GO-G1 peak. Aneuploid tumors (numerical chromosomal abnormality) are characterized by the presence of an additional separate, distinct GO-G1 peak. Histograms were classified as tetraploid only if an apparent subpopulation of GO-G1 cells in the 4C region had a corresponding visible S and G2-M population of cells. This is in contrast to the criteria used by Tribukait, Ronstrom, and Espostis who considered a tumor as tetraploid if the percentage of G2-M ceils in an apparently diploid population exceeded 7 percent. Tetraploid DNA patterns were observed in 21 percent of aneuploid tumors.
UROLOGY
!
MAY 1989
1 VOLUME XxX111. NUMBER 5
Statistical comparison of flow cytometric data was performed with Chi Square test, Wilcoxon test, and linear logistic regression. Actuarial survival curves were prepared with the Kaplan-Meier product limit method and compared using the log rank test. Results The mean age for 51 patients with clinically localized tumors and 18 patients with metastatic Stage D2 disease was 60.5 f 5.7 and 69.9 + 9.6 (age + standard deviation), respectively. The difference in mean ages for these groups was statistically significant (p < 0.05). However, the difference in mean ages for patients with aneuploid tumors compared with diploid tumors in both groups of patients was not significant (p > 0.2). The distribution of aneuploid and diploid tumors based on degree of glandular differentiation for 69 prostate tumors of all stages is depicted in Table I. Tumors were categorized into three groups based on Gleason scores corresponding to well-differentiated (2 to 4), moderately well-differentiated (5 to 7), and poorly differentiated (8 to 10). For 37 patients with surgical Stage B disease, the majority of tumors (45.9%) had an intermediate Gleason score (5 to 7); however many had lesions with low Gleason scores (40.5%) while others (13.5%) had higher scores (highest score was Gleason 8). Among 14 patients with nonlocalized Stages C and Dl tumors and 18 patients with Stage D2 disease, most had poorly or moderately differentiated tumors. The average Gleason grade for aneuploid and diploid tumors broken down by pathologic stage is shown in Table II. Regardless of stage, aneuploid tumors were more poorly differentiated than diploid tmors, though to a lesser extent in Stage D2 lesions. The average Gleason grade for aneuploid lesions was 8.2 f 1.98 compared with 5.5 + 1.89 for diploid tumors (p < 0.01). It is apparent, however, that not all
363
III. Relationship of DNA ploidy to tumor stage for 69 prostate carcinomas
TABLE
Pathologic Stage (n = 37) C and Dl (n = 14) D2 (n = 18)
B
KEY: n =
-No. of TumorsFrequency of Diploid Aneuploid Aneuploidy ( % ) 36 9 5
1
2.7
5
35.7
13
72.2
number of tumors.
high-grade tumors are aneuploid and similarly that not all high-grade lesions have extracapsular spread or regional lymph node involvement. In this series, 5 tumors each with a Gleason sum of 8 (4 diploid) were found to be true surgical Stage B lesions; all 5 patients are alive and had no evidence of disease (NED) at a mean followup of 109 months. The relationship of DNA ploidy to tumor stage is shown in Table III. The increasing frequency of aneuploidy seen with advancing stage is statistically significant (p < 0.001). The upper limit of the 95 percent confidence interval for finding an aneuploid tumor in surgical Stage B disease was 0.12 (based on 1 of 37 patients with diploid pathologic Stage B tumors in this series). The presence of aneuploid stem lines indicated that tumor cells had either spread beyond the prostatic capsule or involved regional lymph nodes in 5 of 6 patients (83.3 %). The 1 patient with an aneuploid pathologic Stage B tumor (Gleason 8) had invasion of the capsule by tumor but no extension outside the capsule and had no evidence of disease at eighty-five months follow-up. Only 1 (diploid tumor) of the 37 patients found to have true Stage B disease progressed to metastatic Stage D2 disease. However, 3 of the 14 patients with nonlocalized lesions progressed to Stage D2 at a mean follow-up interval of 28.3 months. Of these, 1 had an aneuploid tumor and 2 had diploid tumors. Thirty-five of 37 patients with pathologic Stage B lesions were alive at a mean follow-up of 96.4 months (range 75 to 135 months). One patient died after a myocardial infarction (no evidence of disease at autopsy), and 1 patient (diploid tumor) progressed to Stage D2 and died fifty-eight months after radical prostatectomy Eleven of the 14 patients with pathologic Stages C and Dl disease were alive without evidence of metastatic disease at a mean follow-up of 62.5 months. In contrast, 16 of 18 patients with Stage D2 disease were dead of metastases
364
at an average of 30.5 months from presentation Of these, 75 percent had aneuploid stem lines. The influence of DNA ploidy on survival was studied using Kaplan-Meier life table analysis and the log rank test for 69 patients with prostate tumors. After adjusting for stage, a significant difference for survival between aneuploid and diploid tumors was not identified (p > 0.2). The combination of DNA ploidy and Gleason grade for predicting localized versus nonlocalized tumors for 51 patients with clinically localized lesions is shown in Table IV. Tumors were arbitrarily classified into four groups according to degree of glandular differentiation (high Gleason grade = 7 to 10; low Gleason grade = 2 to 6) and ploidy status. A total of 83.8 percent of aneuploid tumors with high Gleason score had spread of tumor cells beyond the prostate compared with only 13.9 percent of diploid low-grade lesions. The relationship of DNA ploidy and Gleason grade to pathologic stage for this group of 51 patients was studied more formally with linear logistic regression treating Gleason score as a continuous predictive variable, ploidy as a binary predictive variable, and pathologic stage (Stage B versus Stages C and Dl) as a binary outcome variable. Both Gleason sum and DNA ploidy considered independently were predictive of pathologic stage (p < 0.01). Although the predictive value of ploidy on pathologic stage when corrected for Gleason grade was not confirmed using this statistical model (p = 0.2) possibly owing to the small total number of aneuploid lesions in this group, the presence of aneuploid stem lines did confer a fourfold increased risk of the tumor having spread beyond the confines of the prostate (Stages C and Dl). Further study on a larger series of patients treated solely by radical retropubic prostatectomy and staging lymphadenectomy is warranted to improve statistical resolution of this finding. Relationship between DNA ploidy, Gleason grade, and pathologic stage for 51 patients with clinically localized tumors
TABLE IV.
DNA Pl. Diploid Diploid Aneuploid Aneuploid
UROLOGY
-No. of TumorsLocalized to Spread Gleason Prostate Beyond Prostate Grade (Stage B) (Stages C, Dl) 2-6 7-10 2-6 7-10
31 (86.1%) 5 (55.6%) : (16.7%)
/ MAY 1989 i
5 (13.9%) 4 (44.4%) 0 5 (83.3%)
Total 36 9 0 6
VOLUME XxX111, NUMBER 5
Comment Flow cytometry for the measurement of cellular DNA content is emerging as an important tool in characterizing the metastatic potential of various malignant human tumorsg~14 To justify the development and application of FCM in the clinical setting, it is necessary not only to show the relationship of DNA ploidy to the currently used pathologic parameters of tumor grade and stage but also to demonstrate that ploidy offers unique or superior information alone or in combination with commonly accepted prognostic parameters. Trib’ukait and associates8 have shown a clear correlation between DNA ploidy and cytologic grade in fine-needle aspiration specimens obtained from 269 untreated prostatic tumors. Our data confirm that in human prostate cancer DNA ploidy is significantly related to the degree of glandular differentiation measured by the Gleason system. Regardless of pathologic stage, the average Gleason sum was higher for aneuploid tumors versus diploid tumors. It is noteworthy that high Gleason score was not necessarily indicative of extraprostatic spread of tumor and poor prognosis; 5 patients with Gleason 8 tumors (4 diploid) were found to have tumor confined to the prostate gland (pathologic Stage B). The frequency of aneuploidy increased significantly with advancing stage of tumor. All but 1 of 37 tumors confined to the prostate (pathologic Stage B) were diploid. This finding corroborates an earlier report by Frankfurt and associates7 in which 11 pathologic Stage B tumors were all diploid (DNA ploidy measurements performed on fresh prostatic tissue). In contrast, 13 of 18 (72.2%) patients with Stage D2 disease had aneuploid tumors. Thus, the presence of aneuploid stemlines appears to be a poor prognostic factor. By combining DNA ploidy and Gleason grade (Table IV), we found that diploid lowgrade (Gleason 2 to 6) tumors were more likely to be localized than a tumor which was only either diploid or Gleason 2 to 6. Five of 6 (83.3 ‘%) aneuploid lesions had tumors which had spread beyond the capsule or involved regional lymph nodes. These data suggest that DNA ploidy in combination with histologic grade may improve our ability to predict the pathologic stage and prognosis of individual patients. A clinical study comparing DNA ploidy and degree of cellular anaplasia (obtained preoperatively by transrectal fine-needle aspira-
UROLOGY
I
MAY 1989
/ VOLUME XXXIII, NUMBER 5
tion in patients with clinically localized tumors) with final pathologic stage would be necessary to substantiate these data and establish the clinical role, if any, of DNA ploidy in the management of this particular group of patients. The influence of DNA ploidy on survival was examined with Kaplan-Meier life-table analysis and the log rank test. Overall, patients with diploid tumors had a survival advantage over patients with aneuploid tumors (p < 0.01); however, when adjusted for stage, the difference in survival curves for aneuploid and diploid tumors was not significant (p > 0.2). Thus, DNA ploidy considered as an independent variable was not a significant predictor of survival in this series. The question of whether or not DNA aneuploidy is present during initial tumor formation and remains stable throughout the course of disease or whether diploid tumors develop into aneuploid lesions during tumor progression has been considered in detail by Frankfurt et al. 7 At this time, neither hypothesis has been proved. Serial biopsy and DNA analysis over many years on large numbers of untreated patients with palpable prostatic nodules would be required to resolve this issue. In summary, our data show that DNA ploidy as an objective pathologic parameter derived from automated measurements correlates well with histologic grade and pathologic stage of human prostate cancers. Of patients who presented with advanced Stage D2 disease, the majority (72.2%) displayed aneuploid DNA patterns. In contrast, 97.3 percent of patients (36 of 37) found to have prostatic malignancy confined to the capsule had diploid tumors. The combination of DNA ploidy and Gleason score may improve the prognostic evaluation of prostatic tumors enhancing the selection of surgically resectable tumor (pathologically confined to the prostate gland) and thus curable lesions. Georgetown University Hospital 3800 Reservoir Road, N.W. Washington, D. C. 20007 (DR. DEJTER) References JT, and Assimos DG: Prognostic significance of tumor grade and stage in the patient with carcinoma of the prostate, Prostate 4: 13 (1983). 2. Gleason DF, Mellinger GT, and Veterans Administration Research Group: Prediction of prognosis for prostatic adenocarcinema by combined histological grading and clinical staging, J Urol 111: 58 (1974). 3. Paulson DF, Piserchia PV, and Gardner IV: Predictors of 1. Grayhack
365
lymphatic spread in prostatic adenocarcinoma. Uro-oncology Research Group study, J Urol 123: 697 (1980). 4. Kraemer SA, et al: Experience with Gleason’s histopathologic grading in prostatic cancer, J Urol 124: 223 (1980). 5. Stephenson RA, Gay H, Fair WR, and Melamed MM: Effect of section thickness on quality of flow cytometric DNA content determinations in paraffin-embedded tissues, Cytometry 7: 41 (1986). 6. Hedley DW, et al: Method for analysis of cellular DNA content of paraffin-embedded pathologic material using flow cytometry, J Histochem Cytochem 31: 1333 (1983). 7. Frankfurt OS, et al: Relationship between DNA ploidy, glandular differentiation, and tumor spread in human prostate cancer, Cancer Res 45: 1418 (1985). 8. Tribukait B, Ronstrom L, and Esposti P: Quantitative and qualitative aspects of flow DNA measurements related to the cytologic grade in prostatic carcinoma, Anal Quant Cytol 5: 107 (1983).
366
9. Balaban G, Merlyn M, Guerry D, and Nowell P: Karyotype stability in the primary tumor and multiple metastases of a human melanoma, Proc Am Assoc Cancer Res 23: 34 (1982). 10. Frankfurt OS, and Huben RP: Clinical applications of DNA flow cytometry for bladder tumors, Urology (Suppl) 23: 29 (1984). 11. Rainwater LM, Hosaka Y, Farrow GM, and Lieber MM: Well-differentiated clear cell renal carcinoma: significance of nuclear deoxyribonucleic acid patterns studied by flow cytometry, J Urol 137: 15 (1987). 12. Frankfurt OS, et al: Flow cytometric analysis of DNA ploidy in primary and metastatic human solid tumors, Cytometry 5: 71 (1984). 13. Fordham MVP et al: Prostatic carcinoma cell DNA content measured by flow cytometry and its relation to clinical outcome, Br J Surg 73: 400 (1986). factors in localized prostatic 14. Pontes JE, et al: Prognostic carcinoma, J Urol 134: 1137 (1985).
UROLOGY
/
MAY 1989
/ VOLUME
XxX111, NUMBER
5