Fax: (011) 351 225084016
Genomic analysis of prostate carcinoma specimens obtained via ultrasound-guided needle biopsy may be of use in preoperative decision-making
Article first published online: 9 SEP 2004
Copyright © 2004 American Cancer Society
Volume 101, Issue 8, pages 1786–1793, 15 October 2004
How to Cite
Teixeira, M. R., Ribeiro, F. R., Eknæs, M., Wæhre, H., Stenwig, A. E., Giercksky, K. E., Heim, S. and Lothe, R. A. (2004), Genomic analysis of prostate carcinoma specimens obtained via ultrasound-guided needle biopsy may be of use in preoperative decision-making. Cancer, 101: 1786–1793. doi: 10.1002/cncr.20527
- Issue published online: 1 OCT 2004
- Article first published online: 9 SEP 2004
- Manuscript Revised: 15 JUN 2004
- Manuscript Accepted: 15 JUN 2004
- Manuscript Received: 25 FEB 2004
- Norwegian Cancer Society. Grant Number: A95068
- Fundação para a Ciência e a Tecnologia
- Comissão de Fomento da Investigação em Cuidados de Saúde (Ministério da Saúde P.I.. Grant Number: 219/01
- prostate carcinoma;
- sextant biopsy;
- molecular cytogenetics
The widespread use of prostate-specific antigen (PSA) testing to screen for prostate carcinoma has led to significant overdiagnosis, due to the frequent detection of indolent malignancies on PSA screening. The detection of abnormal PSA levels typically is followed by ultrasound-guided needle biopsy. Therefore, in an effort to identify genetic markers that augment the information provided by standard histopathologic classification, the authors tested the feasibility of using these minute biopsy samples for genomic profiling via chromosome banding analysis and comparative genomic hybridization (CGH).
Ultrasound-guided needle biopsy specimens obtained preoperatively from 35 patients with prostate carcinoma were analyzed via chromosome banding analysis (after short-term culturing) and CGH. The findings of these analyses then were analyzed for potential correlations with clinicopathologic parameters.
Chromosome banding analysis and CGH were possible in 34 and 33 of the 35 study specimens, respectively. Combined analysis revealed aberrations in 69% of all samples investigated. Copy number losses occurred most commonly at 8p (58% of all abnormal specimens), 16q (42%), and 13q (37%), whereas the only gains detected in more than 1 specimen were those that occurred at 8q (37%). Genomic imbalances and losses at 16q were significantly associated with more poorly differentiated subtypes of prostate carcinoma (P = 0.048 and P = 0.019, respectively), whereas gains at 8q and losses at 16q were significantly correlated with clinically advanced disease (P = 0.048 for the finding of a gain at 8q together with a loss at 16q; P = 0.01 for the finding of either aberration alone).
The authors conclude that genomic analysis of suspected prostate carcinoma specimens obtained via ultrasound-guided needle biopsy is feasible. Thus, it may be possible to use genetic markers to obtain diagnostic and/or prognostic information that is useful in the making of preoperative decisions regarding prostate carcinoma management. Cancer 2004. © 2004 American Cancer Society.
The widespread use of prostate-specific antigen (PSA) testing in prostate carcinoma screening1 has exacerbated the existing dilemma regarding the clinical management of this malignancy. Although the early detection of prostate carcinoma increases the likelihood of potentially curative treatment by prostatectomy in patients with organ-confined disease, there is mounting evidence that PSA screening also leads to considerable overdiagnosis, due to the frequent detection of indolent prostate lesions that probably will not become clinically problematic during patients' lifetimes.2, 3 In fact, autopsy-based series have reported the presence of small prostate carcinomas in as many as one-third of all young men (age < 40 years) who were examined and in approximately two-thirds of all older men who were evaluated.4 This finding is of concern, because prostatectomy is accompanied by major side effects that may lead to reductions in quality of life.5, 6
The genomic changes that characterize and presumably drive neoplastic transformation may hold the key to explaining the clinical heterogeneity of prostate carcinoma. Preliminary data suggest that the number of karyotypic abnormalities detected either on chromosome banding analysis or on comparative genomic hybridization (CGH) possesses prognostic value.7, 8 However, because these data were yielded by the analysis of prostatectomy specimens, the scope of the genetic findings made is limited to clinically organ-confined disease; furthermore, analysis of prostatectomy specimens precludes the use of any potentially informative genetic markers in the preoperative decision-making process.
Because the finding of abnormal PSA levels typically is followed by ultrasound-guided needle biopsy, it would be advantageous to use such biopsy samples to investigate relevant genetic markers and thereby supplement any information obtained on standard histopathologic examination. To address this possibility, we used two genomic screening techniques to analyze preoperative needle biopsy samples obtained from patients with prostate carcinoma.
MATERIALS AND METHODS
Prostate Needle Biopsies
Ultrasound-guided needle biopsy samples were obtained from 35 men who were determined, on the basis of either elevated PSA levels or suspicious findings on digital rectal examination, to have an increased risk of prostate carcinoma. Aside from the sextant biopsy sample that typically is obtained for diagnostic purposes, two additional samples were obtained from the most suspicious region of the prostate for chromosome banding and CGH analyses (one sample for each method). Biopsy samples obtained for karyotyping were transported under sterile conditions to a cytogenetic laboratory, and samples obtained for CGH analysis were frozen at −80 °C until use. Histopathologic diagnoses were made in accordance with World Health Organization guidelines9; histopathologic findings included 2 cases of prostate intraepithelial neoplasia (PIN), 8 cases of well differentiated adenocarcinoma (WDCa), 15 cases of moderately differentiated adenocarcinoma (MDCa), 2 cases of poorly differentiated adenocarcinoma (PDCa), 2 cases of mixed MDCa and PIN, 3 cases of mixed MDCa and WDCa, and 3 cases of mixed PDCa and MDCa (Table 1). For the purposes of our statistical analyses, lesions with mixed morphology were categorized according to their most advanced or least differentiated component. The study was approved by an institutional review board, and informed consent was obtained from all patients.
|Case no.||Age (yrs)||PSA level (ng/mL)||Diagnosisa||T status||Chromosome banding results||CGH results|
|21/98||60||10.2||WDCa||T2||46,XY,del(10)(p13)/46,XY||rev ish enh(8q12q21),dim(3q27qter, 8p12pter)|
|28/98||63||5.9||MDCa||T2||47,XY,+mar/46,XY||rev ish enh(3p22)|
|29/98||53||29.6||MDCa||T3||46,XY||rev ish enh(8q21qter),dim(8p12pter, 13q14q21,16q)|
|35/98||63||31.2||MDCa||T2||46,XY||rev ish enh(8q12qter),dim(8p12pter, 16q21qter)|
|36/98||71||8.6||MD + PDCa||T3||46,XY||No copy number changes|
|37/98||56||17.0||WDCa||T2||45,X,−Y/46,XY,t(1;22)(p13;p11)/46,XY||rev ish enh(17q22qter),dim (17p12pter)|
|46/98||58||7.5||MDCa||T2||46,XY||rev ish dim(8p21pter,13q21q32)|
|48/98||56||10.5||PIN||T2||45,X,−Y/46,XY||No copy number changes|
|50/98||70||6.7||MDCa||T2||n.d.||rev ish dim(10q22q26,13q14qter, 16q,17p12pter)|
|51/98||57||4.7||WDCa||T1||46,XY,del(14)(q13)/47,XY,+Y/46, XY||rev ish enh(4q26),dim(8p23)|
|56/98||57||5.9||MDCa||T2||46,XY||No copy number changes|
|62/98||62||32.9||MDCa + PIN||T2||46,XY,del(10)(p13)/46,XY||No copy number changes|
|64/98||50||11.0||WDCa||T1||47∼48,XY,+7,+18[cp3]/45,X,−Y/46, XY||rev ish dim(8p21p23)|
|66/98||69||n.d.||PIN||n.d.||45,X,−Y/47∼48,XY,+6,+7[cp4]/46, XY||No copy number changes|
|69/98||55||15.8||PDCa||T3||46,XY||rev ish enh(8),dim(12pterq13, 16q22qter,18)|
|70/98||67||n.d.||MDCa||n.d.||47,XY,+Y/46,XY||rev ish dim(8p21pter,13q14,16q)|
|72/98||67||15.0||MD + PDCa||T3||46,XY||rev ish dim(8p21p22,16q24)|
|73/98||63||38.1||PDCa||T3||47,XY,del(7)(p14p21)/46,XY,del(1) (q41q43)/46,XY||No copy number changes|
|77/98||60||20.0||MDCa||T3||46,XY||rev ish enh(8q24,9q33q34), dim(1q42q44,8p12pter)|
|78/98||63||22.0||MDCa||T2||46,XY||rev ish enh(7p13p21,7q31q32)|
|82/98||57||19.0||MDCa||T2||46,XY||rev ish dim(12p12p13)|
|88/98||69||27.7||MDCa||T2||47,XY,+7/46,XY||rev ish enh(8q12qter),dim (1q41qter,6q21q22, 8p21pter,16q)|
|100/98||60||11.0||WD + MDCa||T2||46,XY||n.d.|
|101/98||61||8.0||WDCa||T2||46,XY||No copy number changes|
|3/99||62||18.3||MDCa||T2||46,XY||rev ish dim(1p21p31,13q21q22, 16q22qter)|
|7/99||52||9.0||WDCa||T2||46,XY||No copy number changes|
|10/99||58||3.0||WDCa||T2||46,XY||No copy number changes|
|11/99||62||70.0||MDCa||T3||46,XY||No copy number changes|
|12/99||64||16.0||WDCa||T3||46,XY,add(3)(q24)/46,XY||No copy number changes|
|16/99||48||39.0||MDCa + PIN||T2||45,X,−Y/47,XY,+3/46,XY||rev ish enh(3q23q29,8q12qter), dim(8p21pter,13q14q22)|
|19/99||54||5.0||WD + MDCa||T1||46,XY||rev ish dim(2q14q22,13q21q31)|
|32/99||60||12.0||MDCa||T3||46,XY||No copy number changes|
|38/99||53||7.0||WD + MDCa||T2||46,XY||No copy number changes|
|41/99||59||12.0||MD + PDCa||T2||46,XY||No copy number changes|
Chromosome Banding Analysis
Karyotype analysis after short-term culturing was performed according to the method described previously by Teixeira et al.10 In brief, samples were disaggregated mechanically (using scissors) into pieces of 1–2 mm2 in surface area and then treated with collagenase (CLS 2; Worthington Biochemical Corp., Freehold, NJ) at a concentration of 700 units (U) per mL for 1–2 hours at 37 °C in a humidified atmosphere containing 5% CO2. Cultures were grown in Falcon plastic flasks (surface area, 12.5 cm2; Becton-Dickinson, Franklin Lakes, NJ) coated with 0.1 mg/mL Vitrogen 100 (Collagen Corp., Palo Alto, CA). Cells were grown in a chemically defined medium consisting of Dulbecco Modified Eagle Medium/Ham Nutrient Mixture F12 (1:1 ratio; Life Technologies, Rockville, MD) supplemented with 2 μg/mL dihydrotestosterone, 25 μg/mL transferrin, 20 μg/mL fetuin, 3 μg/mL insulin, 2.6 ng/mL sodium selenite, 0.5 μg/mL hydrocortisone, 10−8 M dibutyryl cyclic adenosine monophosphate, 0.6 ng/mL triiodothyronine, 1 μg/mL cholera toxin, 0.1 mM O-phosphorylethanolamine, 10 μg/mL ascorbic acid, and 0.01% bovine serum albumin (all from Sigma, St. Louis, MO); 100 U/mL penicillin, 100 μg/mL streptomycin, 3 mM L-glutamine, and 100 ng/mL fibronectin (all from Life Technologies); and 20 ng/mL epidermal growth factor (Becton Dickinson). After 5–10 days, Colcemid (Life Technologies) was added to all cultures (final concentration, 0.03 μg/mL), and 4–6 hours later, metaphase cells were harvested after being dislodged with 0.05% trypsin–0.02% ethylenediamine tetraacetic acid (Life Technologies). Harvested cells were subjected to hypotonic shock in 0.05 M KCl, fixed in a 3:1 (v/v) solution of methanol/acetic acid, and dropped onto slides that had been stored in distilled water at 4 °C. Slides were allowed to air-dry, aged overnight at 60 °C in an oven, and then incubated in 2X standard saline citrate (SSC) solution for 3 hours in a 60 °C water bath. G-banding was performed the following day using Wright staining solution. The clonality criteria and karyotype descriptions used were based on the recommendations of the International System for Cytogenetic Nomenclature (ISCN).11 The results of cytogenetic analysis of the first 10 cases (out of 35) presented in Table 1 have been reported.10
The CGH procedure described by Kallioniemi et al.12, 13 was performed with the modifications previously reported by Kraggerud et al.14 and Teixeira et al.15 In brief, test (i.e., prostate biopsy sample) DNA and reference (i.e., healthy male–derived peripheral blood lymphocyte) DNA were extracted using standard methods and labeled in nick-translation reactions, with each reaction involving 2 fluorochrome-conjugated nucleotides (fluoroisothiocyanate [FITC]-12-deoxycytidine triphosphate [dCTP] and FITC-12-deoxyuridine triphosphate [dUTP] for tumor DNA and Texas Red-6-dCTP and Texas Red-6-dUTP for normal DNA; New England Nuclear, Boston, MA); DNA fragments of 300–2000 base pairs were generated by these reactions. Equal amounts of labeled tumor DNA and reference DNA (800 ng each) were mixed with 20 μg unlabeled COT1 DNA (Life Technologies), ethanol-precipitated, dried, and dissolved in hybridization buffer (Vysis, Downers Grove, IL). Normal metaphases were assessed on commercially available slides (Vysis). After chromosomal DNA and the DNA probe were denatured, hybridization was allowed to occur for 2–3 days in a humidified chamber at 37 °C. After a series of washes, slides were mounted in an antifade solution with diaminophenylindole (DAPI; Vectashield; Vector Laboratories, Burlingame, CA).
Ten high-quality metaphase spreads were selected for analysis in each case. Three images, corresponding to the FITC (green) and Texas Red (red) hybridization signals and DAPI counterstain, respectively, were sequentially captured using a coupled-charge device (12-bit grayscale) camera (Cohu 4900; Cohu, San Diego, CA) with an automated filter wheel coupled to a Zeiss Axioplan fluorescence microscope (Zeiss, Oberkochen, Germany) equipped with a CytoVision system (Version 2.7; Applied Imaging, Santa Clara, CA). Chromosomes were identified on the basis of their inverted appearance on DAPI counterstaining, and relative hybridization signal intensities were evaluated along each chromosome. Data obtained from 10 cells were combined to generate average ratio profiles with 99% confidence intervals for each chromosome. Negative (normal vs. normal) and positive control samples were included in each set of experiments. Ten cells from each negative control sample were used to establish the normal ratio profile with 99% confidence intervals.16 Copy number gains or losses were registered and scored whenever the 99% confidence intervals for the test and reference samples did not overlap. Scoring was performed independently by two authors, and interobserver differences were rare; all discrepancies were resolved by joint reevaluation. Changes in CGH copy numbers were characterized according to the guidelines of the ISCN.11
Mann–Whitney and Kruskal–Wallis tests were used to investigate potential correlations of the number of genomic imbalances with clinical stage and histopathologic tumor differentiation. The chi-square test for trend was used to evaluate associations of specific genomic aberrations with histopathologic grade and clinical stage. P values < 0.05 were considered indicative of statistical significance.
Clinicopathologic data and detailed results of the genetic analysis of all cases are summarized in Table 1.
Chromosome Banding Analysis
Successful short-term cultures with pure epithelial morphology were grown from all 35 prostate biopsy samples, and cytogenetic analysis was possible in 34 of these samples (97%). Clonal chromosome abnormalities were detected in 13 of 34 cases (38%; Table 1). The most common copy number aberrations were Y chromosome copy number loss (n = 5; accompanied by other structural or copy number abnormalities in 4 cases), gain at chromosome 7 (n = 3; detected as the lone aberration in only 1 case), and Y chromosome copy number gain (n = 2). Clonal structural abnormalities included the following deletions: del(10)(p13) (n = 2), del(1)(q41q43) (n = 1), del(7)(p14p21) (n = 1), and del(14)(q13) (n = 1).
High-quality DNA for CGH analysis was obtained from 33 of 35 biopsy samples (94%). Copy number aberrations were documented in 19 of 33 cases (58%; Table 1). The average number of genomic aberrations detected in specimens with abnormal findings was 2.7 (mean number of losses, 2.0; mean number of gains, 0.7; Fig. 1). Multiple instances of copy number loss were noted at 8p (58% of all cases with abnormal findings); 16q (42%); 13q (37%); and 1q, 12p, and 17p (11% each). In contrast, the only copy number gains observed in multiple specimens were those occurring at 8q (37%; Fig. 2). The smallest regions of overlap for genomic losses were 8p21–p22 and 8p23 (each in 10 of 11 cases of 8p loss), 16q24 (in all 8 cases of 16q loss), 13q21 (in 6 of 7 cases of 13q loss), 1q42–q44 (in both cases of 1q loss), 12p12–p13 (in both cases of 12p loss), and 17p12–p13 (in both cases of 17p loss). The smallest regions of overlap for genomic gains were 8q21 and 8q24 (each in 6 of 7 cases of 8q gain).
Correlations with Histopathologic Grading
The average number of genomic aberrations detected differed according to histopathologic subtype (0.0, 1.0, 2.1, and 1.2 in PIN, WDCa, MDCa, and PDCa samples, respectively, when all cases were included and 2.0, 2.9, and 3.0 in WDCa, MDCa, and PDCa samples, respectively, when only cases with abnormal findings were considered). These differences were not found to be statistically significant by the Kruskal-Wallis test, probably due to the limited number of PIN and PDCa samples in the current study. When PIN and PDCa samples were grouped together with WDCa and MDCa samples, respectively, the most poorly differentiated tumors were found to have significantly more genomic aberrations compared with other tumors (P = 0.048; Mann–Whitney test).
To evaluate potential correlations between specific genomic abnormalities and histologic tumor differentiation, we assessed the frequencies of the most common aberrations by tumor type (Table 2). Loss at 16q was found to be statistically significantly associated with poorly differentiated phenotypes (frequency: 0%, 46%, and 100% in WDCa, MDCa, and PDCa samples, respectively; P = 0.019).
|Copy number abnormality||Histologic type||T status|
|WDCa (%)||MDCa (%)||PDCa (%)||P||T1 (%)||T2 (%)||T3 (%)||P|
|8q+||1 (25)||5 (39)||1 (50)||0 (0)||4 (36)||3 (75)|
|8p−||3 (75)||7 (54)||1 (50)||2 (67)||5 (46)||3 (75)|
|16q−||0 (0)||6 (46)||2 (100)||0 (0)||4 (36)||3 (75)|
|13q−||0 (0)||7 (54)||0 (0)||1 (33)||4 (36)||1 (25)|
|8q+/16q−||1 (25)||8 (62)||2 (100)||0 (0)||5 (46)||4 (100)|
Correlations with Clinical Stage
Among samples with abnormal findings, the average number of genomic aberrations increased with increasing clinical stage (1.7, 2.7, and 3.5 for T1, T2, and T3 tumors, respectively; P = 0.14). To assess potential correlations between specific genomic aberrations and clinical stage, the frequencies of the most common abnormalities were compared across clinical stage categories (Table 2). The findings of gains at 8q and 16q (Fig. 3) were found to be statistically significantly associated with advanced clinical stage (frequency: 0%, 36%, and 75% in T1, T2, and T3 tumors, respectively; P = 0.048); at least 1 of these 2 aberrations was documented in 0%, 46%, and 100% of T1, T2, and T3 lesions, respectively (P = 0.01).
We have demonstrated for the first time, to our knowledge, that it is possible to consistently obtain complete genomic information by analyzing prostate carcinoma specimens obtained via ultrasound-guided needle biopsy. Screening of the entire genome for clonal genetic changes is essential, as information regarding which specific areas of the genome have greater clinical relevance remains insufficient. Karyotyping and CGH both were feasible in most biopsy samples (97% and 94%, respectively), but CGH detected genetic abnormalities in a larger percentage of specimens (58%) compared with G-banding and karyotyping (38%). Combined use of these techniques revealed genetic rearrangements in 69% of the 35 carcinoma biopsy specimens analyzed and thus yielded a more comprehensive and detailed picture of the chromosomal alterations involved in prostate carcinogenesis. Nonetheless, there often were discordant findings in cases in which both techniques detected clonal aberrations. CGH is especially useful in detecting genomic imbalances that are present in a large proportion of the cells in a test sample, whereas rare clones and balanced rearrangements will remain undetected using this method.13, 15 In contrast, although karyotyping has the potential to detect both balanced and unbalanced chromosomal alterations, its success depends on the mitotic activity of the neoplastic cells in vitro.17 We used a specialized culture medium that selected against nonepithelial cells,10 but nonetheless, we cannot exclude the possibility that culture bias between malignant and nonmalignant epithelial cells affected our findings. These methodologic constraints probably account for most of the observed discrepancies between the results of CGH and the results of karyotyping analysis. In the diagnostic setting, interphase fluorescence in situ hybridization with selected probes may be performed in parallel biopsy samples obtained from specimens that have yielded normal findings on genomic screening.
The pattern of genetic changes detected by CGH analysis and G-banding in prostate carcinoma specimens obtained via ultrasound-guided needle biopsy did not differ significantly from aberration patterns previously documented in prostatectomy specimens containing clinically localized disease.8, 10, 18–21 This finding suggests that genomic aberrations detected in the minute biopsy samples examined in the current study are characteristic of the tumors from which these samples originated. There are two potential explanations for why approximately one-third of all prostate biopsy specimens investigated were not found to have any genetic abnormalities: because morphologically neoplastic lesions were documented in all cases, either the biopsy samples used in the genetic analyses differed from the parallel biopsy samples used in the histopathologic analyses, or else a proportion of morphologically malignant lesions arising in men with elevated PSA levels did not harbor genomic alterations that could be detected on CGH or karyotyping analysis.
Data on the locations and types of chromosome-level aberrations that occur in prostate carcinoma can provide insight into the molecular consequences of these abnormalities, as the smallest regions of overlap for copy number losses and gains correspond to the locations of pathogenetically significant tumor suppressor genes and oncogenes, respectively. Loss at 8p was the most common genomic imbalance observed on CGH, and NKX3.1, a prostate-specific, androgen-regulated homeobox gene located within this region (at 8p21), has been proposed to be a target tumor suppressor gene.22 Nonetheless, observed deletions sometimes encompass 8p23 only (this band is lost in 53% of all abnormal cases [i.e., as frequently as 8p21–p22]), and the only known candidate gene present at this location is CSMD1.23 Known tumor suppressor genes such as CDH1 (16q22), RB1 (13q14), BRCA1 (13q12–q13), and TP53 (17p13) have been proposed to be the relevant targets associated with commonly observed losses at 16q, 13q, and 17p; however, the bands that were most commonly absent from 16q and 13q on CGH analysis were 16q24 (42%) and 13q21 (32%), and to date, no promising candidate targets have been identified at these loci.
Losses at 1q42–q44, 10p13–pter, and 12p12–p13 also were observed in multiple instances. Whereas losses at 1q were detected by CGH in two cases, chromosome banding analysis identified an additional case that harbored del(1)(q41q43). It is noteworthy that the prostate carcinoma susceptibility locus PCAP was mapped to 1q42.2–q43 in a subset of 47 French and German families with early-onset prostate carcinoma,24 although no family history of this malignancy was reported by any of the 3 patients found to have 1q aberrations in the current study. A promising candidate target gene for the del(10)(p13) identified by G-banding is KLF6, a zinc finger transcription factor that recently was found to have 1 allele deleted in 77% of prostate carcinomas and to have accompanying point mutations in the remaining allele in 71% of these cases.25 With regard to the CGH-detected case of loss at 12p, a promising candidate target gene associated with prostate carcinogenesis has not yet been identified.
Copy number gains were rarer than losses, although they were relatively common on the long arm of chromosome 8. Although gains affected most of 8q in the majority of cases, two distinct regions of genomic gain could be discerned: one at 8q21, and another at 8q24. Candidate target genes at 8q23–q24 that have been found to be amplified and overexpressed include EIF3S3 (8q23), MYC (8q24.1), and PSCA (8q24.2).26 To date, however, no candidate target gene has been identified at the more proximal 8q21 locus.
A statistically significant correlation was found between increasing number of genomic abnormalities and more poorly differentiated histologic tumor subtype. This finding is consistent with previous CGH data indicating the presence of a significant correlation between progressive disease and increasing quantity of copy number aberrations, even when stage and grade were taken into account on multivariate analysis.8 In addition, loss at 16q was significantly associated with high histologic grade and advanced clinical stage. Associations between loss at 16q23–q24 and prostate carcinoma progression and metastasis also have been noted on loss of heterozygosity analysis.27 Furthermore, in the current study, a statistically significant association was found between gain at 8q and advanced disease stage, an observation that has been made previously by others.8, 21 Any practical role that these acquired genetic abnormalities might have as prognostic indicators for patients with prostate carcinoma must nonetheless await confirmation in long-term follow-up studies of larger series of patients.
We conclude that genomewide analysis of prostate carcinoma specimens obtained via ultrasound-guided needle biopsy is feasible. Thus, it may be possible to use genetic markers to obtain diagnostic and/or prognostic information that is useful in the making of preoperative decisions regarding prostate carcinoma management. Relevant genetic information may allow physicians to offer potentially curative surgery only to those who will benefit from it while sparing other patients who have less virulent disease from the significant side effects associated with such aggressive treatment.
- 9Histological typing of prostate tumours. Geneva: World Health Organization, 1982., , .
- 11MitelmanF, editor. ISCN (1995): an international system for human cytogenetic nomenclature. Basel: Karger, 1995.