A

p

c

dawaa

m

K

1

(tasifis

H

1d

Original article

Genetic profile identification in clinically localized prostate carcinoma�

Michele Gallucci, M.D.a, Roberta Merola, Ph.D.b, Costantino Leonardo, M.D.a,Piero De Carli, M.D.a, Antonella Farsetti, M.D.c, Steno Sentinelli, M.D.d,

Isabella Sperduti, Ph.D.e, Marcella Mottolese, Ph.D.d, Paolo Carlini, M.D.f,Erika Vico, Ph.D.b, Giuseppe Simone, M.D.a, AnnaMaria Cianciulli, Ph.D.b,*

a Department of Urology, Regina Elena Cancer Institute, Rome, Italyb Department of Clinical Pathology (Cytogenetic Unit), Regina Elena Cancer Institute, Rome, Italy

c Institute of Neurobiology and Molecular Medicine, National Research Council, Rome, Italyd Department of Pathology, Regina Elena Cancer Institute, Rome, Italy

e Biostatistic Unit, Regina Elena Cancer Institute, Rome, Italyf Department of Oncology, Regina Elena Cancer Institute, Rome, Italy

Received 12 February 2008; received in revised form 27 March 2008; accepted 2 April 2008

bstract

Purpose: To confirm our previously obtained results, we genetically characterized prostate cancer from patients undergo radicalrostatectomy in a retrospective study.Materials and methods: Histological sections were evaluated for 106 patients treated with surgery from 1991 to 2004. With fluores-

ence in situ hybridization (FISH) method, the status of LPL (8p22), c-MYC (8q24) genes and 7, 8, X chromosomes was evaluated.Results: Chromosomes 7, 8, X aneusomy was demonstrated in 91.5%, 78.3%, and 51.9% of the samples, respectively, whereas LPL

eletion and MYC amplification were found in 76.0% and 1.6%. A genetic profile was considered as unfavorable when at least twoneusomic chromosomes and one altered gene were present. Tumors with an adverse genetic profile were more frequently present in patientsith higher stages (P � 0.02), biochemical/clinical progression (P � 0.03), and Gleason grade 4 � 3 (P � 0.02). Multiple correspondence

nalysis identified one tumor group characterized by chromosome 8 aneusomy, X polysomy, LPL gene deletion, Gleason � 7 and 4 � 3ssociated with progression.

Conclusions: In this study, we recognized the predictive power of previously identified cytogenetic profiles. Assessment of genetic setay characterize each patient and have influence on postoperative therapeutic strategies. © 2009 Elsevier Inc. All rights reserved.

Urologic Oncology: Seminars and Original Investigations 27 (2009) 502–508

eywords: Prostate cancer; Cancer genetics; Chromosomal abnormalities; Prognostic studies

Pt

ttgtoc([(fA

. Introduction

The prognosis and choice of therapy for prostate cancerPC) are based primarily on 3 parameters obtained at theime of diagnosis: clinical stage, serum prostate-specificntigen (PSA), and degree of tumor differentiation (Gleasoncore), however, these do not provide enough predictivenformation regarding the clinical outcome [1]. The identi-cation of patients who have an increased risk of progres-ion and/or postoperative recurrence is an important goal for

� This work was supported by grants from the Italian Ministry ofealth and the Italian Association for Cancer Research (AIRC).

* Corresponding author. Tel./fax: �39-6-52665966.

rE-mail address: cianciulli@ifo.it (A.M. Cianciulli).

078-1439/09/$ – see front matter © 2009 Elsevier Inc. All rights reserved.oi:10.1016/j.urolonc.2008.04.008

C research, as such patients could be candidates for newerreatments and follow-up strategies.

The goal is to recognize, through emerging technologies,he genetic profiles driving the aggressiveness of PC. Withhe rapid development of molecular cytogenetics, a series ofenetic alterations on multiple chromosomes has been de-ected in PC [2,3]. Deletion in sporadic PC most commonlyccurs at chromosomal locus 8p21–22 [4,5,6]; 8p22 loss,oncurrent with the gain of copy number of chromosome 8aneusomy), may successfully predict disease recurrence7]. The commonly deleted region of 8p22 includes the LPLlipoprotein lipase) gene that is suggested to be responsibleor the initiation or early event in prostate tumorigenesis.nother important genetic alteration in PC is 8q24 overrep-

esentation, which is commonly found in advanced, meta-

stabakpratdic

hati

i(tiafsb

(pipt(Lt

2

2

ei1cnpidtriat

wcCtbcwplswbcpuc

2

Dt(gpXfit2Gt

TC

C

NP

H

G

L

F

503M. Gallucci et al. / Urologic Oncology: Seminars and Original Investigations 27 (2009) 502–508

tatic, and androgen-independent PC [8]. The region con-ains an oncogene MYC, which regulates cell proliferationnd apoptosis [9]. Aneusomies of chromosomes 7 have alsoeen observed frequently with the gain of chromosome 7nd loss of 7q31 [10]. Chromosome 7 alterations werenown to be associated with higher tumor grade, advancedathological stage and poor prognosis [11]. The androgeneceptor gene located on chromosome Xq11–13 encodes thendrogen receptor protein through which androgens exertheir intracellular regulation of prostate growth and cellularifferentiation [12]. Additional androgen receptor gene cop-es are present in patients with PC due to polysomy of thehromosome X [13].

Our previous contribution to this area of investigationas demonstrated that the combined 7, 8, X chromosomesnd MYC (8q24), LPL (8p22) gene anomaly patterns iden-ify cytogenetic profiles as additional markers to patholog-cal features in clinically localized prostate carcinoma [14].

In order to develop a more detailed understanding of thenvolvement of chromosomes 7, 8, X and MYC (8q24), LPL8p22) gene anomalies in human PC and to define the poten-ially predictive biological profiles for a bad or good prognosis,n this study we confirmed our previously obtained results bynalyzing independent set samples from patients with adequateollow-up in a retrospective study. The criteria for patientelection were absence of hormonal neoadjuvant treatmentefore surgery and diagnosis of clinically localized disease.

These results allowed us to identify a poor genetic profileat least 2 aneusomic chromosomes and 1 altered generesent), that can be integrated with the grade and clinicalnformation. The aim would be to clearly divide canceratients into 2 groups, with indolent or aggressive prostateumors. Moreover, the association of specific genetic lesionschromosome X polysomy, chromosome 8 aneusomy, andPL gene deletion) with progression may help to improve

herapeutic options after surgery.

. Materials and methods

.1. Patients

The samples included in the previous investigation werexcluded in this validation study, which was done on anndependent set. Histological sections were evaluated for06 patients treated with radical prostatectomy for clinicallyircumscribed PC from 1991 to 2004. Patients who receivedeoadjuvant treatment were excluded from the study. Com-lete information on pretreatment PSA-values was availablen all patients. Postoperative serial PSA measurements wereone semiannually within the first 2 years and annuallyhereafter. A postoperative PSA-level of 0.1 ng/ml andinsing concentrations were considered as biochemical ev-dence of tumor recurrence. All samples were reviewed by

pathologist with experience in uropathology. The largest

umor focus and/or the focus with the worst Gleason grade h

ere marked on the slides. The pathologic stage for eachase was assigned according to the Union Internationaleontre le Cancer (UICC, 2002) TNM system [15]. In 23

umors with bilateral involvement, the slides obtained fromoth lobes were selected for genetic evaluation, which wasonsequently performed on 129 foci. The Gleason grade (G)as determined for each prostate carcinoma focus. All clinico-athological characteristics are illustrated in Table 1. Fol-ow-up data were obtained from hospital charts and corre-pondence with the referring physicians. In these patients,ith a median follow-up of 48 months, 34 relapses (18iochemical and 16 clinical progression), and 10 prostateancer-related deaths were recorded. As a negative controlopulation, normal prostatic tissue samples from patientsndergoing cystectomy were used. In addition, as positiveontrol we used prostatic cell lines (LNCaP).

.2. Fluorescence in situ hybridization (FISH) analysis

The Vysis ProVysion Multi-color mixture (Vysis, Inc.,owners Grove, IL) was used for detection and quantifica-

ion of 8 chromosome labeled with SpectrumAqua, LPL8p22) gene labeled with SpectrumOrange, and c-MYCene (8q24) labeled with SpectrumGreen. We also em-loyed chromosome enumeration probes (CEP) specific for

and 7 chromosomes (Vysis, Inc.). In brief, after deparaf-nization, specimens were incubated in pretreatment solu-

ion (80°C, 20 min) and then digested with protease (37°C,5 min) (Paraffin Pretreatment Kit II; Vysis, Inc. Downersrove, IL). The probes were applied, a coverslip sealed to

he slide, and the specimens denatured (75°C, 5 min) and

able 1linicopathologic characteristics

haracteristics n Cases %

o patients 106SA0–3.9 8 7.54–9.9 25 23.610–19.9 31 29.2�20 18 17.0

istologic stageT2a–2b 15 14.1T2c 42 39.6T3a–3b 47 44.4T4 1 0.9

leason grade�7 30 23.33 � 4 50 38.84 � 3 22 17.1�7 27 20.9

ymph nodeNegative 99 76.7Positive 2 1.6

ollow-up (months)Median 48.00Range 6–178

ybridized overnight (37°C) in a humidified chamber (Hybrite

DtaMnsmccgFf

2

pb

eavfiiaao[iidgaptcu

3

3s

(t1eaao1(m

Fc

TS

P

U

F

T

504 M. Gallucci et al. / Urologic Oncology: Seminars and Original Investigations 27 (2009) 502–508

enaturation/Hybridization System; Vysis, Inc.). Slides werehen washed, counterstained with DAPI, and analyzed underfluorescence microscope (BX61, Olympus Italia, Segrate,ilano, Italy). In each selected PC focus, an average of 100

uclei was enumerated. Tumoral foci were considered aneu-omic if the percentage of nondisomic nuclei exceeded theean �3SD of any of the signal categories observed in the

ontrol group. We used a ratio of more than 2 oncogene/entromere signals to define gene amplification, whereasene deletion was determined by a ratio lower than 1. TheISH assay was not available for 20 genetic determinationsor technical causes.

.3. Statistics

The association between unfavorable identified geneticatterns and clinico-pathological characteristics was testedy the Pearson �2 test.

ig. 1. Graphic representation of genetically altered neoplastic foci per-entage.

able 2ummary of the combined chromosomes 7, 8, X, and MYC, LPL gene an

atterns Chr. 7 Chr. 8

nfavorable genetic set(at least two aneusomic chromosomes

and one altered gene)A AA AA NormalA AA A

avorable genetic setA NormalA NormalA NormalA AA ANormal ANormal A

otal

A � aneusomic; Amp � amplified; D � deleted.

Multiple correspondence analysis (MCA), a descriptive/xploratory technique designed to analyze simple two-waynd multi-way tables, was used to evaluate which geneticariables are strictly associated with progression in identi-ed profiles. The results provide information that is similar

n nature to those produced by Factor Analysis techniques,llowing one to explore the structure of categorical vari-bles included in the table. The most common kind of tablef this type is the two-way frequency cross-tabulation table16,17]. This representation is aimed at visualizing the sim-larities and/or differences of profiles in a simultaneous way,dentifying those dimensions that contain the majority of theata variability. The position of the points in the MCAraph is informative. Categories plotting close to each otherre statistically related and are similar with regard to theattern of relative frequencies, and this association is sta-istically valuable (Lebart’s statistic) when points are lo-ated far from the origin of the graph that represents a mean,ninformative profile.

. Results

.1. Genetic profiles (chromosomes 7, 8, X and MYC, LPLtatus) in neoplastic examined foci

FISH was successfully performed on 109 out of 129 foci84.5%) for all considered genetic variables. By applyinghe cutoff values previously described [14], we defined18/129 (91.5%), 101/129 (78.3%), and 67/129 (51.9%) ofxamined foci as having aneusomy for chromosome 7, 8nd X, respectively. Only 2/117 (1.6%) samples resultedmplified for the MYC gene with a ratio of more than 2ncogene/centromere signals. As regard LPL deletion, 98/29 (76.0%) foci with a ratio lower than 1 were deletedFig. 1). Table 2 illustrates the summary of combined chro-osome and gene anomaly patterns: the subclassification of

patterns in 109 prostatic cancer foci

hr. X MYC gene LPL gene No. foci %

ormal Normal D 25 22.9Normal D 50 45.9Normal D 3 2.8

ormal Amp D 1 0.9Amp D 1 0.9

Normal Normal 4 3.7ormal Normal D 7 6.4ormal Normal Normal 4 3.7ormal Normal Normal 7 6.4

Normal Normal 4 3.7ormal Normal D 1 0.9ormal Normal Normal 2 1.8

109 100

omaly

C

NAANA

ANNNANN

Flvsp7

3p

tbvswan(c0(Fatgs(

3i

tsqetsasq

TAc

G

����P

S

TTpP

S

G

34P

505M. Gallucci et al. / Urologic Oncology: Seminars and Original Investigations 27 (2009) 502–508

ISH anomalies identified 12 patterns. The presence of ateast 2 aneusomic chromosomes and one altered gene di-ided these profiles into unfavorable and favorable geneticets. A poor genetic pattern with a higher frequency wasresent in 50 foci (45.9%) and was made up of chromosome, 8, X aneusomy and LPL gene deletion.

able 3ssociation between generic profiles with clinico-pathological

haracteristics

leason score No. foci evaluated Unfavorablegenetic patterns

7 22 15 (68.2)7 (3 � 4) 45 29 (64.4)7 (4 � 3) 22 20 (90.9)

7 22 19 (86.4)-value 0.05

tage No. patients evaluated

2a–T2c 46 31 (67.4)3–T4 42 37 (88.1)value 0.02

rogressionNo progression 57 40 (70.2)Biochemical progression 15 13 (86.7)Clinical progression 15 15 (100.0)P-value 0.03

tatusAlive 77 58 (75.3)Dead 10 10 (100.0)P-value 0.07

leason score � 7 No. foci evaluated Unfavorablegenetic patterns

� 4 45 29 (64.4)� 3 22 20 (90.9)-value 0.02

Fig. 2. Graphic representation of comparison between adverse

.2. Association between an unfavorable genetic set withathologic characteristics and clinical outcome

To prove the reproducibility of our previous results, inhe second part of this work we evaluated the associationetween identified genetic profiles and clinico-pathologicariables (Table 3). In evaluated bilateral tumors, we con-idered the worst detected genetic value. In 87 patients itas possible to perform the correlation between follow-up

nd genetic profiles. Unfavorable genetic profile levels sig-ificantly increased in tumors with elevated histologic stageP � 0.02) gradually in patients with biochemical andlinical progression (P � 0.03) and higher grades (P �.05). Moreover, the level increased with the death rateP � 0.07), however, it was not statistically significant. Inig. 2 the graphic representation of the comparison betweendverse genetic profile levels and clinical behavior is illus-rated. When only the heterogeneous prostate cancer G 7roup was considered, the level of the unfavorable geneticet significantly rose in tumors with grade 4 � 3 (P � 0.02)Table 3).

.3. Association of genetic variables with progression indentified profiles

As shown in Fig. 3, the complex interrelationship be-ween the clinico-pathological and genetic variables in ourtudy could be best evaluated by using the MCA. Thisualitative analysis permitted a visual illustration amongvaluated variables and showed the association betweenhese biologic parameters with progression. Values of ab-olute and relative contributions produced by statisticalnalysis (Lebart’s statistic) showed the contrast between 2ets of tumors along the first axis. The first group (upper leftuadrant) is characterized by G � 7, G � 4 � 3, chromo-

genetic profile levels with progression and death rate.

sdssGmtnwbLut

4

cimdtoeisdm

mbpg

otn

psfatcaupoih

o[wpwatalsisamt

uacpoa[eau3Gcntp

Fv7ca

506 M. Gallucci et al. / Urologic Oncology: Seminars and Original Investigations 27 (2009) 502–508

ome X polysomy, chromosome 8 aneusomy, LPL geneeletion. As illustrated in the complementary quadrant, theecond group of tumors is typified by association of theame genetic variables as the first group not altered and

3 � 4. In order to obtain an indication of the prognosticeaning of the graphical configuration produced by MCA,

he absence or the presence of progression was added as aovel, supplementary variable. The presence of progressionas associated with the first group of tumors characterizedy chromosome X polysomy, chromosome 8 aneusomy,PL gene deletion ,and G 4 � 3 and � 7. As shown in thepper right quadrant, the absence of progression was relatedo normal chromosome 7, stage T1-2 and G � 7.

. Discussion

At present, PSA, Gleason score, ethnicity, and the per-ent of positive prostate biopsies appear to be the mostmportant pretreatment predictors of disease recurrence inen after radical prostatectomy [18]. One of the significant

rawbacks of this classification system, however, is thathey only have limited use in predicting the difference in theutcomes observed between patients diagnosed with PCxhibiting similar clinical, histopathological, and biochem-cal features. PC is a highly heterogeneous disease thathould ideally be subcategorized according to the geneticefects, which may reflect prognostic and predictive infor-

ig. 3. MCA plot showing the projections of the identified biologicalariables (active information) and progression (passive information). Chr, 8, X, MYC, LPL; n � normal chromosome and gene, Chr 7, 8; a �hromosomal aneusomy; Chr X p � chromosomal polysomy, MYC amp �mplified, LPL d � deleted, G � Gleason score, T � stage.

ation, in order to assure optimal and individualized treat- c

ent for patients. Interest in novel prognostic markers isased on the fact that a significant number of patients afterotentially curative radical prostatectomy have disease pro-ression.

In our previous study [14], we evaluated the associationf specific molecular cytogenetic characteristics with his-opathological classification and detected unfavorable ge-etic signatures related to histological variables.

In this work our goal was (1) to confirm the predictiveower of the identified profiles by using a set of PC afterurgery from a larger number of patients with adequateollow-up in a retrospective study, (2) to detect the associ-tion of specific genetic variables with progression in iden-ified profiles. To this end, we examined the involvement ofhromosomes 7, 8, X, and MYC (8q24), LPL (8p22) genenomalies in 129 neoplastic foci from 106 men who hadndergone radical prostatectomy for clinically localizedrostate cancer with a median follow-up of 48 months. Eachf these genetic alterations has been analyzed independentlyn this larger set of samples, confirming the association withigher stage and grade (data not shown).

Numerous authors have reported the single involvementf chromosome 7, 8, and X aneusomy in PC. Alers et al.19] demonstrated that extra copies of chromosome 7 and 8ere observed more frequently in tumors derived fromrogressors than in nonprogressors. Moreover, in patientsith negative surgical margins, gain of chromosome 7 was

ssociated with biochemical failure after radical prostatec-omy [20]. Copy number changes of 8p22, centromere 8,nd 8q24 (MYC) in a large cohort of patients with patho-ogic organ-confined PC were associated with poor progno-is, specifically when it was accompanied by an additionalncrease of 8q24 [21]. It showed that additional X chromo-omes are present in a subgroup of primary PC prior tontiandrogen therapy. Additional androgen receptor copiesay be a factor leading to a poor effect of antiandrogen

herapy [13].In our study, we highlighted the potential advantage of

sing a panel of markers over a single marker in diagnosticnd/or prognostic applications. The alteration of at least 2hromosomes and one 1 were considered an unfavorablerofile, and our results showed that this can be an indicatorf poor prognosis, being significantly associated with stagend grade as shown in Table 3. According to Mian et al.22] who demonstrated different molecular genetic patternsxplaining the different biological behavior of the G 3 � 4nd 4 � 3 groups, we found that the percentage of thenfavorable genetic set increased significantly in the G 4 �group. From a clinical point of view, it is known that thescore 7 group is very heterogeneous and there is a distinct

linical outcome between the two subgroups. These resultsot only confirm our previous opinion, but also underlinehat genetic features could help to stratify this group ofatients, and thereby influence the therapeutic decision pro-

ess.

phuccsdtfro

wctbhlmtebMrbv(L8

rtmirtlapstdartbdsaeTLrno

5

bdtmcalccttttit

A

EVp

R

[

507M. Gallucci et al. / Urologic Oncology: Seminars and Original Investigations 27 (2009) 502–508

Moreover, as illustrated in Fig 2, we found that theresence of at least 3 altered genetic variables is sufficient toave significant impact on the clinical outcome. In fact, thenfavorable profile levels significantly and gradually in-reased in the tumors of patients with biochemical andlinical progression (P � 0.03) even though it was nottatistically significant in the tumors of dead patients. Ourata suggest that identified cytogenetic profiles have a po-ential for development of clinical prognostic tests suitableor stratification of PC patients at the time of diagnosis withespect to the likelihood of a negative or positive clinicalutcome after radical prostatectomy.

After definition of genetic profiles and their correlationith clinical outcome, one of our aims was to evaluate the

omplex interrelations among examined variables by testingheir prognostic significance. To this end, we used MCA toetter visualize which phenotypic and genetic parametersave an influence on progression in our series of clinicallyocalized prostate tumors. In our graphical analysis, chro-osome X polysomy, 8 aneusomy, and LPL deletion seem

o be strictly associated with traditional unfavorable param-ters, such as G score 4 � 3 and � 7, indicating adverseiological parameters consistently related to progression.oreover, lower stages are related to no progression. The

esults of this qualitative statistical analysis were confirmedy generating 2 tumor clusters in which the same geneticariables (altered and not altered) were strictly associatedtumors with chromosomes X polysomy, 8 aneusomy, andPL gene deletion vs. tumors with normal chromosomes X,, and LPL gene).

Because the histologic classification of tumors fromadical prostatectomy does not always permit the predic-ion of the clinical course, substantial efforts should beade to further subclassify these types of tumors. The

dentification of genetic prognostic markers that bettereflect tumor biology is eminent. For many cancer types,umor progression can be characterized by the accumu-ation of complex chromosome alterations. Strohmeyer etl. [23] associate genetic aberrations in PC with tumorrogression measured by microvessel density. Recently,everal studies have attempted to stratify PC according toheir molecular profiles, to offer an alternative means toistinguish aggressive tumor biology and to improve theccuracy of outcome prediction for PC patients treated byadical prostatectomy [24,25,26]. We have already iden-ified an adverse genetic prognostic profile in breast andladder cancer [27,28]. In this work, we prove the repro-ucibility of the cytogenetic signatures, obtained by as-embling various well-known PC chromosomal and genelterations, to define a complete panel that could bemployed by urologists for better neoplasia stratification.herefore, chromosomes X polysomy, 8 aneusomy, andPL gene deletion associated with some phenotypic pa-

ameters detected by MCA proved to be a better prog-ostic indicator than the value of either biomarker on its

wn.

. Conclusions

PC remains a heterogeneous and complex disease. Aetter understanding of the molecular genetics of thisisease holds promises for earlier detection and improvedreatment for the many men affected by this tumor. Nu-erous studies remain to be carried out in order to

orrelate the identified genetic profiles and molecularnomalies with tumor prognosis. The assessment of bio-ogic factors predictive of tumor aggressiveness mayharacterize each patient, even within the same histologiclassification, and have an influence on postoperativeherapeutic strategies. Our results provide evidence ofhose who might benefit from particular and innovativeherapeutic interventions. Each of these genetic parame-ers individually and together (biological profiles) couldmprove the understanding of this disease and contributeo stratifying patients with PC.

cknowledgments

The authors thank Mrs. Paula Franke for the formalnglish revision of the manuscript and Giulia Orlandi,alentina Pinetti, and Carlo Del Carlo for technical sup-ort.

eferences

[1] Partin A, Kattan MW, Subong EN, et al. Combination of prostate-specific antigen, clinical stage, and Gleason score to predict patho-logical stage of localized prostate cancer. A multi-institutional up-date. JAMA 1997;277:1445–51.

[2] Das K, Lau W, Sivaswaren CR, et al. Chromosomal changes inprostate cancer: A fluorescence in situ hybridization study. Clin Genet2005;68:40–7.

[3] Karan D, Lin MF, Johansson SL, et al. Current status of moleculargenetics of human prostatic adenocarcinomas. Int J Cancer 2003;103:285–93.

[4] Vocke CD, Pozzetti RO, Bostwick DG, et al. Analysis of 99 micro-dissected prostate carcinomas reveals a high frequency of allelic losson chromosome 8p21–22. Cancer Res 1996;56:2411–6.

[5] Swalwell JI, Vocke CD, Yang Y, et al. Determination of a minimaldeletion interval on chromosome band 8p21 in sporadic prostatecancer. Genes Chromosomes Cancer 2002;33:201–5.

[6] Bethel CR, Faith D, Li X, et al. Decreased NKX3.1 protein expres-sion in focal prostatic atrophy, prostatic intraepithelial neoplasia andadenocarcinoma: Association with Gleason score and chromosome8p deletion. Cancer Res 2006;107:10683–90.

[7] Macoska J, Trybus T, Wojno K. 8p22 Loss concurrent with 8c gainis associated with poor outcome in prostate cancer. Urology 2000;55:776–82.

[8] Tsuchiya N, Kondo Y, Takahashi A, et al. Mapping and gene expres-sion profile of the minimally overrepresented 8q24 region in prostatecancer. Am J Pathol 2002;160:1799–05.

[9] Amati B, Alevizopoulos K, Vlach J. Myc and the cell Cycle. FrontBiosci 1998;3:D250–68.

10] Cui J, Deubler DA, Rohr LR, et al. Chromosome 7 abnormalities inprostate cancer detected by dual color FISH. Cancer Genet Cytogenet

1998;107:51–60.

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

508 M. Gallucci et al. / Urologic Oncology: Seminars and Original Investigations 27 (2009) 502–508

11] Bandyk MG, Zhao L, Troncoso P, et al. Trisomy 7, a potentialcytogenetic marker of human prostate cancer progression. GenesChromosomes Cancer 1994;9:19–27.

12] Gelmann EP. Molecular biology of the androgen receptor. J ClinOncol 2002;20:3001–15.

13] Ropke A, Erbersdobler A, Hammerer P, et al. Gain of androgenreceptor gene copies in primary prostate cancer due to X chromosomepolysomy. Prostate 2004;59:59–68.

14] Gallucci M, Merola R, Farsetti A, et al. Cytogenetic profiles asadditional markers to pathological features in clinically localizedprostate carcinoma. Cancer Lett 2006;237:76–82.

15] UICC. TNM Classification of Malignant Tumors, 6h ed. New York:Wiley-Liss, 2002.

16] Greenacre MJ. Theory and Applications of Correspondence Analysis.London: Academic Press, 1984.

17] Lebart L, Morineau A, Warwick K. Multivariate descriptive statisticalanalysis. New York: Wiley, 1984.

18] Grossfeld GD, Latini DM, Lubeck DP, et al. Predicting recurrenceafter prostatectomy for patients with high risk prostate cancer. J Urol2003;169:157–63.

19] Alers JC, Rochat J, Krijtenburg PJ, et al. Identification of geneticmarkers for prostatic cancer progression. Lab Invest 2000;80:931–42.

20] Skacel M, Ormsby AH, Pettay JD, et al. Aneusomy of chromosomes7, 8, and 17, and amplification of HER-2/neu and epidermal growthfactor receptor in Gleason score 7 prostate carcinoma, a differentialfluorescent in situ hybridization study of Gleason pattern 3 and 4using tissue microarray. Hum Pathol 2001;32:1392–7.

21] Tsuchiya N, Slezak JM, Lieber MM, et al. Clinical significance of

alterations of chromosome 8 detected by fluorescence in situ hybrid-

ization analysis in pathologic organ-confined prostate cancer. GenesChromosomes Cancer 2002;34:363–71.

22] Mian C, Lodde M, Comploj E, et al. Molecular biological analysis ofthe heterogeneous prostate cancer group Gleason Score 7. Prostate2006;66:966–70.

23] Strohmeyer DM, Berger AP, Moore DH, et al. Genetic aberrations inprostate carcinoma detected by comparative genomic hybridizationand microsatellite analysis, association with progression and angio-genesis. Prostate 2004;59:43–58.

24] Stephenson AJ, Smith A, Kattan MW, et al. Integration of geneexpression profiling and clinical variables to predict prostate carci-noma recurrence after radical prostatectomy. Cancer 2005;104:290–8.

25] Nanni S, Priolo C, Grasselli A, et al. Epithelial-restricted gene profileof primary cultures from human prostate tumors. A molecular ap-proach to predict clinical behavior of prostate cancer. Mol Cancer Res2006;4:79–92.

26] Glinsky GV, Glinskii AB, Stephenson AJ, et al. Gene expressionprofiling predicts clinical outcome of prostate cancer. J Clin Invest2004;113:913–23.

27] Mottolese M, Orlandi G, Sperduti I, et al. Biopathologic charac-teristics related to chromosome 11 aneusomy and cyclin D1 genestatus in surgically resected stages I and II breast cancer, identi-fication of an adverse prognostic profile. Am J Surg Pathol 2007;31:247–54.

28] Gallucci M, Vico E, Merola R, et al. Adverse genetic prognosticprofiles define a poor outcome for cystectomy in bladder cancer. Exp

Mol Pathol 2007;83:385–91.