Identification of Numer ‘al Chromosomal Changes Detected by Interphase Fluorescence In Situ Hybridization in High-Grade Prostate Intraepithelial Neoplasia as a Predictor of Carcinoma Jaudah AlMaghrabi, MD, FRCPC; Lada Vorobyova, MD; A. Toi, MD, FRCPC: William Chapman, MD, FRCPC: Maria Zielenska, PhD; Jeremy A. Squire, PhD ‘© Context.—High-grade prostate intraepithelial neoplasia (HPIN) is the most likely precursor of prostate cancer. The condition of many patients with a diagnosis of HPIN in prostate needle core biopsy could, if left untreated, pro- ‘ress to invasive cancer. Currently there is no available I, immunohistochemi that are predictive of this progression, ‘Objective—To determine whether chromosomal insta- bility in these precursor lesions could increase their pre- dictive value for cancer detection. Design.—Dual-color interphase fluorescence in situ hy- bridization analysis was performed on archived prostate needle core biopsies from 34 patent with ital dagnos of isolated HPIN and follow-up of 3 years or more. We used commercially available centromere probes for chi mosomes 4, 7, 8, and 10. We had interpretable results in 44 patients as follows: (1) group A: 24 HPIN patients wi persistent HPIN and/or benign lesions in the follow-up bi- Presta intraepithelial neoplasia (PIN) is characterized by intrahuminal proliferation of epithelial cells and can be divided into high-grade (HPIN) and low-grade le- sions, HPIN is the earliest accepted stage in prostatic car- cinogenesis. HPIN is regarded as the most likely precursor for prostate carcinoma.’ The greatest value of PIN is its high predictive value as a marker for prostate carcinoma, ‘This is particularly true for HPIN. If this lesion is identi- fied, close surveillance and follow-up biopsy are indicat- ed In many patients, HPIN is diagnosed by a prostate needle core biopsy, and if left untreated, such patients’ conditions could progress to invasive carcinoma. HPI when diagnosed in a needle biopsy, is a powerful predic ‘Accepted fo pblieation Septmber 28,2001 Trom the Ontario Cancer Inst, Princess Margaret Hospital, To tonto, Ontario, Canad (Drs AcMaghiab, Vorobyor, Chapman ae Sure); Toronto General Hospital Toronto, Ontario Cad 1 Saue)Unvnty Heath Netwerk, Yorn, Ontario, Cauda Os A Maghab, Vorobyova, Chapman and Saute andthe Deparment Laboratory Medicine and Pathbilogy TOs AL aghabl Vorobyove Chapman, and Squire), Medical iophysc (Or Sq), Redoloy Ors AbMaghiobi and'eb, and Pathology (r Zelemsta, Hospital er Sick ieen,Faciyo Medici, Unies Tro Tto, Ona, Canada Reprints: J A. Sque, PRO, Division of Cellular and Molecular Bi- logy: Ontario Cancer insite, Princess Margaret Hospi 610 Unc ‘erst Ave, Room 9-72, Torani, Ontaria,Canasa M6 2M9 lel Jeremyaqsieunorontoh ‘Arch Pathol Lab Med—Vol 126, February 2002 ‘opsies, and (2) group B: 20 HPIN patients with progression 10 prostate carcinoma. Results —Iwenty-five percent of the patients in group 8 displayed numeric chromosomal aberrations. Only 8.3% of the patients irom group A had chromosomal abnormalities (P = 1). The observed overall chromosomal changes in HPIN were higher than those in normal or hyperplastic ep- itheiom, witha staal sigiticantdference (P< 05), All aberrations were detected in the form of chromosomal gain, Overall, the commonest aberration was gain of chro- mosome 8, followed by gains of chromosomes 7 and 10. Conclusion.—These results indicated that although no single numeric chromosomal abnormality could be as- signed as a predictor of HPIN progression to carcinoma, the overall level of numeric chromosomal abnormalities shows a trend of elevation in HPIN patients whose con- dition subsequently progressed to carcinoma. (Arch Pathol Lab Med. 2002;126:165-169) tor of carcinoma in subsequent needle biopsies. No clinical cr pathologic parameter has been found to be helpful for distinguishing patients who have carcinoma on the next biopsy from those who do not. The term chromosomal instability refers to a high fre- quency of chromosomal loss and gain. It is a type of ge- nomic instability that has been introduced recently to the field of human cancer biology. Recently chromosomal in- stability has been described in many human dysplastic le- sions, and it has been proposed as a marker of progression to cancer and considered as a primary event in neoplastic transformation Whether chromosomal. instability in HPIN can provide additional predictive information for ly cancer progression is still unknown. The objective of this project was fo examine if chromosomal instability can increase the predictive value of HPIN diagnosis by usin interphase fhiorescence in situ hybridization (FISH) applied on prostate biopsies with a diagnosis of HPIN only. MATERIALS AND METHODS Patient Accrual ‘Between 1995 and 1997 in the record of The University Health Network, we had identified 123 patients with HPIN as the pri- mary diagnosis in prostate needle core biopsies, and we com ppared thet with 100 patents with HPIN associated with invasive Cancer simultaneously. HPIN associated with cancer was desig nated as group 1, and isolated HPIN was designated as group 2 Group 2 was subdivided as follows: (Chromosomal Changes to Predict Carcinema—Al-Maghrabi etal 165Table 1. Age and Prostate-Specific Antigen (PSA) Age. igen (PSA) | Levels in Patients Used for Fluorescence In Situ Hybridization Studies | ‘Mean Age, Mean PSA, | croup Y git PValve a 651 387 5.05 8 82 73 305 Group A: patients with HPIN as a primary diagnosis in pros- lati Bapsien who didnot show evidence 0 “Croup B pases ea HPN ali oup Bs paints with HIN asa primary diagnosis in pros- tac biopsies who di show invasive carcinoma on subsequent biopsies. ‘The clinical data were reviewed fr all he paints inhuding age, prostaespecic antigen (PSA) level, digital rectal exami tation (DRE), and transrectal ltasound examination. For the interphase FISH, 30 patents from group A sn 24 patients rom froup B were matched fr age and BSA level (Table). Interphase FSH wee performed only onthe inal opses Materials and Methods All the hematoxlineosin slides of the cates diagnosed as HAIN wc vena y 2 pablo cnn he got tnd to determine the adequacy of te specmen for ISH ancy Only there with suficent material were induded in the eed Interphage FISH was performed on Sun unstained Use se {ons using acjcent henatonyineoinstained ectons a ld tnce Te appropriate section was chosen Directly bed VSS E&P probes {Vg ne, Downers Grove il for Gromesomes 4, 78 fr wer ue, ee dons 7.8 an 0 hae previously been shown tbe aso with numeri change in Prostate cance Parafn prevestnent and FISH procedure fee pefonned according io the maniSctirers intractons Dun robe hybridization was performed For each probe 100 rar Cll were counted by each observer. Ten ofthe eases could not be Interpreted ether becaste of poor hybridisation, Huresence brckground or the decense nize and subsequent disappent ice of the HPIN foc onthe deeper sections In 10 cses eat ne oberver was able fo count only 75199 nce ‘carcinoma on sub- Criteria of Scoring and Evaluation of Numeric ‘Chromosomal Anomalies In preliminary experiments the hybridization efficiency of ev- cry probe was tested on prostate tissues, Slides were evaluated according to the accepted criteria." Briefly, only sections with iyi in at leat 80 of els were veined. Tor eat robe 2 independent investigators counted the numberof Fi Eignas in 100 nnoverlspped intact perc) interphase nicl from foci of HPIN. The rtumber of signals per nucieus was scored 350, 1,2,3, 4, and more than 4 signals per ruicleus, Nuclei from Stromal clement were not enumerated, FISH using a centromere probe for chromosome 4 was usel as @ negative contol, because this chromosome is not commonly seen as aneusomic in prostate carcinoma or prostatic tissues. Normal and hyperplastic glan- ‘dar epithelium present in the biopsies was counted as an in- ternal control. Becuse of truncation ofthe ruck, artifactual loss of signals is expected: however, we applied very conservative criteria 0 detect any significant true numeric changes. For the control cases using the cvomosome 8 centromere probe, the mean + 3 SD percentage of nuclei with 3 oF more signals was "46%, and the mean +3 SD percentage of nuclel with O or 1 signals was 4.5%. Our criteria to evaluate numeric chromosomal sbnormality were as follows: 1, Chromosomal gains were diagnosed when more than 4.6% of the nuclet exhibited more than 2 signals, 2. Chromosomal loss was diagnosed when more than 44.6% ofthe nuclei exhibited a reduction of signal number. 166 Arch Pathol Lab Med—Vol 126, February 2002 3, Tetraploidy was assumed when all chromosomes investi- gated showed signal gains up to 4 These cto values were in Keeping with those published esewhere Statistical Analysis, ‘The Fisher exact test was used to examine the difference be- tween group A and group B regarding the presence of numeric ‘chiomosomal changes. The same test was used to analyze the difference between group 1 and group 2 as well as. Between ‘group A and group B regarding PSA level, DRE, and ultrasound ‘examination. Student ! test was used to compare the age between the different groups. RESULTS linicopathologic Findings ‘The mean ages were 67 years and 64 years for groups 1 nee OS ees ng/ml. and 8 ng/ml. for groups 1 and 2, respectively (P <5). In group 1, the ultrasound and DRE examinations were positive in 50% and 44%, respectively, and in grouy 2, ultrasound and DRE were positive in 30% and 23%, Fespectvely which was statistically significant (P < 05 for DRE and P <1 for ultrasound). The patients with iso- lated HIPIN as the primary diagnosis (n= 123) had atleast 1 follow-up biopsy. Prostate carcinoma was identified in 33 cases (27%) in the follow-up biopsies, with range of follow-up time between 2 and 36 months. The carcinoma was identified in the same side as the HPIN in 55% of the cases; in 33% of cases it was found in both sides; and in 12% of cases it was found in the opposite side. Gleason scores were predominantly 6 and 7. When we divided the patients into group A (did not progress to cancer) and group B (progressed to cancer), we found that the mean ages for groups A and B were 64 years and 62 years (P = ‘Byand thatthe mean PSA values were 8 ng/mL and 78 ing/mL. (P = .4), respectively. Ultrasound was positive in 11 of 33 and in 18 of 90 in groups A and B, respectively (P = 1), Digital rectal examination was positive in 11 of 833 and in 26 of 90 in groups A and B, respectively (P = 5), Pathologic review showed similarity betwen the two ‘groups, with the presence of cribriform, tufting, micro- Papillary, and flat microscopic subtypes (Figure 1, A through D) present in 7%, 55%, 50%, and 11%, respec- tively, in group A and in 6%, 60%, 55%, and 12%, respec- tively, in group B. The most common morphologic HPIN subtypes were the tufting and the micropapillary patterns. No statistically significant difference was observed be- tween group A (no subsequent carcinoma) and group B (developed carcinoma) in any of the clinical parameters examined including age, PSA level, DRE, rectal ultra- sound, or morphologic subtypes. Interphase FISH Findings ‘We had interpretable results in 44 cases (24 from group ‘A-and 20 from group B) for chromosomes 4, 7, 8, and 10. (ur findings indicated the presence of different chromo- somal anomalies in 5 of 20 (25%) cases of group B and in 2 of 24 (8.3%) cases of group A (P = 1). The results from the patient samples with chromosomal anomalies are shown in Table 2. All the chromosomal changes were in the form of a gain (Figures 2 and 3). No chromosomal losses were identified in any case. Gain of chromosome 8 ‘was seen in 5 cases, gain of chromosome 7 in 3 cases, and gain of chromosome 10 in 2 cases. No numeric chromo- somal changes were detected in chromosome 4, No nu- CChromesomal Changes to Predict Carcinoma—Al-Maghyabi etalFigure 1, Hematoxylin-osin sections ofthe ‘morphologic subtypes of high rae intra epithlal neoplasia PIPIND. A, Focus of i Sropapilan pater. 8 Focus of enbiorm pater. bcs of ting pate: D, os flat pate Al of these fot exh the (enerl features of HPIN including hyper hromsss, prominent nucleo, nuceat on Targement and srateation forgnal magn- ‘esti 400) Table 2. Summar ical and Interphase Fluorescence In Situ Hybridization Resulls for Chromosomes 4, 7, 8, and 10 on the Needle Core Prostatic Biopsies From Patients With Chromosomal Abnormalities* Patient Agey Chromosome? Chromosome 8 Chromosome 10 Chromosome 4 Follow-up Biopsy SA, ng/mL 1 8 N c G N PIN TO4 3 62 c N N N PIN, 73 7 6 G N N N cA a7 " 53 N G N N cA na 14 65 ¢ a N N a 18 16 52 N ci a N rN aa 41 55 N G N N a Ba * PSA indicates prostate-speific antigen: N, normaly G gain; PIN, prosate Invaepitheal neoplasia; and CA, carcinoma ‘meric chromosomal anomalies were noticed in the adja- cent hyperplastic or normal prostate glandular epithelium, Which was counted in 30 cases in the study group. We applied the same cytogenetic technique on 5-jzm paratfin- embedded sections from transurethral resection of pros- tte CTURY) specs som patents with nodular hype plasia (benign prostatic hyperplasia) using probe for chro- Trosome 8 and ound thal fore wee no Cheese mic ‘meric anomalies in any of the 14 specimens that were examined. We found that a total of 7 of 44 (16%) cases of HPIN showed chromosomal anomalies in at least 1 chro- mosome. At the time of the diagnosis, 3 of the 7 patients were below the age 60 years (all of them had a cancer in the follow-up biopsies); only 2 of those 7 patients had ab- normal ultrasound results, and none of them had abnor- mal DRE results. Six of 7 had PSA levels above 4 ng/ml. COMMENT ‘The rationale for this study was to determine whether ‘chromosomal aneusomies detected in HPIN foci could be predictive of subsequent carcinoma or could increase the Predictive vale of HDIN ‘In our study, we identified 123 patients with isolated HPIN as the primary diagnosis who had a subsequent follow-up biopsy. Prostate carcinoma was identified in 27% at the first follow-up, with a follow-up period ranging ‘Arch Pathol Lab Med-Vol 126, February 2002 between 2 and 36 months. That figure is consistent with the incidence described in the literature (27% to 100%). Davidson et al" found that the likelihood of finding cancer ‘was greater in patients with PIN undergoing more than 1 follow-up biopsy (44%) than in those with only 1 biopsy (62%). The carcinoma was identified in the same side as, the HPIN in 55% of the cases, in both sides in 33% of cases, and in the opposite side in 12%. Gleason scores, ‘were predominantly 6 and 7. The mean ages were 64 years and 62 years and the mean PSA values were 8 ng/mL and 7.8 ng/mL, respectively, for the group that did not devel- op prostate carcinoma (group A) and for the group that did (group B). No statistically significant difference was, identified between these 2 groups in any of the clinical parameters examined including age, PSA level, DRE, or rectal ultrasound, which is consistent with other authors? However, other investigators have indicated that age, PSA~ positivity, and DRE may be factors that increase the risk ‘of subsequent cancer}? suggesting that HPIN may not always be detectable by elevated serum PSA or by trans: rectal ultrasound or DRE. ‘The underlying mechanism of progression of PIN to prostate carcinoma is not clear. However, the progression ‘ight be caused by the onset of genomic instability, re- silting in a clone that has the ability to invade Cheo- ‘mosomal instability may be defined as an excess of chro- ‘Chromosomal Changes to Predict Carcinoma —Al-Maghrabi etal 167a focus of using dual centromere probes (red signal for ‘chromosome 7 and green signal for chromo some 6). Many cells arrow heads) show more "an 2 signals for both probes (origina mag ieation %600). > Figure 3, brs n (FISH) on a focus of citi Stromal cells, which have been used as intr al controls (original magnifiation > 600) mosome alterations occurring at each cell generation. Chromosomal instability is thought to arise as a result of aberrations in mitot of chromosome consti- tution, leading to excessive numeric chromosomal chang- es. Chromosomal instability is usually manifested in a form of numeric chromosomal changes of 1 or more chro- mosomes, In this study we found the presence of chro- mosomal aneusomies in 5 of 20 (25%) cases of men with HPIN who subsequently developed prostate carcinoma compared with 2 of 24 (8.3%) cases of the men with HPIN who did not develop prostate carcinoma (P = .1). This ference was not statistically significant; however, itis hight re that patients with HPIN with chromo- somal aneusomies have a higher chance of progression to invasive cancer. No numeric chromosomal anomalies have been noticed in the adjacent hyperplastic or normal pros- tate glandular epithelium present in the same biopsies. 168. Arch Pathol Lab Med—Vol 126, February ‘The overall chromosomal changes in HPIN were higher than those in the normal or hyperplastic epithelium, with statistically significant difference (P < 05) The advantage of interphase FISH when applied to his- tologic sections is that the tumor cells can be precisely evaluated, and normal and dysplastic foci can be evali- ated for chromosomal instability even in small biopsies. FISH has been used in PIN to determine chromosomal anomalies in a few studies." However, in those studies it was performed on prostatectomy specimens, which al- ready contain prostate carcinoma foci, Using centromere FISH probes, the most common numeric changes in PIN and prostate carcinoma incluce gain of chromosome 7, 8, and 10 and loss of chromosome ¥. The overall frequency of numeric chromosomal anomalies in PIN and in prostate carcinoma is remarkably similar, which is in keeping with the hypothesis that PIN is a precursor of carcinoma.» The Chromos. ies to Predict Carcinoma—Al-Maghrabi etal‘overall incidence of any numeric chromosomal anomaly ‘was 7 of 44 (16%) in our study. The presence of any anom- aly in PIN ranged in the literature between 12% and 62%. ‘Our relatively lower incidence of chromosomal changes in HPIN foci can be explained by the smaller foci of tissue in biopsies, which might not be representative of the whole HPIN foc. In our study all the chromosomal chang- es were in the form of a gain with no chromosomal losses identified in any case. Gain of chromosome 8 was the most common finding in our study, which is also consistent with the previous studies.""**' Gain of chromosome 7 has ‘been shown to be frequent in prostate carcinoma and is associated with higher tumor grade, advanced stage, and early patient death of prostate carcinoma®>*; however, this change could occur in an early stage of prostate car- cinoma tumorogenesis. Heterogeneity and multifocality are noteworthy features of prostate carcinoma and the associated prostatic tis- sues," which make it relatively difficult to study using traditional bulk mucleic acid extraction methods, It will be important for future studies addressing genetic differenc- es in lesions such as HPIN and microfoci of prostate car cinoma to use laser capture methodologies to provide a ‘more accurate molecular assessment of genetic change in prostate carcinoma and HPIN. ‘We thank Dr Waleed A. Milaat from the Department of Com- ‘munity Medicine and Primary Health Care, King Abdul-Azia University Hospital, Jeddah Saudi Arabia, We aso thank our lab ‘members: Khladoun Alromaih, Ben Behesht, Zong Zhang, Bisera Vukovic, Paula Marrano, Jana Karaskove, Paul Park, lena Kolom- ietz, and Jane Bayani for their contribution to achieve this work. Financial support for this project is funded by the US Army Med- ‘cal Research and Material Command Prostate Cancer Research ‘Program and the National Cancer Institute of Canada, with funds from the Canadian Cancer Society. References 1. van der Kas TH, Labi Te B. Post Iagpiial neoplasia nd ceviine manpulstion. Eur Ue. 1999815 908-510 2 Bostwick Dc. Progression of prosaic nae neoplasia ea in: ‘ative sdenorereoomas Eur Uo 96:30:19 a2 3. LengtuetCKineler KW Vogelstein 8. netic tables in ran cane cats. Nate. 1990386:43-609, “Chil OF Lengauo’ Yu etal. Mutations of mitotic checkpoint genes in human cance sex commer, Nature 1990392 200303, 5. Rabinovich Ps, Deaton 5, Breall a € al, ancoloncchomosoal Instability precedes dap and cance i lve cli, Canc Res 988 sastancdiss 6, ied, HosalmeyerHaddad K, Blegen H, Sctok E, Aus G. Genomic changes defining the genes, progrenion and maligany tena nso he ‘ran tumors» pheatypelgnanpe eorelstion Cenes Chromosomes Cance 1999125:195-208 7. But EC, Jes UA, Greaves M4 Birch IM, Boye JM, Vasey IM. Genomic sectors asoioted with ss of Reterozygoaty fr TP in cramer rome brabass. Br Caner 000 8seF72 {a Hawhins Nl Gorman ® Tomoson I, Bulpt Ward RE Colorectal car ‘nomas arkng in the hyperpaste polyposis syndrome proget reugh the hromosomal salty pata. Arn) Pol 2000187 38592, a, hackney Se, Sankey TV. Common pate gnei vlton in human sold timo. Gromeny 1997255127 10, Wiliams AC, Mile Ic, Cll T, Browne $1, Newbold RF, PaaskevaC The efecto ie 75 mata one mca Sly of human ¢elonc adenoma derived cal ine with rgenou wipe P39 sty be fore and ater DNA damage. Genes Chromosomes Cancer 097 2046-52, 11, Nat Aoosak Fanta Nata Kans Toa F Gnec tnd cigeeti aatins in rutitepcaenogenes oe soma) Castoey terol 260035111118 “ac 12. Taapore . Fulasewa K. p53 mutation and tate ily. Cancer inves. ooo ne de-t35 13, Kichhot , Rose Peteen BL, tl Comparative genomic bia on ieveals a recurent pater af hamosomalsbetons tn sevre span Arch Pathol Lab Med—Vol 126, February 2002 gem nat ce nang cee cha. emanate Taare Se lt a ern mahi cheer coined enmer ens ein Sen ncn el Soi TE REG Roget ete teste ta nace pt =o ae siete gen 9 tae 9. ta sae ers se se TP gat hee ovo pens Ree spe ee net are Seah pend in caraa TT i. eit 5 nde pet 2 el dono caplet 2 dat pee at sre: emer) SS ADeeH beg carpe aha wrestler eats tt te! Soe Seeericrpemaaar ans aa Pe ae cc ts SET et cnn 9, oe Sp inte tr eet parser Fee ren per Ur net pee Sr a gs tenes nc eas re fate apm toma dpe ei Sera ore hipaa Geet no coe ea eek netetae Bd er a oars tia sn cine attain Crate teres eras SEGUE Faraone ee at si 2 ste tf exp Top 0 tse ccpsigeTad sete Reser tetra tre ste aaa can ee ecg fi XFels s oa Gey gac cay Sieur cee WSR SMS te TAR er GS a en scone aes one ed A tae Gees Ee is TS ARES Set ection pa ner Fol St tein a ob Seabees Paes Qenrrnamrtes Seen ree Soe so gen ee es yee tewtwaen er sao SPA yet cr tr etn on let epic foe a SLL een ce co Sa Sgt nae 2 neo on are ieee cis Some ere crt Sy we ka ey sapere pau ern guia EAI one ai na pl gegen me ns cin Es EE oe ta Gn nip Sat Sern eomaee Teteg NEL oa coach pel el mc nk Seuss me ate penance eerie ce Spel each pron ya Wa Bisel ern 2 ff fl Oto ef uns woe aa aa leer te tara SS tae et aT (Chromosomal Changes to Predict Carcinoma—Al-Maghrabi etal 169