Journal of Neuro-Oncology 64: 271–278, 2003.© 2003 Kluwer Academic Publishers. Printed in the Netherlands.

Clinical Study

Loss of heterozygosity for loci on chromosome arms 1p and 10q inoligodendroglial tumors: relationship to outcome and chemosensitivity

Brian Thiessen1, John A. Maguire2, Kelly McNeil 3, David Huntsman3, Montgomery A. Martin4 andDoug Horsman3

1Department of Medical Oncology, British Columbia Cancer Agency, and Division of Neurology, Department ofMedicine; 2Division of Neuropathology, Department of Pathology, Vancouver General Hospital; 3Department ofPathology; 4Department of Diagnostic Imaging, British Columbia Cancer Agency, Vancouver, BC, Canada

Key words: chemotherapy, chromosome 1p, chromosome 10q, oligodendroglioma

Summary

Oligodendroglial tumors frequently have deletions of chromosomal loci on 1p and 19q. Loss of heterozygosity (LOH)of chromosome 10 may be a negative prognostic factor. We reviewed 23 patients with oligodendroglial tumors, toevaluate the frequency of 1p and 10q LOH and correlate with clinical outcome. Three loci (D1S402, D1S1172,MCT118) on 1p and 2 loci (D10S520 and D10S521) on 10q were analyzed for LOH using PCR techniques. Sixteenoligodendrogliomas (6 low grade and 10 anaplastic) and 7 oligoastrocytomas (1 low grade and 6 anaplastic) werestudied. Overall 14/22 (64%) showed 1p LOH and 7/23 (30%) showed 10q LOH. Of 7 patients with some responseto chemotherapy, all showed 1p LOH and none had 10q LOH. Of 5 patients with stable or progressive disease,1 had 1p LOH and 2 showed 10q LOH. The presence of 1p LOH was significantly associated with response tochemotherapy (p = 0.02). Median progression free survival (PFS) was 31 months for 1p intact patients and 118months for the 1p LOH group (p = 0.014). Median PFS for 10q LOH patients was 31 and 118 months for patientswith intact chromosome 10 (p = 0.016). 1p LOH is a predictor of response to chemotherapy and a good prognosticfactor. 10q LOH is less common in oligodendroglial tumors but predicts for worse outcome. Molecular genotypingof oligodendroglial tumors is recommended as part of the standard diagnostic workup.

Introduction

Oligodendroglial tumors have generated significantinterest amongst neuro-oncologists because of theirunique sensitivity to chemotherapy; especially thecombination of procarbazine, lomustine and vincristine(PCV). Still, not all patients with oligodendroglialtumors will respond to chemotherapy and even respon-sive tumors often recur, leading to the demise of thepatient. Predicting which tumors will be chemosensi-tive has clinical implications and may lead to earlierand aggressive treatment with chemotherapy.

Oligodendrogliomas frequently display chromoso-mal alterations including deletions or loss of heterozy-gosity (LOH) for chromosomal loci on 1p, 19q as wellas 9p [1–4]. Recently, Cairncross et al. [2], showedthat LOH for loci on 1p and 19q were predictors of

good response to chemotherapy in anaplastic oligo-dendrogliomas. Smith et al. [5], reviewed alterationsat 1p and 19q in a variety of glial tumors and foundthat the presence of deletions in both chromosomescorresponded with increased survival among oligoden-droglial but not astrocytic neoplasms. These studiesboth suggest that combined histologic and genetic sub-typing may help stratify and manage patients witholigodendroglial tumors.

Loss of heterozygosity on chromosome 10 hasbeen reported in over 70% of glioblastoma multi-forme tumors [6,7] but is less common in anaplasticoligodendrogliomas, occurring in 21–35% [2,8–10].Chromosome 10 deletions appear to be an indicator ofaggressive behavior in astrocytic tumors, and may alsobe prognostic for oligodendrogliomas. Tumor suppres-sor genes such as PTEN are located on the long arm of

272

chromosome 10 and their inactivation may be respon-sible for a more malignant phenotype. PTEN mutationshave been reported in 0–50% of anaplastic oligoden-drogliomas with 10q LOH [9–12], indicating that therole of PTEN is still uncertain and its importance, ascompared to astrocytic tumors, remains to be clarified.

The British Columbia Cancer Agency (BCCA)Neuro-Oncology Clinic in Vancouver is the regionalcenter for treatment and surveillance of patients withgliomas. This afforded access to a cohort of patientswith oligodendroglial tumors in which to evaluategenetic alterations, specifically 1p and 10q LOH. Thisstudy examines the frequency of these genetic alter-ations in this group of patients to determine the impacton survival and response to treatment. It representsa collaborative effort among clinicians, pathologists,molecular geneticists and radiologists in an attemptto bring genetic analysis of brain tumors out of theresearch laboratory and into clinical practice.

Methodology

Patient population

Patients with oligodendroglial tumors (pure oligo-dendrogliomas and mixed oligoastrocytomas) wereidentified prospectively as they presented for initialconsultation or routine follow-up at the BCCA Neuro-Oncology Clinic from 1998 to 1999. Examining a widevariety of oligodendroglial neoplasms was felt to moreaccurately represent the ‘real world’ situation facingneuro-oncologists and as such we included all patientswith significant oligodendroglial components to theirtumors as they presented to the clinic over a one yearspan. Charts were reviewed for clinical outcome dataincluding patient demographics, treatment, time to pro-gression and survival (PFS). Consent was obtained attime of patient identification for use of blood and tumortissue in this analysis.

Pathology

Brain tumor biopsies were reviewed by a neuro-pathologist (JM). Hematoxylin and eosin stainedslides of paraffin embedded tumor tissue were exam-ined for the presence of oligodendroglial and astro-cytic components. Tumors were subclassified intopure oligodendroglioma and mixed oligoastrocytoma.

Within the latter group, the tumors contained both neo-plastic oligodendroglial and astrocytic elements, andthe neoplastic astrocytic elements comprised at least25% of the mixed glioma. Tumors were also gradedas being either low grade or malignant, based uponthe presence or absence of elevated mitotic activity,significant nuclear pleomorphism, vascular neogene-sis and necrosis. In the identification of high gradetumors, great emphasis was placed upon the pres-ence or absence of vascular neogenesis [13]. Vascularneogenesis was identified in all tumors designatedas malignant, with the exception of two small tumorsamples which contained few blood vessels, but dis-played marked nuclear pleomorphism and significantlyelevated mitotic activity. Ancillary diagnostic tech-niques, including immunohistochemistry and electronmicroscopy, were selectively employed.

The most malignant area within each tumor wasidentified: the corresponding area from the block ofparaffin embedded tissue was sampled and submittedfor molecular studies at the BCCA molecular geneticslaboratory.

Molecular genetics

Constitutional DNA was isolated from peripheral bloodsamples by automated nucleic acid extraction (ABI341) using phenol/chloroform. DNA from matchedtumor samples was isolated by overnight proteinase Kdigestion following de-paraffinization. The followingpolymorphic loci were tested (Figure 1): three on chro-mosome 1p (D1S402, D1S1172 MCT118) and twoon chromosome 10q (D10S520 and D10S521). Poly-merase chain reaction (PCR) cycling was undertakenusing an MJ research instrument , annealing as per pre-vious published protocol [14]. PCR products were runon a polyacrylamide gel. A BRL 123bp ladder was runas a size marker. Following electrophoresis, gels werestained and photographed. Constitutional and tumorDNA were run side by side. LOH was determined whenthere was a distinct or complete loss of one allele intumor DNA in a heterozygous (informative) patient(Figure 2).

Radiologic assessment for response andprogression

Twelve patients had received chemotherapy aloneinitially following surgical diagnosis or at the timeof progression. MRI or CT scans were reviewed

273

Figure 1. Representations of chromosomes 1 and 10 indicatingthe approximate location of alleles used to identify LOH for 1pand 10q respectively.

Figure 2. Gel electrophoresis for six representative patients atthe D1S1172 locus on chromosome 1. Constitutional DNA (N)and tumor DNA (T) are represented side by side for each patient.Cases 20 and 26 show LOH on this allele for tumor DNAwhereas cases 23 and 25 show retention of heterozygous alleles(ROH). Case 22 demonstrates homozygosity (HOM) for bothalleles and case 21 shows a failure of PCR amplification with norecognizable products on electrophoresis (DNW).

in these patients. Images were digitized and tumorareas were measured using Scion Corporation’s NIHImage program. Measurements were determined by a

neuro-radiologist (MM) who was blinded to the clini-cal and molecular information regarding each patient.Best response to chemotherapy over at least twocycles was evaluated using Macdonald criteria [15].Macdonald criteria were also used to determine radio-graphic progression in all patients for determinationof PFS.

Statistical methods

The distribution of genetic alterations within prog-nostic factors was tested for using Fisher’s exact testand t-tests. The association of genetic alterations withresponse to chemotherapy was tested for using Fisher’sexact test. Kaplan–Meier curves were generated toexamine overall and PFS. The log rank test was used totest for association between PFS and prognostic factorsand genetic alterations.

Results

Demographics

Twenty-three patients (14 women and 9 men) wereincluded in this study (see Table 1). The mean age atsurgical diagnosis was 41.8 years (range 24–62 years).Mean post-surgical Karnofsky performance score(KPS) was 80 (range 60–100). Sixteen patients hadpure oligodendrogliomas (6 low grade and 10 anaplas-tic) and 7 were mixed oligoastrocytomas (1 low gradeand 6 anaplastic). Thirteen patients were included atinitial presentation of their glioma and 10 patientswere identified during long-term follow-up (range 1–15years).

DNA results (Tables 1 and 2)

Deletions of 1p were detected in 14 (64%) of 22patients (1 tumor failed PCR amplification) includ-ing 10/16 (62%) oligodendrogliomas and 4/6 (67%)mixed gliomas (1 indeterminate). Deletions of 10qwere detected in 7 (30%) of 23 patients, including6/16 (37%) oligodendrogliomas and 1/7 (14%) mixedgliomas.

In relation to histological tumor grade, 10q LOHwas identified in 2/7 (29%) low grade tumors and5/16 (31%) anaplastic tumors. 1p LOH was seenin 3/7 (43%) low grade oligodendrogliomas (LGO)

274

Table 1. Patient demographics and clinical data

Patient Age at Dx Pathology 1p status 10q status Best Progression Overall(years) chemotherapy free survival survival

response (months) (months)(12 evaluated)

1F 47 LGO LOH ROH PR 46+ 46+2M 24 LGO ROH ROH SD 42 48+3F 30 LGO ROH LOH SD 31 36+4F 62 LGO LOH ROH MR 84+ 84+5F 53 LGO ROH LOH 9 96M 30 LGO LOH ROH 118 156+7F 32 MO LOH ROH PR 48 81+8F 35 MO LOH ROH 49+ 49+9M 43 MO LOH ROH SD 124 18910M 50 MO ROH ROH 16 2511F 41 MO LOH LOH 52 76+12F 32 MO LOH ROH PR 51+ 51+13M 57 MO LOH LOH 4 614F 39 MO LOH ROH 28+ 28+15F 52 MO ROH LOH 19 2816F 43 MO ROH LOH PD 12 1717F 45 LOA ROH ROH 32+ 32+18F 31 MOA LOH ROH PR 160 199+19M 38 MOA LOH ROH PR 14 39+20M 44 MOA LOH ROH CR 131 192+21M 53 MOA DNW ROH PD 23 37+22F 39 MOA ROH LOH 44+ 44+23M 42 MOA LOH ROH 42+ 42+Abbreviations: LGO = low grade oligodendroglioma; MO = malignant oligodendroglioma; LOA = lowgrade oligoastrocytoma; MOA = malignant oligoastrocytoma; LOH = loss of heterozygosity; ROH =retention of heterozygosity; DNW = did not work, i.e. failed PCR amplification from tumor;+ = patient stillalive at last follow-up; CR = complete response (no remaining evidence of tumor); PR = partial response(≥50% reduction in tumor size); MR = minor response (25–50% reduction in tumor size); SD = stabledisease (<25% change in tumor size, either reduction or enlargement); PD = progressive disease (≥25%increase in tumor size).

Table 2. Distribution of prognostic factors among patients with and without LOH of 1p and 10q

Prognostic variable Total # ofpatients

Chromosome 1p p value∗ Chromosome 10q p value∗∗

LOH ROH LOH ROH

Mean age 23 41 42 0.7208 45 40 0.994Mean KPS 23 83 85 0.0955 77 86 0.4992

Tumor histology 0.6305 0.9533Pure oligodendroglioma 7 4 (29%) 2 (25%) 1 (14%) 6 (38%)Mixed oligoastrocytoma 16 10 (71%) 6 (75%) 6 (86%) 10 (62%)

Tumor grade 0.4164 0.7612Low grade 9 5 (36%) 4 (50%) 3 (43%) 6 (38%)Anaplastic 14 9 (64%) 4 (50%) 4 (57%) 10 (63%)

ROH = retention of heterozygosity; LOH = loss of heterozygosity; KPS = Karnofsky performance score; ∗p valuesrepresent significant differences in distribution of prognostic variables between patients with 1p LOH and those withintact 1p; ∗∗p values represent significant differences in distribution of prognostic variables between patients with 10qLOH and those with intact 10q.

275

but identified in 11/15 (73%) anaplastic tumors. Therelationship of chromosomal alterations to knownprognostic factors (age, Karnofsky performance sta-tus and tumor histology and grade) is presented inTable 2. No significant difference in prognostic fac-tors was identified between 1p LOH and 1p intactgroups. In fact, anaplastic tumors were over repre-sented in the 1p LOH group (p = 0.42). Similarly,no significant differences were seen in the prog-nostic factors between patients with 10q LOH andintact 10q.

Of the 10 patients identified during long-termfollow-up, 9 had determinate tests and 8 of theseshowed 1p LOH (89%). Among the newly diagnosedpatients, 6 of 13 showed 1p LOH (46%). The differ-ence in 1p LOH frequency was statistically significant(p = 0.0075) between these two groups.

Clinical outcomes

Clinical follow-up ranged from 28 to 199 months.Median overall survival has not yet been reached asonly 6 patients have died at the time of this report(although 1 case was lost to follow-up near death). Forthe entire sample, median PFS was 52 months. MedianPFS for 1p LOH patients was 118 months, and for intact1p median PFS was 31 months (p = 0.0145 log ranktest). Patients with 10q LOH had median PFS of 31months, and those with intact 10q showed a medianPFS of 118 months (p = 0.0159 log rank test). TheKaplan–Meier survival plots indicating these survivaldifferences are displayed in Figure 3. Neither tumorgrade (p = 0.272) nor tumor histology (p = 0.147)were found to significantly predict for survival in thispatient group.

We were able to evaluate the best radiographicresponse to chemotherapy (over at least 2 cycles) in12 patients. Four patients received PCV chemother-apy without irradiation at diagnosis. The remaining8 received a variety of agents at recurrence (PCV,lomustine, temozolomide, carboplatin, tamoxifen).The best radiographic response to chemotherapy isreported in Table 1. Seven patients showed somedegree of chemoresponsiveness (minor response, par-tial response or complete response), and 5 patientsremained stable or progressed on chemotherapy. All7 responders showed 1p LOH (p = 0.02), whereasonly one of the non-responders had 1p LOH. Twoof the non-responders had 10q LOH, whereas allchemosensitive tumors had an intact chromosome 10(p = 0.15). The molecular pattern of 1p LOH and

0 12 24 36 48 60 72 84 96 108 120 132 144 156

0

20

40

60

80

100

Months

Su

rviv

al (%

)

Progression-free survival by chromosome 10q status

ROH

LOH

0 12 24 36 48 60 72 84 96 108 120 132 144 156

0

20

40

60

80

100

Months

Su

rviv

al (%

)

Progression-free survival by chromosome 1p status

LOH

ROH

Figure 3. Kaplan–Meier survival plots representing PFS forthose with and without chromosome 1p deletions (top) andchromosome 10q deletions (bottom).

intact 10q was seen in all seven responders and in onlyone non-responder (p = 0.007)

Discussion

The accurate prediction of which patients with oligo-dendroglial tumors will experience long survival andrespond to treatment has significant therapeutic impli-cations. On its own, histological grading of oligoden-drogliomas has not been as reliable in determiningprognosis when compared to the grading of astrocy-tomas. This has stimulated investigators to exploremolecular markers as predictors of outcome for thesetumors. Oligodendrogliomas are known to have a num-ber of frequent chromosomal alterations, includingLOH of chromosomal loci on 1p, 9p, 10q, and 19q.Cairncross et al. [2] identified 1p and 19q LOH assignificant predictors of response to chemotherapy inanaplastic oligodendrogliomas. In that study, it wasapparent that 1p LOH was the most significant predic-tor and 19q data was less important. Further molecular

276

research by the same group [8] has further subclassifiedanaplastic oligodendrogliomas into four prognosticgroups. Group 1 (1p/19q LOH alone) was found to havethe best prognosis (median survival >123 months).Group 2 (1p LOH and other molecular changes) andGroup 3 (1p intact but TP53 mutation) had similarsurvival (71 months). The worst survival times wereseen in Group 4 with intact 1p and no TP53 mutation(16 months). Other researchers [5] have also identified1p/19q LOH as being important in predicting outcomein oligodendroglial but not astrocytic neoplasms. Addi-tionally, Bauman et al. [16] have shown that 1p LOHalso predicts for good response to radiotherapy.

Our results support these early reports. 1p LOH wasassociated with response to chemotherapy and bet-ter PFS in a wide range of oligodendroglial tumors.While our small numbers preclude extensive multivari-ate analysis, we did not see any evidence that histologyor grade influenced the 1p survival data. In our cohort,malignant tumors and mixed gliomas were as likely toshow 1p LOH as low grade and ‘pure’ oligodendroglialtumors. Despite this mixture of tumor types, we stillsaw a significantly better outcome for patients with 1pLOH tumors. Indeed, it may be that the separation ofpure and mixed oligodendroglial tumors on histologi-cal grounds is less clinically relevant than the identifi-cation of important molecular changes such as 1p LOH.

Mixed oligoastrocytomas have been poorly stud-ied in regards to molecular alterations. Smith et al.[5] included them in their review of all gliomas with1p/19q LOH. Their results indicated that these dele-tions were not predictive for prolonged survival inmixed gliomas. Sasaki et al. [17] showed that oligoden-drogliomas with ‘astrocytic’ features were less likelyto have 1p LOH (6 of 22 patients). Additionally, only3 of those 6 patients with ‘astrocytic’ tumors and 1pLOH responded to chemotherapy. Our results indicatehowever that the mixed glioma group can behave justas favorably as the oligodendroglioma group in thepresence of 1p LOH. We had 4 of 6 mixed tumorsshow 1p LOH and all 3 patients evaluable for responseto chemotherapy showed chemosensitivity. We pro-pose that these favorable mixed oligoastrocytomas mayhave a common progenitor to 1p LOH ‘pure’ oligo-dendrogliomas, but have differentiated histologicallyalong both astrocytic and oligodendroglial lines. Thisis similar to the view held by Kraus et al. [3]. Hope-fully, further molecular profiling of mixed oligoastro-cytomas can further identify patterns that both indicatea favorable prognosis and evaluate their oncogenicrelationship to the ‘pure’ oligodendroglioma.

Our survival data indicate that molecular changesin oligodendroglial neoplasms play an important rolein prognosis. Tumors with 1p LOH showed long PFSdespite this deletion being over-represented in theanaplastic tumors. In contrast, 10q LOH was asso-ciated with significantly shorter PFS despite beingequally represented among low grade and anaplastictumors. These outcome results were statistically sig-nificant whereas both tumor histology (mixed vs. pureoligodendroglial tumors) and tumor grade were notfound to significantly predict for survival in this smallgroup of patients. This again attends to the power ofgenetic testing in its ability to select tumors with betterprognosis.

The prevalence of 1p LOH in this broad group ofoligodendroglial tumors was 67%, identical to the per-centage reported by Cairncross et al. [2,8] among agroup restricted to anaplastic oligodendrogliomas. Ourpatient group represents a cross-sectional survey ofall those with oligodendroglial tumors presenting to aregional Neuro-Oncology clinic over a 1-year period. Itis noteworthy that 89% of patients on long-term follow-up were identified as having 1p LOH, compared to only50% of newly diagnosed patients. This higher preva-lence in long-term survivors is another indicator that1p LOH is a good prognostic factor.

10q LOH has been observed in oligodendrogliomaswith a lower frequency than in astrocytomas. Addi-tionally, 10q LOH in astrocytomas has been associatedwith advanced grade of malignancy, being more fre-quent in glioblastoma than in anaplastic astrocytoma[7]. This association is less certain in oligodendroglialtumors. The frequency of 10q deletions in our patientsample suggested no differences with respect to tumorgrade. However, the presence of 10q LOH was asso-ciated with a worse PFS. No chemoresponsive tumorhad a 10q deletion in our series. This contrasts with thework of Cairncross et al. [2], where over 60% of 10qLOH patients were chemoresponsive. The differencesbetween the two studies may reflect the small num-bers in each study and the fact that our study includedLGOs and mixed gliomas in addition to anaplasticoligodendrogliomas.

Our data indicate that the combination of 1p deletionand intact 10q is significantly associated with responseto chemotherapy. However, 10q status does not providesignificantly more information than 1p results alone.Chromosome 10 data may still be informative as aprognostic indicator in oligodendrogliomas. Our sur-vival results to date as well as those of others wouldindicate that oligodendrogliomas with 10q LOH have

277

worse outcomes [8,10]. Whether this effect of 10q LOHis related to deletions of the PTEN tumor suppressorgene is less certain. Sasaki et al. [9] found PTEN muta-tions in 7 of 14 anaplastic oligodendrogliomas with10q LOH whereas Jeuken et al. [12] found only onePTEN mutation in 7 anaplastic oligodendrogliomaswith 10q LOH and Hoang-Xuan et al. [10] found noPTEN mutations in 26 oligodendrogliomas. Clearly,oligodendroglial tumors remain heterogenous in thisregard and the relevance of PTEN mutation is unclear.

The questions of which tests to perform on thesepatients and which tumors to select remains open fordebate. There is no doubt from previous work and ourresults that 1p LOH is the most important predictor forsurvival and chemosensitivity in oligodendrogliomas,especially those which are anaplastic. Ino et al. [8] hasalso shown that oligodendrogliomas can be subclassi-fied further with 19q LOH and TP53 mutations. Ourresults indicate 10q LOH can also be prognostic inoligodendroglial tumours. Whether the augmentationof 1p LOH results by these other markers will prove rel-evant in the clinic will need larger numbers of patientsand longer analysis of survival and treatment outcomes.With the advent of DNA microarray analysis, we maybe able to classify these groups further based on a largebattery of altered molecular patterns. This may furtheraid clinicians in selecting favorable patients for moreaggressive chemotherapy protocols. For now we wouldrecommend that all oligodendroglial tumors be testedfor 1p LOH.

The increasing information suggesting that molec-ular markers can be used to predict biologic behaviorof oligodendroglial tumors will certainly propel thegenomic era in brain tumor research. It is apparent thatquestions still need to be answered regarding 1p LOH inoligodendrogliomas. A gene on the short arm of chro-mosome 1 has not yet been identified with a clear role ininfluencing sensitivity to alkylating agents. Potential 1pgenes such as CDKN2C, have been investigated but aredeleted in only a small number of oligodendrogliomaswith 1p LOH [18]. At this time we await the discoveryof other candidate genes on 1p which presumably willcome as our genomic technology advances.

In conclusion, our results indicate 1p LOH anal-ysis is perhaps the most clinically relevant tool inidentifying patients with oligodendroglial neoplasmswho have a greater response to chemotherapy anda better prognosis. In our experience, patients witholigodendroglial tumors showing 10q LOH rarely havea good response to chemotherapy and a poorer out-come. We continue to recommend genetic testing in

tumors of oligodendroglial origin as part of the routinepre-treatment investigations.

Acknowledgement

The authors wish to acknowledge the B.C. CancerAgency Surveillance and Outcomes Unit for assistancewith statistical analysis.

References

1. Bello MJ, Vaquero J, de Campos JM, Kusak ME, Sarasa JL,Saez-Castresona J, Pestafia A, Rey JA: Molecular analysisof chromosome 1 abnormalities in human gliomas revealsfrequent loss of 1p in oligodendroglial tumors. Int J Cancer57: 172–175, 1994

2. Cairncross JG, Ueki K, Zlatescu MC, Lisle DK,Finkelstein DM, Hammond RR, Silver JS, Stark PC,Macdonald DR, Ino Y, Ramsay DA, Louis DN: Specificgenetic predictors of chemotherapeutic response and sur-vival in patients with anaplastic oligodendrogliomas. J NatlCancer Inst 90: 1473–1479, 1998

3. Kraus JA, Koopman J, Kaskel P, Maintz D, Bradner S,Schramm J, Louis DN, Wiestler OD, von Diemling A:Shared allelic loss on chromosomes 1p and 19q suggesta common origin of oligodendroglioma and oligoastrocy-toma. J Neuropath Exp Neurol 54: 91–95, 1995

4. Kros JM, van Run PRWA, Alers JC, Beverloo HB,van den Bent MJ, Avezaat CJJ, van Decken H: Geneticaberrations in oligodendroglial tumours: an analysis usingcomparative genomic hybridization. J Pathol 188: 282–288,1999

5. Smith JS, Perry A, Borell TJ, Lee HK, O’Falloon J,Hosek SM, Kimmel D, Yates A, Burger PC,Scheithauer BW, Jenkins RB: Alterations of chromosomearms 1p and 19q as predictors of survival in oligoden-drogliomas, astrocytomas and mixed oligoastrocytomas.J Clin Oncol 18: 636–645, 2000

6. James CD, Carlbom E, Dumanski JP, Hansen M,Nordenskjold M, Collins VP, Cavanee WK: Clonal genomicalterations in glioma malignancy stages. Cancer Res 48:5546–5551, 1988

7. Fults D, Brockmeyer, D, Tullous MW, Pedone CA,Cawthon RM: p53 mutation and loss of heterozygosityon chromosome 17 and 10 during human astrocytomaprogression. Cancer Res 52: 674–679, 1992

8. Ino Y, Betensky RA, Zlatescu MC, Sasaki H,Macdonald DR, Stemmer-Rachamimov AO, Ramsay DA,Cairncross JG, Louis DN: Molecular subtypes of anaplasticoligodendrogliomas: implications for patient managementat diagnosis. Clin Cancer Res 7: 839–845, 2001

9. Sasaki H, Zlatescu ML, Betensky RA, Ino Y, Cairncross JG,Louis DN: PTEN is a target of chromosome 10q lossin anaplastic oligodendroglioma and PTEN alterations areassociated with poor prognosis. Am J Pathol 159: 359–367,2001

278

10. Hoang-Xuan K, He J, Huguet S, Mokhtari K, Marie Y,Kujas M, Leuraud P, Capelle L, Delattre JY, Poirier J,Broet P, Sanson M: Molecular heterogeneity of oligo-dendrogliomas suggests alternative pathways in tumorprogression. Neurology 57: 1278–1281, 2001

11. Bigner SH, Matthews MR, Rasheed BKA, Wiltshire RN,Friedman HS, Friedman AH, Stenzel TT, Dawes DM,McLendon RE, Bigner DD: Molecular genetic aspectsof oligodendrogliomas including analysis by comparativegenomic hybridization. Am J Pathol 155: 375–386, 1999

12. Jeuken JWM, Nelen MR, Vermeer H, van Staveren WCG,Kremer H, van Overbeeke JJ, Boerman RH: PTEN mutationanalysis in two genetic subtypes of high-grade oligoden-droglial tumors: PTEN is only occasionally mutated in oneof the two genetic subtypes. Cancer Genet Cytogenet 119:42–47, 2000

13. Daumas-Duport C, Varlet P, Tucker M-L, Beuvon F,Cervera P, Chodkiewicz J-P: Oligodendrogliomas. Part I:patterns of growth, histological diagnosis, clinical and imag-ing correlations: a study of 153 cases. J Neuro-Oncol 34:37–59, 1997

14. Mullis KB, Faloona FA: Specific synthesis of DNA invitro via a polymerase-catalyzed chain reaction. MethodsEnzymol 155: 335–350, 1987

15. Macdonald DR, Cascino TL, Schold SC, Cairncross JG:Response criteria for phase II studies of malignant glioma.J Clin Oncol 8: 1277–1280, 1990

16. Bauman GS, Ino Y, Ueki K, Zlatescu MC, Fisher BJ,Macdonald DR, Stitt L, Louis DN, Cairncross JG: Allelicloss of chromosome 1p and radiotherapy plus chemother-apy in patients with oligodendrogliomas. Int J Radiat OncolBiol Phys 48: 825–830, 2000

17. Sasaki H, Zlatescu MC, Betensky RA, Johnk LB,Cutone AN, Cairncross JG, Louis DN: Histopathological-molecular genetic correlations in referral pathologist-diagnosed low-grade ‘oligodendroglioma’. J Neuropath ExpNeurol 61: 58–63, 2002

18. Pohl U, Cairncross JG, Louis DN: Homozygous deletions ofthe CDKN2C/p18 gene on the short arm of chromosome 1in anaplastic oligodendrogliomas. Brain Pathol 9: 639–643,1999

Address for offprints: Dr. Brian Thiessen, Department of MedicalOncology, British Columbia Cancer Agency, 600 W 10th Avenue,Vancouver, BC, Canada V5Z 4E6; Tel.: (604)-877-6000; Fax:(604)-877-0585; E-mail: bthiessen@bccancer.bc.ca