SURVIVAL PREDICTION FOR BRAIN TUMOR PATIENTS USING GENE EXPRESSION DATA

Brain tumor is one of the most aggressive types of cancer in humans, with an estimated median survival time of 12 months and only 4% of the patients surviving more than 5 years after disease diagnosis. Until recently, brain tumor prognosis has been based only on clinical information such as tumor grade and patient age, but there are reports indicating that molecular profiling of gliomas can reveal subgroups of patients with distinct survival rates. We hypothesize that coupling molecular profiling of brain tumors with clinical information might improve predictions of patient survival time and, consequently, better guide future treatment decisions. In order to evaluate this hypothesis, the general goal of this research is to build models for survival prediction of glioma patients using DNA molecular profiles (U133 Affymetrix gene expression microarrays) along with clinical information. First, a predictive Random Forest model is built for binary outcomes (i.e. short vs. long-term survival) and a small subset of genes whose expression values can be used to predict survival time is selected. Following, a new statistical methodology is developed for predicting time-to-death outcomes using Bayesian ensemble trees. Due to a large heterogeneity observed within prognostic classes obtained by the Random Forest model, prediction can be improved by relating time-to-death with gene expression profile directly. We propose a Bayesian ensemble model for survival prediction which is appropriate for high-dimensional data such as gene expression data. Our approach is based on the ensemble "sum-of-trees" model which is flexible to incorporate additive and interaction effects between genes. We specify a fully Bayesian hierarchical approach and illustrate our methodology for the CPH, Weibull, and AFT survival models. We overcome the lack of conjugacy using a latent variable formulation to model the covariate effects which decreases computation time for model fitting. Also, our proposed models provides a model-free way to select important predictive prognostic markers based on controlling false discovery rates. We compare the performance of our methods with baseline reference survival methods and apply our methodology to an unpublished data set of brain tumor survival times and gene expression data, selecting genes potentially related to the development of the disease under study. A closing discussion compares results obtained by Random Forest and Bayesian ensemble methods under the biological/clinical perspectives and highlights the statistical advantages and disadvantages of the new methodology in the context of DNA microarray data analysis.

Computer simulation of glioma growth and morphology.

Despite major advances in the study of glioma, the quantitative links between intra-tumor molecular/cellular properties, clinically observable properties such as morphology, and critical tumor behaviors such as growth and invasiveness remain unclear, hampering more effective coupling of tumor physical characteristics with implications for prognosis and therapy. Although molecular biology, histopathology, and radiological imaging are employed in this endeavor, studies are severely challenged by the multitude of different physical scales involved in tumor growth, i.e., from molecular nanoscale to cell microscale and finally to tissue centimeter scale. Consequently, it is often difficult to determine the underlying dynamics across dimensions. New techniques are needed to tackle these issues. Here, we address this multi-scalar problem by employing a novel predictive three-dimensional mathematical and computational model based on first-principle equations (conservation laws of physics) that describe mathematically the diffusion of cell substrates and other processes determining tumor mass growth and invasion. The model uses conserved variables to represent known determinants of glioma behavior, e.g., cell density and oxygen concentration, as well as biological functional relationships and parameters linking phenomena at different scales whose specific forms and values are hypothesized and calculated based on in vitro and in vivo experiments and from histopathology of tissue specimens from human gliomas. This model enables correlation of glioma morphology to tumor growth by quantifying interdependence of tumor mass on the microenvironment (e.g., hypoxia, tissue disruption) and on the cellular phenotypes (e.g., mitosis and apoptosis rates, cell adhesion strength). Once functional relationships between variables and associated parameter values have been informed, e.g., from histopathology or intra-operative analysis, this model can be used for disease diagnosis/prognosis, hypothesis testing, and to guide surgery and therapy. In particular, this tool identifies and quantifies the effects of vascularization and other cell-scale glioma morphological characteristics as predictors of tumor-scale growth and invasion.

Atopic conditions and brain tumor risk in children and adolescents--an international case-control study (CEFALO)

BACKGROUND A number of epidemiological studies indicate an inverse association between atopy and brain tumors in adults, particularly gliomas. We investigated the association between atopic disorders and intracranial brain tumors in children and adolescents, using international collaborative CEFALO data. PATIENTS AND METHODS CEFALO is a population-based case-control study conducted in Denmark, Norway, Sweden, and Switzerland, including all children and adolescents in the age range 7-19 years diagnosed with a primary brain tumor between 2004 and 2008. Two controls per case were randomly selected from population registers matched on age, sex, and geographic region. Information about atopic conditions and potential confounders was collected through personal interviews. RESULTS In total, 352 cases (83%) and 646 controls (71%) participated in the study. For all brain tumors combined, there was no association between ever having had an atopic disorder and brain tumor risk [odds ratio 1.03; 95% confidence interval (CI) 0.70-1.34]. The OR was 0.76 (95% CI 0.53-1.11) for a current atopic condition (in the year before diagnosis) and 1.22 (95% CI 0.86-1.74) for an atopic condition in the past. Similar results were observed for glioma. CONCLUSIONS There was no association between atopic conditions and risk of all brain tumors combined or of glioma in particular. Stratification on current or past atopic conditions suggested the possibility of reverse causality, but may also the result of random variation because of small numbers in subgroups. In addition, an ongoing tumor treatment may affect the manifestation of atopic conditions, which could possibly affect recall when reporting about a history of atopic diseases. Only a few studies on atopic conditions and pediatric brain tumors are currently available, and the evidence is conflicting.

A Hybrid Model for Multimodal Brain Tumor Segmentation

We present a fully automatic segmentation method for multi-modal brain tumor segmentation. The proposed generative-discriminative hybrid model generates initial tissue probabilities, which are used subsequently for enhancing the classi�cation and spatial regularization. The model has been evaluated on the BRATS2013 training set, which includes multimodal MRI images from patients with high- and low-grade gliomas. Our method is capable of segmenting the image into healthy (GM, WM, CSF) and pathological tissue (necrotic, enhancing and non-enhancing tumor, edema). We achieved state-of-the-art performance (Dice mean values of 0.69 and 0.8 for tumor subcompartments and complete tumor respectively) within a reasonable timeframe (4 to 15 minutes).

Contribution of Ectodomain Mutations in Epidermal Growth Factor Receptor to Signaling in Glioblastoma Multiforme

CONTRIBUTION OF ECTODOMAIN MUTATIONS IN EPIDERMAL GROWTH FACTOR RECEPTOR TO SIGNALING IN GLIOBLASTOMA MULTIFORME Publication No._________ Marta Rojas, M.S. Supervisory Professor: Oliver Bögler, Ph.D. The Cancer Genome Atlas (TCGA) has conducted a comprehensive analysis of a large tumor cohort and has cataloged genetic alterations involving primary sequence variations and copy number aberrations of genes involved in key signaling pathways in glioblastoma (GBM). This dataset revealed missense ectodomain point mutations in epidermal growth factor receptor (EGFR), but the biological and clinical significance of these mutations is not well defined in the context of gliomas. In our study, we focused on understanding and defining the molecular mechanisms underlying the functions of EGFR ectodomain mutants. Using proteomic approaches to broadly analyze cell signaling, including antibody array and mass spectrometry-based methods, we found a differential spectrum of tyrosine phosphorylation across the EGFR ectodomain mutations that enabled us to stratify them into three main groups that correlate with either wild type EGFR (EGFR) or the long-studied mutant, EGFRvIII. Interestingly, one mutant shared characteristics of both groups suggesting a continuum of behaviors along which different mutants fall. Surprisingly, no substantial differences were seen in activation of classical downstream signaling pathways such as Akt and S6 pathways between these classes of mutants. Importantly, we demonstrated that ectodomain mutations lead to differential tumor growth capabilities in both in vitro (anchorage independent colony formation) and in vivo conditions (xenografts). Our data from the biological characterization allowed us to categorize the mutants into three main groups: the first group typified by EGFRvIII are mutations with a more aggressive phenotype including R108K and A289T; a second group characterized by a less aggressive phenotype exemplified by EGFR and the T263P mutation; and a third group which shared characteristics from both groups and is exemplified by the mutation A289D. In addition, we treated cells overexpressing the mutants with various agents employed in the clinic including temozolomide, cisplatin and tarceva. We found that cells overexpressing the mutants in general displayed resistance to the treatments. Our findings yield insights that help with the molecular characterization of these mutants. In addition, our results from the drug studies might be valuable in explaining differential responses to specific treatments in GBM patients.

Biologic and molecular analysis of tumor suppressor loci in human glioblastoma multiforme

Molecular and cytogenetic analyses of human glioblastomas have revealed frequent genetic alterations, including major deletions in chromosomes 9, 10, and 17, suggesting the presence of glioma-associated tumor suppressor genes on these chromosomes. To examine this hypothesis, copies of chromosomes 2, 4, and 10 derived from a human fibroblast cell line were independently introduced into a human glioma cell line, U251, by microcell-mediated chromosomal transfer. Successful transfer of chromosomes in each case was confirmed by resistance to the drug G418, indicating the presence of the neomycin-resistance gene previously integrated into each transferred chromosome. The presence of novel chromosomes and or chromosomal fragments was also demonstrated by molecular and karyotypic analyses. The hybrid clones containing either a novel chromosome 4 or chromosome 10 displayed suppression of the tumorigenic phenotype in vivo and suppression of the transformed phenotype in vitro, while cells containing a transferred chromosome 2 failed to alter their tumorigenic phenotype. The hybrid cells containing chromosome 4 or 10 exhibited a significant decrease in their saturation density, altered cellular morphology at high cell density, but only a slight decrease in their exponential growth rate. A dramatic decrease was observed in growth of cells with chromosome 4 or 10 in soft agarose, with the number and size of the colonies being greatly reduced, compared to the parental or chromosome 2 containing cells. The introduction of chromosome 4 or 10 also completely suppressed tumor formation in nude mice. These studies indicate that chromosome 10, as hypothesized, and chromosome 4, a novel finding for gliomas, harbor tumor suppressor loci that may be directly involved in the initiation or progression of normal glial precursors to human glioblastoma multiforme. ^

Analysis of differential gene expression in nontumorigenic glioblastoma multiforme

Loss of chromosome 10 represents the most common cytogenetic abnormality in high grade gliomas (glioblastoma multiforme). To identify genes involved in the malignant progression of human gliomas, a subtractive hybridization was performed between a tumorigenic glioblastoma cell line (LG11) and a nontumorgenic hybrid cell (LG11.3) containing an introduced chromosome 10. LG11 mRNA was subtracted from LG11.3 cDNA to produce cDNA probes enriched for sequences whose expression differs quantitatively from the parental tumorigenic cells. Both known and novel sequences were identified as a result of the subtraction. Northern blot analysis was then used to confirm differential expression of several subtracted clones. One novel clone, clone 17, identified a 2.6 kb message that showed a consistent two to four fold increase in expression in the LG11.3 nontumorigenic cells. Clone 17 (340 bp) was used successfully to screen for a near full-length version, RIG (regulated in glioma), which was 2,569 bp in size. The RIG cDNA sequence showed homology to clone 17 and to an anonymous EST (IB666), but to no previously identified genes. This screening effort also identified several independent clones representing novel sequences, most of which failed to show increased expression in the nontumorigenic GBM cells. Tissue distribution studies of RIG indicated highest levels of expression in human brain with appreciably lower levels in heart and lung. In vitro transcription and translation experiments demonstrated the ability of RIG to direct the synthesis of a 13 kD protein product. However, open reading frame analysis revealed no identify with previously described motifs or any known proteins. Using a combination of somatic cell hybrid panels and in situ hybridization, the RIG gene was mapped to chromosome 11p14-11p15. Further study of RIG and related gene products may provide insight into the negative regulation of glial oncogenesis. ^

Long-term outcome of patients with spinal myxopapillary ependymoma: treatment results from the MD Anderson Cancer Center and institutions from the Rare Cancer Network.

BACKGROUND Spinal myxopapillary ependymomas (MPEs) are slowly growing ependymal gliomas with preferential manifestation in young adults. The aim of this study was to assess the outcome of patients with MPE treated with surgery, radiotherapy (RT), and/or chemotherapy. METHODS The medical records of 183 MPE patients (male: 59%) treated at the MD Anderson Cancer Center and 11 institutions from the Rare Cancer Network were retrospectively reviewed. Mean patient' age at diagnosis was 35.5 ± 15.8 years. Ninety-seven (53.0%) patients underwent surgery without RT, and 86 (47.0%) were treated with surgery and/or RT. Median RT dose was 50.4 Gy. Median follow-up was 83.9 months. RESULTS Fifteen (8.2%) patients died, 7 of unrelated cause. The estimated 10-year overall survival was 92.4% (95% CI: 87.7-97.1). Treatment failure was observed in 58 (31.7%) patients. Local failure, distant spinal relapse, and brain failure were observed in 49 (26.8%), 17 (9.3%), and 11 (6.0%) patients, respectively. The estimated 10-year progression-free survival was 61.2% (95% CI: 52.8-69.6). Age (<36 vs ≥36 y), treatment modality (surgery alone vs surgery and RT), and extent of surgery were prognostic factors for local control and progression-free survival on univariate and multivariate analysis. CONCLUSIONS In this series, treatment failure of MPE occurred in approximately one third of patients. The observed recurrence pattern of primary spinal MPE was mainly local, but a substantial number of patients failed nonlocally. Younger patients and those not treated initially with adjuvant RT or not undergoing gross total resection were significantly more likely to present with tumor recurrence/progression.

miR-125b controls apoptosis and temozolomide resistance by targeting TNFAIP3 and NKIRAS2 in glioblastomas.

Diffusely infiltrating gliomas are among the most prognostically discouraging neoplasia in human. Temozolomide (TMZ) in combination with radiotherapy is currently used for the treatment of glioblastoma (GBM) patients, but less than half of the patients respond to therapy and chemoresistance develops rapidly. Epigenetic silencing of the O(6)-methylguanine-DNA methyltransferase (MGMT) has been associated with longer survival in GBM patients treated with TMZ, but nuclear factor κB (NF-κB)-mediated survival signaling and TP53 mutations contribute significantly to TMZ resistance. Enhanced NF-κB is in part owing to downregulation of negative regulators of NF-κB activity, including Tumor necrosis factor alpha-induced protein 3 (TNFAIP3) and NF-κB inhibitor interacting RAS-like 2 (NKIRAS2). Here we provide a novel mechanism independent of TP53 and MGMT by which oncogenic miR-125b confers TMZ resistance by targeting TNFAIP3 and NKIRAS2. GBM cells overexpressing miR-125b showed increased NF-κB activity and upregulation of anti-apoptotic and cell cycle genes. This was significantly associated with resistance of GBM cells to TNFα- and TNF-related inducing ligand-induced apoptosis as well as resistance to TMZ. Conversely, overexpression of anti-miR-125b resulted in cell cycle arrest, increased apoptosis and increased sensitivity to TMZ, indicating that endogenous miR-125b is sufficient to control these processes. GBM cells overexpressing TNFAIP3 and NKIRAS2 were refractory to miR-125b-induced apoptosis resistance as well as TMZ resistance, indicating that both genes are relevant targets of miR-125b. In GBM tissues, high miR-125b expression was significantly correlated with nuclear NF-κB confirming that miR-125b is implicated in NF-κB signaling. Most remarkably, miR-125b overexpression was clearly associated with shorter overall survival of patients treated with TMZ, suggesting that this microRNA is an important predictor of response to therapy.

Extending resection and preserving function: modern concepts of glioma surgery

Recent studies have demonstrated that the improved prognosis derived from resection of gliomas largely depends on the extent and quality of the resection, making maximum but safe resection the ultimate goal. Simultaneously, technical innovations and refined neurosurgical methods have rapidly improved efficacy and safety. Because gliomas derive from intrinsic brain cells, they often cannot be visually distinguished from the surrounding brain tissue during surgery. In order to appreciate the full extent of their solid compartment, various technologies have recently been introduced. However, radical resection of infiltrative glioma puts neurological function at risk, with potential detrimental consequences for patients' survival and quality of life. The allocation of various neurological functions within the brain varies in each patient and may undergo additional changes in the presence of a tumour (brain plasticity), making intra-operative localisation of eloquent areas mandatory for preservation of essential brain functions. Combining methods that visually distinguish tumour tissue and detect tissues responsible for critical functions now enables resection of tumours in brain regions that were previously considered off-limits, and benefits patients by enabling a more radical resection, while simultaneously lowering the risk of neurological deficits. Here we review recent and expected developments in microsurgery for glioma and their respective benefits.

Predicting the "usefulness" of 5-ALA-derived tumor fluorescence for fluorescence-guided resections in pediatric brain tumors: a European survey.

BACKGROUND Five-aminolevulinic acid (Gliolan, medac, Wedel, Germany, 5-ALA) is approved for fluorescence-guided resections of adult malignant gliomas. Case reports indicate that 5-ALA can be used for children, yet no prospective study has been conducted as of yet. As a basis for a study, we conducted a survey among certified European Gliolan users to collect data on their experiences with children. METHODS Information on patient characteristics, MRI characteristics of tumors, histology, fluorescence qualities, and outcomes were requested. Surgeons were further asked to indicate whether fluorescence was "useful", i.e., leading to changes in surgical strategy or identification of residual tumor. Recursive partitioning analysis (RPA) was used for defining cohorts with high or low likelihoods for useful fluorescence. RESULTS Data on 78 patients <18 years of age were submitted by 20 centers. Fluorescence was found useful in 12 of 14 glioblastomas (85 %), four of five anaplastic astrocytomas (60 %), and eight of ten ependymomas grades II and III (80 %). Fluorescence was found inconsistently useful in PNETs (three of seven; 43 %), gangliogliomas (two of five; 40 %), medulloblastomas (two of eight, 25 %) and pilocytic astrocytomas (two of 13; 15 %). RPA of pre-operative factors showed tumors with supratentorial location, strong contrast enhancement and first operation to have a likelihood of useful fluorescence of 64.3 %, as opposed to infratentorial tumors with first surgery (23.1 %). CONCLUSIONS Our survey demonstrates 5-ALA as being used in pediatric brain tumors. 5-ALA may be especially useful for contrast-enhancing supratentorial tumors. These data indicate controlled studies to be necessary and also provide a basis for planning such a study.

Combination of MRI and dynamic FET PET for initial glioma grading

AIM MRI and PET with 18F-fluoro-ethyl-tyrosine (FET) have been increasingly used to evaluate patients with gliomas. Our purpose was to assess the additive value of MR spectroscopy (MRS), diffusion imaging and dynamic FET-PET for glioma grading. PATIENTS, METHODS 38 patients (42 ± 15 aged, F/M: 0.46) with untreated histologically proven brain gliomas were included. All underwent conventional MRI, MRS, diffusion sequences, and FET-PET within 3±4 weeks. Performances of tumour FET time-activity-curve, early-to-middle SUVmax ratio, choline / creatine ratio and ADC histogram distribution pattern for gliomas grading were assessed, as compared to histology. Combination of these parameters and respective odds were also evaluated. RESULTS Tumour time-activity-curve reached the best accuracy (67%) when taken alone to distinguish between low and high-grade gliomas, followed by ADC histogram analysis (65%). Combination of time-activity-curve and ADC histogram analysis improved the sensitivity from 67% to 86% and the specificity from 63-67% to 100% (p < 0.008). On multivariate logistic regression analysis, negative slope of the tumour FET time-activity-curve however remains the best predictor of high-grade glioma (odds 7.6, SE 6.8, p = 0.022). CONCLUSION Combination of dynamic FET-PET and diffusion MRI reached good performance for gliomas grading. The use of FET-PET/MR may be highly relevant in the initial assessment of primary brain tumours.

Combined ATRX/IDH1 immunohistochemistry predicts genotype of oligoastrocytomas

AIMS To assess whether in oligoastrocytomas ATRX deficiency - as a surrogate of the alternative lengthening of telomeres (ALT) pathway - has a role in predicting the presence or absence of loss of heterozygosity of 1p and 19q (LOH), the genetic signature of oligodendroglial differentiation and a favourable prognostic marker. METHODS AND RESULTS A series of 54 oligoastrocytomas were investigated by immunohistochemistry as well as microsatellite analysis for LOH 1p19q. Genetic findings were correlated with morphological assessment. CONCLUSIONS ATRX deficiency was mutually exclusive with LOH. Conversely, ATRX-proficient tumours immunoreactive for R132H-mutant isocitrate dehydrogenase 1 (IDH1) showed a high rate (85%) of LOH. A more oligodendroglioma-like morphology was associated with a higher rate of LOH even in the morphologically ambiguous group of oligoastrocytomas. Our findings support the concept that oligoastrocytomas represent a morphological grey zone rather than a group of truly "mixed" or "intermediate" tumours. More precise classification of diffuse gliomas may also improve grading of borderline cases. We propose an immunohistochemical algorithm for classification of morphologically ambiguous diffuse gliomas. This article is protected by copyright. All rights reserved.

Comparison of Routine Brain Imaging at 3 T and 7 T.

OBJECTIVE The aim of this study was to compare quantitative and semiquantitative parameters (signal-to-noise ratio [SNR], contrast-to-noise ratio [CNR], image quality, diagnostic confidence) from a standard brain magnetic resonance imaging examination encompassing common neurological disorders such as demyelinating disease, gliomas, cerebrovascular disease, and epilepsy, with comparable sequence protocols and acquisition times at 3 T and at 7 T. MATERIALS AND METHODS Ten healthy volunteers and 4 subgroups of 40 patients in total underwent comparable magnetic resonance protocols with standard diffusion-weighted imaging, 2D and 3D turbo spin echo, 2D and 3D gradient echo and susceptibility-weighted imaging of the brain (10 sequences) at 3 T and 7 T. The subgroups comprised patients with either lesional (n = 5) or nonlesional (n = 4) epilepsy, intracerebral tumors (n = 11), demyelinating disease (n = 11) (relapsing-remitting multiple sclerosis [MS, n = 9], secondary progressive MS [n = 1], demyelinating disease not further specified [n = 1]), or chronic cerebrovascular disorders [n = 9]). For quantitative analysis, SNR and CNR were determined. For a semiquantitative assessment of the diagnostic confidence, a 10-point scale diagnostic confidence score (DCS) was applied. Two experienced radiologists with additional qualification in neuroradiology independently assessed, blinded to the field strength, 3 pathology-specific imaging criteria in each of the 4 disease groups and rated their diagnostic confidence. The overall image quality was semiquantitatively assessed using a 4-point scale taking into account whether diagnostic decision making was hampered by artifacts or not. RESULTS Without correction for spatial resolution, SNR was higher at 3 T except in the T2 SPACE 3D, DWI single shot, and DIR SPACE 3D sequences. The SNR corrected by the ratio of 3 T/7 T voxel sizes was higher at 7 T than at 3 T in 10 of 11 sequences (all except for T1 MP2RAGE 3D).In CNR, there was a wide variation between sequences and patient cohorts, but average CNR values were broadly similar at 3 T and 7 T.DCS values for all 4 pathologic entities were higher at 7 T than at 3 T. The DCS was significantly higher at 7 T for diagnosis and exclusion of cortical lesions in vascular disease. A tendency to higher DCS at 7 T for cortical lesions in MS was observed, and for the depiction of a central vein and iron deposits within MS lesions. Despite motion artifacts, DCS values were higher at 7 T for the diagnosis and exclusion of hippocampal sclerosis in mesial temporal lobe epilepsy (improved detection of the hippocampal subunits). Interrater agreement was 69.7% at 3 T and 93.3% at 7 T. There was no significant difference in the overall image quality score between 3 T and 7 T taking into account whether diagnostic decision making was hampered by artifacts or not. CONCLUSIONS Ultra-high-field magnetic resonance imaging at 7 T compared with 3 T yielded an improved diagnostic confidence in the most frequently encountered neurologic disorders. Higher spatial resolution and contrast were identified as the main contributory factors.

Physical and functional mapping of a novel tumor suppressor locus for prostate cancer within chromosome 10p12.31-q11

Prostate cancer remains the second leading cause of male cancer deaths in the United States, yet the molecular mechanisms underlying this disease remain largely unknown. Cytogenetic and molecular analyses of prostate tumors suggest a consistent association with the loss of chromosome 10. Previously, we have defined a novel tumor suppressor locus PAC-1 within chromosome 10pter-q11. Introduction of the short arm of chromosome 10 into a prostatic adenocarcinoma cell line PC-3H resulted in dramatic tumor suppression and restoration of a programmed cell death pathway. Using a combined approach of comparative genomic hybridization and microsatellite analysis of PC-3H, I have identified a region of hemizygosity within 10p12-p15. This region has been shown to be involved in frequent loss of heterozygosity in gliomas and melanoma. To functionally dissect the region within chromosome 10p containing PAC-1, we developed a strategy of serial microcell fusion, a technique that allows the transfer of defined fragments of chromosome 10p into PC-3H. Serial microcell fusion was used to transfer defined 10p fragments into a mouse A9 fibrosarcoma cell line. Once characterized by FISH and microsatellite analyses, the 10p fragments were subsequently transferred into PC-3H to generate a panel of microcell hybrid clones containing overlapping deletions of chromosome 10p. In vivo and microsatellite analyses of these PC hybrids identified a small chromosome 10p fragment (an estimated 31 Mb in size inclusive of the centromere) that when transferred into the PC-3H background, resulted in significant tumor suppression and limited a region of functional tumor suppressor activity to chromosome 10p12.31-q11. This region coincides with a region of LOH demonstrated in prostate cancer. These studies demonstrate the utility of this approach as a powerful tool to limit regions of functional tumor suppressor activity. Furthermore, these data used in conjunction with data generated by the Human Genome Project lent a focused approach to identify candidate tumor suppressor genes involved in prostate cancer. ^