Miscellaneous Cytogenetic Abnormalities Research Articles

Omacetaxine Mepesuccinate for Patients with Higher-Risk MDS and CMML after Failure of Hypomethylating Agents: A Phase II Trial

Background: The outcome of patients (pts) with myelodysplastic syndromes (MDS) after hypomethylating agent (HMA) failure is poor with a median survival of 4-6 months. Omacetaxine mepesuccinate (OM) has been shown to be safe and effective in acute myeloid leukemia, including in pts who progressed from MDS. We designed a phase II trial to evaluate the safety and efficacy of OM in pts with MDS and chronic myelomonocytic leukemia (CMML) after HMA failure.Methods: Adult pts with MDS or CMML who had not responded to, progressed on, or relapsed after HMA therapy were eligible for this study. Pts were required to be intermediate-2- or high-risk by the International Prognostic Scoring System (IPSS) or be classified as RAEB-1 (5-9% BM blasts), RAEB-2 (10-19% BM blasts) or CMML (5-19% BM blasts). Pts received OM 1.25 mg/m2 SQ every 12 hours x 3 days on a 28-day cycle. The primary efficacy outcomes were overall survival (OS) and the overall response rate (ORR). Secondary outcomes included the remission duration, relapse-free survival (RFS), and the safety of the regimen.Results: Between 6/2015 and 6/2017, 35 pts have been treated. Baseline characteristics are shown in Table 1. The median age of the cohort was 73 years (range, 61-86 years). The median BM blasts was 10% (range, 2-19%) and 15 pts (43%) had poor-risk karyotype. IPSS risk group was intermediate-1 in 8 pts (23%), intermediate-2 in 13 (57%), and high in 7 (20%). 27 pts (77%) were either high risk (n=10) or very high risk (n=17) by IPSS-R. Among 33 pts who underwent comprehensive mutation profiling, 29 (88%) had at least one mutation identified. The most frequently identified mutations were RUNX1 in 10 pts (30%), TP53 in 10 (30%), ASXL1 in 6 (18%) and TET2 in 4 (12%).The median duration of follow-up was 17 months (range, 2-24 months), and the median number of cycles of OM received was 2 (range, 1-18). The ORR was 34%; 3 pts achieved CRp and 9 pts achieved CRi as best response. The median number of cycles to best response was 2 (range, 1-4 cycles). The response rates for pts with MDS and CMML were 41% and 0%, respectively (P=0.02). Responses were seen across cytogenetic risk groups: 5 responders were diploid, 4 had either complex cytogenetics or monosomy 7, and 3 had miscellaneous cytogenetic abnormalities. The response rate for pts with diploid and adverse-risk cytogenetics were 56% (5/9) and 27% (4/15), respectively (P=0.16).Of the 12 responders, 4 remain on study, 5 relapsed, 2 died while in response, and 1 was lost to follow-up. To date, 24 pts (69%) have died. The median OS was 7.6 months, and the 1-year OS rate was 22% (Figure 1). The median RFS was 2.9 months, and the 1-year RFS rate was 28%. Two pts have had ongoing response ≥12 months (15.6 and 19.8 months), both of whom had MDS with diploid cytogenetics.Treatment was overall well-tolerated. Among 26 pts who have received ≥2 cycles of OM, 2 (8%) required dose reduction during subsequent cycles. Most adverse events were grade 1-2; grade ≥3 adverse events occurred in 8 pts (23%). The most frequent non-hematologic adverse events were fatigue in 17 pts (49%), nausea in 12 (34%), mouth sores or mucositis in 6 (17%), infection in 5 (14%) and edema in 5 (14%). The 30-day and 60-day mortality rates were 11% and 14%, respectively.Conclusions: In pts with MDS and CMML who previously failed HMA therapy, OM was safe and resulted in an ORR of 34%. Responses were seen across cytogenetic risk groups but were restricted to patients with MDS. [Display omitted] DisclosuresTakahashi:Symbio Pharmaceuticals: Consultancy. DiNardo:Celgene: Honoraria, Research Funding; Novartis: Honoraria, Research Funding; Daiichi-Sankyo: Honoraria, Research Funding; Agios: Honoraria, Research Funding; AbbVie: Honoraria, Research Funding. Bose:Incyte Corporation: Honoraria. Cortes:BMS: Consultancy, Research Funding; Teva: Research Funding; Sun Pharma: Research Funding; ImmunoGen: Consultancy, Research Funding; Novartis Pharmaceuticals Corporation: Consultancy, Research Funding; ARIAD: Consultancy, Research Funding; Pfizer: Consultancy, Research Funding. Kantarjian:Amgen: Research Funding; Delta-Fly Pharma: Research Funding; ARIAD: Research Funding; Bristol-Meyers Squibb: Research Funding; Novartis: Research Funding; Pfizer: Research Funding. Jabbour:Bristol-Myers Squibb: Consultancy.

Prospective Karyotype Analysis in Adult Acute Lymphoblastic Leukemia: The Cancer and Leukemia Group B Experience

AbstractThe Cancer and Leukemia Group B (CALGB) has been conducting a prospective cytogenetic companion study (CALGB 8461) to all CALGB treatment protocols for newly diagnosed adults with acute lymphoblastic leukemia (ALL). These protocols underwent a significant change in 1988 when a new intensive chemotherapy program was introduced (CALGB 8811). We asked whether karyotype continued to represent a significant prognostic factor in adult ALL patients after the change. A total of 256 patients had adequate pretreatment cytogenetic analyses: 67 before 1988 and 189 subsequently. The complete remission (CR) rate for the whole group was 80%. Patients with t(9;22), t(4;11), −7, or +8 had significantly lower probabilities of continuous CR and survival at 5 years (.11 and .12) than patients with a normal karyotype (.38 and .37) and patients with miscellaneous cytogenetic abnormalities (.52 and .49;P < .001 for each comparison). When analyzed by treatment period, the CR rate before CALGB 8811 was 63%; subsequently, it was 86% (P < .001). Patients with cytogenetic abnormalities other than t(9;22), t(4;11), −7, or +8 had better CR rates, disease-free survival (DFS), and survivals (P = .001,P = .04, and P = .004, respectively) after the change to the more intensive chemotherapy regimens. Patients with normal cytogenetics had improved CR rate but no improved DFS or survival, whereas no significant benefit for patients with t(9;22), t(4;11), −7, or +8 was seen. In a multivariate analysis, karyotype retained its prognostic significance for DFS but not for survival; it remained the most important factor for DFS. We conclude that cytogenetic analysis at diagnosis should be used to guide treatment decisions in adults with ALL.

Abstract 1766: Distinct pattern of alterations in tp53 mutated and wild type acute myeloid leukemia (AML) patients

Abstract Background: Mutations in TP53 gene predict a poor prognosis in patients with AML. The reported TP53 mutation rate in AML is low (2.1%) by contrast AML with a complex karyotype (CK) is higher (69-78%) and have a poor outcome. Quite common is to found paired TP53 mutation together and a segmental 17p deletion. Aims: To investigate the frequency, the types of mutations, the associated cytogenetic, the correlation with known molecular alterations and the prognostic role TP53 mutations in adult AML pts. Moreover we would identify genes/pathways that are mainly affected in the mutated TP53 group compared to the wt one. Patients and Methods: 258 adult AML pts with miscellaneous cytogenetic abnormalities and normal karyotype were examined in our Institution for TP53 mutations using several methods, including Sanger sequencing, NGS and HiSeq2000 platform and were correlated with cytogenetic analysis. 124/258 samples were genotyped with SNP arrays (Affymetrix, 3 250K, 43 SNP 6.0, 78 CytoScan HD). Copy Number Alterations (CNAs) analyses were performed using Chromosome Analysis Suite (Affymetrix) and Nexus Copy Number (BioDiscovery) software. Results: Mutation analysis detected TP53 mutations on 39 patients with 48 different types of mutations (32 deleterious point mutations; 4 deletions); nine pts have 2 mutations. We found 34/48 (70%) missense mutations, 5 mutations in the splice sites, 4 deletions ,2 intronic and and 3 others mutations. The mutation rate is of 15.1%. Mostly (28/39) of the TP53 mutated pts (71.8%) had CK while only 11/39 (28.2%) mutated pts have “no CK” (P&amp;gt;0.0001). Alterations of TP53 were significantly associated with poor outcome (OS and EFS p&amp;lt;0.0001). To take advantage of different methodological characteristics, on TP53 locus, we matched two software and cytogenetic analysis results. We identify 16 mutated pts that were also deleted and one pt that presented only a deletion. Therefore 50% of mutated pts present a concomitant deletion. OS of TP53 mutated and deleted pts is statistically inferior respect to muted pts (p&amp;lt;0.0061). Comparing 32 TP53 mutated and 92 TP53 wt pts CNAs results that: a) chromosomes significantly alterated are 5q, 17p, 12p, 16q, 22q13.33 and 7q; b) over 900 genes are differentially involved (all in Loss); c) and that pathway categories mainly enriched are Signal Transduction, Metabolism, Immune System, Transmembrane transport of small molecules, Gene expression, Cell Cycle. Conclusions: Our data demonstrated that TP53 mutations occur in 15.1% of AML with a higher frequency in the subgroup of CK-AML (p&amp;lt;0.0001). They predicted to be deleterious and significantly correlated with worse prognosis especially if TP53 is both mutated and also deleted. For these reasons, TP53 mutation/deletion screening should be recommended. Different pattern of alterations in mutated and wt groups have to be deeper investigated. ELN, AIL, AIRC, progetto Regione-Università 2010-12 (L. Bolondi), FP7 NGS-PTL project. Citation Format: Anna Ferrari, Eugenio Fonzi, Maria Chiara Fontana, Andrea Ghelli Luserna Di Rorà, Marco Manfrini, Antonella Padella, Carmen Baldazzi, Cristina Papayannidis, Giovanni Marconi, Stefania Paolini, Viviana Guadagnuolo, Margherita Perricone, Valentina Robustelli, Enrica Imbrogno, Eugenia Franchini, Claudia Venturi, Elisa Zuffa, Maria Chiara Abbenante, Giorgia Simonetti, Nicoletta Testoni, Emanuela Ottaviani, Martinelli Giovanni. Distinct pattern of alterations in tp53 mutated and wild type acute myeloid leukemia (AML) patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1766. doi:10.1158/1538-7445.AM2017-1766

Abstract 4507: New JAK2 heterozygous loss: A role in overall survival in acute myeloid leukemia patients

Abstract Acute Myeloid Leukemia (AML) is a clonal hematopoietic disorder characterized by an abnormal proliferation and differentiation of immature blast cells in the bone marrow. SNP microarray approach has resulted in genome-wide screening for genomic alterations with information not previously achievable and also it allows to map all the genes involved in these alterations which may plays a role in oncogenesis. Our objective is to evaluate the prognostic impact of these genetic alterations on clinical outcome. We analyzed 285 AML samples at diagnosis of which 221 by SNP Array 6.0 (Affymetrix) (Affymetrix), while 64 by by CytoScan HD Array (Affymetrix). The results obtained were analyzed by Chromosome Analysis Suite (ChAS) v1.2 (Affymetrix Inc.) and Nexus Copy Number™ v7.5 softwares (BioDiscovery). The 285 AML patients (pts) analyzed are equally distributed between both sexes, they have a median age of 60 years, include miscellaneous cytogenetic abnormalities and normal karyotype. The 44% of pts are treated with chemotherapy (CHT), the 16% of pts with low-dose cytosine arabinoside, the 20% of pts with 5-azacytidine and the 19% of pts are not treated or had only supportive therapy. The ratio between responders and non responders pts was 57/43. Each pt presented an average of 1448 events so distributed: 31% of homozygous deletions, 32% of homozygous amplifications, while irrelevant events are related to heterozygous deletions and amplifications. We also found 36% of uniparental disomy (UPD)events. Among all these macroscopic and submicroscopic alterations, we focused on JAK2 gene, which is located on human chromosome 9 at p24.1 locus. It is composed of 25 exons and encodes a 1132 amino-acid protein of 130.7 kDa. This gene was amplified in 12/285 pts (4%) and deleted in 13/285 (5%). All these pts are treated with CHT but only those characterized by JAK2 deletion are responsive in first line. We showed that the group of pts which present this deletion had a better overall survival rate than the group with the amplification of this gene (p-value &amp;lt; 0,01). The deletion event of JAK2 gene involved two different regions: the first region had a length of 7.429 bp involving the intron 4 and was found in the database genome variant (DGV); the second deletion presented a variable length between different pts: the smallest deleted region had a length of 3342 bp and included the exons 17, 18, 19, while the largest one had a length of 76903 bp which included the portion from exon 9 to exon 19. This second deletion has not been described in DGV and it involved the pseudokinase and kinase domain of JAK2 protein. We have identified Copy Number Variation involving important cancer genes in AML. We have identified a new JAK2 deletion involving its Pseudokinase and Kinase domain and we observed that this deletion of JAK2 correleted with overall survival. ELN, AIL, AIRC, PRIN, progetto Regione-Università 2010-12 (L. Bolondi), FP7 NGS-PTL project Citation Format: Viviana Guadagnuolo, Maria Chiara Fontana, Cristina Papayannidis, Marco Manfrini, Antonella Padella, Giorgia Simonetti, Anna Ferrari, Giovanni Marconi, Andrea Ghelli Luserna di Rorà, Stefania Paolini, MariaChiara Abbenante, Sarah Parisi, Chiara Sartor, Emanuela Ottaviani, Giovanni Martinelli. New JAK2 heterozygous loss: A role in overall survival in acute myeloid leukemia patients. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4507.

Abstract 570: Tp53 mutation screening in adult acute myeloid leukemia (AML) patients shows a strong association with complex karyotype and poor outcome

Abstract Introduction: AML is a heterogeneous disease with various chromosomal aberrations. The karyotype at diagnosis provides important prognostic information that influences therapy and outcome. The TP53 gene is the most frequently mutated gene in human tumors. The reported TP53 mutation rate in AML is low (2.1%). In contrast, the incidence of TP53 mutations in AML with a complex aberrant karyotype (CK-AML) is higher (69-78%). Aims: to investigate the frequency and the prognostic role of TP53 mutations in adult AML patients (pts) focusing the screening on subgroups of pts with chromosome abnormalities. Patients and Methods: 106 adult AML pts with FAB-M0, M1, M2, M3, M4, M5, miscellaneous cytogenetic abnormalities and normal karyotype (nK-AML, 14/106 pts) were examined. Forty-four pts (41.1%) showed 3 or more chromosome abnormalities (CK-AML), 42 (39.6%) presented one or two cytogenetic abnormalities (other-AML) and in 6 cases the karyotype was not available. Genomic DNA and/or cDNA were isolated from mononuclear AML blast cells. TP53 mutation screening was performed on all 106 AML pts, in particular in 42 from exon (ex) 2 to 11, in 48 from ex 4 to 11 and in 16 pts from ex 2 to 8. Analysis was focused on coding sequences (RefSeq GRCh37/hg19 NG_017013.2). Results: By PCR and subsequent Sanger sequencing, mutations of TP53 were detected in 23 pts (21.1%). Seven pts revealed 2 mutations. 83% of all mutated pts had CK (19/23) by contrast the frequency of mutations was very low in “no CK-AML” pts (6.5%). Overall, mutated patients included 19/44 with CK-AML (43.2%); 1/6 (16.7%) with nK and 3/42 (7.1%) with other-AML. 26 TP53 point mutations [19 missense, 1 silent, 5 intron and 1 3’ untranslated region (UTR)] and 4 TP53 deletions were found. Twenty-six out of 30 mutations/deletions (86.6%) were located in the DNA binding domain, 2 in the carboxyl-terminal tetramerization and regulatory domains and 2 in the transcriptional activation domain. All mutations in coding regions were classified by the IARC database (http://p53.iarc.fr/TP53GeneVariations.aspx) as deleterious. Of note, alterations of TP53 were significantly associated with poor outcomes in terms of overall survival and disease free-survival (P &amp;lt; 0.0001). Conclusions: Our data demonstrated that mutations of TP53 occur in 21.7% of AML with a higher frequency in the subgroup of CK-AML (p&amp;lt; 0.0001-Fischer's exact test). Since TP53 mutations have predicted to be deleterious and significantly correlated with prognosis, TP53 mutation screening should be recommended in CK-AML pts. Supported by: EuropeanLeukemiaNet, AIL, AIRC, PRIN 2010-2011, Fondazione del Monte di Bologna e Ravenna, European Union Seventh Framework Programme [FP7/2007-2013]. Citation Format: Anna Ferrari, Ilaria Iacobucci, Cristina Papayannidis, Carmen Baldazzi, Chiara Sartor, Emanuela Ottaviani, Nicoletta Testoni, Valentina Robustelli, Margherita Perricone, Claudia Venturi, Maria Chiara Abbenante, Viviana Guadagnuolo, Antonella Padella, Federica Cattina, Giorgia Simonetti, Domenico Russo, Giovanni Martinelli. Tp53 mutation screening in adult acute myeloid leukemia (AML) patients shows a strong association with complex karyotype and poor outcome. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 570. doi:10.1158/1538-7445.AM2014-570

Abstract 3811: A germline polymorphism in the ANRIL (CDKN2BAS) locus is associated with susceptibility to Philadelphia-positive acute lymphoblastic leukemia (ALL)

Abstract Background: The CDKN2A/B locus is inactivated in several haeamatologic malignancies mainly due to deletion, aberrant repression or epigenetic silencing, but little is known about the potential role of its single nucleotide polymorphisms (SNPs) in leukemia susceptibility. Patients and methods: To investigate whether polymorphisms within this locus can correlate with increased susceptibility to haeamatologic malignancies, an association study was performed by genotyping 23 SNPs spanning the MTAP, CDKN2A/B and CDKN2BAS loci, as well as relative intergenic regions, in a case-control cohort made up of 332 samples: 149 leukemia patients, including Philadelphia positive (Ph+) ALL (n=92) and acute myeloid leukemia (AML) samples (n=57), and 183 unrelated healthy controls. The median age was 52 years (18-78 years) and 53 years (21-71 years) in Ph+ ALL and AML patients, respectively. AML cases included FAB M0-M5, miscellaneous cytogenetic abnormalities and normal karyotype subtypes. 6 SNPs were selected on the basis of their previous association with several diseases, such as coronary artery disease, type 2 diabetes mellitus, frailty. The remaining 17 SNPs were selected among those included in the Affymetrix Genome-Wide Human SNP Array 6.0 to deepen the SNPs coverage for the examined region. Genotyping was performed using iPLEX Gold technology and MassARRAY high-throughput DNA analysis with Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (Sequenom, Inc., San Diego, CA). Results: 17 SNPs, spanning the 9p genomic interval that encompasses the MTAP, CDKN2A/B and CDKN2BAS loci, were successfully genotyped and used for investigating their potential associations with the leukemia phenotypes. Potential population stratification affecting the control sample was ruled out as its genotypes distribution satisfies the Hardy-Weinberg equilibrium criterion. The rs564398, mapping to the CDKN2BAS locus that encodes for ANRIL antisense non-coding RNA, showed a statistically significant correlation with the ALL phenotype, with a risk pattern that was compatible with an overdominant model of disease susceptibility and an Odd Ratio (OR) of 2 (95% CI, 1.20 to 3.33; p = 7.1 × 10-3). Since a co-ordinated regulation of ANRIL and p14/ARF, p16/CDKN2A, p15/CDKN2B transcription has been already observed in both physiologic and pathologic conditions, we hypothesized that rs564398 association reflects a condition of high linkage disequilibrium between such polymorphism and a causative variant that is able to alter CDKN2A/B expression profiles by changing ANRIL dosage, thus leading to abnormal proliferative boosts and consequent increased ALL susceptibility. Supported by European LeukemiaNet, AIL, AIRC, FIRB 2006, Ateneo RFO grants, Project of integrated program, Programma di Ricerca Regione – Università 2007 – 2009. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 3811. doi:10.1158/1538-7445.AM2011-3811

Susceptibility to Philadelphia-Positive acute Lymphoblastic Leukemia (ALL) Is Associated with a Germline Polymorphism In the ANRIL (CDKN2BAS) Locus.

AbstractAbstract 1670▪Among genes that are supposed to be involved in altered molecular pathways leading to leukemogenesis, the CDKN2A/B locus is particularly noteworthy since it encodes for the INK4-class cyclin dependent kinase inhibitors p15 INK4B, p16 INK4A and p14 ARF, which act as tumor suppressors. The CDKN2A/B locus has been found to be inactivated in several haeamatologic malignancies mainly due to deletion, aberrant repression or epigenetic silencing, whereas little is known about the potential role of its single nucleotide polymorphisms (SNPs) in leukemia susceptibility. To investigate whether polymorphisms within this broad genomic region can correlate with increased susceptibility to haeamatologic malignancies, an association study was performed by genotyping 23 SNPs spanning the MTAP, CDKN2A/B and CDKN2BAS loci, as well as relative intergenic regions, in a case-control cohort made up of 332 samples: 149 leukemia patients, including Philadelphia positive (Ph+) ALL (n=92) and acute myeloid leukemia (AML) samples (n=57), and 183 unrelated healthy controls. Ph+ ALL patients showed a median age of 52 years (18-78 years); 24 cases (26%) expressed the p210 oncoprotein, 62 (67%) the p190 and 6 (7%) both the proteins. AML patients showed a median age of 53 years (21-71 years) and they included FAB-M0, M1, M2, M3, M4, M5, miscellaneous cytogenetic abnormalities and normal karyotype subtypes. 6 SNPs were selected on the basis of their previous association with several diseases, such as coronary artery disease (rs2891168, rs518394, rs564398, rs10757278), type 2 diabetes mellitus (rs564398), frailty (rs2811712). The remaining 17 SNPs were selected among those included in the Affymetrix Genome-Wide Human SNP Array 6.0 to deepen the SNPs coverage for the examined region. Genotyping was performed using iPLEX Gold technology and MassARRAY high-throughput DNA analysis with Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (Sequenom, Inc., San Diego, CA). Furthermore, all leukemia samples were also previously genotyped using the GeneChip® Human Mapping 250K NspI and Genome-Wide Human SNP 6.0 (Affymetrix). A total of 17 SNPs, spanning the 9p genomic interval that encompasses the MTAP, CDKN2A/B and CDKN2BAS loci, were successfully genotyped and used for investigating their potential associations with the leukemia phenotypes. Five SNPs (rs1012713, rs10965179, rs34011899, rs3731232, rs3218010) with Minor Allele Frequency (MAF) <0.05 in cases and controls, as well as one SNP (rs3931609) showing >20% missing call rates, were instead excluded from the association analysis. Potential population stratification affecting the control sample was ruled out as its genotypes distribution satisfies the Hardy-Weinberg equilibrium criterion. Among the 17 SNPs, rs564398, mapping to the CDKN2BAS locus that encodes for ANRIL antisense non-coding RNA, showed a statistically significant correlation with the ALL phenotype, with a risk pattern that was compatible with an overdominant model of disease susceptibility and an Odd Ratio (OR) of 2 (95% CI, 1.20 to 3.33; p= 7.1 × 10-3). Since a co-ordinated regulation of ANRIL and p14/ARF, p16/CDKN2A, p15/CDKN2B transcription has been already observed in both physiologic and pathologic conditions, we hypothesized that rs564398 association reflects a condition of high linkage disequilibrium between such polymorphism and a causative variant that is able to alter CDKN2A/B expression profiles by changing ANRIL dosage, thus leading to abnormal proliferative boosts and consequent increased ALL susceptibility. Supported by European LeukemiaNet, AIL, AIRC, FIRB 2006, PRIN 2008, Ateneo RFO grants, Project of integrated program (PIO), Programma di Ricerca Regione – Università 2007 – 2009. Disclosures:Baccarani:Novartis: Consultancy, Honoraria; BMS: Consultancy, Honoraria. Martinelli:Novartis: Consultancy, Honoraria; BMS: Honoraria; Pfizer: Consultancy.

Prognostic Implications of the IDH1synonymous SNP rs11554137 In Pediatric and Adult AML: a Children's Oncology Group and Southwest Oncology Group Study

AbstractAbstract 2737Synonymous SNPs may directly impact gene function through various translational or post-translational mechanisms. Further, such “silent” SNPs in the mutational hotspots of AML-associated genes have recently been reported to carry prognostic impact. We aimed to determine the prevalence, clinical associations, and prognostic significance of a known SNP in exon 4 of IDH1, near the location of the frequently-mutated R132 codon. Diagnostic marrow specimens from 253 pediatric AML patients (treated on the COG trial AAML03P1) and 274 adult AML patients (treated on SWOG trials S9031, S9333, or S9500) were analyzed for the presence of SNP rs11554137 via direct sequencing. In the pediatric cohort (median age 9.8 years), SNP rs11554137 was present in 27 of 253 (10.7%) patients. SNP+ pediatric patients did not differ significantly from wild-type patients in terms of sex, racial distribution (African American patients accounted for 23% of SNP+ patients vs. 14% of wild-type patients, P=0.24), bone marrow blast percentage, or age distribution, except in patients aged 0–2 years, who accounted for 44% of SNP+ patients vs. 23% of wild- type patients (P=0.013). Recurrent cytogenetic abnormalities occurred with similar frequencies in both the SNP+ and wild-type pediatric populations, as did FLT3/ITD, NPMc, and CEBPA mutations. Miscellaneous cytogenetic abnormalities accounted for 33% of SNP+ patients vs. 14% of wild-type patients, P=0.033. IDH1 SNP status had no prognostic impact on survival in the pediatric cohort, as SNP+ and wild-type patients had similar rates of five-year overall survival (OS, 76% vs. 63%, P=0.50), disease-free survival (DFS, 48% vs. 53%, P=0.97), and relapse rate (RR, 39% vs. 39%, P=0.94). In the adult cohort (median age 63 years), the IDH1 SNP was present in 30 of 274 (10.9%) patients. A slight female predominance for the SNP (63% vs. 37%, P=0.052) occurred among adult patients. The SNP was more prevalent in African American patients, who accounted for 30% of the SNP+ patients vs. 7% of wild-type patients, P=0.0046. SNP+ patients also had somewhat higher diagnostic bone marrow blast percentages (medians 80% vs. 70%, P=0.025). The normal karyotype subset accounted for similar proportions of SNP+ vs. wild-type patients (42% vs. 46%, P=0.83). Notably, SNP rs11554137 was not present in adult core-binding factor AML. Miscellaneous cytogenetic abnormalities were significantly more common in SNP+ patients (46% vs. 22%, P=0.022). SNP status was not significantly associated with FLT3/ITD status when all adult patients were considered (P=0.14). However, within the normal karyotype subset, FLT3/ITD was present in 90% of SNP+ patients vs. 59% of wild-type patients (P=0.0053). SNP+ patients had somewhat poorer 5 year OS (10% vs. 18%, hazard ratio [HR]=1.17) though this difference was not statistically significant (P=0.44). Among the 142 patients who achieved complete remission (CR), however, 5-year relapse-free survival (RFS) was significantly worse for SNP+ patients (0% vs. 25%, HR = 2.89, P=0.0014). Of the 14 SNP+ patients who achieved CR, 13 relapsed and the 14th patient died of sepsis in remission after 61 days. In multivariate analysis, after adjusting for the effects of age and cytogenetic group, SNP rs11554137 retained an independent prognostic effect (P=0.0062) regarding RFS. Notably, when FLT3/ITD status is included in multivariate analysis, SNP positivity loses independent prognostic significance (HR=1.72, P=0.18). Genome-wide expression profiling was performed on 134 pediatric AML specimens in whom IDH1 SNP status was known. By comparing SNP+ patients with wild-type patients, we derived a distinct gene-expression signature for patients with SNP rs11554137. Among the most upregulated probe sets in the SNP+ cohort were those representing PEX6 and NFYA, both of which interact with the TGF-beta/SMAD signaling network; the retinoid × receptor beta gene RXRB; and the FER gene, a tyrosine kinase critical to FLT3 signaling. The IDH1 SNP rs11554137 is present in approximately 11% of pediatric and adult AML patients, and gene expression profiling data suggests that leukemia in SNP+ patients may have unique biologic features. The SNP was an independent predictor of decreased RFS in adult AML in univariate analysis, but not in multivariate analysis when adjusting for FLT3/ITD status. Disclosures:No relevant conflicts of interest to declare.

Abstract 656: High-resolution molecular allelokaryotyping identifies novel genomic alterations in acute promyelocytic leukemia (APL)

Abstract Introduction: Experimental evidence obtained in transgenic mice revealed that PML-RARA is necessary but not sufficient for the development of APL, suggesting that additional genetic mutations are also required for its development. Aim: To define whether additional submicroscopic genomic alterations may characterize APL and may be used to better classify the disease by dissection of genomic subsets. Methods: 105 adult patients with acute myeloid leukemia were analyzed. These cases included all French-American-British subtypes, miscellaneous cytogenetic abnormalities and normal karyotype subgroups. Among these, the M3 subtype included 28 patients, representing the 33% of the whole study population. Genomic DNA was isolated from blast cells and applied to Genome-Wide Human SNP 6.0 array (Affymetrix, Santa Clara, CA) following the manufacturer's instructions. Fluorescence in situ hybridization, quantitative PCR and nucleotide sequencing were used to confirm genomic alterations. Results: A wide spectrum of different copy number alterations (CNAs) were identified in all cases and no significant difference in the average number of alterations was detected among different leukemia cytogenetic subgroups except for the complex subgroup, which had an average of 55 CNA/patient. In APL cases an average of 8 CNAs per case (range, 1-24) was found. The macroscopic alterations were rare, confirmed conventional cytogenetics and involved trisomy of chromosome 8 in 3 cases, loss of chromosome 6, loss of chromosome 20 and deletions on chromosome 9 and 7. Microscopic CNAs (&amp;lt; 1.5 Mbps) involved every chromosome at least once and predominantly chromosomes 1, 2, 9, 15 and 17. Genetic gains were more common than losses and their median size was 300 kb (range 0.2- 1.4 Mb). The majority of lesions were not recurrent, being identified in only a single patient. Focal genetic alterations were detected at the breakpoints of t(15;17)(q22;q21) in PML and RARA genes. Hemizygous deletions were identified at at 2q33.3-q34 involving ERBB4 (v-erb-a erythroblastic leukemia viral oncogene homolog 4 avian), at 3p11.2 (EPHA3) and at 12p13 (ETV6). Deletions also affected genes involved in cell regulations as CDKN2A (9p21) and RB1 (13q14.2). Most frequent gains affected the TP73 gene at 1p36.3, the oncogenes MYC and PVT1 at 8q24 (4.33 Mb), PRAME at the 22q11.22. Copy neutral loss of heterozygosity events affected 1p34.2-1p32.3, 10p11.2 (MLLT10), 11p11.2 (WT1, CDKN1C, HRAS). Patients with more than 10 CNAs were associated with a worse prognosis. Conclusions: These data demonstrate that different cooperating events may be involved in the generation of APL. These novel findings may be used to stratify patients according to genomic changes and to facilitate the screening for novel therapeutic targets. Supported by: AIL, AIRC, FIRB 2006, Fondazione del Monte di Bologna e Ravenna, Strategico di Ateneo. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr 656.

Gene Copy Number Changes in Myeloid Leukemias.

Abstract A wide range of genetic alterations contribute to the formation of leukemia. These alterations include, but are not limited to, point mutations, gene rearrangements, deletions, amplifications, and a diverse array of epigenetic changes that influence gene expression. Unfortunately, using present methodologies, the identification of all mutations within a leukemia cell is cost prohibitive. Therefore, a more measured approach is needed. In previous work on pediatric acute lymphoblastic leukemia (ALL), we demonstrated the power of genome-wide copy number analysis for the identification of recurrent mutations. Our analyses revealed deletion, amplification, point mutation, and structural rearrangement in genes encoding key regulators of B lymphocyte development and differentiation in 40% of B-progenitor ALL. PAX5 was the most frequent target of somatic mutation, being altered in 30% of cases. Deletions were also detected in E2A, EBF1, LEF1, Ikaros, and Aiolos. These findings suggest that direct disruption of pathways controlling B cell development and differentiation contributes to ALL pathogenesis. We have now extended our analysis to pediatric acute myeloid leukemia (AML). This analysis revealed a mean number of somatic copy number abnormalities (CNA) of 2.38 per cases, with gains equal to losses. Strikingly, the most frequent recurrent CNA were at the breakpoints of known chromosomal translocations, including t(8;21) and inv(16). Based on this observation, we examined genes with CNA in their 5’ or 3’ regions for their possible involvement in cryptic translocations. Using this approach, two cryptic translocations were identified: t(5;11) [NUP98-NSD1] in four cases, and t(6;11)[MLL-MLLT4] in two cases. A number of other genes were identified as the target of recurrent abnormalities, including amplification of CCDC26, ABCC4, and deletion of FAM20C, BCOR. The most frequent abnormality was the amplification of CCDC26 at 8q24.21 (N=15; focal in four cases, broad in 11 cases), which encodes a putative mediator of retinoic acid receptor signaling. The other recurrent lesions occurred in two or fewer cases. Thus many fewer recurrent lesions were identified in AML than in ALL. In addition to the CNA, we also sequenced genes previously shown to be mutated in AML including NRAS, KRAS, PTPN11, BRAF, SOS1, FLT3, CEPBA, NPM1, AML1, CKIT, ERG, TP53, PTEN, and GATA1. Mutations were identified in all genes except BRAF, or SOS1, with the frequency of mutations varying across the different AML genetic subtypes. FAB M7 cases had the highest frequency of CNA but had sequence mutations limited to GATA1. Mutations of FLT3, CEPBA, and NPM1 were most frequent in cases with no or miscellaneous cytogenetic abnormalities, whereas NRAS mutations were most frequent in t(8;21) and inv(16) cases. Importantly, approximately 20% of the cases with recurring translocations had no other sequence or numerical abnormalities. Collectively, these data demonstrate a marked difference in the spectrum of genetic lesions between AML and ALL, with AML having fewer CNA per cases, a lower frequency of recurrent lesions, and a higher frequency of cases lacking any CNA. These data suggest a fundamental difference in the type of collaborating mutations between AML and ALL.

Genome-wide profiling of DNA copy number abnormalities in adult acute myeloid leukemia (AML) using high density single nucleotide polymorphism array

7023 Background: Acute myeloid leukemia (AML) is a heterogeneous disease with various chromosomal aberrations. The karyotype at diagnosis provides important prognostic information that influences therapy and outcome of this disease. However, using conventional chromosome banding techniques alone, karyotype abnormalities are detected in only half of all AML cases. We performed a genome-wide, single-nucleotide polymorphism (SNP) analysis in order (i) to identify novel genomic regions of interest in normal karyotype AML; (ii) novel candidate regions and disease-related genes, in patients with complex karyotypes. Methods: 20 AML patients were analyzed until now. Cases included FAB-M0, M1, M2, M4, M5, miscellaneous cytogenetic abnormalities and normal karyotype. 250 ng of genomic DNA were processed on 500K SNP array according to protocols provided by the manufacturer. Copy number state was calculated with respect to a set of 48 Hapmap normal individuals and a set of samples obtained from acute leukaemia cases in remission using Partek Genomics Suite. Results: An enormous spectrum of different genetic lesions (gains/losses) involving complete chromosome arms (del 16q, i(13q10), del 3p, del 7p, monosomy 9) or submicroscopic genomic intervals were identified in a substantial proportion of cases without differences in the frequency of losses or gains. The most frequent genomic gains affected: 9p12 (ZNF658B, FOXD4L2), 10q11.2 (PPYR1), 5q31-q33 (CDX1) and 8p23.2 (CSMD1). The most frequent deletions were identified in regions lacking annotated genes. Other recurring genetic lesions were uncommon. Marked differences in the combination of copy number anomalies were identified across the different genetic subtypes of AML. Some lesions affected regions with a single gene, such as: ETAA1, FIGN, STK32B, PRAGMIN, PCM1, GLIS3, MRGPRX1, SESN3, BCL2L14. Conclusions: Different genetic anomalies were identified by using an SNP-based chip array. Genetic abnormalities are associated with disease subtype. Supported by: European LeukemiaNet, AIL, AIRC, FIRB 2006, Fondazione del Monte di Bologna e Ravenna, PIO project 2007, Strategico di Ateneo. No significant financial relationships to disclose.

High-Resolution Genome-Wide Profiling of DNA Copy Number Abnormalities and Gene Resequencing in Pediatric Acute Myeloid Leukemia (AML).

To define the complement of genetic lesions in pediatric AML, we performed high resolution genome-wide analysis on bone marrow blasts from 111 cases using Affymetrix 100K and 500K SNP microarrays and genomic resequencing. Cases included t(8;21) (N=20), inv(16) (N=16), t(15;17) (N=7), MLL rearranged (N=16), FAB-M7 (N=9), miscellaneous cytogenetic abnormalities (N=24) and normal karyotype (N=19). Germline DNA was available for 67 of the cases. The mean number of somatic copy number abnormalities (CNA) was 2.38 (range 0–45) with 1.32 gains (range 0–41) and 1.06 losses (range 0–12). Focal CNAs were detected at the breakpoints of known chromosomal translocations, most commonly in inv(16) cases (7/16) and t(8;21) cases (4/20). Based on this observation, we examined focal CNAs that affected the 5' or 3' regions of genes for their possible involvement in cryptic translocations. CNAs involving NUP98 and MLL led to the identification of two cases with t(5;11)[NUP98-NSD1], and two with t(6;11)[MLL-MLLT4] not evident on cytogenetic analysis. RT-PCR for these fusions identified two additional NUP98-NSD1 cases. Identification of focal CNAs are thus useful in identifying clinically significant translocations in AML. Other recurring somatic CNAs were uncommon. Twelve regions containing a single gene were identified, including amplification of EBF, PRDM5, CCDC26*, GDF10, ABCC4*, and deletion of FAM20C*, TUSC1, GPC5, A2BP1, INSR, BCOR* (* involved in N>2 cases). The most frequent abnormality was amplification of CCDC26 at 8q24.21 (N=15; focal in 4 cases, broad in 11) which encodes a putative mediator of retinoic acid receptor signaling. The focal amplifications are predicted to abolish expression of normal transcripts. Seventeen additional recurring CNAs involving 2 to 25 genes were identified, many of which contain known or putative tumor suppressor genes or oncogenes, including deletions of regions at 7q36, 9p21 (CDKN2A/B), 11p14-p12 (WT1, WIT1), and amplification at 19p13 (JUNB, LYL1). These regions showed significant correlation between copy number and local gene expression in an integrated analysis of CNA and Affymetrix U133A gene expression data. Copy-neutral LOH was uncommon and recurrent only with LOH of chromosome 13 in 3 cases. Sequence analysis has been completed for several genes involved in recurring CNAs (CCDC26, TUSC1, FBXW7) and for genes known to be mutated in AML (N/KRAS, PTPN11, BRAF, SOS1, FLT3, CEPBA, NPM1, AML1, CKIT, GATA1). No mutations were identified in CCDC26, TUSC1, FBXW7, BRAF, or SOS1. Marked differences in the combination of CNA and sequence mutation were identified across the different genetic subtypes of AML. FAB M7 cases had the highest frequency of CNA (mean 9.3 lesions/case) but had sequence mutations limited to GATA1. Mutations of FLT3, CEPBA and NPM1 were most frequent in cases with no or miscellaneous cytogenetic abnormalities, where as RAS mutations were most frequent in t(8;21) and inv(16) cases. Importantly, approximately 30% of the cases with recurring translocations had no other sequence or numerical abnormalities. These data demonstrated that, in contrast to pediatric ALL, AML is characterized by relatively few recurring CNAs, and that spectrum of CNA and sequence mutation is significantly associated with disease subtype.

High-Resolution SNP Array Profiling of Relapsed Acute Leukemia Identifies Genomic Abnormalities Distinct from Those Present at Diagnosis.

Failure of initial remission-induction therapy and disease recurrence remains a major problem in the management of acute leukemia, however the nature of the biologic factors promoting relapse are incompletely understood. To identify genomic abnormalities associated with relapse, we performed high-resolution, genome-wide analysis of DNA copy number abnormalities and loss-of heterozygosity using Affymetrix single nucleotide polymorphism (SNP) microarrays in 33 cases of relapsed acute leukemia. Sixteen ALL (2 ETV6-RUNX1, 2 MLL rearranged, 6 pseudodiploid or cytogenetically normal B-progenitor ALL, and 6 T-lineage) and 17 AML (two RUNX1-RUNX1T1 [AML1-ETO], two MLL-rearranged, one M7, and 12 with normal karyotype or miscellaneous cytogenetic abnormalities) were studied. Samples with less than 80% blasts were flow sorted to at least 90% purity prior to DNA extraction. Diagnostic samples were available for all cases, and germline samples for 22. DNA copy number and LOH analysis was performed using Affymetrix 250k Nsp and Sty arrays was performed for all samples. Data were analysed using a karyotype-guided normalization algorithm, dChipSNP, and circular binary segmentation. In a detailed comparative analysis of paired diagnostic and relapse samples, changes in DNA copy number abnormalities were identified in the relapse sample in 14 of 16 (87.5%) ALL cases. A striking finding was loss of copy number lesions present at diagnosis in 8/16 ALL relapse samples, and the acquisition of new copy number lesions in 4 of these 8 samples. In each case, the pattern of deletions at antigen receptor loci was comparable between relapse and diagnosis, suggesting the emergence of a related leukemic clone, rather than the development of a distinct second leukemia. An additional 8 ALL relapse samples retained the copy number lesions present at diagnosis, but 6 of these acquired additional copy number abnormalities at relapse, a finding further suggestive of significant clonal evolution. Of the newly acquired copy number abnormalities in the relapse samples, deletions (62.5% of cases) were more common than gains (12.5%). In constrast to ALL a more restricted range of copy number abnormalities was seen in AML, with new abnormalities at relapse seen in 5/17 (29.4%) cases, and deletions (29.4%) outnumbering gains (17.6%). The loss of lesions present at diagnosis was only observed in two AML relapse samples. Examining the entire cohort, the CDKN2A/B locus was most commonly involved (N=5), gains of 1q were noted in two cases, otherwise all observed copy number changes were noted in single cases only, and included focal deletions of ERG and RUNX1. Copy neutral loss of heterozygosity was uncommon, with the exception of three AML cases with UPD of the entire chromosome 13. These observations indicate that relapse is frequently the result of the emergence of a leukemic clone that shows significant genetic differences from the diagnostic clone. Whether these represent rare clones present at the time of diagnosis or are the emergence of new clones as the result of ongoing genomic abnormalities can now be determined using genomic probes specific for the newly acquired deletions.

Genome-Wide Profiling of DNA Copy Number Abnormalities and Loss-Of-Heterozygosity in Pediatric Acute Myeloid Leukemia Using High-Resolution Single Nucleotide Polymorphism Microarrays.

To identify a comprehensive registry of oncogenic lesions in pediatric acute myeloblastic leukemia (AML), we used Affymetrix single nucleotide polymorphism (SNP) arrays to examine changes in DNA copy number and loss-of-heterozygosity (LOH) in leukemic blasts from 112 cases of pediatric AML and corresponding remission samples from 63 of these cases. The analyzed cases included t(8;21)[AML1-ETO] (n=20), inv(16)[CBFβ-MYH11](n=16), t(15;17)[PML-RARα] (n=7), MLL rearranged (n=17), FAB M7 (n=9) and normal cytogenetics or miscellaneous cytogenetic abnormalities (n=43). Four SNP arrays (50K Hind and Xba, 250K Sty and Nsp) were used to interrogate over 615,000 markers at a mean inter-marker distance of 4.8 kb. Combined data were analyzed using dChipSNP and a modified array normalization algorithm using only those SNPs from regions known to be diploid by routine karyotyping. These analyses not only detected known whole or partial chromosomal losses or gains, but also detected numerous copy number abnormalities that were not evident by conventional cytogenetics. Somatic DNA copy number abnormalities were identified in 102 (91.1%) cases. The mean number of lesions per patient was 3.2 (range 1–12), with a mean of 2.13 deletions/whole chromosome losses and 1.02 amplifications/whole chromosome gains per patient. Deletions were detected in the leukemic blasts from over 90% of patients, whereas amplifications were only seen in the leukemic blasts from 54% of patients. The vast majority of deletions were focal (<20 Mb) with less than 20% of cases containing larger deletions or losses of whole chromosomes or chromosomal arms. No differences in the frequency of deletions were observed among the different genetic sybtypes of AML. Lesions identified in 2 or more cases included deletions of CDKN2A/B (9p21.3, n=4) and FOXE1 (9p22, n=3), and amplifications of ETS1 (11q24.3, n=3) and MYST4 (10q22.2, n=2). For each of the listed genes, at least one case was identified harboring a focal deletion or amplification confined to the specific gene, thus definitively identifying the gene as the target of the alteration. In addition to these recurrent lesions, copy number changes were identified in regions containing 5 or fewer genes in single cases, including deletions involving the tumor suppressor candidate TUSC3 (8p22), alpha 3 catenin (10q21.3), and amplifications involving FGFR activating protein 1 and RAS homolog G (11p15.4). Importantly, within our cohort of de novo AMLs no focal sub-microscopic lesions involving 5q or 7q were identified. Similarly, copy-neutral LOH (uniparental disomy) that was not adjacent to an identified region of deletion or amplification was uncommon. Taken together, these data demonstrate a surprisingly low frequency of copy number changes in pediatric AML with focal deletions measured at an average resolution of 5-10 KB across the genome predominating over focal amplifications. Moreover, the identified lesions appear to target a rather large number of different genes. Correlating the identified copy number changes with mutations in other genes known to be involved in leukemogenesis including, NRAS, KRAS, FLT3, NPM1, CEBPA, BRAF, PTPN11, AML1 and KIT, should provide valuable insights in the molecular pathology of AML.

Adult Acute Lymphoblastic Leukemia and Cytogenetic Abnormalities in Different Racial and Ethnic Subgroups.

Inferior survival of African American and Hispanic patients with acute lymphoblastic leukemia (ALL) has been reported. The reason for this is unclear. A retrospective analysis was conducted to see if abnormal cytogenetics account for the differences. We have analyzed cytogenetic studies of 39 adults (16 year and older) with newly diagnosed ALL who were consecutively treated at John Stroger Hospital of Cook County between 1997 and 2005. The study population included 13 (33%) African Americans, 18 (46%) Hispanics, 6 (16%) Caucasians, and 2 (5 %) other ethnic background adults. Male to female ratio was 2:1. Mean age at diagnosis was 26 (range between 16 and 71 years). A clonal cytogenetic abnormality was detected in 27 patients, and 12 patients had normal karyotype. Patients with t (9;22)(q34;q11), BCR-ABL + by FISH, t (4;11)(q21;q23), +8, −7 cytogenetic abnormalities were assigned to an unfavorable cytogenetic group. This group was composed of 14 patients. None of our study patients had favorable cytogenetics: del (12p), t(12p), t(14q11-q23), inv(14)(q11;q32) or t(10;14)(q24;q11). Remaining 25 patients with normal karyotype (n=12) or miscellaneous cytogenetic abnormalities (n=13) were classified as normal risk group. 54 % of African Americans had unfavorable cytogenetics, compared with 33 % Caucasians and 17 % Hispanics. Translocation t (9;22) alone or in association with other cytogenetic abnormalities was most commonly seen. 9 patients had single, and 18 had multiple chromosomal changes. Hispanic group had more complex cytogenetic changes when compared with the African American or Caucasian groups. Unfavorable cytogenetic abnormalities may account for inferior survival in African Americans, but other factors such as compliance or pharmacogenetics should be evaluated, especially in the Hispanic patient population.

The clinical significance of tumor necrosis factor-α plasma level in patients having chronic lymphocytic leukemia

Tumor necrosis factor-alpha (TNF-alpha), a cytokine possessing pleiotropic biological activities, is produced by leukemic lymphocytes in patients with chronic lymphocytic leukemia (CLL) and acts as an autocrine and paracrine growth factor in this disease. In this study, TNF-alpha levels were determined in 150 patients with CLL and correlated with disease characteristics, prognostic factors, and survival. The mean TNF-alpha plasma concentration in the patients with CLL was significantly higher than in the healthy control population (16.4 versus 8.7 pg/mL; P <.0001). Patients having an elevated TNF-alpha level had more advanced Rai and Binet stage disease, higher serum beta(2)-microglobulin (beta(2)M) levels, a greater percentage of cells expressing CD38, and lower hemoglobin and platelet levels. Patients having chromosomal abnormalities such as 11q deletion, trisomy 12, and chromosome 17 aberrations had a higher mean TNF-alpha level (27.5 pg/mL) than patients having a diploid karyotype or other miscellaneous cytogenetic abnormalities (14.2 pg/mL; P <.001). The TNF-alpha level was a predictor of survival when the Cox proportional hazards model was used with TNF-alpha entered as a continuous variable (P =.0001). Also, patients having a TNF-alpha level above the mean value of 14 pg/mL had significantly shorter survival duration (P =.00001). The TNF-alpha level remained predictive of survival in Cox multivariate analysis independent of Rai staging and beta(2)M, hemoglobin, prior therapy, white cell count, and platelet level (P =.005). We conclude that the TNF-alpha level serves as a prognostic factor in patients with CLL and that inhibition of TNF-alpha in these patients could have therapeutic importance.