ALBERT

Description: Introduction. Chronic lymphocytic leukemia (CLL) with 11q deletion has been associated to a poor prognosis, but the clinical course of patients carrying this lesion is variable. This aberration, most often monoallelic, is present in 10-17% of newly diagnosed CLL and in 20-30% of patients with progressive or chemorefractory disease. The minimal deleted region (MDR) (2-3 Mbp) is located on the 11q22.3-q23.1 region and includes ATM. Moreover, 30-40% of 11q- CLL have also an inactivating ATM mutation on the other allele. The deleted region on 11q can also include BIRC3, a gene that is often deleted or mutated in advanced/chemorefractory stages of the disease. Although BIRC3 disruption has been associated to a poor prognosis, its prognostic implications in addition to ATM deletion are not well defined. The aim of this study was to perform a copy number aberration (CNA) and gene sequencing analyses on a cohort of CLL patients with 11q- in order to identify subgroups with potential prognostic relevance based on: i) the inclusion of BIRC3 in the deleted region; ii) the presence of BIRC3 mutation; iii) the presence of other CNAs. Methods. The study has included 55 untreated CLL patients followed at our Institution or enrolled in GIMEMA clinical trials (2003-2013). Genomic DNA was extracted from peripheral blood samples. CNA analysis was performed by genomic hybridization on the CytoScan HD array (Affymetrix), which contains more than 2.6 x 106 markers for copy number analysis and 750.000 SNPs. Data were analyzed using both Partek Genomics Suite and ChAS (Chromosome Analysis Suite, Affymetrix) software. The resulting CNAs were verified by visual examination of the plotted copy number profiles. BIRC3 mutations (exons 6-9) were evaluated by Sanger sequencing. Time to first treatment (TFT) was calculated from the date of diagnosis to the date of first therapy or last follow-up; progression-free survival (PFS) from the date of first therapy to the date of progression, death or last follow-up. Results. Baseline characteristics of the 55 cases were as follows: median age at diagnosis 59 years (range 39-84), male gender in 81.8% of patients, progressive disease in 62%. All patients showed 11q- by FISH (median 80%, range 25-99% of nuclei); germline IGHV were present in 96.4% of cases; TP53 deletion in 1 case and TP53 mutation in none; NOTCH1 mutation in 4/40 cases; SF3B1 mutation in 5/40 (all mutually exclusive with only 1 case having both SF3B1 and BIRC3 mutations). By CytoScan HD array, the size of 11q- was very variable, ranging from 0.36 Mbp to 65.14 Mbp; the MDR was located on 11q22.3 region, encompassing 4 genes (ACAT1, ATM, CUL5, NPAT). BIRC3 was included in the deleted region in 45/55 cases (81.8%) and was mutated in 4/54 (7.5%), being always deleted on the other allele. Beside 11q-, 51 cases (92.7%) showed several additional CNAs (average 4.9, range 1-14 per patient), with 5 recurrent lesions: 2p gain in 11 cases, del4(p15.2) in 6, del19(p13.3) in 6, 8q gain in 5 and del4(q22.1) in 4. BIRC3 deletion was not associated to the number of additional CNAs nor to specific CNA. After a follow-up of 59.6 months (range 7.4-229.7), 40 of 47 evaluable patients have received treatment (median TFT 15.8 months, range 0-167). BIRC3 deleted cases (n=37) showed a TFT not significantly different from WT cases (n=10). Conversely, BIRC3 mutation was associated to a shorter TFT (p 3 CNAs larger than 5 Mb (n=14) or 〉10 CNAs (n=5), or the presence of 2p gain, del4(p15.2), del19(p13.3) or 8q gain. So far, 22 patients have been evaluated for PFS after first-line therapy (median 28.7 months). BIRC3 deleted cases (n=17) were not associated to a shorter PFS compared to WT cases (n=5), in line with the results from Rose-Zerilli et al (Haematologica 2014). Conclusions. Among CLL with 11q-: 1) BIRC3 deletion involves more than 80% of cases, whilst the mutation is rare (7.5%); 2) BIRC3 deletion is not associated to a higher genomic complexity; 3) BIRC3 deletion does not seem to influence TFT or PFS of 11q- CLL; 4) BIRC3 mutation is strongly associated to a short TFT; 5) BIRC3 biallelic lesions can be associated to a marked hyperleucocytosis at diagnosis and immediate need of treatment. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 1784 Introduction: The introduction of whole exome sequencing has allowed to unravel novel molecular lesions in CLL. NOTCH1, SF3B1 and BIRC3 mutations are detected, according to the phases of disease, in 4–12%, 5–17% and 4–24% of patients, respectively. In retrospective studies, their presence has been shown to correlate with overall survival (OS) and treatment-free interval shortening. Aims: To define the incidence, correlation with known prognostic factors and clinical impact of NOTCH1, SF3B1 and BIRC3 mutations in CLL patients undergoing first-line treatment. Methods: We evaluated 162 CLL patients enrolled in the GIMEMA LLC0405 protocol (n=80) for patients aged 65 yrs or 60–65 if not eligible for fludarabine-based programs). In the GIMEMA LLC0405 protocol, patients were stratified into low and high-risk: patients with del17p or with del11q plus an unmutated IGHV status and/or CD38 positivity and/or ZAP70 positivity were considered as high-risk (HR) and underwent Fludarabine plus Campath, followed by stem cell transplantation procedures, whereas low-risk patients received Fludarabine and Cyclophosphamide. The MLL21445 protocol consisted of 8 cycles of Chlorambucil and 6 of Rituximab induction treatment. NOTCH1 (exon 34), SF3B1 (exons 14 and 15) and BIRC3 (exons 2–9, including splicing sites) were screened by Sanger sequencing on either genomic DNA (gDNA) or whole genome amplified DNA (WGA) collected at the time of treatment. These studies were not part of the clinical protocols. Results: NOTCH1 mutations were detected at the time of treatment in 18 cases (22%) enrolled in the LLC0405 study. There was a significant association with high-risk stratification (p=0.036), namely with an IGHV unmutated status (p=0.0035), CD38 (p=0.03), +12 (p=0.034) and, partly, ZAP-70 expression (p=0.059). While the overall response rate (ORR) did not differ between NOTCH1 mutated vs wild-type (WT) cases (82% vs 77%, respectively), the complete response (CR) rate was significantly lower in NOTCH1 mutated patients (43% for WT vs 17% for NOTCH1 mutated cases; p=0.05). So far, no significant difference between mutated and WT patients has emerged in terms of OS and progression-free survival (PFS); this may be contributed by the fact that most NOTCH1 mutated cases were HR and were therefore treated more aggressively. SF3B1 mutations were recorded in 9 cases (11%); no significant associations were found with known biological parameters and, so far, with the ORR and CR rate. A single case harbored a BIRC3 mutation; this patient had an IGHV unmutated status, no FISH abnormalities and a concomitant SF3B1 mutation. In the ML21445 cohort, NOTCH1 mutations were found in 12 cases (15%), were associated with an unmutated IGHV status (p=0.047) and ZAP-70 expression (p=0.007), and did not impact on the ORR and CR rate. SF3B1 mutations were found in 11 cases (13%); no significant associations were found with known biological parameters and the ORR rate. Of interest, only 1/11 SF3B1 mutated patients achieved a CR. BIRC3 mutations were recorded in 3 patients (3.6%); of these, 2 were IGHV mutated, 1 had no cytogenetic abnormalities and 1 carried a del11q, while the third patient was IGHV unmutated status and had no cytogenetic abnormalities. No NOTCH1 and/or SF3B1 mutations were detected. Overall, NOTCH1, SF3B1 and BIRC3 mutations were largely mutually exclusive among each other and with TP53 lesions in the whole cohort. Conclusions: This study confirms the association of NOTCH1 mutations with unfavorable biologic markers and +12, while the presence of SF3B1 mutations was not coupled to poor prognostic markers in CLL patients requiring first-line treatment. Furthermore, it suggests that NOTCH1 mutations impact on the CR rate of young patients receiving Fluda-based regimens, while SF3B1 appears to impact on the CR rate of elderly patients treated with Chlorambucil and Rituximab. Given the small numbers of patients harboring BIRC3, it is at present difficult to draw any conclusion on the clinical impact of this mutation in the cohort of patients hereby analyzed. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 294 Rituximab plus fludarabine/cyclophosphamide (R-FC) is currently the standard of care for fit patients with untreated or relapsed CLL. However, patients with CLL are predominantly an elderly population and many of these patients may have comorbidities that make them less suitable to receive fludarabine-containing therapy. Chlorambucil-based treatments are frequently used for these patients despite the fact that clinical benefits are limited. There is a need for well-tolerated and more efficacious treatment regimens for these patients. The ML21445 study evaluated the combination of rituximab and chlorambucil (R-chlorambucil) as first-line treatment for patients with CLL considered ineligible for treatment with the current standard of care, R-FC. Patients aged 〉65 years (or 60–65 years and ineligible for fludarabine) were treated with eight 28-day cycles of chlorambucil (8 mg/m2/day Days 1–7) with rituximab administered on Day 1 of cycle 3 (375 mg/m2) and cycles 4–8 (500 mg/m2). Patients with a response at the end of induction were randomized to rituximab maintenance therapy (375 mg/m2 every 8 weeks for 2 years) or observation. The induction phase of the study is complete while the maintenance phase is still ongoing. The overall response rate (ORR) in 85 patients who received at least one dose of rituximab during induction was 81.2% (n = 69) with 16.5% (n = 14) achieving a complete response (CR) and 2.4% (n = 2) a CR with incomplete bone marrow recovery (CRi). ORR and CR rates were similar across the different Binet stages (ORR: Binet A 86.4%, Binet B 79.6%, Binet C 78.6%) and age categories (ORR: 60–64 years 84.6%, 65–69 years 85.2%, 70–74 years 75.0%, ≥75 years 81.0%). Two of four patients aged ≥80 years responded to induction treatment. Logistic regression analysis revealed no correlation between known biological prognostic factors – CD38, cytogenetics, IGHV mutational status, ZAP-70, thymidine kinase, soluble CD23, and beta-2 microglobulin – and response to treatment. To further investigate possible factors influencing response, pre-treatment patterns of gene expression were analyzed in different patient subgroups. Material was available for 62 patients, including 16 with CR/CRi, 41 partial responders and 5 non-responders. In an exploratory analysis, mRNA expression was examined using Affymetrix® Human Genome U133 microarrays. This revealed marked differences in pre-treatment gene expression profiles between response groups. Non-responders showed a homogeneous gene expression signature involving up-modulation of transcripts involved in anti-apoptotic and pro-proliferative pathways, including K-ras and N-ras. CR/CRi patients also showed a homogeneous pattern of gene expression that was clearly distinct from non-responding patients, while patients with a partial response showed a more heterogeneous pattern of gene expression before treatment. These initial findings reflect the heterogeneity of CLL and suggest that microarray analysis of gene expression may be useful in predicting response to R-chlorambucil in elderly patients with CLL. Disclosures: Foa: Roche: Consultancy, Speakers Bureau. Cuneo:Roche: Consultancy, Speakers Bureau. Montillo:Roche: Membership on an entity's Board of Directors or advisory committees. Alietti:Roche: Employment. Runggaldier:Roche: Employment. Gamba:Roche Italia: Employment.