Initial Central Nervous System Involvement Research Articles

Incidence and prognostic value of central nervous system involvement in infants with B-cell precursor acute lymphoblastic leukemia treated according to the MLL-Baby protocol.

The aim of the study was to evaluate the incidence and prognostic impact of central nervous system (CNS) involvement in infants with B-cell precursor acute lymphoblastic leukemia (BCP-ALL), as well as its relation with minimal residual disease (MRD) data. A total of 139 consecutive infants with BCP-ALL from the MLL-Baby trial were studied. Cerebrospinal fluid (CSF) samples were investigated by microscopy of cytospin slides. MRD was evaluated according to the protocol schedule by flow cytometry and PCR for fusion gene transcripts (FGT). Involvement of the CNS at any level was found in 50 infants (36.0%). The incidence of CNS involvement was higher in patients with KMT2A gene rearrangements (44.0% for KMT2A-r vs. 15.4% for KMT2A-g, p=.003). The outcome of CNS-positive infants was significantly worse than that of CNS-negative infants, although this prognostic impact was limited to the KMT2A-r group (event-free survival 0.21 for CNS-positive vs. 0.48 for CNS-negative infants, p=.044). CNS-positive infants could not be treated successfully by conventional chemotherapy alone, irrespective of the rapidity of MRD response. In contrast, the combination of initial CNS negativity and FGT-MRD negativity identified a group comprising up to one-third of infants with KMT2A-r ALL who can be treated with chemotherapy and achieve very good outcomes (disease-free survival above 95%), and remaining patients should be allocated to receive other types of treatment. We can conclude that this combination of initial CNS involvement and MRD data can significantly improve risk-group allocation in future clinical trials enrolling infants with ALL.

Peripheral T-Cell Lymphomas Involving the Central Nervous System: A Report From the Czech Lymphoma Study Group Registry.

IntroductionWe analyzed the incidence, risk factors of central nervous system (CNS) relapse, and outcome of CNS involvement in patients with peripheral T-cell lymphomas (PTCL) from the Czech Lymphoma Study Group Registry NiHiL (Clinical Trial gov. NCT03199066).Materials and MethodsOut of 1,040 patients with PTCL, we identified 29 patients (2.79%) with CNS involvement: 2 patients with primary CNS T cell lymphoma, 11 patients with CNS and systemic disease at diagnosis, and 16 patients (1.54%) at CNS relapse. The most common histology with CNS disease was PTCL, not otherwise specified. Progression-free survival (PFS) was defined as the time interval from diagnosis to progression or death. PFS-2 was defined as the interval from the date of a new relapse until the next relapse.ResultsPatients with testicular involvement received intrathecal prophylaxis with methotrexate. High-dose methotrexate-based treatment was administered in 44.8% of patients with CNS disease. Median follow-up was 71.3 months. The difference between the median PFS of 1,027 patients without initial CNS disease (32.6 months) and 11 patients with initial CNS and systemic disease (4.8 months) was significant (p = 0.04). The difference between the median PFS2 in CNS relapses (10.1 months) and 493 relapses outside of CNS (9.1 months) was not significant (p = 0.6). Risk factors for CNS relapses included the following: involvement of more than one extranodal site (p = 0.008), soft tissue involvement (p = 0.003), testicular involvement (p = 0.046), and the presence of B symptoms (p = 0.035). The difference between the median OS of 1,027 patients without initial CNS disease (46.0 months) and 11 patients with initial CNS and systemic disease (18.2 months) was significant (p = 0.02). The median OS2 in CNS relapses was 11.8 months and that in relapses outside of CNS was 21.3 months. CNS involvement was not associated with a significantly worse OS compared to relapsed/refractory patients without CNS involvement (p = 0.1).ConclusionsThe incidence of CNS disease at the time of diagnosis and at relapse in PTCL is low and usually associated with other systemic involvement. The prognosis of PTCL with initial CNS involvement is significantly worse when compared to patients without CNS disease at diagnosis. The outcome of CNS relapse is comparable with relapsed PTCL outside of CNS. The optimal treatment is not defined yet.

Impact of Central Nervous System Involvement in Adult Patients with Acute Lymphoblastic Leukemia Treated in a Pediatrics-Inspired Protocol - a Graall Study

▪Background: The prognosis of central nervous system (CNS) involvement in adult patients with acute lymphoblastic leukemia (ALL) has been historically associated with a dismal outcome. Whereas the prognosis of adult patients with ALL has greatly improved since the advent of pediatrics-inspired regimens, the prognostic impact of CNS involvement has not been formerly reevaluated. We report herein the impact of CNS involvement in patients included in the pediatric-inspired prospective GRAALL-2005 study.Methods: All patients received a 5-drug induction therapy with native E. Coli-ASP intravenous injections. Patients in complete remission (CR) received two consolidation courses with alternating cycles including high dose cytarabine (2g/m2/12h on days 1 and 2), high dose methotrexate (3 g/m2 on day 1), and cyclophosphamide. All patients in persistent CR and with no indication for allogeneic stem cell transplantation (SCT) received late intensification, followed by one last consolidation course. Patients with initial CNS involvement, clinically and/or cytologically (cerebrospinal fluid), were recommended to receive an increased number of triple intrathecal therapy, CNS irradiation, and were eligible for allogeneic SCT in first CR. They received less Asp injections during induction therapy to avoid CNS adverse events. CNS irradiation included two lateral fields encompassing the skull, facial, the base of the skull, and the first two cervical vertebrae at a dose of 24 grays for those not receiving allogeneic SCT and 15 grays for those receiving allogeneic SCT.Results: Between 2006 and 2014, 784 adult patients with newly diagnosed Philadelphia-negative ALL were included with 55 (7%) having initial CNS involvement. These patients were more likely to be of T-phenotype (51 versus 32%, p=.004) and had more white blood cells at diagnosis (median 23 G/l versus 11 G/l, p=.02). Most patients (36 pts, 66%) were classified as CNS-3 (> 5 white blood cells/µl and a positive cytospin and/or clinical signs) whereas 5 patients (9%) were CNS-2 (< 5 white blood cells/µl and a positive cytospin), and 14 (25%) have data pending. Among patients with details regarding CNS involvement, 25/41 (61%) had clinical signs including trigeminal anesthesia (9 pts, 36%), facial paralysis (4 pts, 16%), extremities paresthesia (4 pts, 16%), visual signs (2 pts, 8%), meningeal syndrome (2 pts, 8%), and motor deficit (2 pts, 8%), and 4/18 (22%) had radiological signs. Induction death, CR1 rate, and negative minimal residual disease after induction were similar whether patients had CNS involvement or not (6 vs 6%, 89 vs 89%, 73 vs 62%, 26 vs 22%, respectively). Patients with CNS involvement had a worse outcome than those without with a median event-free survival (EFS) of 391 days (versus not reached for patients without CNS involvement, HR: 1.7, 95% CI: 1.2 - 2.5, p=.002) and a median overall survival (OS) of 608 days (versus not reached for patients without CNS involvement, HR: 1.8, 95% CI: 1.3 - 2.6, p=.001) (figure). Similar results were observed when patients who received allogeneic SCT in CR1 were censored at the time of graft. As recommended, patients with CNS involvement were more likely to receive allogeneic SCT than those without (53 versus 34%, p=.01), with a median time of 169 days. A 150-day landmark analysis, excluding 12 patients with an EFS event before 150 days, was performed to study the impact of allogeneic SCT on the outcome of patients with CNS involvement. Allogeneic SCT had no impact on either EFS (HR: .5, 95% CI: .2 - 1.2, p=.15) or OS (HR: .8, 95% CI: .3 - 1.8, p=.53).Conclusion: Despite improved outcome in young adult ALL patients with pediatrics-inspired protocols, CNS involvement remains a poor-risk feature. The historical use of allogeneic SCT does not improve outcome. Specific regimens should be developed for adult ALL patients with CNS involvement. [Display omitted] DisclosuresHuguet: Amgen: Other: Advisor; BMS: Other: Advisor; Celgene: Other: Advisor; Jazz Pharmaceuticals: Other: Advisor; Novartis: Other: Advisor; Pfizer: Other: Advisor. Barbieux: ASTRA-ZENECCA: Consultancy. Vey: Amgen: Honoraria; BMS: Honoraria; BIOKINESIS: Consultancy, Research Funding; NOVARTIS: Consultancy, Honoraria, Research Funding; SERVIER: Consultancy; JAZZ PHARMACEUTICALS: Honoraria; JANSSEN: Consultancy. Dombret: Amgen: Honoraria, Research Funding; Incyte: Honoraria, Research Funding; Jazz Pharmaceuticals: Honoraria, Research Funding; Novartis: Research Funding; Pfizer: Honoraria, Research Funding; Servier: Research Funding; Abbvie: Honoraria; BMS-Celgene: Honoraria; Daiichi Sankyo: Honoraria. Boissel: Bristol-Myers Squibb: Honoraria, Research Funding; Servier: Consultancy, Honoraria; Incyte: Honoraria; Amgen: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Research Funding; JAZZ Pharma: Honoraria, Research Funding; CELGENE: Honoraria; SANOFI: Honoraria; PFIZER: Consultancy, Honoraria. Mathilde: ABBVIE: Consultancy; SERVIER: Consultancy.

Impact of Nelarabine, Intensive L-Asparaginase, and Protracted Intrathecal Therapy on Newly Diagnosed T-Cell Acute Lymphoblastic Leukemia: Results from the Japanese Pediatric Leukemia/Lymphoma Study Group and the Japan Adult Leukemia Study Group

Impact of Nelarabine, Intensive L-Asparaginase, and Protracted Intrathecal Therapy on Newly Diagnosed T-Cell Acute Lymphoblastic Leukemia: Results from the Japanese Pediatric Leukemia/Lymphoma Study Group and the Japan Adult Leukemia Study Group

Socioeconomic Outcomes Among Long-Term Childhood Acute Lymphoblastic Leukemia Survivors Enrolled between 1971 and 1998 in the EORTC Children Leukemia Group Trials 58741, 58831/2 and 58881

BackgroundOver the last 30 years, there has been a considerable progress in the survival of childhood acute lymphoblastic leukemia (ALL). With survival rates exceeding 85%, the long-term effects are of major interest. Exposure to treatments may have severe toxicities such as cognitive deterioration after cranial radiotherapy (CRT) and impaired general physical functioning after hematopoietic stem cell transplantation (HSCT). These late effects may negatively impact the socioeconomic functioning of survivors. The aim of this study was to assess the socioeconomic status of adult survivors of childhood ALL treated in EORTC trials and the effect of patient's and disease characteristics and treatment modalities.MethodsBetween 2012 and 2017, a cross-sectional follow-up study was performed among childhood ALL (and acute lymphoblastic lymphoma) survivors enrolled between 1971 and 1998 in EORTC trials (58741, 58831/2 and 58881) by centers from France and Belgium. Patients who were younger than 18 years at diagnosis and older than 18 years at the time of the follow-up were eligible. Those whose postal address could be retrieved were invited to fill out a questionnaire with information about their socioeconomic situation. Additionally, the treating institutions provided medical updates.The following socioeconomic outcomes were evaluated: being married or living with a partner, having a university degree, having a job, working part-time among those having a job and history of having a paid job.ResultsAmong 1429 eligible patients, 502 (35.1%) filled out the questionnaire, 862 (60.3%) were lost to follow-up and 65 (4.5%) refused to participate. The distribution of age and NCI risk group at diagnosis, initial central nervous system involvement and relapse within one year after diagnosis was comparable between respondents, those lost to follow-up and those who refused to participate.Out of 502 survivors participating in the study, 66 had experienced a relapse (13%), 33 (7%) received HSCT (26 after relapse) and 60 (12%) were treated with CRT (36 in first line treatment, Table 1). The median time between the diagnosis and the socioeconomic evaluation was 20 years (range: 13-42). The median age at follow-up was 25 years (range: 18-53). Among survivors at school at the time of the diagnosis, 22% retook one and 6% two or more years due to the disease and 4% were unable to finish their education. Interestingly, 21% of the survivors had more and only 10% had less educational ambitions due to the disease. More than a half of the survivors had at least a bachelor (undergraduate) degree, 64% had a job and 38% were married or lived with their partner. Among those 25 years of age or older, the percentages for obtaining a bachelor degree, having a job and having a partner were 54%, 84% and 57%, respectively.In a multivariate model including sex, age at diagnosis, country, age at follow-up, relapse, HCST and CRT both treatments were associated with a higher probability of not obtaining a bachelor degree (OR=3.25, 95% CI: 1.27-8.31 for HCST and OR=3.17, 95% CI: 1.35-7.44 for CRT) and HSCT was associated with not having a job (OR=3.55, 95% CI: 1.23-10.21, Table 2). We found no evidence of associations between treatment exposures and having a partner. The associations between age at diagnosis and risk group at diagnosis and the socioeconomic outcomes were not statistically significant. Having a relapse was associated with a higher probability of not having a partner (OR=1.99, 95% CI: 1.05-3.77) and working half-time (OR=2.40, 95% CI: 1.03-5.58), adjusting for sex, age at diagnosis, country and age at follow-up. These findings were consistent with those from a sensitivity analysis performed among survivors aged 25 years or older.ConclusionThis large study showed that, overall, long-term childhood ALL survivors enrolled in EORTC trials between 1971 and 1998 were characterized by a high level of socioeconomic participation. However, HSCT and CRT were associated with negative socioeconomic outcomes. Our study adds to the body of evidence supporting the omission of CRT when other treatment alternatives are sufficient, like in first line treatment. It also indicates the need to provide special support to the most vulnerable groups of childhood ALL survivors, in particular those who received HSCT. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.

Impact of extracranial disease status on survival after initial central nervous system (CNS) involvement and radiation therapy in HER2+ breast cancer brain metastases (BCBM).

1041 Background: BCBMs are very common in metastatic HER2+ breast cancer. CNS-directed local therapy is the gold standard for treatment, followed by systemic HER2-targeted therapies. In patients with HER2+ BCBM and stable extracranial disease (ECD), consensus guidelines recommend continuing current systemic therapy after local therapy. Our goal was to determine the implications of ECD status at time of HER2+ BCBM first CNS involvement on outcomes including intracranial progression-free survival (PFS1) and overall survival (OS). Methods: Retrospective analysis was performed on data extracted from 77 patients with HER2+ BCBM who received CNS radiation at Duke between 2006 and 2020 following initial documentation of CNS involvement. Demographics, dates of metastatic and intracranial diagnosis, ECD status at first CNS involvement, systemic therapy, and outcomes were collected. The primary endpoint was PFS1 defined as the time from first CNS radiation to the subsequent documentation of intracranial progression (RANO-BM). OS was defined as time from first CNS radiation and first metastatic disease to date of death or last known alive. ECD status was defined by RECIST1.1 from systemic staging scans within 30 days of first CNS involvement. Results: In this patient cohort of HER2+ BCBMs undergoing CNS radiation at first CNS involvement, &gt;50% of patients had extracranial disease control: no ECD (25%) or stable/responding disease (31%). 52% of patients’ tumors were ER+. Median age was 50 years (range 27 – 75). Most patients (58%) developed first CNS involvement during adjuvant or first/second line metastatic therapy. For first CNS radiation, 49% received SRS and 48% WBRT. All patients with no ECD presented with isolated CNS disease as first metastatic presentation. Median OS in this cohort from initial metastatic disease to death was markedly worse for patients with no ECD (25.3m, 95% CI: 16.8 to 35.3) compared to those with progressive or stable/responding ECD (48.8m, 95% CI: 28.1 to 65; and 52.9 months, 95% CI: 43.7 to 73.3, respectively; p=0.03). Median OS from first CNS involvement to death was not statistically different amongst groups. This analysis did not detect median PFS1 differences based on ECD after first CNS radiation: progressive ECD (6.3m), no ECD (8.7m), or stable/responding ECD (10.6m) (p=0.13), though clinically meaningful differences were observed. Conclusions: Patients with isolated HER2+ BCBM with no ECD at the time of their initial CNS involvement (25% of population) have substantially worse OS compared to patients who present with ECD and develop CNS metastases later in their disease course. This population with isolated CNS disease at metastatic presentation deserves investigation of novel treatment algorithms, including earlier introduction of brain penetrable HER2-targeted agents.

Differing Outcomes of Patients with High Hyperdiploidy and ETV6-RUNX1 Rearrangement in Korean Pediatric Precursor B Cell Acute Lymphoblastic Leukemia.

Purpose:Recent cooperative trials in pediatric acute lymphoblastic leukemia (ALL) report long-term event-free survival (EFS) of greater than 80%. In this study, we analyzed the outcome and prognostic factors for patients with precursor B cell ALL (n=405) diagnosed during a 10-year period (2005–2015) at our institution.Materials and MethodsAll patients were treated with a uniform institutional regimen based on four risk groups, except for steroid type; patients diagnosed up till 2008 receiving dexamethasone, while subsequent patients received prednisolone. None of the patients received cranial irradiation in first complete remission.ResultsThe 10-year EFS and overall survival was 76.3%±2.3% and 85.1%±1.9%. Ten-year cumulative incidence of relapse, any central nervous system (CNS) relapse and isolated CNS relapse was 20.8%±2.2%, 3.7%±1.1%, and 2.5%±0.9%, respectively. A comparison of established, good prognosis genetic abnormalities showed that patients with high hyperdiploidy had significantly better EFS than those with ETV6-RUNX1 rearrangement (10-year EFS of 91.2%±3.0% vs. 79.5%±4.4%, p=0.033). For the overall cohort, male sex, infant ALL, initial CNS involvement, and Philadelphia chromosome (+) ALL were significant factors for lower EFS in multivariate study, while high hyperdiploidy conferred favorable outcome. For high and very high risk patients (n=231), high hyperdiploidy was the only significant factor for EFS in multivariate study.ConclusionRegarding good prognosis genetic abnormalities, patients with high hyperdiploidy had significantly better outcome than ETV6-RUNX1(+) patients. High hyperdiploidy was a major, favorable prognostic factor in the overall patient group, as well as the subgroup of patients with higher risk.

CD79a Is Associated with Central Nervous System Infiltration of Pediatric B-Cell Precursor Acute Lymphoblastic Leukemia

▪Despite the advances in the treatment of pediatric B-cell precursor acute lymphoblastic leukemia (BCP-ALL), infiltration of the central nervous system (CNS) remains a clinical challenge. Certain cytogenetic subtypes such as E2A-PBX1-and BCR-ABL-positive BCP-ALL confer a higher risk for CNS involvement initially and for CNS relapse. Novel strategies to predict CNS and to eradicate leukemic cells from the CNS are subjects of ongoing research.In order to identify targets with diagnostic and therapeutic relevance, comparative RNA-sequencing was performed with patient derived xenograft (PDX) blasts from 5 E2A-PBX1-positive patients, recovered from the bone marrow (BM) and from the CNS of NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ (NSG) mice. Differential gene expression analysis revealed the upregulation of various genes of the pre-B cell receptor complex, particularly the signaling component CD79a (Igα) in blasts recovered from the CNS as compared to blasts from the BM. We then investigated the impact of CD79a on CNS infiltration in vivo and in patients. CD79a was downregulated by short-hairpin RNA (shRNA) mediated knockdown in the E2A-PBX1 positive cell line 697. Proliferation rates of 697-shCD79a cells and control-transfected 697 cells (697-shCtrl) in vitro were similar. Furthermore, NSG mice injected with 697-shCD79a cells showed comparable survival times, as well as similar blast infiltration in spleen and BM as animals injected with 697-shCtrl cells. However, downregulation of CD79a led to a significantly lower number of CNS-positive mice (4/15, 26%) as compared to control animals (7/10, 70%) (p=0.0486, Figure A). This indicates that CD79a is not critically involved in proliferation and peripheral engraftment, but in CNS infiltration of E2A-PBX1 positive 697 cells in vivo.To test if CD79a also affects CNS involvement in BCR-ABL-positive leukemia, a murine/murine transplantation model was used. B-cells isolated from CD79a-knockout (CD79a-KO) or wildtype mice (CD79a-Ctrl) were stably transfected with a BCR-ABL fusion gene and cultured independent of cytokines, thereby inducing malignant transformation. Both cell lines were subsequently injected into recipient NSG mice (n=8/group) and leukemic development was followed. The experiment was terminated when all control mice had developed leukemic symptoms and mice were analyzed for leukemic engraftment. A further CD79a-KO group was included for survival analysis. Median spleen volume as a surrogate of leukemic infiltration was significantly lower in mice injected with CD79a-KO as compared to CD79a-Ctrl cells (0.35 cm³ vs. 0.06 cm³; p=0.0001). Median blast percentages in spleens and BM were also markedly reduced (75.3% vs. 5.8%; p=0.0001 and 61.0% vs. 4.5%; p=0.0001, respectively). Importantly, none of the animals in the CD79a-KO group showed blasts in the CNS as assessed by histology whereas blasts were present in all of the animals in the CD79a-Ctrl group. Finally and most importantly, NSG-mice injected with CD79a-KO cells showed a highly significant prolongation in median survival as compared to mice with CD79a-Ctrl cells (29 days vs. 95 days; p=0.0001, Figure B). Altogether, these data suggest that in a model of BCR-ABL-positive leukemia, absence of CD79a impacts the engraftment of blasts in vivo, in the CNS and other leukemic niches.To further validate our findings in patient material, we measured CD79a protein expression in PDX cells from an E2A-PBX1- and a BCR-ABL-positive patient serially transplanted into NSG mice for three passages. For both entities and in all passages, CNS blasts showed a higher CD79a expression than blasts isolated from the bone marrow. In order to assess if CD79a can be used as a marker to predict CNS involvement in patients, CD79a mRNA levels were measured in a selected cohort of 98 pediatric BCP-ALL patients, which contained 26 CNS-positive patients matched to 72 CNS-negative patients. CNS-positive patients showed significantly higher mRNA levels of CD79a than CNS-negative patients (p=0.0225, unpaired t-test, Figure C) suggesting that CD79a may be of value as a potential diagnostic marker for initial CNS involvement in BCP-ALL.Our results indicate a role of CD79a in proliferation and CNS infiltration of BCP-ALL blasts in experimental settings and patients. We intend to prospectively evaluate CD79a as a prognostic marker, which may also be a therapeutic target in CNS-positive BCP-ALL. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.

Characteristics and outcome in patients with central nervous system involvement treated in European pediatric acute myeloid leukemia study groups.

There is no consensus on the treatment for pediatric patients with acute myeloid leukemia and initial central nervous system (CNS) involvement. To evaluate different CNS-directed treatment options (intrathecal [IT] therapy, CNS irradiation, hematopoietic stem cell transplantation [HSCT]), 261 patients (excluding acute promyelocytic leukemia) with initial CNS involvement treated in trials with similar intensive chemotherapy by four cooperative European study groups (1998-2013) were studied and compared with CNS-negative patients from the Berlin-Frankfurt-Münster group. Patient characteristics in the different study groups were comparable. Young age, high white blood cell count, extramedullary involvement other than the CNS, monoblastic morphology, and inv(16) were associated with CNS involvement (each P < 0.0001). There were no major differences in outcome between the study groups. The cumulative incidence of relapse (CIR) regarding the CNS was higher in initially CNS-positive versus initially CNS-negative patients (all: 8 ± 2% vs. 3 ± 1%, P(Gray) = 0.001; isolated: 4 ± 1% vs. 1 ± 0%, P(Gray) = 0.03). However, global outcome of the CNS-positive cohort (overall survival, 64 ± 3%; event-free survival 48 ± 3%; and CIR 33% ± 3%) did not differ significantly from CNS-negative patients. Risk groups defined by cytogenetics were of likewise prognostic significance in CNS-positive and -negative patients. CNS treatment with cranial irradiation was not superior compared to IT therapy and systemic chemotherapy (± HSCT). Although CNS relapses occurred more frequently in initially CNS-positive patients, their global outcome was similar as in CNS-negative patients. Intensified IT therapy was heterogeneous; however, at least eight applications, preferably with triple IT chemotherapy, seem to be appropriate to accompany dose-intensive systemic chemotherapy.

Acute myeloid leukemia in children: the study of clinical and biological features and results of the treatment at the Department of pediatrics, Clinical Hospital Centre Rijeka

Objective: To investigate clinical and biological (cytomorphology, immunophenotyping, cytogenetics) features and the survival rate of pediatric patients with acute myeloid leukemia (AML) in a tertiary centre. Material and Methods: Epidemiological, retrospective, descriptive study involving patients with newly diagnosed AML younger than 18 years, from January 1st 2002 to December 31st 2013. Results: In a 12-year period, 9 children with AML were hospitalized, 5 (55.6%) girls and 4 (44.4%) boys. The median age was 11 (± 3.8 SD) years. The most common symptoms were asthenia (66.7%), persistent fever (55.6%), pallor (44.4%) and bleeding (44.4%). Lymphadenopathy and hepatomegaly were present in 3 patients (33.3%), and splenomegaly in one (11.1%) patient. One (11.1%) patient had initial central nervous system involvement. Thrombocytopenia was present in 8 (88.9%) patients, anemia in 7 (77.8%) patients, leukocytosis in 4 (44.5%) patients, and leukopenia in 3 (33.3%) patients. All patients had blasts in the peripheral blood smear. Seven cytomorphologic subtypes with characteristic immunophenotypic features were found. Seven (77.8%) patients had numeric and structural chromosome aberrations. Three (33.3%) patients were classified in standard-risk group and 6 (66.7%) in high-risk group. Remission was achieved in 6 (66.7%) patients on day 15 of chemotherapy. Allogeneic hematopoietic stem cell transplantation was performed in 3 (33.3%) patients. Eight (88.9%) patients are in complete remission with a median follow-up of 10.1 years. The death occurred in one (11.1%) patient due to septic shock. Conclusions: Pediatric AML is a heterogeneous hematologic neoplasm. Aggressive chemotherapy and hematopoietic stem cell transplantation provide a high cure rate.

NKX2‐5, SIL/TAL and TLX3/HOX11L2 expression in Egyptian pediatric T‐cell acute lymphoblastic leukemia

Cohorts of T-cell acute lymphoblastic leukemia (T-ALL) patients show regional geographic differences in incidence, biological features and clinical outcome, implying that in different populations, cases may harbor different genetic lesions than those reported elsewhere. In this study, we prospectively evaluated the frequency and the clinical relevance of NKX2-5, TLX3/HOX11L2 and SIL/TAL expression in Egyptian childhood T-ALL. NKX2-5, TLX3/HOX11L2 and SIL/TAL expression were tested in peripheral blood and/or bone marrow of 83 newly diagnosed Egyptian childhood T-ALL patients. NKX2-5 expression was detected in 11/83 cases (13%), TLX3/HOX11L2 (5/83, 6%) and SIL/TAL (4/83, 5%). Initial central nervous system involvement was significantly higher in the NKX2-5 positive versus negative patients (P = 0.009). The follow-up period was a median of 65.5 months. The 5-year leukemia-free and event-free survival rates of the whole T-ALL population were 70 ± 6% and 58 ± 6%, respectively. The 5-year leukemia-free and event-free survival rates of NKX2-5 were 86 ± 13% and 60 ± 16%, respectively. There were no statistically significant differences in clinical presentation, biological features, initial response to chemotherapy, or subsequent treatments between the subgroups and the total population. Egyptian T-ALL cases seemed to have a different genetic pattern compared to other populations, with a lower incidence of TLX3/HOX11L2 and SIL/TAL but a higher incidence of NKX2-5 expression than recorded in Western countries.

Predictive factors for oligometastatic versus non-oligometastatic involvement of the central nervous system by brain metastases from breast cancer.

123 Background: To assess rates of oligometastatic (1-3 metastases) versus non-oligometastatic (4 or more) involvement of the central nervous system (CNS) from breast cancer and determine whether rates vary for different breast cancer subtypes. Methods: We reviewed records of 65 patients with brain metastases from breast cancer. Fisher’s Exact Test compared incidence of oligometastatic versus nonoligometastatic CNS involvement and presence of visceral metastatic disease between groups. Paired Student's t test compared mean age and interval from initial diagnosis to development of brain metastases. Results: 25 patients (38.5%) developed oligometastatic and 40 (61.5%) developed non-oligometastatic CNS disease. Subtypes predictive of oligometastatic disease included ER+ (p .04) and PR+ (p .008). Subtypes predictive of non-oligometastatic disease included HER2+ (p .05). HER2+ patients trended toward a lower incidence of visceral metastases compared to ER+ or PR+ patients (p .07). Non-oligometastatic and HER2+ patients developed brain metastases sooner after initial diagnosis (p .046 and .03, respectively). Conclusions: ER+ breast cancer patients are more likely to develop oligometastatic CNS involvement and may be excellent candidates for stereotactic radiosurgery (SRS) alone. HER+ patients are likely to develop non-oligometastatic CNS involvement and may be better treated initially with whole brain radiotherapy (WBRT) even with apparent initial oligometastatic CNS involvement.

Long‐term results of AIEOP‐8805 protocol for acute B‐cell lymphoblastic leukemia of childhood

Acute B-cell leukemia (B-ALL) is a rare form of pediatric leukemia characterized by a very high-proliferation index, rapid clinical progression, and a high frequency of central nervous system (CNS) involvement. Commonly, it is treated in the clinical trials for Burkitt lymphoma, of which it represents the leukemic counterpart. Children with B-ALL diagnosed between 1988 and 1999 were enrolled in the AIEOP-8805 protocol. Treatment included six high-dose chemotherapy courses. No prophylactic CNS irradiation was administered. Sixty-five consecutive patients were enrolled in the study. L3 morphology was observed in 57 of 65 patients (88%). Twenty-five children (38%) had tumor mass in addition to massive bone marrow infiltration; 11 children (17%) had CNS disease at diagnosis. Sixty-two patients obtained complete morphological remission of which 13 suffered a relapse, including 3 with initial CNS involvement. Ten-year overall survival and event-free survival were 77% and 75%, respectively. Neither relevant long-term toxicity nor second malignancies were observed. The AIEOP-8805 confirmed that short high-dose chemotherapy is highly effective for the treatment of B-ALL without significant long-term adverse sequelae. Therapy modifications to reduce relapse rate, such as the use of anti-CD20 monoclonal antibody and more effective CNS treatment, are being tested.

Prognostic significance of the initial cerebro-spinal fluid (CSF) involvement of children with acute lymphoblastic leukaemia (ALL) treated without cranial irradiation: Results of European Organization for Research and Treatment of Cancer (EORTC) Children Leukemia Group study 58881

Aim of the study To evaluate the prognostic significance of the initial cerebro-spinal fluid (CSF) involvement of children with ALL enrolled from 1989 to 1996 in the EORTC 58881 trial. Patients and methods Patients (2025) were categorised according to initial central nervous system (CNS) status: CNS-1 (CNS negative, n = 1866), CNS-2 (<5 leucocytes/mm 3, CSF with blasts, n = 50), CNS-3 (CNS positive, n = 49), TLP+ (TLP with blasts, n = 60). CNS-directed therapy consisted in intravenous (i.v.) methotrexate (5 g/sqm) in 4–10 courses, and intrathecal methotrexate injections (10–20), according to CNS status. Cranial irradiation was omitted in all patients. Results In the CNS1, TLP+, CNS2 and CNS3 group the 8-year EFS rate (SE%) was 69.7% (1.1%), 68.8% (6.2%), 71.3% (6.5%) and 68.3% (6.2%), respectively. The 8-year incidence of isolated CNS relapse (SE%) was 3.4% (0.4%), 1.7% (1.7%), 6.1% (3.5%) and 9.4% (4.5%), respectively, whereas the 8-year isolated or combined CNS relapse incidence was 7.6% (0.6%), 3.5% (2.4%), 10.2% (4.4%) and 11.7% (5.0%), respectively. Patients with CSF blasts had a higher rate of initial bad risk features. Multivariate analysis indicated that presence of blasts in the CSF had no prognostic value: (i) for EFS and OS; (ii) for isolated and isolated or combined CNS relapse; WBC count < 25 × 10 9/L and Medac E-coli asparaginase treatment were each related to a lower CNS relapse risk. Conclusions The presence of initial CNS involvement has no prognostic significance in EORTC 58881. Intensification of CNS-directed chemotherapy, without CNS radiation, is an effective treatment of initial meningeal leukaemic involvement.

Clinical Features and Prognostic Relevance of Ovarian Involvement in Non-Hodgkin's Lymphoma: a Consortium for Improving Survival of Lymphoma (CISL) Report.

Abstract 4999 IntroductionOvary can be involved as a primary ovarian lymphoma or secondarily involved by disseminated disease of non-Hodgkin's lymphoma. However, ovarian involvement is an extremely rare event in non-Hodgkin's lymphoma. Thus, it clinical features and prognostic relevance has rarely been addressed, and most publications refer to a single or a few cases. Thus, we retrospectively analyzed patients with ovarian involvement Patients and methods32 patients with ovarian involvement were assembled from 8 hospitals affiliated with the CISL (Consortium for Improving Survival of Lymphoma), a Korean lymphoma study group. Primary ovarian involvement was defined as a lymphoma confined to ovary with or without involvement of adjacent lymph nodes and contiguous organs. Secondary ovarian lymphoma was defined as a secondary involvement of ovary in disseminated disease of lymphoma at initial diagnosis. ResultsTwelve patients had primary ovarian lymphoma (37.5%) while twenty patients (62.5%) had secondary ovarian involvement by systemic disease. The clinical manifestations of ovarian involvement were similar to that of other ovarian tumors, namely an abdominal pain (31%), abdominal distension (19%) or lower abdominal palpable mass (16%). Pathological review according to the WHO classification showed that the most common histological subtype was diffuse large B-cell lymphoma (DLBCL, 75.0%, 24/32), and the frequency of other subtypes was as follows: Burkitt lymphoma (BL, 12.5%, 4/32), lymphoblastic lymphoma (6.3%, 2/32), marginal zone B-cell lymphoma (MZL, 3.1%, 1/32), peripheral T-cell lymphoma, unspecified (PTCL-U, 3.1%, 1/32). The median age (43 years, range 18-80) was younger than that of previously reported other organs such as uterus or prostate. The presence of B symptoms was only observed in 31.3%, and the performance status was good (84.4% of patients had less than grade II of ECOG performance status). The cases involving two or more than two extranodal sites were 68.8% while cases with elevated level of serum LDH were 59.4%. Thus, 59.4% of patients had the low or low-intermediate score of IPI score. Bilateral ovarian involvement was found in 12/32 (38%) while unilateral involvement was 20/32 (63%, 9 right and 11 left side. Three patients showed the involvement of central nervous system (CNS) at diagnosis (3/32, 9.4%). These three patients had DLBCL histology and unfavorable parameters including stage IV, high IPI score and bone marrow BM involvement. Thus, the initial CNS involvement might be associated with advanced stage of lymphoma not with ovarian involvement itself. Surgical removal of involved ovary was performed in 20 patients (62%), and then they were treated with systemic chemotherapy. Twelve patients (38%) were treated with chemotherapy alone. The comparison of outcomes according to the treatment modalities showed the outcomes of chemotherapy-based treatment versus surgery-based treatment were not significantly different (2 year overall survival; 66% vs. 68%). With a median follow-up of 25 months (range 3-185), 13 patients (40.6%) relapsed. Two patients were relapsed in single lesion and 11 were relapsed in multiple lesions. The majority relapsed at various extranodal sites (11/13, 84.6%) and only 2 cases relapsed at nodal sites. Most common relapse site was CNS (4 cases among 13 cases of relapse, 31%). All CNS relapsed patients had DLBCL histology. Ovarian relapse observed in one case that had been involved both ovary at the time of diagnosis. The 2 year overall survivals (OS) were 67% (95% CI: 50 to 83%) and the 2 year progression free survivals (PFS) were 61% (95% CI: 44 to 78%). In univariate analysis, high IPI score, 2 or more extranodal sites involvement and elevated LDH level were statistically significant parameters for lower PFS; moreover, 2 or more extranodal sites involvement and elevated LDH level associated with poor OS. ConclusionOvarian involvement of non-Hodgkin's lymphoma showed a dismal prognosis despite active treatment. Therefore, more optimal treatment strategy should be warranted. DisclosuresNo relevant conflicts of interest to declare.

Prognostic Significance of Central Nervous System (CNS) Status of Children with Acute Lymphoblastic Leukemia (ALL) Treated without Cranial Irradiation: Results of European Organization for Research and Treatment of Cancer (EORTC) Children Leukemia Group Study 58951

Background: Cranial irradiation (XRT) in combination with intrathecal chemotherapy has markedly reduced the incidence of meningeal relapse in childhood acute lymphoblastic leukemia (ALL), and is still recommended for patients (pts) at high risk of relapse within the CNS. Because of late adverse effects of XRT, the previous CLG trial (EORTC 58881, 1989–1996) deleted XRT from the treatment regimen of all ALL patients. Nevertheless, the results (5-year event-free survival (EFS) rates) of 58881 trial were good in CNS-1 (72.1%), dubious CNS-2 (62.2%), surreptitious CNS-2 (64.6%) and particularly good in pts with CNS leukemia involvement (70.3%, SE 6.2%), emphasizing the importance of systemic chemotherapy in such pts, of whom, 50% were treated according to a very high risk group protocol (ASH, 2006). In order to validate this therapeutic strategy, we evaluate the prognostic significance of CNS status at diagnosis in ALL children enrolled from 12/1998 to 2/2008 in the EORTC CLG 58951 randomized phase III trial.Methods: Treatment design was according to BFM. Three randomizations were programmed:R1) the value of Dexamethasone (DEX) vs Prednisolone (PRED) during induction (all pts);R2) the value of prolonged (24 injections) vs conventional (12 injections) duration of L-asparaginase courses during consolidation and late intensification (all except very high risk pts);R3) the value of 6 (DEX or PRED + vincristine) pulses every 10 weeks in maintenance therapy (average risk pts).CNS-directed therapy consisted in i.v. methotrexate (MTX) (5 g/sqm over 24 hours) in 4 to 10 courses, according to the degree of initial CNS involvement, and intrathecal MTX. No radiotherapy was used. According to CNS status, pts were classified in 4 groups:CNS-1: no blasts and WBC<6/μl, RBC<100/μl;dubious CNS-2: presence of blasts, RBC □100/μl;surreptitious CNS-2: presence of blasts, WBC<6/μl, RBC<100/μl;CNS-3: presence of blasts, WBC>5/μl, RBC<100/μl.Only CNS-3 pts were to receive 10 courses of i.v. MTX, but some of dubious (n=21) and surreptitious CNS-2 pts (n=38) did eventually receive 10 courses as well.Results: Between 12/1998 and 2/2008, a total of 1853 ALL pts were registered in 58951 trial. On 2/2008 a total of 1686 were evaluable for EFS; the distribution according to initial CNS status was: CNS-1 (n=1494, 88.6%), dubious CNS- 2 (n=100, 5.9%), surreptitious CNS-2 (n=59, 3.5%) and CNS-3 (n=33, 2%). The higher the degree of CNS involvement, the higher the incidence of unfavourable features: initial WBC□100000/μl, T-lineage, NCI high risk; very high risk (VHR) initial features (peripheral blasts ≥1000/μl post prephase, high-risk cytogenetics). In CNS-3 group, 10/33 (30%) pts were treated according to VHR protocol. Median follow-up was 4 years.CNS-1Dubious CNS-2Surrept. CNS-2CNS-3No CR17 (1.1%)0 (0%)0 (0%)1 (3.0%)CCR1286 (86.1%)84 (84.0%)50 (84.7%)22 (66.7%)Isolated CNS rel.17 (1.1%)3 (3.0%)1 (1.7%)2 (6.1%)Combined CNS r.23 (1.5%)1 (1.0%)2 (3.4%)2 (6.1%)Non-CNS relapse20 (13.4%)2 (2.0%)1 (1.7%)5 (15.2%)Death in CR31 (2.1%)2 (2.0%)1 (1.7%)1 (3.0%)4-yr EFS (%, SE)84.0% (1.1%)88.2% (3.5%)82.4% (5.4%)55.1% (10.8%)4-yr OS (%, SE)90.7% (0.9%)92.0% (2.9%)88.0% (4.6%)63.4% (12.8%)Conclusion: In comparison with previous EORTC 58881 trial, and according to CNS status, EFS, OS, and isolated and combined CNS relapses rates improved in 58951 trial, except for CNS-3 pts. These “disappointing” results in CNS-3 pts (representing 2% of the entire population), less heavily treated in 58951 than in 58881, confirm the importance of systemic chemotherapy in such pts.

Central Nervous System Involvement in Adult Acute Lymphoblastic Leukemia (ALL) at Diagnosis and/or at First Relapse: Results from the GET-LALA Group.

Outcome of adult ALL with central nervous system (CNS) involvement is not clearly defined. We studied 104 patients presenting with CNS involvement at diagnosis among 1493 patients (7%) included into the LALA-87 or LALA-94 trials, and 109 patients (9% of first remitters) presenting CNS disease at the time of first relapse among the 709 relapsing patients (15%) included initially in these trials. Treatment of patients presenting CNS involvement at diagnosis consisted in initial chemotherapy completed by 18 double or triple intrathecal injections associated with 15 to 20 Gy cranial irradiation, followed when possible by intensification by allogeneic or autologous stem cell transplantation (SCT). At diagnosis, 43 patients (41%) presenting with CNS involvement had T-lineage ALL, 53 (51%) had B-lineage ALL (of whom 9 were diagnosed as Philadelphia (Ph) chromosome positive ALL), 8 had undifferentiated ALL or unknown immunophenotype. Eighty-seven of 104 (84%) patients with CNS involvement at diagnosis achieved complete remission (CR). Fifty-three patients underwent SCT (25 allogeneic SCT from matched related or unrelated donor, 28 autologous SCT). Overall survival at 7 years was 34% in those with CNS involvement at diagnosis versus 29% (p = NS) for those without. DFS at 7 years was 35% versus 28% (p = NS). There were no significant differences between patients with CNS involvement and those without CNS involvement regarding T lineage ALL, B lineage ALL (including or not Ph ALL). There were also no significant differences regarding patients who underwent transplantation as consolidation intensification, while in patients receiving only chemotherapy patients without initial CNS involvement had a better outcome (p = 0.01). Among the 709 patients with primary relapse, 66 patients (61%) presented a CNS relapse combined with bone marrow relapse, whereas 17 relapses (15%) and 26 relapses (24%) were CNS relapses combined with another extramedullary relapse or isolated CNS relapses respectively. Median time to relapse was 6.7 months (range, 1–62) in patients with CNS relapse versus 11.2 months (1.7–111) in relapsing patients without CNS involvement. Eleven patients (10%) with CNS relapse had CNS involvement at diagnosis, while 98 patients were diagnosed with CNS disease only at the time of first relapse. Overall, 38 out of 109 patients with CNS relapse (35%) achieved CR. The median OS was 6.3 months. Outcome was similar in terms of CR proportion and OS in relapsing patients without CNS involvement. The 2-year OS rates did not show any difference among patients with CNS relapse who had CNS involvement at diagnosis and those with CNS disease only diagnosed at the time of first relapse.Overall, CNS leukemia in adult ALL is uncommon at diagnosis. Patients have a similar outcome than those who did not present with CNS involvement. However, patients benefit from intensification therapy by autologous or allogeneic SCT. CNS leukemia at first relapse are also uncommon but probably underestimated. Outcome is particularly poor as this of all adult ALL in first relapse.

Prognostic Significance of Blasts with/without Pleiocytosis in the Cerebro-Spinal Fluid (CSF) of Children with Acute Lymphoblastic Leukemia (ALL) Treated without Cranial Irradiation: Results of European Organization for Research and Treatment of Cancer (EORTC) Children Leukemia Group Study 58881.

To evaluate the prognostic significance of blasts in the CSF at diagnosis in children with ALL, 2049 patients (pts) enrolled from 1989 to 1996 in EORTC 58881 trial were retrospectively studied. Treatment design was according to BFM. Central nervous system (CNS)-directed therapy consisted in i.v. methotrexate (MTX) (5 g/sqm over 4 hours) in 4 to 10 courses, according to grade of initial CNS involvement, and intrathecal MTX. No radiotherapy was used. Three randomizations were programmed: Erwinia vs Medac E.coli asparaginase (all pts); addition or not of i.v. Ara-C to i.v. MTX (for increased-risk pts); addition of monthly courses of i.v. 6-MP in maintenance therapy (all pts). According to CNS status, pts were classified in 4 groups: 1) CNS-1: <6 WBC/μl, RBC<100/μl, no blasts; 2) dubious CNS-2: presence of blasts, RBC>100/μl; 3) surreptitious CNS-2: presence of blasts, <6 WBC/μl, RBC<100/μl; 4) CNS-3: presence of blasts, >5 WBC/μl, RBC<100/μl. Only CNS-3 pts were to receive 10 courses of i.v. MTX, but some of dubious (N=21) and surreptitious CNS-2 pts (N=19) did eventually receive 10 courses as well. Dubious CNS-2 (n=53), surreptitious CNS-2 (n=52), and CNS-3 (n=54) contained a higher rate of pts with unfavourable features than CNS-1 pts: WBC > 100000/μl; T-lineage; NCI high risk; very high risk (VHR) features (≥1000 peripheral blasts/μl post prephase, high-risk cytogenetics). Median follow-up was 7.5 years. The 5-yr overall event-free survival (EFS) and overall survival (OS) rates (SE%) were 71.6 % (1.0 %) and 82.6 % (0.8%) respectively. The 5-yr EFS rate (SE%) was 72.1 % (1.0%) for CNS-1, 62.2 % (6.6%) for dubious CNS-2, 64.7 % (6.7%) for surreptitious CNS-2, and 70.3 % (6.2%) for CNS-3 group. Overall, pts with blasts in the CSF (dubious CNS-2, surreptitious CNS-2 or CNS-3) had a significantly (p=0.02) shorter EFS than those in the CNS-1 group: 5-yr EFS rate 65.6% (3.8%) vs 72.1%. Multivariate analysis indicated that low WBC, Medac E-Coli asparaginase, absence of VHR features, middle age group were, together, predictive for longer EFS, whereas CNS involvement (CNS-2/-3 vs CNS-1) lost its prognostic value (p=0.87). Out of 2018 pts who reached CR, a total of 71 isolated and 78 combined CNS relapses were reported. The 5-yr isolated CNS relapse rate was 3.8%: 3.5% in CNS-1, 6.7% in dubious CNS-2, 10.5% in surreptitious CNS-2 and 7.1% in CNS-3 group. The 5-yr isolated or combined CNS relapse rate was 7.9%; in the 4 CNS-groups it was 7.6%, 11.1%, 14.7% and 9.2% respectively. The 5-yr OS rate (SE%) was 83.5% (0.9%) in CNS-1 vs 72.4% (3.9%) in CNS-2/-3: p=0.0003. Prognostic importance was lost (p=0.23) in multivariate analysis. Conclusion: the presence of blasts in the CSF, with or without pleiocytosis, is associated with unfavorable prognostic features and with worse outcome. Intensification of CNS-directed chemotherapy, without CNS radiation, is an effective treatment of initial meningeal leukemic involvement.

Retrospective study of prognostic factors in non-Hodgkin lymphoma secondarily involving the central nervous system

The aim of this retrospective single-center study was to analyze the clinical characteristics and outcome of non-Hodgkin lymphoma (NHL) patients with central nervous system (CNS) involvement and to identify prognostic factors for survival. We searched our hospital records for NHL patients diagnosed with CNS involvement from 1982 to 2004, and 43 patients were identified. The median age was 63 years (range 23-88) and the median Karnofsky performance status was 55% (range 10-90). Treatment of CNS lymphoma included intrathecal chemotherapy in 33 patients (77%), systemic chemotherapy in 25 (58%), and radiotherapy in 16 (37%). Twenty-six patients showed a CNS response. The median survival after CNS manifestation was 5 months (range 2 days-82.5+months). Nine patients achieved long-term survival. Low lactate dehydrogenase (LDH) at CNS manifestation and a CNS response to therapy were favorable independent prognostic factors for survival in multivariate analysis (p = 0.051 and p < 0.0005, respectively), whereas a young age at initial diagnosis, initial CNS involvement, an initially normal LDH, and high-dose chemotherapy for CNS involvement were significant in univariate analysis. In conclusion, long-term survival can be achieved in patients with secondary CNS lymphoma. LDH at CNS manifestation and a CNS response to therapy were significantly associated with survival.

Central nervous system involvement in non-Hodgkin’s lymphoma: A single-center analysis of prognostic factors for survival

6718 Background: The prognosis of patients with non-Hodgkin’s lymphoma (NHL) and central nervous system (CNS) involvement is generally regarded as poor. Very little is known about prognostic factors in these patients. The aim of this retrospective single-center study was to analyze the clinical characteristics and outcome of NHL patients with CNS involvement and identify prognostic factors for survival. Methods: We searched our hospital records for NHL patients with CNS involvement diagnosed from 1982 to 2003, and 43 were identified. Prognostic factors for survival were assessed by univariate and multivariate analysis. Results: The median age at CNS manifestation was 61 (range, 22–87); the median Karnofsky performance status was 55 (range 10–90). Thirty-five patients had CNS relapse and 8 patients CNS involvement at first presentation. Isolated CNS disease was found in 6 patients and concomitant systemic disease in 36. Initial management included systemic chemotherapy (n=25), chemo-/radiotherapy (n=16), dexamethasone (n=1), and no treatment (n=1). Chemotherapy consisted of CNS-penetrating drugs (high-dose methotrexate or ifosfamide) in 24 patients. Intrathecal chemotherapy was given to 34 patients. Twenty-six patients showed a CNS response. The median survival from the time of CNS manifestation and the median overall survival (OAS) were 5.5 (0–82.5+) and 37 (0–181) months, respectively. Lactate dehydrogenase (LDH) at CNS manifestation (p = .051) and CNS response to therapy were identified as independent prognostic factors for survival on multivariate analysis (p = <.001), whereas age at initial diagnosis, initial CNS involvement, initial LDH, and treatment with CNS-penetrating chemotherapy for CNS involvement were only significant on univariate analysis. All but one of 9 patients who are still alive received substances able to sufficiently penetrate the blood-brain barrier. All survivors were disease-free at the last follow-up. Conclusions: Long-term survival can be achieved in patients with NHL and CNS involvement. LDH at CNS manifestation and CNS response to therapy were significantly associated with survival. No significant financial relationships to disclose.