Primary CNS lymphoma in HIV positive and negative patients: comparison of clinical characteristics, outcome and prognostic factors

Transcription

1 J Neurooncol (2011) 101: DOI /s CLINICAL STUDY PATIENT STUDY Primary CNS lymphoma in HIV positive and negative patients: comparison of clinical characteristics, outcome and prognostic factors Soley Bayraktar Ulas D. Bayraktar Juan C. Ramos Alexandra Stefanovic Izidore S. Lossos Received: 11 January 2010 / Accepted: 20 May 2010 / Published online: 6 June 2010 Ó Springer Science+Business Media, LLC Abstract Primary central nervous system lymphoma (PCNSL) accounts for approximately 4% of all primary brain tumors and has a poor prognosis in both immunocompetent as well as in immunocompromised patients. We conducted a retrospective analysis to examine the clinical characteristics and prognostic factors in HIV-negative and HIV-positive patients with PCNSL and to assess the effect of highly active antiretroviral therapy (HAART) therapy on the outcome of HIV-positive patients. Patients diagnosed with PCNSL between 1999 and 2008 at our institution were divided into two groups based on their HIV status. Their demographic and clinical characteristics were compared using the chi-square test. Kaplan Meier survival curves were constructed employing the univariate log-rank test. Multivariate analyses of survival were performed by Cox proportional hazards models incorporating the prognostic factors identified in the univariate log rank test. Forty-one HIV-positive patients and 45 HIV-negative patients were identified. HIV-positive patients were younger, more likely to present with seizures and elevated serum LDH levels. There were significant differences in This article is dedicated to Dr. William J. Harrington who devoted his life for treatment of lymphoma in HIV-positive patients. S. Bayraktar U. D. Bayraktar J. C. Ramos A. Stefanovic I. S. Lossos (&) Division of Hematology and Oncology, Department of Medicine, University of Miami, Sylvester Comprehensive Cancer Center, 1475 NW 12th Ave (D8-4), Miami, FL 33136, USA Ilossos@med.miami.edu I. S. Lossos Department of Molecular and Cellular Pharmacology, Miller School of Medicine, University of Miami, Miami, FL 33136, USA complete remission (CR) rates (P = 0.010) and overall survival (OS) (P = 0.034) in favor of the HIV-negative group. In the HIV-positive group, OS was better in patients with KPS [ 70 and patients who received HAART, but remained inferior to that in the HIV-negative patients. HIV-positive patients had a worse prognosis compared to HIV-negative patients despite similar clinical characteristics. Better performance status (KPS [ 70) and treatment with HAART conferred better OS in HIV-positive patients. Keywords Primary CNS lymphoma HIV Prognostic factors EBV AZT Introduction Primary central nervous system lymphoma (PCNSL) accounts for approximately 4% of all primary brain tumors and 1% of all cases of non-hodgkin lymphoma (NHL), with an incidence of 0.43/100,000 persons per year [1]. The ageadjusted incidence of PCNSL in the U.S. has increased substantially since the 1970s, from a rate of 0.16 per 100,000 in to 0.48 per 100,000 in [2]. The observed increase in the incidence of PCNSL may be at least partially attributed to the emergence of acquired immune deficiency syndrome (AIDS). PCNSL is diagnosed in % of patients with AIDS [3, 4] and represents the second most common intracranial mass lesion in this population. As the susceptibility to PCNSL is inversely proportional to the CD4 count, since the introduction of highly active antiretroviral therapy (HAART), the incidence of PCNSL in the HIV infected population declined [5, 6]. The prognosis of PCNSL remains poor compared to other extranodal lymphomas, with a 5-year survival rate of 22 40% in highly selected patients enrolled in clinical

2 258 J Neurooncol (2011) 101: studies [6, 7]. In unselected 2,462 patients with PCNSL reported to the Surveillance, Epidemiology, and End Result (SEER) database between 1975 and 1999, the median survival was 9 months for immunocompetent patients, but it was only 4 months for the entire cohort that included patients with AIDS [7]. While the introduction of highly active antiretroviral therapy (HAART) therapy has led to marked improvement in the outcome of HIV-positive patients with systemic NHL, resulting in similar survival to HIV-negative NHL patients [8], it is not clear whether it confers a similar survival benefit in patients with PCNSL. The relative rarity of PCNSL precludes large-scale randomized trials and therefore, the optimal treatment for PCNSL in both HIV-negative and HIV-positive patients have not yet been determined. Several Phase II studies in HIV-negative patients have shown that chemotherapy in combination with whole brain radiotherapy (WBRT) as first-line treatment appears to improve both disease-free survival (DFS) and overall survival (OS) compared to radiation therapy alone [9 13]. Systemic chemotherapy without WBRT which was omitted to reduce the risk of neurological toxicity also demonstrated encouraging results in HIV-negative patients with PCNSL, leading to a median survival time of months [14, 15]. The most common approach has been to use agents such as methotrexate and cytarabine that cross the blood brain barrier and have activity in non-hodgkin s lymphoma [10, 15]. However, the best chemotherapy regimen in HIV-positive and HIV-negative PCNSL is still a matter of debate. PCNSL in HIV-positive patients is unique among intracranial neoplasms because of its strong association with Epstein-Barr virus (EBV) infection [16]. The presence of EBV in PCNSL has led to the development of antiviral therapy-based regimens including nucleoside analogs such as azidothymidine (AZT) and ganciclovir that demonstrated promising initial results [17]. The tumor manifestations of PCNSL in HIV-negative and HIV-positive patients are similar, however more diffuse and multifocal involvement is frequently detected in HIV-related PCNSL [18, 19]. Despite similar clinical presentation, there have been no comparative studies evaluating the prognostic factors and treatment outcomes in HIV-negative and HIV-positive patients with PCNSL treated at the same institution. The current analysis was conducted retrospectively to examine the clinical characteristics and prognostic factors in HIV-negative and HIVpositive patients with PCNSL treated at the same institutions and to assess effect of HAART therapy on the outcome of HIV-positive patients. We also describe the different treatment approaches adopted in our HIV-negative and HIV-positive patient population and assess OS and progression-free survival (PFS) for each cohort of patients. Materials and methods After approval by our institutional review board to perform retrospective review of clinical charts of patients with PCNSL, we identified patients diagnosed with PCNSL between January 1999 and December 2008 at Jackson Memorial Hospital and Sylvester Comprehensive Cancer Center through the institutional tumor registry and the pathology database. Using a standardized data collection form, the medical records of all patients were reviewed for the following data: age, sex, ethnicity, HIV status, CD4 cell count at the time of PCNSL and HIV diagnosis, treatment with HAART, presenting clinical symptoms, Karnofsky performance status (KPS), serum lactate dehydrogenase (LDH) level, staging information including neuroimaging studies and cerebrospinal fluid (CSF) analysis (cytology and EBV DNA PCR), treatment modalities, response to treatment, and follow-up data. Survival data and date of death were obtained from our institutional tumor registry, chart reviews, and social security death index searches. Biopsy confirmation was required for diagnosis of PCNSL in HIV-negative patients. While the brain biopsy is the gold standard for diagnosis of PCNSL also in the HIVpositive patients, it is associated with considerable morbidity (8%) and mortality (3%) in this population [20] and thus frequently substituted by clinical diagnostic criteria [21, 22]. Therefore, in the HIV-positive patients, the diagnosis was established by fulfillment of the following 3 clinical criteria: (1) compatible magnetic resonance (MRI) scan; (2) compatible positive SPECT thallium-201 study; (3) unresponsiveness to toxoplasmosis treatment or positive EBV-DNA PCR in CSF. EBV infection was assessed by Epstein-Barr virus encoded RNA (EBER) in pathology specimens and by EBV-DNA PCR in the CSF [23, 24]. Systemic evaluation included CT scan of the chest, abdomen and pelvis and bone marrow biopsy. Complete response (CR) was defined as a complete disappearance of all enhancing lesion(s) on MRI or CT scan imaging in patients that discontinued use of all corticosteroids for at least 2 weeks and had negative CSF fluid and no evidence of active ocular lymphoma (if present at initial presentation). Partial response (PR) was defined as [50% decrease in the largest diameter of the largest lesion. CR and PR comprised overall response (OR) [25]. OS was defined as the time from diagnosis until death from any cause. Progression free survival (PFS) was defined as the time from diagnosis until documented disease progression or disease-related death. Patients not treated for their lymphoma were excluded from survival analyses. Patients with less than 30 days of follow-up were excluded from survival analyses unless they died or their disease progressed within the 30 days.

3 J Neurooncol (2011) 101: Patients were divided into two groups based on their HIV status. The two groups demographic and clinical characteristics were compared using the Student s t test and chi-square test. Kaplan Meier survival curves were constructed and compared between the two groups using the log rank test. Multivariate analyses of survival were performed by Cox proportional hazards models incorporating the prognostic factors identified in the univariate log rank test. A P value less than 0.05 was considered statistically significant. SPSS 16.0 (SPSS Inc., Chicago, IL, USA) was used for all statistical analyses. Results Forty-one patients with positive HIV serology and 45 HIVnegative patients were included in the study. PCNSL was diagnosed by histological examination of resection or biopsy specimens in all 45 HIV-negative patients and 7 of the HIV-positive patients. In the remaining 34 HIV-positive patients, PCNSL was diagnosed based on the criteria described in materials and methods. Additionally, the diagnosis was confirmed by autopsy in two HIV-positive patients. Of the 9 HIV-positive patients for whom we had a histological diagnosis, 5 had DLBCL and the remaining four had T-cell lymphoma not otherwise specified, small lymphocytic, marginal zone, and follicular lymphomas. Of the 45 HIV-negative patients, 42 (93%) had diffuse large B-cell lymphoma (DLBCL) and the remaining three had Burkitt s, small lymphocytic, and marginal zone lymphomas. Demographic and selected characteristics of the patients in the two groups are compared cumulatively in all the patients as well as in patients treated for PCNSL (Table 1). HIV-positive patients were younger at the time of PCNSL diagnosis and there were more non-hispanic whites in the HIV-negative group. HIV-positive patients were more likely to present with elevated serum LDH level and have EBV-related PCNSL. None of the HIV-positive patients tested had positive CSF cytology findings. Table 2 demonstrates the presenting symptoms and signs in HIVnegative and positive patients. HIV-positive patients were more likely to present with seizures and less likely to present with hemiparesis. Of note, ocular involvement (retinal infiltrates or presence of cells in the vitreus) was detected in 9% of HIV-negative patients with PCNSL, but in none of the HIV-positive patients. HAART treatment data was available for 37 HIV positive patients, of which 13 (35%) were on HAART at the time of PCNSL diagnosis and 17 (46%) were started on HAART after the PCNSL diagnosis (Table 3). Thirty-three (92%) of 35 HIV-positive patients with available CD4 count had CD4 counts less than 200/mm [3] at the time of PCNSL diagnosis. The mean interval between HIV and PCNSL diagnoses was 55 months with a range of months. Therapy was given at the discretion of the treating physician and varied between the patients (Table 4). Radiotherapy (alone or combined with chemotherapy) was more frequently utilized in the HIV-negative group (68% vs. 27%, P \ 0.001). Furthermore, HIV-negative patients were more likely to receive radiotherapy at doses above 36 Gy compared to HIV-positive patients (68% vs. 18%, P = 0.005). No clinical response was seen in 2 HIVpositive patients who received radiotherapy only. Chemotherapy was incorporated into the therapeutic approach in 27 (60%) HIV-negative and 25 (61%) HIV-positive patients. The most commonly used chemotherapy regimen in the HIV-negative group was methotrexate-based, following the DeAngelis protocol [26], with an ORR of 73%. All 25 HIV-positive patients were treated with AZT combinations, with a cumulative ORR of 56% [17, 27]. These included 12 patients treated with high-dose AZT (HD- AZT) combined with high dose IL-2 and ganciclovir, 11 patients treated with HD-AZT combined with hydroxyurea, and 2 patients treated with HD-AZT in combination with high-dose methotrexate (HD-MTX) alternating with or without R-EPOCH. Five of these patients are still alive in continuous remission up to 8 years post diagnosis and therapy. There was no statistical difference in ORR between HIV-negative and positive patients. However, CR was achieved in 59 and 26% of HIV-negative and HIVpositive patients respectively who received therapy for PCNSL (P = 0.010). For the combined cohort of the HIV negative and positive patients, 8 patients were lost to follow-up before their response to treatment could be evaluated. With a median follow up of 10.3 months (range, ), the median PFS and OS of HIV negative patients was 7.9 (95% CI: ) and 14.6 months (95% CI: ), respectively. With a median follow up of 3.8 months (range, ), the median PFS and OS of HIV positive patients was 3.4 (95% CI: ) and 4.0 months (95% CI: ), respectively (Table 5; Fig. 1). There was a statistically significantly difference in OS (P = 0.034) between HIV positive and negative patients who received therapy for PCNSL. In the combined cohort of the HIV negative and positive patients, shorter OS was associated with KPS \ 70 (Table 5), as previously reported [28]. Comparison of OS of HIV-negative and HIV-positive patients with PCNSL with distinct clinical variables demonstrated that HIV positivity was associated with numerically worse OS in all patient strata (Table 6) reaching statistically inferior OS among patients with KPS \ 70 and in patients treated with chemotherapy and radiation. In the HIVpositive group, OS was shorter in patients with KPS \ 70 and patients who did not receive HAART (Table 7).

4 260 J Neurooncol (2011) 101: Table 1 Demographic and selected clinical characteristics of HIV-negative and positive patients (cumulatively and in patients who received therapy for PCNSL) Characteristics All patients n (%) Treated patients n (%) HIV- (n = 45) HIV? (n = 41) P HIV- (n = 38) HIV? (n = 27) P Male sex 22 (49) 24 (58) NS 19 (50) 18 (67) NS Age at diagnosis* 59.8 ± ± 1.3 \ ± ± 1.3 \0.001 Ethnicity \0.001 \0.001 White-Hisp 22 (49) 18 (44) 21 (55) 12 (44) White-non-Hisp 18 (40) 0 (0) 12 (32) 0 (0) Black 3 (7) 23 (56) 3 (8) 15 (56) Asian 2 (4) 0 (0) 2 (5) 0 (0) LDH [ normal 16/36 (44) 24/29 (83) /33 (45) 17/21 (81) KPS C70 25/43 (58) 19/39 (49) NS 21/36 (58) 15/26 (58) NS C50 \ (28) 7 (18) NS 9 (25) 4 (15) NS \50 6 (14) 13 (33) (17) 7 (27) NS DLBCL histology 41/44 (93) 5/10 (50) 36/38 (95) 3/8 (38) Size, largest lesion (cm)* 2.7 ± ± 0.2 NS 2.7 ± ± 0.2 NS Number of lesions* 1.9 ± ± 0.2 NS 1.9 ± ± 0.2 NS EBV-detected 3/26 (11) 11/19 (58) /24 (13) 11/16 (69) * Median ± standard error Table 2 Selected clinical characteristics of HIV-negative and positive patients with PCNSL a 39 HIV-negative and 17 HIVpositive patients underwent ophthalmologic examination PCNSL primary CNS lymphoma Characteristics HIV- (n = 45), n (%) HIV? (n = 41), n (%) P Signs and symptoms Headache 23 (51) 24 (60) NS Fatigue 9 (20) 8 (20) NS Ataxia 17 (38) 15 (38) NS Blurry vision 16 (36) 9 (22) NS Hemiparesis 17 (38) 6 (15) Nausea/vomiting 11 (24) 13 (33) NS Memory loss 14 (31) 20 (50) NS Seizure 5 (11) 15 (38) Speech difficulties 12 (27) 5 (13) NS Involved sites Ophthalmic a 4 (9) 0 (0) NS Frontal lobe 22 (49) 22 (54) NS Occipital lobe 8 (18) 5 (12) NS Temporal lobe 6 (13) 9 (22) NS Parietal lobe 9 (20) 7 (17) NS Cerebellar 9 (20) 9 (22) NS Periventricular 12 (27) 19 (46) NS Pituitary region 4 (9) 1 (2) NS Midbrain/hindbrain 2 (4) 5 (12) NS Moreover, in patients who were not on HAART before PCNSL diagnosis, initiation of HAART was associated with better OS (median (95% CI): 4.0 ( ) vs. 1.1 ( ) months, P = 0.007). There was no difference in OS and PFS between HIV-positive patients with histologically confirmed PCNSL and those with clinically diagnosed PCNSL. In the HIV-negative group, we did not identify any significant prognostic factors for OS, although patients with KPS \ 70 had numerically shorter median OS (13.2 vs months).

5 J Neurooncol (2011) 101: Table 3 Selected HIV-related characteristics of HIV positive patients with PCNSL a Mean ± SE (range) HAART highly-active antiretroviral treatment, PCNSL primary CNS lymphoma Characteristics n? /n (%) HAART at diagnosis 13/37 (35) HAART added after PCNSL diagnosis 17/24 (71) CD4 at diagnosis of PCNSL \ /35 (92) CD4 at diagnosis of PCNSL \ 50 23/35 (66) CD4 at PCNSL a 56 ± 14 (2-374) CD4 at HIV diagnosis a 75 ± 13 (2-247) Interval between HIV and PCNSL diagnoses (mos) a 55.2 ± 8.8 ( ) Table 4 Overall response (OR) rates and overall survival (OS) achieved in HIV negative and positive patients according to the treatment modalities used Treatment n OR P Median OS ± SE (mos) WBRT only HIV NS 10.3 ± 4.5 HIV? 2 0 Not calculated Chemotherapy only HIV- 7 4 NS 21.3 ± 8.5 NS HIV? ± 4.3 WBRT? chemotherapy HIV NS 23.8 ± HIV? ± 1.5 No treatment HIV ± 10.9 NS HIV? ± 0.3 HIV- N = 45; HIV? N = 41 SE standard error, WBRT whole brain radiotherapy Chemotherapy regimens used: DeAngelis protocol, fludarabine? cytarabine, high dose methotrexate with or without temozolomide, Azidothymidine in combination with hydroxyurea or high dose methotrexate or high dose IL-2? ganciclovir Discussion In this retrospective study of 86 patients with PCNSL, we compared clinical characteristics and outcome of HIVnegative and HIV-positive patients and assessed the impact of HAART on HIV-positive patients with these tumors. Our HIV-positive cohort did not differ from prior study populations in terms of CD4 cell count at presentation [29] (mean 56 cells/l) and tumor histology [30]. Younger age, worse performance status, higher LDH level at presentation and shorter OS were observed more commonly in the HIVpositive compared to the HIV-negative patients with PCNSL. Although few studies suggested a higher incidence of multiple lesions in AIDS-related PCNSL [18, 19], we did not identify any differences in the size and number of lesions between the two groups. While there was a predilection of HIV-related PCNSL to involve the P Table 5 Prognostic factors for overall survival (OS) and progressionfree survival (PFS) in HIV-positive and negative patients treated for PCNSL Factors (n) OS (months) PFS (months) a Median ± SE P HR (95% CI) Median ± SE P Age B50 (28) 8.7 ± 3.1 NS 4.7 ± 2.3 NS [50 (34) 11.7 ± ± 1.1 HIV Neg (35) 14.6 ± ( ± 0.5 NS Pos (27) 4.0 ± ) 3.4 ± 1.1 LDH B ± 5.2 NS 7.0 ± 1.6 NS (21) [ ± ± 2.2 (32) KPS C70 (26) 15.8 ± ( ± 1.0 NS \70 (35) 7.9 ± ) 6.4 ± 2.3 Tumor size B3 cm 7.9 ± 2.3 NS 6.8 ± 0.6 NS (32) [3 cm 10.3 ± ± 1.1 (20) No. of lesions B2 (42) 13.2 ± 2.9 NS 7.3 ± 0.4 NS [2 (18) 6.5 ± ± 1.8 WBRT Yes (41) 10.3 ± 3.0 NS 6.9 ± 0.5 NS No (21) 8.7 ± ± 1.2 a Number of patients included in PFS analyses is lower due to patients lost to follow-up before response assessment SE standard error, HR hazard ratio, LDH lactate dehydrogenase, KPS Karnofsky performance status, WBRT whole brain radiotherapy, PCNSL primary CNS Lymphoma periventricular areas and a majority of HIV-positive patients presented with memory loss and seizures, the prevalence of these findings did not differ statistically in

6 262 J Neurooncol (2011) 101: Fig. 1 Kaplan Meier plot showing the overall survival (OS) curves (a) and progression-free survival (PFS) curves (b) in treated HIVpositive and negative patients with PCNSL. Statistically significant difference in OS (P = 0.034) was observed between HIV-negative and positive patients Table 6 Comparison of OS between treated HIV negative and positive patients in various strata Strata HIV negative HIV positive P KPS C 70 (n = 42) 16.0 ± ± 4.1 NS KPS \ 70 (n = 36) 13.2 ± ± Received only chemotherapy (n = 21) 21.3 ± ± 4.2 NS Received only radiotherapy (n = 12) 10.3 ± NS Received chemo and radiotherapy (n = 29) 23.8 ± ± Median OS ± SE in months and P value from log-rank tests are shown KPS Karnofsky performance status Table 7 Factors impacting OS in HIV-positive patients treated for PCNSL (data in months) Factors Median OS ± SE Median OS ± SE P Age B 50 (n = 19) vs. [50 (n = 8) 6.9 ± ± 1.6 NS CD4 C 50 (n = 8) vs. \50 (n = 17) 6.8 ± ± 1.2 NS HAART (n = 23) vs. no HAART (n = 4) 6.9 ± ± LDH B 618 (n = 4) vs. [618 (n = 17) ± 1.0 NS KPS C 70 (n = 15) vs. \70 (n = 11) 8.7 ± ± Histologically confirmed (n = 8) vs. clinically diagnosed (n = 19) 6.9 ± ± 0.9 NS Tumor size B 3cm(n= 11) vs. [3 cm(n= 9) 6.8 ± ± 1.6 NS Number of lesions B 2(n= 17) vs. [2 (n= 8) 8.7 ± ± 1.5 NS Received WBRT (n = 11) vs. not (n = 16) 3.4 ± ± 4.2 NS PCNSL primary CNS lymphoma, HAART highly-active anti-retroviral treatment, LDH lactate dehydrogenase, KPS Karnofsky performance status, WBRT whole brain radiotherapy comparison to the HIV-negative patients. To the best of our knowledge, this is the first comparison of HIV-negative and positive PCNSL patients treated at the same institution. Currently, there is no uniform approach to the treatment of PCNSL in both immunocompetent and immunocompromised patients. Previous studies conducted in HIVnegative PCNSL have shown OR rates approaching 100% and median OS ranging between 30 and 60 months when a combined-modality approach involving chemotherapy and radiation was used [15, 31 34]. However, it is possible that these promising results that were observed in patients enrolled in clinical trials may not represent the general population of PCNSL patients, since rigorous enrollment criteria applied in these studies may select more favorable patients. Indeed, most reported PCNSL studies excluded patients with very poor performance status as well as HIVpositivity. In our HIV-negative cohort, median OS in patients who received a combined modality approach was relatively short at 23.8 months. The retrospective nature, inclusion of all the patients without any exclusion criteria and use of different therapy regimens in our study may make a head-to-head comparison with the previous reports difficult and may contribute to inferior outcome of our patients compared to patients enrolled on studies.

7 J Neurooncol (2011) 101: The therapeutic options for HIV-positive patients include high-dose WBRT and HD-MTX which provide modest improvement (\12 months) in DFS and are often associated with devastating morbidity [35, 36]. One small series suggested a survival benefit of HD-MTX with a median OS of 9.5 months in HIV-related PCNSL [35]. Addition of WBRT to systemic therapy was not associated with improved OS, likely because only 18% of HIV-positive patients received WBRT C 36 Gy. EBV is implicated in the pathogenesis of the HIV-positive PCNSL and thus presents a unique opportunity to develop targeted therapeutic strategies [16, 27]. EBV-encoded thymidine kinase may increase activity of nucleoside analogues, AZT and ganciclovir, selectively potentiating their apoptotic effects in EBV-infected lymphoma cells. In a small number of studies, a regimen consisting of interleukin-2, ganciclovir and high-dose AZT was shown to improve ORR and OS [17, 37]. In our study, all systemic therapies used in HIVpositive patients included HD-AZT and yielded ORR, OS and PFS of 56%, 4.0 and 3.4 months, respectively. These results may seem to be poorer compared to the outcome achieved with HD-MTX, suggesting that this therapeutic approach may be inferior, although in the absence of sideby-side comparison, it is impossible to reach such conclusion. The median OS was 8.7 months in our HIVpositive patients with KPS of C70, comparable to the OS reported with HD-MTX. Combination therapy with AZT, IL-2 and ganciclovir was investigated in a phase II study conducted by AIDS Malignancy Consortium; however the trial was closed due to lack of accrual as only 4 patients were recruited for this study. The addition of anti-viral therapy to chemotherapy and/or RT has a strong theoretical rationale in AIDS patients but treatment efficacy, safety profile and toxicity need to be compared to the standard chemotherapeutic approaches in those patients who are suitable to receive aggressive treatment. We identified significant differences in CR rate and OS between HIV-negative and HIV-positive patients. Previously, in a cohort of 338 HIV-negative PCNSL patients, only age and performance status were identified as prognostic factors [28]. We did not find age and performance status to be prognostic factors for OS in HIV-negative patients, likely due to the low number of patients in the cohort. While we identified KPS as a good prognostic factor in both the combined and HIV-positive cohorts, age was not found to be a prognostic factor, likely due to the younger age and poor survival in HIV-positive patients and inadequate power of the study in the HIV-negative patients. Inferior outcome of HIV-positive compared to HIV-negative patients with PCNSL could not be attributed to age since HIV positive patients were younger. Consequently, the inferior survival of the HIV-positive patients is most probably attributed to more aggressive nature of their lymphomas, immunosuppression leading to increased incidence of infections, and decreased immune response against EBV associated with the pathogenesis of the PCNSL in those patients. The majority of HIV-related PCNSL cases arise in patients with CD4 counts \ 50 cells/ l, and development of EBV-associated PCNSL is preceded by a loss of functional EBV-specific cytotoxic T cells, demonstrating the role of immune response in PCNSL development [38]. A previous population-based cohort study, similar to our analysis, showed poorer survival outcome than that observed in clinical trials and hospitalbased cohort studies, perhaps reflecting systematic differences between study subjects and patients in the general population [7, 39]. Clinical trial participants and patients treated at specialized centers may have better performance status and other observable and non-observable advantageous prognostic characteristics compared with patients treated in a community setting. In our HIV-positive cohort, treatment with HAART was found to be associated with better OS, concurrent with the findings from two other small published series [40, 41]. Although the role of HAART in the treatment of HIVrelated PCNSL has not been clearly established, it may improve survival by restoring immune response against EBV and PCNSL. Therefore, we recommend immediate start of HAART therapy to all newly diagnosed HIVpositive patients with PCNSL. However, even with HA- ART therapy, the survival of HIV-positive patients with PCNSL is inferior to survival of HIV-negative patients, in marked contrast to systemic lymphoma. Different pathogenesis of systemic and PCNSL in HIV-positive patients and more pronounced immunosuppression of patients with PCNSL, as reflected by lower CD4 counts, may underlie inferior outcomes even following HAART therapy. This study has certain limitations. It is a retrospective and non-randomized study with various different treatment regimens used. Use of clinical diagnostic criteria for PCNSL in HIV-positive patients may have resulted in inclusion of non-pcnsl patients, however, we found no difference in OS and PFS between HIV-positive patients with histologically confirmed and clinically diagnosed PCNSL. Although there is almost 100% co-association between EBV and HIV-related PCNSL, the retrospective review of our records could identify the positive EBV status in only 58% of the cases; documented by either positive CSF EBV-PCR or immunohistochemistry. Although HIV-positive patients who were treated with HAART had better OS, patients who had a better performance status or less co-morbidities at baseline may have been more likely to receive therapy. On the other hand, therapeutic trials usually limit enrollment based on inclusion criteria such as performance status, vital organ status and thus are representative of specific subpopulations of

8 264 J Neurooncol (2011) 101: PCNSL patients that may inherently have better outcome. Our retrospective cohort of patients may be more similar to the general patient population encountered by practitioners in the clinic and thus may better reflect the realistic outcome of these patients that may be inferior compared to outcomes reported in clinical trials. In summary, HIV-positive patients had a worse prognosis compared to HIV-negative patients, although the tumor manifestations and clinical characteristics were similar. In keeping with previous studies, we demonstrated that patients with better performance status had improved OS. Additionally, treatment with HAART conferred better OS in HIV-positive patients. In patients with HIV-related PCNSL, treatments with dual efficacy against HIV and EBV merit further investigation. Acknowledgment ISL is supported by the United States Public Health Service, National Institutes of Health [grant numbers R01- CA109335, R01-CA122105], Fidelity Foundation and the Dwoskin Family Foundation. References 1. Lister A, Abrey LE, Sandlund JT (2002) Central nervous system lymphoma. Hematol Am Soc Hematol Educ Program 2002: Olson JE, Janney CA, Rao RD et al (2002) The continuing increase in the incidence of primary central nervous system non- Hodgkin lymphoma: a surveillance, epidemiology, and end results analysis. Cancer 95: Rosenblum ML, Levy RM, Bredesen DE et al (1988) Primary central nervous system lymphomas in patients with AIDS. Ann Neurol 23(Suppl):S13 S16 4. Welch K, Finkbeiner W, Alpers CE et al (1984) Autopsy findings in the acquired immune deficiency syndrome. JAMA 252: Sparano JA, Anand K, Desai J et al (1999) Effect of highly active antiretroviral therapy on the incidence of HIV-associated malignancies at an urban medical center. J Acquir Immune Defic Syndr 21(Suppl 1):S18 S22 6. Inungu J, Melendez MF, Montgomery JP (2002) AIDS-related primary brain lymphoma in Michigan, January 1990 to December AIDS Patient Care STDS 16: Panageas KS, Elkin EB, DeAngelis LM et al (2005) Trends in survival from primary central nervous system lymphoma, : a population-based analysis. Cancer 104: Mounier N, Spina M, Gisselbrecht C (2007) Modern management of non-hodgkin lymphoma in HIV-infected patients. Br J Haematol 136: Bessell EM, Lopez-Guillermo A, Villa S et al (2002) Importance of radiotherapy in the outcome of patients with primary CNS lymphoma: an analysis of the CHOD/BVAM regimen followed by two different radiotherapy treatments. J Clin Oncol 20: LMSW DeAngelis, Seiferheld W, Schold SC (2002) Combination chemotherapy and radiotherapy for primary central nervous system lymphoma: Radiation Therapy Oncology Group Study J Clin Oncol 20: Gavrilovic IT, Hormigo A, Yahalom J et al (2006) Long-term follow-up of high-dose methotrexate-based therapy with and without whole brain irradiation for newly diagnosed primary CNS lymphoma. J Clin Oncol 24: O Brien PC, Roos DE, Pratt G et al (2006) Combined-modality therapy for primary central nervous system lymphoma: long-term data from a Phase II multicenter study (Trans-Tasman Radiation Oncology Group). Int J Radiat Oncol Biol Phys 64: Yamanaka R, Shinbo Y, Sano M et al (2007) Salvage therapy and late neurotoxicity in patients with recurrent primary CNS lymphoma treated with a modified ProMACE-MOPP hybrid regimen. Leuk Lymphoma 48: Freilich RJDJ, Delattre J-Y, Monjour A, DeAngelis LM (1996) Chemotherapy without radiation therapy as initial treatment for primary CNS lymphoma in older patients. Neurology 46: Glass J, Gruber ML, Cher L et al (1994) Preirradiation methotrexate chemotherapy of primary central nervous system lymphoma: long-term outcome. J Neurosurg 81: Tanner JE, Alfieri C (2001) The Epstein-Barr virus and posttransplant lymphoproliferative disease: interplay of immunosuppression, EBV, and the immune system in disease pathogenesis. Transpl Infect Dis 3: Raez L, Cabral L, Cai JP et al (1999) Treatment of AIDS-related primary central nervous system lymphoma with zidovudine, ganciclovir, and interleukin 2. AIDS Res Hum Retrovir 15: Schwaighofer BW, Hesselink JR, Press GA et al (1989) Primary intracranial CNS lymphoma: MR manifestations. AJNR Am J Neuroradiol 10: Gage JT, Vance EA, Hildenbrand PG et al (2000) Brain lesion and AIDS. Proc (Bayl Univ Med Cent) 13: Antinori A, Ammassari A, Luzzati R et al (2000) Role of brain biopsy in the management of focal brain lesions in HIV-infected patients. Gruppo Italiano Cooperativo AIDS and Tumori. Neurology 54: Antinori A, Ammassari A, De Luca A et al (1997) Diagnosis of AIDS-related focal brain lesions: a decision-making analysis based on clinical and neuroradiologic characteristics combined with polymerase chain reaction assays in CSF. Neurology 48: Fitzsimmons A, Upchurch K, Batchelor T (2005) Clinical features and diagnosis of primary central nervous system lymphoma. Hematol Oncol Clin N Am 19: , vii 23. De Luca A, Antinori A, Cingolani A (1995) Evaluation of cerebrospinal fluid EBV-DNA and IL-10 as markers for in vivo diagnosis of AIDS-related primary central nervous system lymphoma. Br J Hematol 90: Antinori A, De Rossi G, Ammassari A et al (1999) Value of combined approach with thallium-201 single-photon emission computed tomography and Epstein-Barr virus DNA polymerase chain reaction in CSF for the diagnosis of AIDS-related primary CNS lymphoma. J Clin Oncol 17: Abrey LE, Batchelor TT, Ferreri AJ et al (2005) Report of an international workshop to standardize baseline evaluation and response criteria for primary CNS lymphoma. J Clin Oncol 23: DeAngelis LM (1999) Primary CNS lymphoma: treatment with combined chemotherapy and radiotherapy. J Neurooncol 43: Roychowdhury S, Peng R, Baiocchi RA et al (2003) Experimental treatment of Epstein-Barr virus-associated primary central nervous system lymphoma. Cancer Res 63: Abrey LE, Ben-Porat L, Panageas KS et al (2006) Primary central nervous system lymphoma: the Memorial Sloan-Kettering Cancer Center prognostic model. J Clin Oncol 24: Corti M, Villafane F, Trione N et al (2004) Primary central nervous system lymphomas in AIDS patients. Enferm Infecc Microbiol Clin 22: Raez LE, Patel P, Feun L et al (1998) Natural history and prognostic factors for survival in patients with acquired immune

9 J Neurooncol (2011) 101: deficiency syndrome (AIDS)-related primary central nervous system lymphoma (PCNSL). Crit Rev Oncog 9: Gabbai AA, Hochberg FH, Linggood RM et al (1989) High-dose methotrexate for non-aids primary central nervous system lymphoma. Report of 13 cases. J Neurosurg 70: Littman P, Wang CC (1975) Reticulum cell sarcoma of the brain. A review of the literature and a study of 19 cases. Cancer 35: Dahlborg SA, Henner WD, Crossen JR et al (1996) Non-AIDS primary CNS lymphoma: first example of a durable response in a primary brain tumor using enhanced chemotherapy delivery without cognitive loss and without radiotherapy. Cancer J Sci Am 2: Abrey LE, Yahalom J, DeAngelis LM (2000) Treatment for primary CNS lymphoma: the next step. J Clin Oncol 18: Jacomet C, Girard PM, Lebrette MG et al (1997) Intravenous methotrexate for primary central nervous system non-hodgkin s lymphoma in AIDS. AIDS 11: Khoo VS, Liew KH (1999) Acquired immunodeficiency syndrome-related primary cerebral lymphoma. Clin Oncol (R Coll Radiol) 11: Aboulafia DM (2002) Interleukin-2, ganciclovir, and high-dose zidovudine for the treatment of AIDS-associated primary central nervous system lymphoma. Clin Infect Dis 34: Gasser O, Bihl FK, Wolbers M et al (2007) HIV patients developing primary CNS lymphoma lack EBV-specific CD4? T cell function irrespective of absolute CD4? T cell counts. PLoS Med 4:e Nelson DF, Martz KL, Bonner H et al (1992) Non-Hodgkin s lymphoma of the brain: can high dose, large volume radiation therapy improve survival? Report on a prospective trial by the Radiation Therapy Oncology Group (RTOG): RTOG Int J Radiat Oncol Biol Phys 23: Hoffmann C, Tabrizian S, Wolf E et al (2001) Survival of AIDS patients with primary central nervous system lymphoma is dramatically improved by HAART-induced immune recovery. AIDS 15: Skiest DJ (2003) Survival is prolonged by highly active antiretroviral therapy in AIDS patients with primary central nervous system lymphoma. AIDS 17: