Human immunodeficiency virus-related primary central nervous system lymphoma

Human Immunodeficiency Virus–Related PrimaryCentral Nervous System LymphomaFactors Influencing Survival in 111 Patients

Mark E. Newell, M.B.B.S., M.P.H1

Jennifer F. Hoy, M.B.B.S.2,3

Stephen G. Cooper, M.B.B.S.4

Bernadette DeGraaff, B.Sc.2,3

Andrew E. Grulich, Ph.D.1

Melissa Bryant, B.Sc.2,3

Jeremy L. Millar, M.B.B.S.5

Bruce J. Brew, M.B.B.S., M.D.1

David I. Quinn, M.B.B.S., Ph.D.6

1 National Centre in HIV Epidemiology and ClinicalResearch, University of New South Wales, Sydney,Australia.

2 Infectious Diseases Unit, Alfred Hospital, Mel-bourne, Australia.

3 Department of Medicine, Monash University,Melbourne, Australia.

4 Department of Radiation Oncology, St. Vincent’sHospital, Sydney, Australia.

5 William Buckland Radiation Therapy Centre, Al-fred Hospital, Melbourne, Australia.

6 Department of Clinical Pharmacology and Toxi-cology, University of New South Wales, Sydney,Australia.

Presented in part at the 10th Annual Conference ofthe Australian Society for HIV Medicine, Adelaide,Australia, November 18–21, 1998.

David I. Quinn’s current address: Division of Med-ical Oncology, Keck School of Medicine, Universityof Southern California, Los Angeles, California.

Address for reprints: David I. Quinn, Division ofMedical Oncology, Keck School of Medicine, Uni-versity of Southern California, 1441 Eastlake Ave-nue, Suite 3453, Los Angeles, CA 90033; Fax:(323) 865-0061; E-mail: [email protected]

Received October 23, 2003; revision received Feb-ruary 11, 2003; accepted March 24, 2004.

BACKGROUND. The current study evaluated factors influencing survival in patients

diagnosed with human immunodeficiency virus (HIV)-related primary central

nervous system lymphoma (PCNSL), with a focus on the effects of therapeutic

radiotherapy (RT) and highly active antiretroviral therapy (HAART).

METHODS. A retrospective chart review of patients with a diagnosis of HIV-related

PCNSL at one of five university hospitals between 1987 and 1998 was performed.

Clinical details including antiretroviral agent use, brain imaging scan results, RT

use, and survival outcomes were recorded.

RESULTS. One hundred eleven patients with HIV-related PCNSL were identified.

The annual incidence decreased significantly between 1992 and 1995 and between

1996 and 1998 (P � 0.04). The median survival period was 50 days (mean, 109 days;

range, 4 –991 days), with improved survival for patients diagnosed after 1993.

Patients treated with two or more antiretroviral agents had improved survival (P

� 0.01), as did patients who received RT (P � 0.0001). For patients who received

RT, completion of the prescribed course and treatment to � 30 Gray (Gy) inde-

pendently predicted a more favorable outcome. RT used in conjunction with

antiretroviral therapy involving two or more agents had an additive positive effect

on survival. For patients who did not receive RT, poor performance status and

encephalopathy predicted a shorter survival duration.

CONCLUSIONS. The results of the current study suggest that HAART and treatment

with RT to � 30 Gy improve survival for patients with HIV-related PCNSL. This

combination of therapies may provide a standard of care as the basis for further

trials of chemotherapy, novel adjunctive treatment, and quality of life assessment.

Cancer 2004;100:2627–36. © 2004 American Cancer Society.

KEYWORDS: central nervous system lymphoma, acquired immunodeficiency syn-drome, antiretroviral therapy, radiotherapy.

Before the onset of the human immunodeficiency virus (HIV) ep-idemic, primary central nervous system lymphoma (PCNSL) was

a rare malignancy, accounting for � 1% of all primary intracranialtumors and only 1–2% of all non-Hodgkin lymphomas (NHL).1,2 Afterthe onset of the HIV epidemic in the early 1980s, the number ofreported cases of PCNSL increased. Among HIV-infected patients,PCNSL is the second most common cerebral mass lesion after toxo-plasmosis.3 It is still uncommon as an initial manifestation of ac-quired immunodeficiency syndrome (AIDS). For example, � 1% ofHIV-infected patients are likely to develop PCNSL as the primaryAIDS-defining illness. However, it is more common after the onset ofAIDS, occurring in 2–13% of patients with a previous AIDS diagnosis.4

PCNSL tends to develop in patients with more severe immunodefi-

2627

© 2004 American Cancer SocietyDOI 10.1002/cncr.20300Published online 14 May 2004 in Wiley InterScience (www.interscience.wiley.com).

ciency compared with patients with systemic NHL,with the median CD4-positive cell counts at diagnosisbeing 10 cells/mm3 and 189 cells/mm3, respectively,in these patients.5 Greater than 95% of AIDS-relatedPCNSLs are of B-cell origin and are associated withEpstein-Barr virus coinfection. The large cell immu-noblastic pattern is the most common histopathologicpattern observed.4,6

The median survival of patients with HIV-relatedPCNSL is poor. Several studies report a median sur-vival of 3 months for those who receive radiotherapy(RT) and 1 month for those who do not receive treat-ment.7,8 A number of cohort studies have suggestedthat RT, performance status, previous AIDS-definingillness, and CD4 count may predict outcome. It isunclear whether the survival difference in patientsreceiving RT is a primary beneficial effect that is at-tributable to the RT itself or to selection bias based onother prognostic factors that, in turn, influence theclinical decisions leading to the use of this therapeuticmodality. One study suggests that HIV-infected pa-tients with PCNSL who receive RT are less likely to dieof PCNSL than of other mortal manifestations of AIDS,whereas patients not treated with RT are more likely todie of PCNSL.9 Moreover, the effect of highly activeantiretroviral therapy (HAART) in patients who al-ready have HIV-related PCNSL is not well defined; toour knowledge, a study in which survival improved insix patients who received HAART and a case reportrepresent the only published experience in the medi-cal literature.10,11

The aims of the current study were to determinewhether the incidence of HIV-related PCNSL has de-creased with the introduction of combination antiret-roviral therapy, to determine the prognostic factorsassociated with increased survival, to determinewhether an independent survival benefit associatedwith RT was evident when other prognostic factorswere controlled, and to attempt to identify a subgroupof patients who would be most likely to benefit fromRT.

MATERIALS AND METHODSWith the approval of an ethics committee, caserecords of all patients diagnosed with PCNSL duringthe period 1987–1998 were audited from the medicalrecords of five Australian university hospitals: AlfredHospital (Melbourne), Fairfield Hospital (Melbourne),St. Vincent’s Hospital (Sydney), Prince of Wales Hos-pital (Sydney), and Royal Perth Hospital (Perth). Eachof these centers was a designated state referral centerfor the treatment of HIV disease and had on-site con-sultative and treatment services in neurosurgery, neu-rology, and radiation oncology. These centers treated

54% of the patients with AIDS in Australia during thestudy period. Patients with PCNSL were identified us-ing a combination of hospital coding records, hospi-tal-based HIV databases, the Australian national AIDSdatabase, and databases from radiation oncology andanatomic pathology departments. Patients at the twoSydney hospitals were matched by linking a databaseof all AIDS-related NHL with the New South WalesCancer Registry.12. Patients initially considered eligi-ble for inclusion in the study but subsequently foundineligible were documented but not included in fur-ther analysis. Patients who were excluded had a diag-nosis of CNS lymphoma but were found to have aprevious or concurrent diagnosis of systemic NHL.Patients who presented with lymphoma confined tothe leptomeninges but who did not have evidence ofcerebral lesions after computed tomographic (CT) ormagnetic resonance imaging (MRI) scans were ex-cluded because of heterogeneity of clinical presenta-tion and a significant chance of concurrent systemicNHL. This process yielded 111 evaluable patients withHIV infection with a diagnosis of PCNSL. Analyseswere performed for these patients. Patients who com-pleted RT planning were deemed to have commencedRT treatment for the purposes of the analysis.

Data CollectionBaseline data were collected, including gender, date ofbirth, date of AIDS diagnosis, list and date of otherprevious AIDS-defining illnesses if not PCNSL (includ-ing previously suspected toxoplasmosis), Eastern Co-operative Oncology Group (ECOG) performance sta-tus at admission, date of PCNSL diagnosis, presentingsymptoms (classified as headache, seizure, ataxia, fo-cal neurologic symptoms, and/or nonfocal neurologicsymptoms [e.g., confusion, cognitive decline, and/ormemory loss]), anatomic site(s) of PCNSL (classified ascerebral cortex, periventricular site, basal ganglia,brain stem, and/or cerebellum), antiretroviral therapy,and previous therapy for suspected toxoplasmosis.The date of diagnosis of PCNSL was defined as thedate of the first abnormal CT or MRI scan showing alesion considered to be PCNSL after nonresponse to acourse of antitoxoplasma therapy (if applicable). Clin-ical data collected subsequent to a PCNSL diagnosisincluded the dates and details of brain biopsies; RTtreatment details, including fields, prescribed and de-livered doses, and fractionation; outcome; and sur-vival. All patient data were stored in a confidentialmanner using a name code and date of birth in apassword-protected computer database.

2628 CANCER June 15, 2004 / Volume 100 / Number 12

Statistical MethodsStatistical analyses were performed with Statview soft-ware (Version 4.5; Abacus Concepts, Berkeley, CA),using Cox regression univariate and multivariate anal-yses,13 Kaplan–Meier analysis,14 and log-rank tests, aswell as the Fisher exact test as appropriate. Statisticalsignificance in the current study was defined by P� 0.05. All reported P values are two sided. Univariateanalysis was performed to identify baseline factorsthat had a significant impact on survival. Multivariatesurvival analysis was undertaken using the variablesidentified on univariate analysis as having prognosticsignificance. The effect of RT was tested by generatingmultivariate models that included as variables the ad-ministration of RT (RT treatment model), completionof RT (RT completion model), and receipt of a mini-mum radiation dose of 30 Gray (Gy) (RT dose model)to determine whether any survival benefit persistedwhen other factors were controlled for. Given poten-tial concerns and differing opinions in the medicalliterature regarding treatment effects in patients witha clinical but not a histologic diagnosis of PCNSL,15–20

a subgroup analysis of the 47 patients with biopsy-proven PCNSL was performed using the same threemodels.

RESULTSBaseline DemographicsOne hundred eleven patients with PCNSL were in-cluded in the database for analysis. One hundred eightpatients were male, with a mean age of 38.0 years(median, 36.6 years; range, 22.0 – 63.2 years). Otherbaseline characteristics are shown in Tables 1 and 2.The original data abstraction identified an additional14 patients presenting with lymphoma involving theCNS. However, a subsequent review of case notesrevealed that these patients did not have PCNSL, butrather systemic NHL.

IncidenceThe incidence of new cases of PCNSL, both as primaryand subsequent AIDS-defining illnesses, increasedmarkedly to a peak of 16 cases per year in 1992 and1993, followed by a subsequent decline, with only twocases per year in 1997 and 1998 (Fig. 1). In assessingdifferences in incidence between time periods in thecohort, the overall trend in incidence was found to besignificant (P � 0.03), as were the increases between1987 and 1990 and between 1991 and 1993 (P � 0.03)and the downward trends between 1992 and 1995 andbetween 1996 and 1998 (P � 0.03). This latter changein incidence coincides with the onset of the use of

HAART in Australia and a concurrent decrease in theincidence of AIDS dementia complex.21

Presentation, Diagnosis, and Treatment of PrimaryCentral Nervous System LymphomaAll patients had neurologic symptoms and signs doc-umented at presentation. For example, 61.3% of pa-tients had focal neurologic symptoms, 51.4% of pa-tients had nonfocal neurologic symptoms, 43.2% ofpatients had headache, 21.6% of patients had seizure,and 18% of patients had ataxia. All patients received aCT scan on presentation, with additional MRI for 18patients (16.2%), 8 of whom had normal or nonspe-cific findings on CT scans. The anatomic sites of PC-NSL at presentation were the cerebral cortex in 65% ofpatients, the periventricular region in 56% of patients,the basal ganglia in 33% of patients, the cerebellum in

TABLE 1Baseline Demographics for Patients with HIV-Related PCNSL(n � 111)

VariableEntirecohort

RT

P valuebNo(n � 57)

Yesa

(n � 54)

GenderMale 108Female 3

Age (yrs) 0.09Median 38.0 39.0 36.9Mean 36.6 39.7 35.3Range 22.0–63.2 22.0–58.4 23.9–63.2

ECOG performance status 0.007Median 2.5 2.7 2.3Mean 3.0 3.0 2.0Range 0–4.0 0–4.0 0–4.0

No. of previous AIDS-definingillnesses 0.04

Median 1.8 2.0 1.52Mean 2.0 2.0 1.00Range 0–5.0 0–4.0 0–5.0

Time from AIDS diagnosis toPCNSL diagnosis (days) 0.32

Median 487 493 480Mean 256 295 206Range 0–3828 0–3828 0–3232

CD4 lymphocyte count (cells/mm3) 0.77Median 37.4 25.5 50.1Mean 11 11.0 10.5Range 1–750 1–150 1–750

% not receiving antiretroviraltherapy at PCNSL diagnosis 34.2 33.4 35.2 0.84

HIV: human immunodeficiency virus; PCNSL: primary central nervous system lymphoma; RT: radio-

therapy; ECOG: Eastern Cooperative Oncology Group; AIDS: acquired immunodeficiency syndrome.a Completed planning of treatment with radiotherapy.b P value for difference between patients who received radiotherapy and patients who did not.

HIV-Related Primary CNS Lymphoma/Newell et al. 2629

7% of patients, and the brain stem in 4% of patients.Approximately one-half (48.6%) of the lesions werefound in one region only, and the remaining lesionswere found in two or more anatomic areas. Data onthe number of lesions present on pretreatment CT orMRI scans were available for 90 patients. A mean of 2.8lesions were present (median, 2 lesions; range, 1–16

lesions). No significant difference in the number oflesions at presentation was found when the periodsbefore and after January 1993 were compared (P� 0.54). There was a significant association betweenthe presence of periventricular PCNSL and nonfocalneurologic symptoms and signs, such as confusion,cognitive decline, and memory loss (P � 0.02). Therealso was an association between such symptoms andthe presence of three or more lesions on imagingscans (P � 0.04). There were trends toward signifi-cance between seizures and the presence of PCNSL atperiventricular (P � 0.06) and cerebral cortical sites (P� 0.1). Contemporaneous cerebrospinal fluid exami-nation was performed for nine patients and yieldedcytologically negative findings in all of them.

For 91% of patients, a presumptive diagnosis ofcerebral toxoplasmosis based on CT or MRI findingswas made, and these patients were treated with py-rimethamine and either sulfadiazine or clindamycinfor a median treatment period of 15 days; dexameth-asone therapy also was administered to the majority ofthese patients (88.3%). The rate of use of toxoplasmo-sis therapy before biopsy or before the start of treat-ment for lymphoma did not vary between the pre-1993 and post-1993 groups (P � 0.66). Subsequently, adiagnosis of PCNSL was made in 57.4% of patientsbased on a lack of imaging response to antitoxoplasmatherapy. In the remaining 42.6% of patients, brainbiopsy provided a histologic diagnosis. Eleven patients(10%) were judged to have concurrent PCNSL andtoxoplasmosis. Data on histopathologic classificationwere available for 41 of the 47 patients who underwentbiopsy. Nineteen patients had high-grade immuno-blastic lymphoma, 21 patients had intermediate-gradediffuse large cell lymphoma, and 1 patient had small

TABLE 2Univariate Analysis Results for Patients with HIV-Related PCNSL(n � 111)

VariableNo. ofpatients (%)

Hazard ratio(95% CI) P a

Mediansurvival(days)

Age (yrs)� 40 42 (37.8) 1.41 (0.95–2.09) 0.093 53� 40 69 (62.2) 50

Date of diagnosis� 1/1/93 63 (56.8) 1.56 (1.06–2.30) 0.026 49� 1/1/93 48 (43.2) 60

CD4 count (cells/mm3)� 125 104 (93.6) 2.31 (1.05–5.09) 0.038 49� 125 7 (6.4) 110

ARV agents0 or 1 86 (77.5) 1.88 (1.16–3.02) 0.0098 492 or 3 25 (22.5) 59

Nonfocal CNSsymptoms

Yes 57 (51.8) 2.17 (1.47–3.20) � 0.0001 40No 53 (48.2) 98

Focal CNS symptomsYes 42 (38.2) 0.91 (0.62–1.34) 0.63 50No 68 (61.8) 51

ECOG performancestatus

� 3 67 (60.4) 1.87 (1.26–2.76) 0.0018 40� 2 44 (39.6) 103

Biopsy-proven NHLNo 64 (58.2) 2.00 (1.35–2.96) 0.0005 40Yes 47 (41.8) 79

No. of previous AIDS-defining illnesses

� 3 63 (56.8) 1.47 (1.00–2.17) 0.0497 48� 2 48 (43.2) 65

Received RTNo 57 (51.4) 3.05 (2.01–4.63) � 0.0001 51Yes 54 (48.6) 92

Completed RTNo 73 (65.8) 3.50 (2.24–5.47) � 0.0001 40Yes 38 (34.2) 116

RT dose (Gy) received(n � 54)

� 30 18 (33.3) 2.91 (1.59–5.33) � 0.0001 43� 30 36 (66.7) 116

HIV: human immunodeficiency virus; PCNSL: primary central nervous system lymphoma; CI: confi-

dence interval; ARV: antiretroviral; CNS: central nervous system; ECOG: Eastern Cooperative Oncology

Group; NHL: non-Hodgkin lymphoma; AIDS: acquired immunodeficiency syndrome; RT: radiotherapy;

Gy: grays.a P value for survival difference based on univariate analysis.

FIGURE 1. Incidence of human immunodeficiency virus–related primary

central nervous system lymphoma per calendar year in the study cohort.


and cleaved cell lymphoma. Histopathologic classifi-cation did not influence survival (P � 0.48).

Approximately one-half of all patients (n � 54[48.6%]) underwent planning for RT. Of these 54 pa-tients, 36 (66.7%) completed the prescribed RT course.Of the patients who completed RT, 67% (n � 24)received a total dose � 30 Gy. The total delivered doserange for patients completing the prescribed RTcourse was 15–56 Gy, all administered through op-posed right and left lateral fields to the whole brain. Ofthe 54 patients receiving RT, 12 (22.2%) died of a CNSsyndrome, presumably related to PCNSL, whereas 40of 57 patients (61.4%) who did not receive RT died ofa CNS syndrome (P � 0.0001). Of the 25 patients whoreceived two or three antiretroviral drugs, 7 (28%)developed a new AIDS-related infection after the di-agnosis of PCNSL, whereas 36 of 86 patients (42.8%)who received 0 or 1 antiretroviral drug developed suchan infection (P � 0.008).

SurvivalUnivariate analysisThe median survival of the cohort was 50 days (mean,109 days; range, 4 –991 days). On univariate analysis(Table 2, Figs. 2 and 3), baseline parameters that hada significant positive impact on survival were diagno-sis after January 1, 1993, a CD4-positive lymphocytecount � 125 cells/mm3, previous antiretroviral ther-apy with two or more drugs, the absence of nonfocalCNS symptoms, an ECOG performance status of � 2,biopsy-proven NHL, receipt of any RT, completion ofRT, and a received RT dose of � 30 Gy. Neuroana-tomic site and number of lesions present had no sig-nificant impact on survival (P � 0.2; log-rank test).Patients who did not receive antiretroviral therapybefore PCNSL diagnosis had an outcome similar tothat of patients who received previous antiretroviraltherapy (P � 0.35; log-rank test). Among patients whosubsequently received more than two antiretroviraldrugs, RT (P � 0.01), completion of prescribed RT (P� 0.003), RT to � 30 Gy (P � 0.01), and biopsy con-firmation of PCNSL diagnosis (P � 0.003) were asso-ciated with improved survival.

Multivariate analysisTable 3 summarizes the multivariate analyses under-taken using the three RT models: the RT treatmentmodel, the RT completion model, and the RT dosemodel (which is based on a total dose cutoff of 30 Gy,the dose considered the standard of care in mostAustralian and international centers7). Different cutoffpoints for RT doses received were explored on univar-iate analysis for their effects on survival: 20 Gy, P� 0.01; 30 Gy, P � 0.0006; and 40 Gy, P � 0.43.

Increasing radiation dose was associated with im-proved survival as a continuous variable (P � 0.008).Additional significant independent predictors of im-proved survival in the RT treatment model were a CD4lymphocyte count � 125 cells/mm3 and the absenceof nonfocal neurologic symptoms and signs. Using theRT completion model, only the absence of nonfocal

FIGURE 2. Survival according to (A) radiotherapy (RT) use (n � 111; log-rank

P � 0.0001), (B) completion of the prescribed RT course (n � 111; log-rank

P � 0.0001), and (C) RT dose received (n � 54; log-rank P � 0.0001). Gy:

grays.


neurologic symptoms and signs was found to be asignificant predictor of improved survival in additionto RT dose � 30 Gy. However, there was a nonsignif-icant association between improved survival and CD4count � 125 cells/mm3 (P � 0.08) and between im-proved survival and ongoing therapy with at least twoantiretroviral drugs (P � 0.09). In the RT dose model,the significant predictors of improved survival thatwere used included ongoing combination antiretrovi-ral therapy with at least two agents, the absence ofnonfocal neurologic syndrome, and no more than twoprevious AIDS-defining illnesses, in addition to intentto treat with RT. On backward-stepwise regressionmodeling, the most predictive factors of improvedsurvival were found to be receipt of RT, RT comple-tion, and total RT dose � 30 Gy, followed by (in de-creasing order of effect on survival), ongoing therapywith two or more antiretroviral drugs, no more thantwo previous AIDS-defining illnesses, and the absenceof nonfocal CNS symptoms and signs. Using multivar-iate analysis for the subgroup of patients who receivedtreatment with two or more antiretroviral drugs, eachof the RT-related variables was found to be indepen-dently predictive of an improved outcome, whereas noother variable achieved significance.

Additional analysesIn the subgroup analysis of patients who had biopsy-proven NHL (n � 47), univariate analysis revealed thatthe parameters that had a significant positive impacton survival included diagnosis after January 1, 1993,ongoing antiretroviral therapy with at least two agents,receipt of RT, completion of RT, and RT dose � 30 Gy(Table 4). On multivariate analysis, ongoing antiretro-viral therapy with at least two agents remained a sig-nificant predictor of survival along with RT in all threeRT models. The significant correlation between poorsurvival and diagnosis before 1993 that was identifiedon univariate analysis was no longer significant whenthe number of antiretroviral agents was controlled foron multivariate analysis.

Additional analysis was performed to identify po-tential prognostic factors associated with failure tocomplete RT. Significant factors included ECOG per-formance status � 3 and age � 40 years (P � 0.04 and0.05, respectively). In the subgroup of patients whodid not receive any RT, the absence of nonfocal neu-rologic symptoms and ECOG performance status � 2were significant positive predictors of survival (P� 0.009 and 0.002 respectively; log-rank test). No otherfactors were predictive of survival in this subgroup.

FIGURE 3. Survival according to ongoing antiretroviral therapy. (A) Entire

cohort (n � 111; log-rank P � 0.008). (B) Subset of patients receiving

radiotherapy (RT; n � 54; log-rank P � 0.07). (C) Subset of patients receiving

a total RT dose � 30 grays (n � 36; log-rank P � 0.02). ARVA: antiretroviral

agent.


DISCUSSIONThe current study evaluated a large series of patientswith HIV-related primary cerebral lymphoma usingsurvival as the primary endpoint. The salient findingswere that 1) combination antiretroviral therapy, espe-cially with HAART, has markedly decreased the inci-dence of PCNSL in the study population; 2) ongoingtreatment with combination antiretroviral therapywas associated with improved survival; 3) treatmentwith RT to a dose of 30 Gy was associated with im-proved survival; and 4) older patients (age � 40 years)and patients with a poor performance status (ECOGperformance status � 3) were less likely to complete aprescribed course of RT.

Our reported overall median survival of 50 days,with significantly improved survival for patients re-ceiving any RT (median, 92 days) compared with pa-tients who did not receive RT (median, 38 days), issimilar to the findings of other studies of HIV-relatedPCNSL. In a study of 55 patients performed early inthe AIDS epidemic, the reported median survival was119 days for patients who received RT, compared withonly 27 days for patients who did not.8 In a cohort of17 patients treated with varying doses of RT, the re-ported median survival was 72 days.22,23 In a 10-yearanalysis of all patients with HIV-related PCNSL (n� 41) treated at the University of California–San Fran-cisco (San Francisco, CA), the reported median sur-

TABLE 3Multivariate Analysis summary for Patients with HIV-Related PCNSL (n � 111)

Variable

RT treatment model (n � 110) RT completion model (n � 110) RT dose model (n � 54)

Hazard ratio(95% CI) P value



Age (yrs)� 40 0.98 (0.62–1.57) 0.96 0.88 (0.56–1.41) 0.6 0.96 (0.45–2.01) 0.91� 40

Date of diagnosis� 1/1/93 0.73 (0.41–1.30) 0.29 0.80 (0.45–1.41) 0.44 0.57 (0.21–1.58) 0.28� 1/1/93

CD4 count (cells/mm3)� 125 2.73 (1.15–6.50) 0.02 2.24 (0.92–5.46) 0.076 2.65 (0.85–8.22) 0.09� 125

ARV agents0 or 1 1.45 (0.81–2.56) 0.21 1.66 (0.93–2.97) 0.087 2.31 (1.05–5.06) 0.0372 or more

Nonfocal CNS syndromeYes 1.87 (1.20–2.92) 0.006 1.87 (1.18–2.94) 0.007 1.49 (0.72–3.09) 0.28No

ECOG performance status� 3 1.44 (0.93–2.23) 0.10 1.38 (0.89–2.13) 0.15 1.40 (0.70–2.80) 0.34� 2

Biopsy-proven NHLNo 1.35 (0.79–2.30) 0.27 1.21 (0.71–2.05) 0.48 1.07 (0.45–2.55) 0.89Yes

No. of previous AIDS-defining illnesses

� 3 1.33 (0.86–2.06) 0.2 1.29 (0.84–1.96) 0.24 2.53 (1.30–4.95) 0.007� 2

Received RTNo 1.62 (0.96–2.71) 0.069 — — — —Yes

Completed RTNone/not — — 2.51 (1.44–4.36) 0.001 — —Completed

RT dose (Gy) received(n � 54)

� 30 — — — — 3.25 (1.47–7.21) 0.004� 30

HIV: human immunodeficiency virus; PCNSL: primary central nervous system lymphoma; CI: confidence interval; ARV: antiretroviral; CNS: central nervous system; ECOG: Eastern Cooperative Oncology Group; NHL:

non-Hodgkin lymphoma; AIDS: acquired immunodeficiency syndrome; RT: radiotherapy; Gy: grays.


vival durations were 3 months and 4.5 months forpatients who received RT.24 A review of survival datafor 174 patients in 32 series of AIDS-related PCNSLreported an overall mean survival of 2.6 months, 3months for patients receiving RT and 0.9 months forpatients not receiving RT.7 These individual studies,like ours, clearly were confounded by selection bias inthe assignment of therapy, but they were performedbefore the advent of combination antiretroviral ther-apy.

To our knowledge, the current study is one of thefirst to demonstrate improved outcome for patientswith HIV-related PCNSL treated with combination an-tiretroviral therapy. However, recent reports suggestthat this effect may be relatively widespread, at least inthe developed world.25,26 In addition, there have beenreports of clinical responses and disease remission inpatients with PCNSL after the commencement ofHAART with or without RT.10,11 It is possible that theeffect of antiretroviral therapy was influenced by thestandard practice in the period before 1993 of ceasingadministration of zidovudine (AZT) when RT was ini-tiated due to concerns regarding myelosuppressionand anemia during RT. However, the data in the cur-rent retrospective analysis do not allow evaluation ofthis effect.

The data regarding the recent decrease in inci-dence of HIV-related primary lymphoma in the cur-rent cohort are encouraging. It is likely that the sharpdecrease in the incidence of cases following the rapid

increase in the early 1990s is attributable to combina-tion antiviral therapy, which became the standard ofcare in Australia after 1994 (Fig. 1). This sharp de-crease in incidence has also been observed for othervirally mediated opportunistic infections, such as Ka-posi sarcoma and systemic NHL, in patients with HIVboth in Australia27,28 and elsewhere29,30 and supportsthe hypothesis that improved immune status due toeffective antiviral therapy resulted in the observeddecreases in incidence. The sequential decreases inincidence in the current cohort between 1993 and1994 and then between 1996 and 1997 may also reflectsequential changes in antiretroviral therapy during thestudy period. For example, the addition of didanosineto zidovudine occurred between 1992 and 1993, fol-lowed by the addition of protease inhibitors (HAART)beginning in 1995.21 The evidence to support this con-tention is limited, but these changes in therapy alsocoincide with downward inflection points in the over-all incidence of AIDS in Australia.31–33 Similar patternsin the incidence of PCNSL and AIDS have been ob-served at a number of North American centers (TupuleA. Unpublished data),34 whereas other groups havereported only small decreases in PCNSL incidenceover a similar time period.25,26,30,35,36 Further evidencein support of this sequenced effect on incidence isprovided by data on 1255 patients diagnosed withAIDS between 1994 and 1997 in eight U.S. cities.37

That study demonstrated that increases in the inten-

TABLE 4Multivariate Analysis for the Subgroup of Patients with a Biopsy-Proven Diagnosis of PCNSL (n � 47)

Variable

Univariate analysis

Multivariate analysis

RT treatment RT completion RT dose

HR (95% CI) P HR (95% CI) P HR (95% CI) P HR (95% CI) P

Date of diagnosis� 1/1/93 2.34 (1.12–4.10) 0.02 0.77 (0.36–1.63) 0.49 1.04 (0.48–2.25) 0.93 1.13 (0.44–2.87) 0.80� 1/1/93

ARV agents� 1 2.17 (1.13–4.15) 0.02 1.90 (0.91–3.98) 0.09 2.90 (1.30–6.49) 0.01 2.81 (1.13–7.00) 0.03� 2

Received RTNo 2.81 (1.40–5.67) 0.004 2.54 (1.21–5.31) 0.01 — — — —Yes

Completed RTNo 2.87 (1.53–5.37) 0.001 — — 3.74 (1.83–7.62) 0.0003 — —Yes

RT dose (Gy)� 30 2.75 (1.26–5.97) 0.008 — — — — 4.40 (1.79–10.82) 0.001� 30

PCNSL: primary central nervous system lymphoma; HR: hazard ratio; CI: confidence interval; RT: radiotherapy; Gy: grays; ARV: antiretroviral.


sity of antiretroviral therapy were associated withstepwise reductions in morbidity and mortality.37

Improved survival in patients selected for RT maybe at least partially attributable to selection bias. Pa-tients who received any RT had a better mean ECOGperformance status (P � 0.007), fewer previous AIDS-defining illnesses (P � 0.04), and a nonsignificanttrend toward younger age (P � 0.09; Table 1). In theUniversity of California–San Francisco series, therewas a significant correlation (P � 0.02) between base-line performance status and duration of survival.24

These findings are supported by a large multicenterstudy of 163 patients with HIV-related PCNSL treatedwith RT and evaluated for response.9 Good perfor-mance status, younger age, and higher doses of RTwere correlated significantly with higher responserates,9 whereas other groups have demonstrated theprevailing significance of performance status as a pre-dictor of outcome.38 The reported improvement insurvival among patients who received � 30 Gy ofradiation in the current study confirms the rationalefor recommending a total dose of 30 – 40 Gy for effec-tive treatment of PCNSL.7–9,22,23 Fine and Mayer7 haverecommended a shorter course of 30 Gy administeredin 10 fractions for patients with poorer performancestatus. This dose is considered standard at most Aus-tralian and international centers.7 In the current se-ries, an ECOG performance status of � 3 was expectedto be predictive of noncompletion of RT as well aspoor survival for patients not receiving RT, but thiswas not the case. The presence of diffuse neurologicsymptoms and low CD4 lymphocyte counts remainedsignificant predictors of outcome independent of anyRT effect. Nonetheless, the treatment benefit of RTpersisted when these two factors were controlled foron multivariate analysis. Our analyses suggest that thepatients most likely to benefit from RT were those whohad a good pretreatment performance status and whowere able to receive ongoing therapy with at least twoantiretroviral agents. Improved survival was associ-ated with a total RT dose of � 30 Gy. Since the com-pletion of the current study, we have routinely offeredRT and HAART to patients with good performancestatus. In general, patients with poorer performancestatus are offered similar therapy, provided that ourclinical impression is that they will survive beyond theplanned period of RT.

Although we have demonstrated that treatmentwith RT, completion of RT, receipt of a total dose of 30Gy, ongoing combination antiretroviral therapy, nomore than one previous AIDS-defining illness, and theabsence of nonfocal CNS symptoms and signs werepredictors of prolonged survival, we have also con-firmed the poor prognosis of patients with HIV-related

PCNSL. The data presented in the current study sug-gest that RT and HAART represent the standard of carefor patients with good performance. This treatmentcombination may serve as a baseline upon which thisgroup of patients can be offered adjunctive novel ther-apies, including chemotherapy,39 – 41 in large multi-center clinical trials. The development of trial strate-gies for patients with poor performance status or forpatients who do not have access to or cannot receivecombination antiretroviral therapy presents an evengreater challenge.

REFERENCES1. Freeman CR, Shustik C, Brisson ML, Meagher-Villemure K,

Dylewski I. Primary malignant lymphoma of the centralnervous system. Cancer. 1986;58:1106 –1111.

2. Hochberg FH, Miller DC. Primary central nervous systemlymphoma. J Neurosurg. 1988;68:835– 853.

3. Gray F, Gherardi R, Scaravilli F. The neuropathology of theacquired immune deficiency syndrome (AIDS). A review.Brain. 1988;111:245–266.

4. Forsyth PA, DeAngelis LM. Biology and management ofAIDS-associated primary CNS lymphomas. Hematol OncolClin North Am. 1996;10:1125–1134.

5. Pluda JM, Venzon DJ, Tosato G, et al. Parameters affectingthe development of non-Hodgkin’s lymphoma in patientswith severe human immunodeficiency virus infection re-ceiving antiretroviral therapy. J Clin Oncol. 1993;11:1099 –1107.

6. Meeker TC, Shiramizu B, Kaplan L, et al. Evidence for mo-lecular subtypes of HIV-associated lymphoma: division intoperipheral monoclonal, polyclonal and central nervous sys-tem lymphoma. AIDS. 1991;5:669 – 674.

7. Fine HA, Mayer RJ. Primary central nervous system lym-phoma. Ann Intern Med. 1993;119:1093–1104.

8. Baumgartner JE, Rachlin JR, Beckstead JH, et al. Primary cen-tral nervous system lymphomas: natural history and responseto radiation therapy in 55 patients with acquired immunode-ficiency syndrome. J Neurosurg. 1990;73:206–211.

9. Corn BW, Donahue BR, Rosenstock JG, et al. Performancestatus and age as independent predictors of survival amongAIDS patients with primary CNS lymphoma: a multivariateanalysis of a multi-institutional experience. Cancer J Sci Am.1997;3:52–56.

10. Hoffmann C, Tabrizian S, Wolf E, et al. Survival of AIDSpatients with primary central nervous system lymphoma isdramatically improved by HAART-induced immune recov-ery. AIDS. 2001;15:2119 –2127.

11. McGowan JP, Shah S. Long-term remission of AIDS-relatedprimary central nervous system lymphoma associated withhighly active antiretroviral therapy. AIDS. 1998;12:952–954.

12. Grulich AE, Wan X, Coates M, Day P, Kaldor JM. Validationof a non-identifying method of linking cancer and AIDSregister data. J Epidemiol Biostat. 1996;1:207–212.

13. Cox DR. Regression models and life tables (with discussion).J R Stat Soc. 1972;34:187–189.

14. Kaplan EL, Meier P. Nonparametric estimation from incom-plete observations. J Am Stat Assoc. 1958;53:457– 481.

15. Holloway RG, Mushlin AI. Intracranial mass lesions in ac-quired immunodeficiency syndrome: using decision analy-sis to determine the effectiveness of stereotactic brain bi-opsy. Neurology. 1996;46:1010 –1015.


16. Luzzati R, Ferrari S, Nicolato A, et al. Stereotactic brainbiopsy in human immunodeficiency virus-infected patients.Arch Intern Med. 1996;156:565–568.

17. Viale P, Pagani L, Mori M, et al. Diagnosis of HIV-relatedprimary central nervous system lymphoma: is there still arole for brain biopsy? A case history. Ann Oncol. 1997;8:491–496.

18. Ammassari A, Scoppettuolo G, Murri R, et al. Changingdisease patterns in focal brain lesion-causing disorders inAIDS. J Acquir Immune Defic Syndr Hum Retrovirol. 1998;18:365–371.

19. Antinori A, Ammassari A, De Luca A, et al. Diagnosis ofAIDS-related focal brain lesions: a decision-making analysisbased on clinical and neuroradiologic characteristics com-bined with polymerase chain reaction assays in CSF. Neu-rology. 1997;48:687– 694.

20. Bower M, Fife K, Sullivan A, et al. Treatment outcome inpresumed and confirmed AIDS-related primary cerebrallymphoma. Eur J Cancer. 1999;35:601– 604.

21. Dore GJ, Correll PK, Li Y, Kaldor JM, Cooper DA, BrewBJ. Changes to AIDS dementia complex in the era of highlyactive antiretroviral therapy. AIDS. 1999;13:1249 –1253.

22. Goldstein JD, Dickson DW, Moser FG, et al. Primary centralnervous system lymphoma in acquired immune deficiencysyndrome. A clinical and pathologic study with results oftreatment with radiation. Cancer. 1991;67:2756 –2765.

23. Goldstein JD, Zeifer B, Chao C, et al. CT appearance ofprimary CNS lymphoma in patients with acquired immuno-deficiency syndrome. J Comput Assist Tomogr. 1991;15:39 –44.

24. Ling SM, Roach M III, Larson DA, Wara WM. Radiotherapyof primary central nervous system lymphoma in patientswith and without human immunodeficiency virus. Tenyears of treatment experience at the University of CaliforniaSan Francisco. Cancer. 1994;73:2570 –2582.

25. Sacktor N, Lyles RH, Skolasky R, et al. HIV-associated neu-rologic disease incidence changes: Multicenter AIDS CohortStudy, 1990-1998. Neurology. 2001;56:257–260.

26. Chow KU, Mitrou PS, Geduldig K, Helm EB, Hoelzer D,Brodt HR. Changing incidence and survival in patients withAIDS-related non-Hodgkin’s lymphomas in the era of highlyactive antiretroviral therapy (HAART). Leuk Lymphoma.2001;41:105–116.

27. Grulich AE, Li Y, McDonald AM, Correll PK, Law MG, KaldorJM. Decreasing rates of Kaposi’s sarcoma and non-Hodgkin’s lymphoma in the era of potent combination anti-retroviral therapy. AIDS. 2001;15:629 – 633.

28. Dore GJ, Li Y, Grulich AE, et al. Declining incidence and lateroccurrence of Kaposi’s sarcoma among persons with AIDS inAustralia: the Australian AIDS cohort. AIDS. 1996;10:1401–1406.

29. Bower M, Fox P, Fife K, Gill J, Nelson M, Gazzard B. Highly

active anti-retroviral therapy (HAART) prolongs time totreatment failure in Kaposi’s sarcoma. AIDS. 1999;13:2105–2111.

30. Conti S, Masocco M, Pezzotti P, et al. Differential impact ofcombined antiretroviral therapy on the survival of Italianpatients with specific AIDS-defining illnesses. J Acquir Im-mune Defic Syndr. 2000;25:451– 458.

31. Law MG, Li Y, McDonald AM, Cooper DA, Kaldor JM. Esti-mating the population impact in Australia of improved an-tiretroviral treatment for HIV infection. AIDS. 2000;14:197–201.

32. Correll PK, Law MG, McDonald AM, Cooper DA, Kaldor JM.HIV disease progression in Australia in the time of combi-nation antiretroviral therapies. Med J Aust. 1998;169:469 –472.

33. Dore GJ, Hoy JF, Mallal SA, et al. Trends in incidence of AIDSillnesses in Australia from 1983 to 1994: the Australian AIDScohort. J Acquir Immune Defic Syndr Hum Retrovirol. 1997;16:39 – 43.

34. Sparano JA, Anand K, Desai J, Mitnick RJ, Kalkut GE, HanauLH. Effect of highly active antiretroviral therapy on the in-cidence of HIV-associated malignancies at an urban medi-cal center. J Acquir Immune Defic Syndr. 1999;21 Suppl1:S18 –S22.

35. Ledergerber B, Egger M, Erard V, et al. AIDS-related oppor-tunistic illnesses occurring after initiation of potent antiret-roviral therapy: the Swiss HIV Cohort Study. JAMA. 1999;282:2220 –2226.

36. Brew BJ, Dore G. Decreasing incidence of CNS AIDS defin-ing events associated with antiretroviral therapy. Neurology.2000;55:1424.

37. Palella FJ Jr., Delaney KM, Moorman AC, et al. Decliningmorbidity and mortality among patients with advanced hu-man immunodeficiency virus infection. HIV OutpatientStudy Investigators. N Engl J Med. 1998;338:853– 860.

38. Raez LE, Patel P, Feun L, Restrepo A, Raub WA Jr., CassilethPA. Natural history and prognostic factors for survival inpatients with acquired immune deficiency syndrome(AIDS)-related primary central nervous system lymphoma(PCNSL). Crit Rev Oncog. 1998;9:199 –208.

39. Forsyth PA, Yahalom J, DeAngelis LM. Combined-modalitytherapy in the treatment of primary central nervous systemlymphoma in AIDS. Neurology. 1994;44:1473–1479.

40. Jacomet C, Girard PM, Lebrette MG, Farese VL, Monfort L,Rozenbaum W. Intravenous methotrexate for primary cen-tral nervous system non-Hodgkin’s lymphoma in AIDS.AIDS. 1997;11:1725–1730.

41. Slobod KS, Taylor GH, Sandlund JT, Furth P, Helton KJ,Sixbey JW. Epstein-Barr virus-targeted therapy for AIDS-related primary lymphoma of the central nervous system.Lancet. 2000;356:1493–1494.


Human immunodeficiency virus-related primary central nervous system lymphoma

Documents