J Cancer Res Clin Oncol (2010) 136:1729–1735

DOI 10.1007/s00432-010-0831-x

ORIGINAL PAPER

Prognostic factors and treatment options in patients with leptomeningeal metastases of diVerent primary tumors: a retrospective analysis

Karin Oechsle · Victoria Lange-Brock · Andreas Kruell · Carsten Bokemeyer · Maike de Wit

Received: 19 October 2009 / Accepted: 8 February 2010 / Published online: 4 March 2010© Springer-Verlag 2010

AbstractPurpose Leptomeningeal metastases (LM) are associatedwith very poor prognosis and data on outcome are limited.We evaluated prognostic factors and treatment options inpatients (pts) with LM of diVerent malignancies in a singlecenter experience.Methods Single center data on characteristics, treatmentand outcome of 135 consecutive pts (73 solid tumors and62 hematologic malignancies) with LM between 1989 and2005 were retrospectively analyzed.Results Treatment consisted of systemic chemotherapy(SC) plus intrathecal chemotherapy (ITC) in 28%, ITCalone in 22%, radiotherapy (RT) plus ITC in 12% and othermodalities (SC, RT, SC + RT) in 7%. Thirteen percent ofpts received supportive care only (4% not evaluable ontreatment). Median survival from diagnosis of LM was2.5 months. Univariate analysis revealed age >50, intervalbetween diagnosis of primary tumor and LM ·12 months,lung cancer and malignant melanoma, and Karnofsky per-formance status ·70 as signiWcant negative predictors foroverall survival. Positive predictive factors were response

in cerebrospinal Xuid and application of SC. In multivariateanalysis, only SC was signiWcantly associated with longermedian survival (5.6 vs. 1.7 months).Conclusions In patients with LM an age >50, perfor-mance status ·70%, interval between diagnosis of primarytumor and LM ·12 months, primary tumor (lung cancer,malignant melanoma) and lack of cytologic response pres-ent negative prognostic factors. Systemic chemotherapy issigniWcantly associated with longer survival time than localtreatment modalities.

Keywords Leptomeningeal metastases · Meningeosis carcinomatosa · Intrathecal chemotherapy · Radiotherapy

Introduction

Leptomeningeal metastases are a rare clinical pictureoccurring in only 5–8% of all patients with malignant dis-eases (Chamberlain 2005; De Angeles and Boutros 2005).Their occurrence always imply a very limited prognosiswith median survival times of 1.8–5.8 months (Chamber-lain 2005; De Angeles and Boutros 2005; Pentheroudakisand Pavlides 2005; Herrlinger et al. 2004; Taillibert andJerzy 2006). Mostly, leptomeningeal metastases are associ-ated with advanced disease and are observed in patientswith primary tumor entities, which tend to form out solidparenchymal cerebral metastases frequently, e.g. malignantmelanoma, lung and breast cancer with rates of develop-ment of leptomeningeal metastases up to 20% (Chamber-lain 2005). Due to optimized treatment strategies in patientswith malignant diseases leading to prolonged survival timesand due to improved medical imaging technology, the inci-dence of leptomeningeal metastases increased over the lastdecades (Pentheroudakis and Pavlides 2005; Herrlinger

K. Oechsle (&) · V. Lange-Brock · C. Bokemeyer · M. de WitDepartment of Oncology/Hematology/Bone Marrow Transplantation/Pneumology, University Cancer Center Hamburg-Eppendorf, Martinistr. 52, 20246 Hamburg, Germanye-mail: k.oechsle@uke.uni-hamburg.de

M. de WitClinic of Internal Medicine–Hematology and Oncology, Vivantes Klinikum Neukoeln, Berlin, Germany

A. KruellDepartment of Radio- Oncology, University Cancer Center Hamburg-Eppendorf, Martinistr. 52, 20246 Hamburg, Germany

123

1730 J Cancer Res Clin Oncol (2010) 136:1729–1735

et al. 2004). Once malignant cells have entered into thecerebrospinal Xuid (CSF) by hematogenous spread, endo-or perineural spread along peripheral nerves, or by directspread from parenchymal cerebral metastases, they canextend at every location in the neurospinal axis (Taillibertand Jerzy 2006). Therefore, leptomeningeal metastasis cancause almost any neurologic symptom originating in thediVerent areas of the central nervous system, brain, cranialnerves and spinal cord (Chamberlain 2005; Taillibert andJerzy 2006).

Usually, the treatment of leptomeningeal metastasesincludes diVerent treatment modalities, such as intrathecaland systemic chemotherapy, radiotherapy and surgery(Chamberlain 2005; De Angeles and Boutros 2005; Pent-heroudakis and Pavlides 2005; Herrlinger et al. 2004;Taillibert and Jerzy 2006). Surgery and radiotherapy aremainly applied for palliation by ventriculoperitonealshunting in patients with hydrocephalus or for the reduc-tion of symptoms resulting from focal lesions (Chamber-lain 2005; De Angeles and Boutros 2005; Pentheroudakisand Pavlides 2005; Herrlinger et al. 2004; Taillibert andJerzy 2006). Systemic chemotherapy remains an optionwhich can treat the systemic disease including the wholeneurospinal axis and might be eVective in on both, symp-tom release and outcome, but data on systemic chemo-therapy alone in patients with leptomeningeal metastasesare rare. However, the eYcacy of systemic chemotherapymay be limited as most systemically applied cytotoxicdrugs do not penetrate the blood-CSF-barrier. Subarach-noidale leptomeningeal metastases may open the blood-CSF-barrier and allow the drugs to penetrate in the cere-brospinal Xuid (Bokstein et al. 1998; Boogerd et al.2004). The eYcacy of intrathecal chemotherapy can bereduced by impaired cerebrospinal Xuid due to intrathecaltumor lesions and increased neurotoxicity rates up to 47%are described (Boogerd et al. 2004). Therefore, the impactof intrathecal chemotherapy is discussed controversiallyand the results of randomized studies on the optimal regi-men of intrathecally applied cytotoxic agents are hetero-geneous (Grossman et al. 1993; Kim et al. 2003; Glantzet al. 1999).

Until now, optimal treatment of leptomeningeal metasta-ses remains in discussion and no standard treatment guide-lines are available. Additionally, little is known concerningprognostic factors in these patients which might help toselect treatment and to spare those patients from intensivetherapy who will not proWt.

This report of our single center experiences tries todetermine prognostic factors in patients with leptomenin-geal metastasis and to retrospectively analyze the eYcacyof diVerent treatment modalities administered within dailyclinical practice in patients with leptomeningeal metastasesof diVerent malignant diseases.

Patients and methods

In this analysis, all consecutive patients with leptomenin-geal metastases of all kinds of primary solid or hematologicmalignancies at the Department of Oncology, Hematology,Bone marrow transplantation with Section Pneumology andthe Department of Radio-Oncology at University HospitalHamburg-Eppendorf, Germany between 1989 and 2005were retrospectively analyzed. As the detection of malig-nant cells in the cerebrospinal Xuid and magnetic resonanceimaging with gadolinium enhancement remains the stan-dard prognostic procedures for diagnosis of leptomeningealmetastasis (Chamberlain 2005; De Angeles and Boutros2005; Pentheroudakis and Pavlides 2005; Herrlinger et al.2004), patients were included in the underlying analysis ifleptomeningeal metastases were proven either by detectionof malignant cells in the cerebrospinal Xuid in microscopiccytology, by typical Wndings at magnetic resonance imag-ing (MRI) or both.

The data of a total of 135 consecutive patients were eval-uated regarding diVerent treatment regimens and possibleprognostic factors on survival. Performance status accord-ing to Karnofsky, presence of clinical symptoms, includingpain and neurologic Wndings at the time of diagnosis of lep-tomeningeal metastases were retrospectively analyzed. Inaddition, all treatment modalities, intrathecal and systemicchemotherapy as well as radiotherapy, were evaluated andcorrelated with the patients’ symptoms, treatment outcomeand survival time. Subgroup analyses were performed inorder to observe diVerences in prognostic factors and treat-ment eYcacy between diVerent hematologic and oncologicmalignancies.

Response to therapy was observed by monitoring ofmalignant cells in the cerebrospinal Xuid and was assumedif cerebrospinal Xuid was completely cleared from malig-nant cells in at least one examination at any time after thestart of leptomeningeal metastases therapy. In patients withpathologic Wndings at MRI without diagnostics of cerebro-spinal Xuid response to treatment was observed by MRI.

Statistical analysis

Data were analyzed according to the following statisticalmethods: univariate analyses of potential prognostic factorsfor survival time according to Kaplan and Meier (1958),and comparison of diVerences in survival between groupsfor statistical signiWcance using the log-rank test. For multi-variate analysis, all factors with a univariate relevant(P · 5%) inXuence on survival were included in the Coxregression model. Results were indicated for estimation ofthe Hazard ratio (relative risk) including the 95%-conW-dence interval. The McNemar’s test was used to estimatethe course of clinical Wndings.

123

J Cancer Res Clin Oncol (2010) 136:1729–1735 1731

Results

A total of 135 patients with a median age of 54.4 (range,16–85) were cared for leptomeningeal metastases at theUniversity Hospital of Hamburg-Eppendorf between 1989and 2005. Malignant diseases had resulted from diVerentsolid tumors in 73 patients (54%) and from hematologicmalignancies in 62 patients (46%). Detailed data onpatient’s primary diseases are presented in Table 1.

Median time between the diagnosis of the primarytumor and Wrst diagnosis of leptomeningeal metastaseswas 12 months (range, 0–22.9 years) with wide variationsbetween diVerent primary tumor entities (e.g. breast can-cer 45 months; high-grade non-Hodgkin’s lymphoma5 months). At time of diagnosis of leptomeningeal metas-tases, 54 of 73 patients with solid primary tumors (86%)presented with concurrent systemic metastases and 25patients (39%) with concurrent solid parenchymal cere-bral metastases.

Diagnosis of leptomeningeal metastases was based onthe demonstration of malignant cells in the cerebrospinalXuid in 132 patients (98%). In three cases, diagnosis of lep-tomeningeal metastases was based on typical Wndings inmagnetic resonance imaging (MRI) at the time of presenceof clinical symptoms without pathologic Wndings in cere-bral spinal Xuid. In these patients, further MRI was per-formed for treatment response evaluation.

Clinical presentation

Data on performance status at the time of Wrst diagnosis ofleptomeningeal metastases were retrospectively availablein 57 patients. Thirty-eight patients (67%) presented with aKarnofsky performance status (KPS) ·70% and 19 patients(33%) with a KPS >70%.

The analysis of clinical symptoms at the time of diagnosisof leptomeningeal metastases revealed the presence of painin 76% of patients. Prior to start of treatment for leptomenin-geal metastases, neurologic Wndings other than unspeciWcpain were observed in 119 of 135 patients (88%). Neurologicsymptoms were brain pressure in 45%, cranial nerve palsy in41%, dizziness in 14%, radicular symptoms in 56%,impaired orientation in 19% and somnolence in 7%. Despiteof treatment, clinical symptoms revealed no signiWcantchanges throughout the treatment course with 69% ofpatients still presenting with pain and 86% still suVeringfrom neurologic symptoms after treatment. Subgroup analy-sis revealed no statistically relevant diVerence in theimprovement of clinical symptoms between patients withcomplete clearance from malignant cells in the cerebrospinalXuid after treatment and patients with incomplete results.Performance status tended to increase due to oncologic treat-ment with 33% of patients presenting with performance sta-tus (Karnofsky index) >70 prior to treatment and 48% aftertreatment, but no statistical signiWcance could be achieved.

Treatment modalities

Combined systemic plus intrathecal chemotherapy for lep-tomeningeal metastases was applied in 38 patients (29%).Intrathecal chemotherapy alone was administered in 29patients (22%). A three-fold combination treatment modal-ity consisting of radiotherapy, systemic and intrathecal che-motherapy was applied in 21 patients (16%). Sixteenpatients (12%) were treated with a combination of radio-therapy and intrathecal chemotherapy. Other treatment reg-imens, either radiotherapy only or the combination ofsystemic chemotherapy and radiotherapy, were performedin eight patients (7%). Eighteen patients (13%) receivedsupportive care only without any antitumoral treatment.Unfortunately, in Wve patients (4%) no data on treatmentwere retrospectively evaluable.

Table 1 Patients’ characteristics (n = 135 patients)

SCLC small cell lung cancer; NSCLC non-small cell lung cancer; CUPcancer of unknown primary; NHL non-Hodgkin’s lymphoma; ALLacute lymphatic leukemia; AML acute myeloic leukemia; CML chronicmyeloic leukemia; LM leptomeningeal metastases

No. of patients (%)

Median age 54.5 years (range, 16–85)

Sex

Female 80 (59%)

Male 55 (41%)

Primary malignancies

Solid tumors 73 (54%)

Breast cancer 43 (32%)

Lung cancer [SCLC/NSCLC] 14 (10%) [8 (6%)/6 (4%)]

Cerebral tumors 5 (4%)

Melanoma 3 (2%)

CUP 4 (3%)

Others 4 (3%)

Hematologic malignancies 62 (46%)

High-grade NHL 39 (29%)

Low-grade NHL 2 (2%)

ALL 8 (6%)

AML 6 (4%)

Blast crisis of CML 3 (2%)

Multiple Myeloma 3 (2%)

Others 1 (1%)

Time between primary tumor and diagnosis of LM

Median 12 months (range, 0–22.9 years)

Performance status at study inclusion (n = 57)

·70% 38 (67%)

>70% 19 (33%)

123

1732 J Cancer Res Clin Oncol (2010) 136:1729–1735

Intrathecal chemotherapy was the most frequent treat-ment modality applied in 104 patients (77%) and wasadministered two or three times per week in dosages of 10–15 mg methotrexate (MTX) and 40 mg cytosine arabino-side (Ara-C). Most frequently, either MTX (44%) or com-bination therapy with MTX and Ara-C (49%) was used forintrathecal therapy. Systemic chemotherapy alone or incombination with other therapy modalities was applied in65 patients (48%). The applied chemotherapy regimensvaried depending on the type of primary tumor. Chemother-apy regimen used for systemic treatment contained antracy-lines in 64%, cyclophosphamide in 59% and Ara-C in 13%of patients. Cytostatic drugs penetrating the blood-CSF-barrier, mainly Methorexat, were applied systemically in23% of patients (13% breast cancer and 10% high-gradenon-Hodgkin’s lymphoma).

Systemic chemotherapy at the time of leptomeningealmetastases was Wrst-line chemotherapy in 16 patients(25%), second-line for relapsed or progressive disease in 42patients (65%) and more than second-line in 7 patients(10%). Radiotherapy was applied in 41 patients (30%) andconsisted of cerebral irradiation, either whole brain radio-therapy or focal irradiation of single lesions, in 24 of these41 patients (59%) or of irradiation of the whole craniospi-nal axis in 17 patients (41%). Radiotherapy was adminis-tered in fractions of 1.5–3 Gray to a total maximum dose of40 Gray.

Outcome and survival analysis

After treatment for leptomeningeal metastases, 48% ofpatients had achieved a response treatment and presentedwithout any microscopically detectable malignant cells incerebrospinal Xuid in at least one examination after the startof leptomeningeal metastases therapy, but in 52% ofpatients malignant cells still could be detected in cerebro-spinal Xuid after treatment.

Median survival time from the time of Wrst diagnosis ofleptomeningeal metastases in the entire cohort of 135patients was 2.5 months (95% conWdence interval 1.9–3.2 months), and one year survival rate was 7.4%. Concern-ing primary malignant disease, median survival was longestin breast cancer patients with 3.1 months and patients withhigh-grade non-Hodgkin’s lymphoma with 3.0 months inmedian. The shortest median survival times were observedin patients with lung cancer with 0.8 months and in patientswith malignant melanoma with 0.9 months.

Based on the treatment modalities applied to thesepatients despite of primary disease median survival waslongest in patients undergoing systemic plus intrathecalchemotherapy (5.6 months) and patients with three-foldcombination treatment including additional irradiation(5.8 months). In patients undergoing systemic chemother-apy alone or within a combined modality treatment mediansurvival time was signiWcantly longer compared to patientsnot receiving any systemic chemotherapy, (5.6 vs.1.8 months). Further details are presented in Table 2.

Prognostic factors

Univariate analysis according to Kaplan–Meier revealeddiVerent signiWcant prognostic factors for reduced survival:type of primary disease (lung cancer and malignant mela-noma), time between diagnosis of primary tumor and ofleptomeningeal metastases <12 months, age <50, Karnof-sky performance status <70%, application of systemic che-motherapy and lack of cytologic response in thecerebrospinal Xuid. Detailed results are presented inTable 3.

Table 2 Survival data according to Kaplan–Meier analysis

NHL non-Hodgkin’s lymphoma; ITC intrathecal chemotherapy; SCsystemic chemotherapy; RT radiotherapy

No. of patients (%)

Median survival (months)

Primary tumor (n = 135)

Breast 43 (32%) 3.1

Lung 14 (10%) 0.8

Melanoma 3 (2%) 0.9

High-grade NHL 39 (29%) 3.0

Others 36 (27%) 2.5

Treatment (n = 130)

ITC + SC 38 (29%) 5.6

ITC 29 (22%) 1.4

ITC + SC + RT 21 (16%) 5.8

ITC + RC 16 (12%) 2.0

Others (RT, SC + RT) 8 (7%) 1.8

Supportive care only 18 (14%) 0.5

Systemic chemotherapy (n = 130)

Yes 65 (50%) 5.6

No 65 (50%) 1.8

Table 3 Univariate analysis of prognostic factors

KPS Karnofsky performance status; LM Leptomeningeal metastases

Condition 1 Condition 2 P-value

Type of primary tumor 0.001

Time between primary tumor and LM

·12 months >12 months 0.029

Patient’s age ·50 years >50 years 0.009

KPS ·70% >70% 0.051

Cytological response Yes No 0.0101

Systemic chemotherapy Yes No 0.0001

123

J Cancer Res Clin Oncol (2010) 136:1729–1735 1733

In contrast, gender, concurrent systemic or solid paren-chymal brain metastases, type of radiotherapy (local irradi-ation vs. whole brain irradiation vs. irradiation ofcraniospinal axis), line chemotherapy of chemotherapy(Wrst vs. second-line) and regimen of intrathecal chemo-therapy (MTX alone vs. MTX + Ara-C) revealed no signiW-cant impact on survival.

Treatment modality was the only factor showing a sig-niWcant inXuence on survival in multivariate Cox regres-sion analyses. Additionally, survival times were evidentlyshorter in patients who were treated with intrathecal che-motherapy alone compared to patients who received com-bined intrathecal and systemic chemotherapy. Theapplication of systemic chemotherapy, either as singletreatment or within combined treatment modalities, wasassociated with a signiWcant longer median survival time(5.6 months) compared to patients with other treatmentmodalities excluding systemic chemotherapy (1.7 months)or patients with supportive care only (0.5 months). Detaileddata on the diVerent treatment modalities and curvesaccording to Kaplan–Meier are presented in Table 4 andFig. 1, respectively.

Discussion

In the underlying single center analysis, the data of 135consecutive patients with leptomeningeal metastases ofdiVerent malignant diseases were retrospectively analyzed.

The evaluation of clinical presentation demonstrated thatclinical symptoms mainly remained stable during treatmentperiod despite of the treatment given and indicating thatestablished neurologic defects are rarely improved evenwith eVective therapy. Although only about half of patientshad reached an adequate response to treatment with com-plete clearance from malignant cells in the cerebrospinal

Xuid, subgroup analysis revealed no diVerence in improve-ment of clinical symptoms in patients with or without treat-ment response in the cerebrospinal Xuid. Therefore, ourdata suggest that palliation by antitumor treatment remainslimited in patients with neurologically evident leptomenin-geal metastases. The phenomenon of Wxed neurologic deW-cits refractory to eVective local and/or systemic treatmentwas described previously and seems to present a negativeprognostic factor on outcome and survival (Taillibert andJerzy 2006; Siegal et al. 1994). The Wxation of neurologicsymptoms despite of oncologic treatment for leptomenin-geal metastases might be inXuenced by the time betweenWrst occurrence of neurologic symptoms and the start oftreatment. Unfortunately, no detailed information on timeinterval between Wrst occurrence of neurologic symptomsand start of treatment for leptomeningeal metastases couldbe presented in the underlying analysis. But this retrospec-tive analysis of consecutive patients might represent usualtime Xow in daily practice. Therefore, early cytologic diag-nostics of cerebrospinal Xuid seems to be of primary impor-tance in patients with advanced malignant diseases andoccurrence of unspeciWc neurologic symptoms to improvethe patients’ chance to proWt from oncologic treatment.

In the current series, univariate analysis revealed thetype of primary disease, time between diagnosis of primarytumor and of leptomeningeal metastases, age, Karnofskyperformance status, treatment, application of systemic che-motherapy and cytologic response in the cerebrospinal Xuidas prognostic factors with signiWcant impact on patient’ssurvival. Previous analysis also had demonstrated thatperformance status and primary tumor histology presentrelevant prognostic factors in patients with leptomeningeal

Table 4 Results of multivariate Cox regression analyses (relative riskon shorter survival comparing to the base categories)

ITC intrathecal chemotherapy; SC systemic chemotherapy; RT radio-therapy

No. of patients (%)

Hazard ratio (95%-CI)

P-value

ITC + SC (basis category)

38 (29) 1

Supportive care only 18 (14) 14.26 (6.96–29.22) 0.0001

ITC 29 (22) 5.06 (2.80–9.14) 0.0001

ITC + RT 16 (12) 2.94 (1.52–5.72) 0.001

ITC + ST + RT 21 (16) 0.80 (0.45–1.42) 0.44

Others (RT, SC + RT) 8 (7) 1.90 (0.78–4.62) 0.16

SC (basis category) 65 (50) 1

Supportive care only 18 (14) 14.99 (7.61–29.51) 0.0001

Fig. 1 Overall survival in patients with systemic chemotherapy (SC)(n = 65) versus treatment without SC (n = 47) or supportive care only(n = 18)

60 50 40 30 20 10 0

Observation time (months)

1,0

0,8

0,6

0,4

0,2

0,0

Cum

ulat

ive

prop

ortio

n su

rviv

ing

p = 0,0001

Systemic chemotherapy (SC)Supportive care only Treatment without SC

123

1734 J Cancer Res Clin Oncol (2010) 136:1729–1735

metastases of malignant diseases. In addition, previousanalyses had observed the status of systemic disease, com-partmentalization of cerebrospinal Xuid, coexistent bulkymetastatic disease in the central nervous system and thepresence of carcinomatous encephalopathy as negativeprognostic factors (Chamberlain 2005; Herrlinger et al.2004; Chamberlain and Kormanik 1997; Altundag et al.2007). Similarly, in patients with solid parenchymal cere-bral metastases, higher age, low Karnofsky performancestatus, uncontrolled systemic malignant disease and a shortinterval between Wrst diagnosis of primary disease and ofleptomeningeal metastases are known as negative prognos-tic factors on patient’s outcome (Gaspar et al. 1997). How-ever, the inXuence of these above mentioned prognosticfactors was not validated in multivariate Cox regressionanalyses, which might be explained by overlapping eVectsin statistical evaluation.

In multivariate Cox regression analyses, only the typeof treatment was relevant as prognostic factor for sur-vival time. Analogous to previous analysis (Herrlingeret al. 2004; Glantz et al. 1998; Lassman et al. 2006) inour experience, the application of systemic chemother-apy revealed to be the most important positive prognos-tic factor for survival (P = 0.0001), although cytostaticdrugs penetrating the blood-CSF-barrier regularly wereused for systemic chemotherapy in only 23% of patients.This strengthens the hypothesis that the blood-CSF-bar-rier is at least reduced in patients with leptomeningealtumor lesions and cytotoxic agents can penetrate intocerebrospinal Xuid (Bokstein et al. 1998; Boogerd et al.2004). In addition, systemic chemotherapy is not onlytreating the leptomeningeal but also all other extra-leptomeningeal tumor lesions, which might be theexplanation of the signiWcant impact of systemic chemo-therapy on the patient’s prognosis in general. Contrarily,the decision to give systemic chemotherapy might havebeen inXuenced by the patient’s overall performancestatus particularly, introducing a selection bias by thisvariable. Unfortunately, in the underlying analysis, dataon patient’s Karnofsky performance status were avail-able in less than half of patients, which might explain thelack of statistical impact of the performance status onoverall survival in multivariate analysis and limits eval-uation of the impact of selection bias between the appli-cation of systemic chemotherapy and performancestatus.

Subgroup analysis demonstrated that this eVect is mostsigniWcant in patients with breast cancer with 7.0 vs.1.4 months and high-grade non-Hodgkin’s lymphoma (4.4vs. 1.8 months), which present chemotherapy-sensitivemalignant entities. Our data conWrm previous analysesdemonstrating eYcacy of systemic and local chemotherapyin patients with leptomeningeal metastases from breast

cancer (Altundag et al. 2007; Fizazi et al. 1996; Rudnickaet al. 2007) and lymphatic malignancies including high-dose chemotherapy regimen with autologous peripheralstem cell transplantation (Glantz et al. 1999; Gleissner andChamberlain 2007).

In a similar series, study by Herrlinger et al. (2004), ofbreast cancer patients with leptomeningeal metastasesachieved the longest survival times up to 12 months. Theshortest survival times were observed in patients withleptomeningeal metastases of malignant melanoma andlung cancer with an overall survival time of less than5 months (Herrlinger et al. 2004), which is comparable tothe current series with shortest median survival times inpatients with lung cancer (0.8 months) and patients withmalignant melanoma (0.9 months). Especially malignantmelanoma is not chemotherapy sensitive, and a recentevaluation by Raizer et al. (2008) revealed leptomenin-geal metastases as a negative prognostic factor in mela-noma patients.

In contrast to systemic chemotherapy, the impact ofintrathecal chemotherapy in the treatment of leptomenin-geal metastases is discussed controversially. There are ran-domized trials comparing diVerent single or multiple drugintrathecal chemotherapy regimens (Grossman et al. 1993;Kim et al. 2003), but the results are heterogeneous. Sum-marizing the current evidence, the application of intrathecalchemotherapy seems to have has no relevant eVect on sur-vival compared to other treatment modalities, but is associ-ated with an increased rate of therapy-associatedcomplications up to 47% (Boogerd et al. 2004; Grossmanet al. 1993; Kim et al. 2003). Our data strengthen thishypothesis demonstrating no signiWcant impact of intrathe-cal chemotherapy alone on overall survival and limitedeVects on the patient’s symptoms.

In conclusion, age of more than 50, reduced performancestatus <70%, interval between diagnosis of primary tumorand leptomeningeal metastases <12 months, lung cancer ormalignant melanoma as primary tumor and lack of cyto-logic response present negative prognostic factors inpatients with leptomeningeal metastases. The most signiW-cant positive eVect on survival time is caused by the appli-cation of systemic chemotherapy with highest impact inpatients with chemotherapy-sensitive primary tumors, espe-cially breast cancer or high-grade non-Hodgkin’s lym-phoma. The palliative value of treatment modalities aimingto reduce already established neurologic Wndings isextremely modest in our Wndings.

ConXict of interest statement We all declare that none of us hasany Wnancial and personal relationships with other people or organiza-tions that could inappropriately inXuence the work on the underlyingretrospective analysis titled “Prognostic factors and treatment optionsin patients with leptomeningeal metastases of diVerent primarytumors.”

123

J Cancer Res Clin Oncol (2010) 136:1729–1735 1735

References

Altundag K, Bondy ML, Mirza NQ, Kau SW, Broglio K, HortobagyiGN, Rivera E (2007) Clinicopathologic characteristics and prog-nostic factors in 420 metastatic breast cancer patients with centralnervous system metastasis. Cancer 110:2640–2647

Bokstein F, Siegal T, Lossos A (1998) Leptomeningeal metastasesfrom solid tumors, a comparison of two prospective series treatedwith and without intra–cerebrospinal Xuid chemotherapy. Cancer82:1756–1763

Boogerd W, van den Bent MJ, Koehler PJ, Heimans JJ, van der SandeJJ, Aaronson NK, Hart AAM, Benraadt J, Vecht CJ (2004) Therelevance of intraventricular chemotherapy for leptomeningealmetastases in breast cancer: a randomised study. Eur J Cancer40:2726–2733

Chamberlain MC (2005) Meningitis neoplastica. J Clin Oncol23:3605–3613

Chamberlain MC, Kormanik PR (1997) Carcinomatous meningitissecondary to breast cancer: predictors of response to combinedmodality therapy. J Neuro Oncol 35:55–64

De Angeles LM, Boutros D (2005) Leptomeningeal metastases. Can-cer Invest 23:145–154

Fizazi K, Asselain B, Vincent-Salomon A, Jouve M, Dieras V, Palan-gie T, Beuzeboc P, Dorval T, Pouillart P (1996) Meningeal carci-nomatosis in patients with breast carcinoma. Clinical features,prognostic factors, and results of a high-dose intrathecal metho-trexate regimen. Cancer 77:1315–1323

Gaspar L, Scott C, Rotmann M, Asbell S, Phillips T, Wasserman T, McKenna WG, Byhardt R (1997) Recursive partitioning analysis(RPA) of prognostic factors in three radiation therapy oncologygroup (RTOG) brain metastases trials. Int J Radiat Oncol BiolPhys 37:745–751

Glantz MJ, Cole B, Recht L, Akerley W, Mills P, Saris S, Hochberg F,Calabresi P, Egorin MJ (1998) High-dose intravenous methotrex-ate for patients with nonleukemic leptomeningeal cancer: is intra-thecal chemotherapy necessary? J Clin Oncol 16:1561–1567

Glantz MJ, LaFollette S, Jaeckle KA, Shapiro W, Swinnen L, RozentalJR, Phuphanich S, Rogers LR, Gutheil JC, Batchelor T, Lyter D,Chamberlain MC, Maria BL, SchiVer C, Bashir R, Thomas D,Cowens W, Howell SB (1999) Randomized trial of a slow-releaseversus a standard formulation of cytarabin for the intrathecal

treatment of lymphomatous meningitis. J Clin Oncol 17:3110–3116

Gleissner B, Chamberlain M (2007) Treatment of CNS disseminationin systemic lymphoma. J Neurooncol 84:107–117

Grossman SA, Finkelstein DM, Ruckdeschel JC, Trump DL, Moyni-han T, Ettinger DS for the Eastern Cooperative Oncology Group(1993) Randomized prospective comparison of intraventricularmethotrexate and thiotepa in patients with previously untreatedneoplastic meningitis. J Clin Oncol 11:561–569

Herrlinger U, Förschler H, Küker W, Meyermann R, Bamberg M,Dichgans J, Weller M (2004) Leptomeningeal metastasis: sur-vival, prognostic factors in 155 patients. J Neurol Sci 223:167–178

Kaplan EL, Meier P (1958) Non-parametric estimations from incom-plete observations. J Am Stat Assoc 53:457–481

Kim DY, Lee KW, Yun T, Park SR, Jung JY, Kim DW, Kim TY, HeoDS, Bang YJ, Kim NK (2003) Comparison of intrathecal chemo-therapy for leptomeningeal carcinomatosis of a solid tumor:methotrexate alone versus methotrexate in combination withcytosine arabinoside and hydrocortisone. Jpn J Clin Oncol33:608–612

Lassman AB, Abrey LE, Shah GG, Panageas KS, Begemann M, Mal-kin MG, Raizer JJ (2006) Systemic high–dose intravenous meth-otrexate for central nervous system metastases. J Neurooncol78:255–260

Pentheroudakis G, Pavlides N (2005) Management of leptomeningealmalignancy. Expert Opin Pharmacother 6:1115–1125

Raizer JJ, Hwu WJ, Panageas KS, Wilton A, Baldwin DE, Bailey E,von Althann C, Lamb LA, Alvarado G, Bilsky MH, Gutin PH(2008) Brain and leptomeningeal metastases from cutaneous mel-anoma: survival outcomes based on clinical features. Neuro On-col 10:199–207

Rudnicka H, Niwijska A, Murawska M (2007) Breast cancer leptome-ningeal metastasis—the role of multimodality treatment. J NeuroOncol 84:57–62

Siegal T, Lossos A, PfeVer MR (1994) Leptomeningeal metastases:analysis of 31 patients with sustained oV-therapy response follow-ing combined modality treatment. Neurology 44:1463–1469

Taillibert S, Jerzy H (2006) Treatment of central nervos system metas-tases: parenchymal, epidural, and leptomeningeal. Curr Opin On-col 18:637–643

123