Clin Exp Metastasis (2015) 32:383–391 DOI 10.1007/s10585-015-9716-3
RESEARCH PAPER
Diagnostic and prognostic significance of flow cytometry immunophenotyping in patients with leptomeningeal carcinomatosis D. Subira´1 • M. Simo´2 • J. Illa´n3 • C. Serrano4 • S. Castan˜o´n4 • R. Gonzalo4 • J. J. Granizo5 • M. Martı´nez-Garcı´a6 • M. Navarro7 • J. Pardo8 • J. Bruna2
Received: 2 February 2015 / Accepted: 16 March 2015 / Published online: 21 March 2015 Ó Springer Science+Business Media Dordrecht 2015
Abstract Some patients with epithelial-cell cancers develop leptomeningeal carcinomatosis (LC), a severe complication difficult to diagnose and with an adverse prognosis. This study explores the contribution of flow cytometry immunophenotyping (FCI) to the diagnosis and prognosis of LC. Cerebrospinal fluid (CSF) samples from patients diagnosed with LC were studied using FCI. Expression of the epithelial-cell adhesion molecule (EpCAM) was the criterion used to identify the epithelial cells. To test the diagnostic precision, 144 patients (94
diagnosed with LC) were included. The prognostic value of FCI was evaluated in 72 patients diagnosed with LC and eligible for therapy. Compared with cytology, FCI showed greater sensitivity and negative predictive value (79.79 vs. 50 %; 68.85 vs. 51.55 %, respectively), but lower specificity and positive predictive value (84 vs. 100 %; 90.36 vs. 100 %, respectively). The multivariate analysis revealed that the percentage of CSF EpCAM? cells predicted an increased risk of death (HR: 1.012, 95 % CI 1.000–1.023; p = 0.041). A cut-off value of 8 %
& D. Subira´ [email protected]
1
Department of Hematology, Flow Cytometry Division, Hospital Universitario de Guadalajara, c/Donantes de sangre s.n., 19002 Guadalajara, Spain
2
Unit of Neuro-Oncology, Departments of Oncology and Neurology, Hospital Universitario de Bellvitge-ICO Duran i Reynals, Avda. Gran Vı´a s/n km 2.7, Hospitalet de Llobregat, 08907 Barcelona, Spain
3
Unilabs Diagno´sticos, SLU, c/Juan Esplandiu´ 15, 28007 Madrid, Spain
4
Department of Hematology, Fundacio´n Jime´nez Dı´az, Plaza Cristo Rey 1, 28040 Madrid, Spain
5
Clinical Epidemiology Unit, Hospital Infanta Cristina, Avda. 9 de Junio, 2, Parla, 28981 Madrid, Spain
6
Department of Oncology, Hospital del Mar, Paseo Marı´timo 25-29, 8003 Barcelona, Spain
7
Department of Oncology, Hospital Universitario de Salamanca, Instituto de Investigacio´n Biome´dica de Salamanca (IBSAL), Paseo de San Vicente, 58-182, Salamanca 37007, Spain
8
Department of Neurology, Hospital Infanta Elena/Hospital General de Villalba/Hospital Universitario Rey Juan Carlos, c/Gladiolo s/n., 28933 Mo´stoles, Spain
M. Simo´ [email protected] J. Illa´n [email protected] C. Serrano [email protected] S. Castan˜o´n [email protected] R. Gonzalo [email protected] J. J. Granizo [email protected] M. Martı´nez-Garcı´a [email protected] M. Navarro [email protected] J. Pardo [email protected] J. Bruna [email protected]
123
384
EpCAM? cells in the CSF distinguished two groups of patients with statistically significant differences in overall survival (OS) (p = 0.018). This cut-off value kept its statistical significance regardless of the absolute CSF cell-count. The FCI study of the CSF improved the sensitivity for diagnosing LC, but refinement of the technique is needed to improve specificity. Furthermore, quantification of CSF EpCAM? cells was revealed to be an independent prognostic factor for OS in patients with LC eligible for therapy. An 8 % cut-off value contributed to predicting clinical evolution before initiation of therapy. Keywords Flow cytometry Immunophenotype Cerebrospinal fluid Leptomeningeal disease Diagnosis Prognosis Abbreviations CSF Cerebrospinal fluid EpCAM Epithelial-cell adhesion molecule FCI Flow cytometry immunophenotyping IQR Interquartile range IT Intrathecal chemotherapy KPS Karnofsky performance status scale MRI Magnetic resonance imaging NPV Negative predictive value LC Leptomeningeal carcinomatosis OS Overall survival PPV Positive predictive value RDT Radiotherapy SC Systemic chemotherapy
Introduction Neoplastic meningitis is a devastating complication that arises in 5–15 % of patients with cancer, and which involves a poor prognosis. Overall survival (OS) has been associated with several clinical, tumor, and cerebrospinal fluid (CSF) features, sometimes with contradictory findings due to the limitations of these studies [1, 2]. The most widely accepted prognostic factors are performance status, histology, and degree of neurological deficits. Early detection and start of therapy might prevent the onset of irreversible symptoms and therefore prolong survival with an acceptable quality of life [1, 2]. Once clinical symptoms appear, the diagnosis of neoplastic meningitis relies on the cytological identification of malignant cells in the CSF, and/or compatible magnetic resonance imaging (MRI) findings and/or suggestive biochemical CSF findings [3]. However, the sensitivity of cytology and the specificity of MRI are not optimal [1, 2, 4, 5]. Moreover, incidental MRI features suggestive of meningeal dissemination can be found in asymptomatic and paucysymptomatic patients, increasing the diagnostic
123
Clin Exp Metastasis (2015) 32:383–391
challenge. Therefore, several studies have focused on improving the efficiency of the available diagnostic tools, and on introducing new potentially useful biomarkers [6, 7]. Flow cytometry immunophenotyping (FCI) has been used for detecting neoplastic cells in the CSF, but most studies have focused on patients with hematological malignancies [8–13]. However, in our experience, FCI can also improve the sensitivity of conventional CSF cytology in diagnosing leptomeningeal involvement from epithelialcell neoplasias (leptomeningeal carcinomatosis, LC) [14]. In this study, the first goal was to determine whether our previous findings regarding better sensitivity and negative predictive value of FCI compared to CSF cytology could be validated in a larger series of patients with epithelialcell tumors. As a second step, we also explored whether FCI data might offer any prognostic information.
Materials and methods Study design To test the diagnostic precision of FCI, 166 patients with clinical data suspicious of LC from epithelial-cell malignancies were enrolled between January 2011 and December 2012 in 14 Spanish hospitals. Sensitivity, specificity, and the positive predictive value (PPV) and negative predictive value (NPV) for diagnosing LC were calculated for CSF cytology, FCI and MRI. These data were compared with our previous report on a series of 78 patients with cancer who were recruited between 2009 and 2010 [14]. To explore the prognostic value of CSF FCI data, we monitored all patients diagnosed with LC recruited during the 2-year period of our study: from March 2009 to July 2010 [14], and from January 2011 to December 2012 (validation study). Only patients who were candidates to receive therapy were selected and OS was registered. Criteria for exclusion were: patients eligible only for palliative care, those without a complete collection of clinical data, and those whose CSF samples had macroscopic blood contamination after sample centrifugation. In the end, 72 patients were included (Fig. 1). Demographic data collected were age, sex, Karnofsky performance status scale (KPS), localization and histology of primary tumor, time from diagnosis of primary tumor to development of LC, presence of brain metastasis, and therapy received. The CSF parameters obtained with FCI analysis were the absolute number of CSF cells, detection of epithelial cells, percentage of epithelial cells, and percentage and absolute number of accompanying inflammatory cells (lymphocytes, monocytes, and polymorphonuclear cells). Additional CSF parameters such as cytology results, and concentration of glucose and proteins, were also evaluated.
Clin Exp Metastasis (2015) 32:383–391
385
January 2011 to December 2012 Patients with clinical suspicious of leptomeningeal carcinomatosis n=166
Patients excluded n=22 • Incomplete data or no information
Validation study n=144
Diagnosed with leptomeningeal carcinomatosis n=94
Leptomeningeal carcinomatosis excluded (other diagnoses established) n=50
Diagnosed with leptomeningeal carcinomatosis n=48
March 2009 to July 2010 Clinical suspicion of leptomeningeal carcinomatosis n=78
Diagnosed with leptomeningeal carcinomatosis n=142 Patients excluded n=70 • Not eligible for therapy n=38 • Incomplete or no follow-up data n=15 • Macroscopic blood contamination of the cerebrospinal fluid n=17 Prognostic study n=72
Leptomeningeal carcinomatosis eligible for therapy n=72
Fig. 1 Summary of the criteria used for the final inclusion of patients in the validation and prognostic studies
Local ethics committees of participating centers approved the study, and informed consent was obtained from each participant before their enrollment. All procedures were in accordance with the ethical standards of the Helsinki Declaration. FCI studies CSF samples were obtained with lumbar puncture or from an Ommaya reservoir and collected in tubes with EDTA containing an immunofixative reagent (TransFix, Cytomark), for safe transportation [15] from the hospital of origin to the Fundacio´n Jime´nez Dı´az (Madrid, Spain), where all FCI studies were performed. A median of 3 days from extraction to study was recorded. Samples were processed using a previously reported FCI protocol for CSF samples with suspected LC [14]. Briefly, for cell count, an aliquot of the CSF sample (100 ll) was added to the fluorescent dye DRAQ5 (100 ll previously diluted 1:1000) for DNA staining (Biostatus Limited) and Perfect-count microspheres (Cytognos SL, Salamanca, Spain). The remaining CSF sample was centrifuged and the
cell-pellet was stained with a 2-color (fluorescein isothyocyanate, FITC/phycoerythrin, PE) monoclonal antibody (mAb) combination directed against the epithelial-cell adhesion molecule (EpCAM), [clones BerEP4 from DAKO; and EBA-1 from Becton–Dickinson Biosciences (BDB)]. After 20 min of incubation at room temperature in darkness, samples were washed once, and the cell-pellet was resuspended in phosphate buffered saline (PBS). Then, DRAQ5 was added and after 10 min of incubation, the whole volume of sample was acquired on a FACSCanto II flow cytometer (BDB) using the FACSDiva software (version 6.1). The FSC area/FSC height dot-plot was used to help exclude aspirated air and doublets. Analyses were performed using the INFINICYT software program (Cytognos SL, Salamanca, Spain). Cells were separated from debris using positive staining with DRAQ5, and then epithelial and inflammatory cells were identified. For both diagnostic and prognostic studies, the FCI criterion used to define a sample as positive for malignancy was a cluster of at least 16 events (modified from Subira´ et al. [16]) positive for EpCAM expression with the 2 mAb used (EpCAM positive cells). The result was expressed as percentage of
123
386
epithelial cells referred to the total CSF cells. Non-malignant inflammatory cells were identified on the basis of their forward scatter (FSC) and side-scatter (SSC) characteristics. Lymphocytes were FSClow/SSClow, monocytes FSChigh/SSCintermediate/high, and polymorphonuclear cells FSClow/SSChigh. The relative percentage of each population was referred to the total inflammatory cells (EpCAM negative cells). Statistics Qualitative variables were expressed as median and interquartile range. Quantitative variables were expressed as percentage. The usefulness of cytology, FCI and MRI to diagnose LC was compared by calculating the sensitivity, specificity, PPV and NPV. OS analysis was determined by constructing probability curves according to the Kaplan–Meier method and comparing them using a log-rank test. Variables that had a p value \ 0.05 in the univariate analysis were subsequently put in a backward stepwise proportional-hazard analysis (Multivariate Cox regression model) to explore their independent value as predictor of survival. The cut-off used for the continuous variables was their median value. A receiver-operator characteristics (ROC) curve analysis was applied to determine the best cut-off value to assess the prognostic significance of EpCAM? cells. Calculations were performed using SPSS version 14.0 (SPSS Inc, Chicago, IL).
Results Validation study Appropriate clinical data were obtained in 144/166 patients (Fig. 1). The diagnosis of LC was established on the basis of CSF cytology, and/or compatible clinical signs plus MRI findings with biochemical CSF abnormalities. LC was established in 94 patients (61.7 % female) with a median age of 58 years (Interquartile range, IQR 48–66). The distribution of primary tumors was as follows: breast (n = 39), lung (n = 35), gastrointestinal (n = 6), ovarian (n = 4), prostate (n = 3), and others (n = 5). Two and three neoplasms (breast/lung, and gastrointestinal/breast/lung) were simultaneously diagnosed in 2 patients. The most common histology subtype was adenocarcinoma (85.1 %). Forty-six patients (48.94 %) had a KPS value C70 % and 39 (41.49 %) associated parenchymal metastases. A positive cytological detection of malignant cells in the CSF was found in 47 of the 94 patients diagnosed with LC. MRI data were collected in 90 patients and findings were consistent with LC in 68 patients (35 with negative CSF
123
Clin Exp Metastasis (2015) 32:383–391
cytology). Malignant cells were detected with FCI in 75 CSF samples, including 8 out of the 11 samples from patients with both negative CSF cytology and MRI findings (Fig. 2). Sensitivity, specificity, PPV and NPV were calculated for cytology, MRI and FCI (Table 1). FCI showed better sensitivity and NPV compared with cytology (79.79 vs. 50 % and 68.85 vs. 51.55 %, respectively). In contrast, compared to cytology, FCI showed lower specificity and PPV (84 % and 90.36 %, respectively vs. 100 %). MRI (performed in 129 patients) showed 75.56 % sensitivity, 63.04 % specificity, 87.18 % PPV, and 58.86 % NPV. Prognostic study Patient characteristics Seventy-two patients (68.05 % female), with a median age of 59.5 years (IQR 50–66), were eligible for therapy and recruited for the prognostic study. Median KPS score was 70 % (IQR 60–80). The modalities of therapy used were: intrathecal chemotherapy, ITC (n = 33), with (n = 2) or without radiotherapy (RDT, focal spinal or whole brain) (n = 31), systemic chemotherapy, SC (n = 13), with (n = 3) or without RDT (n = 10), whole brain RDT (n = 5), and combination of ITC plus SC (n = 21), with (n = 9) or without RDT (n = 12). All patients included in the study received at least 1 dose of IT. Primary tumors were distributed as follows: breast (n = 35), lung (n = 24), gastrointestinal (n = 5), ovarian (n = 3), and others (n = 2). In the remaining 3 patients, 2 neoplasms (breast/gastrointestinal, n = 2; breast/lung, n = 1) were simultaneously diagnosed. The most common histology subtype was adenocarcinoma (94.44 %). Median time between diagnosis of primary tumor and development of LC was 2.41 years (IQR 0.56–5.54). The diagnosis of primary neoplasia was synchronic with diagnosis of LC in 9 patients (12.5 %). Concurrent brain metastases were detected in 34 of the 69 patients who could be evaluated with MRI. Demographics and CSF data are summarized in Table 2. At the end of the study, after a median follow-up of 46 weeks (range 30.3–146), 8 patients were still alive. Median OS of the entire cohort of patients was 11 weeks (range 1–146.7 weeks). Depending on the therapy received, the median OS was 3.29 weeks for patients with whole brain RDT, 10.14 weeks with SC (±RDT), 10.43 weeks with ITC (±RDT), and 35 weeks for patients receiving ITC ? SC (±RDT). CSF data Median CSF sample volume was 2.6 ml (IQR 2.1–2.8), and median CSF cell count was 6 cells/ll (IQR 2.3–17.8). Cytology identified malignant cells in 35 CSF samples
Clin Exp Metastasis (2015) 32:383–391
387
Flow cytometry immunophenotyping
Cytology
11
Diagnostic tools for leptomeningeal carcinomatosis Cytology/Magnetic resonance imaging Cytology/Flow cytometry immunophenotyping Magnetic resonance imaging/Flow cytometry immunophenotyping Cytology/Magnetic resonance imaging/Flow cytometry immunophenotyping
8
% of cases detected 87.77 82.22 96.66 96.66
31 2
22
*3
13
Magnetic resonance imaging Fig. 2 Distribution of the cerebrospinal fluid samples studied in the validation cohort according to the results of cytology, magnetic resonance imaging and flow cytometry immunophenotyping. Positive cases for leptomeningeal carcinomatosis of each technique are specified. Data shown belong to the 90 patients with available
Table 1 Sensitivity, specificity, positive and negative predictive value for the diagnosis of leptomeningeal carcinomatosis according to cerebrospinal fluid cytology, flow cytometry immunophenotyping and magnetic resonance imaging Cytology (n = 144)
Magnetic resonance imaging (n = 129)
Flow cytometry immunophenotyping (n = 144)
Sensitivity
50.00
75.56
79.79
Specificity
100
63.04
84.00
Positive predictive value
100
87.18
90.36
Negative predictive value
51.55
56.86
68.85
(48.6 %), and FCI detected EpCAM? cells in 56/72 cases (77.7 %). Median percentage of EpCAM? cells was 2.25 % (IQR 0.2–18.6). Median number of EpCAM? cells/ll of CSF was 0.26 (IQR 0.01–2.21). Lymphocytes (median value 56 %) were the predominant population in 40 samples, and the percentage of monocytes exceeded 30 % in 36 samples (50 %). Polymorphonuclear cells were described in 49/72 samples (68 %), and their percentage was [1 % of the inflammatory cell compartment in 37/72 CSF samples (51.4 %). Survival prognosis analysis The univariate analysis revealed that age, sex, histology, localization of primary tumor, time between primary tumor and LC, and presence or absence of parenchymal
cytology, flow cytometry immunophenotyping, and magnetic resonance imaging data. The 4 patients in whom magnetic resonance imaging was not performed are excluded from this diagram. *Patients with negative cytology, flow cytometry immunophenotyping and magnetic resonance imaging
metastases had no impact on OS. Conversely, a KPS value C70 % indicated favorable outcome (p = 0.001; HR = 0.968, 95 % CI 0.950–0.987), and treatment with ITC (±RDT) (p = 0.01) or ITC ? SC (±RDT) (p = 0.001) presented a significantly better OS compared to patients treated only with whole brain RDT. Comparison of the remaining modalities of therapy (SC ± RDT, ITC ± RDT, and ITC ? SC ± RDT) did not show statistical differences in OS. Regarding CSF variables, the univariate analysis showed that the percentage of CSF EpCAM? cells (p = 0.049), the absolute number of EpCAM? cells/ll (p = 0.011), number of monocytes/ll (p = 0.018), and number of PMN cells/ll (p = 0.001) were associated with worse OS. However, no statistical differences in OS were observed after comparing patients with positive versus negative CSF cytology, and positive versus negative CSF FCI findings. Also, neither the absolute number of CSF cells nor the quantitative value of CSF proteins and glucose had any impact on OS. In addition, using the cut-off value of 48.5 mg/dl CSF glucose (2.7 mol/l), as previously reported, was an independent prognostic factor [17], and a trend towards significance was observed (p = 0.064). The multivariate analysis, performed using all the factors that were significant in the univariate analysis, revealed that a median KPS value \70 % (HR = 1.033, 95 % CI 1.013–1.053; p = 0.001), and receiving whole brain RDT (HR = 3.963, 95 % CI 1.445–10.871; p = 0.007) instead of any chemotherapy treatment, predicted poor survival. No statistical differences were observed after comparison of patients receiving ITC versus
123
388
Clin Exp Metastasis (2015) 32:383–391
Table 2 Clinical data and cerebrospinal fluid characteristics of patients included in the prognostic study group
Patients enrolled Male:female
n
%
72
100
59.5 (50–66)
Median interval in years from primary diagnosis to leptomeningeal carcinomatosis (interquartile range)
2.41 (0.56–5.54)
Adenocarcinoma
68
94.44
Breast
35
48.61
Lung
24
33.33
Gastrointestinal
5
6.94
Ovary
3
4.17
Bladder
1
1.39
Kidney
1
1.39
Primary tumor
Breast ? gastrointestinal
2
2.77
Breast ? lung
1
1.39
Yes
34
47.22
No
35
48.61
3
4.17
\70 %
29
40.28
C70 %
42
58.33
Unknown
1
1.39
72
100
Central nervous system metastasis
Unknown Karnofsky performance status
Susceptible to therapy Therapy received Whole brain radiotherapy
5
6.94
Intrathecal therapy alone
31
43.06
Intrathecal therapy ? focal spinal or whole brain radiotherapy
2
2.78
Systemic chemotherapy alone
10
13.89
Systemic chemotherapy ? focal spinal or whole brain radiotherapy
3
4.17
Intrathecal therapy ? systemic chemotherapy
12
16.67
Intrathecal therapy ? systemic chemotherapy ? focal spinal or whole brain radiotherapy
9
12.5
After identification of the percentage of CSF EpCAM? cells useful as a parameter related to OS in patients with LC receiving therapy, we looked for an optimal cut-off value. The receiver-operator characteristic curve (ROC) showed that when using a threshold value of 8 % EpCAM? cells, 2 groups of patients with statistically significant differences in OS were identified (p = 0.018) (Fig. 3). Patients with \8 % EpCAM? cells (n = 47) had a median OS of 24.57 versus 7.71 weeks for patients with C8 % EpCAM? cells (n = 25). The median age and KPS value were 61 years and 70 % for patients with \8 % EpCAM? cells, and 56 years and 60 % for patients with C8 % EpCAM? cells. The number of patients with breast and lung cancer was 20 (42.5 %) and 17 (31.9 %), respectively, in the group of\8 % EpCAM? cells, and 15 (60 %) and 3 (12 %) in the group of C8 % EpCAM? cells. Considering that the absolute number of cells in the CSF of patients diagnosed with LC is highly variable, we explored whether the cut-off value of 8 % EpCAM? cells might be useful to predict OS in patients with a normal or high CSF cell count. No differences in OS were observed after comparison of CSF samples with [5 and B5 cells/ll (p = 0.432). In contrast, the variable percentage of EpCAM? cells (\8 and C8 %) kept its statistical significance after being stratified according to the absolute number of CSF cells (p = 0.020).
Discussion
Median value (interquartile range) Cerebrospinal fluid data 2.6 (2.1–2.8)
Cell count per ml
6 (2.3–17.8)
Proteins in mg/dl Glucose in mg/dl
58 (38–76.5) 73 (47–170)
SC (p = 0.770), ITC versus combination of ITC ? SC (p = 0.153), and SC versus the combination ITC ? SC (p = 0.158). The only CSF parameter that retained
123
Influence of EpCAM? cells and OS
23:49
Median age (interquartile range)
Sample volume in ml
statistical significance for an increased risk of death in the multivariate analysis was the percentage of EpCAM? cells (HR = 1.012, 95 % CI 1.000–1.023; p = 0.041).
The increasing survival of cancer patients implies an increased risk of developing LC [18]. However, diagnosing LC is still a challenge, especially in the early stages. New methods that can help physicians diagnose the disease are under development, but CSF cytology remains the keystone for diagnosing LC because it identifies malignant cells in the CSF. The rationale in using FCI for the diagnosis of LC in patients with epithelial-cell tumors is also its ability to identify epithelial cells in the CSF, using surface expression of the epithelial-cell antigen EpCAM as the marker of this population [14, 19–24]. A similar strategy is used for the detection of circulating tumor cells, and this recently was used to detect tumor cells in the CSF [25, 26]. However, at the time of writing, clinical laboratories are more familiar with the use of FCI techniques, and it is available in many centers. In our study we tested whether the FCI protocol for detecting EpCAM? cells in the CSF [14] was useful to identify
Clin Exp Metastasis (2015) 32:383–391
389
Fig. 3 a Kaplan–Meier Curves for overall survival according to the 8 % cut-off level of cerebrospinal fluid EpCAM? cells. Median overall survival was 24.57 versus 7.71 weeks for patients with\8 and C8 % EpCAM? cells, respectively. A representative example of 2
cerebrospinal fluid samples is shown, one with a low cell-count and C8 % EpCAM? cells (b), and another with a high cell-count and \8 % EpCAM? cells (c)
LC in a larger validation cohort of patients diagnosed with epithelial-cell tumors. In accordance with our previous findings, the CSF FCI study improved the sensitivity of classical CSF cytology. However, using FCI to analyze the CSF did not eliminate the false negative cases. It is likely that some of the reasons given to explain negative cytology results may also justify negative FCI results: negative or transient seeding in the CSF and suboptimal collection with samples obtained far from the site of origin of clinical symptoms [1, 2, 4, 5]. In addition, down regulation of EpCAM expression described in some aggressive tumors [27] and epithelial-mesenchymal-transition should also be considered [28]. In order to solve these problems, future studies will need to use more powerful flow cytometers capable of evaluating nuclear dyes in combination with a greater number of cell surface antigens, including different epithelial-cell markers. This will help improve the specificity and PPV of flow cytometry studies, even when data are better than those described for MRI. In the meantime, the inclusion of FCI among the LC diagnostic procedures adds specificity to MRI data, and, combined with CSF cytology, significantly decreases the number of undiagnosed cases. Regarding the prognostic study, and in keeping with previous reports, our series also confirmed that a good performance status remains as an independent factor associated with better survival [1, 2, 4, 5, 29–32]. The levels of glucose in the CSF showed a trend towards significance [17], and receiving only whole brain RDT was the only therapeutic choice associated with a bad prognosis. There was no advantage for survival after comparing the different therapeutic options but this fact should be interpreted with caution since this was not a randomized trial evaluating the efficacy of different therapies.
The novelty of this study is the identification of quantification of CSF EpCAM? cells as an independent prognostic factor for patients diagnosed with LC who are to be treated with any of the currently available therapeutic modalities. The prognostic value of the CSF FCI data has been studied in patients diagnosed with high grade B cell lymphoma, but most authors have focused on determining whether occult leptomeningeal disease (defined as positive FCI, negative cytology) might be an indicator of central nervous system relapse [8–13]. In our study, and in keeping with previous reports [33] concerning epithelial-cell neoplasms, this ‘qualitative’ information (positive vs. negative cytology and positive vs. negative FCI data) had no prognostic value. In contrast, we found an association between higher percentages of CSF EpCAM? cells and a shorter OS. To the best of our knowledge, this is the first study to propose using a cut-off level of malignant cell infiltration in the CSF to stratify the OS in patients with leptomeningeal involvement. The cut-off value of C8 % EpCAM? cells identified patients diagnosed with LC with a two-fold risk of death. This correlation between percentage of CSF EpCAM? cells and OS is probably a measurement of disease burden in CSF at the time of diagnosing LC (early or advanced disease). A recent study by Mun˜iz et al. [34], describing a clear association between higher sCD19 (a soluble B-cell specific protein) CSF levels and OS in patients with diffuse large B-cell lymphoma and Burkitt lymphoma, also supports this hypothesis. In cancer patients, leptomeningeal neoplastic cell involvement is usually associated with advanced systemic disease [1, 2], but our observations also support the hypothesis that tumor burden in the CSF affects the patient’s prognosis.
123
390
Clin Exp Metastasis (2015) 32:383–391
From a technical point of view, the 8 % threshold value is high enough to be easily implemented in clinical flow cytometry laboratories. It also has the advantage of maintaining its prognostic value independently of the total number of CSF cells, considering the wide range of values that can be found in patients with LC. This would mean that, for every case, the really important measurement is the relationship between malignant and inflammatory cells. However, none of the inflammatory cell populations evaluated in the multivariate analysis was associated with differing rates of survival among patients with LC. In a previous study, we reported that the percentage of polymorphonuclear cells in the CSF was the only significant distinctive parameter between the inflammatory cell compartment of cancer patients with and without LC [35]. In the present study, the percentage of polymorphonuclear cells was cited in the univariate analysis, but lacked statistical significance in the multivariate analysis. In summary, an initial evaluation of the CSF using FCI is an easy way to obtain both diagnostic and prognostic information about patients with LC. On the one hand, FCI improves the sensitivity for the diagnosis, and on the other, quantification of the percentage of CSF EpCAM? cells is prognostic of OS. According to our data, a cut-off value of 8 % CSF EpCAM? cells is a useful variable to help physicians select patients who might benefit from therapy and to stratify patients in treatment trials. This finding needs to be externally validated and its clinical value has to be verified with additional clinical studies. Acknowledgments The authors would like to thank all the patients and participants of the Meningeal Carcinomatosis Study Cooperative Group who made this study possible. We also wish to thank Mundipharma Spain for supporting the study, and Lawrence J.C. Baron for editing the manuscript. Conflict of interest of interest.
4. 5.
6.
7.
8.
9.
10.
11.
The authors declare that they have no conflict
Ethical standards Local ethics committees of participating centers approved the study, and all patients gave their informed consent prior to their inclusion in the study. All procedures were performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. Funding Mundipharma contributed financial support for sample transportation.
12.
13.
References 1. Gleissner B, Chamberlain MC (2006) Neoplastic meningitis. Lancet Neurol 5:443–452 2. Chamberlain MC (2010) Leptomeningeal metastasis. Curr Opin Oncol 22:627–635. doi:10.1097/CCO.0b013e32833de986 3. Brem SS, Bierman PJ, Brem H, Butowski N, Chamberlain MC, Chiocca EA, DeAngelis LM, Fenstermaker RA, Friedman A, Gilbert MR, Hesser D, Junck L, Linette GP, Loeffler JS, Maor
123
14.
MH, Michael M, Moots PL, Morrison T, Mrugala M, Nabors LB, Newton HB, Portnow J, Raizer JJ, Recht L, Shrieve DC, Sills AK Jr, Vrionis FD, Wen PY, National Comprehensive Cancer Network (2011) Central nervous system cancers. J Natl Compr Cancer Netw 9:352–400 Bruna J, Simo´ M, Velasco R (2012) Leptomeningeal metastases. Curr Treat Options Neurol 14:402–415 Strik H, Pro¨mmel P (2010) Diagnosis and individualized therapy of neoplastic meningitis. Expert Rev Anticancer Ther 10:1137–1148 Groves MD, Hess KR, Puduvalli VK, Colman H, Conrad CA, Gilbert MR, Weinberg J, Cristofanilli M, Yung WKA, Liu TJ (2009) Biomarkers of disease: cerebrospinal fluid vascular endothelial growth factor (VEGF) and stromal cell derived factor (SDF)-1 levels in patients with neoplastic meningitis (NM) due to breast cancer, lung cancer and melanoma. J Neurooncol 94:229–234 Herrlinger U, Wiendl H, Renninger M, Fo¨rschler H, Dichgans J, Weller M (2004) Vascular endothelial growth factor (VEGF) in leptomeningeal metastasis: diagnostic and prognostic value. Br J Cancer 91:219–224 Hegde U, Filie A, Little RF, Janik JE, Grant N, Steinberg SM, Dunleavy K, Jaffe ES, Abati A, Stetler-Stevenson M, Wilson WH (2005) High incidence of occult leptomeningeal disease detected by flow cytometry in newly diagnosed aggressive B-cell lymphomas at risk for central nervous system involvement: the role of flow cytometry versus cytology. Blood 105:496–502 Subira´ D, Go´rgolas M, Castan˜o´n S, Serrano C, Roma´n A, Rivas F, Toma´s JF (2005) Advantages of flow cytometry immunophenotyping for the diagnosis of central nervous system nonHodgkin’s lymphoma in AIDS patients. HIV Med 6:21–26 Benevolo G, Stacchini A, Spina M, Ferreri AJ, Arras M, Bellio L, Botto B, Bulian P, Cantonetti M, Depaoli L, Di Renzo N, Di Rocco A, Evangelista A, Franceschetti S, Godio L, Mannelli F, Pavone V, Pioltelli P, Vitolo U, Pogliani EM, Linfomi Fondazione Italiana (2012) Final results of a multicenter trial addressing role of CSF flow cytometric analysis in NHL patients at high risk for CNS dissemination. Blood 120:3222–3228. doi:10. 1182/blood-2012-04-423095 Epub 2012 Aug 27 Sancho JM, Orfao A, Quijano S, Garcı´a O, Panizo C, Pe´rez-Ceballos E, Deben G, Salar A, Gonza´lez-Barca E, Alonso N, Garcı´aVela JA, Capote J, Pen˜alver FJ, Provencio M, Arias J, Plaza J, Caballero D, Morado M, Feliu E, Ribera JM, Spanish Group for the Study of CNS Disease in NHL (2010) Clinical significance of occult cerebrospinal fluid involvement assessed by flow cytometry in non-Hodgkin’s lymphoma patients at high risk of central nervous system disease in the rituximab era. Eur J Haematol 85:321–328. doi:10.1111/j.1600-0609.2010.01478.x Epub 2010 Jul 28 Alvarez R, Dupuis J, Plonquet A, Christov C, Copie-Bergman C, Hemery F, Gaillard I, El Gnaoui T, Kuhnowski F, Bedoui M, Belhadj K, Brugie`res P, Haioun C (2012) Clinical relevance of flow cytometric immunophenotyping of the cerebrospinal fluid in patients with diffuse large B-cell lymphoma. Ann Oncol 23:1274–1279 Wilson WH, Bromberg JE, Stetler-Stevenson M, Steinberg SM, Martin-Martin L, Mun˜iz C, Sancho JM, Caballero MD, Davidis MA, Brooimans RA, Sanchez-Gonzalez B, Salar A, Gonza´lezBarca E, Ribera JM, Shovlin M, Filie A, Dunleavy K, Mehrling T, Spina M, Orfao A (2014) Detection and outcome of occult leptomeningeal disease in diffuse large B-cell lymphoma and Burkitt lymphoma. Haematologica 99:1228–1235 Subira´ D, Serrano C, Castan˜o´n S, Gonzalo R, Illa´n J, Pardo J, Martı´nez-Garcı´a M, Millastre E, Aparisi F, Navarro M, Do´mine M, Gil-Bazo I, Pe´rez Segura P, Gil M, Bruna J (2012) Role of flow cytometry immunophenotyping in the diagnosis of leptomeningeal carcinomatosis. NeuroOncol 14:43–52
Clin Exp Metastasis (2015) 32:383–391 15. de Jongste AH, Kraan J, Broek PD, Brooimans RA, Bromberg JE, Montfort KA, Smitt PA, Gratama JW (2014) Use of TransFixTM cerebrospinal fluid storage tubes prevents cellular loss and enhances flow cytometric detection of malignant hematological cells after 18 hours of storage. Cytom B Clin Cytom 86:272–279 16. Subira´ D, Castan˜o´n S, Aceituno E, Herna´ndez J, Jime´nez-Garo´fano C, Jime´nez A, Jime´nez AM, Roma´n A, Orfao A (2002) Flow cytometric analysis of cerebrospinal fluid samples and its usefulness in routine clinical practice. Am J Clin Pathol 117:952–958 17. Bruna J, Gonza´lez L, Miro´ J, Velasco R, Gil M, Tortosa A (2009) Leptomeningeal carcinomatosis: prognostic implications of clinical and cerebrospinal fluid features. Cancer 115:381–389. doi:10.1002/cncr.24041 18. Glass JP, Melamed M, Chernik NL, Posner JB (1979) Malignant cells in cerebrospinal fluid (CSF): the meaning of a positive CSF cytology. Neurology 29:1369–1375 19. Chang A, Benda PM, Wood BL, Kussick SJ (2003) Lineagespecific identification of nonhematopoietic neoplasms by flow cytometry. Am J Clin Pathol 119:643–655 20. Pillai V, Cibas ES, Dorfman DM (2013) A simplified flow cytometric immunophenotyping procedure for the diagnosis of effusions caused by epithelial malignancies. Am J Clin Pathol 139:672–681. doi:10.1309/AJCP4HIFSHMO9WTK 21. Davidson B, Dong HP, Holth A, Berner A, Risberg B (2007) Flow cytometric immunophenotyping of cancer cells in effusion specimens: diagnostic and research applications. Diagn Cytopathol 35:568–578 22. Kentrou NA, Tsagarakis NJ, Tzanetou K, Damala M, Papadimitriou KA, Skoumi D, Stratigaki A, Anagnostopoulos NI, Malamou-Lada E, Athanassiadou P, Paterakis G (2011) An improved flow cytometric assay for detection and discrimination between malignant cells and atypical mesothelial cells, in serous cavity effusions. Cytom B Clin Cytom 80:324–334. doi:10.1002/ cyto.b.20608 Epub 2011 Jun 21 23. Krishan A, Ganjei-Azar P, Hamelik R, Sharma D, Reis I, Nadji M (2010) Flow immunocytochemistry of marker expression in cells from body cavity fluids. Cytom A 77:132–143. doi:10.1002/cyto. a.20824 24. Risberg B, Davidson B, Dong HP, Nesland JM, Berner A (2000) Flow cytometric immunophenotyping of serous effusions and peritoneal washings: comparison with immunocytochemistry and morphological findings. J Clin Pathol 53:513–517. doi:10.1136/ jcp.53.7.513 25. Patel AS, Allen JE, Dicker DT, Peters KL, Sheehan JM, Glantz MJ, El-Deiry WS (2011) Identification and enumeration of circulating tumor cells in the cerebrospinal fluid of breast cancer patients with central nervous system metastases. Oncotarget 2:752–760
391 26. Nayak L, Fleisher M, Gonzalez-Espinoza R, Lin O, Panageas K, Reiner A, Liu CM, Deangelis LM, Omuro A (2013) Rare cell capture technology for the diagnosis of leptomeningeal metastasis in solid tumors. Neurology 80:1598–1605. doi:10.1212/WNL. 0b013e31828f183f Epub 2013 Apr 3 27. Songun I, Litvinov SV, van de Velde CJ, Pals ST, Hermans J, van Krieken JH (2005) Loss of Ep-CAM (CO17-1A) expression predicts survival in patients with gastric cancer. Br J Cancer 92:1767–1772 28. Klymkowsky MW, Savagner P (2009) Epithelial-mesenchymal transition: a cancer researcher’s conceptual friend and foe. Am J Pathol 174:1588–1593. doi:10.2353/ajpath.2009.080545 29. Chamberlain MC, Johnston SK, Glantz MJ (2009) Neoplastic meningitis-related prognostic significance of the Karnofsky performance status. Arch Neurol 66:74–78 30. Oechsle K, Lange-Brock V, Kruell A, Bokemeyer C, de Wit M (2010) Prognostic factors and treatment options in patients with leptomeningeal metastases of different primary tumors: a retrospective analysis. J Cancer Res Clin Oncol 136:1729–1735. doi:10.1007/s00432-010-0831-x Epub 2010 Mar 4 31. de Azevedo CR, Cruz MR, Chinen LT, Peres SV, Peterlevitz MA, de Azevedo Pereira AE, Fanelli MF, Gimenes DL (2011) Meningeal carcinomatosis in breast cancer: prognostic factors and outcome. J Neurooncol 104:565–572. doi:10.1007/s11060010-0524-y Epub 2011 Jan 14 32. Gauthier H, Guilhaume MN, Bidard FC, Pierga JY, Girre V, Cottu PH, Laurence V, Livartowski A, Mignot L, Dieras V (2010) Survival of breast cancer patients with meningeal carcinomatosis. Ann Oncol 21:2183–2187 33. Chamberlain MC, Johnston SK (2009) Neoplastic meningitis: survival as a function of cerebrospinal fluid cytology. Cancer 115:1941–1946. doi:10.1002/cncr.24210 34. Mun˜iz C, Martı´n-Martı´n L, Lo´pez A, Sa´nchez-Gonza´lez B, Salar A, Almeida J, Sancho JM, Ribera JM, Heras C, Pen˜alver FJ, Go´mez M, Gonza´lez-Barca E, Alonso N, Navarro B, Olave T, Sala F, Conde E, Ma´rquez JA, Cabezudo E, Cladera A, Garcı´aMalo M, Caballero MD, Orfao A, Spanish Group for the Study of Central Nervous System Disease in Non-Hodgkin Lymphoma (2014) Contribution of cerebrospinal fluid sCD19 levels to the detection of CNS lymphoma and its impact on disease outcome. Blood 123:1864–1869. doi:10.1182/blood-2013-11-537993 Epub 2014 Feb 5 35. Illa´n J, Simo M, Serrano C, Castan˜o´n S, Gonzalo R, Martı´nezGarcı´a M, Pardo J, Go´mez L, Navarro M, Pe´rez Altozano J, ´ lvarez R, Bruna J, Subira´ D (2014) Differences in the cereA brospinal fluid inflammatory cell reaction of patients with leptomeningeal involvement by lymphoma and carcinoma. Transl Res 164:460–467. doi:10.1016/j.trsl.2014.03.012
123
RESEARCH PAPER
Diagnostic and prognostic significance of flow cytometry immunophenotyping in patients with leptomeningeal carcinomatosis D. Subira´1 • M. Simo´2 • J. Illa´n3 • C. Serrano4 • S. Castan˜o´n4 • R. Gonzalo4 • J. J. Granizo5 • M. Martı´nez-Garcı´a6 • M. Navarro7 • J. Pardo8 • J. Bruna2
Received: 2 February 2015 / Accepted: 16 March 2015 / Published online: 21 March 2015 Ó Springer Science+Business Media Dordrecht 2015
Abstract Some patients with epithelial-cell cancers develop leptomeningeal carcinomatosis (LC), a severe complication difficult to diagnose and with an adverse prognosis. This study explores the contribution of flow cytometry immunophenotyping (FCI) to the diagnosis and prognosis of LC. Cerebrospinal fluid (CSF) samples from patients diagnosed with LC were studied using FCI. Expression of the epithelial-cell adhesion molecule (EpCAM) was the criterion used to identify the epithelial cells. To test the diagnostic precision, 144 patients (94
diagnosed with LC) were included. The prognostic value of FCI was evaluated in 72 patients diagnosed with LC and eligible for therapy. Compared with cytology, FCI showed greater sensitivity and negative predictive value (79.79 vs. 50 %; 68.85 vs. 51.55 %, respectively), but lower specificity and positive predictive value (84 vs. 100 %; 90.36 vs. 100 %, respectively). The multivariate analysis revealed that the percentage of CSF EpCAM? cells predicted an increased risk of death (HR: 1.012, 95 % CI 1.000–1.023; p = 0.041). A cut-off value of 8 %
& D. Subira´ [email protected]
1
Department of Hematology, Flow Cytometry Division, Hospital Universitario de Guadalajara, c/Donantes de sangre s.n., 19002 Guadalajara, Spain
2
Unit of Neuro-Oncology, Departments of Oncology and Neurology, Hospital Universitario de Bellvitge-ICO Duran i Reynals, Avda. Gran Vı´a s/n km 2.7, Hospitalet de Llobregat, 08907 Barcelona, Spain
3
Unilabs Diagno´sticos, SLU, c/Juan Esplandiu´ 15, 28007 Madrid, Spain
4
Department of Hematology, Fundacio´n Jime´nez Dı´az, Plaza Cristo Rey 1, 28040 Madrid, Spain
5
Clinical Epidemiology Unit, Hospital Infanta Cristina, Avda. 9 de Junio, 2, Parla, 28981 Madrid, Spain
6
Department of Oncology, Hospital del Mar, Paseo Marı´timo 25-29, 8003 Barcelona, Spain
7
Department of Oncology, Hospital Universitario de Salamanca, Instituto de Investigacio´n Biome´dica de Salamanca (IBSAL), Paseo de San Vicente, 58-182, Salamanca 37007, Spain
8
Department of Neurology, Hospital Infanta Elena/Hospital General de Villalba/Hospital Universitario Rey Juan Carlos, c/Gladiolo s/n., 28933 Mo´stoles, Spain
M. Simo´ [email protected] J. Illa´n [email protected] C. Serrano [email protected] S. Castan˜o´n [email protected] R. Gonzalo [email protected] J. J. Granizo [email protected] M. Martı´nez-Garcı´a [email protected] M. Navarro [email protected] J. Pardo [email protected] J. Bruna [email protected]
123
384
EpCAM? cells in the CSF distinguished two groups of patients with statistically significant differences in overall survival (OS) (p = 0.018). This cut-off value kept its statistical significance regardless of the absolute CSF cell-count. The FCI study of the CSF improved the sensitivity for diagnosing LC, but refinement of the technique is needed to improve specificity. Furthermore, quantification of CSF EpCAM? cells was revealed to be an independent prognostic factor for OS in patients with LC eligible for therapy. An 8 % cut-off value contributed to predicting clinical evolution before initiation of therapy. Keywords Flow cytometry Immunophenotype Cerebrospinal fluid Leptomeningeal disease Diagnosis Prognosis Abbreviations CSF Cerebrospinal fluid EpCAM Epithelial-cell adhesion molecule FCI Flow cytometry immunophenotyping IQR Interquartile range IT Intrathecal chemotherapy KPS Karnofsky performance status scale MRI Magnetic resonance imaging NPV Negative predictive value LC Leptomeningeal carcinomatosis OS Overall survival PPV Positive predictive value RDT Radiotherapy SC Systemic chemotherapy
Introduction Neoplastic meningitis is a devastating complication that arises in 5–15 % of patients with cancer, and which involves a poor prognosis. Overall survival (OS) has been associated with several clinical, tumor, and cerebrospinal fluid (CSF) features, sometimes with contradictory findings due to the limitations of these studies [1, 2]. The most widely accepted prognostic factors are performance status, histology, and degree of neurological deficits. Early detection and start of therapy might prevent the onset of irreversible symptoms and therefore prolong survival with an acceptable quality of life [1, 2]. Once clinical symptoms appear, the diagnosis of neoplastic meningitis relies on the cytological identification of malignant cells in the CSF, and/or compatible magnetic resonance imaging (MRI) findings and/or suggestive biochemical CSF findings [3]. However, the sensitivity of cytology and the specificity of MRI are not optimal [1, 2, 4, 5]. Moreover, incidental MRI features suggestive of meningeal dissemination can be found in asymptomatic and paucysymptomatic patients, increasing the diagnostic
123
Clin Exp Metastasis (2015) 32:383–391
challenge. Therefore, several studies have focused on improving the efficiency of the available diagnostic tools, and on introducing new potentially useful biomarkers [6, 7]. Flow cytometry immunophenotyping (FCI) has been used for detecting neoplastic cells in the CSF, but most studies have focused on patients with hematological malignancies [8–13]. However, in our experience, FCI can also improve the sensitivity of conventional CSF cytology in diagnosing leptomeningeal involvement from epithelialcell neoplasias (leptomeningeal carcinomatosis, LC) [14]. In this study, the first goal was to determine whether our previous findings regarding better sensitivity and negative predictive value of FCI compared to CSF cytology could be validated in a larger series of patients with epithelialcell tumors. As a second step, we also explored whether FCI data might offer any prognostic information.
Materials and methods Study design To test the diagnostic precision of FCI, 166 patients with clinical data suspicious of LC from epithelial-cell malignancies were enrolled between January 2011 and December 2012 in 14 Spanish hospitals. Sensitivity, specificity, and the positive predictive value (PPV) and negative predictive value (NPV) for diagnosing LC were calculated for CSF cytology, FCI and MRI. These data were compared with our previous report on a series of 78 patients with cancer who were recruited between 2009 and 2010 [14]. To explore the prognostic value of CSF FCI data, we monitored all patients diagnosed with LC recruited during the 2-year period of our study: from March 2009 to July 2010 [14], and from January 2011 to December 2012 (validation study). Only patients who were candidates to receive therapy were selected and OS was registered. Criteria for exclusion were: patients eligible only for palliative care, those without a complete collection of clinical data, and those whose CSF samples had macroscopic blood contamination after sample centrifugation. In the end, 72 patients were included (Fig. 1). Demographic data collected were age, sex, Karnofsky performance status scale (KPS), localization and histology of primary tumor, time from diagnosis of primary tumor to development of LC, presence of brain metastasis, and therapy received. The CSF parameters obtained with FCI analysis were the absolute number of CSF cells, detection of epithelial cells, percentage of epithelial cells, and percentage and absolute number of accompanying inflammatory cells (lymphocytes, monocytes, and polymorphonuclear cells). Additional CSF parameters such as cytology results, and concentration of glucose and proteins, were also evaluated.
Clin Exp Metastasis (2015) 32:383–391
385
January 2011 to December 2012 Patients with clinical suspicious of leptomeningeal carcinomatosis n=166
Patients excluded n=22 • Incomplete data or no information
Validation study n=144
Diagnosed with leptomeningeal carcinomatosis n=94
Leptomeningeal carcinomatosis excluded (other diagnoses established) n=50
Diagnosed with leptomeningeal carcinomatosis n=48
March 2009 to July 2010 Clinical suspicion of leptomeningeal carcinomatosis n=78
Diagnosed with leptomeningeal carcinomatosis n=142 Patients excluded n=70 • Not eligible for therapy n=38 • Incomplete or no follow-up data n=15 • Macroscopic blood contamination of the cerebrospinal fluid n=17 Prognostic study n=72
Leptomeningeal carcinomatosis eligible for therapy n=72
Fig. 1 Summary of the criteria used for the final inclusion of patients in the validation and prognostic studies
Local ethics committees of participating centers approved the study, and informed consent was obtained from each participant before their enrollment. All procedures were in accordance with the ethical standards of the Helsinki Declaration. FCI studies CSF samples were obtained with lumbar puncture or from an Ommaya reservoir and collected in tubes with EDTA containing an immunofixative reagent (TransFix, Cytomark), for safe transportation [15] from the hospital of origin to the Fundacio´n Jime´nez Dı´az (Madrid, Spain), where all FCI studies were performed. A median of 3 days from extraction to study was recorded. Samples were processed using a previously reported FCI protocol for CSF samples with suspected LC [14]. Briefly, for cell count, an aliquot of the CSF sample (100 ll) was added to the fluorescent dye DRAQ5 (100 ll previously diluted 1:1000) for DNA staining (Biostatus Limited) and Perfect-count microspheres (Cytognos SL, Salamanca, Spain). The remaining CSF sample was centrifuged and the
cell-pellet was stained with a 2-color (fluorescein isothyocyanate, FITC/phycoerythrin, PE) monoclonal antibody (mAb) combination directed against the epithelial-cell adhesion molecule (EpCAM), [clones BerEP4 from DAKO; and EBA-1 from Becton–Dickinson Biosciences (BDB)]. After 20 min of incubation at room temperature in darkness, samples were washed once, and the cell-pellet was resuspended in phosphate buffered saline (PBS). Then, DRAQ5 was added and after 10 min of incubation, the whole volume of sample was acquired on a FACSCanto II flow cytometer (BDB) using the FACSDiva software (version 6.1). The FSC area/FSC height dot-plot was used to help exclude aspirated air and doublets. Analyses were performed using the INFINICYT software program (Cytognos SL, Salamanca, Spain). Cells were separated from debris using positive staining with DRAQ5, and then epithelial and inflammatory cells were identified. For both diagnostic and prognostic studies, the FCI criterion used to define a sample as positive for malignancy was a cluster of at least 16 events (modified from Subira´ et al. [16]) positive for EpCAM expression with the 2 mAb used (EpCAM positive cells). The result was expressed as percentage of
123
386
epithelial cells referred to the total CSF cells. Non-malignant inflammatory cells were identified on the basis of their forward scatter (FSC) and side-scatter (SSC) characteristics. Lymphocytes were FSClow/SSClow, monocytes FSChigh/SSCintermediate/high, and polymorphonuclear cells FSClow/SSChigh. The relative percentage of each population was referred to the total inflammatory cells (EpCAM negative cells). Statistics Qualitative variables were expressed as median and interquartile range. Quantitative variables were expressed as percentage. The usefulness of cytology, FCI and MRI to diagnose LC was compared by calculating the sensitivity, specificity, PPV and NPV. OS analysis was determined by constructing probability curves according to the Kaplan–Meier method and comparing them using a log-rank test. Variables that had a p value \ 0.05 in the univariate analysis were subsequently put in a backward stepwise proportional-hazard analysis (Multivariate Cox regression model) to explore their independent value as predictor of survival. The cut-off used for the continuous variables was their median value. A receiver-operator characteristics (ROC) curve analysis was applied to determine the best cut-off value to assess the prognostic significance of EpCAM? cells. Calculations were performed using SPSS version 14.0 (SPSS Inc, Chicago, IL).
Results Validation study Appropriate clinical data were obtained in 144/166 patients (Fig. 1). The diagnosis of LC was established on the basis of CSF cytology, and/or compatible clinical signs plus MRI findings with biochemical CSF abnormalities. LC was established in 94 patients (61.7 % female) with a median age of 58 years (Interquartile range, IQR 48–66). The distribution of primary tumors was as follows: breast (n = 39), lung (n = 35), gastrointestinal (n = 6), ovarian (n = 4), prostate (n = 3), and others (n = 5). Two and three neoplasms (breast/lung, and gastrointestinal/breast/lung) were simultaneously diagnosed in 2 patients. The most common histology subtype was adenocarcinoma (85.1 %). Forty-six patients (48.94 %) had a KPS value C70 % and 39 (41.49 %) associated parenchymal metastases. A positive cytological detection of malignant cells in the CSF was found in 47 of the 94 patients diagnosed with LC. MRI data were collected in 90 patients and findings were consistent with LC in 68 patients (35 with negative CSF
123
Clin Exp Metastasis (2015) 32:383–391
cytology). Malignant cells were detected with FCI in 75 CSF samples, including 8 out of the 11 samples from patients with both negative CSF cytology and MRI findings (Fig. 2). Sensitivity, specificity, PPV and NPV were calculated for cytology, MRI and FCI (Table 1). FCI showed better sensitivity and NPV compared with cytology (79.79 vs. 50 % and 68.85 vs. 51.55 %, respectively). In contrast, compared to cytology, FCI showed lower specificity and PPV (84 % and 90.36 %, respectively vs. 100 %). MRI (performed in 129 patients) showed 75.56 % sensitivity, 63.04 % specificity, 87.18 % PPV, and 58.86 % NPV. Prognostic study Patient characteristics Seventy-two patients (68.05 % female), with a median age of 59.5 years (IQR 50–66), were eligible for therapy and recruited for the prognostic study. Median KPS score was 70 % (IQR 60–80). The modalities of therapy used were: intrathecal chemotherapy, ITC (n = 33), with (n = 2) or without radiotherapy (RDT, focal spinal or whole brain) (n = 31), systemic chemotherapy, SC (n = 13), with (n = 3) or without RDT (n = 10), whole brain RDT (n = 5), and combination of ITC plus SC (n = 21), with (n = 9) or without RDT (n = 12). All patients included in the study received at least 1 dose of IT. Primary tumors were distributed as follows: breast (n = 35), lung (n = 24), gastrointestinal (n = 5), ovarian (n = 3), and others (n = 2). In the remaining 3 patients, 2 neoplasms (breast/gastrointestinal, n = 2; breast/lung, n = 1) were simultaneously diagnosed. The most common histology subtype was adenocarcinoma (94.44 %). Median time between diagnosis of primary tumor and development of LC was 2.41 years (IQR 0.56–5.54). The diagnosis of primary neoplasia was synchronic with diagnosis of LC in 9 patients (12.5 %). Concurrent brain metastases were detected in 34 of the 69 patients who could be evaluated with MRI. Demographics and CSF data are summarized in Table 2. At the end of the study, after a median follow-up of 46 weeks (range 30.3–146), 8 patients were still alive. Median OS of the entire cohort of patients was 11 weeks (range 1–146.7 weeks). Depending on the therapy received, the median OS was 3.29 weeks for patients with whole brain RDT, 10.14 weeks with SC (±RDT), 10.43 weeks with ITC (±RDT), and 35 weeks for patients receiving ITC ? SC (±RDT). CSF data Median CSF sample volume was 2.6 ml (IQR 2.1–2.8), and median CSF cell count was 6 cells/ll (IQR 2.3–17.8). Cytology identified malignant cells in 35 CSF samples
Clin Exp Metastasis (2015) 32:383–391
387
Flow cytometry immunophenotyping
Cytology
11
Diagnostic tools for leptomeningeal carcinomatosis Cytology/Magnetic resonance imaging Cytology/Flow cytometry immunophenotyping Magnetic resonance imaging/Flow cytometry immunophenotyping Cytology/Magnetic resonance imaging/Flow cytometry immunophenotyping
8
% of cases detected 87.77 82.22 96.66 96.66
31 2
22
*3
13
Magnetic resonance imaging Fig. 2 Distribution of the cerebrospinal fluid samples studied in the validation cohort according to the results of cytology, magnetic resonance imaging and flow cytometry immunophenotyping. Positive cases for leptomeningeal carcinomatosis of each technique are specified. Data shown belong to the 90 patients with available
Table 1 Sensitivity, specificity, positive and negative predictive value for the diagnosis of leptomeningeal carcinomatosis according to cerebrospinal fluid cytology, flow cytometry immunophenotyping and magnetic resonance imaging Cytology (n = 144)
Magnetic resonance imaging (n = 129)
Flow cytometry immunophenotyping (n = 144)
Sensitivity
50.00
75.56
79.79
Specificity
100
63.04
84.00
Positive predictive value
100
87.18
90.36
Negative predictive value
51.55
56.86
68.85
(48.6 %), and FCI detected EpCAM? cells in 56/72 cases (77.7 %). Median percentage of EpCAM? cells was 2.25 % (IQR 0.2–18.6). Median number of EpCAM? cells/ll of CSF was 0.26 (IQR 0.01–2.21). Lymphocytes (median value 56 %) were the predominant population in 40 samples, and the percentage of monocytes exceeded 30 % in 36 samples (50 %). Polymorphonuclear cells were described in 49/72 samples (68 %), and their percentage was [1 % of the inflammatory cell compartment in 37/72 CSF samples (51.4 %). Survival prognosis analysis The univariate analysis revealed that age, sex, histology, localization of primary tumor, time between primary tumor and LC, and presence or absence of parenchymal
cytology, flow cytometry immunophenotyping, and magnetic resonance imaging data. The 4 patients in whom magnetic resonance imaging was not performed are excluded from this diagram. *Patients with negative cytology, flow cytometry immunophenotyping and magnetic resonance imaging
metastases had no impact on OS. Conversely, a KPS value C70 % indicated favorable outcome (p = 0.001; HR = 0.968, 95 % CI 0.950–0.987), and treatment with ITC (±RDT) (p = 0.01) or ITC ? SC (±RDT) (p = 0.001) presented a significantly better OS compared to patients treated only with whole brain RDT. Comparison of the remaining modalities of therapy (SC ± RDT, ITC ± RDT, and ITC ? SC ± RDT) did not show statistical differences in OS. Regarding CSF variables, the univariate analysis showed that the percentage of CSF EpCAM? cells (p = 0.049), the absolute number of EpCAM? cells/ll (p = 0.011), number of monocytes/ll (p = 0.018), and number of PMN cells/ll (p = 0.001) were associated with worse OS. However, no statistical differences in OS were observed after comparing patients with positive versus negative CSF cytology, and positive versus negative CSF FCI findings. Also, neither the absolute number of CSF cells nor the quantitative value of CSF proteins and glucose had any impact on OS. In addition, using the cut-off value of 48.5 mg/dl CSF glucose (2.7 mol/l), as previously reported, was an independent prognostic factor [17], and a trend towards significance was observed (p = 0.064). The multivariate analysis, performed using all the factors that were significant in the univariate analysis, revealed that a median KPS value \70 % (HR = 1.033, 95 % CI 1.013–1.053; p = 0.001), and receiving whole brain RDT (HR = 3.963, 95 % CI 1.445–10.871; p = 0.007) instead of any chemotherapy treatment, predicted poor survival. No statistical differences were observed after comparison of patients receiving ITC versus
123
388
Clin Exp Metastasis (2015) 32:383–391
Table 2 Clinical data and cerebrospinal fluid characteristics of patients included in the prognostic study group
Patients enrolled Male:female
n
%
72
100
59.5 (50–66)
Median interval in years from primary diagnosis to leptomeningeal carcinomatosis (interquartile range)
2.41 (0.56–5.54)
Adenocarcinoma
68
94.44
Breast
35
48.61
Lung
24
33.33
Gastrointestinal
5
6.94
Ovary
3
4.17
Bladder
1
1.39
Kidney
1
1.39
Primary tumor
Breast ? gastrointestinal
2
2.77
Breast ? lung
1
1.39
Yes
34
47.22
No
35
48.61
3
4.17
\70 %
29
40.28
C70 %
42
58.33
Unknown
1
1.39
72
100
Central nervous system metastasis
Unknown Karnofsky performance status
Susceptible to therapy Therapy received Whole brain radiotherapy
5
6.94
Intrathecal therapy alone
31
43.06
Intrathecal therapy ? focal spinal or whole brain radiotherapy
2
2.78
Systemic chemotherapy alone
10
13.89
Systemic chemotherapy ? focal spinal or whole brain radiotherapy
3
4.17
Intrathecal therapy ? systemic chemotherapy
12
16.67
Intrathecal therapy ? systemic chemotherapy ? focal spinal or whole brain radiotherapy
9
12.5
After identification of the percentage of CSF EpCAM? cells useful as a parameter related to OS in patients with LC receiving therapy, we looked for an optimal cut-off value. The receiver-operator characteristic curve (ROC) showed that when using a threshold value of 8 % EpCAM? cells, 2 groups of patients with statistically significant differences in OS were identified (p = 0.018) (Fig. 3). Patients with \8 % EpCAM? cells (n = 47) had a median OS of 24.57 versus 7.71 weeks for patients with C8 % EpCAM? cells (n = 25). The median age and KPS value were 61 years and 70 % for patients with \8 % EpCAM? cells, and 56 years and 60 % for patients with C8 % EpCAM? cells. The number of patients with breast and lung cancer was 20 (42.5 %) and 17 (31.9 %), respectively, in the group of\8 % EpCAM? cells, and 15 (60 %) and 3 (12 %) in the group of C8 % EpCAM? cells. Considering that the absolute number of cells in the CSF of patients diagnosed with LC is highly variable, we explored whether the cut-off value of 8 % EpCAM? cells might be useful to predict OS in patients with a normal or high CSF cell count. No differences in OS were observed after comparison of CSF samples with [5 and B5 cells/ll (p = 0.432). In contrast, the variable percentage of EpCAM? cells (\8 and C8 %) kept its statistical significance after being stratified according to the absolute number of CSF cells (p = 0.020).
Discussion
Median value (interquartile range) Cerebrospinal fluid data 2.6 (2.1–2.8)
Cell count per ml
6 (2.3–17.8)
Proteins in mg/dl Glucose in mg/dl
58 (38–76.5) 73 (47–170)
SC (p = 0.770), ITC versus combination of ITC ? SC (p = 0.153), and SC versus the combination ITC ? SC (p = 0.158). The only CSF parameter that retained
123
Influence of EpCAM? cells and OS
23:49
Median age (interquartile range)
Sample volume in ml
statistical significance for an increased risk of death in the multivariate analysis was the percentage of EpCAM? cells (HR = 1.012, 95 % CI 1.000–1.023; p = 0.041).
The increasing survival of cancer patients implies an increased risk of developing LC [18]. However, diagnosing LC is still a challenge, especially in the early stages. New methods that can help physicians diagnose the disease are under development, but CSF cytology remains the keystone for diagnosing LC because it identifies malignant cells in the CSF. The rationale in using FCI for the diagnosis of LC in patients with epithelial-cell tumors is also its ability to identify epithelial cells in the CSF, using surface expression of the epithelial-cell antigen EpCAM as the marker of this population [14, 19–24]. A similar strategy is used for the detection of circulating tumor cells, and this recently was used to detect tumor cells in the CSF [25, 26]. However, at the time of writing, clinical laboratories are more familiar with the use of FCI techniques, and it is available in many centers. In our study we tested whether the FCI protocol for detecting EpCAM? cells in the CSF [14] was useful to identify
Clin Exp Metastasis (2015) 32:383–391
389
Fig. 3 a Kaplan–Meier Curves for overall survival according to the 8 % cut-off level of cerebrospinal fluid EpCAM? cells. Median overall survival was 24.57 versus 7.71 weeks for patients with\8 and C8 % EpCAM? cells, respectively. A representative example of 2
cerebrospinal fluid samples is shown, one with a low cell-count and C8 % EpCAM? cells (b), and another with a high cell-count and \8 % EpCAM? cells (c)
LC in a larger validation cohort of patients diagnosed with epithelial-cell tumors. In accordance with our previous findings, the CSF FCI study improved the sensitivity of classical CSF cytology. However, using FCI to analyze the CSF did not eliminate the false negative cases. It is likely that some of the reasons given to explain negative cytology results may also justify negative FCI results: negative or transient seeding in the CSF and suboptimal collection with samples obtained far from the site of origin of clinical symptoms [1, 2, 4, 5]. In addition, down regulation of EpCAM expression described in some aggressive tumors [27] and epithelial-mesenchymal-transition should also be considered [28]. In order to solve these problems, future studies will need to use more powerful flow cytometers capable of evaluating nuclear dyes in combination with a greater number of cell surface antigens, including different epithelial-cell markers. This will help improve the specificity and PPV of flow cytometry studies, even when data are better than those described for MRI. In the meantime, the inclusion of FCI among the LC diagnostic procedures adds specificity to MRI data, and, combined with CSF cytology, significantly decreases the number of undiagnosed cases. Regarding the prognostic study, and in keeping with previous reports, our series also confirmed that a good performance status remains as an independent factor associated with better survival [1, 2, 4, 5, 29–32]. The levels of glucose in the CSF showed a trend towards significance [17], and receiving only whole brain RDT was the only therapeutic choice associated with a bad prognosis. There was no advantage for survival after comparing the different therapeutic options but this fact should be interpreted with caution since this was not a randomized trial evaluating the efficacy of different therapies.
The novelty of this study is the identification of quantification of CSF EpCAM? cells as an independent prognostic factor for patients diagnosed with LC who are to be treated with any of the currently available therapeutic modalities. The prognostic value of the CSF FCI data has been studied in patients diagnosed with high grade B cell lymphoma, but most authors have focused on determining whether occult leptomeningeal disease (defined as positive FCI, negative cytology) might be an indicator of central nervous system relapse [8–13]. In our study, and in keeping with previous reports [33] concerning epithelial-cell neoplasms, this ‘qualitative’ information (positive vs. negative cytology and positive vs. negative FCI data) had no prognostic value. In contrast, we found an association between higher percentages of CSF EpCAM? cells and a shorter OS. To the best of our knowledge, this is the first study to propose using a cut-off level of malignant cell infiltration in the CSF to stratify the OS in patients with leptomeningeal involvement. The cut-off value of C8 % EpCAM? cells identified patients diagnosed with LC with a two-fold risk of death. This correlation between percentage of CSF EpCAM? cells and OS is probably a measurement of disease burden in CSF at the time of diagnosing LC (early or advanced disease). A recent study by Mun˜iz et al. [34], describing a clear association between higher sCD19 (a soluble B-cell specific protein) CSF levels and OS in patients with diffuse large B-cell lymphoma and Burkitt lymphoma, also supports this hypothesis. In cancer patients, leptomeningeal neoplastic cell involvement is usually associated with advanced systemic disease [1, 2], but our observations also support the hypothesis that tumor burden in the CSF affects the patient’s prognosis.
123
390
Clin Exp Metastasis (2015) 32:383–391
From a technical point of view, the 8 % threshold value is high enough to be easily implemented in clinical flow cytometry laboratories. It also has the advantage of maintaining its prognostic value independently of the total number of CSF cells, considering the wide range of values that can be found in patients with LC. This would mean that, for every case, the really important measurement is the relationship between malignant and inflammatory cells. However, none of the inflammatory cell populations evaluated in the multivariate analysis was associated with differing rates of survival among patients with LC. In a previous study, we reported that the percentage of polymorphonuclear cells in the CSF was the only significant distinctive parameter between the inflammatory cell compartment of cancer patients with and without LC [35]. In the present study, the percentage of polymorphonuclear cells was cited in the univariate analysis, but lacked statistical significance in the multivariate analysis. In summary, an initial evaluation of the CSF using FCI is an easy way to obtain both diagnostic and prognostic information about patients with LC. On the one hand, FCI improves the sensitivity for the diagnosis, and on the other, quantification of the percentage of CSF EpCAM? cells is prognostic of OS. According to our data, a cut-off value of 8 % CSF EpCAM? cells is a useful variable to help physicians select patients who might benefit from therapy and to stratify patients in treatment trials. This finding needs to be externally validated and its clinical value has to be verified with additional clinical studies. Acknowledgments The authors would like to thank all the patients and participants of the Meningeal Carcinomatosis Study Cooperative Group who made this study possible. We also wish to thank Mundipharma Spain for supporting the study, and Lawrence J.C. Baron for editing the manuscript. Conflict of interest of interest.
4. 5.
6.
7.
8.
9.
10.
11.
The authors declare that they have no conflict
Ethical standards Local ethics committees of participating centers approved the study, and all patients gave their informed consent prior to their inclusion in the study. All procedures were performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. Funding Mundipharma contributed financial support for sample transportation.
12.
13.
References 1. Gleissner B, Chamberlain MC (2006) Neoplastic meningitis. Lancet Neurol 5:443–452 2. Chamberlain MC (2010) Leptomeningeal metastasis. Curr Opin Oncol 22:627–635. doi:10.1097/CCO.0b013e32833de986 3. Brem SS, Bierman PJ, Brem H, Butowski N, Chamberlain MC, Chiocca EA, DeAngelis LM, Fenstermaker RA, Friedman A, Gilbert MR, Hesser D, Junck L, Linette GP, Loeffler JS, Maor
123
14.
MH, Michael M, Moots PL, Morrison T, Mrugala M, Nabors LB, Newton HB, Portnow J, Raizer JJ, Recht L, Shrieve DC, Sills AK Jr, Vrionis FD, Wen PY, National Comprehensive Cancer Network (2011) Central nervous system cancers. J Natl Compr Cancer Netw 9:352–400 Bruna J, Simo´ M, Velasco R (2012) Leptomeningeal metastases. Curr Treat Options Neurol 14:402–415 Strik H, Pro¨mmel P (2010) Diagnosis and individualized therapy of neoplastic meningitis. Expert Rev Anticancer Ther 10:1137–1148 Groves MD, Hess KR, Puduvalli VK, Colman H, Conrad CA, Gilbert MR, Weinberg J, Cristofanilli M, Yung WKA, Liu TJ (2009) Biomarkers of disease: cerebrospinal fluid vascular endothelial growth factor (VEGF) and stromal cell derived factor (SDF)-1 levels in patients with neoplastic meningitis (NM) due to breast cancer, lung cancer and melanoma. J Neurooncol 94:229–234 Herrlinger U, Wiendl H, Renninger M, Fo¨rschler H, Dichgans J, Weller M (2004) Vascular endothelial growth factor (VEGF) in leptomeningeal metastasis: diagnostic and prognostic value. Br J Cancer 91:219–224 Hegde U, Filie A, Little RF, Janik JE, Grant N, Steinberg SM, Dunleavy K, Jaffe ES, Abati A, Stetler-Stevenson M, Wilson WH (2005) High incidence of occult leptomeningeal disease detected by flow cytometry in newly diagnosed aggressive B-cell lymphomas at risk for central nervous system involvement: the role of flow cytometry versus cytology. Blood 105:496–502 Subira´ D, Go´rgolas M, Castan˜o´n S, Serrano C, Roma´n A, Rivas F, Toma´s JF (2005) Advantages of flow cytometry immunophenotyping for the diagnosis of central nervous system nonHodgkin’s lymphoma in AIDS patients. HIV Med 6:21–26 Benevolo G, Stacchini A, Spina M, Ferreri AJ, Arras M, Bellio L, Botto B, Bulian P, Cantonetti M, Depaoli L, Di Renzo N, Di Rocco A, Evangelista A, Franceschetti S, Godio L, Mannelli F, Pavone V, Pioltelli P, Vitolo U, Pogliani EM, Linfomi Fondazione Italiana (2012) Final results of a multicenter trial addressing role of CSF flow cytometric analysis in NHL patients at high risk for CNS dissemination. Blood 120:3222–3228. doi:10. 1182/blood-2012-04-423095 Epub 2012 Aug 27 Sancho JM, Orfao A, Quijano S, Garcı´a O, Panizo C, Pe´rez-Ceballos E, Deben G, Salar A, Gonza´lez-Barca E, Alonso N, Garcı´aVela JA, Capote J, Pen˜alver FJ, Provencio M, Arias J, Plaza J, Caballero D, Morado M, Feliu E, Ribera JM, Spanish Group for the Study of CNS Disease in NHL (2010) Clinical significance of occult cerebrospinal fluid involvement assessed by flow cytometry in non-Hodgkin’s lymphoma patients at high risk of central nervous system disease in the rituximab era. Eur J Haematol 85:321–328. doi:10.1111/j.1600-0609.2010.01478.x Epub 2010 Jul 28 Alvarez R, Dupuis J, Plonquet A, Christov C, Copie-Bergman C, Hemery F, Gaillard I, El Gnaoui T, Kuhnowski F, Bedoui M, Belhadj K, Brugie`res P, Haioun C (2012) Clinical relevance of flow cytometric immunophenotyping of the cerebrospinal fluid in patients with diffuse large B-cell lymphoma. Ann Oncol 23:1274–1279 Wilson WH, Bromberg JE, Stetler-Stevenson M, Steinberg SM, Martin-Martin L, Mun˜iz C, Sancho JM, Caballero MD, Davidis MA, Brooimans RA, Sanchez-Gonzalez B, Salar A, Gonza´lezBarca E, Ribera JM, Shovlin M, Filie A, Dunleavy K, Mehrling T, Spina M, Orfao A (2014) Detection and outcome of occult leptomeningeal disease in diffuse large B-cell lymphoma and Burkitt lymphoma. Haematologica 99:1228–1235 Subira´ D, Serrano C, Castan˜o´n S, Gonzalo R, Illa´n J, Pardo J, Martı´nez-Garcı´a M, Millastre E, Aparisi F, Navarro M, Do´mine M, Gil-Bazo I, Pe´rez Segura P, Gil M, Bruna J (2012) Role of flow cytometry immunophenotyping in the diagnosis of leptomeningeal carcinomatosis. NeuroOncol 14:43–52
Clin Exp Metastasis (2015) 32:383–391 15. de Jongste AH, Kraan J, Broek PD, Brooimans RA, Bromberg JE, Montfort KA, Smitt PA, Gratama JW (2014) Use of TransFixTM cerebrospinal fluid storage tubes prevents cellular loss and enhances flow cytometric detection of malignant hematological cells after 18 hours of storage. Cytom B Clin Cytom 86:272–279 16. Subira´ D, Castan˜o´n S, Aceituno E, Herna´ndez J, Jime´nez-Garo´fano C, Jime´nez A, Jime´nez AM, Roma´n A, Orfao A (2002) Flow cytometric analysis of cerebrospinal fluid samples and its usefulness in routine clinical practice. Am J Clin Pathol 117:952–958 17. Bruna J, Gonza´lez L, Miro´ J, Velasco R, Gil M, Tortosa A (2009) Leptomeningeal carcinomatosis: prognostic implications of clinical and cerebrospinal fluid features. Cancer 115:381–389. doi:10.1002/cncr.24041 18. Glass JP, Melamed M, Chernik NL, Posner JB (1979) Malignant cells in cerebrospinal fluid (CSF): the meaning of a positive CSF cytology. Neurology 29:1369–1375 19. Chang A, Benda PM, Wood BL, Kussick SJ (2003) Lineagespecific identification of nonhematopoietic neoplasms by flow cytometry. Am J Clin Pathol 119:643–655 20. Pillai V, Cibas ES, Dorfman DM (2013) A simplified flow cytometric immunophenotyping procedure for the diagnosis of effusions caused by epithelial malignancies. Am J Clin Pathol 139:672–681. doi:10.1309/AJCP4HIFSHMO9WTK 21. Davidson B, Dong HP, Holth A, Berner A, Risberg B (2007) Flow cytometric immunophenotyping of cancer cells in effusion specimens: diagnostic and research applications. Diagn Cytopathol 35:568–578 22. Kentrou NA, Tsagarakis NJ, Tzanetou K, Damala M, Papadimitriou KA, Skoumi D, Stratigaki A, Anagnostopoulos NI, Malamou-Lada E, Athanassiadou P, Paterakis G (2011) An improved flow cytometric assay for detection and discrimination between malignant cells and atypical mesothelial cells, in serous cavity effusions. Cytom B Clin Cytom 80:324–334. doi:10.1002/ cyto.b.20608 Epub 2011 Jun 21 23. Krishan A, Ganjei-Azar P, Hamelik R, Sharma D, Reis I, Nadji M (2010) Flow immunocytochemistry of marker expression in cells from body cavity fluids. Cytom A 77:132–143. doi:10.1002/cyto. a.20824 24. Risberg B, Davidson B, Dong HP, Nesland JM, Berner A (2000) Flow cytometric immunophenotyping of serous effusions and peritoneal washings: comparison with immunocytochemistry and morphological findings. J Clin Pathol 53:513–517. doi:10.1136/ jcp.53.7.513 25. Patel AS, Allen JE, Dicker DT, Peters KL, Sheehan JM, Glantz MJ, El-Deiry WS (2011) Identification and enumeration of circulating tumor cells in the cerebrospinal fluid of breast cancer patients with central nervous system metastases. Oncotarget 2:752–760
391 26. Nayak L, Fleisher M, Gonzalez-Espinoza R, Lin O, Panageas K, Reiner A, Liu CM, Deangelis LM, Omuro A (2013) Rare cell capture technology for the diagnosis of leptomeningeal metastasis in solid tumors. Neurology 80:1598–1605. doi:10.1212/WNL. 0b013e31828f183f Epub 2013 Apr 3 27. Songun I, Litvinov SV, van de Velde CJ, Pals ST, Hermans J, van Krieken JH (2005) Loss of Ep-CAM (CO17-1A) expression predicts survival in patients with gastric cancer. Br J Cancer 92:1767–1772 28. Klymkowsky MW, Savagner P (2009) Epithelial-mesenchymal transition: a cancer researcher’s conceptual friend and foe. Am J Pathol 174:1588–1593. doi:10.2353/ajpath.2009.080545 29. Chamberlain MC, Johnston SK, Glantz MJ (2009) Neoplastic meningitis-related prognostic significance of the Karnofsky performance status. Arch Neurol 66:74–78 30. Oechsle K, Lange-Brock V, Kruell A, Bokemeyer C, de Wit M (2010) Prognostic factors and treatment options in patients with leptomeningeal metastases of different primary tumors: a retrospective analysis. J Cancer Res Clin Oncol 136:1729–1735. doi:10.1007/s00432-010-0831-x Epub 2010 Mar 4 31. de Azevedo CR, Cruz MR, Chinen LT, Peres SV, Peterlevitz MA, de Azevedo Pereira AE, Fanelli MF, Gimenes DL (2011) Meningeal carcinomatosis in breast cancer: prognostic factors and outcome. J Neurooncol 104:565–572. doi:10.1007/s11060010-0524-y Epub 2011 Jan 14 32. Gauthier H, Guilhaume MN, Bidard FC, Pierga JY, Girre V, Cottu PH, Laurence V, Livartowski A, Mignot L, Dieras V (2010) Survival of breast cancer patients with meningeal carcinomatosis. Ann Oncol 21:2183–2187 33. Chamberlain MC, Johnston SK (2009) Neoplastic meningitis: survival as a function of cerebrospinal fluid cytology. Cancer 115:1941–1946. doi:10.1002/cncr.24210 34. Mun˜iz C, Martı´n-Martı´n L, Lo´pez A, Sa´nchez-Gonza´lez B, Salar A, Almeida J, Sancho JM, Ribera JM, Heras C, Pen˜alver FJ, Go´mez M, Gonza´lez-Barca E, Alonso N, Navarro B, Olave T, Sala F, Conde E, Ma´rquez JA, Cabezudo E, Cladera A, Garcı´aMalo M, Caballero MD, Orfao A, Spanish Group for the Study of Central Nervous System Disease in Non-Hodgkin Lymphoma (2014) Contribution of cerebrospinal fluid sCD19 levels to the detection of CNS lymphoma and its impact on disease outcome. Blood 123:1864–1869. doi:10.1182/blood-2013-11-537993 Epub 2014 Feb 5 35. Illa´n J, Simo M, Serrano C, Castan˜o´n S, Gonzalo R, Martı´nezGarcı´a M, Pardo J, Go´mez L, Navarro M, Pe´rez Altozano J, ´ lvarez R, Bruna J, Subira´ D (2014) Differences in the cereA brospinal fluid inflammatory cell reaction of patients with leptomeningeal involvement by lymphoma and carcinoma. Transl Res 164:460–467. doi:10.1016/j.trsl.2014.03.012
123
