J Neurooncol (2015) 121:177–183 DOI 10.1007/s11060-014-1622-z
CLINICAL STUDY
Diagnostic value of interleukin-10 in cerebrospinal fluid for diffuse large B-cell lymphoma of the central nervous system Yasuo Sasagawa • Takuya Akai • Osamu Tachibana Hideaki Iizuka
•
Received: 4 May 2014 / Accepted: 21 September 2014 / Published online: 26 September 2014 Ó Springer Science+Business Media New York 2014
Abstract A biomarker for early diagnosis of central nervous system (CNS) lymphoma would permit early treatment for attenuation of disease progression and neurological deterioration. High interleukin-10 (IL-10) or an IL-10/IL-6 ratio [1.0 are informative parameters for discriminating intraocular lymphomas from uveitis. Recent reports have also shown that CSF IL-10 is a potential diagnostic biomarker for CNS lymphoma. The purpose of this study was to evaluate the diagnostic value of IL-10 in cerebrospinal fluid (CSF) in patients with CNS lymphoma compared with other CNS diseases, including CNS tumors and inflammatory diseases. CSF IL-10, IL-6, beta-2 microglobulin, soluble IL-2 receptor and FDG-PET SUVmax were measured in 19 patients with CNS lymphoma (15 primary and 4 secondary diffuse large B-cell lymphomas) and 26 non-lymphoma patients with various brain tumors and inflammatory diseases. The diagnostic accuracy of the respective examinations for differentiation of CNS lymphomas from non-lymphomas was evaluated by receiver operating characteristic (ROC) curve analysis. The area under the ROC curve (AUC) was calculated. CSF IL-10 was detected at significant levels (median, 28 pg/ml; range \2–4,100 pg/ml) in all except one patient with CNS lymphoma, but not detected in any non-lymphoma patients. CSF IL-10 had the highest diagnostic accuracy with AUC = 0.974. At an IL-10 cutoff of 3 pg/ml, the sensitivity and specificity were 94.7 and 100 %, respectively. These results indicate that CSF IL-10 is a superior biomarker for initial screening for patients with CNS lymphoma. Y. Sasagawa (&) T. Akai O. Tachibana H. Iizuka Department of Neurosurgery, Kanazawa Medical University, 1-1 Daigaku, Uchinada, Ishikawa 920-0293, Japan e-mail: [email protected]
Keywords Central nervous system lymphoma Diffuse large B-cell lymphoma Interleukin-10 Cerebrospinal fluid Sensitivity Specificity
Introduction Malignant lymphomas of the central nervous system (CNS) comprise primary CNS lymphoma (PCNSL) and secondary CNS involvement by systemic lymphomas, mostly diffuse large B-cell lymphoma (DLBCL) [1]. Open surgery or stereotactic biopsy including histopathology is the standard diagnostic procedure for patients with suspected CNS lymphoma [1, 2]. Early diagnosis and treatment of CNS lymphoma attenuates disease progression and neurological deterioration, and thus reliable radiological parameters or biomarkers in blood/cerebrospinal fluid (CSF) samples are needed. F-18 fluorodeoxyglucose positron emission tomography (18F-FDG PET) can be used to distinguish between CNS lymphoma and other brain tumors such as malignant glioma or metastatic brain tumors [3]. A high SUVmax value in 18F-FDG PET may be useful for diagnosis of CNS lymphoma [3, 4]. Biomarkers such as beta-2 microglobulin (b2MG) and soluble interleukin-2 receptor (sIL-2R) in CSF have also been correlated with the presence of CNS lymphoma [5]. However, these CSF markers are not specific and may be elevated in other neurologic disorders, including inflammatory disease and meningitis. High IL-10 levels are found in the vitreous of most patients with primary ocular lymphoma [6–11] and high IL-10 or a IL-10/IL-6 ratio [1.0 are useful for discriminating intraocular lymphoma from uveitis [8, 11]. Thus, measurement of these cytokines is helpful in initial screening. The purpose of this study was to evaluate the
123
178
diagnostic value of CSF IL-10 in patients with CNS lymphoma, following recent reports of the potential value of CSF IL-10 as a diagnostic biomarker for this disease [12, 13].
Patients and methods Patients The study population consisted of 19 consecutive patients with CNS lymphoma (15 cases of PCNSL and 4 of secondary CNS lymphoma) and 26 control patients with various brain tumors and neurologic disorders that were suspected to be CNS lymphoma in initial magnetic resonance imaging (MRI). All patients were of Japanese origin and none had a history of immunodeficiency. The patients were treated at Kanazawa Medical University from January 2008 through June 2014. All cases of CNS lymphoma were histologically diagnosed as diffuse large B-cell lymphoma by biopsy or in open surgery. All controls except patients with multiple sclerosis underwent surgery and a pathological diagnosis was obtained. For multiple sclerosis, myelin basic protein and oligoclonal bands in CSF and the clinical course (disappearance of the lesion) were used as evidence for the diagnosis. CSF analysis After informed consent was obtained, CSF samples were collected by diagnostic lumbar puncture in 45 patients with no signs of increased intracranial pressure. These samples were routinely assessed for cell counts, total protein, glucose and cytology, and levels of IL-10, IL-6, b2MG and sIL-2R were measured. IL-10 and sIL-2R were determined by human enzyme-linked immunosorbent assays, IL-6 by a human chemiluminescent enzyme immunoassay, and b2MG by a latex agglutination-turbidimetric immunoassay. The lower limits of quantification for IL-10, IL-6, sIL2R, and b2MG were 2 pg/ml, 0.3 pg/ml, 54.5 U/ml, and 0.2 mg/L, respectively. 18
F-FDG PET
Fifteen of 19 lymphoma patients and 17 of 26 non-lymphoma patients underwent 18F-FDG PET with a dedicated PET camera (Headtome IV, Shimadzu, Japan) before surgery. All patients fasted for 6 h before scanning. Transmission scans were obtained before 18F-FDG administration for attenuation correction using a 68Ge ring source. Blood (1 mL) was drawn for the baseline blood glucose estimation. Immediately after the transmission scan, 18F-FDG (mean dose: 185 MBq) was administered intravenously. After a
123
J Neurooncol (2015) 121:177–183
40-60 min uptake period, the patient was repositioned in the scanner. An emission scan was acquired for 10 min per bed position and a whole body scan was performed for each patient using several bed positions according to the height of the patient. 18F-FDG accumulation within lesions was analyzed semi-quantitatively by evaluating the standardized uptake value (SUV, activity concentration/injected dose/ body weight). A region of interest was set manually by an observer around the hottest area of each lesion or the lesion center located on MRI. The maximal value of SUV (SUVmax) was regarded as the representative value for each lesion. Data analysis All analyses were performed using Excel 2010 (Microsoft, Redmond, WA, USA) add-in software Ekuseru-Toukei 2010 for Windows (Social Survey Research Information, Tokyo, Japan). The IL-10/IL-6 ratio was calculated from the IL-10 and IL-6 levels. The diagnostic accuracy of the IL-10, IL-10/IL-6 ratio, b2MG, sIL-2R and SUVmax values for differentiation of CNS lymphomas from non-lymphomas was evaluated by receiver operating characteristic (ROC) curve analysis [14]. The area under the ROC curve (AUC) was calculated. A cutoff was considered optimal when the product of paired values for sensitivity and specificity reached a maximum on the ROC curve. Sensitivity and specificity for each test were estimated using a Bayesian approach implemented in a Markov Chain Monte Carlo algorithm [15].
Results The clinical characteristics and levels of IL-10, IL-6, sIL-2R, b2MG and SUVmax are given in Table 1. The median CSF levels of IL-10, IL-6, sIL-2R, and b2MG and the SUVmax were 28 pg/ml (range B2–4,100 pg/ml), 10.8 pg/ml (range 1.2–127 pg/ml), 225 U/ml (range \54.5–2,750 U/ml), 3.9 mg/L (range 1.7–11.8 mg/L), and 17.4 (range 9.1–30.8) in CNS lymphoma patients. In contrast, in controls, these values were \2.0 pg/ml, 3.9 pg/ml (range 0.4–1,790 pg/ml), \54.5 U/ml (range \54.5–152 U/ml), 1.5 mg/L (range 0.3–6.6 mg/ L), and 9.8 (range 3–23.7), respectively. CSF IL-10 was detected at significant levels in all except one patient with CNS lymphoma, but not detected in any non-lymphoma patients. No patients with CNS lymphoma except one (case L2) had malignant cells (class C3) in cytology analysis. ROC curves for IL-10, IL-10/IL-6 ratio, sIL-2R, b2MG and SUVmax to distinguish CNS lymphomas are shown in Fig. 1. IL-10 showed the highest diagnostic accuracy with AUC = 0.974. The IL-10/IL-6 ratio, sIL-2R, b2MG and SUVmax had diagnostic accuracies of AUC = 0.795,
J Neurooncol (2015) 121:177–183
179
Table 1 Clinical characteristics, CSF and FDG-PET data in CNS lymphoma and non-lymphoma patients No.
Age/Sex
Diagnosis
Location
IL-10 (pg/ml)
IL-6 (pg/ml)
sIL-2R (U/ml)
b2MG (mg/L)
FDG-PET SUVmax
CNS lymphoma patients L1
64/M
Primary DLBCL
Temporal lobe
1,040
17.2
190
7.5
17.4
L2
83/F
Primary DLBCL
Brainstem
278
127
204
11.8
NP
L3
73/F
Primary DLBCL
Cerebellum
116
5.1
63
3.9
NP
L4
78/M
Primary DLBCL
Temporal lobe
59
11.2
507
5.7
30.8
L5
61/M
Primary DLBCL
4th ventricle
86
5.2
299
5.7
21.8
L6
50/M
Primary DLBCL
Parietal lobe
51
10.8
102
2.6
13.6
L7
79/F
Primary DLBCL
3rd ventricle
43
6.8
2,750
9.7
15.8
L8
65/F
Primary DLBCL
Right cp angle
28
6.6
\54.5
2.4
12.4
L9
76/F
Primary DLBCL
Cervical cord
10
3.2
69
6.3
NP
L10
46/M
Primary DLBCL
Frontal lobe
12
179
260
3.2
19.8
L11
53/M
Primary DLBCL
Abducens nerve
8
1.2
498
4.1
9.1
L12
64/M
Primary DLBCL
Corpus callosum
7
12.2
286
3.9
28.9
L13
56/M
Primary DLBCL
Basal ganglia
6
8.9
199
1.7
14.6
L14
70/M
Primary DLBCL
Frontal lobe
3
49.8
344
2.8
25
L15
67/M
Primary DLBCL
3rd ventricle
\2
11.2
130
2.5
10.5
L16
50/M
Secondary DLBCL
Temporal lobe basal ganglia
4,100
68.7
1,000
2.6
17.7
L17
68/F
Secondary DLBCL
Cervical cord
212
4.2
158
2.5
10.5
L18
79/F
Secondary DLBCL
Frontal lobe
17
5.1
225
7.5
18.5
L19
80/F
Secondary DLBCL
Cp angle cerebellum
13
23.8
620
1.8
NP
Non-lymphoma patients N1
70/F
Glioblastoma
Temporal lobe
\2
5.9
\54.5
2
18.9
N2
68/F
Glioblastoma
Frontal lobe
\2
8.2
\54.5
1.4
NP
N3
61/M
Glioblastoma
3rd ventricle
\2
835
\54.5
2.2
11.6
N4
60/F
Glioblastoma
Occipital lobe
\2
3.5
\54.5
1.2
10.9
N5
77/M
Glioblastoma
Parietal lobe
\2
4.5
\54.5
1.9
7.8
N6
65/F
Anaplastic astrocytoma
Corpus callosum
\2
1,790
\54.5
2.6
15.4
N7
46/M
Anaplastic astrocytoma
Frontal lobe
\2
1.8
\54.5
1.1
10.2
N8
59/M
Ependymoma
Cervical cord
\2
1.8
\54.5
1.6
10.5
N9
59/M
Glioma
Pons
\2
1.7
\54.5
1.4
NP
N10
58/F
Metastasis (breast Ca)
3rd ventricle
\2
0.4
\54.5
0.3
6.8
N11
44/F
Metastasis (breast Ca)
Parietal lobe
\2
15.5
\54.5
1.8
5.7
N12
47/F
Metastasis (lung Ca)
Frontal lobe cerebellum
\2
4.3
\54.5
1
NP NP
N13
67/M
Metastasis (lung Ca)
Frontal lobe
\2
4.2
\54.5
2.2
N14
73/M
Metastasis (colon Ca)
Frontal lone
\2
1.3
\54.5
0.3
4.6
N15
21/M
Immature teratoma
Suprasellar region
\2
271
\54.5
3.3
NP
N16
60/M
Atypical meningioma
Cp angle
\2
1.6
\54.5
1.1
8.5
N17
39/F
Undifferentiated sarcoma
Cavernous sinus
\2
18.5
\54.5
1.3
9.8 3.58
N18
44/F
Multiple sclerosis
Basal ganglia
\2
3.2
\54.5
1.7
N19
36/M
Multiple sclerosis
Pons
\2
1.3
60.4
0.9
6.6
N20
30/M
Multiple sclerosis
Cervical cord
\2
2.3
67.1
1
NP
N21
26/M
Sarcoidosis
Frontal lobe temporal lobe
\2
62.9
100
2.1
23.7
N22
69/M
Hypertrophic pachymeningitis
Cervical spine
\2
2
\54.5
1.2
NP
N23
52/F
Myelitis
Cervical cord
\2
5.8
\54.5
1.1
3
N24
69/M
Leukoencephalopathy
Frontal lobe
\2
9.2
152
6.6
NP
N25
61/F
Rosai–Dorfman disease
Temporal lobe
\2
3.4
70
1.8
NP
N26
37/M
Cerebral vasculitis
Parietal lobe
\2
3
\54.5
2.1
21.8
CSF cerebrospinal fluid, FDG-PET fluorodeoxyglucose-positron emission tomography, IL-10 interleukin, 10 IL-6 interleukin 6, sIL-2R soluble interleukin 2 receptor, b2MG beta-2 microglobulin, DLBCL diffuse large B-cell lymphoma, NP not performed, cp cerebellopontine, Ca cancer
123
180
J Neurooncol (2015) 121:177–183
Fig. 1 ROC curves for IL-10, IL-10/IL-6 ratio, b2MG, sIL-2R and FDG-PET SUVmax for differentiation of CNS lymphomas from nonlymphomas. IL-10 showed the highest diagnostic accuracy with AUC = 0.974
0.950, 0.908 and 0.806, respectively. The sensitivity and specificity of each parameter using the optimal cutoff value are shown in Table 2. IL-10 and sIL-2R had the highest sensitivity of 94.7 % and IL-10 had the highest specificity of 100 %. Illustrative cases Case L13 (Fig. 2): A 56-year-old man presented with a 3-week history of recent memory disturbance and difficulty walking. MRI revealed a heterogeneously enhanced tumor with peritumoral edema in the right basal ganglia. A radiologist diagnosed the tumor as glioblastoma based on MRI findings and the patient was admitted to our hospital. No primary lesion of the metastasis was revealed on 18FFDG PET, and SUVmax of the tumor was 14.6. The CSF levels of IL-10, IL-6, sIL-2R and b2MG were 6 pg/ml, 8.9 pg/ml, 199 U/ml, and 1.7 mg/L, respectively. Frontal craniotomy with an interhemispheric transcallosal approach was performed to obtain a pathological diagnosis. The biopsy sample contained lymphoma cells and the pathological diagnosis was diffuse large B-cell lymphoma. Chemotherapy with high-dose methotrexate was initiated 8 days postoperatively. MRI after two courses of chemotherapy showed disappearance of the tumor and the patient made a complete recovery.
Case N6 (Fig. 3): A 65-year-old woman presented with a 2-week history of recent memory disturbance. MRI revealed a homogeneously enhanced tumor in the corpus callosum. No primary lesion of the metastasis was revealed on 18F-FDG PET, and SUVmax of the tumor was 15.4. A radiologist diagnosed the tumor as primary CNS lymphoma based on MRI and 18F-FDG PET findings. The CSF levels of IL-10, IL-6, sIL-2R and b2MG were \2 pg/ml, 1,790 pg/ml, \54.5 U/ml, and 2.6 mg/L, respectively. An endoscopic biopsy was performed via the right anterior horn of the lateral ventricle to obtain a pathological diagnosis. The biopsy sample contained glioma cells and the pathological diagnosis was anaplastic astrocytoma. Chemotherapy with temozolomide and extended local radiation with 60 Gy was initiated 12 days postoperatively. MRI after initial chemoradiotherapy revealed reduction of the tumor and the patient made a slight recovery.
Discussion The results of this study show that the CSF IL-10 level has diagnostic value for distinguishing CNS lymphomas from other CNS lesions. CSF IL-10 levels were elevated in almost all patients with CNS lymphomas, but were undetectable in patients with non-lymphomas, including CNS tumors or inflammatory diseases. This biomarker was superior to traditional parameters used in initial screening for CNS lymphoma patients. In ocular lymphomas, an IL-10/IL-6 ratio [1.0 in vitreous specimens has been suggested as a diagnostic criterion. The sensitivity based on this criterion is 75–90 % [8, 9, 11]. Thus, the sensitivity of 68.4 % in patients with CNS lymphoma in the current study was inferior to that reported for ocular lymphoma. In CNS lymphoma, two other recent studies of the diagnostic value of IL-10 alone in cohorts of moderate size have been reported (Table 3). In an analysis of CSF IL-10 in 26 patients with PCNSL and 40 patients with other brain tumors, Sasayama et al. showed that CSF IL-10 may be a useful diagnostic biomarker for PCNSL [12], since the CSF IL-10 was significantly higher in PCNSL than in other brain tumors. The sensitivity and specificity at an IL-10 cutoff level of 9.5 pg/ml were 71.0 and 100 %, respectively [12], and thus the sensitivity was lower than that of 94.7 % found in the present study. The
Table 2 Sensitivity, specificity and optimal cutoff values of parameters for differentiation of CNS lymphomas from non-lymphomas IL-10
IL10/IL-6 ratio
sIL-2R
b2MG
FDG-PET SUVmax
Sensitivity (%)
94.7
68.4
94.7
89.4
80
Specificity (%)
100
96.1
84.6
88.5
Cutoff value
3 pg/ml
123
2.2
60.4 U/ml
2.4 mg/L
75 12.4
J Neurooncol (2015) 121:177–183
181
Fig. 2 a Gadolinium-enhanced T1-weighted MRI showed a heterogeneously enhanced tumor in the right basal ganglia. b Hematoxylin and eosin staining of a specimen from the tumor showed proliferation of lymphoma cells (original magnification 9100)
difference in sensitivity may be due to the use of different cutoff values in the two studies (9.5 vs. 3 pg/ml). Also, some T cell and unclassified lymphomas were included in the population in Sasayama et al., and these type of lymphomas have undetectable CSF IL-10 levels [12]. Five of the 24 PCNSL cases had CSF IL-10 at an undetectable level, including two cases of DLBCL [12]. CSF IL-10 was also undetectable in one lymphoma in the present study. Thus, the detection limit (2 ng/ml) for IL-10 analysis may cause a false negative result. A specificity of 100 % was found in Sasayama et al. [12], consistent with our result. These results suggest that screening using the CSF IL-10 level may be useful for selection of a subsequent surgical procedure, as described below. Rubenstein et al. [13] examined the diagnostic value of CSF IL-10 in 53 newly diagnosed cases of CNS lymphoma (43 primary and 10 secondary lymphomas) and 91 control patients (8 cases of other primary brain tumors, 12 metastatic brain tumors, and 71 neuroinflammatory/infection diseases). This study included a relatively large cohort, but the sensitivity was lower and the median IL-10 levels were higher than those in Sasayama et al. and in the current study. The median IL10 levels were 282.9 ng/ml in PCNSL patients in Rubenstein et al., compared to 27 ng/ml in Sasayama et al. and 28 ng/ml in our study. In addition, CSF IL-10 was detected in non-lymphoma patients in Rubenstein et al., but not in the other two studies. The reasons for the differences in IL10 levels among the three studies are unclear, but these levels may have been affected by the treatment method and the antibodies used in the measurement. In our study, CSF sIL-2R also had higher sensitivity. Thus, the combination of multiple biomarkers such as IL-10 plus sIL-2R may give
a more accurate diagnosis. Regardless, there is a need to accumulate studies to determine the optimal cutoff value for CSF IL-10 to differentiate CNS lymphomas from nonlymphomas. CNS lymphomas are difficult to distinguish from other CNS tumors and from rare inflammatory diseases such as multiple sclerosis, neurosarcoidosis and cerebral vasculitis [16, 17] because inflammatory diseases occur in multifocal and other patterns that are similar to CNS lymphomas. However, CSF IL-10 was undetectable in inflammatory diseases, while b2MG, sIL-2R and FDG-PET SUVmax were elevated in some cases. Cytokines are deregulated and play important roles in the pathogenesis of various cancers via autocrine or paracrine mechanisms [18]. IL-10 was first described as a growth and differentiation factor for B lymphocytes that induced activated B-cells to secret large amounts of immunoglobulin [19]. IL-10 also has broad anti-inflammatory properties through suppression of macrophage and dendritic cell function [20]. The IL-10 level has also been associated with inflammation in overweight and obese patients [21], but our cases did not include obese patients (body mass index C30). In malignant lymphomas, IL-10 is produced by lymphoma cells and further enhances proliferation of these cells [22], although the exact role of IL-10 in proliferation of CNS lymphoma has yet to be established. In neurosurgery, preoperative diagnosis is very important because the surgical method depends on the tumor type. Gross total resection is the standard procedure for malignant gliomas or metastatic brain tumors, whereas biopsy for pathological diagnosis is sufficient in CNS lymphoma because chemoradiotherapy is effective [2]. Therefore, burr hole or minimum craniotomy can be
123
182
J Neurooncol (2015) 121:177–183
Fig. 3 a Gadolinium-enhanced T1-weighted MRI showed a homogeneously enhanced tumor in the corpus callosum. b 18F-FDG PET showed high FDG uptake (SUVmax: 15.4) in the tumor.
c Hematoxylin and eosin staining of a specimen from the tumor showed proliferation of glioma cells (original magnification 9200)
Table 3 Summary of studies of median level, cuttoff value, sensitivity and specificity of IL-10 for differentiation of CNS lymphomas from nonlymphomas Author (year)
Median IL-10 level (pg/ml)
Cutoff value (pg/ml)
Sensitivity (%)
Specificity (%)
Sasayama [12]
27: PCNSL (n = 26) \ 2.0: other BTs (n = 40)
9.5
71
100
Rubenstein [13]
282.9: PCNSL (n = 43) 57: SCNSL (n = 10) 10.9: other primary BTs (n = 8) 5.3: metastatic BTs (n = 12) 5.6: neuro-infla/infe (n = 71)
16.2
64
94.1
Current study (2014)
28: PCNSL (n = 15) 114.5: SCNSL (n = 4) \ 2.0: other BTs (n = 17) \ 2.0: neuro-infla (n = 9)
3
94.7
100
PCNSL primary central nervous system lymphoma, n number, BTs brain tumors, SCNSL secondary central nervous system lymphoma, neuroinfla neuro-inflammatory diseases, neuro-infe neuro-infection diseases
selected for a preoperative diagnosis of CNS lymphoma [12]. Thus, preoperative measurement of CSF IL-10 can give useful information for planning of a neurosurgical procedure.
123
In summary, our results show the diagnostic value of the CSF IL-10 level in patients with CNS lymphomas, especially DLBCL. The sample size limits the strength of our conclusions and accumulation of studies is required to give
J Neurooncol (2015) 121:177–183
definitive conclusion on the impact of CSF IL-10 in diagnosis of CNS lymphoma. Acknowledgments This work was supported by a Grant for Promoted Research from Kanazawa Medical University (S2013-7). Conflict of interest of interest.
The authors declare that they have no conflict
References 1. Olson JE, Janney CA, Rao RD, Cerhan JR, Kurtin PJ, Schiff D, Kaplan RS, O’Neill BP (2002) The continuing increase in the incidence of primary central nervous system non-Hodgkin lymphoma: a surveillance, epidemiology, and end results analysis. Cancer 95:1504–1510 2. Pels H, Juergens A, Schirgens I, Glasmacher A, Schulz H, Engert A, Schackert G, Reichmann H, Kroschinsky F, Vogt-Schaden M, Egerer G, Bode U, Deckert M, Fimmers R, Urbach H, SchmidtWolf IG, Schlegel U (2010) Early complete response during chemotherapy predicts favorable outcome in patients with primary CNS lymphoma. Neuro Oncol 12:720–724 3. Kosaka N, Tsuchida T, Uematsu H, Kimura H, Okazawa H, Itoh H (2008) 18F-FDG PET of common enhancing malignant brain tumors. AJR Am J Roentgenol 190:365–369 4. Yamashita K, Yoshiura T, Hiwatashi A, Togao O, Yoshimoto K, Suzuki SO, Abe K, Kikuchi K, Maruoka Y, Mizoguchi M, Iwaki T, Honda H (2013) Differentiating primary CNS lymphoma from glioblastoma multiforme: assessment using arterial spin labeling, diffusion-weighted imaging, and 18F-fluorodeoxyglucose positron emission tomography. Neuroradiology 55:135–143 5. Scott BJ, Douglas VC, Tihan T, Rubenstein JL, Josephson SA (2013) A systematic approach to the diagnosis of suspected central nervous system lymphoma. JAMA Neurol 70:311–319 6. Cassoux N, Giron A, Bodaghi B, Tran TH, Baudet S, Davy F, Chan CC, Lehoang P, Merle-Be´ral H (2007) IL-10 measurement in aqueous humor for screening patients with suspicion of primary intraocular lymphoma. Investig Ophthalmol Vis Sci 48:3253–3259 7. Cassoux N, Merle-Beral H, Leblond V, Bodaghi B, Mile´a D, Gerber S, Fardeau C, Reux I, Xuan KH, Chan CC, LeHoang P (2000) Ocular and central nervous system lymphoma: clinical features and diagnosis. Ocul Immunol Inflamm 8:243–250 8. Kimura K, Usui Y, Goto H; Japanese Intraocular Lymphoma Study Group (2012) Clinical features and diagnostic significance of the intraocular fluid of 217 patients with intraocular lymphoma. Jpn J Ophthalmol 56:383–389
183 9. Wolf LA, Reed GF, Buggage RR, Nussenblatt RB, Chan CC (2003) Vitreous cytokine levels. Ophthalmology 110:1671–1672 10. Chan CC, Buggage RR, Nussenblatt RB (2002) Intraocular lymphoma. Curr Opin Ophthalmol 13:411–418 11. Chan CC (2003) Molecular pathology of primary intraocular lymphoma. Trans Am Ophthalmol Soc 101:275–292 12. Sasayama T, Nakamizo S, Nishihara M, Kawamura A, Tanaka H, Mizukawa K, Miyake S, Taniguchi M, Hosoda K, Kohmura E (2012) Cerebrospinal fluid interleukin-10 is a potentially useful biomarker in immunocompetent primary central nervous system lymphoma (PCNSL). Neuro Oncol 14:368–380 13. Rubenstein JL, Wong VS, Kadoch C, Gao HX, Barajas R, Chen L, Josephson SA, Scott B, Douglas V, Maiti M, Kaplan LD, Treseler PA, Cha S, Hwang JH, Cinque P, Cyster JG, Lowell C (2013) CXCL13 plus interleukin 10 is highly specific for the diagnosis of CNS lymphoma. Blood 121:4740–4748 14. Zweig MH, Campbell G (1993) Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin Chem 39:561–577 15. Branscum AJ, Gardner IA, Johnson WO (2005) Estimation of diagnostic-test sensitivity and specificity through Bayesian modeling. Prev Vet Med 68:145–163 16. Trojano M, Paolicelli D (2001) The differential diagnosis of multiple sclerosis: classification and clinical features of relapsing and progressive neurological syndromes. Neurol Sci 22(Suppl 2):98–102 17. Spencer TS, Campellone JV, Maldonado I, Huang N, Usmani Q, Reginato AJ (2005) Clinical and magnetic resonance imaging manifestations of neurosarcoidosis. Semin Arthritis Rheum 34:649–661 18. Li N, Grivennikov SI, Karin M (2011) The unholy trinity: inflammation, cytokines, and STAT3 shape the cancer microenvironment. Cancer Cell 19:429–431 19. O’Garra A, Stapleton G, Dhar V, Pearce M, Schumacher J, Rugo H, Barbis D, Stall A, Cupp J, Moore K (1990) Production of cytokines by mouse B cells: b lymphomas and normal B cells produce interleukin 10. Int Immunol 2:821–832 20. Moore KW, de Waal Malefyt R, Coffman RL, O’Garra A (2001) Interleukin-10 and the interleukin-10 receptor. Annu Rev Immunol 19:683–765 21. McGuire TR, Brusnahan SK, Bilek LD, Jackson JD, Kessinger MA, Berger AM, Garvin KL, O’Kane BJ, Tuljapurkar SR, Sharp JG (2011) Inflammation associated with obesity: relationship with blood and bone marrow endothelial cells. Obesity (Silver Spring) 19:2130–2136 22. Spits H, de Waal Malefyt R (1992) Functional characterization of human IL-10. Int Arch Allergy Immunol 99:8–15
123
CLINICAL STUDY
Diagnostic value of interleukin-10 in cerebrospinal fluid for diffuse large B-cell lymphoma of the central nervous system Yasuo Sasagawa • Takuya Akai • Osamu Tachibana Hideaki Iizuka
•
Received: 4 May 2014 / Accepted: 21 September 2014 / Published online: 26 September 2014 Ó Springer Science+Business Media New York 2014
Abstract A biomarker for early diagnosis of central nervous system (CNS) lymphoma would permit early treatment for attenuation of disease progression and neurological deterioration. High interleukin-10 (IL-10) or an IL-10/IL-6 ratio [1.0 are informative parameters for discriminating intraocular lymphomas from uveitis. Recent reports have also shown that CSF IL-10 is a potential diagnostic biomarker for CNS lymphoma. The purpose of this study was to evaluate the diagnostic value of IL-10 in cerebrospinal fluid (CSF) in patients with CNS lymphoma compared with other CNS diseases, including CNS tumors and inflammatory diseases. CSF IL-10, IL-6, beta-2 microglobulin, soluble IL-2 receptor and FDG-PET SUVmax were measured in 19 patients with CNS lymphoma (15 primary and 4 secondary diffuse large B-cell lymphomas) and 26 non-lymphoma patients with various brain tumors and inflammatory diseases. The diagnostic accuracy of the respective examinations for differentiation of CNS lymphomas from non-lymphomas was evaluated by receiver operating characteristic (ROC) curve analysis. The area under the ROC curve (AUC) was calculated. CSF IL-10 was detected at significant levels (median, 28 pg/ml; range \2–4,100 pg/ml) in all except one patient with CNS lymphoma, but not detected in any non-lymphoma patients. CSF IL-10 had the highest diagnostic accuracy with AUC = 0.974. At an IL-10 cutoff of 3 pg/ml, the sensitivity and specificity were 94.7 and 100 %, respectively. These results indicate that CSF IL-10 is a superior biomarker for initial screening for patients with CNS lymphoma. Y. Sasagawa (&) T. Akai O. Tachibana H. Iizuka Department of Neurosurgery, Kanazawa Medical University, 1-1 Daigaku, Uchinada, Ishikawa 920-0293, Japan e-mail: [email protected]
Keywords Central nervous system lymphoma Diffuse large B-cell lymphoma Interleukin-10 Cerebrospinal fluid Sensitivity Specificity
Introduction Malignant lymphomas of the central nervous system (CNS) comprise primary CNS lymphoma (PCNSL) and secondary CNS involvement by systemic lymphomas, mostly diffuse large B-cell lymphoma (DLBCL) [1]. Open surgery or stereotactic biopsy including histopathology is the standard diagnostic procedure for patients with suspected CNS lymphoma [1, 2]. Early diagnosis and treatment of CNS lymphoma attenuates disease progression and neurological deterioration, and thus reliable radiological parameters or biomarkers in blood/cerebrospinal fluid (CSF) samples are needed. F-18 fluorodeoxyglucose positron emission tomography (18F-FDG PET) can be used to distinguish between CNS lymphoma and other brain tumors such as malignant glioma or metastatic brain tumors [3]. A high SUVmax value in 18F-FDG PET may be useful for diagnosis of CNS lymphoma [3, 4]. Biomarkers such as beta-2 microglobulin (b2MG) and soluble interleukin-2 receptor (sIL-2R) in CSF have also been correlated with the presence of CNS lymphoma [5]. However, these CSF markers are not specific and may be elevated in other neurologic disorders, including inflammatory disease and meningitis. High IL-10 levels are found in the vitreous of most patients with primary ocular lymphoma [6–11] and high IL-10 or a IL-10/IL-6 ratio [1.0 are useful for discriminating intraocular lymphoma from uveitis [8, 11]. Thus, measurement of these cytokines is helpful in initial screening. The purpose of this study was to evaluate the
123
178
diagnostic value of CSF IL-10 in patients with CNS lymphoma, following recent reports of the potential value of CSF IL-10 as a diagnostic biomarker for this disease [12, 13].
Patients and methods Patients The study population consisted of 19 consecutive patients with CNS lymphoma (15 cases of PCNSL and 4 of secondary CNS lymphoma) and 26 control patients with various brain tumors and neurologic disorders that were suspected to be CNS lymphoma in initial magnetic resonance imaging (MRI). All patients were of Japanese origin and none had a history of immunodeficiency. The patients were treated at Kanazawa Medical University from January 2008 through June 2014. All cases of CNS lymphoma were histologically diagnosed as diffuse large B-cell lymphoma by biopsy or in open surgery. All controls except patients with multiple sclerosis underwent surgery and a pathological diagnosis was obtained. For multiple sclerosis, myelin basic protein and oligoclonal bands in CSF and the clinical course (disappearance of the lesion) were used as evidence for the diagnosis. CSF analysis After informed consent was obtained, CSF samples were collected by diagnostic lumbar puncture in 45 patients with no signs of increased intracranial pressure. These samples were routinely assessed for cell counts, total protein, glucose and cytology, and levels of IL-10, IL-6, b2MG and sIL-2R were measured. IL-10 and sIL-2R were determined by human enzyme-linked immunosorbent assays, IL-6 by a human chemiluminescent enzyme immunoassay, and b2MG by a latex agglutination-turbidimetric immunoassay. The lower limits of quantification for IL-10, IL-6, sIL2R, and b2MG were 2 pg/ml, 0.3 pg/ml, 54.5 U/ml, and 0.2 mg/L, respectively. 18
F-FDG PET
Fifteen of 19 lymphoma patients and 17 of 26 non-lymphoma patients underwent 18F-FDG PET with a dedicated PET camera (Headtome IV, Shimadzu, Japan) before surgery. All patients fasted for 6 h before scanning. Transmission scans were obtained before 18F-FDG administration for attenuation correction using a 68Ge ring source. Blood (1 mL) was drawn for the baseline blood glucose estimation. Immediately after the transmission scan, 18F-FDG (mean dose: 185 MBq) was administered intravenously. After a
123
J Neurooncol (2015) 121:177–183
40-60 min uptake period, the patient was repositioned in the scanner. An emission scan was acquired for 10 min per bed position and a whole body scan was performed for each patient using several bed positions according to the height of the patient. 18F-FDG accumulation within lesions was analyzed semi-quantitatively by evaluating the standardized uptake value (SUV, activity concentration/injected dose/ body weight). A region of interest was set manually by an observer around the hottest area of each lesion or the lesion center located on MRI. The maximal value of SUV (SUVmax) was regarded as the representative value for each lesion. Data analysis All analyses were performed using Excel 2010 (Microsoft, Redmond, WA, USA) add-in software Ekuseru-Toukei 2010 for Windows (Social Survey Research Information, Tokyo, Japan). The IL-10/IL-6 ratio was calculated from the IL-10 and IL-6 levels. The diagnostic accuracy of the IL-10, IL-10/IL-6 ratio, b2MG, sIL-2R and SUVmax values for differentiation of CNS lymphomas from non-lymphomas was evaluated by receiver operating characteristic (ROC) curve analysis [14]. The area under the ROC curve (AUC) was calculated. A cutoff was considered optimal when the product of paired values for sensitivity and specificity reached a maximum on the ROC curve. Sensitivity and specificity for each test were estimated using a Bayesian approach implemented in a Markov Chain Monte Carlo algorithm [15].
Results The clinical characteristics and levels of IL-10, IL-6, sIL-2R, b2MG and SUVmax are given in Table 1. The median CSF levels of IL-10, IL-6, sIL-2R, and b2MG and the SUVmax were 28 pg/ml (range B2–4,100 pg/ml), 10.8 pg/ml (range 1.2–127 pg/ml), 225 U/ml (range \54.5–2,750 U/ml), 3.9 mg/L (range 1.7–11.8 mg/L), and 17.4 (range 9.1–30.8) in CNS lymphoma patients. In contrast, in controls, these values were \2.0 pg/ml, 3.9 pg/ml (range 0.4–1,790 pg/ml), \54.5 U/ml (range \54.5–152 U/ml), 1.5 mg/L (range 0.3–6.6 mg/ L), and 9.8 (range 3–23.7), respectively. CSF IL-10 was detected at significant levels in all except one patient with CNS lymphoma, but not detected in any non-lymphoma patients. No patients with CNS lymphoma except one (case L2) had malignant cells (class C3) in cytology analysis. ROC curves for IL-10, IL-10/IL-6 ratio, sIL-2R, b2MG and SUVmax to distinguish CNS lymphomas are shown in Fig. 1. IL-10 showed the highest diagnostic accuracy with AUC = 0.974. The IL-10/IL-6 ratio, sIL-2R, b2MG and SUVmax had diagnostic accuracies of AUC = 0.795,
J Neurooncol (2015) 121:177–183
179
Table 1 Clinical characteristics, CSF and FDG-PET data in CNS lymphoma and non-lymphoma patients No.
Age/Sex
Diagnosis
Location
IL-10 (pg/ml)
IL-6 (pg/ml)
sIL-2R (U/ml)
b2MG (mg/L)
FDG-PET SUVmax
CNS lymphoma patients L1
64/M
Primary DLBCL
Temporal lobe
1,040
17.2
190
7.5
17.4
L2
83/F
Primary DLBCL
Brainstem
278
127
204
11.8
NP
L3
73/F
Primary DLBCL
Cerebellum
116
5.1
63
3.9
NP
L4
78/M
Primary DLBCL
Temporal lobe
59
11.2
507
5.7
30.8
L5
61/M
Primary DLBCL
4th ventricle
86
5.2
299
5.7
21.8
L6
50/M
Primary DLBCL
Parietal lobe
51
10.8
102
2.6
13.6
L7
79/F
Primary DLBCL
3rd ventricle
43
6.8
2,750
9.7
15.8
L8
65/F
Primary DLBCL
Right cp angle
28
6.6
\54.5
2.4
12.4
L9
76/F
Primary DLBCL
Cervical cord
10
3.2
69
6.3
NP
L10
46/M
Primary DLBCL
Frontal lobe
12
179
260
3.2
19.8
L11
53/M
Primary DLBCL
Abducens nerve
8
1.2
498
4.1
9.1
L12
64/M
Primary DLBCL
Corpus callosum
7
12.2
286
3.9
28.9
L13
56/M
Primary DLBCL
Basal ganglia
6
8.9
199
1.7
14.6
L14
70/M
Primary DLBCL
Frontal lobe
3
49.8
344
2.8
25
L15
67/M
Primary DLBCL
3rd ventricle
\2
11.2
130
2.5
10.5
L16
50/M
Secondary DLBCL
Temporal lobe basal ganglia
4,100
68.7
1,000
2.6
17.7
L17
68/F
Secondary DLBCL
Cervical cord
212
4.2
158
2.5
10.5
L18
79/F
Secondary DLBCL
Frontal lobe
17
5.1
225
7.5
18.5
L19
80/F
Secondary DLBCL
Cp angle cerebellum
13
23.8
620
1.8
NP
Non-lymphoma patients N1
70/F
Glioblastoma
Temporal lobe
\2
5.9
\54.5
2
18.9
N2
68/F
Glioblastoma
Frontal lobe
\2
8.2
\54.5
1.4
NP
N3
61/M
Glioblastoma
3rd ventricle
\2
835
\54.5
2.2
11.6
N4
60/F
Glioblastoma
Occipital lobe
\2
3.5
\54.5
1.2
10.9
N5
77/M
Glioblastoma
Parietal lobe
\2
4.5
\54.5
1.9
7.8
N6
65/F
Anaplastic astrocytoma
Corpus callosum
\2
1,790
\54.5
2.6
15.4
N7
46/M
Anaplastic astrocytoma
Frontal lobe
\2
1.8
\54.5
1.1
10.2
N8
59/M
Ependymoma
Cervical cord
\2
1.8
\54.5
1.6
10.5
N9
59/M
Glioma
Pons
\2
1.7
\54.5
1.4
NP
N10
58/F
Metastasis (breast Ca)
3rd ventricle
\2
0.4
\54.5
0.3
6.8
N11
44/F
Metastasis (breast Ca)
Parietal lobe
\2
15.5
\54.5
1.8
5.7
N12
47/F
Metastasis (lung Ca)
Frontal lobe cerebellum
\2
4.3
\54.5
1
NP NP
N13
67/M
Metastasis (lung Ca)
Frontal lobe
\2
4.2
\54.5
2.2
N14
73/M
Metastasis (colon Ca)
Frontal lone
\2
1.3
\54.5
0.3
4.6
N15
21/M
Immature teratoma
Suprasellar region
\2
271
\54.5
3.3
NP
N16
60/M
Atypical meningioma
Cp angle
\2
1.6
\54.5
1.1
8.5
N17
39/F
Undifferentiated sarcoma
Cavernous sinus
\2
18.5
\54.5
1.3
9.8 3.58
N18
44/F
Multiple sclerosis
Basal ganglia
\2
3.2
\54.5
1.7
N19
36/M
Multiple sclerosis
Pons
\2
1.3
60.4
0.9
6.6
N20
30/M
Multiple sclerosis
Cervical cord
\2
2.3
67.1
1
NP
N21
26/M
Sarcoidosis
Frontal lobe temporal lobe
\2
62.9
100
2.1
23.7
N22
69/M
Hypertrophic pachymeningitis
Cervical spine
\2
2
\54.5
1.2
NP
N23
52/F
Myelitis
Cervical cord
\2
5.8
\54.5
1.1
3
N24
69/M
Leukoencephalopathy
Frontal lobe
\2
9.2
152
6.6
NP
N25
61/F
Rosai–Dorfman disease
Temporal lobe
\2
3.4
70
1.8
NP
N26
37/M
Cerebral vasculitis
Parietal lobe
\2
3
\54.5
2.1
21.8
CSF cerebrospinal fluid, FDG-PET fluorodeoxyglucose-positron emission tomography, IL-10 interleukin, 10 IL-6 interleukin 6, sIL-2R soluble interleukin 2 receptor, b2MG beta-2 microglobulin, DLBCL diffuse large B-cell lymphoma, NP not performed, cp cerebellopontine, Ca cancer
123
180
J Neurooncol (2015) 121:177–183
Fig. 1 ROC curves for IL-10, IL-10/IL-6 ratio, b2MG, sIL-2R and FDG-PET SUVmax for differentiation of CNS lymphomas from nonlymphomas. IL-10 showed the highest diagnostic accuracy with AUC = 0.974
0.950, 0.908 and 0.806, respectively. The sensitivity and specificity of each parameter using the optimal cutoff value are shown in Table 2. IL-10 and sIL-2R had the highest sensitivity of 94.7 % and IL-10 had the highest specificity of 100 %. Illustrative cases Case L13 (Fig. 2): A 56-year-old man presented with a 3-week history of recent memory disturbance and difficulty walking. MRI revealed a heterogeneously enhanced tumor with peritumoral edema in the right basal ganglia. A radiologist diagnosed the tumor as glioblastoma based on MRI findings and the patient was admitted to our hospital. No primary lesion of the metastasis was revealed on 18FFDG PET, and SUVmax of the tumor was 14.6. The CSF levels of IL-10, IL-6, sIL-2R and b2MG were 6 pg/ml, 8.9 pg/ml, 199 U/ml, and 1.7 mg/L, respectively. Frontal craniotomy with an interhemispheric transcallosal approach was performed to obtain a pathological diagnosis. The biopsy sample contained lymphoma cells and the pathological diagnosis was diffuse large B-cell lymphoma. Chemotherapy with high-dose methotrexate was initiated 8 days postoperatively. MRI after two courses of chemotherapy showed disappearance of the tumor and the patient made a complete recovery.
Case N6 (Fig. 3): A 65-year-old woman presented with a 2-week history of recent memory disturbance. MRI revealed a homogeneously enhanced tumor in the corpus callosum. No primary lesion of the metastasis was revealed on 18F-FDG PET, and SUVmax of the tumor was 15.4. A radiologist diagnosed the tumor as primary CNS lymphoma based on MRI and 18F-FDG PET findings. The CSF levels of IL-10, IL-6, sIL-2R and b2MG were \2 pg/ml, 1,790 pg/ml, \54.5 U/ml, and 2.6 mg/L, respectively. An endoscopic biopsy was performed via the right anterior horn of the lateral ventricle to obtain a pathological diagnosis. The biopsy sample contained glioma cells and the pathological diagnosis was anaplastic astrocytoma. Chemotherapy with temozolomide and extended local radiation with 60 Gy was initiated 12 days postoperatively. MRI after initial chemoradiotherapy revealed reduction of the tumor and the patient made a slight recovery.
Discussion The results of this study show that the CSF IL-10 level has diagnostic value for distinguishing CNS lymphomas from other CNS lesions. CSF IL-10 levels were elevated in almost all patients with CNS lymphomas, but were undetectable in patients with non-lymphomas, including CNS tumors or inflammatory diseases. This biomarker was superior to traditional parameters used in initial screening for CNS lymphoma patients. In ocular lymphomas, an IL-10/IL-6 ratio [1.0 in vitreous specimens has been suggested as a diagnostic criterion. The sensitivity based on this criterion is 75–90 % [8, 9, 11]. Thus, the sensitivity of 68.4 % in patients with CNS lymphoma in the current study was inferior to that reported for ocular lymphoma. In CNS lymphoma, two other recent studies of the diagnostic value of IL-10 alone in cohorts of moderate size have been reported (Table 3). In an analysis of CSF IL-10 in 26 patients with PCNSL and 40 patients with other brain tumors, Sasayama et al. showed that CSF IL-10 may be a useful diagnostic biomarker for PCNSL [12], since the CSF IL-10 was significantly higher in PCNSL than in other brain tumors. The sensitivity and specificity at an IL-10 cutoff level of 9.5 pg/ml were 71.0 and 100 %, respectively [12], and thus the sensitivity was lower than that of 94.7 % found in the present study. The
Table 2 Sensitivity, specificity and optimal cutoff values of parameters for differentiation of CNS lymphomas from non-lymphomas IL-10
IL10/IL-6 ratio
sIL-2R
b2MG
FDG-PET SUVmax
Sensitivity (%)
94.7
68.4
94.7
89.4
80
Specificity (%)
100
96.1
84.6
88.5
Cutoff value
3 pg/ml
123
2.2
60.4 U/ml
2.4 mg/L
75 12.4
J Neurooncol (2015) 121:177–183
181
Fig. 2 a Gadolinium-enhanced T1-weighted MRI showed a heterogeneously enhanced tumor in the right basal ganglia. b Hematoxylin and eosin staining of a specimen from the tumor showed proliferation of lymphoma cells (original magnification 9100)
difference in sensitivity may be due to the use of different cutoff values in the two studies (9.5 vs. 3 pg/ml). Also, some T cell and unclassified lymphomas were included in the population in Sasayama et al., and these type of lymphomas have undetectable CSF IL-10 levels [12]. Five of the 24 PCNSL cases had CSF IL-10 at an undetectable level, including two cases of DLBCL [12]. CSF IL-10 was also undetectable in one lymphoma in the present study. Thus, the detection limit (2 ng/ml) for IL-10 analysis may cause a false negative result. A specificity of 100 % was found in Sasayama et al. [12], consistent with our result. These results suggest that screening using the CSF IL-10 level may be useful for selection of a subsequent surgical procedure, as described below. Rubenstein et al. [13] examined the diagnostic value of CSF IL-10 in 53 newly diagnosed cases of CNS lymphoma (43 primary and 10 secondary lymphomas) and 91 control patients (8 cases of other primary brain tumors, 12 metastatic brain tumors, and 71 neuroinflammatory/infection diseases). This study included a relatively large cohort, but the sensitivity was lower and the median IL-10 levels were higher than those in Sasayama et al. and in the current study. The median IL10 levels were 282.9 ng/ml in PCNSL patients in Rubenstein et al., compared to 27 ng/ml in Sasayama et al. and 28 ng/ml in our study. In addition, CSF IL-10 was detected in non-lymphoma patients in Rubenstein et al., but not in the other two studies. The reasons for the differences in IL10 levels among the three studies are unclear, but these levels may have been affected by the treatment method and the antibodies used in the measurement. In our study, CSF sIL-2R also had higher sensitivity. Thus, the combination of multiple biomarkers such as IL-10 plus sIL-2R may give
a more accurate diagnosis. Regardless, there is a need to accumulate studies to determine the optimal cutoff value for CSF IL-10 to differentiate CNS lymphomas from nonlymphomas. CNS lymphomas are difficult to distinguish from other CNS tumors and from rare inflammatory diseases such as multiple sclerosis, neurosarcoidosis and cerebral vasculitis [16, 17] because inflammatory diseases occur in multifocal and other patterns that are similar to CNS lymphomas. However, CSF IL-10 was undetectable in inflammatory diseases, while b2MG, sIL-2R and FDG-PET SUVmax were elevated in some cases. Cytokines are deregulated and play important roles in the pathogenesis of various cancers via autocrine or paracrine mechanisms [18]. IL-10 was first described as a growth and differentiation factor for B lymphocytes that induced activated B-cells to secret large amounts of immunoglobulin [19]. IL-10 also has broad anti-inflammatory properties through suppression of macrophage and dendritic cell function [20]. The IL-10 level has also been associated with inflammation in overweight and obese patients [21], but our cases did not include obese patients (body mass index C30). In malignant lymphomas, IL-10 is produced by lymphoma cells and further enhances proliferation of these cells [22], although the exact role of IL-10 in proliferation of CNS lymphoma has yet to be established. In neurosurgery, preoperative diagnosis is very important because the surgical method depends on the tumor type. Gross total resection is the standard procedure for malignant gliomas or metastatic brain tumors, whereas biopsy for pathological diagnosis is sufficient in CNS lymphoma because chemoradiotherapy is effective [2]. Therefore, burr hole or minimum craniotomy can be
123
182
J Neurooncol (2015) 121:177–183
Fig. 3 a Gadolinium-enhanced T1-weighted MRI showed a homogeneously enhanced tumor in the corpus callosum. b 18F-FDG PET showed high FDG uptake (SUVmax: 15.4) in the tumor.
c Hematoxylin and eosin staining of a specimen from the tumor showed proliferation of glioma cells (original magnification 9200)
Table 3 Summary of studies of median level, cuttoff value, sensitivity and specificity of IL-10 for differentiation of CNS lymphomas from nonlymphomas Author (year)
Median IL-10 level (pg/ml)
Cutoff value (pg/ml)
Sensitivity (%)
Specificity (%)
Sasayama [12]
27: PCNSL (n = 26) \ 2.0: other BTs (n = 40)
9.5
71
100
Rubenstein [13]
282.9: PCNSL (n = 43) 57: SCNSL (n = 10) 10.9: other primary BTs (n = 8) 5.3: metastatic BTs (n = 12) 5.6: neuro-infla/infe (n = 71)
16.2
64
94.1
Current study (2014)
28: PCNSL (n = 15) 114.5: SCNSL (n = 4) \ 2.0: other BTs (n = 17) \ 2.0: neuro-infla (n = 9)
3
94.7
100
PCNSL primary central nervous system lymphoma, n number, BTs brain tumors, SCNSL secondary central nervous system lymphoma, neuroinfla neuro-inflammatory diseases, neuro-infe neuro-infection diseases
selected for a preoperative diagnosis of CNS lymphoma [12]. Thus, preoperative measurement of CSF IL-10 can give useful information for planning of a neurosurgical procedure.
123
In summary, our results show the diagnostic value of the CSF IL-10 level in patients with CNS lymphomas, especially DLBCL. The sample size limits the strength of our conclusions and accumulation of studies is required to give
J Neurooncol (2015) 121:177–183
definitive conclusion on the impact of CSF IL-10 in diagnosis of CNS lymphoma. Acknowledgments This work was supported by a Grant for Promoted Research from Kanazawa Medical University (S2013-7). Conflict of interest of interest.
The authors declare that they have no conflict
References 1. Olson JE, Janney CA, Rao RD, Cerhan JR, Kurtin PJ, Schiff D, Kaplan RS, O’Neill BP (2002) The continuing increase in the incidence of primary central nervous system non-Hodgkin lymphoma: a surveillance, epidemiology, and end results analysis. Cancer 95:1504–1510 2. Pels H, Juergens A, Schirgens I, Glasmacher A, Schulz H, Engert A, Schackert G, Reichmann H, Kroschinsky F, Vogt-Schaden M, Egerer G, Bode U, Deckert M, Fimmers R, Urbach H, SchmidtWolf IG, Schlegel U (2010) Early complete response during chemotherapy predicts favorable outcome in patients with primary CNS lymphoma. Neuro Oncol 12:720–724 3. Kosaka N, Tsuchida T, Uematsu H, Kimura H, Okazawa H, Itoh H (2008) 18F-FDG PET of common enhancing malignant brain tumors. AJR Am J Roentgenol 190:365–369 4. Yamashita K, Yoshiura T, Hiwatashi A, Togao O, Yoshimoto K, Suzuki SO, Abe K, Kikuchi K, Maruoka Y, Mizoguchi M, Iwaki T, Honda H (2013) Differentiating primary CNS lymphoma from glioblastoma multiforme: assessment using arterial spin labeling, diffusion-weighted imaging, and 18F-fluorodeoxyglucose positron emission tomography. Neuroradiology 55:135–143 5. Scott BJ, Douglas VC, Tihan T, Rubenstein JL, Josephson SA (2013) A systematic approach to the diagnosis of suspected central nervous system lymphoma. JAMA Neurol 70:311–319 6. Cassoux N, Giron A, Bodaghi B, Tran TH, Baudet S, Davy F, Chan CC, Lehoang P, Merle-Be´ral H (2007) IL-10 measurement in aqueous humor for screening patients with suspicion of primary intraocular lymphoma. Investig Ophthalmol Vis Sci 48:3253–3259 7. Cassoux N, Merle-Beral H, Leblond V, Bodaghi B, Mile´a D, Gerber S, Fardeau C, Reux I, Xuan KH, Chan CC, LeHoang P (2000) Ocular and central nervous system lymphoma: clinical features and diagnosis. Ocul Immunol Inflamm 8:243–250 8. Kimura K, Usui Y, Goto H; Japanese Intraocular Lymphoma Study Group (2012) Clinical features and diagnostic significance of the intraocular fluid of 217 patients with intraocular lymphoma. Jpn J Ophthalmol 56:383–389
183 9. Wolf LA, Reed GF, Buggage RR, Nussenblatt RB, Chan CC (2003) Vitreous cytokine levels. Ophthalmology 110:1671–1672 10. Chan CC, Buggage RR, Nussenblatt RB (2002) Intraocular lymphoma. Curr Opin Ophthalmol 13:411–418 11. Chan CC (2003) Molecular pathology of primary intraocular lymphoma. Trans Am Ophthalmol Soc 101:275–292 12. Sasayama T, Nakamizo S, Nishihara M, Kawamura A, Tanaka H, Mizukawa K, Miyake S, Taniguchi M, Hosoda K, Kohmura E (2012) Cerebrospinal fluid interleukin-10 is a potentially useful biomarker in immunocompetent primary central nervous system lymphoma (PCNSL). Neuro Oncol 14:368–380 13. Rubenstein JL, Wong VS, Kadoch C, Gao HX, Barajas R, Chen L, Josephson SA, Scott B, Douglas V, Maiti M, Kaplan LD, Treseler PA, Cha S, Hwang JH, Cinque P, Cyster JG, Lowell C (2013) CXCL13 plus interleukin 10 is highly specific for the diagnosis of CNS lymphoma. Blood 121:4740–4748 14. Zweig MH, Campbell G (1993) Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin Chem 39:561–577 15. Branscum AJ, Gardner IA, Johnson WO (2005) Estimation of diagnostic-test sensitivity and specificity through Bayesian modeling. Prev Vet Med 68:145–163 16. Trojano M, Paolicelli D (2001) The differential diagnosis of multiple sclerosis: classification and clinical features of relapsing and progressive neurological syndromes. Neurol Sci 22(Suppl 2):98–102 17. Spencer TS, Campellone JV, Maldonado I, Huang N, Usmani Q, Reginato AJ (2005) Clinical and magnetic resonance imaging manifestations of neurosarcoidosis. Semin Arthritis Rheum 34:649–661 18. Li N, Grivennikov SI, Karin M (2011) The unholy trinity: inflammation, cytokines, and STAT3 shape the cancer microenvironment. Cancer Cell 19:429–431 19. O’Garra A, Stapleton G, Dhar V, Pearce M, Schumacher J, Rugo H, Barbis D, Stall A, Cupp J, Moore K (1990) Production of cytokines by mouse B cells: b lymphomas and normal B cells produce interleukin 10. Int Immunol 2:821–832 20. Moore KW, de Waal Malefyt R, Coffman RL, O’Garra A (2001) Interleukin-10 and the interleukin-10 receptor. Annu Rev Immunol 19:683–765 21. McGuire TR, Brusnahan SK, Bilek LD, Jackson JD, Kessinger MA, Berger AM, Garvin KL, O’Kane BJ, Tuljapurkar SR, Sharp JG (2011) Inflammation associated with obesity: relationship with blood and bone marrow endothelial cells. Obesity (Silver Spring) 19:2130–2136 22. Spits H, de Waal Malefyt R (1992) Functional characterization of human IL-10. Int Arch Allergy Immunol 99:8–15
123
