Central nervous system (CNS) involvement in patients diagnosed with non-Hodgkin lymphoma (NHL) occurs in about 5% of all patients, either at the time of diagnosis of NHL or during the course of progressive disease. In addition to the well-recognized poor prognosis of the patients with NHL and CNS involvement, this complication has been associated with a reduction in the quality of life due to both CNS involvement itself and to the side effects of CNS-directed therapies such as cranial irradiation, intrathecal (IT) drugs and systemic therapy with drugs able to cross the blood-brain barrier1,2.
Risk factors for lymphomatous meningeosis have been identified from the analysis of large case series3-11. Histology is cited as one of the main risk factors, and thus, patients with certain high-grade histologic subtypes such as lymphoblastic lymphoma and Burkitt's lymphoma systematically received CNS prophylaxis because of the well known propensity of these lymphomas to involve the CNS. In contrast, the risk of meningeal involvement from unstransformed low-grade NHL is low, and CNS prophylaxis is generally not recommended1,2,12. Controversy on CNS prophylaxis is open in other aggressive lymphomas, especially diffuse large B-cell lymphomas (DLBCL), in which no consensus on CNS prophylaxis in specific situations has been reached. On the other hand, the new diagnostic tools for the detection of meningeal involvement, such as flow cytometry of cerebrospinal fluid (CSF), have led to the detection of a greater number of patients with occult CNS disease compared to conventional cytology13-15. In some of these studies patients with occult CNS disease showed a trend to high-risk of CNS relapse13.
In addition to this poorly clarified situation, and contrary to the proven efficacy of CNS prophylaxis in patients with acute lymphoblastic leukemia (ALL), conclusions regarding the efficacy of CNS prophylaxis on CNS disease remain controversial in some studies1,16,17. This is mainly due to the heterogeneity in drugs and schedules administered, the small number of patients and/or the lack of a control arm in most studies.
Besides the heterogeneity for indications of CNS prophylaxis, schedules of prophylaxis and treatment of CNS involvement in NHL homogeneous in different countries are lacking, even within the same country, and have been scarcely investigated18,19. The QUIT (Registro Español de Pacientes que Reciben Quimioterapia Intratecal) registry was an initiative of the PETHEMA (Programa Español de Tratamiento de las Hemopatías Malignas, Asociación Española de Hematología y Hemoterapia), GEL/TAMO (Grupo Español de Linfomas/Trasplante Autólogo de Médula Ósea) and GOTEL (Grupo Oncológico para el Tratamiento y Estudio de los Linfomas) groups designed to prospectively know the current practice of CNS prophylaxis and therapy in patients diagnosed with hematological malignancies (acute leukemias [AL] and lymphoproliferative disorders) in Spain. The results in AL patients have recently been reported20. The objective of this study was to report the practice of CNS prophylaxis and treatment in patients with lymphoproliferative disorders.
Patients and methods
The QUIT study was presented at meetings of the Spanish PETHEMA, GELTAMO and GOTEL groups to select participant hospitals as has been previously described20. Briefly, from June 2005 to June 2006 adult patients ($ 18 yr) diagnosed with NHL or other lymphoproliferative disorders who received CNS prophylaxis or CNS treatment were consecutively included through online registration in 33 hospitals. Each patient was included only once. The questionnaire included sociodemographic variables (age, sex, hospital), NHL subtype (or type of lymphoproliferative disorder according to WHO classification)21 and date of diagnosis, stage (Ann Arbor)22, B symptoms, ECOG score, presence or absence of extranodal involvement (testis, liver, lung, gastrointestinal tract, bone marrow, skin, paranasal sinus, orbita and others) and number of extranodal sites and lymph node regions involved, bulky mass (over 10 cm in diameter), serum levels of lactate dehydrogenase (LDH) (raised or normal) and International Prognostic Index (IPI) score23.
With regard to meningeal involvement, the timing of the CNS involvement (at diagnosis or at relapse), date of diagnosis of neoplastic meningeosis, neurological symptoms at the time of CNS involvement (including headache, nausea and vomiting, mental status abnormalities, muscular weakness, impairment of march, sensorial neuropathy, dyplopia and others) as well as the method used for diagnosis of CNS involvement (detection of blast cells in a sample of CSF by cytology or by flow cytometry, computed tomography [CT] scan or magnetic resonance imaging [MRI]) were recorded.
The following data of CNS therapy were recorded: reason of CNS-directed therapy (prophylaxis or treatment, and in the case of prophylaxis, the reason [hystologic subtype, extranodal involvement (1 or $ 2), site of extranodal involvement, raised LDH, age > 60 yr, bone marrow involvement, bulky mass, intermediate-high IPI score, and others]), date of the onset of IT therapy (methotrexate [MTX], cytarabine, triple therapy [including MTX, cytarabine and hydrocortisone]), route (IT through lumbar puncture or intraventricular) and schedule of administration, as well as other CNS-directed therapies such as cranial irradiation or the use of new drugs.
As reported for the results of AL patients included in the QUIT study20, an intermediate analysis was performed 6 months after starting the registry and the results were reported at meetings of PETHEMA, GEL/TAMO and GOTEL groups. Results were expresed as crude frequencies and percentages, and were analysed separately for CNS prophylaxis and therapy. For the comparison of type and route of administration of IT therapy between patients receiving CNS prophylaxis and therapy, a Fisher exact test was used. A p value of less than 0.05 was considered statistically significant.
A total of 228 patients diagnosed with lymphoproliferative disorders from 33 hospitals were included. The mean (standard deviation [SD]) age was 52 (16) yr. (range 18-83), with 63 (34%) patients older than 60 years, and 144 (63%) males. DLBCL was the most frequent histologic subtype (133 patients [58%]), followed by Burkitt's lymphoma (37 patients [16%]) and MCL (11 cases [4.8%]). Table 1 provides a complete list of the lymphoproliferative disorders. At the time of diagnosis, half of the patients presented with B symptoms and three quarters had advanced disease (stage IV in 149 [65%]). A total of 228 patients (79%) showed extranodal involvement, being bone marrow the most frequent extranodal site involved (81 patients) followed by CNS, liver and lung (30 cases each one). The LDH level was increased in 63% of the cases and 62% had an IPI score higher than 2. The main demographic and clinical characteristics are listed in table 2.
IT therapy was administered for neoplastic meningeosis in 41 cases and as CNS prophylaxis in 187. Thirty out of 41 patients showed CNS involvement at the time of diagnosis of lymphoproliferative disorders, whereas 11 showed CNS relapse. Diagnosis of CNS involvement was established by the observation of malignant cells in CSF in 33 cases, with a median of cell count of 18 cells/µl (range 2-100), or by imaging techniques (MRI in 15 and CT scan in 8). In 17 patients (41%) CSF involvement was confirmed by flow cytometry. Neurological symptoms included: impairment in march in 18, headache in 17, mental status impairment in 14 cases, cranial nerve palsy in 14, muscle weakness in 12, nausea and vomiting in 10, diplopy in 9, sensorial neuropathy in 7, loss of vision in 6 and seizures in 1.
As can been observed in table 3, the most common CNS-directed therapy reported in cases with CNS involvement was TIT, although it is of note that almost half of the patients were treated with IT depot liposomal cytarabine. Regarding IT administration schedule, in most of the cases (n = 15) IT therapy was administered every 15 days (to almost all the patients corresponding to those receiving IT depot liposomal cytarabine) or 2-3 times weekly, whereas lumbar puncture was the preferred route for IT administration (37 out of the 41 patients). Cranial irradiation was used in 4 cases (local radiotherapy in 3 and craniospinal radiotherapy in the remaining).
In the group of patients who received CNS prophylaxis (n = 187), TIT was also the most frequent schedule (table 3), followed by IT MTX. However, significant differences were observed in the type of IT therapy administered as prophylaxis and therapy (table 3), and, in fact, only three patients received IT depot liposomal cytarabine as CNS prophylaxis. Two were patients diagnosed with DLBCL (one with elevated LDH) whereas the remaining was a patient diagnosed with follicular lymphoma who had involvement of 2 extranodal sites, raised LDH and bulky disease at diagnosis. In most of the patients in which the schedule of administration was available, IT prophylaxis was administered with each cycle of chemotherapy, followed by those who received just one IT administration, probably during screening lumbar puncture. On the other hand, although the most frequent route for IT administration was, again, lumbar puncture, an Ommaya reservoir was used for intraventricular administration in only one patient. Radiotherapy was administered as CNS prophylaxis in 2 cases. The main reasons for CNS prophylaxis cited by the investigators included extranodal involvement in 89 patients, raised LDH level in 87, IPI score higher than 2 in 62, bulky mass in 43, extranodal involvement in more than one organ in 33, age over 60 yr in 28 and human immunodeficency virus (HIV) infection in 13. Table 4 shows reasons for CNS prophylaxis by the main histologic subtypes.
Similar to a previous study by our group focused on CNS prophylaxis and therapy for AL patients20, this is the first study that prospectively describes the current practice of prophylaxis and therapy for neoplastic meningeosis by lymphomas and other lymphoproliferative disorders in Spain. As observed in AL study, the most frequent therapy for CNS involvement or CNS prophylaxis was TIT, and lumbar puncture was the preferred route for IT administration. The introduction of new drugs, especially liposomal depot cytarabine for therapy of CNS infiltration and the scarce use of radiotherapy are also of note. On the other hand, and conversely to patients with AL, one of the most relevant findings of this study was the marked differences between investigators in the indications of CNS prophylaxis for most of the cases, reflecting the disparity in criteria for CNS prophylaxis, a feature also described in other studies18,19.
In most of the studies on the risk of CNS relapse in patients with lymphoma, the histologic subtype has been indicated as one of the main risk factors. Thus, there is a consensus on the need of CNS-directed therapy in some aggressive lymphomas such as Burkitt's lymphoma and lymphoblastic lymphoma, in which the risk of CNS relapse in the absence of prophylaxis could achieve 20%-25%8,12,24. This explains why, in the QUIT registry, patients diagnosed with these histologies received homogeneous IT prophylaxis following specific Spanish protocols that include systematic IT prophylaxis, generally with TIT25. Conversely, the risk of CNS relapse in the untransformed low-grade lymphomas is low8, and there is no evidence for the routine use of CNS prophylaxis, with the exception of high-grade transformation4. However, 10 patients diagnosed with follicular or marginal zone lymphoma in the QUIT registry received IT prophylaxis, mostly due to the presence of other risk factors for CNS disease identified in studies focused on aggressive NHL (table 4).
The controversy in the criteria used for CNS prophylaxis especially arises in other aggressive lymphomas such as the blastic variant of MCL, anaplastic large cell lymphoma, peripheral T cell lymphomas and, especially, DLBCL. Case-series studies have estimated that the risk of CNS relapse for patients diagnosed with DLBCL is about 5%5,8,9. The use of systematic prophylaxis in these cases could lead to the treatment of patients who may never have CNS relapse. For this reason, many of these studies have tried to identify which of these patients are at risk of developing CNS disease based on the presence of other risk factors. One of the largest studies is that by Hollender et al8 in which more than 2,500 patients diagnosed with NHL were studied for the incidence and risk factors for CNS disease. These authors described a score system based on the findings of five independent risk factors: raised serum LDH level, serum albumin over 35 g/L, age less than 60 yr, retroperitoneal lymph node involvement and involvement of more than one extranodal site. The presence of 4 or 5 of these factors increased the risk of CNS recurrence by 25%. In the study by Haioun et al7, the risk of CNS relapse increased in patients with raised LDH or with involvement of more than one extranodal site; although when IPI was added to multivariate analysis it remained as the only parameter with statistical significance, identical to the report by Feugier et al9. In a recent article10, increased serum LDH level or involvement of more than one extranodal site were, again, the main risk factors associated with higher CNS relapse.
Although many patients with DLBCL included in the QUIT registry received CNS prophylaxis according to some of the above mentioned reasons (raised LDH levels and intermediate-high IPI score), some findings were of note. The first is that the main criteria for CNS prophylaxis was the presence of only one extranodal site in 55 patients with this histologic subtype, a risk factor not cited in the aforementioned studies. In fact, this finding probably reflects the trend to use CNS prophylaxis in patients with involvement of anatomical sites associated, in some studies, with a higher risk of CNS involvement, such as bone marrow3, paranasal sinuses26, breast27, epidural space28,29 or testicular involvement30,31, yet the significance of this increased involvement is difficult to detect by multivariate analysis12. In this sense, the guidelines of the National Comprehensive Cancer Network (NCCN)32 recommend CNS prophylaxis in patients with DLBCL and involvement of these mentioned sites. Identical results to those of our study were observed in similar studies conducted in the UK18 and Canada19, in which only a minority of clinicians used a high IPI score, increased serum LDH level or more than one extranodal site as criteria for CNS prophylaxis. The second remarkable finding in our study is that bulky disease was also mentioned as a reason for CNS prophylaxis in 24 patients with DLBCL. Bulky disease has not been associated with a high-risk of CNS involvement and, in fact, in the study by Tomita et al33, CNS relapse was observed more frequently in the group of patients with less frequency of bulky masses. Third, HIV infection was included as a reason for CNS prophylaxis in 13 cases (12 diagnosed with DLBCL), despite the results of the study by Desai et al34, suggesting that CNS prophylaxis should be administered only for aggressive lymphomas or bone marrow involvement. In addition, in one study35, the frequency of CNS involvement had significantly decreased in patients with HIV-related lymphomas who were receiving highly active antiretroviral therapy (HAART). Finally, although a high risk of CNS disease has been associated with younger age3,8,28, in our study age over 60 yr was included as criteria for CNS prophylaxis in 28 patients (23 diagnosed with DLBCL), probably reflecting the fact that advanced age is a risk factor included in the IPI score.
As described for patients diagnosed with AL20, there was a massive use of TIT as the prefered schedule for CNS prophylaxis, conversely to reports in other foreign studies in which IT MTX alone or combined with soluble steroids was the most frequent IT therapy7,8,18,33. The reason has been previously discussed20 and was a consequence of the wide use of the risk-adapted protocols from the PETHEMA group in Spain for treatment of patients with ALL, Burkitt's and lymphoblastic lymphoma, that could exert a mimetic effect leading to a generalization of this schedule for CNS prophylaxis for the remaining lymphoid malignancies.
Another remarkable feature of the QUIT study was the administration of IT depot liposomal cytarabine in almost 50% of the cases of lymphomas with CNS involvement. Due to its long half time in CSF, IT depot liposomal cytarabine can be administered every two weeks, allowing fewer IT administrations36,37. The efficacy of this drug in lymphomatous meningeosis has been established in several studies38,39, and it indeed constitutes an accepted indication by the regulatory agencies. It is note, however, that 3 patients also received IT depot liposomal cytarabine as CNS prophylaxis despite the scarce information available in this setting. Mc Clune et al40 included 14 patients with ALL or high-grade non-Hodgkin's lymphoma treated with systemic chemotherapy. IT depot liposomal cytarabine was administered in 40 out of 81 cycles as CNS prophylaxis and at the time of the analysis no CNS relapse was observed. Several ongoing trials are attempting to establish the role of IT liposomal cytarabine in the prophylaxis of lymphomatous meningeosis41.
The incorporation of new diagnostic tools for CNS involvement in the QUIT study is of note. Although in most of the cases of our registry (33 patients) the diagnosis was based on the presence of atypic lymphoid cells detected in a cytocentrifuged sample of CSF, more sensitive techniques for the detection of neoplastic cells in CSF, such as flow cytometry for the detection of aberrant phenotypes were employed. In our study, flow cytometry confirmed the presence of neoplastic cells in CSF in 17 out of 41 patients with CNS involvement. The use of these more sensitive techniques has been associated with the detection of more frequent CSF involvement compared to conventional cytology13-15. In a recent study by Hedge et al13, the detection of CSF disease by flow cytometry was associated with the presence of more than one extranodal site, one of the main risk factors for CNS relapse reported in several studies5,7,8,10, suggesting that patients with CNS involvement detected by cytometry could be at risk for meningeal relapse. In addition, among the patients with evidence of occult CSF lymphoma by flow cytometry, CNS relapse was observed in 45% vs 8% in patients with negative CSF flow cytometry13. It is of note, however, that this technique is not currently accepted as standard for CNS involvement detection, but could become a standard procedure in the future.
The lack of the systematic inclusion of all the cases in which CNS-directed therapy was administered constitutes the main limitation of the present study. However, the design of this survey (cross sectional and consecutive) and the similar results obtained in other surveys conducted in foreign countries18,19 indicate that this registry probably was a representative sample of the practice of CNS prophylaxis and treatment in patients with lymphoproliferative disorders in Spain.
In conclusion, the results of this study point out the generalized use of IT therapy both for CNS prophylaxis and therapy in patients with NHL and other lymphoproliferative disorders, as well as the increasing use of new formulations of drugs, such as IT depot liposomal cytarabine, and the scarce use of radiotherapy. As has been reported in similar studies, the absence of homogeneous criteria for CNS prophylaxis in some aggressive lymphomas, especially DLBCL, is of note.
The following institutions and clinicians participated in the study: Hospital Universitari Germans Trias i Pujol, Badalona: J.M. Ribera, J.M. Sancho, M. Morgades. Complejo Hospitalario Universitario de Santiago, Santiago de Compostela: N. Alonso. Hospital Juan Canalejo, La Coruña: G. Deben. ICO-Hospital Duran i Reynals, L'Hospitalet de Llobregat: A. Fernández de Sevilla. Hospital Universitario de Salamanca, Salamanca: L. Vázquez. Hospital Universitario Central de Asturias, Oviedo: C. Nicolás. Hospital Universitario de Getafe, Madrid: J.A. García Vela. Hospital Universitario de la Princesa, Madrid: R. Arranz. Hospital del Mar, Barcelona: E. Abella. Hospital Universitario La Paz, Madrid: M.A. Canales. Hospital General Universitario Gregorio Marañón, Madrid: P. Miralles, J. Berenguer. Hospital Vall d'Hebron, Barcelona: E. Sánchez García. Hospital Josep Trueta, Lleida: M. Hermosilla. Clínica Universitaria de Navarra, Pamplona: C. Panizo. Hospital 12 de Octubre, Madrid: R. Toscazo. Fundación Jiménez Díaz, Madrid: F. Lobo. Hospital Universitario Príncipe de Asturias, Madrid: J. García Suárez, D. de Miguel. Hospital Universitario de Canarias, Santa Cruz de Tenerife: M.J. Rodríguez Salazar, M. Llanos. Hospital General Universitario de Alicante, Alicante: P. Fernández-Abellán. Hospital Marqués de Valdecilla, Santander: E. Conde. Hospital Xeral, Lugo: J. Arias. Complejo Hospitalario de Jaén, Jaén: J.A. López. Hospital Universitario La Fe, Valencia: M. Blanes. Hospital Universitario Miguel Servet, Zaragoza: A. Rubio. Hospital de Fuenlabrada, Madrid: J.A. Hernández. Hospital Virgen del Camino, Pamplona: M.C. Mateos. Complexo Hospitalario Universitario de Vigo, Pontevedra: C. Albo. Hospital Clínico Universitario de Valladolid, Valladolid: F.J. Fernández. Hospital de Pontevedra, Pontevedra: F.R. García Arroyo. Hospital de Basurto, Bilbao: J.A. Márquez. Hospital Clínico de Zaragoza, Zaragoza: L. Palomera. Hospital Virgen de la Arrixaca, Murcia: P. Rosique. Hospital Clínico Virgen de la Victoria, Málaga: A. Rueda.