Fax: (011) 81-47-099-1198
High frequency of neurolymphomatosis as a relapse disease of intravascular large B-cell lymphoma
Article first published online: 29 MAR 2011
Copyright © 2011 American Cancer Society
Volume 117, Issue 19, pages 4512–4521, 1 October 2011
How to Cite
Matsue, K., Hayama, B. Y., Iwama, K.-I., Koyama, T., Fujiwara, H., Yamakura, M., Takeuchi, M. and O'uchi, T. (2011), High frequency of neurolymphomatosis as a relapse disease of intravascular large B-cell lymphoma. Cancer, 117: 4512–4521. doi: 10.1002/cncr.26090
- Issue published online: 16 SEP 2011
- Article first published online: 29 MAR 2011
- Manuscript Accepted: 7 FEB 2011
- Manuscript Revised: 21 JAN 2011
- Manuscript Received: 29 SEP 2010
- intravascular large B-cell lymphoma;
- 18F-fluoro-2-deoxy-D-glucose-positron emission tomography/computed tomography;
- magnetic resonance imaging;
- high-dose methotrexate
Intravascular large B-cell lymphoma (IVL) is characterized by lymphoma cell proliferation in the lumina of small vessels in various organs. A high incidence of neurologic symptoms associated with the central nervous system has been reported, but peripheral nerve involvement (neurolymphomatosis [NL]) rarely has been described.
The medical records from patients who were diagnosed with IVL over the past 4 years were reviewed. A diagnosis of NL was made based on the combination of neurologic symptoms and their correspondence with imaging studies, such as magnetic resonance imaging (MRI), 18F-fluoro-2-deoxy-D-glucose (FDG)-positron emission tomography/computed tomography (PET/CT), and/or the histologic confirmation of lymphoma cells within the peripheral nerves, nerve root/plexuses, or cranial nerves.
Four patients with NL were identified among 11 patients who had IVL. All cases of NL occurred as relapsed disease during or shortly after the completion of chemotherapy. Although MRI studies of the brains and whole spines revealed nerve infiltration by gadolinium enhancement in 2 patients, the technology was not sensitive enough to detect such infiltration in the remaining 2 patients. In contrast, FDG-PET/CT studies successfully revealed cranial or peripheral nerve lesions in all 4 patients and was useful for evaluating therapeutic response. Patients received treatment with high-dose methotrexate with or without other systemic chemotherapy, which achieved varied success. Further studies will be needed to determine the optimal treatment.
Considering the rarity of IVL and NL, the current observations suggested that IVL may have a predilection not only for the vessels but also for both the central and peripheral nervous systems. Cancer 2011. © 2011 American Cancer Society
Intravascular lymphoma (IVL) is a rare type of extranodal, large B-cell lymphoma that is characterized by the preferential proliferation of neoplastic cells in the lumina of small vessels with no tendency to form tumors.1, 2 The variety of nonspecific clinical symptoms, including pyrexia, neurologic abnormalities, respiratory distress, and lack of lymphadenopathy, makes it difficult to diagnose lymphomas.3-5 However, recent recognition of the disease and the introduction of various diagnostic modalities, including random skin biopsy,6, 7 and 18F-fluoro-2-deoxy-D-glucose (FDG)-positron emission tomography (PET)/computed tomography (CT) (PET/CT)8, 9 have facilitated the in vivo diagnosis of IVL and prompt initiation of chemotherapy. The use of rituximab in combination with standard cyclophosphamide, doxorubicin, vincristine, and prednisolone (R-CHOP)10 has resulted in a marked improvement in the survival rate of patients with IVL.11
Neurolymphomatosis (NL) also is a rare manifestation of non-Hodgkin lymphoma in which lymphoma cells infiltrate the peripheral nerves.12, 13 NL is characterized clinically by a painful polyneuropathy or polyradiculopathy, cranial neuropathy, painless polyneuropathy, and peripheral mononeuropathy. Unfortunately, the diagnosis of NL often is difficult and requires histologic confirmation of the affected nerve, which may not be feasible and often is nondiagnostic.12, 14 Enhanced magnetic resonance imaging (MRI) is useful in diagnosing NL before histopathologic confirmation by revealing the enlargement and enhancement of the affected nerve, although it may not always provide optimal visualization of lymphomatous involvement.12, 15 Several recent studies have demonstrated the advantages of FDG-PET/CT in the diagnosis of NL.13, 16-20
The clinical characteristics of IVL are extremely variable; however, a high incidence of neurologic symptoms associated with central nervous system (CNS) involvement has been reported,3, 21 although peripheral nerve involvement in IVL rarely has been described. Here, we report 4 patients with IVL who had cranial and/or peripheral neuropathy as a manifestation of recurrent IVL. Imaging studies, including MRI and FDG-PET/CT, and neurologic findings supported the diagnosis of NL in these patients. The current observations suggest that IVL may have a predilection not only for the vessels but also for both the CNS and the peripheral nervous system.
MATERIALS AND METHODS
We reviewed the medical records of patients who were diagnosed with IVL over the past 4 years at Kameda General Hospital, Kamogawa-shi, Japan. There were 11 patients with pathologically confirmed IVL. Nine of those 11 patients were diagnosed with IVL by random skin biopsy, and the remaining 2 patients were diagnosed by bone marrow biopsy alone6, 7 with immunohistochemistry. Diagnoses of IVL were made by an expert hematopathologist in accordance with the latest World Health Organization classification.1, 2 Upon admission, all patients were examined by imaging modalities, including whole-body CT and/or brain MRI and FDG-PET/CT, as the initial evaluation of IVL in addition to routine laboratory examinations. The diagnosis of NL required 1) clinical symptoms and neurologic findings related to peripheral neuropathy of the cranial or spinal nerves; and either 2) CT/MRI-based demonstration of enhancement and/or enlargement of peripheral nerve(s) or nerve roots that also revealed accumulation of FDG by FDG-PET/CT12, 13 or 3) histologic confirmation of malignant lymphoma cells within the peripheral nerves, nerve root/plexuses, or cranial nerves.
A man aged 74 years was admitted to our hospital because of progressive dyspnea and unexplained fever. Laboratory data were unremarkable except for the presence of elevated serum lactate dehydrogenase (LDH) (1166 IU/L; normal level, 119-229 IU/L) and mild thrombocytopenia (107 × 103/mm3). Chest radiograms revealed ground-glass opacities in both lung fields. A histologic diagnosis of IVL was made on the basis of a random skin biopsy and a lung biopsy. The patient was treated with R-CHOP and had rapid improvement of clinical symptoms; however, after 3 courses of R-CHOP, the patient developed progressive paresthesia with pain in the upper and lower extremities, predominantly in the left leg. Neurologic examination indicated the absence of tendon reflexes in the extremities and glove-and-stocking type motor and sensory disturbances. An examination of cerebrospinal fluid (CSF) revealed slight pleocytosis (24/mm3) but was inconclusive for lymphoma cell invasion. The findings on brain and whole-spine MRI studies also were inconclusive. Although FDG-PET/CT revealed hypermetabolic lesions in the left iliopsoas and brachial plexus of both arms (Fig. 1a), NL was not diagnosed at that time; instead, a diagnosis of vincristine-induced peripheral neuropathy was considered. Subsequently, CSF became positive for cytology. The patient was treated with a high-dose of methotrexate (MTX) (1.5 g/m2) combined with weekly intrathecal MTX injections, which resulted in a brief improvement of neurologic symptoms. However, the patient developed renal failure and pneumonia, and his neurologic symptoms deteriorated during the treatment of these complications. The patient and his family refused further treatment. The patient died of aspiration pneumonia 4 months after the onset of neurologic symptoms. An autopsy revealed disseminated lymphomatous lesions involving the perineurium of the cauda equina, muscles, subcutaneous lesions, bone marrow, liver, adrenal gland, and testis, although no lymphomatous involvement was observed in the cerebral parenchyma. Unfortunately, the autopsy did not focus on the peripheral nerves, such as the brachial plexus or the lumbar plexus, and only the cauda equina was examined. Figure 1b reveals the cauda equina, where the perineurium of the individual nerve fascicles was infiltrated by atypical lymphoid cells that were positive for cluster of differentiation 79A (CD79a) (an immunoglobulin-associated alpha) (Fig. 1c) and negative for B-lymphocyte antigen CD20 (an activated-glycosylated phosphoprotein expressed on the surface of all mature B-cells). The loss of CD20 may have been caused by rituximab.
A woman aged 64 years who previously was in good health was referred to our hospital with a fever, consciousness disturbance, and elevated LDH (1210 IU/L). She had anemia (hemoglobin 7.9 g/dL), thrombocytopenia (34 × 109/mm3), and marked elevation of soluble interleukin-2 receptor (sIL2R) (24,900 U/mL; normal level, 230-620 U/mL). A diagnosis of IVL was made on the basis of a random skin biopsy. R-CHOP chemotherapy was initiated with prompt recovery of hematologic and neurologic findings. The patient achieved complete remission (CR) and was discharged with full recovery of neurologic abnormalities after 8 courses of R-CHOP. Soon after discharge, the patient developed weakness and burning pain in her lower legs. Neurologic examination revealed peripheral neuropathy with a glove-and-stocking distribution. Laboratory data, including results from a bone marrow biopsy, did not support the recurrence of lymphoma. CSF was negative for malignant cells. Although the results of both brain and whole-spine MRI and FDG-PET/CT were negative for the recurrence of a lymphoma, clinical signs and symptoms strongly suggested NL of the sciatic nerve in this case. The patient received 8 cycles of high-dose MTX (2.5 g/m2) every 2 weeks. This resulted in a gradual improvement of her neurologic symptoms. She then underwent autologous peripheral blood transplantation with high-dose cyclophosphamide (120 mg/kg) and total-body irradiation (12 grays). She remained in CR and was able to walk with the assistance of a frame for 3 years after the completion of treatment. A retrospective reassessment of the FDG-PET/CT studies at the onset of her neurologic symptoms revealed a linear uptake of FDG along both sciatic nerves corresponding to her clinical symptoms, which disappeared after treatment (Fig. 2a,b).
A woman aged 74 years was admitted to our hospital because of unexplained fever, hypoxia, splenomegaly, and elevation of serum LDH (2620 IU/L). A diagnosis of IVL was made based on a bone marrow biopsy. She achieved CR and was discharged after 6 cycles of R-CHOP followed by 1 course of high-dose MTX (3.5 g/m2) as prophylaxis to prevent CNS relapse. She was readmitted 1.5 months after discharge because of left facial hypoesthesia with pain, which was followed by right lower leg paralysis over the next few days. Neurologic examination revealed mononeuropathy multiplex of the left trigeminal nerve and the right deep peroneal nerve. The results of hematologic examinations, including complete blood count, bone marrow smear, and blood chemistry, all were within the respective normal limits. The findings of CSF examination were unremarkable. An axial, T1-weighted brain MRI revealed gadolinium enhancement of the thickened left trigeminal nerve (Fig. 3a). An FDG-PET/CT study revealed increased 18F-FDG uptake in the left Gasserian ganglion of the trigeminal nerve (Fig. 3b, center of the crossed red lines) and the right peroneal nerve (Fig. 3c,d), which corresponded to the clinical symptoms of NL. The patient responded well to salvage chemotherapy with ifosfamide, etoposide, cytarabine, and high-dose MTX22 followed by 2 courses of high-dose MTX (5.0 g/m2). After a transient improvement of clinical symptoms and imaging findings, she again developed right peroneal nerve palsy. Repeat FDG-PET/CT examination revealed intense uptake in the right peroneal nerve, but not in the trigeminal nerve. Subsequent CSF analysis was positive for cytology. Currently, she is being treated with irradiation (30 grays [Gy]) of the right peroneal nerve and intrathecal MTX infusion.
A woman aged 84 years was referred to our hospital with unexplained fever, confusion, progressive worsening of renal function, and deterioration of her general condition. On examination, she had pancytopenia with a hemoglobin level of 6.5 g/dL, a white blood cell count of 1700/mm3, and a platelet count of 28,000/mm3. Blood chemistry revealed marked elevation of LDH (1112 IU/L) and sIL2R (19,558 U/mL). A whole-body CT scan revealed pleural effusion, an adrenal mass, and mild splenomegaly. No peripheral or central lymphadenopathy was noted. A diagnosis of IVL was made by random skin biopsy. The patient received 7 cycles of a 50% to 80% dose of R-CHOP chemotherapy every 3 weeks and was discharged 5 months after admission in CR. However, she was readmitted because of weakness of the left lower leg 2 weeks after discharge. Neurologic examination revealed peripheral nerve impairment from the third lumbar segment to the first sacral segment. The results of a laboratory examination were unremarkable. An FDG-PET/CT study revealed intense uptake of FDG in the left lumbar plexus (Fig. 4b), which had not been observed at the time of discharge (Fig. 4c). Thin-slice coronal MRI images of the lumbar plexus revealed gadolinium enhancement of the enlarged sciatic nerve (Fig. 4a), which confirmed the presence of lymphomatous infiltration. Examination of CSF revealed a slight increase in protein and positive cytology. She was treated with high-dose MTX (2.0 g/m2) and weekly intrathecal MTX injection without benefit. Therefore, she received irradiation of the left lumbosacral lesion and had partial improvement. After irradiation, the accumulation of FDG in the left lumbar plexus disappeared completely (Fig. 4c).
The clinical characteristics of the 4 patients with NL are provided in Table 1. All NL developed as recurrent disease during or shortly after R-CHOP chemotherapy. LDH and sIL2R levels, both of which are sensitive tumor markers for IVL, did not change remarkably when NL occurred. CSF analyses revealed mild to moderate increases in both cell count (>5 cells/mm3) and protein concentrations in all 4 patients. Although only 1 patient (Patient 4) had positive cytology when NL occurred, subsequent CSF analysis detected leptomeningeal spread in an additional 2 patients (Patients 1 and 3).
|Patient No.||Age, y||Sex||Time of NL Development, mo||LDH, IU/L||IL2R, U/mL||Cells/ mm3||Protein, mg/dL||Clinical Signs|
|1||75||Man||3.0||318||482||24||142||Paresthesia; left arm weakness, speech and swallowing disturbance; oculomotor nerve palsy followed by quadriplegia|
|2||65||Woman||3.2||263||558||6||64||Lower leg paresthesia with pain and decreased vibration|
|3||74||Woman||6.0||211||259||6||54||Left trigeminal nerve palsy, right peroneal nerve palsy|
|4||84||Woman||6.0||696||220||24||58||Absence of knee and ankle jerks; flaccid paralysis of left leg; left sciatic nerve palsy|
The results of MRI and FDG-PET/CT studies in 4 patients with NL are provided in Figures 1 through 4 and are summarized in Table 2. Peripheral nerve palsies of the lower extremities were common in these patients. All patients exhibited peripheral nerve neuropathy corresponding to the affected nerve or nerve root, as determined by FDG-PET/CT studies. Routine whole-spine MRI in Patients 1 and 2 failed to detect NL lesions even after FDG-PET/CT. In contrast, we successfully discovered NL lesion by thin-sliced coronal imaging in Patients 3 and 4 after detecting suspected NL lesions by FDG-PET/CT. FDG-PET/CT studies indicated that NL lesions disappeared or reappeared in response to chemotherapy or radiation therapy. There were no significant differences between the patients with and without NL with regard to age, sex, clinical presentation, laboratory data (including anemia, thrombocytopenia, serum LDH, and sIL2R), the presence or absence of hemophagocytic histiocytosis, positivity of random skin biopsy, or CNS recurrence during the course of disease; although positivity for CD5 tended to be related to the development of NL (Table 3), this correlation was not statistically significant. However, it should be noted that a comparison of the 2 groups of patients who had IVL with or without NL was of limited value because of the small number of patients.
|Patient No.||MRI Findings||PET Findings|
|1||Negative for lymphoma invasion by whole spinal MRI||Hyperintense lesion in both brachial and left lumbar plexuses|
|2||Negative for lymphoma invasion by whole spine MRI||FDG uptake in left sciatic nerve|
|3||Enhancement and enlargement of Gasserian ganglion||FDG uptake in trigeminal and peroneal nerves|
|4||Enhancement and enlargement of left sciatic nerve||Intense FDG uptake at left sciatic nerve|
|Total No. of IVL Patients, n=11|
|Characteristic||Patients With NL, n=4||Patients Without NL, n=7||Pa|
|Median age (range), y||73 (63-83)||70 (48-79)||.57|
|Clinical features at presentation|
|Unexplained fever >38.0°C||4||7||NAb|
|Brain parenchymal lesion by MRI||1||1||1.00|
|Bone marrow infiltration||4||6||1.00|
|Laboratory data at presentation|
|Anemia <12.0 g/dL||3||5||1.00|
|Median serum LDH (range), IU/mL||1744 (1120-2054)||1216 (891-1728)||.57|
|Median serum sIL2-R (range), U/mL||12424 (3416-30,200)||7128 (2105-22,600)||.57|
|Positive random skin biopsy||3||6||1.00|
|Median survival after diagnosis of IVL, mo||49.5||13.9||.45c|
IVL represents an aggressive lymphoma with a rapid progression and a high incidence of systemic dissemination. Among the involved organs, it has been demonstrated that IVL has a predilection for CNS involvement at either presentation or relapse; however, neurologic manifestations of IVL are extremely heterogeneous and can include sensory and motor deficits as well as altered consciousness. Neuroimaging findings vary widely and range from the diffuse involvement of the deep white matter to infarct-like lesions,23 reflecting the heterogeneity of CNS involvement in terms of true lymphoma infiltration and vascular occlusion. In addition, many patients with IVL have a deteriorated performance status at diagnosis, and the diagnosis of CNS involvement often may be made by observing symptoms alone. Therefore, despite the frequency of CNS involvement in previous studies, there may be heterogeneity in CNS involvement within patients who have IVL. In addition, previous studies have not addressed peripheral nerve involvement in patients with IVL,3, 4, 21 although some case reports have described peripheral nerve involvement.24-27
NL is a rare clinical entity that has been described primarily in case reports and small case series. Given the extreme rarity of both IVL and NL, the current study highlights the high incidence of NL in patients with IVL. The diagnosis of NL often is delayed or is not made at all because of a lack of familiarity with the disease. In addition, NL mimics many other conditions; the diagnosis requires the exclusion of other causes of neuropathy and positive imaging findings to detect neural involvement. If the diagnosis remains in doubt, then histologic confirmation of the affected structure is necessary; otherwise, autopsy findings may be required to make the final diagnosis.13 The diagnoses of NL in our patients were made based on neurologic symptoms and identification of the corresponding lesion(s) on imaging studies, including MRI and FDG-PET/CT. Pathologic findings were obtained in Patient 1, who had nerve fascicles surrounded by tumor cell infiltrates. Intravascular lymphoma lesions were not observed in the cauda equina but were observed in other organs. There have been several reports indicating direct lymphoma cell infiltration within the nerve fascicles in patients with NL. Therefore, the pattern of lymphoma cell infiltration may vary between patients.12, 28 These differences may have implications regarding the clinical features and treatment responses in these patients.
The MRI radiographic characteristics of NL include nerve enlargement and enhancement, which were observed best on contrast-enhanced, thin-slice, T1-weighted coronal images of the spine.12, 29, 30 However, MRI may not be a reliable and sensitive diagnostic method, because NL lesions have a patchy distribution in long peripheral nerves. In our series, routine whole-spine MRI studies with 1.0-cm slice images failed to detect NL lesions in Patients 1 and 2, but 0.5-cm, thin-slice images detected nerve enlargement in Patients 3 and 4. In contrast, FDG-PET/CT studies detected NL lesions in all 4 patients and appeared to be more sensitive for the identification of NL, as reported previously.13 FDG-PET/CT studies also provided information on the distribution of the affected area(s) and treatment responses, as indicated by nodular lymphomas. Despite the usefulness of FDG-PET/CT studies in the diagnosis of NL, the number of reported cases diagnosed by FDG-PET/CT remains small, and a positive FDG-PET/CT study does not specifically suggest invasion of the lymphoma. The nonspecific accumulation of FDG in muscles, adjunct soft tissue, and vessels may mimic positivity. Biopsies of the affected nerve(s) may be necessary for the definite diagnosis of NL, but permanent defects in biopsied nerves and false-negative results often prevent this approach in the clinical setting. The concordance between neurologic findings and MRI and/or FDG-PET/CT findings could indicate a diagnosis of NL, as suggested by Baehring et al12 and Grisariu et al.13 In the current series, MRI was useful for confirming the FDG-PET/CT findings of cranial neuropathy and/or neuropathy because of spinal cord involvement, although a dedicated method for the disclosure of the affected nerve(s), such as using thin-slice coronal images and gadolinium enhancement, was necessary for the diagnosis of NL. In addition to the imaging studies and clinical symptoms, a high index of suspicion and familiarity with the clinical manifestations of NL are needed to ensure its early diagnosis.
Positive CSF cytology was observed in only 1 of the 4 patients when NL occurred, although almost all patients had variable increases in protein concentrations and exhibited slight CSF pleocytosis. These findings were consistent with previous studies indicating that many of patients with NL have negative CSF cytology at presentation.12 However, the use of multiparameter flow cytometry in addition to conventional cytology will improve the identification of leptomeningeal disease in patients with IVL, as reported recently in patients with aggressive non-Hodgkin lymphoma.31, 32
The most striking feature of recurrent NL is its timing. It is noteworthy that all 4 of our patients developed NL during or shortly they completed chemotherapy. It is possible that these patients had subclinical NL lesions at diagnosis although they had no clinical symptoms. Similar observations have been made in patients with CNS relapse of aggressive non-Hodgkin lymphoma.33, 34 Recently, Gan et al35 reported 4 patients who had diffuse large B-cell lymphomas that were complicated by the presence of NL. It is noteworthy that all NL occurred as recurrent disease during or shortly after those patients completed chemotherapy, as observed in our patients. These observations support the idea that chemotherapeutic agents, including rituximab, may not adequately penetrate the presumed blood-nerve barrier as well as the blood-brain barrier.36
The multiple potential sites of involvement mean that chemotherapy usually is treatment of choice in patients with NL. Intravenous high-dose MTX, either alone or in combination with other agents been used for the treatment of NL.13 Various doses of MTX ranging from 1.5 to 5.0 g/m2 with or without other systemic chemotherapy were used in our patients, but severe renal toxicity and pneumonia prevented further treatment in Patient 1. In addition, high-dose MTX with intrathecal MTX injection could not prevent or treat NL successfully, as observed in Patients 3 and 4. These observations suggest that these treatments may be suboptimal or inadequate for NL, as also reported in patients with aggressive B-cell lymphoma, in whom 5 of 9 patients who had flow cytometry-positive CSF developed overt CNS disease despite aggressive CNS-oriented treatment.34 Radiation therapy was successful in Patients 3 and 4, who had high-dose, MTX-resistant, peroneal, and lumbosacral involvement. These observations suggest that radiation therapy with curative or palliative intent may be of value in patients who have a high tumor burden or localized disease control.
In summary, we report 4 patients with IVL who presented with NL, which developed either as an initial manifestation or as recurrent disease. Their presenting symptoms consisted of progressive peripheral and/or cranial neuropathy. Although brain MRI and whole-spine MRI studies revealed nerve infiltration by demonstrating enlargement and enhancement of the nerve or nerve root, the sensitivity of this technology was limited by the patchy distribution or small lesion size of NL. FDG-PET/CT studies successfully revealed both cranial and peripheral nerve involvement in all patients and allowed us to evaluate the distribution of involvement and the therapeutic responses to chemotherapy. Considering the rarity of IVL and NL, the current observations suggest that IVL may have a predilection not only for the vessels but also for both the CNS and the peripheral nervous system. Further studies will be required to clarify the optimal dose and duration of treatment and to determine the long-term outcome of these patients.
We are grateful to the personnel of the Hematology/Oncology Ward and the Departments of Radiology and Clinical Pathology who assisted in this work. We are grateful to Dr. Kengo Takeuchi (Department of Pathology, Japanese Foundation for Cancer Research, Tokyo, Japan) for his contributions regarding pathology. We also thank Drs. Toshio Fukutake (Department of Neurology, Kameda General Hospital), Kazuo Oshimi (Eizai Pharmaceutical Co. Ltd., Boston, Mass), and Keitaro Matsuo (Department of Epidemiology, Aich Cancer Center, Nagoya, Japan) for their suggestions and careful reading of our article.
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.
- 2Intravascular Large B-Cell Lymphoma. Lyon, France: JARA Press; 2008., , .
- 3Intravascular lymphoma: clinical presentation, natural history, management and prognostic factors in a series of 38 cases, with special emphasis on the “cutaneous variant.” Br J Haematol. 2004; 127: 173-183., , , et al.