Authors: Dylan Asmani (Service de Neurologie, Centre de Référence Neuropathies Périphériques Rares, Centre Hospitalier Universitaire de Limoges, Limoges, France), Jean‐Baptiste Chanson (Centre de Référence Des Maladies Neuromusculaires Nord/Est/Ile‐de‐France, Service de Neurologie, Hôpitaux Universitaires de Strasbourg, Strasbourg, France), Céline Tard (Centre de Reference des Maladies Rares Neuromusculaires Nord Est Ile‐de‐France Centre Constitutif de Lille, Lille, France), Juliette Svahn (Service ENMG et de Pathologies Neuromusculaires, Centre de référence Des Maladies Neuromusculaires PACA‐Réunion‐Rhône Alpes, Hôpital Neurologique P. Wertheimer, Hospices Civils de Lyon, Bron, France), Antoine Pegat (Service ENMG et de Pathologies Neuromusculaires, Centre de référence Des Maladies Neuromusculaires PACA‐Réunion‐Rhône Alpes, Hôpital Neurologique P. Wertheimer, Hospices Civils de Lyon, Bron, France), Armelle Magot (Laboratoire d'Explorations Fonctionnelles, Centre de Référence Maladies Neuromusculaires Rares AOC, Nantes, France), Yann Péréon (Laboratoire d'Explorations Fonctionnelles, Centre de Référence Maladies Neuromusculaires Rares AOC, Nantes, France), Jean‐Philippe Camdessanché (Neurology Department, University Hospital of Saint‐Etienne, Saint‐Etienne, France), Samuel Naudin (Service de Neurologie, Centre Hospitalier de Brive‐la‐Gaillarde, Brive‐la‐Gaillarde, France), Olivier Colin (Service de Neurologie, Centre Hospitalier de Brive‐la‐Gaillarde, Brive‐la‐Gaillarde, France), Guillaume Taieb (Department of Neurology, CHU Montpellier, Hôpital Gui de Chauliac, Montpellier, France), Idrissa Coulibaly (Department of Internal Medicine, Hospital Center of Cholet, Cholet, France), Laurent Magy (Service de Neurologie, Centre de Référence Neuropathies Périphériques Rares, Centre Hospitalier Universitaire de Limoges, Limoges, France), Simon Frachet (Service de Neurologie, Centre de Référence Neuropathies Périphériques Rares, Centre Hospitalier Universitaire de Limoges, Limoges, France)
Categories: Clinical Research Article, CIDP, conduction block, demyelinating parameters, neurolymphomatosis, neuropathy
Source: Muscle & Nerve
Doi: 10.1002/mus.28343
Authors: Dylan Asmani, Jean‐Baptiste Chanson, Céline Tard, Juliette Svahn, Antoine Pegat, Armelle Magot, Yann Péréon, Jean‐Philippe Camdessanché, Samuel Naudin, Olivier Colin, Guillaume Taieb, Idrissa Coulibaly, Laurent Magy, Simon Frachet
Neurolymphomatosis is a hematological condition defined by the direct infiltration of malignant lymphomatous cells into the peripheral nervous system. Since nerve conduction studies may disclose demyelinating features, clinicians may misdiagnose neurolymphomatosis as chronic inflammatory demyelinating polyneuropathy (CIDP). This study aimed to determine whether patients with neurolymphomatosis met the 2021 revised criteria for CIDP.
We retrospectively analyzed 23 patients with primary or secondary neurolymphomatosis from nine French hospitals. We analyzed whether patients with a diagnosis of neurolymphomatosis met the 2010 and 2021 CIDP criteria.
152 motor nerves were analyzed, and conduction blocks were found in 13.8% of them. Eight patients (34.7%) demonstrated at least one conduction block. Other demyelinating parameters fulfilling the 2021 CIDP criteria were rare, including reduced conduction velocities (1.7%), prolonged distal motor latencies (3%), absent F‐waves (12.4%), and prolonged F‐wave latencies (7.7%). Five patients met the 2010 CIDP criteria, whereas only one met the 2021 CIDP criteria.
Demyelinating features are rare in neurolymphomatosis, and a conduction block is the most frequent abnormality. Consequently, only one patient met the 2021 criteria for CIDP diagnosis. This is likely attributable to the inclusion of sensory criteria. Furthermore, the new criteria emphasize the importance of identifying red flags, such as pain or monoclonal gammopathy, which could suggest an alternative diagnosis to CIDP. Clinicians should consider neurolymphomatosis in patients who present with red flags or atypical CIDP.
Neurolymphomatosis is a rare presentation of hematological malignancies caused by the infiltration of malignant lymphomatous cells into the peripheral nervous system, including the cranial nerves. It is most commonly observed in non‐Hodgkin's lymphoma (NHL), particularly in diffuse large B‐cell lymphoma [1]; however, researchers have also reported cases of neurolymphomatosis secondary to T‐cell lymphoma, NK‐lymphoma, chronic lymphocytic leukemia, and Waldenström's macroglobulinemia [2, 3, 4, 5]. Neurolymphomatosis may manifest as the initial presentation of an as‐yet‐undetected lymphoma (primary neurolymphomatosis) or may develop during the progression of a pre‐existing lymphoma (secondary neurolymphomatosis). Its clinical presentation is highly heterogeneous, with four different phenotypes traditionally painful involvement of nerves or roots, painless involvement of peripheral nerves, cranial neuropathy, and involvement of a single peripheral nerve [6]. In the case of polyneuropathies, involvement is typically sensory‐motor, although instances of pure motor or sensory involvement have also been documented [7]. Neurolymphomatosis may also present with clinical features that are similar to those observed in Guillain‐Barré syndrome [8, 9]. Therefore, the diagnosis of this condition can be complex. It can be based on histological results, with a nerve sample showing tumor cell infiltration in the perineurium and endoneurium. In the absence of histological data, diagnosis requires the integration of clinical data, imaging findings, cerebrospinal fluid (CSF) analysis, and information on clinical evolution and response to treatment. Magnetic resonance imaging (MRI) and fludeoxyglucose‐18 positron emission tomography (FDG‐PET) are the most sensitive, non‐invasive tools, with MRI showing nerve, root, or plexus enlargement with enhancement after gadolinium injection and FDG‐PET displaying a linear or fusiform mass indicating nerves infiltrated by lymphoma. CSF analysis has a rather limited diagnostic value, with malignant cells being identified in only 20% to 40% of patients [6, 10].
Researchers have described three main electrophysiological profiles in patients with primary multiple mononeuropathies, and symmetrical and asymmetrical sensory‐motor polyneuropathies [11]. A partial conduction block may be an early marker of neurolymphomatosis [12]. Electrodiagnostic evidence of demyelinating patterns has also been described, explaining why chronic inflammatory demyelinating polyneuropathy (CIDP) is one of the main alternative diagnoses for neurolymphomatosis [13]. This study also revealed that one‐third of the patients diagnosed with neurolymphomatosis fulfilled the 2010 European Federation of Neurological Societies/Peripheral Nerve Society (EFNS/PNS) CIDP criteria, but only 15 patients were studied. The underlying pathophysiological mechanism is believed to be secondary to focal infiltration of the nerve between the proximal and distal stimulation sites, which results in segmental demyelination, with axonal degeneration occurring secondarily [13]. This segmental demyelinating pattern has been proven for sarcoidosis, another disease involving cell infiltration into nerves [14]. The present study aimed to describe the clinical and electrophysiological features of patients with neurolymphomatosis and to determine whether they met the revised criteria from 2021 for CIDP.
This retrospective multicenter study was conducted in nine French hospitals (located in Limoges, Lyon, Strasbourg, Nantes, Lille, Montpellier, Saint‐Etienne, Brive‐La‐Gaillarde, and Cholet) and involved patients diagnosed with neurolymphomatosis between October 2004 and September 2023.
All patients included in this study had been diagnosed with neurolymphomatosis by a neurologist and/or a hematologist, either histologically by nerve biopsy or based on a combination of clinical presentation, imaging results, cerebrospinal fluid and blood analyses, response to treatment, and exclusion of mimics. Three levels of diagnostic certainty were Definite Neurolymphomatosis (when a neuropathological examination confirmed the diagnosis), Probable Neurolymphomatosis (defined by a typical clinical picture and abnormal CSF cytology/PET‐FDG/MRI evidence of nerve infiltration by abnormal tissue), and Possible Neurolymphomatosis (defined by the coexistence of a pathologically proven lymphoma and a clinical picture suggestive of neurolymphomatosis without a better explanation after extensive testing).
The study included patients with primary or secondary neurolymphomatosis. Patients for whom electrodiagnostic data was unavailable were excluded.
Clinical data were retrospectively collected from the medical records of patients at the participating centers. The collected clinical information included sex, age at symptom onset, pain level, and neurological signs and symptoms (muscle weakness, sensory deficits), as well as their distribution. We also considered other causes of neuropathy. Onset was arbitrarily defined as acute if symptoms developed over 4 weeks or less, subacute if they developed over 4–12 weeks, and chronic if they developed over a period longer than 12 weeks. We used the Overall Neuropathy Limitation Scale (ONLS) for evaluation, which ranges from 0 (no impairment) to 12. Prognostic parameters according to the Eastern Cooperative Oncology Group Performance Status (ECOG‐PS) ranging from 0 (fully active) to 4 (completely disabled) were also recorded [15]. Additionally, we examined the results of blood work and cerebrospinal fluid analyses, nerve biopsies, and imaging findings (FDG‐PET/CT and brain, spinal cord, and plexus MRI).
The nerve conduction studies (NCS) were performed at each center in accordance with local standard operating procedures. All included patients underwent NCS at least once. Distal motor latencies, motor nerve conduction velocities (MNCVs), and the amplitudes of compound muscle action potentials (CMAPs) were evaluated for the median, ulnar, fibular, and tibial nerves. MNCVs were calculated using the following nerve wrist‐to‐elbow for the median nerve; wrist‐to‐below the elbow for the ulnar nerve; ankle‐to‐below the neck of the fibula for the fibular nerve; and ankle‐to‐popliteal fossa for the tibial nerve. The presence and latency of the F‐waves were assessed. For the sensory nerves, conduction velocity and action potentials were recorded from the median, ulnar, radial, sural, and superficial fibular nerves. Sensory nerve action potential (SNAP) amplitudes were assessed peak to peak for the median and ulnar nerves and recorded orthodromically or antidromically based on local practice, whereas radial, sural, and superficial peroneal SNAP amplitudes were recorded antidromically.
A conduction block was defined as a decrease of > 30% in the negative‐peak amplitude of the CMAP between distal and proximal stimulations, except in the case of the tibial nerve, where a 50% decrease was required. Proximal amplitudes of the ulnar and fibular nerves were measured below the elbow and fibular neck, respectively, to avoid blocks caused by entrapment. Temporal dispersion was defined as an increase in the CMAP duration of > 30% between distal and proximal stimulations. Needle electromyographic analysis was performed on several muscles in each patient to determine spontaneous resting and voluntary activity.
The criteria used were those outlined in the 2010 and 2021 revisions by the EFNS and European Academy of Neurology, respectively [16, 17]. In both cases, clinical, electrodiagnostic, and supportive criteria were subjected to rigorous analysis.
The study was conducted in accordance with the principles of the Declaration of Helsinki. Individual written consent was not required because of the retrospective design of the study, and the study was approved by the Ethics Committee of the University Hospital of Limoges (registration number 05–2024‐02).
All statistical analyses were performed using Rstudio (version 1.4.1717). We described the characteristics of the study population using descriptive statistics. Quantitative values are expressed as medians (IQR). We used the Fisher exact test to compare the key qualitative variables between patients with primary and secondary neurolymphomatosis. Statistical analyses were two‐tailed, with an adjusted significance level of 0.005 according to the Bonferroni correction for multiple comparisons.
Twenty‐three patients with pathologically confirmed lymphoma were enrolled. Thirteen of the 23 patients (56%) were classified as primary neurolymphomatosis, and the remaining 10 patients as secondary neurolymphomatosis. Tables 1 and 2 show the main clinical and electrophysiological features of the patients. Almost all the patients had sensory‐motor deficits with areflexia. Among the hematological malignancies, the highest proportion corresponded to NHL, with diffuse large B‐cell lymphoma being the most common form. Further details regarding the hematological disease are shown in Table S1.
Seven patients had Definite Neurolymphomatosis proven by nerve biopsy, which was conducted in 14 patients. Only one biopsy was performed under positron emission tomography guidance, which yielded a positive result. Based on the results of FDG‐PET/CT and MRI examinations, 11 patients were diagnosed with Probable Neurolymphomatosis and five with Possible Neurolymphomatosis. The results of the various laboratory, histopathological, and imaging examinations that led to the diagnosis of neurolymphomatosis are presented in Table S1. Fifteen patients underwent 18‐FDG‐PET/CT. Of these, five showed FDG avidity on the nerve tract or nerve roots (33%) and seven showed lymphoma dissemination. The findings were unremarkable in the remaining three patients. Abnormalities were detected in 14 of the 17 patients (82%) who underwent plexus or spinal cord MRI. Of these, 12 showed abnormal enlargement of the roots/nerves and contrast enhancement, whereas the other two showed associated medullary lymphoma and lymphoma‐related bone involvement. Serum protein electrophoresis data were obtained from 15 patients, of whom six (40%) had a monoclonal gammopathy. Of those with primary neurolymphomatosis, four patients underwent bone marrow biopsy, two of which were positive for lymphoma.
Based on the results of NCS, we identified six distinct electrodiagnostic patterns, of which the most common ones indicated multiple mononeuropathies (26.1%) and polyradiculopathy (21.7%). Four patients exhibited abnormalities in multiple nerve roots and in the cervical or lumbosacral plexus, consistent with radiculoplexus neuropathy (17.4%). Four patients (17.4%) had axonal length‐dependent polyneuropathy, three (13%) had polyradiculoneuropathy, and one (4.3%) had isolated mononeuropathy of the left fibular nerve.
A total of five patients met the 2010 criteria for CIDP, while only one of them met the 2021 criteria. The most common reasons for not meeting the criteria were the presence of a red flag, such as pain or monoclonal gammopathy, and not fulfilling the sensory criteria added in the 2021 revision.
We studied a total of 152 motor nerves. We found that 20 distal CMAPs were absent. The results of the motor nerve studies are shown in Table 3. We assessed each electrical parameter to determine whether it met the 2021 CIDP criteria. Out of the 132 distal motor latencies analyzed, only 4 (3%) were prolonged. Two out of 116 nerves tested (1.7%) showed motor conduction velocities that were sufficiently slowed to meet criteria. A total of 16 (13.8%) conduction blocks were observed in eight of our patients (34.7%). Additionally, we observed six (5.2%) instances of temporal dispersion. Eleven out of the 89 tested F‐waves were absent (12.4%), and six (7.7%) out of 78 had prolonged latencies (≥ 20% above the upper normal limit).
Out of the 124 sensory nerves tested, responses were absent in 27 (21.7%). In the remaining 97 sensory nerves, 23 (23.7%) showed a decreased conduction velocity according to local normative values. Only three (3.1%) showed sufficient slowing to be suggestive of a demyelinating process according to the 2021 CIDP criteria.
The key clinical and neurophysiological characteristics of primary and secondary neurolymphomatosis were compared (Table 4). We found no significant differences in the presence or asymmetry of symptoms. No differences in the type of electrodiagnostic pattern, presence of conduction blocks or temporal dispersion, or fulfillment of CIDP criteria were detected in the NCSs.
Among the 23 patients diagnosed with neurolymphomatosis included in this retrospective study, only one patient met the 2021 CIDP criteria, whereas 21.7% (5/23) met the 2010 criteria. The reason for this change was the inclusion of sensory criteria and the need to systematically identify red flags before diagnosing CIDP according to the 2021 criteria. Thus, the risk of overlooking underlying neurolymphomatosis is reduced by strictly applying the most recent CIDP criteria, which appear to be more specific. These results are consistent with recent observations in the literature. Although a previous study reported that 33% of patients with neurolymphomatosis met the 2010 CIDP criteria [13], a more recent study involving nine patients with primary neurolymphomatosis found that none of them fulfilled the 2010 criteria [11]. An incorrect initial diagnosis of CIDP can result in inappropriate and ineffective treatment. The diagnostic delay for primary neurolymphomatosis is frequently longer than that for secondary forms, likely because primary forms are less aggressive, and the absence of a known lymphoma reduces the likelihood of considering this diagnosis [18]. None of our patients who were initially suspected of having CIDP showed improvement with intravenous immunoglobulin, resulting in a significant delay in diagnosis. In our cohort, neurolymphomatosis was seldom confused with typical CIDP. Rather, it was predominantly mistaken for multifocal CIDP, as shown in Table S1. It is therefore imperative that, in the event of a suspected case of multifocal CIDP, the presence of any potential red flags, such as pain, a family history of neuropathy, a monoclonal gammopathy, or a positive antineutrophil cytoplasmic antibody (ANCA) test, should be checked in order to rule out any alternative diagnoses. Lack of response to intravenous immunoglobulin, rapid worsening of symptoms, subacute onset, cranial nerve involvement, and general signs should also always be monitored in any suspected variant of CIDP. Furthermore, it is important not to give disproportionate weight to neurophysiological abnormalities suggestive of a demyelinating process if the context does not favor a diagnosis of CIDP.
Seven (30%) had a histological diagnosis based on nerve biopsy. The diagnostic yield of the nerve biopsy in our cohort was 50%, consistent with what was reported in a previous study [1]. This emphasizes the need for imaging guidance before biopsy. This study confirms that the presentation of neurolymphomatosis is heterogeneous, making the diagnosis challenging. Some patients presented with radiculoplexus neuropathy, a rare peripheral neuropathy, and its presence should always prompt consideration of a malignancy as an underlying cause [19].
The present work suggests that patients with neurolymphomatosis exhibit minimal demyelinating abnormalities on NCS. The only frequently encountered abnormality was the presence of conduction block, which was found in 34.7% of the patients and in 13.8% of the motor nerves examined. This proportion is lower than that reported in a previous study on a smaller sample, in which 58% of the patients were found to have at least one partial conduction block [12]. However, these blocks showed no significant temporal dispersion, as previously reported.
Few studies have compared the characteristics of primary and secondary neurolymphomatosis. In 2022, Tan et al. [20] found no differences in sex, age at symptom onset, lymphoma type, clinically involved nerve structures, or PET abnormalities, but no comparisons were made based on NCS results. Our study found no significant differences in NCS characteristics between primary and secondary neurolymphomatosis.
This study was retrospective and included a rather small number of patients. The retrospective analysis of NCS may have been susceptible to bias, considering the lack of a standardized protocol among different centers. Additionally, this retrospective study did not thoroughly analyze the presence of proximal conduction block with stimulation at Erb's point. We recruited a substantial cohort of patients primarily through registries of nerve biopsies, resulting in a higher prevalence of mononeuropathy multiplex compared to polyradiculoneuropathy. This discrepancy is likely because mononeuropathy multiplex is one of the key indications for nerve biopsy. Altogether, these issues may have contributed to an underestimation of the frequency of conduction block. Furthermore, a weakness of our study is that the diagnosis of neurolymphomatosis was not confirmed by histopathological examination in all of our patients. This is due to the invasive nature and limited availability of biopsies, which were not routinely performed. Moreover, some biopsies may have targeted an unaffected nerve trunk due to the inherent difficulty in accessing nerve roots. Additionally, the sensitivity of nerve biopsies remains imperfect. Consequently, the diagnosis of neurolymphomatosis was occasionally challenging to confirm antemortem, despite comprehensive evaluation. However, we believe that the rigorous criteria used for diagnosing neurolymphomatosis (see Methods) allowed for a retrospective diagnosis of neurolymphomatosis with an acceptable level of confidence.
In conclusion, demyelinating abnormalities are exceedingly rare in patients diagnosed with neurolymphomatosis, except for conduction block without significant temporal dispersion. Only one of the patients included in the study met the revised CIDP criteria from 2021, which appears to be more specific than those of 2010. One of the main improvements of the more recent CIDP criteria is the inclusion of red flags to account for the main alternative diagnoses for this disorder. The present results suggest that clinicians should consider alternative diagnoses before settling on a diagnosis of CIDP and should rigorously search for red flags such as severe pain or monoclonal gammopathy even before performing NCS and applying the electrodiagnostic criteria.
Dylan Asmani: conceptualization, investigation, methodology, validation, visualization, writing – review and editing, data curation, formal analysis, writing – original draft. Jean‐Baptiste Chanson: investigation, writing – review and editing. Céline Tard: investigation, writing – review and editing. Juliette Svahn: investigation, writing – review and editing. Antoine Pegat: investigation, writing – review and editing. Armelle Magot: investigation, writing – review and editing. Yann Péréon: investigation, writing – review and editing. Jean‐Philippe Camdessanché: investigation, writing – review and editing. Samuel Naudin: investigation, writing – review and editing. Olivier Colin: investigation, writing – review and editing. Guillaume Taieb: investigation, writing – review and editing. Idrissa Coulibaly: investigation, writing – review and editing. Laurent Magy: investigation, conceptualization, writing – original draft, methodology, validation, visualization, writing – review and editing, supervision. Simon Frachet: conceptualization, investigation, writing – original draft, methodology, validation, visualization, writing – review and editing, software, formal analysis, data curation, supervision.
We confirm that we have read the journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
The authors declare no conflicts of interest.