Introduction

Neurolymphomatosis (NL) is a disease process in which lymphoma cells invade the peripheral nervous system. It is estimated to occur in up to 3% of patients with lymphoma [1]. The mechanism of perineural spread (PNS) in NL is still being elucidated, but is believed to result from intraneural invasion of neurolymphoma and proximal and/or distal spread of disease. Patients can present with one limb being primarily affected, but, on diagnostic workup, are sometimes seen to have disease of the contralatera limb as well. We sought to explain the findings of bilateral upper or bilateral lower limb disease and understand if there was a connection to the initial, symptomatic side of disease. We hypothesized there may be evidence of previously unrecognized circumdural extension of PNS similar to that reported in metastatic solid organ disease as an etiology of bilateral disease [3, 4]. This has important prognostic and pathomechanistic implications that can improve our understanding of this rare disease. This finding could suggest the existence of unappreciated contralateral disease in patients initially thought to have only one affected limb.

Materials and methods

After institutional review board approval, patients with bilateral disease were identified from a previously published cohort of patients with tumefactive NL with the addition of more recent cases seen at our institution [2]. All cases were biopsy proven in the primarily affected limb with corresponding evidence of imaging abnormalities. We excluded one case of multi-limb involvement as these causes likely had systemic disease that could have diffuse hematogenous or lymphatic spread. Magnetic resonance imaging (MRI) and positron emission tomography/computed tomography (PET/CT) were reviewed by a musculoskeletal trained radiologist (KKA) with expertise in PNS. Circumdural extension was defined as post-gadolinium dural enhancement at the corresponding level of the primary affected nerve. We reviewed demographics (sex and age), clinical data (primary or secondary disease and biopsy-proven diagnosis), and imaging findings (primary involved nerve, contralateral nerve(s) affected, and location of circumdural extension).

Results

Eight out of 22 patients (2 women and 6 men) were identified to have bilateral upper or bilateral lower limb disease and all eight had previously unrecognized evidence of circumdural extension on MRI (Table 1). Average age at diagnosis was 61 years (range, 37 to 76 years) and average time from symptom onset to diagnosis was 12 months (range, 2 to 60 months). Five cases occurred in patients with a history of lymphoma as secondary disease, while three cases occurred as primary disease leading to initial diagnosis. Only four cases had evidence of bilateral limb involvement at the same dermatome on imaging, while the remaining had only proximal, contralateral disease. All patients had at least one primarily symptomatic limb leading to workup and imaging. Five patients had bilateral symptoms, but still had a primarily affected limb. Of these patients, two had intermittent contralateral symptoms. All cases had post-gadolinium enhancement of the dominantly affected nerve, circumdural enhancement, and contralateral proximal nerve enhancement. PET/CT showed corresponding avidity in cases of diffuse large B cell lymphoma (DLBCL) and was negative in the case of low-grade B cell lymphoma (LGBCL). PET/CT was negative on the contralateral side in all cases, except for one case in which PET/CT was not performed. One patient had intermittent bilateral symptoms over the course of years, but imaging of the contralateral brachial plexus remained negative despite significant, temporary pain and weakness (Fig. 1).

Table 1 Summary of clinical and imaging data of the series
Full size table
Fig. 1
figure 1

Axial spoiled gradient fat-suppressed post-contrast MR images depicting circumdural extension (arrowheads). a Patient with DLBCL with primary involvement of right S1 nerve with circumdural extension at L5-S1 spinal level. b Patient with DLBCL with primary involvement of left C8 nerve with circumdural extension at C7-T1 spinal level. c Patient with DLBCL with primary involvement of right C7 nerve with circumdural extension at C6-C7 spinal level

Full size image

In our cases, patients presented with a dominant affected limb and in cases with bilateral upper or bilateral lower limb disease, most often had mild contralateral symptoms. Typical nerve symptoms were present, including pain, sensory changes, and weakness. PET/CT corresponded with the primary symptomatic limb, without avidity in the contralateral limb even with mild symptoms, or evidence of involvement on MRI. Cerebrospinal fluid analysis and flow cytometry were performed in four cases and were negative.

Discussion

We believe our series of cases illustrate an unrecognized pathomechanism of bilateral involvement in NL. Circumdural extension of PNS has been described in other cancers, including breast and prostate, but not previously in NL [3, 4]. In our cases, after intraneural invasion of the primary limb, PNS occurs proximally toward the midline. It then spreads circumdurally to the contralateral side. Within the primarily affected limb, we see that multiple segments of the plexus can be affected, but that there is a more dominant nerve proximally that appears most involved. This dominant nerve corresponds to the primary crossing level around the dura, which also corresponds to the primarily affected contralateral portion of the plexus (Fig. 2).

Fig. 2
figure 2

Illustration of circumdural extension of NL with dotted lines and arrows indicating proposed path of PNS. In this illustration, tumor cells spread from anatomic right to left with cells in the dominant nerve spreading centrally toward the dura via PNS. Cells then enter the spinal canal via the dura and move around the dura (circumdural) to cross to the contralateral side. Cells continue to spread via perineural spread, but now on the contralateral side of initial disease. Used with permission of Mayo Foundation for Medical Education and Research. All rights reserved

Full size image

We believe in these cases of an initially single dominant limb with NL, our proposed pathomechanism of bilateral involvement applies. Our cases all show a primary limb and significant ipsilateral plexus involvement with circumdural enhancement confined to the corresponding spinal level and otherwise negative spinal imaging. The sensitivity of CSF in detecting hematologic malignancies varies but has been reported upwards of 90% [5]. Given that CSF and spinal imaging were otherwise negative, we do not believe CNS dissemination of the disease to be the mechanism of contralateral spread. Rather, our findings strongly support a physical, anatomic spread of tumor cells around or within the dural layer and propagation to the contralateral nerve.

We believe this finding of circumdural extension is under-recognized, and at our institution, this was noticed only on re-review of imaging given the initially low clinical suspicion of bilateral disease. Only one limb was typically imaged (e.g., right or left brachial plexus) so examination of the spinal canal was focused on ruling out diffuse leptomeningeal disease or obvious spinal lesions. As a result, further characterization of distal bilateral limb involvement may not have always been appreciated. We are unaware of a previous report in the literature for lymphoma with this mechanism of PNS. Because of limitations in the available published imaging of the spinal canal and dura in patients with NL reported by other groups, it was difficult for us to confirm other cases that could have circumdural involvement.

Conclusion

PNS of malignancy is becoming increasingly recognized and its mechanism, while still largely a mystery, is beginning to be investigated. We have described PNS and the tumefactive appearance in neurolymphoma [2], but have been previously unaware of the significance of bilateral disease and this pathomechanism. We have described the pathomechanism of spread in bilateral disease, where NL spreads along a dominant nerve toward the spinal canal and moves circumdurally to affect the corresponding contralateral nerve. We believe this mechanism follows an anatomic, logical progression of disease and follows the clinical presentation of our patients. We believe this information is useful to further understand the spread of NL, as well as offering important diagnostic and prognostic information for patients who may have more widespread disease than initially appreciated.