Neurolymphomatosis of the lumbosacral plexus and its branches: case series and literature review

The term neurolymphomatosis (NL) specifically describes a process of direct endoneurial invasion by lymphoma cells [1‐3]. All segments of the peripheral nervous system can be targeted by NL, including cranial nerves, spinal roots, brachial or lumbosacral plexus, and individual peripheral nerve branches [4]. Such focal invasion occurs in roughly 5% of patients with lymphoma, with a strong preponderance of the non-Hodgkin’s B-cell subtype known as diffuse large B-cell lymphoma (DLBCL). NL is a particularly difficult diagnostic dilemma when it is the presenting focal manifestation of occult lymphoma [5].

We present here 3 patients representative of the spectrum of NL of the lumbosacral plexus and its branches. We have surveyed the relevant literature and discuss the role and sensitivity of neuroimaging techniques.

In advanced biopsy-proven recurrent or treatment-resistant diffuse large B-cell lymphoma, the occurrence of multifocal radiculopathy, plexopathy or mononeuropathies will immediately suggest a diagnosis of NL [6]. In contrast, the diagnosis may be elusive in the setting of isolated focal progressive peripheral neuropathy if the patient has no history of malignancy or has achieved prolonged lymphoma remission after therapy. Our series is representative of this clinical spectrum of NL presentation. In case 3, lumbosacral NL was the first symptomatic focus of lymphoma, while in cases 1 and 2, lumbosacral NL was a secondary finding in disseminated or upper limb predominant disease. We also present the use of multiple imaging modalities for diagnosis and work-up of the patient during different stages and severity of the disease. On the other hand, we do not present an extremely granular timeline of all events for each patient given the redundancy with the provided clinical information and also the complexity of the cases.

In the absence of documented systemic malignancy, the differential diagnosis of progressive focal neuropathy will include a long list of benign or malignant peripheral nerve sheath tumors [7], perineurioma [8], entrapment syndromes, multifocal syndromes of inflammatory demyelinating neuropathy [9], nerve vasculitis and sarcoidosis [10]. Lymphomatous infiltration should however always be also considered.

In Table 1, we have summarized 23 case reports where isolated lumbosacral plexus NL occurred de novo or as the index manifestation of recurrent lymphoma. As expected, B-cell histopathology was reported in 21 cases (92.3%), and T-cell lymphoma was noted in only 2 cases [19, 28]. The median age at presentation was 62 years, compared to 70 years in a large UK epidemiological survey for diffuse large B-cell lymphoma [32]. Primary sciatic localization was by far most frequent (17, or 73.9%), followed by posterior tibial nerve (8.6%) [11, 12], common or deep fibular nerve (8.6%) [13, 17] and the pelvic segment of the femoral nerve (4.3%) [15]. Isolated lumbosacral plexus branch NL infiltration was the first clinical manifestation of lymphoma in 17 of 23 cases, and was reported in the setting of recurrent lymphoma in six cases. Sciatic involvement presented as a discrete or lobulated nerve infiltration at any level of the thigh. In three cases, infiltration of the sciatic nerve was diffuse at time of diagnosis, extending from the ischial region to the popliteal fossa, and then further caudally into the tibial nerve [26, 27, 29]. In contrast, the most discrete examples of lower limb NL were cases of isolated short-segment infiltration of the tibial nerve at the ankle [11] and deep fibular nerve in the mid lower leg [13]. Our table does not list treatment outcomes, as most reports only documented the initial favorable response to radiotherapy or chemotherapy without providing sufficient longitudinal data. Death from disseminated disease within a few months to 3 years was however specifically mentioned in 5 cases in this series and occurred in two of our cases.

In the imaging assessment of nerve tumors and the detection of NL, ultrasound is still valuable because of its easy access, safety, lack of contraindications and favorable patient tolerability [14]. It can reliably distinguish cystic from solid masses, the anatomical localization including the relationship to blood vessels and help guide safe biopsy. Additionally, color Doppler ultrasound has been proposed to demonstrate increased blood flow in NL, but less in entrapment or inflammatory neuropathy [16, 18]. Enhanced CT imaging has been largely superseded by MRI, when the latter is not contraindicated. CT may however more easily detect tumor calcifications and may also help characterize tumors in relation to neighboring bony structures, particularly at the level of the spine and thorax [20]. CT has also been the modality most commonly combined with PET to improve spatial resolution.

MRI neurography is emerging as a powerful tool to help detect and characterize nerve pathology [21]. Normal nerves show gradual tapering, remaining typically smaller than the accompanying artery. They are usually outlined by fat (best appreciated on T1-weighted images) and are isointense to skeletal muscle on both T1 and T2-weighted images. They have an even fascicular distribution, with no appreciable gadolinium enhancement other than at the level of dorsal root ganglia. The more prevalent forms of diffuse metabolic or toxic axonal polyneuropathy do not show significant deviation from this normal pattern. Many focal inflammatory, infiltrative, hypertrophic and neoplastic neuropathies however show a fairly consistent pattern of MRI abnormality [22, 23]. Such disorders often feature focal or diffuse nerve enlargement, often exceeding the accompanying artery. In these disorders, abnormal nerves remain isointense to muscle on T1-weighted images, but are often significantly hyperintense on T2-weighted or STIR images. They may also show marked fascicular disorganization, deviation from their normal course and significant enhancement. MRI changes of denervation of neighboring muscles may provide an additional clue to the presence of neuropathy. These criteria however offer relatively poor specificity to help discriminate between NL and other etiologies such as acute or chronic inflammatory demyelination, inherited hypertrophic neuropathy, radiation changes, perineuroma, focal hypertrophic neuritis, amyloidosis or neurofibromatosis. As a general rule, gadolinium enhancement is absent or minimal in amyloid or genetic neuropathy. It tends to be modest in acute inflammation or radiation, but it may be quite prominent in chronic inflammatory demyelinating polyneuropathy (CIDP), infectious neuropathy, perineuroma and NL [23]. Within the limits of resolution of MRI, it may be challenging to distinguish an extraneural soft tissue lymphomatous infiltration from true intraneural spread. The demonstration of thickening and enhancement of distal peripheral nerve branches or proximal radiculo-plexus elements may strongly suggest NL [2].

Baehring reported an MRI diagnostic sensitivity of 70% in a series of 40 cases of NL, combining the experience at the Massachusetts General Hospital and a review of the literature up to the year 2000 [1]. Most modern MR protocols routinely take advantage of more comprehensive image protocols (T1, T2, fat saturation, STIR), reduced slice thickness (1–3 mm), 1.5 T- 3 T magnet strength and multi-planar capability. In a review covering the period of 2001–2008, Grisariu reported a slightly higher sensitivity of 80%, likely reflecting such technical improvements in image acquisition [2].

¹⁸FDG-PET/CT is a very sensitive technique to reveal hypermetabolic foci, with one study reporting detection of at least one active site in 97% of patients with diffuse large B-cell lymphoma [24]. ¹⁸FDG-PET/CT has also been shown to be superior to MRI to detect bone marrow involvement in lymphoma [25]. PET imaging routinely provides a whole-body field of imaging and is uniquely suited to detect both nodal and extranodal tumor invasion. In a literature case review covering the 2001–2008-time period, NL was detected by PET in 90% of cases. In our review of lumbosacral NL (Table), PET was positive in 7/7 reports where it was done. Note that all such reports were from the current decade, reflecting the now recent use of PET in lymphoma screening. ¹⁸FDG-PET by itself is a highly sensitive screening tool which offers poor focal spatial resolution, thus it is commonly correlated with CT. Targeted MRI neurography at sites of hypermetabolism can be expected to substantially further improve the anatomical characterization of intraneural invasion. In addition, ¹⁸FDG-PET has been demonstrated to be helpful in monitoring the response to therapy [30, 31].

Neurolymphomatosis was once a diagnosis often only reached at autopsy. Advances in neuroimaging offer hope for early recognition, at a stage of discrete unifocal or even entirely asymptomatic lymphoma invasion. Survival rates remain poor with recurrent disease but newer protocols offer hope by combining high-intensity chemotherapy, targeted radiotherapy, immunotherapy (such as chimeric antigen receptor T-Cell) and salvage bone marrow transplantation [33].