Myelopathy following intrathecal chemotherapy in adults: a single institution experience

Introduction

Intrathecal (IT) chemotherapy in combination with systemic chemotherapy is an effective treatment modality used in the prophylaxis and treatment of central nervous system (CNS) disease in patients with hematological malignancies. IT agents commonly used in various chemotherapeutic regimens include methotrexate (MTX) and cytarabine arabinoside (Ara-C). Neurological side effects from IT chemotherapy vary from mild asymptomatic forms of arachnoiditis diagnosed on MRI to leukoencephalopathy and myelopathy [1, 2]. The more severe neurological complications have been reported more frequently in the pediatric literature. [3] The aim of this paper is to describe rare but debilitating neurological complications that may occur as sequelae of IT chemotherapy. Herein, we describe the clinical manifestations, radiological features and clinical outcomes in 7 adult patients with lymphoid leukemia who developed myelopathy following IT chemotherapy.

Methods

Results

The systemic disease status, frequency and dose of IT chemotherapy, time to onset of symptoms, and clinical characteristics of the 7 patients are listed in Table 1. There were four men and three women, with a median age of 48 (range, 18–70 years). Five of the seven patients were being treated for B-cell ALL, one for Burkitt’s leukemia/lymphoma, and one for CML-lymphoid blast crises. All patients received both IT Ara-C and MTX for the prophylaxis or treatment of CNS disease.

We reviewed the radiographic findings in the 611 patients treated at MDACC. 489 patients (470 ALL + 19 CML patients) (81%) did not have a spine MRI, 91 patients (86 ALL and 5 CML patients) had a an MRI spine which did not show any spinal cord abnormality (15%), 14 patients (all ALL patients) had evidence of enhancement suggestive of LMD (0.02%), 13 patients (all ALL patients) had evidence of leukemic infiltration (0.02%) and 5 patients (4 ALL patients and 1 CML patient) had preferential non-enhancing T2 hyperintensity involving the dorsal columns (<0.01%).

All seven patients who developed a myelopathy, had spinal cord abnormalities seen on MRI with 5 of the 7 patients manifesting T2 abnormalities involving the dorsal columns of the spinal cord as shown in Fig. 1 (MRI from patient #4). In the remaining 2 cases, enhancement along the conus medullaris and cauda equina nerve roots was noted and was initially thought to represent leptomeningeal disease, but the CSF was repeatedly negative.

The time to onset of neurological symptoms was variable, ranging from 2 days to 7 months from the initiation of IT chemotherapy. The time to maximal neurological deficit from onset of symptoms ranged from 10 days to 1 month (median = 14 days).Median survival was 3.1 months from the onset of myelopathy. Table 2 summarizes the investigations done on these 7 patients. Cerebrospinal fluid (CSF) analysis revealed elevated protein at presentation in 5 patients, decreased glucose in 0 patients and greater than 5 white blood cells (WBC) in CSF in 1 patient. All 7 patients eventually developed elevated protein levels in the CSF during the course of the myelopathy. CSF myelin basic protein (MBP), a marker of CNS inflammation and trauma, was assessed in 2 patients and was significantly elevated in both. Cytological evaluation of the CSF was repeatedly normal in all 7 patients (patients had CSF sent for cytology between 3–13 times per patient) whilst CSF flow cytometry was performed in 4 out of the 7 cases and was also negative. Despite normal cytology, the high clinical suspicion that the neurological symptoms were secondary to CNS leukemia prompted continuation of IT-chemotherapy for at least 1 more cycle beyond the onset of myelopathy symptoms in 5 of 7 patients. Lower extremity weakness of varying degrees was noted in all patients with 5 of 7 patients progressing to paraparesis. Two patients temporarily improved after discontinuation of IT chemotherapy (patient 1 and 7) but subsequently deteriorated with multi-organ failure (patient 1) and leptomeningeal disease (patient 7). The treatments for these 7 patients varied from the use of high dose steroids to intravenous immunoglobulin to radiation therapy with no difference in the outcome (Table 1). Patient #5 was the only patient to have an autopsy. The autopsy revealed almost complete transverse necrosis at the upper thoracic level, with small remnants of central gray matter showing ischemic (“red neuron”) changes. The lower thoracic cord also showed necrosis, with slightly greater preservation of the central gray matter and ischemic change similar to those seen in the upper thoracic cord. The cervical cord showed degeneration of the posterior columns, consistent with a Wallerian-like degeneration. The lumbar and sacral cord sections showed patchy edematous vacuolation and degeneration of the white matter columns. These histopathologic findings correlate accurately with the anatomic distribution of spinal cord abnormalities seen on antemortem MRI.

Table 2.

Results of relevant investigations done on the patients described in this study

	Patient 1	Patient 2	Patient 3	Patient 4	Patient 5	Patient 6	Patient 7
MRI brain	-	-	Areas of T2/FLAIR hyperintensity with focal restricted diffusion involving bilateral centrum semiovale/corona radiata, right greater than left	Extensive leuko- encephalopathy	Normal	T2/FLAIR abnormalities with enhancement in R inf. frontal lobe. Abnormal enhancement pineal and L in internal auditory canal	No evidence leuko- encephalopathy
MRI spine	Diffuse enhancement of the distal conus and the nerve roots of the cauda equina	Diffuse enhancement along the conus medullaris and around the cauda equina	T2 hyperintensity along dorsal columns of T & L spine	T2 hyperintensity along dorsal and lat columns of T & L spine, dorsal columns C-spine	Expansion of cord at T3-8. T2 hyperintensity along dorsal columns from cervical to thoracic region	Diffusely expanded cord with abnormal T2 signal abnormality of the dorsal columns from the cervical through the conus terminale	T2 signal abnormality in C and T spine
EMG	(Day 13) Severe motor> sensory generalized (worse distally) polyneuropathy mostly axonal	(Day 15) Normal EMG (5 months) Severe axonal polyradiculo- neuropathy or anterior horn cell disease	N/A	N/A	(Day 60) Mild axonal predominantly motor neuropathy (Day 77) Moderate predominantly axonal motor>sensory neuropathy with some demyelinating features	(Day 43) Severe chronic and ongoing sensori- motor polyneuropathy	(Day 39) Mild distal sensorimotor axonal poly- neuropathy
PET CT	No definite leptomeningeal or IT hyper- metabolism	No definite leptomeningeal or IT hyper- metabolism	N/A	No definite leptomeningeal or IT hyper- metabolism	No definite leptomeningeal or IT hyper- metabolism	N/A	No definite leptomeningeal or IT hyper- metabolism
CSF	WCC:16, RBC: 4450, protein: (over 2 weeks), glucose: 221, cytology:−ve (×4) Flow cytometry: −ve	WCC:0, RBC: 1, protein: (over 1 month), glucose:59, cytology: −ve (×4) Flow cytometry: N/A	WCC:0, RBC: 14875, protein: (over 2 weeks), glucose: 51, cytology:−ve (×5) Flow cytometry: N/A	WCC:0, RBC: 5, Protein: (over 2 weeks), glucose: 91, cytology: −ve (×3) Flow cytometry:−ve	WCC:0, RBC: 25, protein: (over 3 months), glucose: 60, cytology: −ve (×13) Flow cytometry: N/A	WCC:5,RBC: 210, protein:, glucose: 58, cytology: −ve (×3) Flow cytometry: −ve	WCC:0, RBC:4, protein: (over 2 months), glucose:65, cytology:−ve (×5) Flow cytometry: − ve
CSF PCR HSV, HSV6, CMV, VZV, West Nile, toxoplasma	−ve,N/A,−ve,−ve, N/A, N/A	−ve, N/A, −ve, −ve, − ve, −ve	−ve, −ve,−ve, −ve, −ve, −ve	−ve,−ve,−ve, N/A, − ve	−ve,−ve,−ve,−ve, N/A, N/A	−ve, −ve, −ve, −ve, N/A, N/A	−ve, −ve, −ve, −ve, N/A, −ve
CSF fungal assays: Cryptococcal Ag	−ve	−ve	−ve	−ve	−ve	−ve	−ve
Vit B12, MMA, copper, folate, homocysteine	790,N/A,N/A, 9.4 , N/A	278,0.26, 1.29,N/A,N/A	235, 0.09, 0.91, 50.2,	514, 0.24, N/A,	335,0.17, N/A, 12.3,	1312,0.1,N/A, 12.8, 9.2	1329,0.15, N/A, 13.2,
CSF MBP	-	-	-		-	-

Open in a new tab

^*

Abbreviations: Ag: Antigen, CMV: Cytomegalovirus, CSF: Cerebrospinal fluid, EMG: Electromyography, HSV: Herpes Simplex Virus, MBP: Myelin Basic Protein, MMA: Methylmalonic acid, N/A: not available, PET: Positron Emission Tomography, PCR: Polymerase Chain Reaction, RBC: Red Blood Cells, Vit: Vitamin, VZV: Varicella Zoster Virus, WCC: White Cell Count

^*

Results marked in red indicate abnormal values.

^*

In all 7 patients, an infectious workup in the CSF was negative

Discussion

The etiology of non-leukemic myelopathy in patients receiving intrathecal chemotherapy remains unclear. The neurotoxicity may be a result of aseptic meningitis incited by the preservative used in the diluent or alternatively in the case of methotrexate, it may be a result of local depletion of folate secondary to methotrexate. However, similar neurologic complications have been reported in patients who were given intrathecal preservative-free methotrexate or in those receiving Ara-C, suggesting that alternate mechanisms may be involved. [7]

Methotrexate is known to inhibit the enzyme dihydrofolate reductase which catalyzes the conversion of dihydrofolate (DHF) to tetrahydrofolate (THF) [8]. In the biologically active form, 5-methyl-THF is a donor of the methyl group for producing methionine and, in turn, S-adenosylmethionine (SAM). Therefore, IT methotrexate treatment can induce demyelination, by a pathogenic process similar to subacute combined degeneration (SACD), i.e. deficiency of methionine and SAM. The development of neurotoxic effects with IT-chemotherapy has been correlated with a high CSF level of methotrexate. [9] However, in our patients symptoms developed not only in patients who had received multiple doses of IT MTX, but also among those who had received one dose of IT chemotherapy. Also, myelopathy developed over a variable time range, from 2 days to up to 7 months after initiation of IT chemotherapy. In the pediatric literature a similar wide range has been reported. [10] These findings suggest that the toxic effects of methotrexate do not seem to be dose-related but may be related to individual sensitivity.

Why these 7 cases occurred in a single institution over a short period of time is unknown. We discussed these cases with other cancer institutions where the occurrence of this relatively rare complication of IT chemotherapy has not changed and no recent spikes in incidence of myelopathy have been observed. A few potential reasons for the unexpectedly high frequency of these cases in our institution include:

1. the high volume of B-cell ALL patients seen at MDACC (from January 2010 to March 2014 we had a total of 574 newly diagnosed B-cell ALL and 28 CML-lymphoid blast phase patients),

2. the increased survival of B-cell ALL patients [11–13] and

3. the increased awareness and attention to this phenomenon in our institution.

In our patients, we have no proof that MTX was the cause of the progressive myelopathy especially since all of the patients also received IT cytarabine. Liposomal cytarabine has been described in the literature as causing myelopathy as a complication of treatment in 2.5% of cases [14]. However in our case series, given the radiographic findings of preferential posterior column involvement, and knowing the biochemical similarities between methotrexate and vitamin B12 deficiency induced SACD, we believe that methotrexate was the more likely potential cause of neurological symptoms.

MRI is the most sensitive radiologic technique in the diagnosis of myelopathy. MRI abnormalities in patients receiving IT chemotherapy are non-specific and include cord swelling, T2 hyperintensity, and contrast enhancement. Five of 7 patients demonstrated longitudinal hyperintensities involving mostly the dorsal columns and to a varying degrees the lateral corticospinal tracts. This pattern has been described in other small case series and case reports. However a normal spine MRI does not exclude myelopathy [15]. This demonstrates that to make a diagnosis of IT induced myelopathy, a high index of suspicion is required in the appropriate clinical setting. In the remaining 2 of 7 cases, enhancement along the conus medullaris and cauda equina nerve roots was present.

EMG studies were performed on 5 out of the 7 patients. In 3 of the 5 patients, there was evidence of a severe polyneuropathy or polyradiculoneuropathy on EMG. In the two patients who did not have evidence of severe polyradiculoneuropathy (patients 5 and 7) one possible reason these severe changes were not seen could be the fact that the EMG was done relatively early in the course of the clinical spectrum (at 10 weeks and 2 weeks from the onset of weakness). This neurophysiological worsening over time was also reflected in the elevation of CSF protein in the same patients: 1, 2, 5 and 7 (patient 6 had only 1 spinal tap), indicating a correlation between the two studies and suggesting a possible measure to monitor treatment efficacy. Although EMG does not allow the clear distinction between a peripheral process and a central cause of weakness early on during the course of the disease, discordance on needle EMG showing mild denervation changes but a total absence of motor units (as seen in patient 6) should alert the treating physician to the possibility of a central cause. Early on, acute denervation changes such as fibrillation potentials and positive sharp waves are non-specific and do not allow localization.

One of the concerning questions in these patients is the ability to successfully distinguish between leukemic infiltration of the spinal cord and neurotoxic spinal cord damage since the treatment is drastically different for the two scenarios. This is especially difficult to determine when there is an elevated protein level in the CSF such as occurred in 5 of the 7 patients at presentation. In fact in 5 of 7 patients, IT chemotherapy was continued for at least 1 cycle after the onset of neurological symptoms due to the concern that the neurological symptoms were secondary to leukemic infiltration. This diagnostic uncertainty is underscored by the fact that the sensitivity of CSF cytopathological analysis to detect tumor cells is only 70% after the first lumbar puncture, approaching 90% and 95% if the puncture is repeated two or three times, respectively [16]. There is some evidence that elevated levels of MBP in the CSF could reflect spinal cord damage rather than leukemic infiltration since MBP is a protein produced by oligodendrocytes and is the major constituent of the myelin sheath of axons. [17] Though this was checked in only 2 of our cases, the levels of MBP were significantly elevated in both patients. Other causes of elevated CSF protein include Guillain Barre, and infectious processes which were excluded in all 7 patients by EMG and CSF analysis, respectively. CSF flow obstruction leading to pooling of intrathecal methotrexate is another cause that can be investigated by performing CSF flow studies prior to the administration of intrathecal medications. However CSF flow studies are not routinely done in most institutions and are not well standardized. Unfortunately none of our patients had this test done, although the likelihood that CSF flow obstruction was a principal cause of myelopathy in our cluster of cases is very low.

Treatment with high dose folate metabolites has shown some promising results in patients with methotrexate induced myelopathy and may be a potential consideration for future patients suspected to have chemotherapy induced myelopathy. [18]. Drachtman et al reported that dextromethorphan was effective in treating subacute methotrexate toxicity in children [19]. Interestingly the 2 patients in our series that received dextromethorphan improved temporarily. However, they also received other concurrent treatments, making it difficult to draw a direct conclusion. Further research is required to establish an effective and definite diagnostic and treatment approach for these patients.