Leptomeningeal metastasis (LMD) is a lethal complication caused by a variety of cancers, typically developing late in the disease course. It is associated with major neurologic disabilities and short survival. The incidence of LMD may increase because of longer survival of patients who have cancer, and because of the use of newer large-molecule therapies with poor central nervous system penetration. To achieve improved outcomes for patients who have LMD, new treatments need to reach the meninges and cerebrospinal fluid and interact with relevant molecular targets. Some of the agents currently in testing may contribute to this goal. To allow for better outcomes through earlier treatment, advances in diagnosis are needed. By using agents with higher therapeutic indices, in patients with a lower burden of disease (identified earlier with clinical or molecular markers) it should be possible to achieve gradual improvements in outcomes for patients suffering from this devastating disease.
Leptomeningeal metastasis (LMD) refers to the dissemination of cancer to the arachnoid mater, cerebrospinal fluid (CSF), and pia mater, which occurs in approximately 5% to 8% of all patients with cancer. The incidence of LMD may increase with better treatments and the overall longer survival of patients with cancer in general. The management of LMD requires the expertise of medical and neuro-oncologists, neurosurgeons, and radiation oncologists. LMD is typified by multifocal neurologic deficits and short survival in the 3- to 6-month range. The current standard of care includes external beam radiotherapy, and systemic and intrathecal (IT) chemotherapy. Future improvements in outcomes depend on: advances in the understanding of the molecular changes that allow for central nervous system seeding of cancer, the identification of subgroups of patients who can be predicted to develop LMD and in whom preventative measures can be instituted, and the development of effective drugs that penetrate, or can be directly administered into, the CSF.
Epidemiology
It is estimated that between 5% and 8% of patients with cancer will develop LMD. Of patients with cancer who have neurologic symptoms and who undergo autopsy, as many as 19% can show evidence of leptomeningeal seeding by their cancer. Based on recent data from the National Cancer Institute (NCI) Surveillance Epidemiology and End Results (SEER) Web site, the age-adjusted incidence for all cancers is 461.6 per 100,000 men and women per year. Using 300 million as the population of the United States, this equates to approximately 1.4 million cancer cases, and as many as 110,000 LMD cases, per year in the United States.
The most common cancers that result in LMD include lung cancer, breast cancer, melanoma, acute lymphoblastic leukemia (ALL), and non-Hodgkin lymphoma (NHL), but virtually any cancer can metastasize to the leptomeninges. The incidence of LMD may increase in the future because of improved survival for cancers in general. The increased use of large-molecule systemic therapies that create a sanctuary site for cancer cells behind the blood-brain barrier and blood-CSF barrier (BCSFB) may also promote the development of LMD.
Pathophysiology
The seeding of cancer cells to the leptomeninges and CSF can be considered from an anatomic or from a molecular pathophysiology perspective. Anatomically, tumor cells must reach the leptomeninges to disseminate through the system. This dissemination occurs through either hematologic dissemination, via direct extension from a tumor mass adjacent to the meninges, or from a preexisting parenchymal central nervous system (CNS) metastasis (synchronous or preexisting CNS metastases are found in 28%–75% of patients with LMD).
Intraoperative spread of tumor cells from a CNS metastasis is uncommon but does occur as shown by the way piecemeal tumor resection (vs en bloc resection or stereotactic radiation) increases CSF dissemination rates 2.4- to 5.8-fold. Primary CNS malignancies, such as medulloblastomas, pineal region tumors, and spinal cord tumors that reside near the CSF pathways, are particularly prone to CSF dissemination.
Once tumor cells reach the leptomeninges, they can disseminate throughout the CNS. Tumor cells can travel along the pia mater and invade the subpial parenchyma, penetrate nerves, and produce masses in the subarachnoid space. Cells can be carried in the CSF flow through the nervous system, resulting in multifocal CNS disease. The skull base and the sacral thecal sac are particularly prone to tumor cell build up because of their gravitationally dependent positions.
The molecular pathophysiology of LMD is not well understood. However, the principles that apply to metastatic disease in general apply to LMD. To successfully metastasize, tumor cells must (1) possess unlimited growth potential and a high level of genomic instability; (2) be able to invade basement membranes and mobilize bone marrow constituents; (3) be able to remodel blood vessels, survive in the circulatory system and evade immune surveillance, and extravasate out of the vascular system; and (4) invade, proliferate, and establish a blood supply in the host organ. These cellular activities require the hijacking of complex molecular machinery typically active during embryogenesis. Even though specific genetic mutations that produce LMD have not been identified, some molecular signals related to the disease have been found. Homing proteins, such as stromal derived factor-1 alpha and angiogenesis-related molecules, such as vascular endothelial growth factor (VEGF) are present in increased levels in the CSF of patients with LMD. These proteins may be related to the biology and development of the disease, or simply epiphenomena.
Pathophysiology
The seeding of cancer cells to the leptomeninges and CSF can be considered from an anatomic or from a molecular pathophysiology perspective. Anatomically, tumor cells must reach the leptomeninges to disseminate through the system. This dissemination occurs through either hematologic dissemination, via direct extension from a tumor mass adjacent to the meninges, or from a preexisting parenchymal central nervous system (CNS) metastasis (synchronous or preexisting CNS metastases are found in 28%–75% of patients with LMD).
Intraoperative spread of tumor cells from a CNS metastasis is uncommon but does occur as shown by the way piecemeal tumor resection (vs en bloc resection or stereotactic radiation) increases CSF dissemination rates 2.4- to 5.8-fold. Primary CNS malignancies, such as medulloblastomas, pineal region tumors, and spinal cord tumors that reside near the CSF pathways, are particularly prone to CSF dissemination.
Once tumor cells reach the leptomeninges, they can disseminate throughout the CNS. Tumor cells can travel along the pia mater and invade the subpial parenchyma, penetrate nerves, and produce masses in the subarachnoid space. Cells can be carried in the CSF flow through the nervous system, resulting in multifocal CNS disease. The skull base and the sacral thecal sac are particularly prone to tumor cell build up because of their gravitationally dependent positions.
The molecular pathophysiology of LMD is not well understood. However, the principles that apply to metastatic disease in general apply to LMD. To successfully metastasize, tumor cells must (1) possess unlimited growth potential and a high level of genomic instability; (2) be able to invade basement membranes and mobilize bone marrow constituents; (3) be able to remodel blood vessels, survive in the circulatory system and evade immune surveillance, and extravasate out of the vascular system; and (4) invade, proliferate, and establish a blood supply in the host organ. These cellular activities require the hijacking of complex molecular machinery typically active during embryogenesis. Even though specific genetic mutations that produce LMD have not been identified, some molecular signals related to the disease have been found. Homing proteins, such as stromal derived factor-1 alpha and angiogenesis-related molecules, such as vascular endothelial growth factor (VEGF) are present in increased levels in the CSF of patients with LMD. These proteins may be related to the biology and development of the disease, or simply epiphenomena.
Diagnosis
Clinical Findings
Symptoms and signs of LMD are related to the location of tumor deposits, or obstruction of CSF flow pathways, and are usually organized into 3 categories: cerebral, cranial nerve, and spinal. Common cerebral symptoms and signs include headache, altered mental status, gait difficulty, nausea or vomiting, incoordination, syncope, and cerebellar signs. The most frequent cranial nerve–related symptoms are diplopia, vision loss, hearing changes, and facial weakness. Common spinal symptoms include lower motor neuron weakness, paresthesias, radicular pain, neck or back pain, and bladder or bowel dysfunction. Multifocal symptoms and findings are typical in patients with LMD.
Imaging
Because of its higher sensitivity, gadolinium contrast-enhanced magnetic resonance imaging (MRI) of the CNS has replaced computed tomography as the imaging modality of choice in patients suspected of having LMD. Imaging of the entire neuraxis is necessary to properly quantify the extent of CNS disease and to allow for a coherent and organized treatment plan. Imaging findings on MRI that are suggestive of LMD include contrast enhancement of the leptomeninges, subependyma, cranial and spinal nerves, as well as communicating hydrocephalus. Typical imaging findings of LMD are depicted in Figs. 1–4 . Neuroimaging is more likely to be abnormal in patients with LMD from solid tumors (72%–80%) versus those with LMD caused by hematologic malignancies (48%–62%).
Leptomeningeal enhancement is suggestive of LMD, but not diagnostic. Therefore, one must consider other conditions that can produce similar imaging features, such as the effects of intracranial hypotension after craniotomy or lumbar puncture, as well as infectious or inflammatory diseases.
Up to 61% of patients with LMD develop CSF flow blocks at various levels along the CSF pathway. Such flow blocks can result in loculation of intrathecally administered drugs, causing severe toxicity and impaired survival. Therefore, once a decision is made to treat a patient who has LMD with intrathecal (IT) chemotherapy, CSF flow is assessed by injecting a radioactive tracer (usually pentaacetic acid labeled with indium 111). Focal radiation that opens areas of CSF flow obstruction improves survival.
CSF
Cytologic examination of the CSF identifying malignant cells is the sine qua non of the diagnosis of LMD. However, the sensitivity of a single CSF cytologic examination can be as low as 45%, and 5% of patients with LMD have normal CSF. Four simple measures can increase the likelihood of finding malignant cells from a CSF sample: (1) obtain at least 10.5 mL of CSF for analysis, (2) immediately process the sample, (3) obtain CSF from a site adjacent to the affected CNS region, and (4) repeat the CSF sampling and analysis. By repeating the CSF cytological analysis, one can increase the sensitivity of finding malignant cells to more than 77%. Because of its higher sensitivity, lumbar space (rather than ventricular) CSF, should be used for diagnostic and treatment response purposes.
Discordance between imaging and CSF findings is common. In patients with LMD who have solid tumors, CSF is negative in 17% to 23% in the face of positive MRI findings. Conversely, in patients with LMD who have liquid tumors, MRI is unrevealing in 38% to 52% of patients whose CSF contains malignant cells. Because of their lack of specificity, other standard CSF analyses performed at the time of lumbar puncture can only be used to direct attention toward additional testing, rather than provide a basis for treatment. Traditional CSF analyses include CSF opening pressure, increased (>20 cm water) in 50% of patients with LMD; CSF white blood cell count, increased (>5 per mm 3 ) in 64% of patients with LMD; CSF protein concentration, increased (>50 mg/dL) in 59% of patients with LMD; and CSF glucose concentration, low (<40 mg/dL) in 31% of patients with LMD.
To increase the yield of CSF analysis, several soluble protein and enzyme markers of LMD in the CSF have been identified. Potential usefulness has been noted for CSF levels of β glucuronidase, lactate dehydrogenase, β-2-microglobulin, carcinoembryonic antigen, and β-human chorionic gonadotrophin. However, none of these molecules have become widely used, primarily because of their lack of specificity. If identified, these markers can be used to track disease progress, but the serum levels of the molecules must also be followed to ensure that there is true CSF production rather than transfer across the BCSFB.
Recent studies have identified increased levels of molecules in the CSF of patients with LMD related to cellular homing and migration, and angiogenesis. These may prove useful in tracking or targeting the disease, or in better understanding its biology.
In hematological malignancies, flow cytometry analysis of the CSF improves the sensitivity compared with standard CSF morphologic cytologic analysis, and is the preferred testing method when LMD is a concern.
Treatment
Aggressive treatment of patients with LMD is controversial; this is because of the low efficacy of most treatments, the overall poor prognosis, and the risks of treatment-related toxicities. Many clinicians advise against aggressive treatment for patients with LMD. Even though neurologic deficits rarely improve, one argument in favor of treatment is that early intervention may improve quality of life and prevent neurologic death.
Treatment of LMD is considered palliative, because even aggressive treatment directed at the leptomeninges may only increase survival by 1 to 3 months. Without intervention, median survival is typically 4 to 6 weeks, after which death is often caused by progressive neurologic decline.
Therapy for LMD can provide control of the meningeal disease. However, even with treatment, one-quarter of patients with LMD die as a direct consequence of the LMD, and up to 60% die with simultaneous progression of the LMD and the systemic cancer.
Treatment of LMD should target the entire neuraxis because tumor cells disseminate throughout the CSF pathway. For a comprehensive approach, both bulky disease and malignant cells floating in the CSF, as well as the patient’s systemic disease, are treated. This most comprehensive approach requires focal (radiotherapy to bulky meningeal disease), local (IT chemotherapy for the disease present in the CSF), and systemic (systemic chemotherapy for cancer outside the nervous system) treatments.
Guidelines published by the National Comprehensive Cancer Network (2006) suggest stratifying patients as either poor or good risks for survival to aid in determining how aggressive the treatment should be. These guidelines define poor-risk patients as those with (1) a low Karnofsky performance scale (KPS) score; (2) multiple, serious, fixed neurologic deficits; and (3) extensive systemic disease with few treatment options. These patients may be best served by supportive care and radiotherapy to symptomatic sites of disease. Patients who are a good risk and who may warrant more aggressive treatment are those with (1) a high KPS score, (2) no fixed neurologic deficits, (3) limited systemic disease, and (4) effective systemic treatment options.
Because of physical or cognitive impairment, patients with LMD may require more assistance than is readily available. For practicality and safety, modifications of the home and special means of transportation may be necessary. Counseling to address the psychological burden of LMD on patients and their caregivers is often warranted.
Radiotherapy
External beam radiotherapy is often used in patients with LMD. Involved field radiotherapy can be delivered to the meninges if bulky disease is present and causing symptoms (often seen with involvement of cranial nerves) or when CSF flow obstruction occurs. Thirty Gray given in 10 fractions is a typical radiation dosing scheme. Because of the higher likelihood of response, patients with breast cancer, leukemia, and lymphoma are considered best for this therapy. Even though the entire CSF system might be considered a target for radiotherapy, craniospinal radiotherapy is rarely administered because of toxicity concerns.
Systemic Chemotherapy
In adults with ALL (aALL), systemic chemotherapy (SC) agents possessing high levels of CNS penetration, along with IT chemotherapy, are now standard in the prevention of LMD. By incorporating CNS prophylactic regimens, the 5-year CNS relapse rate in aALL has been reduced from 58% to 8%. The studies supporting the CNS prophylactic approach in adults were undertaken to try to emulate the outcomes seen in childhood ALL (cALL). In cALL, before the incorporation of CNS prophylaxis, the risk of the development of CNS leukemia at 4 years was 75%. After CNS prophylactic treatments were added, the 5-year event free survival improved to 80%.
Evidence to support the use of SC (or hormonal therapy) in the treatment or prophylaxis of solid tumor LMD is less strong. However, there are reports of responses and improved survival in patients with LMD treated with SC, often without IT chemotherapy.
Despite reports of usefulness, the use of SC in LMD is questioned because therapeutic CSF concentrations of most SCs are usually not considered to be achievable. Table 1 lists commonly used SCs that achieve greater than 0.05 CSF/plasma ratio (a ratio <0.05 signifies nonspecific leakage of drug across the BCSFB). Reasonable approaches to treating LMD include the use of SCs with high levels of CSF penetration either with or without IT chemotherapy. However, there is a risk for increased neurotoxicity when combining such agents.
| Drug | CSF: Plasma Ratio | Reference |
|---|---|---|
| Triethylenethiophosphoramide (thiotepa) | 1.0 | |
| Busulfan | 0.95 | |
| Temozolomide | 0.20 | |
| Tiazofurin | 0.28 | |
| 6-Mercaptopurine | 0.27 | |
| 5-Fluorouracil | 0.155 | |
| Arabinosyl-5-azacytidine | 0.15 | |
| Cytosine arabinoside | 0.06–0.22 | |
| Topotecan (lactone) | 0.29–0.42 | |
| Hydroxyurea | 0.24 a | |
| Cyclophosphamide Ifosfamide | 0.20 (0.00–1.1) 1.2 (0.4–1.6) b |
Intrathecal Chemotherapy
The theory behind using IT chemotherapy is to treat (1) subclinical leptomeningeal deposits and (2) any viable tumor cells floating in the CSF to prevent further leptomeningeal seeding, thereby preserving neurologic functioning and improving survival. As noted earlier, not all clinicians advise the use of IT chemotherapy because of limited randomized data showing benefit, and concerns about toxicity.
Route of administration of intrathecal chemotherapy
To avoid frequent lumbar punctures and injury, and for possible improved efficacy of treatment, most patients with LMD have a ventricular reservoir placed for drug delivery. Delivery of chemotherapeutic drugs by lumbar puncture can result in drug placement outside the thecal sac and accompanying tissue damage. More variability of ventricular drug concentrations is seen after intralumbar drug administration. For drugs with a short half-life, intraventricular administration results in better outcomes, probably because of the more even distribution of the drug throughout the thecal sac. In CNS leukemia, drug delivery via a ventricular reservoir improves the durability of remission compared with intralumbar delivery. Ventricular reservoirs are usually well tolerated, but complications such as misplacement, catheter tip occlusion, and infection can occur in as many as 5% of patients.
Commonly used intrathecal chemotherapies
The most commonly used IT chemotherapies for LMD from all causes include the antimetabolite, methotrexate (MTX); the pyrimidine analogue, cytarabine (Ara-C) and its longer-acting liposomal version; and the alkylator, thiotepa (triethylenethiophosphoramide). Typical schedules of administration of IT chemotherapies include a high-dose induction phase (dose based on specific drug, usually twice weekly × 4–6 weeks), followed by a less intensive consolidation phase (usually once-weekly dosing) and an even less intensive maintenance phase. Dosing schemes using a concentration-times-time technique, prolonging tumor cell drug exposure, may improve outcomes.
Few randomized data exist comparing the efficacy of these drugs. In the randomized studies that do exist, nonsignificant differences in survival have been shown when comparing IT chemotherapies with each other or in various combinations. However, in lymphomatous meningitis, liposomal Ara-C results in a higher degree of tumor cell clearance from the CSF and a longer median time to neurologic progression and survival duration compared with the standard formulation of Ara-C. The benefits of liposomal Ara-C seen in patients with lymphomatous meningitis were not seen in patients with LMD who had solid tumors.
Related posts:
Review of Imaging Techniques in the Diagnosis and Management of Brain Metastases
Radiotherapy for Brain Metastases
Management of Skull Base Metastases
Radiosurgical Management of Brain Metastases
Investigational Therapies for Brain Metastases
Evidence-Based Guidelines for the Management of Brain Metastases
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
