Version May 2024
INTRODUCTION — Leptomeningeal disease (LMD; also referred to as leptomeningeal metastases or carcinomatous meningitis) is a rare but frequently devastating complication of advanced cancer from solid tumors, most commonly lung cancer, breast cancer, and melanoma. Patients can present with a broad range of signs and symptoms due to simultaneous involvement of multiple areas of the craniospinal axis. Diagnosis often requires a high index of suspicion and is confirmed by neuroimaging and/or cerebrospinal fluid (CSF) analysis.
Patients with LMD have complex needs and are at risk for rapid deterioration. Patients often require multidisciplinary care to achieve optimal symptom control and quality of life. In addition to multifocal neurologic signs and symptoms, patients may have hydrocephalus, seizures, and dysfunction of the hypothalamic-pituitary axis.
LMD is traditionally a late-stage complication of advanced cancer with a median survival of several months. However, with improvements in systemic cancer therapies, the natural history of central nervous system (CNS) metastases is evolving, particularly for certain cancer subtypes. Prolonged survival is occasionally achieved, and selected patients with LMD have a wider range of treatment options available to them than they did previously.
The treatment of LMD from solid tumors is reviewed here. The pathophysiology, clinical manifestations, and diagnosis of LMD, and leptomeningeal involvement in hematologic malignancies, especially large cell lymphomas and acute leukemias, are discussed separately. (See "Clinical features and diagnosis of leptomeningeal disease from solid tumors" and "Secondary central nervous system lymphoma: Clinical features and diagnosis", section on 'Leptomeningeal dissemination' and "Involvement of the central nervous system (CNS) with acute myeloid leukemia (AML)".)
OVERVIEW OF APPROACH
Prognosis — LMD from solid tumors is usually a late-stage development associated with a prognosis of three to four months [1-5]. Patients often have widespread systemic disease, and approximately 40 percent have prior or concurrent parenchymal brain metastases [5]. Patients may succumb to LMD, progressive brain metastases, complications of systemic disease, or a combination. Nonetheless, sustained tumor control is occasionally achieved, especially in younger patients with breast cancer or oncogene-driven non-small cell lung cancer (NSCLC), as well as those with controlled systemic neoplasm and good remaining treatment options.
Tumor histology and molecular subtype are important factors influencing prognosis. In a large series of 318 patients with breast cancer and LMD, median overall survival was 3.5 months for the entire cohort [6]. Survival was longest among those with human epidermal growth factor receptor 2 (HER2)-positive disease and shortest for those with triple-negative disease (5.2 versus 2.5 months). Similarly, in NSCLC, patients with LMD from epidermal growth factor receptor (EGFR)-mutant or anaplastic lymphoma kinase (ALK)-rearranged tumors have the potential for more extended survival (eg, 12 to 18 months) compared with patients without a driver mutation [7-12]. (See 'Patients with non-small cell lung cancer' below.)
Diagnostic certainty and the radiographic pattern of LMD also impact prognosis. In a multicenter retrospective study of 254 patients with newly diagnosed solid tumor LMD, median survival was shortest in patients with confirmed LMD, defined by positive cerebrospinal fluid (CSF) cytology or biopsy (2.3 months, compared with 2.8 months for the overall cohort) [5]. Older age and non-breast histology were also associated with shorter survival. Among patients with possible or probable LMD (clinical and/or radiologic evidence but equivocal or negative cytology), those with nodular disease on imaging (image 1) had shorter survival compared with those who had linear or absent leptomeningeal enhancement.
Identifying patients with a particularly poor prognosis can help to spare unnecessary or futile interventions and maximize supportive care and comfort (table 1). Patients with a Karnofsky Performance Status (KPS) <60 (table 2); multiple, serious, fixed neurologic deficits; untreated symptomatic hydrocephalus; or extensive systemic cancer with limited remaining therapeutic options have a poor prognosis (eg, one to two months) even with active treatment.
In the absence of these factors, the potential for sustained tumor control depends upon an individualized assessment of tumor histology and molecular subtype, extent of disease, and prior therapies. Patients with reasonable systemic treatment options, particularly agents with central nervous system (CNS) activity, do best.
Treatment plan and sequencing — The goals of treatment in patients with LMD are to stabilize or improve neurologic function, palliate symptoms, and improve quality of life. All therapies are given with palliative intent, and certain urgent procedures, such as ventriculoperitoneal (VP) shunting for hydrocephalus, can also extend life in selected patients.
For relatively good-risk patients with a cancer for which reasonable treatment options remain, a treatment hierarchy is helpful to sequentially organize multiple required interventions (algorithm 1):
●Bulky, symptomatic disease – Bulky and neurologically symptomatic areas of LMD are unlikely to respond to systemic or intrathecal (IT) therapy but often improve when treated with radiation therapy (RT). RT is typically delivered to the most symptomatic area. Radiation to the full neuroaxis (ie, craniospinal RT) utilizing protons may be considered in selected patients with better prognosis. (See 'Radiation to symptomatic sites' below.)
●Hydrocephalus – A high index of suspicion for hydrocephalus is necessary in patients with LMD, as ventricles may not appear dilated, overtly obstructive disease may be absent, and symptoms may be attributed to other sites of disease. Papilledema, an opening pressure of >20 cm H2O on lumbar puncture (LP), and plateau waves are clues to the diagnosis. Plateau waves are frequently confused with seizures, even by experienced clinicians. (See "Clinical features and diagnosis of leptomeningeal disease from solid tumors", section on 'Plateau waves'.)
In some cases, urgent management of severe hydrocephalus may require delaying RT. In mild cases, RT can lead to symptom resolution and forestall or avoid shunt placement. (See 'Hydrocephalus and increased intracranial pressure' below.)
●Further CNS-directed therapy – For poor-risk patients (low KPS, widespread systemic disease without good systemic treatment options), treatment focuses upon alleviating symptoms without the use of systemic or IT anticancer therapy. For good-risk patients, selection of therapy is individualized based on extent of disease, cancer type, cancer genomics, and prior systemic treatment responsiveness.
Control of both CNS and systemic disease may require multiple agents, although many newer therapies have demonstrated dual CNS and systemic activity. If such a treatment option exists for a patient, then we proceed with systemic therapy rather than, or in rare cases along with, IT therapy. The preference for a systemic agent is strongest for patients with active extracranial disease or concurrent parenchymal brain metastases, for whom IT therapy would not address all sites of disease. (See 'Systemic therapy options' below.)
For patients without good systemic therapy options (eg, past adverse effects, prior lack of response, or poor reserve) who remain eligible for CNS-directed therapy (high KPS, isolated site of disease, limited extracranial disease burden), we offer IT chemotherapy. The best candidates for IT therapy are those with linear, nonbulky LMD; no hydrocephalus; and absent brain metastases. (See 'Intrathecal therapy' below.)
Supportive care needs — Intensive supportive care is important and often essential to achieve acceptable quality of life in patients with LMD. Treatment of LMD involves complex care coordination, the potential for several highly morbid complications of the disease and its treatment, and thoughtful balancing of symptom relief with the principles of self-determination.
●Palliative care – At each visit, a careful history and detailed neurologic evaluation should focus on frequent complications of LMD in order to anticipate palliative care needs. Cauda equina syndrome, hyponatremia, seizures, and pain require vigilance and urgent treatment. Importantly, LMD can evolve rapidly, and functional status may decline quickly in the setting of uncontrolled symptoms.
Frequent clinical evaluations are often necessary to evaluate new symptoms, identify treatable causes, and reevaluate goals of care. Early involvement of palliative care clinicians is encouraged, whether or not a patient is pursuing disease-modifying therapy.
●Pain management – LMD can cause headaches, neck pain or stiffness, back pain, and radicular nerve pain. Analgesics (acetaminophen, nonsteroidal antiinflammatory drugs, opioids) are appropriate for initial management in most cases. Glucocorticoids rarely reverse neurologic deficits from LMD but may improve headaches and radicular pain more effectively than analgesics.
VP shunting is occasionally necessary as a palliative procedure for refractory headaches that are due to increased intracranial pressure (ICP). In patients with radicular pain, RT plays an important palliative role, especially for bulky cauda equina disease (image 2) (see 'Radiation to symptomatic sites' below). Neuropathic agents such as gabapentin, pregabalin, and duloxetine can be useful as adjunctive therapies. (See "Cancer pain management: Role of adjuvant analgesics (coanalgesics)", section on 'Patients with neuropathic pain'.)
●Seizures – Seizures occur in approximately one-quarter of patients with LMD. Patients who experience a seizure should be treated with an antiseizure medication indefinitely to decrease the risk of recurrent seizures. Levetiracetam is often a good first-line choice due to ease of administration and low potential for drug interactions [13,14]. (See "Seizures in patients with primary and metastatic brain tumors", section on 'Approach to management'.)
Plateau waves may be mistaken for seizures and can often be recognized because they occur during positional changes. Plateau waves indicate increased ICP and will not respond to antiseizure medications. (See "Clinical features and diagnosis of leptomeningeal disease from solid tumors", section on 'Plateau waves'.)
As in patients with brain metastases, routine prophylaxis in patients who have never had a seizure is not indicated and exposes patients to risk of side effects of antiseizure medications. (See "Seizures in patients with primary and metastatic brain tumors", section on 'Approach to management'.)
Response evaluation — Response assessment should include the triad of clinical signs and symptoms, review of neuroimaging, and CSF evaluation. Most patients will require frequent clinical assessments (eg, every two to four weeks) along with brain and spine magnetic resonance imaging (MRI) every two to three months.
The frequency of CSF evaluation depends on its ready availability. CSF can be drawn from an Ommaya or may require an LP. For those who are not receiving IT therapy, we generally coordinate CSF cytology with neuroimaging and sometimes more frequently.
For patients receiving intra-CSF therapy, we evaluate CSF for routine studies (ie, protein, glucose, cell count, Gram stain, and culture) at the time of each IT treatment and send cytology at every two to four weeks. Both ventricular and thecal CSF should be evaluated whenever possible because absence of malignant cells in one compartment (eg, ventricular) does not guarantee resolution of neoplasm in the other (eg, lumbar).
Other testing on CSF may be useful in some cases, either diagnostically or to follow response. These techniques, such as cell-free DNA analysis on CSF or measurement of tumor markers specific to a given tumor type, are reviewed separately. (See "Clinical features and diagnosis of leptomeningeal disease from solid tumors", section on 'Cerebrospinal fluid'.)
URGENT SYMPTOMATIC INTERVENTIONS
Radiation to symptomatic sites — Radiation therapy (RT) should be used whenever possible to treat symptomatic areas of LMD, especially for bulky deposits in the cauda equina causing weakness, pain, or incontinence [15,16].
Involved field radiation — The specific fields are dictated by symptoms, as reviewed below. RT frequently alleviates radicular pain and sometimes improves focal neurologic deficits such as cranial neuropathies or diffuse symptoms such as encephalopathy. It may also delay or prevent progression of neurologic deficits. By contrast, systemic therapy alone rarely reverses neurologic signs and symptoms from solid tumor LMD.
The RT plan should be individualized based on symptoms and disease distribution:
●Patients with cauda equina syndrome and/or painful radiculopathy – Lower-extremity weakness and bladder or bowel dysfunction are often seen with radiographic evidence of bulky enhancing neoplasm in the cauda equina (image 3 and image 2). Prompt lumbosacral spine RT may recover function (image 4). Occasionally, such symptoms occur without radiographic findings. These patients may benefit from the same palliative lumbosacral spine RT despite the absence of radiographic targets.
For patients with painful radiculopathies above the level of the cauda equina, cervical or thoracic spine RT to the area of radiographic and clinical involvement is administered.
●Patients with cranial neuropathies – Isolated cranial neuropathies may resolve with whole brain RT or limited skull base RT. The typical schedule for either whole brain RT or more limited skull base fields is 30 Gy in 10 daily fractions. Skull base RT may spare temporal lobe regions to avoid memory deficits and can sometimes be given in patients who have already received whole brain RT in the past [17].
●Patients with diffuse encephalopathy – Patients with diffusely symptomatic intracranial LMD are candidates for whole brain RT, especially in the presence of concurrent parenchymal brain metastases. As with other forms of RT in patients with LMD, the goal of treatment is symptom control, and overall survival may not be prolonged [18-20]. A typical schedule for whole brain RT is 30 Gy in 10 daily fractions; for poor-prognosis patients, 20 Gy in five daily fractions is sometimes used to minimize patient burden and treatment time.
Major adverse effects during or after focal RT alone are rare. With more extensive fields (eg, craniospinal or whole brain RT), major effects of RT may include myelosuppression, mucositis, esophagitis, and leukoencephalopathy. Whole brain RT can result in cognitive decline, which typically takes months or even years to become clinically apparent. However, toxicity may arise earlier when RT is given concurrently with intrathecal (IT) methotrexate. Although many patients do not live long enough to suffer the delayed effects of RT, some do. (See "Acute complications of cranial irradiation" and "Delayed complications of cranial irradiation".)
Proton craniospinal radiation — Radiation to the full neuraxis (ie, craniospinal RT) with photon radiation is typically not appropriate for solid tumor LMD due to high risk of toxicity, short life expectancy, and low likelihood of meaningful benefit. However, where available, proton craniospinal RT, which has an improved toxicity profile compared with photon craniospinal RT, is an option in selected patients with adequate functional status [21,22].
In a randomized phase II trial of proton craniospinal RT versus involved field RT in 63 patients with LMD from either breast cancer or non-small cell lung cancer (NSCLC), proton craniospinal RT improved both median central nervous system (CNS) progression-free survival (7.5 versus 2.3 months) and overall survival (9.9 versus 6.0 months) compared with involved field RT, with no difference in the rate of serious adverse effects [22]. Approximately one-half of patients in each group had active systemic disease at the time of enrollment, and nearly one-half in each group had a targetable variant (eg, HER2, epidermal growth factor receptor [EGFR]). In an exploratory cohort of 35 patients with other histologies treated with proton craniospinal RT, median CNS progression-free and overall survival were 5.8 and 6.6 months, respectively, and there were more systemic disease progression-related deaths compared with the breast and NSCLC patients.
Hydrocephalus and increased intracranial pressure — Hydrocephalus and increased intracranial pressure (ICP), even without ventriculomegaly, are significant causes of intractable headache, cognitive slowing, gait abnormalities, and incontinence in patients with LMD. Hydrocephalus from LMD is a subacute process, and symptoms usually evolve over the course of weeks rather than days. With a high index of suspicion and early recognition, decisions about shunting can often be made in the outpatient setting.
Patients with LMD may have either communicating or noncommunicating hydrocephalus, and treatment approaches vary depending on the type.
●In communicating hydrocephalus from LMD, the ventricles are symmetrically dilated above and below the third ventricle without focal sites of obstruction. The mechanism is thought to be an interruption of cerebrospinal fluid (CSF) resorption by malignant cell clumping at the level of the arachnoid granulations, which in most cases is microscopic and not visible on magnetic resonance imaging (MRI) (image 3).
●Less commonly in LMD, noncommunicating (obstructive) hydrocephalus occurs due to bulky parenchymal or leptomeningeal tumors obstructing the cerebral aqueduct, fourth ventricle, or spinal canal. CSF flow studies occasionally identify an otherwise radiographically occult regional block that may be a target for focal RT [23]. Rarely, hydrocephalus may occur without intracranial disease in the setting of nodular spinal cord deposits and respond to focal RT to the spine.
Glucocorticoids — A trial of glucocorticoids is appropriate for patients with symptoms of elevated ICP from LMD, although steroids are not a definitive treatment for hydrocephalus. Nonetheless, some patients will respond symptomatically with a decrease in headaches.
A typical starting dose is dexamethasone 8 to 16 mg per day in one or two divided doses. Maximal clinical response should occur within two to three days of starting steroids. A taper can then be attempted, aiming for the lowest effective dose necessary to control symptoms. (See "Management of vasogenic edema in patients with primary and metastatic brain tumors", section on 'Symptomatic treatment'.)
For patients with frank hydrocephalus on imaging, glucocorticoids are less likely to improve symptoms and are generally only a temporizing measure until definitive treatment with CSF diversion or RT.
Lumbar puncture — Because most LMD-related hydrocephalus is communicating and arises without an overtly obstructive mass lesion, urgent lumbar puncture (LP) typically resolves hydrocephalus temporarily without risk for herniation. LP then allows time for further management and provides CSF for analysis and opening pressure measurement. Although CSF is produced at a rate of 20 mL/hour, relief of symptomatic hydrocephalus typically lasts for days.
Shunting versus radiation therapy — In patients for whom further disease-modifying therapy is planned, hydrocephalus from LMD requires either CSF diversion via ventriculoperitoneal (VP) shunting or RT. Systemic and IT therapy alone do not address hydrocephalus.
The appropriate therapy depends on the severity of hydrocephalus and the type (communicating or noncommunicating).
●Severe hydrocephalus – Patients with severe hydrocephalus of either type require a VP shunt for relief of symptoms, even if symptoms improve with steroids, pain medications, and antiemetics. Although overall survival may not be improved by CSF diversion, shunting commonly relieves suffering and improves quality of life. (See 'Ventriculoperitoneal shunting' below.)
Decisions about a surgical intervention in patients with a terminal condition such as LMD should include patients' known wishes, caregiver input, and ethical considerations governing end-of-life care [24]. Hydrocephalus causes cognitive deficits, gait impairment, incontinence, and symptoms of increased ICP such as headache and intractable nausea and vomiting. While not all of these symptoms will improve following CSF diversion, none will improve without intervention. (See "Ethical issues in palliative care".)
●Less severe, communicating hydrocephalus – For patients with milder degrees of communicating hydrocephalus, and for those with nonspecific symptoms that may or may not be the result of hydrocephalus, an LP with measurement of opening pressure is often useful as an initial diagnostic and therapeutic procedure. A therapeutic response to high-volume LP (eg, removal of greater than 20 mL of CSF) offers reasonable assurance that VP shunting will provide palliative benefit. (See 'Ventriculoperitoneal shunting' below.)
For patients who undergo LP and respond with a decrease in headaches or other symptoms, we typically offer VP shunting. In patients who do not respond clinically to high-volume LP, or for those who prefer to avoid a surgical procedure, whole brain or skull base RT may forestall and sometimes resolve hydrocephalus and elevated ICP.
●Less severe, noncommunicating (obstructive) hydrocephalus – Unlike patients with communicating hydrocephalus, patients with obstructive hydrocephalus are at risk for herniation with removal of lumbar CSF and should not undergo LP. RT to the site of obstruction may be reasonable in patients who are asymptomatic and can be observed closely for signs of deterioration. However, we have a low threshold for VP shunting, as obstructive tumors may rapidly recur after an initial positive response to RT.
Ventriculoperitoneal shunting — VP shunting is a relatively safe and effective way to relieve hydrocephalus and refractory symptoms of elevated ICP in patients with LMD [25-29].
VP shunting is usually performed in the presence of frank hydrocephalus, with dilated ventricles and evidence of transependymal flow of CSF on MRI (image 5). However, in some cases, a communicating hydrocephalus from LMD can cause elevated ICP without significantly increasing the size of the ventricles. Such patients are also candidates for VP shunting to relieve refractory headaches or other symptoms of ICP as documented by an increased opening pressure (>20 cm H20). However, the surgical procedure may be more difficult because the ventricles are relatively small.
Outcomes associated with VP shunting are illustrated by a meta-analysis that identified 10 retrospective studies in a total of 420 patients undergoing shunting for LMD from solid tumors, mostly lung and breast malignancies [30]. Overall survival ranged from two to three months after shunt placement. Symptom relief was usually seen (67 to 100 percent) with higher improvement rates for headaches, nausea, and confusion than for focal neurologic deficits and cranial nerve palsies. Among 109 patients in four studies reporting pre- and postoperative Karnofsky performance status (KPS), a mean KPS improvement of 16.7 points was seen. Overall complication rates ranged from 10 to 20 percent. The most frequent complications were mechanical shunt failure (8.6 percent), overdrainage and other manifestations of functional failure (5 percent), and infection (3.8 percent). Approximately half the patients suffering a complication required an additional surgical procedure. Only one case of peritoneal metastasis after surgery was identified, and the authors attributed it to systemic cancer progression, not the VP shunting. Intestinal perforation was identified in 36 patients (8.6 percent). Separately, peritoneal spread of germinoma, sarcoma, medulloblastoma, and glioblastoma with primitive neuroectodermal features has been reported in children [31,32] but almost never occurs in adult patients [24,33].
Of note, standard VP shunt placement will divert IT therapy from intracranial tumor into the abdominal cavity. Several investigators have used programmable VP shunts or on/off valves to deliver chemotherapy, stop CSF outflow in order to expose leptomeningeal tumor to the drug, and then reopen CSF diversion 24 hours later to treat hydrocephalus [34,35]. The effectiveness of IT therapy delivered by this method has not been established and the risk of toxicity may be higher, since the injected agent remains in the ventricular system and diffuses into the surrounding brain, increasing potential toxicity, until the valve is opened. In one study, such modified shunts appeared as effective as administering chemotherapy through an intraventricular access device in the absence of hydrocephalus [35], but in another study, 17 of 70 shunted patients required shunt revision due to malfunction or infection [24].
SYSTEMIC THERAPY OPTIONS — A growing number of anticancer therapies, especially molecularly targeted drugs and immunotherapies, penetrate into the central nervous system (CNS) and have some reported utility in patients with LMD.
Systemic therapy in patients with LMD is individualized. There is no single drug with guaranteed activity across multiple different histologies. In general, we choose an agent (chemotherapy, targeted therapy, or immunotherapy) with the highest likelihood of both CNS penetration and efficacy against the primary cancer.
The options are organized below by primary cancer type, although for some therapies, such as immunotherapy and therapies targeting a common genetic mutation, patients with multiple cancer types may be eligible for the same agent. Should disease progress on one targeted agent, responses are sometimes achieved using an alternative medication targeting the same mutation.
Patients with breast cancer — Patients with LMD from breast cancer are considered similarly to those with parenchymal brain metastases in terms of systemic therapy options, where a choice of systemic therapy is guided by response to prior therapies, the hormone receptor status, human epidermal growth factor receptor 2 (HER2) status, programmed death ligand 1 (PD-L1) expression, and other susceptibility gene mutations found in the tumor.
For HER2-positive disease, several agents or combinations have sufficient CNS penetration to be considered for use in patients with LMD following radiation therapy (RT) to symptomatic sites of disease and control of hydrocephalus if present. Frequently, brain metastases are HER2-positive, even in the setting of HER2-negative extracranial disease [36]. In the DEBRAH prospective study, intracranial responses to HER2-directed therapies occurred later than systemic responses (median time to response for intracranial and extracranial disease was 3.4 versus 1.3 months, respectively) [37].
A paucity of data remains regarding chemotherapeutic options for patients with LMD despite ongoing investigations, but rationale for their use may be extrapolated from treatment of patients with active brain metastases, reviewed in detail elsewhere. Examples of demonstrably CNS active regimens include trastuzumab-emtansine (T-DM1), fam-trastuzumab deruxtecan [T-DXd], capecitabine alone, tucatinib in combination with capecitabine and trastuzumab, and capecitabine with lapatinib or neratinib. The ongoing ROSET-BM trial of T-DXd reported one-year progression-free survival of 61 percent and one-year overall survival of 87 percent among 19 patients with LMD, of whom 17 had concurrent brain metastases [38]. (See "Brain metastases in breast cancer", section on 'HER2-positive disease'.)
Absent a targeted therapy option, we consider use of either high-dose intravenous methotrexate or single-agent capecitabine.
●High-dose methotrexate – Systemic administration of high-dose intravenous methotrexate with leucovorin rescue is an alternative to intrathecal (IT) methotrexate administration in patients who cannot receive IT therapy. Although studies assessing the effectiveness of high-dose methotrexate in patients with LMD from solid tumors are mixed [39-44], clinical experience supports a beneficial role in selected patients with breast cancer LMD.
Most studies have used a dose of 3 to 3.5 g/m2 given every two weeks. In the two larger series composed primarily of patients with LMD from breast cancer, the rate of stable disease ranged from 30 to 45 percent, with no complete responses and median survival of five to six months [40,41]. More prolonged survival has been described mostly in case reports [41-43].
Although ambulatory regimens are under investigation [45], high-dose methotrexate presently requires hospitalization for hydration, urinary alkalinization, and leucovorin rescue at a center experienced with the regimen to avoid systemic methotrexate toxicity. If patients develop early renal insufficiency and plasma methotrexate levels remain high for a prolonged period despite these measures, early treatment with glucarpidase (carboxypeptidase G2 [CPDG2]) to metabolize the methotrexate is crucial to preventing severe toxicity and death. The use of high-dose methotrexate and crystal-induced renal failure in patients treated with methotrexate are discussed elsewhere. (See "Therapeutic use and toxicity of high-dose methotrexate", section on 'Alternative rescue techniques for patients with renal failure' and "Crystal-induced acute kidney injury".)
●Capecitabine – Capecitabine, an oral prodrug of fluorouracil, has induced responses and disease stabilization in some patients with LMD from breast and other solid tumors [46-50]. Some responses continued for more than 12 months [49,51].
Patients with non-small cell lung cancer — Systemic administration of genotype-directed targeted therapies can result in clinical benefit in appropriately selected patients with non-small cell lung cancer (NSCLC) with CNS disease, including LMD.
●Epidermal growth factor receptor (EGFR)-mutant NSCLC – The EGFR tyrosine kinase inhibitors have CNS activity in patients with EGFR-mutant NSCLC. Most data are in patients with brain metastases and are reviewed separately. (See "Brain metastases in non-small cell lung cancer", section on 'Patients with oncogenic drivers'.)
Available data in patients with LMD include the following:
•Osimertinib – Osimertinib is a third-generation EGFR inhibitor with significant intracranial activity. In a multicenter phase I trial of double-dose osimertinib (160 mg) in patients with LMD from EGFR-mutant lung cancer that had progressed on prior EGFR targeted therapy, 17 of 41 enrolled patients (41 percent) were alive and progression free at 12 months [7]. Recurrences were seen more often outside the CNS, and drug dose reduction, discontinuation, and interruption were observed in 5, 9, and 24 patients, respectively. Other studies have confirmed activity with osimertinib 80 mg daily, including a study of 22 EGFR T790M-positive patients with LMD in which 52 percent of patients were alive at 18 months [8,52].
•Afatinib – Afatinib is a second-generation drug that has activity in patients with LMD from EGFR-mutant NSCLC [53-56]. Multiple case reports describe responses in afatinib-naïve patents, including at least one report of a patient with LMD responding to afatinib after progression on osimertinib [56].
•Gefitinib and erlotinib – In case reports, patients with LMD from lung adenocarcinoma improved clinically and radiographically after treatment with standard-dose [57-60] or high-dose [61,62] gefitinib. Two reports observed a good response using erlotinib at a higher dose (either 600 mg daily or 1500 mg once weekly) [63,64]. Occasional responses to higher-dose erlotinib following failure with standard-dose regimens have also been described [65,66].
In a single-center study involving 22 EGFR-mutant lung adenocarcinoma patients with LMD, erlotinib-treated patients had a longer overall survival than gefitinib-treated patients (6.6 versus 2.1 months, respectively) [67]. Rechallenge with a second EGFR tyrosine kinase inhibitor led to improved survival in one study of 92 EGFR-mutant lung cancer patients with LMD after progression on the first drug; overall survival in patients rechallenged with a tyrosine kinase inhibitor was 7.6 months compared with 4.2 months without rechallenge [68].
●Anaplastic lymphoma kinase (ALK)-rearranged NSCLC – ALK inhibitors crizotinib, ceritinib, alectinib, brigatinib, and lorlatinib have activity in patients with ALK fusion oncogene-positive NSCLC. The reported prevalence of ALK rearrangements in unselected NSCLC is approximately 4 percent. Up to 40 percent of patients with ALK-positive lung cancer will have CNS spread at diagnosis.
•Lorlatinib – Lorlatinib is a third-generation ALK inhibitor with high CNS penetration and intracranial activity in treatment-naïve as well as pretreated ALK-positive NSCLC. Lorlatinib was active in two patients with LMD in one report, including one patient with a complete and ongoing response at 22 months [69].
•Second-generation drugs – Alectinib, brigatinib, and ceritinib have important clinical activity in patients with brain metastases, including those with LMD.
-In a phase I dose-escalation study of alectinib, 21 patients had CNS metastases at baseline, of whom 52 percent achieved an objective response, including 29 percent with complete responses [70]. Only two patients (10 percent) had radiographic progression.
-As with the EGFR inhibitors, switching from one ALK inhibitor to another has shown efficacy in several LMD case reports [9-11]. Dose escalation can also prove effective [12]. One case report described a durable (>15 month) clinical and radiographic response to alectinib after symptomatic LMD developed on crizotinib [71].
-In early trials, measurable reduction in brain metastases was seen with brigatinib therapy in patients previously treated with crizotinib and in ALK inhibitor-naïve patients. CNS progression-free survivals exceed 14 months on this therapy with modestly improved outcomes on 180 mg/day over 90 mg/day [72,73].
•Crizotinib – Crizotinib is a first-generation ALK inhibitor with modest CNS activity compared with later-generation drugs. In a pooled retrospective analysis of patients with asymptomatic brain metastases from ALK-rearranged NSCLC enrolled in two clinical trials of crizotinib, intracranial disease control rates were similar to rates of systemic disease control [74]. Such responses tend not to be durable, however.
Patients with melanoma — Despite significant progress in melanoma therapeutics, the prognosis of patients with LMD remains quite limited, and treatment decisions must be individualized [75,76]. In a retrospective study of 52 patients with melanoma and LMD diagnosed between 2011 and 2019, the median overall survival was 2.9 months. Most patients received either targeted therapy or immunotherapy; however, among patients who received systemic therapy, overall survival was only slightly longer (3.7 months).
In BRAF-mutant melanoma patients, case reports describe durable responses to vemurafenib and dabrafenib in several patients with LMD [77-82].
In an early case series of 39 melanoma LMD patients, treatment with targeted agents and/or immunotherapies led to longer survivals compared with untreated patients. Median overall survival was 16.4 weeks, and 3 of 25 (12 percent) treated patients were alive at one year (range 2 to 235.1 weeks) [83]. A randomized phase II trial of nivolumab and ipilimumab in 79 melanoma patients with brain metastases demonstrated intracranial responses to therapy [84].
Prolonged survival after checkpoint inhibitor therapy (pembrolizumab, ipilimumab, or nivolumab) in patients with LMD from melanoma has also been described [85,86]. Checkpoint therapy-naïve patients may have better response to these agents [87-89], which have shown efficacy in patients with primary leptomeningeal melanomatosis as well [90,91]. (See "Management of brain metastases in melanoma", section on 'Systemic therapy'.)
Other tumor types — Patients with LMD related to a less common primary tumor type, such as ovarian cancer, colon cancer, gastric cancer, renal cell carcinoma, are often heavily pretreated and have limited systemic chemotherapy options. Such patients often have limited overall survival, and RT to symptomatic areas plays a primary role in palliation.
For patients who have a good functional status and are candidates for further systemic therapy, agents are generally selected based knowledge of CNS penetration. Drugs that have reasonable CNS penetration and activity against some solid tumors include temozolomide and nitrosoureas (gliomas), topotecan (ovarian, small cell lung cancer), capecitabine (colon, pancreatic cancer), and carboplatin (ovarian, lung cancer) [92-94]. IT chemotherapy may also be an option in some of these patients. (See 'Intrathecal therapy' below.)
Role of immunotherapy — For patients who are immunotherapy naïve with a range of histologies, checkpoint inhibitors have some evidence of activity in LMD, although median survival in prospective studies of patients with LMD has been modest. Glucocorticoids are typically tapered as low as possible to maximize responsiveness to immunotherapy. In a review of case reports and retrospective series that included 61 patients with LMD from a variety of neoplasms treated with checkpoint inhibitors, 12-month survival was 32 percent, highlighting occasional successes [95].
Nivolumab induced dramatic, durable, and prolonged responses in two patients with LMD due to lung cancer and in one patient with renal cell cancer LMD, with ongoing disease-free survivals of three, four, and three years at publication [96-98]. Long-term survival has also been reported in melanoma after multimodal therapy that included immunotherapy [88]. In a single-arm phase II study of pembrolizumab in 20 patients with solid tumor LMD (17 patients with breast cancer, 2 with lung cancer, and 1 with ovarian cancer), approximately 50 percent of patients achieved stable disease and 60 percent were alive at three months [99]. The median overall survival remained limited at 3.6 months.
Combination therapy has also been studied. In a phase II study of ipilimumab plus nivolumab in 18 patients with LMD from breast cancer (n = 8) and a range of other solid tumor histologies, 39 percent had stable disease in the CNS as a best response, and there was one complete response; median overall survival was 2.9 months [100]. The rate of grade 3/4 toxicities was 33 percent, and two patients discontinued therapy due to toxicity (colitis and hepatitis).
INTRATHECAL THERAPY — Intrathecal (IT) chemotherapy has historically been a primary therapy for LMD in patients with solid tumors, although its efficacy is modest and its superiority compared with systemic treatment has not been established in randomized trials [101]. As discussed above, we generally favor systemic therapy over IT therapy for patients who have a reasonable systemic option (algorithm 1). (See 'Treatment plan and sequencing' above.)
The drugs most commonly used for IT chemotherapy in patients with solid tumor LMD are methotrexate, thiotepa, and cytarabine. Methotrexate is preferred in patients with breast cancer and is a reasonable first-line choice for IT therapy in other solid tumor types as well.
Candidates for IT therapy — The best candidates for IT therapy are those without bulky LMD, with no hydrocephalus, and with absent parenchymal brain metastases [16]. The presence of hydrocephalus, even when mild or asymptomatic, is a sign that cerebrospinal fluid (CSF) flow dynamics are abnormal, and the risk of toxicity from administration of IT chemotherapy is increased. A minority of centers use a pretreatment CSF flow study to further investigate CSF flow dynamics before treating with IT therapy. (See 'Pretreatment CSF flow study' below.)
IT chemotherapy is more likely to be effective for small leptomeningeal deposits and individual tumor cells floating in the CSF [102,103]. It cannot reliably treat bulky disease, because diffusion of drug into tumor deposits thicker than 1 to 3 mm is limited. IT chemotherapy does not treat dural-based disease or parenchymal disease.
Pretreatment CSF flow study — Prior to radiation therapy (RT) or IT chemotherapy, a CSF flow study via a radionuclide cisternogram may be helpful although not available or used in all centers. This study will identify areas of obstruction that will prevent the chemotherapy from being homogeneously distributed, potentially decreasing efficacy and increasing toxicity. As an example, the administration of intraventricular methotrexate through an Ommaya reservoir can cause leukoencephalopathy even at conventional doses if there is a block to egress of the drug from the ventricular system. (See "Overview of neurologic complications of conventional non-platinum cancer chemotherapy", section on 'Leukoencephalopathy'.)
Abnormal CSF flow is seen in up to two-thirds of patients with LMD, often without evidence of hydrocephalus or other abnormalities on conventional neuroimaging studies [104-106]. Common sites of obstructed CSF flow include the base of the brain (ventricular outlet obstruction), within the spinal canal at the thoracic level, and over the cortical convexities. For centers that obtain CSF flow studies, RT can be delivered to areas of obstruction, even though no lesion is identified by magnetic resonance imaging (MRI), as it can restore CSF flow. A flow study is then repeated after RT to determine eligibility for IT chemotherapy.
Route of administration — IT chemotherapy involves the injection of antitumor agents into the CSF, either directly into the lateral ventricle through a subcutaneous reservoir and ventricular catheter (ie, an Ommaya device) or into the lumbar thecal sac by lumbar puncture (LP). If an Ommaya is placed, it is important to check for correct placement before chemotherapy is initiated, as administration of drug into the brain parenchyma may cause focal parenchymal injury.
Whenever possible, we suggest that IT chemotherapy be administered through a subcutaneous reservoir and ventricular catheter directly into the lateral ventricle (eg, an Ommaya device), rather than by LP [16]. Ventricular administration generally results in better distribution and avoids the approximately 10 percent risk that lumbar injections do not enter the CSF space [107,108]. Repeat LP is required in patients who unable or unwilling to undergo surgical placement of a reservoir.
Ommaya access is possible at lower platelet levels, so drug may be delivered in the setting of chemotherapy-induced thrombocytopenia. An Ommaya reservoir may be accessed with a platelet count as low as 20,000/microL, although special precautions are necessary, such as gently handling the syringe and taking care not to exert too much force in withdrawing fluid. However, administering methotrexate if the platelet count is below 50,000/microL is not recommended. (See 'Methotrexate' below.)
Limited data suggest that the complication rate of intraventricular reservoirs is low in patients with LMD from solid tumors [109,110]. In a single-center prospective study of 112 such patients treated with a mean of nine intraventricular chemotherapy injections each (mostly liposomal cytarabine), the overall catheter-related complication rate was 10 percent [109]. There were seven infections; of these, three occurred within 30 days of surgery and four were likely related to access of the reservoir for chemotherapy injection. All affected patients presented with a febrile meningeal syndrome, and bacterial meningitis was confirmed in the majority; one case was fatal.
Technique of administration — Careful attention to minimizing any change in CSF volume with IT therapy is critical. Serious complications (eg, headache, nausea and vomiting, obtundation, herniation) can develop acutely if the total volume is increased even in asymptomatic patients with LMD, because they may be on the edge of their CSF ventricular compliance ("pressure-volume") curve. Thus, equivalent volumes of CSF (usually 7 to 10 mL) should be removed prior to instilling IT chemotherapy.
Conversely, decreases in CSF volume can cause patients to develop a pressure sensation (usually frontal or retroorbital) during the withdrawal of large volumes of CSF from the ventricles. This sensation disappears promptly after fluid is replaced during the chemotherapy infusion and is less common if the procedure is performed with the patient supine.
Administration of intraventricular chemotherapy follows a standard procedure (movie 1). After sterile preparation of the skin with povidone iodine, a 23-gauge butterfly needle attached to a three-way stopcock is inserted into the ventricular reservoir. Approximately 15 to 20 mL of CSF are then withdrawn gradually over two to five minutes (depending upon patient tolerance) into syringes no larger than 10 mL. Syringes should be partially opened and closed prior to their attachment to the stopcock. The first 10 mL syringe is left attached to the stopcock and is used to flush the chemotherapy through the butterfly needle and Ommaya catheter following drug administration. The remaining CSF is sent for cytology and other studies.
Once the CSF has been withdrawn, the syringe containing the drug is attached to a stopcock port. The port pointing to the ceiling is best, to avoid injecting air into the ventricle. A small amount of CSF is withdrawn into the syringe to ensure continued correct placement of the butterfly needle. The chemotherapy is then administered slowly, generally over two to three minutes. In some patients, particularly those with small ventricles or elevated intracranial pressure (ICP), a slower rate of administration is necessary to avoid headache.
CSF from the original 10 mL syringe is then drawn into the chemotherapy syringe by repositioning the stopcock and is subsequently injected through the reservoir over two to three minutes, in order to flush any drug remaining through the chemotherapy syringe or reservoir into the ventricular system in a sterile fashion. The needle is then removed, and light pressure is applied with a sterile gauze pad.
The use of IT hydrocortisone (15 to 30 mg) in conjunction with IT chemotherapy was popularized in the 1970s and thought to reduce toxicity [111]. However, there are no controlled trials proving benefit, and most investigators do not use IT glucocorticoids for patients with solid tumor LMD receiving water-soluble cytarabine, methotrexate, or other IT chemotherapy. IT hydrocortisone can itself cause a chemical meningitis and should be avoided.
Specific agents
Methotrexate — Methotrexate is the chemotherapeutic agent most commonly used for IT therapy in LMD. It has activity against breast cancer and hematologic malignancies but is less active against other solid tumors such as lung cancer and sarcoma. In the CSF, methotrexate has a half-life of 4.5 hours, declining to subtherapeutic levels within four days.
In small prospective series of IT methotrexate in patients with mostly solid tumor LMD, the cytologic response rate ranges from 20 to 61 percent, with few or no patients showing neurologic improvement or recovery of function [2,39,112]. The experience with IT methotrexate in conjunction with RT is illustrated by several retrospective series [113-115]. In one report, median survival ranged from three months in patients with breast cancer to eight months in those with lymphoma, although responding patients occasionally survived substantially longer [113]. In another series, patients who responded to treatment by the end of one month exhibited a median additional survival of six months, while those not responding survived an additional two months [114].
●Dosing – A standard induction regimen consists of a fixed dose of 10 or 12 mg twice a week for four weeks or until cytology becomes negative for malignant cells. For IT use, preservative-free methotrexate is diluted to a concentration of up to 4 mg/mL in an appropriate sterile, preservative-free medium such as 0.9 percent sodium chloride injection.
To minimize the risk of systemic methotrexate toxicity, we routinely give concurrent oral leucovorin (10 mg twice daily for three days) to all patients to minimize myelosuppression related to methotrexate leaving the CSF and entering the bloodstream before being eliminated by the kidneys. Leucovorin does not cross the blood-brain barrier and does not affect the activity of methotrexate against LMD. Thus, it may be started at the same time as the IT methotrexate.
If no cytologic response to methotrexate is seen after four weeks, therapeutic options include another four weeks of induction therapy or a trial with an alternative agent. If clinical response occurs, the frequency of administration is decreased to weekly for four to eight weeks, and then a maintenance regimen is continued with drug administration every two weeks for several months, and then monthly for two to four months. The optimal duration of therapy in responding patients is uncertain; however, treatment beyond six months may be unnecessary [112]. (See 'Response evaluation' above.)
●Side effects – Continuous exposure to low systemic concentrations of methotrexate can cause severe myelosuppression. Methotrexate is not metabolized in the CSF. It is absorbed slowly by the choroid plexus [116], released into the systemic circulation, and then excreted by the kidneys.
The risk of serious myelosuppression is increased in patients with one or more of the following:
•Renal insufficiency, resulting in delayed excretion of methotrexate
•Pleural effusion or ascites, which can accumulate methotrexate and then slowly release drug into the systemic circulation (the "third space" effect)
•Abnormal CSF flow, prolonging release of methotrexate from the CSF into the systemic circulation
Methotrexate is partially bound to serum albumin. Thus, toxicity can be increased when drugs that displace methotrexate from albumin are coadministered (eg, aspirin, phenytoin, sulfonamides, and tetracycline).
We do not administer methotrexate to patients with a platelet count ≤50,000/microL. If the platelet count is between 50,000 and 150,000/microL, we give concurrent oral leucovorin (10 mg four times daily for three days). Leucovorin does not cross the blood-brain barrier but will counteract systemic methotrexate effects on the bone marrow and gastrointestinal epithelium. (See "Therapeutic use and toxicity of high-dose methotrexate", section on 'Rationale for leucovorin rescue'.)
A variety of neurologic complications can result from IT methotrexate therapy. These include chemical (aseptic) meningitis, delayed leukoencephalopathy, acute encephalopathy, and transverse myelopathy. These issues are discussed in detail separately. (See "Overview of neurologic complications of conventional non-platinum cancer chemotherapy", section on 'Methotrexate'.)
●Accidental overdose – Massive accidental overdose with IT methotrexate must be treated vigorously since it can cause acute myelopathy, encephalopathy, and death. Patients with significant overdoses are treated with the combination of IT administration of glucarpidase (carboxypeptidase G2 [CPDG2]) to metabolize the methotrexate, and ventriculolumbar perfusion to reduce the amount of methotrexate in the CSF. The approach to such overdoses in discussed in detail elsewhere. (See "Overview of neurologic complications of conventional non-platinum cancer chemotherapy", section on 'Accidental overdose of intrathecal methotrexate'.)
Thiotepa — Thiotepa is highly lipid soluble, rapidly diffuses out of the CSF (within four hours), and has the shortest half-life of the agents used for IT chemotherapy.
The efficacy of IT thiotepa is not as well established as that of IT methotrexate. In a multicenter randomized trial of IT thiotepa versus IT methotrexate in 59 patients with LMD, complete cytologic clearance was seen in approximately one-third of patients treated with each agent, and median survivals were similar (14 versus 16 weeks) [2]. A single-center retrospective study that included 66 breast cancer patients with LMD treated with IT thiotepa every two weeks reported similar results, with a cytologic response in 16 of 36 evaluable patients and a median overall survival of 4.5 months [117]. Another retrospective study of 24 patients with breast cancer who had failed liposomal cytarabine yielded a median progression-free survival of 3.1 months and an overall survival of 4.0 months, with minimal toxicity [118].
●Dosing – The usual regimen of IT thiotepa is 10 mg IT on the same twice-weekly schedule as methotrexate, with subsequent reductions in dose in responding patients [2,119], although less frequent dosing (eg, every two weeks) has also been reported [117]. Thiotepa has been used in patients who have failed methotrexate, have methotrexate-induced leukoencephalopathy, or need concurrent brain radiation.
●Side effects – IT thiotepa tends to be well tolerated. Rare cases of myelopathy have been reported. It does have the potential to cause systemic myelosuppression [2].
Cytarabine — Conventional cytarabine is relatively ineffective in patients with LMD due to solid tumors, and its use is usually restricted to those with LMD from leukemia or lymphoma. The conventional formulation has a half-life of less than four hours in the CSF and is entirely eliminated within one to two days [120,121]. (See "Secondary central nervous system lymphoma: Clinical features and diagnosis" and "Involvement of the central nervous system (CNS) with acute myeloid leukemia (AML)".)
Liposomal cytarabine (DepoCyt), which had a longer half-life, was used in the past for LMD [3,122-124] but is no longer commercially available [125].
Trastuzumab — IT delivery of trastuzumab is an emerging option in patients with human epidermal growth factor receptor 2 (HER2)-positive cancers, particularly breast cancer. Overall, the evidence supports IT trastuzumab 80 mg twice weekly as a reasonable and well-tolerated IT therapy.
Multiple small series have demonstrated that 25 to 150 mg of trastuzumab in preservative-free saline is well tolerated [126-131]. In the largest phase I/II multicenter study of 34 patients with HER2-positive cancers, only one grade 4 toxicity (arachnoiditis) was encountered at the 80 mg twice-weekly IT trastuzumab dose level [132]. Pharmacokinetic analysis suggested a CSF half-life of 4.1 hours after a single dose of 80 mg.
The phase II portion of the study included 23 patients with HER2-positive breast cancer LMD treated with 80 mg (twice weekly for the first four weeks, then weekly for four weeks, then every one to two weeks as maintenance) [132]. Rates of partial response, stable disease, and progressive disease were 19, 50, and 30 percent, respectively. With a median follow-up of 10.5 months, median progression-free and overall survival were 2.8 and 10.5 months, respectively (measured from study enrollment). Four patients received concomitant CNS-penetrating therapies (high-dose methotrexate in one patient and lapatinib in three patients).
Investigational intrathecal therapies — The limited effectiveness of currently available agents for IT administration has led to the evaluation of several alternatives.
●Etoposide – At least three reports have evaluated IT etoposide for LMD [133-135]. Treatment was well tolerated, but there are inadequate data to draw any conclusions regarding its relative efficacy. Topotecan is occasionally given IT and may be a reasonable second-line choice for solid tumors such as lung and ovarian [136].
●Pemetrexed – A first-in-human study of IT pemetrexed plus IT dexamethasone in patients with LMD due to epidermal growth factor receptor (EGFR)-mutant non-small cell lung cancer (NSCLC) arrived at a phase II dose of 50 mg of pemetrexed given twice per week for the first week, then weekly for three weeks, then once monthly [137]. Myelosuppression was the most common side effect (37 percent), followed by nausea/vomiting and neurotoxicity (eg, limb pain, headache). Among 30 patients in the phase II study, median overall survival was nine months and clinical responses were reported in the majority of evaluable patients. These encouraging results suggest that further study is indicated.
●Immunotherapy – Administering concurrent IT and intravenous nivolumab has been evaluated in an initial clinical trial of patients with metastatic melanoma and LMD, but the approach remains investigational [138]. IT ipilimumab is also under investigation in at least one trial [139].
SUMMARY AND RECOMMENDATIONS
●Goals of therapy – Leptomeningeal disease (LMD) from solid tumors is usually a late-stage development associated with a prognosis of three to four months. Goals of treatment include stabilizing or improving neurologic function and palliating symptoms. Extended survival is occasionally achieved. (See 'Introduction' above.)
●Prognostic assessment – Patients with a Karnofsky Performance Status (KPS) ≤60 (table 2); multiple, serious, fixed neurologic deficits; untreated, symptomatic hydrocephalus; and extensive systemic cancer with limited remaining therapeutic options have a poor prognosis even with active treatment.
In the absence of these factors, the potential for sustained tumor control is most often signaled by the presence of reasonable remaining systemic treatment options, particularly ones with central nervous system (CNS) activity. (See 'Prognosis' above.)
●Urgent symptomatic interventions – For relatively good-risk patients with a cancer for which reasonable treatment options remain, treatment is directed at palliating symptomatic sites of disease with radiation therapy (RT) when possible, correcting structural problems such as hydrocephalus, and then attempting further CNS disease control with systemic or intrathecal (IT) therapy (algorithm 1) (see 'Treatment plan and sequencing' above):
•RT to symptomatic sites – RT is used whenever possible to treat symptomatic areas of LMD, since systemic or IT therapy alone rarely reverses neurologic signs and symptoms. The most commonly used forms of RT for LMD are lumbosacral spine RT (for symptomatic cauda equina disease) and whole brain or skull base RT (for cranial neuropathies, diffuse encephalopathy, and mild hydrocephalus). (See 'Involved field radiation' above.)
Craniospinal RT with photon radiation is typically not appropriate in patients with LMD from solid tumors due to high toxicity and short life expectancy. However, where available, proton craniospinal RT, which has an improved toxicity profile compared with photon craniospinal RT, is an option in selected patients with better prognosis and adequate functional status. (See 'Proton craniospinal radiation' above.)
•Severe hydrocephalus – Most patients with severe hydrocephalus from LMD require placement of a ventriculoperitoneal (VP) shunt before attempting further CNS-directed therapy. Patients with a poor prognosis are usually managed supportively without shunting, unless symptoms cannot be controlled with steroids, pain medications, and antiemetics. (See 'Shunting versus radiation therapy' above and 'Ventriculoperitoneal shunting' above.)
•Less severe hydrocephalus – Management of less severe forms of hydrocephalus and increased intracranial pressure (ICP) is individualized. Symptomatic glucocorticoids, shunting, and RT all play a role. The appropriate therapy depends on the severity of hydrocephalus and the type (communicating or noncommunicating), as described above. (See 'Glucocorticoids' above and 'Shunting versus radiation therapy' above.)
●Further CNS-directed therapy – Although local therapies are often indicated for palliation, LMD is a disseminated process, and further systemic or IT anticancer therapy with the ability to reach the full neuroaxis must be given if disease is to be controlled.
•Systemic therapy – For most patients with a systemic therapy option that has both efficacy against the primary tumor and a reasonable likelihood of CNS penetration or activity, we suggest systemic therapy rather than IT therapy (Grade 2C).
There is no single drug with guaranteed activity across different histologies in patients with LMD. In general, we choose an agent (chemotherapy, targeted therapy, or immunotherapy) with the highest likelihood of both CNS penetration and efficacy against the primary cancer. (See 'Systemic therapy options' above.)
•Intrathecal therapy – For patients who do not have a good systemic therapy option but are nonetheless candidates for further CNS-directed therapy, we suggest IT chemotherapy via an Ommaya reservoir (Grade 2C). The best candidates for IT therapy are those without bulky LMD, with no hydrocephalus, and with no parenchymal brain metastases. (See 'Intrathecal therapy' above.)
For most patients selected for IT therapy, we suggest IT methotrexate rather than other agents (Grade 2C). IT trastuzumab is a reasonable alternative in patients with HER2-positive cancers. When placement of an Ommaya reservoir is difficult or impossible, systemic chemotherapy with high-dose methotrexate is an alternative to IT therapy in patients with breast cancer. (See 'Methotrexate' above and 'Trastuzumab' above.)
●Supportive care – Intensive supportive care is important and often essential to achieve acceptable quality of life in patients with LMD, regardless of goals of care. (See 'Supportive care needs' above.)
For patients with a very poor prognosis who are not candidates for further tumor-directed therapy, glucocorticoids, pain medications, and antiseizure medications are used to control symptoms. RT directed to symptomatic sites of disease and even VP shunting may be necessary in selected patients for optimal symptom control. (See 'Radiation to symptomatic sites' above.)
آیا می خواهید مدیلیب را به صفحه اصلی خود اضافه کنید؟
