Leptomeningeal disease (LMD), also known as neoplastic meningitis, leptomeningeal carcinomatosis, or carcinomatous meningitis, is a rare cancer complication in which malignant cells infiltrate the layers of the central nervous system (CNS), known as meninges, and lead to significant morbidity and mortality. This disorder was first diagnosed in 1870, by Eberth et al. and was noted to be a rare complication of malignancy as it was uncommonly diagnosed before death (1). Today, it is known that this condition occurs in ~5% of all cancer patients, presenting most commonly in primary diagnoses of breast cancer (41%), lung cancer (24%), gastrointestinal tract malignancies (13%), and melanoma (12%) (2–4). However, incidence continues to rise with continuous improvements in systemic therapies that allow patients diagnosed with cancer to live longer, as well as improvements in both technology and availability of imaging of the neuroaxis leading to continued increases in the detection and diagnosis of LMD (3).

The anatomy of the neuroaxis consists of the brain and spinal cord, covered by the meninges, which are comprised of dura mater, arachnoid membrane, and pia mater. The leptomeninges refers to the two most inner layers, arachnoid membrane and pia matter, including the subarachnoid space, which separates these two sheets, and is the location of the cerebrospinal fluid (CSF). The CSF is the location of circulating tumor cells in patients with LMD. The pathogenesis of LMD is multifaceted, and can include direct extension from pre-existing CNS tumors or systemic tumors that follow peripheral nerves into the subarachnoid space, as well as infiltration through hematogenous dissemination, or even seeding of the subarachnoid space during surgical procedures (3, 5–8). Once malignant cells gain access to the CSF, they can spread along the meningeal surface by direct extension and be carried by the CSF flow and form new metastatic deposits in other locations.

The clinical presentation of LMD can be widely variable with signs and symptoms of increased intracranial pressure, cranial nerve palsies, radiculopathies, or other focal neurologic deficits, seizures, changes in brain metabolism, or encephalopathy (9–12). A subset of patients present asymptomatic with an incidental imaging finding. Commonly, patients may report a headache or worsening back or radicular pain, as well as signs and symptoms that suggest involvement of multiple anatomic neurologic sites, including: incontinence, lower motor neuron weakness, and sensory abnormalities. In addition to clinical suspicion and signs of LMD on imaging, the most informative diagnostic study is evaluation of the CSF from a lumbar puncture. The opening pressure, cytology, cell count, and measurements of protein and glucose should be performed (13). However, normal lumbar puncture (LP) studies have been reported in 5% of confirmed LMD cases, and, therefore, the requirement of a positive LP is not necessary for diagnosis (14, 15).

Standard treatments for LMD include neuroaxis directed therapies in addition to optimal systemic therapy for the primary and extra-CNS disease. These treatments may include intrathecal chemotherapy (ITC) to target microscopic tumor on the surfaces of the leptomeninges and cells floating in the CSF to prevent further seeding. Radiotherapy (RT) is also commonly used either focally to treat symptomatic sites and areas of bulky disease that will be unlikely to be adequately treated with chemotherapy or in some settings to treat the entire neuroaxis.

Regardless of systemic and local control of the primary disease, prognosis in the setting of LMD is very poor, with reported average survivals of ~2–4 months, even with treatment (16–19). There are currently no standardized guidelines for the treatment of LMD, and many clinicians are hesitant to treat patients with clinically advanced neoplastic meningitis given the high risk of toxicity and unknown value of treatment. Furthermore, in the modern era of immunotherapy, it is unclear if the incidence or prevalence of LMD is increasing.

Therefore, the objective of this systematic review was to examine the outcomes of the use of RT to treat LMD from any primary cancer histology. We reviewed clinical trials, retrospective reviews, and case series with ≥2 subjects that utilized radiation therapy techniques in the treatment of LMD, and report our findings below.

The present systematic review was designed to evaluate the published evidence underlying the use of RT treatment for LMD.

The Population Intervention Comparator Outcome Study (PICOS) design framework was used to structure the research question for the review and the corresponding published data search. The population of interest included patients of any age with a diagnosis of LMD who had RT as a part of their treatment regimen once LMD was diagnosed. No restrictions in terms of age, comorbidities, previous treatment received, or method of diagnosis were used. Patients must have received RT as a part of their LMD treatment, but were allowed to receive any other treatment in addition to RT. All doses, preparations, and frequencies of administration were considered. The primary outcome of interest was overall survival (OS). All study types were permitted, including randomized and non-randomized interventional studies, observational studies, and case series with ≥2 human subjects. Only studies with full articles published in English, or with an English translation were considered.

The report documenting our systemic review as prepared in consultation with the Preferred Reporting Items for Systemic Review and Meta-Analysis (PRISMA) statement for systematic review and meta-analyses (20).

An information specialist (G.P.) designed and executed an electronic data search to seek relevant citations for the present systematic review from Medline (1946-present), EMBASE (1947-present), and Cochrane (all years) databases. The last search was completed on 7/30/2018. The following search terms to search all trials registers and databases were used: randomized controlled trial (RCT), controlled clinical trial, leptomeningeal carcinomatosis, leptomeningeal metastasis, neoplastic meningitis, meningeal metastasis, meningeal carcinomatosis, radiation therapy, radiotherapy, radiation. A full summary of the search strategies used is provided in Supplementary Appendices 1–3.

Stage 1 screening consisted of an independent review of the titles and abstracts by two reviewers (S.B., and C.C.). Discrepancies between studies identified as potentially eligible were reconciled between the two reviewers by consensus. Stage 2 screening consisted of an independent, full-text review of all potentially relevant articles by two reviewers. Data collection of the final articles was performed by reviewer S.B. using a standardized data extraction template. The following information was collected: characteristics of trial participants that had LMD, the trials inclusion/exclusion criteria, study type (retrospective, prospective, randomized controlled trial), number of participants treated with RT for LMD, primary cancer histology, type of intervention, survival results if reported, length of follow up, and study conclusion.

Due to limitations in the available data, the number of prospective series and randomized controlled trials, and variation in treatments and outcome reporting, it was not possible to perform a detailed meta-analysis. A descriptive approach to summarizing the data was therefore used.

A consort diagram of the study selection process is shown in Figure 1. The search of Medline, EMBASE, and Cochrane provided a total of 716 citations; 18 additional citations were identified through other sources including Google and Pubmed. After adjusting for duplicates, 547 citations were reviewed, of which 400 studies were discarded for the following reasons: 253 studies did not include the words or variations of the words: “radiation” or “leptomeningeal disease” in the title or abstract, and thus they were deemed unlikely to be the focus of the publications, 2 citations had no traceable abstract or full text available, and 20 manuscripts reported on only 1 patient with LMD treated with RT. Of the remaining citations, 115 citations did not evaluate the use of RT treatment for LMD and 10 citations excluded patients with LMD and therefore, these 125 citations were excluded. An additional 147 full text articles were subsequently reviewed for inclusion. Of these, 85 were excluded for the following reasons: 12 citations did not have a full text available in English or an English translation; four full texts were found to have a single patient treated with RT for their LMD, thus not meeting inclusion criteria; 16 citations did not evaluate the use of RT to treat LMD; four citations excluded patients with LMD from their study; 43 manuscripts did not report the outcomes of patients with LMD treated with RT; and finally, six studies did not include any details of the RT treatment, and were excluded. Therefore, a total of 62 studies met inclusion criteria and were included in the systematic review.

The included studies were first categorized according to the primary histology included, and then by the type of study (randomized, observational [retrospective review, prospective review, and case series], guidelines, NCDB analyses, and practice patterns). The results of the following histologies are detailed, below: mixed, breast cancer, non-small cell lung cancer, gastrointestinal (GI) cancer, adult central nervous system (CNS) gliomas, melanoma, leukemia/lymphoma, gynecologic cancer, esthesioneuroblastoma, pediatric CNS disease, and pediatric rhabdomyosarcoma.

Due to the small number of randomized trials found, all trials that met eligibility criteria were included.

Leptomeningeal disease is an important, but devastating complication of cancer that leads to significant morbidity and mortality. Since its initial discovery in the 1800s, the incidence of this condition has continued to increase as our methods of diagnosis continue to improve and patients with various cancer histologies are experiencing longer survival, even with metastatic disease. Based on the single histology studies included in this systematic review, the incidence of LMD treated with RT from most common to least common was: lung (n = 893) > breast (n = 799) > melanoma (n = 140) > GI (n = 127) > adult CNS gliomas (n = 106), which is similar to the reported frequencies of the incidence of LMD in all patients, in the literature (2–4). The studies included in this review included patients treated from 1944 to 2015, and all of these studies were included given the small number of studies meeting inclusion criteria in the more modern era. The imaging (or lack thereof) and treatment techniques utilized in the older studies do not reflect how we think about LMD today. Therefore, we have separated these studies into those that included a majority of patients treated with RT prior to the year 2000 and those more modern studies that incorporated a majority of patients treated with RT in the year 2000 and after when evaluating the incidence of LMD treated with RT and the OS seen in these studies, as these outcomes have evolved over time. Prior to the year 2000, the incidence of LMD treated with RT is as follows: melanoma (n = 126) > breast (n = 75) > GI (n = 45) > lung (n = 32) > adult CNS gliomas (n = 16). After the year 2000, the incidence of LMD treated with RT is as follows: lung (n = 861) > breast (n = 724) > adult CNS gliomas (n = 90) > GI (n = 82) > melanoma (n = 14). The median overall survival ranges reported in these studies have also evolved over time, and are demonstrated in Figure 2. Prior to the year 2000, OS in patients with LMD treated with RT increased from shortest to longest in primary histologies of GI (1–1.75 months) < melanoma (2.3–4 months) < lung (5 months) < breast (5.4–7.5 months) < adult CNS gliomas (2.8–7.6 months). After the year 2000, OS increased from shortest to longest in GI (1.2–3 months) < breast (3.8–5.4 months) < melanoma (5.2 months) < lung (3–8.7 months) < adult CNS gliomas (3.5–10.2 months).

The optimal management of this condition remains unknown, with limited data to guide standard of care treatment. Current management approaches include a range of treatments including systemic therapy (intravenous or intrathecal), targeted therapy, focal or non-focal radiotherapy, and supportive care. A recent survey study of oncologists across Europe evaluated their diagnosis and treatment patterns for patients with LMD from solid tumors. A total of 115 physicians submitted responses (19% radiation oncologists, 23% medical oncologists, 34% neuro-oncologists, and 10% neurosurgeons) reporting that only 31.5% always administer systemic treatment when feasible, 15.5% felt that WBRT should always be performed, while 73% of medical oncologists, 56% of neuro-oncologists, 50% of radiation oncologists, and 39% of neurosurgeons felt that WBRT should be offered in the setting multifocal nodular disease only. A total of 73% of respondents declared that focal RT should only be performed in cases of neurologic symptoms only when they could be linked to an MRI abnormality and not for neurological signs or symptoms alone (88). This wide range in practice patterns highlights the need for more formal guidelines and recommendations for the treatment of LMD.

Recent developments, to assist in the standardization of diagnosis of LMD, were performed by an international panel of experts from the United States and Europe, termed the RANO group. This group recently published their results of using an MRI scorecard to assess response to treatment in patients with leptomeningeal disease. Unfortunately, after MRIs of 22 patients with LMD were scored by neuro-oncologists and neuroradiologists, many raters experienced difficulty with the instructions of the scorecard and no acceptable alpha concordance coefficient was obtained. The authors are currently working on a more simplified scorecard that will require validation, but will be a great step forward in the standardization of this disease process, hopefully leading to more comparable trials in the future (89).

The National Comprehensive Cancer Network (NCCN) has recently incorporated risk stratification and guidelines for the treatment of LMD (90). Patients with LMD can be defined as Good Risk if they have all of the following characteristics: Karnofsky Performance Status (KPS) ≥60, no major neurologic deficits, minimal systemic disease, and if they have a reasonable systemic treatment options should they need them. Patients in this category are typically diagnosed with LMD based on routine imaging and are being evaluated on an outpatient basis. Ultimately, the level of aggressiveness of the treatment provided should include a discussion with the patient and their wishes. RT is typically included in the multimodal treatment of patients with Good Risk LMD. In contrast to Good Risk LMD, patients with LMD can be defined as Poor Risk if they have any of the following characteristics: KPS <60, multiple, serious, or major neurologic deficits, extensive systemic disease with few treatment options, bulky central nervous system (CNS) disease, or encephalopathy. Some exceptions to these symptoms, where a patient may still be considered for a more aggressive treatment course, include patients with exceptionally chemosensitive tumors (e.g., small cell lung cancer, lymphoma).

Traditionally, radiotherapy has been recommended for patients with symptomatic disease or bulky metastatic disease, regardless of clinical symptoms (4, 13, 15, 89–91). To date, there is no level I evidence available on the use of RT for LMD, as the only randomized controlled trial included in this systematic review sought to evaluate the use of ITC in patients with LMD from breast cancer (39). However, a large majority (69%) of the observational studies found an improvement or likely improvement in OS with the addition of RT to the treatment regimen of LMD of those studies that commented on the influence of a single or combined treatment regimen on survival (n = 18/26 studies).

To the best of our knowledge, this is the first systemic review that has attempted to assimilate the reported data for the use of radiotherapy in the treatment of LMD across a range histologies. In our own clinical experience, radiotherapy for the treatment of LMD is consistently considered and discussed in the multidisciplinary setting. The treatment of RT is often recommended for patients with bulky disease seen on imaging or causing symptoms to prevent further neurologic compromise, or for asymptomatic patients with LMD seen on MRI who have well-controlled extracranial disease.

The results of our review indicate that for the majority of studies, any treatment was felt to be superior to best supportive care in the management of LMD, but this may reflect a selection bias of patients with better prognosis and performance status to receive treatment. The use of RT monotherapy was found to be beneficial in patients with LMD from pediatric CNS gliomas, leukemia/lymphoma, gynecologic, esthesioneuroblastoma, and pediatric rhabdomyosarcoma histologies. For patients with LMD from breast cancer, melanoma and adult CNS cancer, multimodality therapy, including the use of radiotherapy, was found to be most beneficial. In terms of prognostic factors, MVA across the included series in this systematic review consistently revealed that a good performance status at the time of diagnosis of LMD and normal CSF flow predicted for improved survival. In patients with breast cancer, hormone receptor positive disease, limited prior systemic therapy exposure, and LMD as the first site of metastases additionally predicted for improved survival.

Due to the vast clinical and methodologic differences between the included studies, a meta-analysis could not be performed to identify the optimal treatments for patients with LMD. Most studies only reported whether RT was used or not, and what type of RT was used (WBRT, focal brain/spine RT, etc.) with a complete lack of details about RT dose, fields, or timing. Additionally, the role of targeted therapies and immunotherapy has greatly increased in the treatment of LMD in the past decade (91). For example, for NSCLC with LMD, current therapies target epidermal growth factor receptor (EGFR) mutations and anaplastic lymphoma kinase (ALK) rearrangement via tyrosine kinase inhibitors (TKIs); in breast cancer, HER2 positive disease may be targeted via an intrathecal injection; and finally, melanoma with BRAF V600E mutations may also be treated with targeted therapies. Given that the included studies only included patients treated through 2015, the role and impact of targeted therapies is not reflected in this systematic review. It will be imperative to re-evaluate the use of RT, paying special attention to the dose of RT, RT fields, and timing of RT, in conjunction with these targeted and immunotherapies, in future studies.

Additionally, new classifications of LMD are being reported that may influence survival outcomes and help aid in the most optimal therapy selection. A recent report by Prabhu et al. categorized the LMD pattern of their postoperative stereotactic radiosurgery cohort as nodular or the more classical “sugarcoating” pattern. (92) The authors found that nodular LMD was less likely to be symptomatic and had better OS outcomes. Furthermore, when the patients with nodular LMD were treated with focal RT, as opposed to WBRT, they had a higher risk of second LMD recurrence, but no detriment in OS, providing a reasonable, less toxic, treatment strategy for these patients. Unfortunately, LMD was not subclassified in the articles included in this review. Future studies should consider further subclassification of this heterogeneous disease.

In conducting this systematic review, it became very apparent that there is a paucity of prospective, interventional studies in the treatment of LMD, particularly for the use of radiotherapy for the treatment of LMD. The single RCT that met inclusion criteria was a randomized trial evaluating intraventricular chemotherapy for LMD that also reported that the use of RT with systemic chemotherapy was feasible in this population (39).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.