Table 1.

Key Points From the International Leptomeningeal Metastasis Collaborative

Category	Key points
Biology	• Translational research of the mechanisms of LM metastatic spread, survival within the nutrient-spare CSF, local immune interactions, and therapeutic resistance lag behind that of metastases to the brain and extracranial sites, and should be prioritized. • Preclinical models of LM from solid tumor primaries can be created through iterative in vivo selection of cell lines to create a subpopulation of leptomeningeal-homing cells, which when injected into allograft and xenograft models recapitulate human disease. • Standardization of CSF collection and processing guidelines across institutions to preserve cellular and soluble components would aid in research discovery and streamline CSF banking procedures. • Warm autopsy programs should be encouraged as a means of studying pial-adherent leptomeningeal cells, which are not represented in routine CSF collections.
Epidemiology	• The incidence of LM is challenging to quantify due to its relative rarity, limitations associated with disease diagnosis, and exclusion of LM from most modern population-based epidemiologic studies. • The prognosis of LM is poor but heterogeneous, with certain subgroups of patients (eg, EGFR-mutant NSCLC, HER2-positive breast cancer) demonstrating superior survival. The historical survival benchmark of 2–4 months does not apply to all patients, and patient counseling and clinical trial design should be individualized based on cancer- and patient-specific factors. • Creation of a multinational registry of patients with LM, with inclusion of a de-identified biorepository, would provide a means of analyzing LM biology and prognosis on a grander scale and could serve as a sample population for synthetic controls in clinical trial design.
Radiology	• As a disease state encompassing the entire neuroaxis, patients with LM should undergo high quality MRI brain and total spine with and without contrast at the time of diagnosis and for routine disease monitoring. • Standardization of imaging practices in patients with LM is lacking. Technical imaging recommendations to enhance LM monitoring include the use of 1.5 and 3 Tesla MRI scanners only, consistency of MRI scanner at diagnosis and surveillance time points, inclusion of 3D T1 post-contrast images with isotropic 1 mm voxels, and a reformatted slice thickness of 3 mm to enhance detection of small deposits. • CSF sampling should be performed in all patients with suspicious leptomeningeal enhancement to confirm the diagnosis of LM and rule out radiographic mimics.
Liquid biopsy	• CSF cytology remains the gold standard for LM diagnosis and should be collected in all patients, both at diagnosis and to monitor disease progression at the discretion of the treating clinician. • Due to the low sensitivity of CSF cytology, a negative result in a patient with a strong suspicion of LM should prompt a second CSF puncture with optimal collection and processing procedures (ie, minimum cytology volume of 5–10 mL, laboratory processing within 30 minutes, sampling adjacent to regions of abnormal enhancement when feasible). • CSF biomarkers with improved sensitivity and potential to guide therapeutic decisions, including rare cell capture techniques and circulating tumor DNA, are available at a limited number of tertiary cancer centers. These techniques require validation and CLIA certification before they can be routinely incorporated into clinical practice and made more widely available.
Radiation therapy	• LM is a disseminated neuraxial process. Consequently, conventional IFRT (WBRT or focal cranial or spinal RT) are palliative interventions, and have not been proven to improve survival in LM. • Proton CSI has demonstrated reasonable toxicity and superior survival compared to IFRT in adults with LM in phase I/II studies, and so may be a therapeutic option with careful consideration of a patient’s performance status, extracranial disease, and goals of care. Timely access to proton centers is limited and hinders the widespread applicability of this treatment option at present.
Systemic therapy	• Systemic therapies with CNS bioactivity and blood-CSF barrier permeability should be prioritized in all patients with LM, both to treat active disease and prevent leptomeningeal reseeding following local therapies. • Investigations of systemic therapies in LM should include CSF pharmacokinetic and pharmacodynamic analyses to better understand blood-CSF penetration and local bioactivity, particularly for modern agents with demonstrated CNS efficacy (eg, small molecule inhibitors, immunotherapies, antibody-drug conjugates).
Intrathecal therapy	• Intrathecal chemotherapies are most effective in patients with thin linear LM deposits and unobstructed CSF flow. • Untargeted intrathecal therapies (eg, methotrexate, cytarabine, thiotepa, topotecan) conferred a median survival of 2–4 months in historical clinical trials. More recently investigated intrathecal therapies (eg, trastuzumab, nivolumab, pemetrexed) have reasonable safety profiles and a median survival of 4–9 months in more modern phase I/II trials. • Intrathecal drug development would benefit from consistency and standardization of preclinical testing, drug preparation, and conversion to human dosing as new therapeutics are investigated.
Surgical interventions	• Ventricular access devices (eg, Ommaya reservoir) are preferred over lumbar drug delivery in patients receiving intrathecal chemotherapy, due to ease of administration, enhanced drug circulation, and association with superior survival. Early device placement may also be considered in those requiring frequent CSF sampling or in clinical research settings for improvement in pharmacokinetic monitoring. • Modern investigations of ventricular access devices suggest lower complication rates (2%–10%) than previously reported (10%–15%). • CSF diversion devices (eg, ventriculoperitoneal shunts) relieve symptoms of elevated intracranial pressure in the majority of treated patients, and should be offered as a palliative procedure based on a patient’s goals of care and availability of further tumor-directed therapy. • Clinical trials of intrathecal therapies should refrain from enrollment of patients with CSF diversion devices as these risk interference with pharmacokinetic endpoints.
Clinical trial design	• Clinical trials in LM are both feasible and necessary. • Standardization of consistency in LM diagnostic criteria for enrollment, endpoint selection, and response assessments should be a priority. • Overall survival remains the most objective outcome measurement in LM clinical trials due to the lack of validation of LM response assessments. A revised RANO MRI scorecard has been validated with the moderate interobserver agreement; however, validation of other RANO proposed response criteria (ie, cytologic response, neurologic examination) remains pending. • When using progression-free survival as a clinical endpoint in trials of systemic therapies with CNS bioactivity, data should be collected regarding the site of progression (extracranial versus intracranial) and whether intracranial progression was due to parenchymal or leptomeningeal metastases. • Phase I trials in LM should consider broad eligibility in cancer subtypes when the primary objective is safety. Phase II/III trials in LM would benefit from stricter criteria, including restriction to a single cancer subtype and complete neuraxial surveillance at enrollment and follow-up, to best evaluate efficacy. • Clinical protocols evaluating drug efficacy in the CNS should be designed with dedicated brain and leptomeningeal metastasis cohorts and include CSF pharmacokinetic analyses. • Multicenter participation in clinical trial execution is encouraged to foster collaboration, improve patient access, and prevent early study closure due to slow enrollment.
Novel therapeutics	• Advocacy efforts should continue to encourage LM patient inclusion in industry-sponsored clinical trials of emerging systemic therapies with potential CNS activity. • Active areas of translational research in LM to be considered include: approaches targeting LM metabolism, therapeutics which enhance LM interactions with the local immune microenvironment, intraventricularly delivered radioactive particles, and reformulated or conjugated drugs which enhance blood-CSF barrier penetration.

Abbreviations: LM, leptomeningeal metastases; CSF, cerebrospinal fluid; EGFR, epidermal growth factor receptor; NSCLC, non-small cell lung cancer; HER2, human epidermal growth factor receptor 2; MRI, magnetic resonance imaging; IFRT, involved field radiation therapy; WBRT, whole-brain radiation therapy; RT, radiation therapy; CSI, craniospinal irradiation; CNS, central nervous system; RANO, Response Assessment in Neuro-Oncology.