Leptomeningeal Disease Is Becoming More Common, and Also More Treatable

Key Ideas

Leptomeningeal disease (LMD) is becoming more common simply because patients with actionable mutations are living longer, but survival for these same patients is also improving, driven by better systemic (not local) therapy. CSF-based molecular testing, including circulating tumor cells and ctDNA, is now formally part of NCCN guidelines and is changing both diagnosis and monitoring, and multimodality treatment, particularly targeted therapy dose escalation and intrathecal-systemic combinations, is where the most promising outcomes are emerging.

Assessing a Patient With Suspected LMD

A practical framework for evaluating a patient with suspected LMD starts with a few core questions: how symptomatic is the patient (do they need urgent CSF diversion, high-volume shunting, or urgent radiation for hydrocephalus); what prior therapies have they received, or are they newly diagnosed; what is their systemic disease status; and, critically, has the tumor ever been comprehensively sequenced, and at what point in the disease course.

What LMD Actually Is, and Why Incidence Is Rising

LMD occurs when tumor cells spread into the CSF, which circulates throughout the entire CNS axis, producing headaches, nausea, cranial nerve symptoms, or even focal neurologic symptoms that can mimic parenchymal brain metastases. Incidence is roughly 5-11% in NSCLC and rising, specifically because patients with actionable mutations are living longer on effective systemic therapy that delays, but doesn't prevent, LMD. The encouraging counterpart: overall survival after an LMD diagnosis is also improving, from a historical range of weeks to months, now considerably longer, and what actually drives that improvement in retrospective datasets is systemic targeted therapy, not chemotherapy and not whole-brain radiation.

Risk varies meaningfully by driver mutation: LMD incidence is about 3.6% in patients without an actionable driver, versus roughly 11% with EGFR, ALK, or HER2, and even higher with ROS1. Notably, even within EGFR-mutant disease, risk isn't uniform: one recent publication found a 24-month LMD incidence of 15% with EGFR L858R and atypical EGFR mutations treated with osimertinib, versus 5.7% with exon 19 deletions, a difference not seen in older, pre-osimertinib data, reinforcing how much modern systemic therapy has changed this disease's natural history.

Diagnosis Remains Genuinely Imperfect

CSF cytology, the diagnostic gold standard, has poor sensitivity (roughly 50%), and MRI brain interpretation shows substantial inter-observer disagreement, one radiologist calls a finding normal, another calls the same image LMD-positive. Advanced CSF testing (circulating tumor cells and CSF tumor DNA) is now incorporated into NCCN guidelines and adds real diagnostic value alongside clinical exam findings. LMD itself is generally divided into two types: type 1 (cytology-positive, worse prognosis) and type 2 (radiographically or clinically positive, cytology-negative, generally better prognosis). Response assessment is similarly imperfect: RECIST doesn't capture the true picture of LMD response, which led to the RANO working group's LMD-specific criteria (combining cytology, imaging, and exam findings), though even this remains imperfect; a subsequent smaller retrospective study found similar overall survival between imaging-defined stable disease and partial response, suggesting the distinction may matter less than assumed.

EGFR-Mutant LMD: Where Most of the Prospective Data Exists

Most prospective (still small) LMD studies have been done in EGFR-mutant disease. BLOOM (double-dose osimertinib, 160 mg, in patients previously treated with first- or second-generation EGFR TKIs, though about half with ECOG performance status 2) achieved median OS of 11 months. BLOSSOM (standard-dose osimertinib, similar prior-TKI population, better baseline performance status) achieved median OS of 15 months. A phase 2 study of amivantamab plus lazertinib (in patients previously treated primarily with osimertinib, a third-generation TKI, mostly without prior radiation) achieved median OS of 14.4 months, with a 67% reduction in ctDNA. Current NCCN CNS guidelines list osimertinib (standard or double dose) and amivantamab plus lazertinib as category 2A recommendations for LMD. In FLAURA2, 18 patients with LMD were included (13 on the osimertinib-chemotherapy arm, 5 on osimertinib monotherapy); no formal LMD-specific outcome measures were collected, but 8 of the 13 patients on the combination arm were alive and progression-free at data cutoff.

Beyond EGFR

Third-generation ALK TKIs are directly associated with both delayed LMD development and improved overall survival; a large retrospective dataset found that switching from a second- to a third-generation ALK TKI significantly prolonged survival. For other drivers, evidence is limited mostly to case reports: tepotinib is listed in NCCN CNS guidelines for MET exon 14 skipping LMD, and in the KRYSTAL trial, one patient with focal LMD treated with adagrasib. For non-driver disease, immunotherapy is associated with improved overall survival based on smaller prospective and retrospective studies; in one 32-patient cohort, single-agent versus combination immunotherapy wasn't directly associated with a survival difference, though better baseline performance status was. Intrathecal chemotherapy alone has never been directly correlated with improved OS in LMD generally, though it showed a reasonably good median OS (about 12 months) specifically in a heavily pretreated EGFR-mutant population studied in phase 1/2 trials.

Radiation and Emerging Combination Approaches

Whole-brain radiation has never been associated with improved LMD-specific outcomes. Proton craniospinal irradiation (CSI), by contrast, improved overall survival from 4.9 to 11 months in a phase 2 trial of primarily lung and breast cancer LMD, with a manageable safety profile, and is now one of the standard of care options reflected in NCCN CNS guidelines. An emerging approach combines systemic therapy with intrathecal therapy: small prospective studies have tested intrathecal immunotherapy combined with intrathecal pemetrexed (achieving median OS around 11 months in one such study), and a separate intrathecal chemo-immunotherapy combination was presented at this year's ASCO meeting, showing safety and activity even in a heavily pretreated LMD population.

To see the visualized data points discussed in this article, you can watch the lecture presentation from the 2026 Brooklyn Lung Cancer Conference here: https://youtu.be/YPc0i_yMHaQ?si=zigmHR3YNHSvC1fs.

For Patients

Leptomeningeal disease, cancer cells spreading into the fluid surrounding the brain and spinal cord, is becoming more common simply because patients with certain genetic mutations are living longer on effective treatment, but survival after this diagnosis is also improving for the same reason: better targeted drugs, not chemotherapy, are what's actually extending life. A specific pill dose, or a combination of a targeted pill with another targeted drug, can meaningfully help depending on which genetic mutation is driving the cancer. Newer testing methods on spinal fluid (looking for actual tumor cells and tumor DNA, not just traditional cytology) are improving diagnosis and monitoring. Ask your care team whether comprehensive genetic testing has been done, and whether newer CSF-based testing might help monitor your specific situation.

Key Takeaways

- LMD incidence is rising in NSCLC because patients with actionable mutations live longer on effective systemic therapy, which delays but doesn't prevent LMD; overall survival after an LMD diagnosis is improving for the same reason.

- CSF cytology has poor sensitivity (about 50%); circulating tumor cells and CSF tumor DNA testing are now part of NCCN guidelines and add meaningful diagnostic and monitoring value.

- In EGFR-mutant LMD, osimertinib (standard or double dose, per BLOOM and BLOSSOM) and amivantamab plus lazertinib are NCCN-recommended options, each achieving median overall survival in the 11-15 month range in small prospective studies.

- Proton craniospinal irradiation improved overall survival from 4.9 to 11 months in a phase 2 trial and is now one of standard of care options; whole-brain radiation has not shown LMD-specific overall survival benefit.

- Combining systemic therapy with intrathecal treatment is an emerging, actively studied approach showing early promise even in heavily pretreated patients.

References

1. Yang JCH, et al. Osimertinib 160 mg in EGFR-mutated NSCLC with leptomeningeal metastases (BLOOM).

2. Ahn MJ, et al. Osimertinib in EGFR-mutated NSCLC with leptomeningeal metastases (BLOSSOM).

3. Amivantamab plus lazertinib in EGFR-mutated NSCLC with leptomeningeal disease, phase 2.

4. Planchard D, et al. Osimertinib with or without chemotherapy in EGFR-mutated NSCLC (FLAURA2).

5. Le X, et al. Proton craniospinal irradiation for leptomeningeal metastases from solid tumors, phase 2 trial.