Pathophysiology and treatment of leptomeningeal metastases in lung cancer.

Introduction
Cerebrospinal cells are segregated from rest of the body by the blood-brain barrier (BBB), blood-cerebrospinal fluid (CSF) barrier, blood-retinal barrier, and blood-spinal barrier (1,2). The BBB contributes to central nervous system (CNS) homeostasis by preventing toxic substances from entering the CNS (3). Meanwhile, leptomeningeal metastases (LMs) occur due to the entry of cancer cells into the subarachnoid space, either through hematogenous dissemination via the choroid plexus, retrograde venous extension via Batson’s plexus, or direct seeding from brain and dural metastases, resulting in significant neurological impairment (4). At the time of diagnosis, patients with LMs undergo high quality magnetic resonance imaging (MRI) of the brain and spine, while CSF sampling is collected from patients with leptomeningeal enhancement to confirm the presence of LMs. CSF cytology is the gold standard for diagnosing LMs; however, it has a low sensitivity. Advances in the understanding of LM pathophysiology and novel treatment options are gradually shifting the way LMs are treated and managed. The European Association of Neuro-oncology-European Society for Medical Oncology (EANO-ESMO) guideline (5) provides expert opinion-based recommendations for the diagnosis and the management of LMs from solid tumors, and Society for Neuro-Oncology (SNO) and American Society of Clinical Oncology (ASCO) have reported the international professional discussion and consensus on the topics of LMs such as biology, epidemiology, diagnosis, management, clinical research, and future directions (1).

Pathophysiology of LMs

Meninges
The meninges are fibrous membranes covering the CNS, composed of three layers and containing various cells; dura, arachnoid, and pia. The dural component is primarily composed of fibroblasts, perforated blood vasculature, lymphatic system, and immune cells different from those in the CNS. Separating the outer and inner meningeal section is the arachnoid cells, that control pathogens, molecules, and cells movement between the CSF and brain parenchyma. Nearest to the brain parenchyma is the pia matter (Figure 1). Although the meninges were recognized as a structure that primarily provides support to the CNS and protection from trauma, recent evidence suggests that the meninges act as an important connection between the CNS and the periphery, contribute to brain homeostasis, and are associated with various diseases (1,2,4).

Meninges as a site of cancer metastasis
The estimated incidence of LMs is approximately 5–20% based on population and autopsy data (4), and small reviews and autopsy studies have reported the incidence of LMs to approximate 9–25% for lung cancer (1). Cancer cells could metastasize into the meninges via the brain or choroid plexus, by crossing pial blood vessels, or through vascular channels which connect the bone marrow and meninges. To cross the BBB, cancer cells attach to endothelial cells and break tight junctions (1,2). Upon invading the leptomeninges, cancer cells may also regulate the CSF component to facilitate their growth. The CSF is acellular and has low levels of cytokine, glucose and protein, making it unfavorable for cancer cell growth. However, cancer cells can proliferate utilizing several mechanisms. Tumor cells in the CSF release complement component 3, which interferes with the barrier functions, thus enabling nutrients and macromolecules into the CSF (2). Another mechanism to facilitate cancer cell survival within the space is through expression of iron-scavenging protein and receptor pair, lipocalin-2 (LCN2) and SCL22A17 (6). Cancer cells secrete LCN2 into the CSF to sequester iron from CSF macrophage, thereby impairing macrophage function and supporting metastatic growth. Within the leptomeningeal space, cancer cells may adhere to the leptomeninges, creating a plaque-like appearance or be free-floating detected by CSF cytology. Free-floating cancer cells has been associated with poorer survival compared to plaque-like adherent LMs in lung cancer. Cytology positive for cancer cells was noted as a negative prognostic factor. Among patients with cytology negative for cancer cells, presence of nodular disease on MRI was a negative prognostic factor compared to radiographic linear enhancement on MRI (6).
Yao et al. (3) reported that blood cell lineage cells may transfer from the skull bone marrow or vertebral to the leptomeninges by migrating along the laminin-rich external surface of emissary veins (EVs). EVs, which are part of the vertebral-calvarial bone marrow vasculature, pass through foramina in the bone surface and emerge into the leptomeninges as blood vessels, thereby directly connecting the bone marrow compartment to the CNS. The vascular migration circumvents the necessity for cells to enter the circulation, cross the BBB or blood-CSF barrier, or destroy tissue boundaries. Whiteley et al. (7) reported that cancers expressing integrin α6 can invade the leptomeninges by traversing the outer surface of the blood vessels that normally connect the adjacent vertebral and skull bone marrows to the CNS meninges. Thereafter, cancer cells can induce resident meningeal macrophages to secrete a neuronal prosurvival protein, namely glial-derived neurotrophic factor, that stimulates tumor growth and promotes metastasis depending on the binding of integrin α6 to laminin. Zhao et al. (8) reported that laminin-bound integrin α6β4 promotes lung cancer progression via the activation of Yes-associated protein and transcriptional co-activator with PDZ-binding motif. Breast cancer cells employ a hematopoietic cell trafficking route to gain access into the leptomeninges. Once accessed, these cells communicate with resident myeloid cells to set up an advantageous microenvironment in the leptomeninges. The process combines neuronal imitation, immune cell takeover, and a restructuring of the leptomeninges to enable a more protective stromal niche. The core of this process consists of integrin α6 and NCAM. As lung cancer cells also express integrin α6 and/or NCAM, it is likely that the pathways are tumor agnostic. Therefore, future research that target these signaling mechanisms may benefit patients with LMs. Altogether, these findings highlight the tumor-host interactions that contribute to LM onset, which may provide novel therapeutic targets for its treatment (7).

Immune surveillance in the meninges
CNS is an immunologically special organ because of the BBB and absence of lymphatic vessels (9). However, functional lymphatic vessels are present within the CNS, particularly near the dural sinuses. These show lymphatic endothelial cell markers and can transport immune cells and CSF. Additionally, they are intimately associated with CSF circulation and are linked to the deep cervical lymph nodes (10). CSF is exchanged with the brain’s interstitial fluid, aiding in the clearance of the brain’s metabolic waste product such as beta-amyloid. The CSF then circulates through the subarachnoid space, ventricles, cisterns, and central canal. Meanwhile, the lymphatic vessels collect the CSF and large molecular substances such as antigens, and transport them to the deep cervical lymph nodes to trigger an immune response. Immune cells in the CNS originate from both the supply from the heart and the cranial marrow. The cranial marrow produces various immune cells, such as monocytes, neutrophils, and B cells, that may move to the meninges utilizing the channels between the meninges and skull; they can even enter the CNS parenchyma in a state of illness (11). The meninges harbor several immune cells, such as T cells, B cells, macrophages, and dendritic cells, all of that may directly contact the CSF and handle pathological issues in the CNS. Research scientists have revealed “arachnoid capillary exits” points that allow the exchange of molecules and CSF between the subarachnoid space and the dura mater. Immune cells in the meninges also assist in humoral immunity through dural-associated lymphoid tissues. In cases of CNS damage, levels of inflammatory mediators and antigens in the CSF increase, subsequently activating immunocytes in the meninges, that then move to the lymph nodes through the lymphatic system while starting antibody production and activating T cells (12). Remsik et al. (13) examined CSF specimens from patients with cancer through various methods, including multi-omic analyses. In the CSF of patients with LMs, they noted increased cellularity and various immune cell populations as well as elevated levels of several soluble inflammatory mediators. Among these, interferon γ (IFNγ) was consistently enriched across LMs originating from different primary tumors, including lung cancer, breast cancer, and melanoma. They suggested that the antitumoral role of IFNγ signaling in LMs is independent of the antigen presentation process. Furthermore, IFNγ-activated dendritic cells may play a role in inhibiting LM growth, underscoring the importance of immune-targeted therapies customized to this compartment (13).

Treatment of LMs
Therapeutic strategies for LMs include systemic therapy, intrathecal therapy, and radiotherapy (Figure 1). A deeper understanding of LMs have resulted in the development of brain-penetrating therapies, and novel systemic treatments, including oral tyrosine kinase inhibitors (TKIs), antibody-drug conjugates (ADCs), and immune checkpoint inhibitors (ICIs), exhibits remarkable intracranial efficacy that translates to clinical benefits for patients with LMs (1,5).

Systemic therapy
Some TKIs, ADCs, immunotherapies that have CNS penetrance, and systemic therapies with CNS pharmacological activity and blood-CSF barrier permeability may be considered for patients with LMs. The CSF penetrance of systemic therapies is dose-dependent, and dose escalation that focuses on CNS response rates and pharmacokinetic analysis should optimize therapy and prevent adverse events.
Systemic treatments for CNS diseases have evolved due the inclusion of molecularly targeted therapies into first-line treatment including epidermal growth factor receptors (EGFRs) and anaplastic lymphoma kinases (ALKs). Third-generation EGFR-TKIs have showed activity in CNS with favorable response rates and tolerability; however, the data largely involved patients with brain metastasis only and not LMs. Regarding LMs, data on the effectiveness of these treatments are limited. In trials, high doses of osimertinib resulted in acceptable response rates and persistence of responses, suggesting its toxicity and the necessity for medication dosage. In the BLOSSOM study, the standard 80-mg osimertinib also effectively controlled LMs. In preliminary reports, the combined use of amivantamab and lazertinib appears to be effective in both brain metastases and LMs (14). Notably, most of these studies involved patients who were treated with active agents and have developed LMs. Accordingly, using drugs with a different chemical structure or mode of action may be beneficial. Furthermore, it should also be determined whether first-line therapy, such as the combination of chemotherapy plus osimertinib or the combination of amivantamab plus lazertinib, would reduce the incidence of LMs. Both second and third-generation ALK inhibitors have better intracranial efficacy compared to the first-generation ALK inhibitors. Third-generation ALK inhibitors were designed to efficiently cross the BBB and to accomplish higher concentrations in the CSF, thus supplying a superior ability to suppress CNS extension. These effects were achieved by reducing the molecular weight, increasing lipophilicity, and revising the number of available hydrogen bond donors. Lorlatinib is a highly potent and selective third-generation ALK inhibitor that was specifically developed to pass through the BBB and has broad ALK mutational coverage. The effects of lorlatinib were verified in trial wherein the CSF to plasma ratio was 0.75 in patients. Another studyalso reported that analysis on samples of patients showed a CSF-to-free plasma concentration ratio of 0.77 (15). A phase II study of lorlatinib included two patients with LMs; one achieved complete intracranial response for 21.9 months, and the other achieved partial intracranial response for 11 months (16). In the largest study with LMs (n=11), lorlatinib resulted in an intracranial response rate of 45%, with a disease control rate of 91% (17). Two patients achieved complete intracranial response, and the median progression-free survival (PFS) was 9.3 months. Among nine patients with LMs, a German study reported a partial response of 77.8% (18). In the Crown study, the median PFS has yet to be reached in the lorlatinib group even after 5 years of follow-up, indicating prolonged intracranial efficacy (19).
Clinically relevant ADC activity has been documented in brain metastases and LMs (20). Trastuzumab deruxtecan has exhibited robust parenchymal brain metastasis responses, suggesting that it may also be effective for LMs in patients with breast cancer regardless of its vagueness behind the mechanisms of penetration of monoclonal antibody into the CSF. The TUXEDO-3 trial was an international, multicenter, phase 2 study investigating patritumab deruxtecan (HER3-DXd) for various cancers focusing on patients with treatment-naive LMs or LMs progressing after radiotherapy (21). According to the results from cohort 3 of the phase 2 TUXEDO-3 trial, patritumab deruxtecan has clinically relevant activity and a manageable safety profile in patients with LMs from primary solid tumors. Among the 20 participants, with the majority having breast (60%) and lung (30%) cancers, 65% of patients were alive after 3 months, offering a new potential avenue for patients with LMs. Given the poor prognosis of patients with LMs and the limited treatment landscape, the positive results from TUXEDO-3 are encouraging.
ICIs are effective in various malignancies, including lung cancer, and they have also demonstrated efficacy for brain metastases. In patients with LMs, ICIs had a favorable safety profile and showed encouraging activity (22). Additionally, a pooled CSF analysis showed the increasing of CD8+ T-lymphocytes and gamma-interferon signaling following ICI therapy (23); however, the activity of ICIs within LMs were less robust. Thus, their role in LMs remain limited. As single agents, ICIs have not markedly improved survival of patients with LMs. In large cohorts of patients with cancer, Grippin et al. reported that commercially available mRNA vaccines against COVID condition patients’ immune systems to eliminate cancer cells, and combining these vaccines with ICIs overwhelms tumor defenses and improves survival (24). The effectivity of combination therapies, such ICIs with proton-based craniospinal radiation, dual ICIs, or other systemic and intrathecal therapies with ICIs are currently under investigation.

Intrathecal therapy
Intrathecal chemotherapy is administered by directly injecting agents into the CSF via lumbar puncture or using intracerebral catheters such as an Omaya reservoir. Intrathecal drugs are generally effective against floating and thin linear deposits of LMs but may be less responsive to bulky and nodular LMs. Intrathecal chemotherapy is limited to drugs such as methotrexate, thiotepa, cytarabine, and topotecan, and recent evidence has shown that pemetrexed, which is a multi-target anti-folate metabolic agent, or nivolumab have reasonable safety profiles. Zhao et al. conducted a meta-analysis of 8 clinical trials with 306 participants to evaluate the efficacy and safety of intrathecal pemetrexed for LMs in lung cancer. The pooled analysis indicated a response rate of 57.6%, disease control rate of 85.4%, and overall survival of 8.12 months (25). Studies on intrathecal ICIs, immune effector cells, pertuzumab, and deferoxamine for solid tumors are ongoing (1). Recently, the Food and Drug Administration has granted orphan drug designation to rhenium (186Re) obisbemeda (26), a novel injectable radiotherapy, for LMs in lung cancer. Rhenium (186Re) obisbemeda delivers highly-targeted high-dose radiation directly to CNS tumors. The agent was recently evaluated for LMs in the phase 1/2a ReSPECT-LM trial (NCT05034497). Currently, a phase 2 dose expansion study and a phase 1 multiple-dose study will evaluate the drug in patients with lung cancer and LMs.
If tumor cells are present in the CSF, intrathecal therapy is usually recommended regardless of the MRI findings. Furthermore, an alteration of systemic therapy is considered if LMs are confirmed. Focal radiation therapy is considered for symptomatic nodular involvement or in the presence of symptomatic lesions of cranial nerves or the cauda equina. Ventricular access devices, such as the Ommaya reservoir, are advantageous over drug delivery from lumbar puncture in patients with intrathecal chemotherapy because of the simplicity in management. CSF shunt devices such as ventriculoperitoneal shunts also reduce symptoms of elevated intracranial pressure and should be considered as a palliative procedure.

Radiation therapy
Involved-field RT (IFRT), whole brain RT, and focal spine RT are effective modalities for symptom management. Meanwhile, conventional photon-based craniospinal irradiation (CSI) targeting the CNS may result in off-target damage to several internal organs and myelosuppression of the spinal column. These palliative interventions do not improve survival in LMs because LMs tumor dissemination affects the entire neural axis, making it necessary to target the craniospinal canal. Proton-beam CSI has demonstrated reasonable toxicity and superior survival compared to photon IFRT; it may be a therapeutic alternative following deliberate consideration (1). Proton-beam CSI offers more selective treatment of CNS structures due to a tighter range of radiation delivery with limited exit dosage and has become recognized as a safer choice. It may also offer superior LM control and improve patient survival because of its capability of cytoreduction of leptomeningeal dissemination in comparison to IFRT. In a randomized phase 2 study of proton-beam CSI vs. photon IFRT which included patients with lung cancer and breast cancer with LMs, Yang et al. reported a considerable benefit in CNS-PFS with proton-beam CSI compared with photon IFRT, leading to the early termination of the study (27). In the final analysis, a notable advantage was continually gained for CNS-PFS with proton-beam CSI (median, 8.2 months) vs. photon IFRT (median, 2.3 months; P<0.001). Moreover, a significant and clinically meaningful overall survival benefit with proton-beam CSI (median, 11.3 months) vs. photon IFRT (median, 4.9 months; P=0.04) was also appreciated. The study demonstrated improved CNS-PFS and overall survival following proton-beam CSI than with photon IFRT, indicating that proton-beam CSI should be considered when available.
LMs mean an aggressive and advanced stage of cancer that causes substantial neurologic illness. Managing LMs is challenging due to leptomeningeal cancer cell biology, immunosuppressive CSF microenvironment, and limitations of drug delivery. However, recent advances in both diagnostic methodologies and molecularly targeted therapies, intrathecal therapies, immunotherapies, cell-based therapies, proton CSI and ongoing innovative clinical trials provide expectation for improving outcomes in patients with LMs. Identifying cancer cell molecular weakness and mechanistic analysis of immune-exposed cancer cell function may discover the novel targets and CSF-specific immunomodulators. Improved understandings of cancer-immune interactions within LMs propose dendritic and NK cell-engaging therapies which may act as more effective strategies to regulate LMs. Intensified cooperation among medical- and neuro-oncology, nuclear medicine, and radiation oncology as well as basic research are warranted to further develop novel therapies for LMs.

Conclusions
Recent cutting-edge research and study on LMs may open new possibilities for significantly improving patient prognosis and could contribute to the development of novel treatments.

Supplementary
The article’s supplementary files as