Long-term control of leptomeningeal disease after radiation therapy and nivolumab in a metastatic melanoma patient

Abstract

Background:

Leptomeningeal disease (LMD) from melanoma is rapidly fatal with median overall survival between 6.9 weeks and 3.5 months. It is not known whether immune checkpoint inhibitors have a role in treating LMD.

Case presentation:

We report a 33-year-old male patient who developed LMD from a BRAF V600E-mutated melanoma brain metastasis, despite prior treatment with surgical resection, radiotherapy and dabrafenib/trametinib. He underwent whole brain radiotherapy with stereotactic radiotherapy to the lumbosacral spine, and was started on nivolumab, which led to prolonged remission lasting 2 years and 3 months, before disease progression and death.

Conclusion:

This is the first case report to highlight a potential long-term efficacy of radiotherapy and anti-PD-1 immunotherapy, in treating LMD from metastatic melanoma that is resistant to targeted therapy.

Background

Leptomeningeal disease (LMD) occurs in between 5 and 25% of metastatic melanoma patients [1]; it is rapidly fatal with reported median overall survival between 6.9 weeks and 3.5 months [2,3]. Palliative craniospinal irradiation and intrathecal chemotherapy have been used in the past, though with only marginal benefits [4]. In two prior case reports, long-term responses up to 18–19 months were observed in two melanoma patients with LMD treated with BRAF or BRAF/MEK targeted therapies [5,6], though it is not clear what therapeutic option is available upon disease progression. Intrathecal administration of IL-2 has been attempted, though this is limited to highly selected patient population due to the toxicity profile [7]. More recently, Glitza et al., reported some efficacy against LMD after adoptive transfer of autologous tumor-infiltrating lymphocytes via the intrathecal route in a metastatic melanoma patient [8]. Immune checkpoint inhibitors have contributed improvements in the treatment of brain metastases from metastatic melanoma based on a case series and reports from recent ongoing clinical trials [9–14]. Murine studies in mice have shown the synergistic potential of different immune checkpoint therapies with radiotherapy: it was shown that anti-CTLA4 blockade promoted expansion of CD8⁺ Teff while inhibiting T-regulatory cells (Tregs), whereas radiation enhanced the diversity of the T-cell receptor (TCR) repertoire of the expanded peripheral and intratumoral T-cell clones [15]. However, PD-L1 upregulation on melanoma tumor cells has emerged as a resistance mechanism to radiation and anti-CTLA4 therapy in both murine models and of melanoma patients. Hence, the addition of anti-PD-1/PD-L1 blockade was found to reverse T-cell exhaustion, mitigate depression in the CD8⁺ Teff/Treg ratio, and encourage further oligoclonal T-cell expansions [15]. Currently, it is not known whether anti-PD-1 immune checkpoint therapy has a role in treating LMD. There are multiple ongoing clinical trials (NCT02939300, NCT03025256, NCT03719768 and NCT03091478) to address this important question. At the present time, there is no known optimal approach for treating LMD from metastatic melanoma that has become resistant to the BRAF/MEK targeted therapy.

Case presentation

The patient was a 33-year-old Caucasian male who was initially diagnosed in 2010 with a T2b cutaneous melanoma of the trunk with 2 mm Breslow thickness. He underwent a wide local excision of the primary lesion and a sentinel lymph node biopsy, which showed evidence of micrometastases in 2 loco-regional lymph nodes (N2a). He later underwent complete lymph node dissection of 35 loco-regional lymph nodes, which were negative for additional metastases. His clinical stage was IIIB (T2b, N2a, M0). He was followed with close surveillance without adjuvant therapy at an outside institution until 3 years later, when he presented with sudden onset of left-sided weakness of his upper and lower extremities. He presented to a local hospital initially and was subsequently transferred to the UPMC Shadyside Hospital for further care. MRI of the brain revealed a solitary right frontal mass. Computed tomography chest demonstrated a left lung mass and multiple equivocal sub-centimeter nodules (not shown). He underwent right frontal craniotomy and tumor resection. Pathology confirmed metastatic melanoma, positive for BRAF V600E mutation. He underwent stereotactic radiosurgery (SRS) to the right frontotemporal resection bed (24 Gy in 3 fractions), and was started on systematic therapy with the BRAF and MEK kinase inhibitors, dabrafenib and trametinib, after completion of SRS [16]. He continued to receive treatment for 2 months until a surveillance MRI brain showed a suspicious 7 mm enhancing lesion on the left VIII cranial nerve and 2 other subcentimeter lesions. This was treated with additional SRS to the right posterior mesial frontal lobe (24 Gy in 1 fraction), cranial nerve VIII (16 Gy in 1 fraction) and the left rolandic metastasis (24 Gy in 1 fraction).

He was continued on dabrafenib/trametinib and did well for a year until he suddenly developed headache, diplopia and left leg weakness and was admitted to the hospital after a syncopal episode. MRI brain demonstrated diffuse LMD with prominent lateral and third ventricles (Figure 1A & B). Initial cerebrospinal fluid (CSF) demonstrated presence of melanoma cells (cytology in Figure 2A, with detailed CSF analysis data in Table 1). Concurrent PET/computed tomography imaging demonstrated a ¹⁸F-fluorodeoxyglucose (FDG)-avid lesion within the lumbosacral junction (Figure 3A), likely causing significant pain in his left lower extremity. He was treated with whole brain radiation therapy (30 Gy in 10 fractions) without complications. He was initiated 2 weeks later on systematic anti-PD-1 immunotherapy, nivolumab, at a dose of 3 mg/kg every 2 weeks. He also received SRS in 1 fraction to a lumbosacral spine lesion concurrently (information on radiation dosage unavailable). He was not on systematic steroid at the time. He did well through the first four cycles (2 months) until he developed worsening left eye vision and papilledema. He was found to have hydrocephalus due to LMD and underwent placement of a ventriculoperitoneal shunt. CSF cytology was interpreted as suspicious for the presence of melanoma cells (Figure 2B & Table 1). He continued to receive nivolumab therapy. After eleven cycles of treatment, three serial follow-up MRI of the brain have demonstrated sustained improvement of his LMD in brain and lumbar spine (representative figures shown in Figure 1C & Figure 3B). His temporary remission of LMD lasted approximately 2 years and 3 months since starting nivolumab. During this time, he had no significant neurologic symptom other than mild left thigh weakness. Despite continuing therapy with nivolumab, he unfortunately presented in August 2017 with severe headache, diplopia and lower extremity weakness. Restaging MRI brain revealed carcinomatous meningitis that required adjustment of his VP shunt setting. Cytology on the CSF fluid confirmed the presence of melanoma cells (Figure 2C & Table 1). He was placed on a quick taper of high-dose systematic steroid due to transient mental status changes. He was restarted on dabrafenib and trametinib, though these were discontinued a month later due to cardiotoxicity. Nivolumab was restarted every 2 weeks. However, his CNS disease continued to progress despite these interventions. He was eventually transitioned to hospice and died in April 2018.

Figure 1. . MRI brain imaging of the patient's leptomeningeal disease prior to and after treatment with radiation therapy and nivolumab.

(A) Fluid-attenuation inversion recovery axial image demonstrates abnormal signal scattered through the sulci (arrows), indicating abnormal protein content in the cerebrospinal fluid. (B) Contrast-enhanced spoiled gradient axial image obtained at the same time as (A) reveals abnormally increased enhancement in a pial pattern (arrows), indicating leptomeningeal spread of tumor. (C) Contrast-enhanced spoiled gradient axial image after treatment shows resolution of pial enhancement.

Figure 2. . Cerebrospinal fluid cytology of the patient prior to, during, and after treatment with radiation therapy and nivolumab.

(A) Bloody cerebrospinal fluid (CSF) sample, at the time of leptomeningeal disease (LMD) diagnosis, showing an individual melanoma cell (cytospin, diff quick stain, magnification ×600). (B) CSF sample, taken after whole-brain radiotherapy and four cycles of nivolumab, showing melanosis due to macrophages heavily laden with melanin pigment. (C) Bloody CSF sample, taken at the time of LMD disease progression, showing an individual melanoma cell (cytospin, diff quick stain, magnification ×600).

Table 1. . Results of cerebrospinal fluid studies at different treatment time points.

Cell counts	30 April 2015 (at the time of LMD diagnosis)	19 August 2015 (after WBRT and 4 cycles of nivolumab)	19 August 2017 (At the time of disease relapse postimmunotherapy)
WBC (/mm × 103)	49	10	273
RBC (/mm × 103)	38,250	0	295,362
% Neutrophils	58	8	75
% Lymphocytes	32	90	15
% Monocytes	10	2	0
% ‘Other’ Cells	3	Not reported	10
Protein	752	46	485

LMD: Leptomeningeal disease; RBC: Red blood cell; WBC: White blood cell.

Figure 3. . MRI lumbar spine imaging of the patient's leptomeningeal disease prior to and after treatment with nivolumab and radiation therapy.

(A) Contrast-enhanced sagittal T1-weighted image reveals extensive abnormal enhancement in the thecal sac, including nodules in the cauda equina, posterior dural enhancement and a large mass at the tip of the thecal sac. (B) Contrast-enhanced sagittal T1-weighted image performed after treatment reveals near-complete resolution of lumbar spinal disease with only a few linear areas of enhancement that remained stable for several years.

Discussion

There is paucity of data on the efficacy of immune checkpoint therapy against LMD from metastatic melanoma. There has only been one other case report describing the successful use of another immune checkpoint therapy, ipilimumab, in combination with WBRT, against LMD from melanoma [17]. However, rare neurologic syndromes have been reported in metastatic melanoma patients treated with ipilimumab [18]. There is currently no data in the literature on the efficacy of anti-PD-1 immune checkpoint therapy, which has a tolerable side effect profile and different immunologic mechanisms of action versus ipilimumab, against LMD from melanoma, including those that were resistant to targeted BRAF/MEK therapy. We wish to bring this to the attention of our colleagues of the potential efficacy of anti-PD-1 immunotherapy in combination with radiotherapy, in treating LMD from melanoma resistant to targeted therapy. Though the patient was later re-challenged with the BRAF/MEK targeted therapy for 1 month, this was unlikely to contribute to his survival of another 7–8 months due to the short duration of therapy. We think that radiotherapy likely synergized with the immunotherapy in this patient, to promote an effective T-cells’ response against the released tumor antigen(s). In this patient, there was a 2-week interval between the completion of the WBRT and initiation of nivolumab. We think that the initial WBRT served to release tumor antigens to prime the antitumor T-cell responses, while the initiation of anti-PD-1 immune checkpoint therapy within a 2-week time frame helped to prevent the expanded T-cell clones from becoming exhausted in order to sustain their antitumor activity. He also received concurrent SRS to a lumbosacral metastatic lesion at the time of nivolumab initiation, which may have helped to release additional tumor antigens to further enhance the diversity of TCR repertoire targeting tumor antigens.

It was difficult to define the patient’s overall disease response based on the RECIST criteria, as there is not a uniform definition of leptomeningeal disease. Although radiographically he could have been described as having a CR (complete response), there were still melanoma cells that persisted in his CSF despite him having had significant clinical responses elsewhere (i.e., brain, lumbosacral area). Thus, his overall best disease response could only be characterized as SD (stable disease) based on the RECIST criteria.

There is now data supporting T-cells’ access to the meninges via the functional lymphatic vessels lining the dural sinuses [19]. These structures, which are connected to the deep cervical lymph nodes, express molecular markers of lymphatic endothelial cells that allow for the trafficking of immune cells from the CSF. We hypothesize that T cells activated by exposure to anti-PD-1 immunotherapy likely have differential expressions of adhesion molecules as compared with the nonactivated T cells. Consequently, the activated effector T cells enter the CNS from the blood via the leptomeninges, from which they may be released to CSF and re-attach to leptomeninges, whereas nonactivated T cells are released into the CSF. This potential mechanism is being evaluated in an ongoing clinical study of anti-PD-1 immunotherapy and SRS by others [20].

In sum, the overall effects of anti-PD-1 immunotherapy in combination with radiotherapy may help to reverse T-cell exhaustion as well as influence the CNS trafficking pattern of the activated antitumor T-cells. These phenomenons may explain the potential synergistic efficacy of anti-PD-1 immunotherapy and radiotherapy against LMD from metastatic melanoma resistant to the BRAF/MEK targeted therapy in this patient.

Conclusion

The most interesting aspect of this case is that the combined modalities of anti-PD-1 immunotherapy, together with radiation therapy, were able to sustain response of LMD in near remission for 2 years and 3 months. This is in contrast to the expected average life expectancy between 6.9 weeks and 3.5 months for melanoma patients with LMD without treatment [2-3]. As the majority of ongoing cancer immunotherapy clinical trials still listed presence of LMD as an exclusion criteria, we hope that the case reported here will spur clinical investigations to further explore more efficacious and tolerable immunotherapy agent in combination with either WBRT versus targeted radiotherapy in the treatment of LMD, as there is currently no data on the optimal therapeutic approaches. It will also be of immense interest in the field to further delineate the molecular mechanism behind the T-cell trafficking pattern in CNS/leptomeninges in response to immunotherapy.

Summary points.

• Leptomeningeal disease (LMD) occurs in between 5 and 25% of metastatic melanoma patients. It is rapidly fatal with a median overall survival between 6.9 weeks and 3.5 months since the time of diagnosis.

• Numerous approaches have been attempted previously, such as palliative craniospinal radiation, intrathecal chemotherapy/immunotherapy (i.e., IL-2), however these have largely been abandoned due to lack of demonstrable benefits or could only be used in highly selected patient population due to significant toxicities.

• Targeted therapy targeting BRAF/MEK have been previously in reported in two cases to lead to prolonged remission of LMD. It is not known what treatment option is available if patients’ disease is resistant to BRAF/MEK inhibitors.

• Immune checkpoint inhibitors have contributed to improvements in the treatment of brain metastases from metastatic melanoma based on a case series and reports from recent ongoing clinical trials. However, it is not known whether this form of immunotherapy has a role in treating LMD.

• Murine studies in mice have shown the synergistic potential of different immune checkpoint therapies with radiotherapy. It was shown that anti-CTLA4 blockade promoted expansion of CD8⁺ Teff while inhibiting T-regulatory cells, whereas radiation enhanced the diversity of the TCR repertoire of the expanded peripheral and intratumoral T-cell clones. Upregulation of PD-L1 on melanoma tumors has emerged as a resistance mechanism by promoting T-cell exhaustion. Anti-PD-1 therapy was shown to reverse T-cell exhaustion and reverse depressed CD8⁺ Teff/T-regulatory cells ratios. Priming of tumor antigen-specific T cells after abscopal effect from radiotherapy has been shown previously as well.

• We report a case of a 33-year-old male patient who developed a solary brain metastasis with BRAF V600E mutation from previously resected stage IIIB melanoma. Despite treatment with surgical resection followed by stereotactic radiotherapy to the resection bed and targeted therapy (dabrafenib/trametinib), he developed extensive LMD a year later. He was started on anti-PD-1 immunotherapy, nivolumab, at 3 mg/kg every 2 weeks. He achieved a prolonged remission lasting 2 years and 3 months. His LMD eventually progressed despite addition of dabrafenib/trametinib to nivolumab, which led to his death at age 38-year old.

• In conclusion, the synergistic effects between anti-PD-1 immune checkpoint therapy and radiotherapy may help to reverse T-cell exhaustion and improve trafficking of the antitumor T-cells to the CNS/meninges. This may explain the efficacy of such combination treatment approach against LMD from metastatic melanoma resistant to the BRAF/MEK targeted therapy.