Metastatic myxopapillary ependymoma treated with immunotherapy achieving durable response

Abstract

Myxopapillary ependymoma (MPE) is a rare glial tumour mainly located in the areas of the conus medullaris, cauda equina and filum terminale of the spinal cord. Ectopic MPE tends to behave more aggressively and distant metastases are often seen. Unfortunately, no standard treatment options are established as only small series of treated patients and a few reported cases are available in the literature. We report the case of a 25-year-old woman who was initially diagnosed with a metastatic MPE, with multiple bilateral lung metastases. She was treated with an investigational monoclonal antibody antiprogrammed cell death protein 1, called tislelizumab (BGB-A317), following surgical resection of the perisacral primary mass. The response was long-lasting and side effects nil. Immunotherapy is a treatment modality to be considered in patients with rare tumours.

Background

Myxopapillary ependymoma (MPE) is a rare subtype of glial tumours mainly located in the areas of the conus medullaris, cauda equina and filum terminale of the spinal cord.¹ The incidence of MPE is estimated of 1.00 per million person-years with a peak incidence among young patients aged 20–30 years.² Back pain is the most common presenting symptom due to nerve root compression, although other sensory abnormalities such as lower limb weakness and sphincter dysfunction can also be seen.³ MPE is currently classified by WHO 2016 as a grade I glioma,⁴ a slow-growing ependymal tumour which is usually associated with a favourable prognosis with complete surgical resection, the mainstay of treatment.^{5 6} An extradural location of this tumour seems to confer worse prognosis with higher risk of distant metastatic deposits.⁷ Due to the rarity of this tumours, the management of patients with recurrent or de novo metastatic MPE is unclear because there is no standard salvage therapy and the available data on the role of systemic oncology treatments is scarce. In these situations, the prognosis of these patients is sadly much poorer with survival estimated around 15% at 5 years.^{8 9}

We report the case of a 25-year-old woman who was initially diagnosed with a metastatic MPE with bilateral lung metastases. She was treated with an investigational monoclonal antibody (MAb) antiprogrammed cell death protein 1 (PD-1), called tislelizumab (BGB-A317),¹⁰ following surgical resection of the perisacral primary mass. Our patient treated with this new checkpoint inhibitor in a phase Ib clinical trial achieved stabilisation of her disease for more than 18 months without experiencing any side effects. To the best of our knowledge, this is the first reported case of a metastatic extradural MPE patient treated (successfully) with palliative immunotherapy.

Case presentation

Our patient was 23 years of age when she started complaining of lumbosacral pain. She told her general practitioner (GP) that her pain was insidious in onset and had been treated unsuccessfully by over-the-counter analgesics. She denied any previous trauma and her clinical examination was unremarkable. At the time of initial presentation to her GP radiological sacral and coccygeal examinations showed no structural abnormality and laboratory findings were within normal reference ranges. On a background of no medical history and previously being fit and healthy, her pain was presumed to be a lumbar strain. She was instructed to rest and more anti-inflammatories were prescribed.

By July 2015, 9 months after the initial visit to her GP, the pain had become constant and disabling, radiating to the right buttock and posterolateral thigh down to knee.

Investigations

With a working diagnosis of L5–S1 nerve root impingement MRI and a CT scans of the lumbosacral region were requested. The imaging revealed a large sacral mass (10×8 ×8.7 cm) with associated bone destruction and gluteus maximus invasion. This tumour was displacing but not invading pelvic organs (rectum, vagina and uterus) (figure 1). The radiological differential diagnoses for this sacral mass were wide including: chordoma, giant cell tumour, schwannoma or chondrosarcoma. Her staging chest/abdomen/pelvis CT scan (August 2015) unexpectedly identified multiple lung deposits throughout the lung fields.

Figure 1.

Sagittal postcontrast fat saturated image at the sacral level demonstrating a large sacral mass with bone destruction and right gluteus maximus invasion. Pelvic organs (rectum, uterus and ovaries) are displaced anteriorly by the mass but are not macroscopically invaded.

The definitive diagnosis was established by histological examination of core biopsies of the sacral mass, as well as one of the lung lesions demonstrating grade 1 MPE with vascular invasion present.

In September 2015, our patient was referred to the Royal Prince Alfred Hospital in Sydney where she underwent a major surgical resection of the perisacral mass including partial removal of the sacrum with clear margins (figure 2). Macroscopically, the tumour measured 12 x 9.5 x 10 cm and it was encapsulated with soft to rubbery and greyish to reddish-brown appearance. Microscopic study showed a fibrous capsule and variable architecture demonstrating gland-like spaces and arrays reminiscent of papillae with fibrovascular cores. There were also sheets and nests of cuboidal to elongated cells with round to oval nuclei, speckled chromatin and eosinophilic cytoplasm. Some of the blood vessels were surrounded by relatively anuclear eosinophilic zones forming perivascular pseudorosettes. In addition, there were occasional eosinophilic spheroid structures in the more cellular areas, reminiscent of ‘balloons’ described in MPE. Within and beyond the fibrous capsule, there was vascular invasion with tumour plugs demonstrating endothelial attachment (figure 3). Immunohistochemistry was positive for glial fibrillary acidic protein (GFAP), CD56, synaptophysin (SYP), vimentin and negative for epithelial membrane antigen (EMA), Pan-CK, chromo A and S100. Neither expression of hormone receptors nor c-kit (CD117) was seen. CD99 showed weak and patchy staining and Ki-67% was approximately 10%–20%. Somatic EGFR, RAS (small GTPase involved in transmitting signals within cells) and BRAF (v-Raf murine sarcoma viral oncogene homolog B) mutations were not detected in tumour tissue. PD-L1 expression on tumour-infiltrating ICs was 5% and on tumour cells 0% (figure 4).

Figure 2.

Macroscopic view of the resected sacral tumour. The specimen is sectioned sagittally to reveal a tumour with cut surface dimensions of 120 mm (anterior to posterior) x 95 mm (superior to inferior) x 100 mm (left to right). The tumour appears well defined and has a multinodular pale to mucoid and haemorrhagic appearance. The bony margins appear macroscopically clear.

Figure 3.

Pathological findings of the resected tumour. H&E stain original magnification ×40. Picture showing the tumour with variable architecture, demonstrating gland-like spaces and arrays reminiscent of papillae with fibrovascular cores. In addition, there are occasional eosinophilic spheroid structures in the more cellular area, reminiscent of ‘balloons’ described in myxopapillary ependymomas. Immunohistochemistry was positive for GFAP, CD-56, SYP, vimentin and negative for EMA, Pan-CK, chromo A and S100. CD-99 showed weak and patchy staining and Ki-67% was approximately 10%–20%. Somatic EGFR, RAS and BRAF mutations were not detected in tumour tissue. PD-L1 expression on tumour-infiltrating immune cells was 5% and on tumour cells 0%. BRAF, v-Raf murine sarcoma viral oncogene homolog B; EGFR, epidermal growth factor receptor; EMA, epithelial membrane antigen; GFAP, glial fibrillary acidic protein; RAS, small GTPase involved in transmitting signals within cells; SYP, synaptophysin.

Figure 4.

PD-L1 expression immunostaining. Tumour specimen was stained using VENTANA PD-L1 (SP263) assay. Images were scanned at ×40 magnification on VENTANA iscan high-throughput (HT) scanner.

Postoperative course was uneventful and she was discharged home on the 22nd postoperative day (although requiring self-catheterisations due to postsurgical urinary incontinence and tapentadol, pregabalin and tramadol for postoperative pain).

Following resection of the primary tumour, her chest CT scans were reviewed by thoracic surgeons. Disappointingly they felt that her pulmonary lesions were too numerous for an optimal resection and therefore any subsequent treatment would be focused on maintaining quality of life and managing symptoms. She did not receive adjuvant radiotherapy to the sacral area because of this new incidental finding of lung metastases. Although she was asymptomatic from the chest perspective, the lung lesions were insidiously increasing in size and number. For example, a right middle lobe metastasis grew from 29×23 mm in August 2015 to 82×72 mm in May 2016 and a left lower lobe nodule grew from 45×48 mm in August 2015 to 90×63 mm in May 2016 (figure 5). The largest lung lesions doubled in size in just 9 months.

Figure 5.

Soft-tissue and lung window axial postcontrast CT demonstrated bulky bilateral lung metastases. Baseline scan preimmunotherapy treatment.

Treatment

Unfortunately this patient had a rare subtype of a rare tumour making treatment recommendations difficult. Palliative treatment, generally with platinum-based chemotherapy and etoposide, is considered one possible option for this type of tumour, based on a series of cases, however, response rates with this combination are low and the risk of toxicity is significant. As she was fit and well and given the fact that her case did not require chemotherapy urgently, our patient was also open for discussion on phase I clinical trials. At that time, we had an open phase I clinical trial at our institution evaluating a new anti-PD1 monoclonal antibody for the treatment of refractory advanced tumours. She wished to explore this option further.

The BGB-A317-001 clinical trial is a multicentre, open-label, multiple-arm, phase Ib dose expansion study (NCT02407990) evaluating the safety, tolerability, pharmacokinetics and treatment effects of a new anti-PD1 checkpoint inhibitor, tislelizumab (BGB-A317), in patients with advanced cancers who have failed standard-care treatment.^{10 11} BGB-A317 is a humanised IgG4 variant monoclonal antibody against PD-1 and is being developed by BeiGene. On administration, BGB-A317 binds to PD-1 and inhibits the binding of PD-1 to the PD-1 ligands 1 and 2 (PD-L1 and PD-L2).

After considering the options, our patient signed an informed consent to participate in this clinical trial on 2 June 2016. After meeting all the inclusion and exclusion criteria she was registered into the study on 16 June 2016. She was recruited under the cancer indication arm 13 for subjects with renal cell carcinoma, bladder cancer, melanoma, Merkel cell carcinoma, sarcoma or any other solid tumours not covered in the other 12 arms of the study. Her treatment was started on 17 June 2016 (figure 6).

Figure 6.

Interval CT scan while on immunotherapy showing stable disease as per RECIST criteria V1.1 (CT scan date April 2017). RECIST, Response Evaluation Criteria in Solid Tumour.

In this study, tislelizumab was given at a dose of 5 mg/kg (290 mg in the case of our patient) by 30 min intravenous infusion every 3 weeks. Subjects were prophylactically pre-treated with oral H1 blockers (cetirizine 10 mg) and oral acetaminophen (paracetamol 500 mg) 30–60 min prior to each infusion.¹¹ Our patient neither experienced infusion reactions nor significant side effects and she tolerated treatment well. The best radiological response to treatment was stable disease. She had completed 18 doses of tislelizumab when her restaging CT scan in June 2017 showed progressive disease (as per Response Evaluation Criteria in Solid Tumours (RECIST) V.1.1 criteria¹²) within her lungs. However, she was considered for treatment beyond progression due to clinical benefit, after reconsenting. There is increasing evidence that a minority of patients treated with checkpoint inhibitors such as tislelizumab may derive clinical benefit despite initial documentation of progressive disease by RECIST. This pseudoprogression is thought to occur due to immune cell (IC) infiltration and tumour response to immunotherapy agents.¹³ This was our rationale for maintaining the patient in the trial. More treatment was given on a 3 weekly basis with excellent tolerance.

Sadly, 7 months later, in January 2018, her CT scan showed further progressive lung disease. She had received 27 doses over 18 months, with the last dose given on 13 December 2017. Her best radiological response during immunotherapy was stabilisation of her disease. At this time she continued working and remained clinically asymptomatic apart from recurrent urinary tract infections likely secondary to late morbidity from the previous surgery. However, she had to come off the study and therefore second line options were sought.

Outcome and follow-up

Based on the Collaborative Ependymoma Research Network (CERN) phase II clinical trial (CERN08-02) for patients with recurrent low-grade and anaplastic supratentorial, infratentorial and spinal cord ependymomas,¹⁴ our patient was started on oral temozolomide on days 1–7 & 15–21 of a 28-day-cycle at 125 mg/m² (temozolomide 200 mg daily). She started this treatment in January 2018. Ten months later her lung disease progressed despite temozolomide. In January 2019, she was started on third line palliative treatment with 4 weekly intravenous carboplatin area under the concentration-time curve 5 (750 mg) and so far she has not experienced any major side effects of note—apart from mild anticipatory emesis and moderate degree of fatigue.^{15 16} At present, she continues on this chemotherapy treatment (figure 7).

Figure 7.

Most recent CT scan capture taken at the time of the case report submission (March 2020).

Discussion

MPEs are rare tumours, probably arising from extradural remnants of the filum terminale, the coccygeal medullary vestige or ectopic ependymal cells nests^{1 3} although there are reports of MPEs found within the brain, the cervicothoracic spinal cord and, in very rare cases, extraspinally. Outside the central nervous system (CNS), they mostly arise from the sacrococcygeal and presacral regions.^{17 18} This was particularly the case of our patient. The incidence of MPEs is thought to be 1.00 per million person-years and young adults (18–29 years) are most affected.² The most common presenting symptoms are pain in the lower back, legs and sacrum. Up to 25% of patients may present with weakness of the lower limbs or sphincter dysfunction.

MPE was first described by Mallory¹⁹ and is currently classified by the WHO 2016 as a grade I glioma: a slowly growing ependymal tumour.⁴ The WHO grading system assigns a grade to ependymomas based on cellular pleomorphism, mitotic count, tumour necrosis and cellularity, endothelial proliferation, grade of invasion and molecular and genetic features. MPEs have historically been considered benign glial tumours and if excised completely, the prognosis is excellent, with 5-year survival over 98%. However, extradural MPEs seldom disseminate within the CNS (unlike intradural spinal ependymomas), but pose a significant risk for systemic spread leading to a much worse prognosis. In one review of extraspinal ependymomas, the prevalence of extraneural metastases was found to be approximately 5% for presacral ependymomas and >15% for postsacral ependymomas.^{7 20} Other studies have reported prevalence rates ranging between 18% and 27%.²¹ The lungs are the most common metastatic site, but metastases may also be seen in the liver, bones and lymph nodes. Another feature of these tumours is that distant metastases can be diagnosed several years following initial presentation.²² There are no available prognostic biomarkers predicting the biological behaviour of ependymomas and studies of genetic changes and potential treatment targets in adults are sparse.^{23 24}

For those patients with widespread disease, choosing a treatment strategy becomes challenging. First, most of the pertinent literature is based on small series of patients or reported cases rather than on randomised controlled trials. Second, the scarce available data tend to mix low-grade supratentorial, infratentorial and spinal cord ependymomas (which behave differently). In addition, some of these studies combine paediatric and adult populations making it very difficult to draw conclusions for a specific age group; as a result, there are neither standard systemic treatment options nor consensual guidelines for the management of metastatic MPE patients. A variety of chemotherapeutic agents have been mainly used in small series for the treatment of recurrent intracranial ependymomas: platinum-based chemotherapy, oral temozolomide and oral topotecan have shown very modest efficacy with median progression-free survivals (mPFS) ranging from 5 to 15 months. It is uncertain whether this data can be extrapolated to metastatic spinal cord and ectopic MPEs.8 15 25 26

Chamberlain, for instance, conducted a prospective phase II study enrolling 10 consecutive adult patients with recurrent low-grade ependymoma of the spinal cord. Patients were treated with oral etoposide (50 mg/m²/day in a 3-week on schedule and 2-week off schedule). The mPFS was 15 months (range 2.5–45 months) and the median overall survival (mOS) was 17.5 months (range 3–45 months). It should be noted that all the patients had ependymomas of the cervicothoracic regions of the spine only.²⁶

Regarding targeted therapies for the treatment of MPE the data is even more limited. Fegerl reported an anecdotal case on a 59-year-old woman with metastatic MPE treated with imatinib (a multityrosine kinase inhibitor (TKI) targeting cKIT (human receptor tyrosine kinase referred to as stem cell factor receptor or CD117), the BCR-ABL fusion protein, and the platelet derived growth factor receptor (PDGFR)) and sorafenib (small multi-TKI, such as vascular endothelial growth factor receptor (VEGFR), PDGFR and RAF family kinases). With the later drug, that patient achieved disease stabilisation for a year, although she did experience side effects such as gait disturbance and proximal muscle weakness. With imatinib the response was short lived.²⁷

Bevacizumab either alone or combined with different chemotherapy agents has also been tested for the treatment of recurrent ependymomas, given that these tumours tend to show VEGFR expression. Green²⁸ retrospectively analysed the outcomes of eight patients treated with bevacizumab-based regimens in third-line palliative treatment. Six of these patients achieved partial response with mOS of 9.4 months. Further trials with bevacizumab in patients with recurrent spinal cord tumours are warranted.

More recently, Gilbert explored the efficacy of the treatment combination with lapatinib (a dual TKI which interrupts the HER2/neu and epidermal growth factor receptor (EGFR) pathways) and dose dense temozolomide.¹⁴ This is the first prospective clinical trial in adult recurrent low-grade and anaplastic supratentorial, infratentorial and spinal cord ependymomas. Out of 50 patients enrolled, eight with MPE were included. For MPE patients, mPFS was 9 months with a 12-month PFS rate of 75%. Treatment was well tolerated with only modest myelotoxicity and rash. This was the rationale for starting our patient on dose-dense temozolomide. The clinical benefit of the addition of lapatinib to this combination is as yet unknown and in our context these tablets are not funded for this indication by the Australian Government’s Pharmaceutical Benefits Scheme.

To the best of our knowledge, this is the first reported case of a metastatic MPE treated with a checkpoint inhibitor (tislelizumab) who achieved stabilisation of her disease for 18 months, a much longer PFS than previously reported with chemotherapy agents and/or targeted therapies for this tumour type. In recent years, immunotherapy has become a new forefront in oncology treatment. PD-1 is an inhibitory receptor expressed on the surface of T lymphocytes, B lymphocytes, activated monocytes, dendritic cells and natural-killer cells. The immune responses of T cells that express PD-1 in the tumour microenvironment can be downregulated on binding of its ligands PD- L1 or PD-L2, which are expressed by the tumourus cells. Disruption of PD-1/PD-L1 and/or cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) signalling pathway by MAbs to release the anticancer T-cell immunity has become a major breakthrough in cancer treatment. These checkpoint inhibitors, anti-PD-1 MAbs such as nivolumab or pembrolizumab—or in this case tislelizumab; anti-PD-L1 MAbs such as atezolizumab, avelumab or durvalumab; and CTLA-4 inhibitors like ipilimumab or tremelimumab, may achieve durable as well as clinically meaningful responses in patients with different tumour types, including melanoma, non-small-cell lung cancer, renal cell carcinoma, bladder cancer, Hodgkin’s lymphoma and any cancer with high microsatellite instability status.^29–31

Clinical data on the use of checkpoint inhibitors in patients with rare tumours are accumulating rapidly and several immunotherapy trials are currently open. Among these trials, the DART (dual anti-CTLA-4 and anti-PD-1 blockade in rare tumors) trial³² is worth discussing here because it is testing the combined use of ipilimumab and nivolumab for rare cancers. It is managed by the Southwest Oncology Group and the National Cancer Institute (NCI) and is dedicated to more than 30 types of rare cancers. To be eligible for the DART trial, patients must first be registered with its companion trial NCI-MATCH,³³ a basket study attempting to match patients with treatments that target a specific tumour mutation or signalling pathway. Patients with rare tumours can enrol in the DART trial if they have no treatment options in the NCI-MATCH or do not respond to treatment in that trial. Early this year, Naing et al³⁴ demonstrated in a phase II study that antiPD-1 inhibitor pembrolizumab had acceptable toxicity and antitumour effect in oncology patients with four types of advanced, hard-to-treat rare cancers. One hundred and twenty-seven patients were included in the study: squamous cell carcinoma of the skin, carcinoma of unknown primary, adrenocortical carcinoma and paraganglioma-pheochromocytoma. Overall response rate after 4 months of 3-weekly pembrolizumab was 38%. Treatment-related adverse events occurred in half of the patients, with the most common immune-related side effects being fatigue and skin rash. In the coming years, more information will be available for the treatment of rare diseases with immunotherapy agents.

Clearly future studies in ependymomas are a priority and further investigation into the role of checkpoint inhibitors, temozolomide, bevacizumab and other targeted therapies is eagerly awaited. Combined efforts from the medical community, patients and foundations such as CERN³⁵ will help to develop further research approaches and treatment opportunities for these orphan tumours.

Learning points.

• Myxopapillary ependymomas (MPEs) are rare tumours where no standard of treatment has been established to date.

• Ectopic MPE confers more aggressive behaviour (compared with other MPEs) with tendency to metastasise distally.

• Checkpoint inhibitors, in general, and tislelizumab, in particular, may represent an efficacious treatment option for metastatic MPE patients as has been illustrated with this case report.

• Clearly future studies for this orphan tumour are a priority and further prospective investigation into the role of checkpoint inhibitors for this indication is eagerly awaited.