A Molecular Update and Review of Current Trials in Paediatric Low-Grade Gliomas

Paediatric low-grade gliomas (pLGGs) are the most common primary brain tumour in children, accounting for 25–30% of brain of all brain and spinal cord tumours [1]. This is a highly heterogeneous group, classified under the World Health Organisation (WHO) histopathological classification system as grade 1 or 2 tumours of the central nervous system, with a 10-year overall survival (OS) of 90%. Most cases are sporadic, arising from a single location within the cerebellum, cerebral cortex, brainstem, spinal cord, hypothalamus, optic chiasm, or optic nerves, or in 5–10% of cases, disease is disseminated [2]. A particularly affected patient group are those with the genetic predisposition syndrome neurofibromatosis type 1 (NF1), in whom around 20% develop the hallmark optic pathway tumours [3].

Though considered benign, slow-growing lesions with excellent OS, their long-term morbidity can be significant, both from the tumour and secondary to treatment. These include neurosensory, motor, and visual deficits, cognitive and endocrine sequelae [4]. Often following a chronic disease course with multiple recurrences, several lines of treatment may be indicated over many years with a cumulative treatment burden. This poses further challenges for management, with a need to balance disease control against treatment toxicity and preservation of neurodevelopmental, visual, and endocrine functional outcomes. Thus, the goal of future therapies is to improve the quality of survival and prevent long-term sequelae of treatment. Vast progress has been made in recent years to better understand the molecular biology underlying pLGGs, with promising implications for new targeted therapeutic strategies. It remains to be seen if conventional management approaches can be replaced with less “toxic” therapies, or whether new agents will be effective when prior therapies fail. The purpose of this review article is to summarise our current understanding of the biology and molecular classification of pLGGs, their clinical implications, and to provide an overview of current management approaches, novel agents, and open trials in children and adolescent patients with pLGG.

pLGGs are typically characterised by glial or mixed glial-neuronal morphology and in the majority of cases, lack overtly high-grade features (such as necrosis and high mitotic index), hence are classed as either WHO grade 1 or 2 (formerly known as grades I and II). Histopathological entities historically classed as pLGG include pilocytic astrocytoma (WHO grade 1) and pilomyxoid astrocytoma; subependymal giant cell astrocytoma (WHO grade 1); pleomorphic xanthoastrocytoma (WHO grade 2 or 3); paediatric-type diffuse astrocytoma (WHO grade 1) and the mixed low-grade glial-neuronal tumours (LGGNTs) encompassing dysembryoplastic neuroepithelial tumours (WHO grade 1), ganglioglioma (WHO grade 1), and desmoplastic infantile astrocytoma/ganglioglioma (WHO grade 1).

It is increasingly evident that classification based purely on histopathology does not capture the heterogeneity of molecular subgroups in pLGG. Large-scale collaborative studies around the globe have documented genomic alterations in gliomas across ages, grades, and histologies. The 2016 WHO Classification of Tumours of the Central Nervous System (4th edition update) first introduced molecular parameters for tumour classification, incorporating “integrated” histological and molecular diagnoses. The most recent 2021 5th edition (WHO CNS5) summarised by Louis et al. [5], groups tumours into more biologically and molecularly defined entities, and introduces new tumour types and subtypes in paediatric patients, with implications for molecular testing. In particular, it presents a distinction between paediatric-type and adult-type diffuse gliomas, given their well-established molecular and clinical differences.

The majority of pLGGs harbour a variety of genetic abnormalities linked with aberrant intracellular signalling via the RAS/mitogen-activated protein kinase (RAS/MAPK) pathway [6]. The most frequent is the oncogenic KIAA1549-BRAF fusion, a tandem duplication involving the BRAF gene, detected in 70–80% of pilocytic astrocytomas [7], followed by germline and somatic mutations in neurofibromin 1 and the BRAF V600E mutation (present in 15–20% pLGG) [8]. Other rarer alterations of fibroblast growth factor 1 (FGFR1), FGFR2, NTRK, MYB/MYBL1, and others are also seen [9]. While diffuse low grade gliomas in adults usually harbour IDH-gene mutations and co-deletion of chromosome arms 1p/19q, the paediatric-type diffuse gliomas are usually IDH1/2 wild-type and lack 1p/19q-co-deletion; instead they may harbour varying alterations of BRAF, FGFR1, MYB, or MYBL1 genes [10]. The characteristic genes and molecular profiles associated with histopathological tumour types adapted from WHO CNS5 are summarised in Table 1 [5].

These differences in molecular profiles have been used to inform pLGG treatment planning, with separation of patients on the basis of “targetable” BRAF mutations and fusions underpinning clinical trials [11]. The identification of these molecular alterations can have important prognostic and therapeutic implications. The presence of BRAF V600E has been correlated with lower objective response rates to chemotherapy and poorer prognosis relative to patients with BRAF V600 wild-type pLGG. The co-existence of CDKN2A deletion has been shown to contribute independently to poor outcome in BRAF V600E pLGG [8]. The WHO CNS5 classification recognises that in some gliomas, such as pilocytic astrocytomas with piloid features, there may be concomitant mutations (e.g., CDKN2A/B, ATRX) in addition to BRAF alterations, with prognostic implications [12].

The use of molecular biomarkers in tumours of the central nervous system is complex: in some cases, a defining molecular feature consistently characterizes the tumour; in others, molecular features are not required but may support tumour classification; and in others, molecular approaches are rarely used for diagnosis. Current nosological classification is therefore mixed and represents the current state of the field. This is seen in the case of the newly termed “paediatric-type diffuse low-grade gliomas” (see Table 1), whereby some tumour types encompass several subtypes sharing one molecular feature, and other types are specifically defined by a single feature.

This multi-layered classification system of biologic subgroups in pLGG, integrating distinct molecular and histological features is crucial to further our clinical understanding of these heterogeneous tumours. It provides prognostic information to better direct treatment planning in specific patient groups that may benefit from targeted approaches. Yet it also poses challenges for clinicians, whereby different molecularly targeted therapies may be required for each tumour type, hence the need for rational clinical trials directed towards rare tumour subtypes [13].

Surgery is the mainstay treatment of choice for pLGGs, where gross total resection is possible with minimal morbidity. This is particularly the case for hemispheric or cerebellar lesions, with 5-year progression-free survival (PFS) of 75–100% [14]. For other lesions with closer proximity to vital structures that preclude total resection (e.g., optic pathway, hypothalamus, or pituitary), a partial resection for mechanical decompression or biopsy is usually undertaken, with the exception of children with NF1 and an optic pathway tumour [15]. As such, more than 40% of children with pLGG need additional non-surgical treatments, either for disease that is threatening neurological function, or progressive disease post-surgery.

The use of adjuvant therapy is determined by age, tumour location, and presence of NF1. For children with NF1 and optic pathway glioma, an observational approach may be appropriate in the first instance, as these tumours can remain indolent in up to 50% of cases. Radiotherapy, though highly effective, with 5-year PFS estimates over 80%, must be considered carefully in view of risks of significant late neurotoxicity affecting growth, endocrine, and neurocognitive development, as well as subsequent stroke related to vasculopathy, and second malignancy [16, 17]. Treatment-related morbidities have particularly been observed in NF1 patients, in whom irradiation is generally avoided [18, 19]. The age of the child, tumour location, radiation dose, and volume should be considered. Increasingly conformal focal approaches, such as proton beam radiotherapy, should be considered to reduce treatment volumes and associated toxicity by minimising irradiation to normal brain. The Children’s Oncology Group (COG) ACNS0221 study found that children with unresectable progressive/recurrent LGG treated with conformal radiation therapy (with a reduced clinical target volume margin of 5 mm) had a 5-year PFS of 71% and did not incur high rates of marginal relapse [20]. Proton therapy, which can reduce the radiation dose to uninvolved brain tissue, has been observed to maintain disease control (5-year PFS 84%, OS 92%) and reduce acute toxicity in children with non-metastatic LGG when compared to photon series [21]. A small series of children with LGG treated with proton therapy between 1995 and 2007 achieved excellent clinical outcomes (8-year PFS of 82.8%) and improved visual acuity in 83.3% of cases at risk for radiation-induced injury to the optic pathways. However, significant decline in neurocognitive outcomes were observed in young children (<7 years) and those with significant doses to the left temporal lobe and hippocampus [22]. The HALGG study (NCT04065776) is a phase 2 study which will investigate the feasibility of reducing radiotherapy doses to the hippocampi using proton therapy in midline and suprasellar LGGs in order to reduce late deficits in memory. The results of another phase 2 trial of proton radiotherapy in children with brain tumours are awaited (NCT01288235) and the long-term toxicity in larger numbers remains to be evaluated.

Chemotherapy is the usual treatment in newly diagnosed patients (typically under 7 years old) with unresectable or partially resected progressive disease and/or neurological/visual impairment, in NF1 patients in whom radiotherapy is to be avoided or delayed, and in symptomatic metastatic disease. Results of chemotherapy trials have demonstrated very good OS of between 70 and 95% but a low 5-year PFS of around 45%. The combination of vincristine and carboplatin remains the current standard first-line treatment in both Europe and the USA. The first SIOP LGG study, a multi-national single-arm pilot study, demonstrated that a 12-month chemotherapy regimen with vincristine and carboplatin (VC) could obviate the need for radiotherapy, previously the mainstay of treatment for incompletely resected pLGG. NF1 patients achieved prolonged tumour stabilisation with chemotherapy, compared to those without NF1 [23]. The more recent SIOP LGG 2004 study, a randomised first-line chemotherapy strategy for non-NF1 patients with progressive/symptomatic unresectable tumours, confirmed high rates of non-progression (93% at 24 weeks) using this chemotherapy combination for 18 months, and concluded no additional benefit in further intensification with etoposide. The 5-year PFS and OS with VC were 46% and 89%. However, there is significant neurotoxicity associated with this regimen and ongoing uncertainty regarding optimal duration of treatment. The investigators identified a need for less toxic and more effective treatments, especially with regards to improving visual outcomes for visual pathway gliomas [24].

Disease progression or treatment-related toxicity means that many patients require alternative and/or multiple lines of treatment. Other current options include single-agent vinblastine, irinotecan, and bevacizumab, and TPCV regimens (thioguanine, procarbazine, lomustine, and vincristine). Trials evaluating these agents have been performed with variable results. A trial of weekly vinblastine in chemotherapy-naïve children with unresectable or progressive LGG demonstrated a 5-year PFS of 53.2%, with more favourable outcomes for NF1 patients (5-year PFS 85.1% in patients with NF1 vs. 42% without NF1) and excellent tolerability [25]. The combination of irinotecan and the VEGF-targeted antibody bevacizumab has been explored in small studies of recurrent/progressive LGG, with reported 2-year PFS of 47.8% as well as associated recovery of vision in patients with optic pathway glioma [26, 27]. The A9952 COG trial compared carboplatin and vincristine (CV) with thioguanine, procarbazine, lomustine, and vincristine (TPCV) in children less than 10 years old with newly diagnosed LGGs (without NF1). The 5-year EFS rates were 39% ± 4% for CV and 52% ± 5% for TPCV, a difference which did not reach statistical significance (p = 0.1) [28]. Despite reasonable efficacy of both regimens, the potential for greater toxicity using TPCV has precluded its usage as first-line therapy. Vinorelbine, a semi-synthetic vinca-alkaloid, has been investigated as an agent with potentially reduced neurotoxicity compared to related agents. A study of 23 patients with optic pathway glioma receiving single agent vinorelbine reported a low toxicity profile and 3- and 5-year PFS of 64% ± 19% and 37% ± 20% [29]. Another chemotherapy agent utilised in relapse is temozolomide, however robust data supporting its use in pLGG is scarce. In a phase 2 study, 41% of patients with low-grade astrocytoma treated with temozolomide achieved disease stability through 12 courses, however only 1/21 patients achieved an objective partial response and temozolomide did not demonstrate activity according to the “protocol definition of efficacy” [30]. In a further series of 13 children with LGG treated with temozolomide, the reported 3-year EFS was 57%, with the advantage of oral administration and good tolerability however 2 patients progressed on treatment in this small cohort [31]. Therefore, the need for more effective therapies and larger studies is ongoing.

One challenge in interpreting the results of previous studies and comparing with outcomes of newer agents, has been a lack of standardised response criteria in paediatric clinical trials. For this reason, the Radiologic Assessment in Pediatric Neuro-Oncology (RAPNO) working group, consisting of international experts in paediatric neuro-oncology, radiology, and neurosurgery developed consensus recommendations for the response assessment of pLGG. As well as recommendations for imaging response assessment, they include guidelines for assessing visual and functional outcomes. These recommendations aim to standardise response definitions for pLGG and will be evaluated and validated in prospective clinical trials [32].

There are several upcoming trials investigating chemotherapy agents or multi-agent combinations, both upfront and in progressive disease, at various stages of completion. A randomised phase 2 multicentre study will investigate the addition of bevacizumab to vinblastine in chemotherapy-naïve patients with progressive pLGG (NCT02840409). As well as monitoring efficacy, the study will include visual outcome measures (in patients with optic pathway glioma), cognitive function, and quality of life assessment. Furthermore, the planned LOGGIC-FIREFLY-2 trial (EudraCT 2022-001363-27) is a phase 3 randomised multicentre global study evaluating the efficacy, safety, and tolerability of the targeted treatment Day101 (tovorafenib), a pan-RAF inhibitor directed at MAPK activation, versus standard of care chemotherapy (carboplatin and vincristine or single agent vinblastine) in patients <25 years old with LGG harbouring a RAF alteration and requiring first-line systemic therapy. It will mandate collection of fresh frozen tumour tissue for molecular characterisation in all patients. In addition to tumour response rates and PFS as outcome measures, the trial will include neurological and visual function as endpoints [33].

In conclusion, outcomes for children with pLGG using chemotherapy remain suboptimal, while the long-term effects of radiotherapy make this a “last resort” option in younger children. There is ongoing need to not only minimize treatment burden and treatment-related toxicity but to improve quality of survival and monitor functional outcomes in future studies. The inclusion of biologic information in current trials will provide additional information that may shed light on reasons for therapy successes and failures. In addition, the use of molecularly targeted therapies in isolation or in combination with chemotherapy holds promise as a future management strategy.

Up to 50% of pLGG will progress after conventional chemotherapy. Therapies targeting common genetic aberrations in the RAS/MAPK pathway, thereby disrupting tumour growth, have been the focus of many recent studies and offer new therapeutic possibilities. Clinical trials have focussed on their use in the recurrent or refractory setting, however their potential as upfront therapy is to be determined.

Inhibition of the MAPK enzymes MEK1 and/or MEK2 has emerged as a promising treatment strategy, with a number of MEK inhibitors currently under investigation at various stages in clinical trials. A phase 1 study demonstrated that selumetinib, a MEK1/2 inhibitor, induced tumour response or stability in patients with progressive/recurrent pLGG. The 2-year PFS at the recommended phase 2 dose was 69 ± SE 9.8% [34]. An ongoing phase 2 study of selumetinib is evaluating its use in refractory/progressive pLGG, across six histologically and biologically defined strata. It was provided as oral capsules taken twice daily in 28-day courses for up to 26 courses. Published reports to date have been positive, with a sustained partial response seen in 35% of patients with pilocytic astrocytoma harbouring a BRAF alteration (KIAA1549-BRAF fusion or BRAF V600E mutation) with a 2-year PFS of 70%. Of patients with NF1, 40% sustained partial response with 2-year PFS of 96% [35]. Another subgroup of patients with sporadic and recurrent/progressive optic pathway and hypothalamic gliomas achieved partial response in 24% and stable disease in 56%, with 2-year PFS of 78%. Where evaluable, visual acuity was stable in 68% and improved in 21% of cases [36]. The most frequent adverse events were elevated creatinine phosphokinase, maculopapular rash, nausea, and diarrhoea. Given these positive results, selumetinib shows promise as a tolerable alternative to standard chemotherapy in these subgroups, associated with response, and disease stability. It is currently being investigated in two Children’s Oncology Group randomised phase 3 studies as front-line therapy in newly diagnosed pLGG, in children with and without NF1 (ACNS1831 and ACNS1833). The ACN1833 trial will exclude patients with BRAF V600E mutations.

Other MEK inhibitors currently under investigation include trametinib, cobimetinib, and binimetinib, though published experience is limited. A case series of six children with pilocytic astrocytoma treated with trametinib for 4–20 months following progression after conventional treatment observed two partial and three minor responses [37]. Early clinical trials using trametinib (in patients with BRAF fusion) and binimetinib have reported partial responses and stable disease in children with refractory/recurrent pLGG [38, 39]. Treatments appear to be relatively well tolerated, with common adverse events among agents including diarrhoea, elevated creatinine kinase, and rash. The iMATRIXcobi study (NCT02639546), a phase 1/2 dose-finding study of cobimetinib in paediatric patients with refractory/relapsed tumours and known or potential RAS/RAF/MEK/ERK pathway activation demonstrated partial responses in only 3/32 patients with LGG, with an overall response rate (ORR) of 9.4% in this cohort [40]. Current open trials include the TRAM-01 study (NCT03363217), a phase 2 study investigating the use of trametinib as a single agent in progressive/refractory tumours with MAPK/ERK pathway activation, and will further determine response rates. Eligible patients for given subgroups will receive oral trametinib daily, at full dose for a total of 18 cycles of 28 days [41]. A phase 1/2 study investigating binimetinib or MEK162™ (NCT02285439) in progressive/refractory pLGG is still active. Another ongoing study is a randomised and controlled phase 2 study comparing daily oral trametinib with weekly vinblastine (18 courses of each) as first-line therapy in sporadic BRAF wild-type (non-NF1) pLGG (NCT05180825). It is yet to be determined which of these agents will be the most efficacious and best tolerated in pLGG.

Direct inhibitors of BRAF (dabrafenib, vemurafenib) are another important targeted agent under investigation. It has been observed that patients with BRAF mutations have poorer prognoses relative to those with BRAF wild-type, NF1, or BRAF fusion when treated with conventional chemotherapy [8]. Dramatic clinical and radiological responses to BRAF inhibitors have been observed in cases harbouring BRAF V600E mutation [42‒44]. Dabrafenib is a potent and selective inhibitor of the V600-mutant form of the BRAF kinase. Initial results from a multicentre phase 1 clinical study using dabrafenib reported an impressive ORR of 41% in patients with recurrent or progressive BRAFV600E mutant pLGG. Median duration of response was 11 months [45]. Reported toxicities included low-grade pyrexia, vomiting, fatigue, headache, and rash. A follow-up trial (NCT01677741) exploring the activity and safety of dabrafenib in a similar cohort reported a consistent ORR of 44% in 32 patients, with 1-year PFS of 85%; the most common adverse event was fatigue [46]. Thus, dabrafenib has demonstrated meaningful clinical activity and acceptable tolerability and holds potential to transform the outcomes for children with pLGG and BRAF V600E mutation.

A recent phase 1/2 study (NCT02124772) demonstrated that trametinib, used as a monotherapy or combined with dabrafenib, demonstrated a manageable safety profile and good tolerability in patients with relapsed/refractory BRAF V600 mutant pLGG [47]. The activity of dabrafenib in combination with trametinib was compared to that of standard of care chemotherapy (carboplatin and vincristine) in a recent global, open-label randomised phase 2 study (NCT02684058) in paediatric patients with BRAF V600E mutant pLGG. The study assessed ORRs and PFS in chemotherapy-naïve patients who were unamenable to surgery and those with progressive disease after surgery. The independently assessed ORR (CR + PR) was 47% (95% CI, 35–59%) with dabrafenib and trametinib compared to 11% (95% CI, 3–25%) with carboplatin and vincristine (p < 0.001; odds ratio, 7.2 [95% CI, 2.3–22.4]). The 1-year PFS rates were 67% versus 26%. Patients in the dabrafenib and trametinib arm also incurred less grade ≥3 adverse events and fewer treatment discontinuations, suggesting this as a tolerable and promising alternative first-line systemic therapy in patients with BRAF V600E mutant LGG [48, 49]. Larger scale studies will be needed to further determine the effectiveness, optimal duration of therapy, and long-term toxicity. Current active trials include a phase 1 study investigating the safety and optimal dosage of vemurafenib in recurrent/refractory BRAF V600E mutant gliomas (NCT01748149).

Though effective in BRAF V600E tumours, first-generation BRAF inhibitors exhibit a paradoxical effect in KIAA1549-BRAF fusion or BRAF wild-type tumours, through activation of RAS/MAPK signalling. This was observed in a study investigating sorafenib, where 9 of 11 patients had rapid and unexpected radiological disease progression, causing the trial to be terminated early [50]. To overcome this, second-generation BRAF inhibitors were devised, to inhibit BRAF without causing RAS/MAPK activation. A single patient study using one such product, TAK-580™ (now known as DAY101™, a pan-RAF inhibitor), demonstrated stable disease after 33 weekly cycles and good tolerability [51]. The PNOC014 study further explored its potential in a phase 1 setting, with promising preliminary results: of eight relapsed pLGG patients with RAF fusions, 2 patients achieved a complete radiological response, three had partial response, and two achieved prolonged stable disease. Thus, this oral agent shows promise in pLGG patients harbouring known BRAF alterations and its use will be investigated further in the recurrent/progressive disease setting in the currently open phase 2 multicentre FIREFLY-1 study for patients with pLGG (NCT04775485).

While MEK and BRAF inhibitors show potential as emerging therapies, the exploration of new treatments or combinations that work outwith the MAPK pathway are warranted, given the complex molecular landscape of pLGG. Studies are investigating the use of these targeted agents in combination with other therapies, such as the PNOC021 trial, a dose-finding study investigating use of trametinib when combined with mechanistic target of rapamycin (mTOR) inhibitor everolimus, in recurrent pLGG (NCT04485559). mTOR is a protein kinase regulating cell growth, survival, and metabolism. mTOR inhibition may have an additional role in preventing tumour growth and metastases. Indeed, mTOR pathway activation is implicated in the formation of subependymal giant cell astrocytoma associated with tuberous sclerosis complex, and mTOR inhibitors including everolimus are an established treatment option in these patients [52]. Hydroxychloroquine, an autophagy inhibitor, is another agent being evaluated in combination studies (NCT04201457). It may play an additional role in preventing tumour growth when combined with targeted therapies. The FGFR receptors are another potential target for novel therapies, playing a key role in signal transduction via activation of the intramembranous tyrosine kinase domain. FGFR1 alterations comprise the second most commonly altered gene in pLGG, occurring in 5–10% of patients [53]. Small molecular inhibitors of FGFR have been developed, demonstrating encouraging clinical activity in tumours harbouring FGFR alterations in adults in phase 1 study. Zoligratinib (Debio1347™), an oral FGFR 1–3 inhibitor, was observed to induce objective imaging responses in three paediatric patients with refractory FGFR-altered LGG and functional response, with improved vision in 1 patient. Debio1347™ demonstrated tolerable toxicity with commonest adverse effects being hyperphosphatemia, increased ALT and hypoalbuminemia [54]. A phase 2 study evaluating the efficacy of erdafitinib, a potent FGFR 1–4 inhibitor, in adult and paediatric patients with advanced solid tumours (irrespective of histology) and FGFR alterations is ongoing (NCT04083976). Other novel agents examined under phase 1 study include fluvastatin and celecoxib, shown to have anticancer activity in preclinical and clinical reports. Celecoxib is a Cox-2 inhibitor with anti-angiogenic properties, while statins increase the sensitivity of gliomas to anti-tumour agents. This multicentre phase 1 included 10 patients with relapsed/refractory pLGG, of whom four maintained stable disease beyond 12 cycles of treatment with limited toxicity, suggesting preliminary activity. The researchers suggest a possible role of this therapeutic combination in a maintenance setting, given its good tolerance and low cost for children living in low- and middle-income countries [55]. The immunostimulant Poly-ICLC, a synthetic double-stranded RNA complex, may induce innate immune responses to kill tumour cells and is the subject of phase 2 study in NF1 patients with progressive pLGG (NCT04544007). Results from such trials will inevitably inform larger studies and the use of these novel agents in pLGG. These studies highlight the importance of adequate molecular characterisation of pLGG for guiding management decisions when considering targeted and novel therapies, and enrolling patients in clinical trials where available and appropriate.

pLGG is a common tumour of the central nervous system which can be associated with a chronic disease course, significant treatment burden, and morbidity. There is a need for treatment strategies that provide lasting tumour control and better quality of survival through minimising toxicity and protecting against neurological, cognitive, and endocrine deficits. Integrated histological and molecular characterisation of these tumours has enhanced our clinical understanding, serving as a predictive and prognostic tool that may enable better risk stratification of patients, and consideration for targeted therapies in the future. There may be a role for MAPK inhibitors and other targeted therapies in front-line therapy – as single agents or in combination with chemotherapy – and other agents such as bevacizumab, which may not only be able to control disease but also improve or protect neurological and visual integrity. Adequate molecular characterisation of pLGG is therefore crucial in today’s clinical trials, as well as evaluation of neurological, visual, and quality of life outcomes to truly understand the benefits and costs of future treatments.

The authors declare they have no competing interests.

The authors did not receive any funding for this work.

Sarah Al-Jilaihawi and Stephen Lowis conceptualised the article and reviewed the literature. Sarah Al-Jilaihawi wrote the main text of the article and prepared tables. Both authors read, edited, and approved the final manuscript.