Medulloblastoma in the Modern Era: Review of Contemporary Trials, Molecular Advances, and Updates in Management

Abstract

Critical discoveries over the past two decades have transformed our understanding of medulloblastoma from a single entity into a clinically and biologically heterogeneous disease composed of at least four molecularly distinct subgroups with prognostically and therapeutically relevant genomic signatures. Contemporary clinical trials also have provided valuable insight guiding appropriate treatment strategies. Despite therapeutic and biological advances, medulloblastoma patients across the age spectrum experience tumor- and treatment-related morbidity and mortality. Using an updated risk stratification approach integrating both clinical and molecular features, ongoing research seeks to (1) cautiously reduce therapy and mitigate toxicity in low-average risk patients, and (2) thoughtfully intensify treatment with incorporation of novel, biologically guided agents for patients with high-risk disease. Herein, we review important historical and contemporary studies, discuss management updates, and summarize current knowledge of the biological landscape across unique pediatric, infant, young adult, and relapsed medulloblastoma populations.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13311-022-01273-0.

Introduction

Medulloblastoma, an embryonal tumor arising from the cerebellum with potential for neuroaxis dissemination, is the most common central nervous system (CNS) malignancy in childhood [1–3]. With multimodal treatment including maximal surgical resection, craniospinal irradiation (CSI), and adjuvant platinum- and alkylator-based chemotherapy, cure is achieved in approximately 70% of all patients diagnosed at ≥ 3 years of age [4–8]. Recent discoveries have provided valuable insight into substantial clinical and biological heterogeneity within this disease—encompassing at least four molecularly distinct subgroups (Wingless [WNT], Sonic Hedgehog [SHH], group 3, group 4), each with prognostically relevant genomic, methylomic, and transcriptomic signatures [9–14]—for which future therapy must be tailored [1, 3]. Historical risk stratification has been defined largely by the presence or absence of diffuse anaplastic histology, metastatic status, and extent of resection. Given the high incidence of treatment-related neurocognitive, endocrine, and auditory sequelae among survivors of medulloblastoma [15–19], contemporary clinical trials for patients with average-risk disease seek to mitigate toxicity while maintaining excellent 5-year event-free survival (EFS) and overall survival (OS) in excess of 80% [4, 7, 8, 20]. Outcomes remain inferior for patients with high-risk disease (60–70% 5-year EFS and OS), motivating investigation of treatment intensification strategies [5–8], with a critical need for incorporation of novel, biologically guided agents. Infants and adults with medulloblastoma represent distinct patient populations with unique clinical, molecular, and toxicity considerations [21–28]. Finally, there is a paucity of curative options for relapsed medulloblastoma [29–31], driving ongoing efforts to augment upfront regimens for high-risk patients and identify new effective approaches at recurrence.

This article aims to (a) summarize key historical and recently published contemporary studies which have transformed the treatment paradigm for patients across the age spectrum with medulloblastoma; (b) review the impact of molecular advances on risk stratification, upcoming clinical trial design, and the relentless search for novel targeted therapies; and (c) discuss innovative techniques for toxicity reduction, early detection of minimal residual disease, and expanding treatment modalities.

Historical and Contemporary Clinical Trial Review

Pediatric and Adolescent Patients with Newly Diagnosed Medulloblastoma (Tables 1 and 2)

Table 1.

Key historical and contemporary clinical trials for pediatric and adolescent patients (> 3–5 years of age) with newly diagnosed average-risk medulloblastoma

Study	Years conducted	Number of evaluable patients	Age at diagnosis	Treatment investigated	Event-free survival (EFS)	Overall survival (OS)
SIOP PNET-3 [36]	1992–2000	179	3–16 years	Randomized trial evaluating addition of neoadjuvant chemotherapy (4 cycles of alternating carboplatin + etoposide and cyclophosphamide + etoposide) to craniospinal irradiation (CSI) [35 Gy + 55 Gy posterior fossa (PF) boost] vs. CSI alone	5-year EFS: 74.2% (chemo) vs. 59.8% (no chemo) (p = 0.0366*)	5-year OS: 76.7% (chemo) vs. 64.9% (no chemo) (p = 0.0928)
COG-A9961 [20]	1996–2000	313	3–21 years	Randomized trial evaluating adjuvant cyclophosphamide- vs. CCNU (lomustine)-containing maintenance chemotherapy regimens (both also with cisplatin + vincristine), following reduced dose CSI in all patients (23.4 Gy CSI + 55.8 Gy PF boost with weekly vincristine)	5-year EFS: 82% (CCNU) vs. 80% (cyclo-phosphamide) (p = 0.40)	Not reported
SJMB96 [7] average-risk arm	1996–2003	86	3–21 years	Non-randomized trial, in which average-risk patients received 23.4 Gy CSI + 55.8 Gy PF boost, followed by shortened, dose-intensive chemotherapy with autologous stem cell rescue (4 cycles: high-dose cyclophosphamide + cisplatin + vincristine)	5-year EFS: 83%	5-year OS: 85%
HIT-SIOP PNET 4 [41]	2001–2006	340	4–21 years	Randomized trial among average-risk patients, evaluating hyperfractionated CSI (36 Gy CSI + 60 Gy PF boost + 58 Gy tumor bed boost) vs. conventional fractionated CSI (23.4 Gy CSI + 54 Gy PF boost), followed by 8 cycles of maintenance chemotherapy (cisplatin + CCNU + vincristine)	5-year EFS: 78% (hyper-fractionated) vs. 77% (conventional)	5-year OS: 85% (hyper-fractionated) vs. 87% (conventional)
SJMB03 [8] average-risk arm	2003–2013	227	3–21 years	Non-randomized trial, in which average-risk patients received 23.4 Gy CSI + 55.8 Gy PF boost, followed by shortened, dose-intensive chemotherapy with autologous stem cell rescue (4 cycles: high-dose cyclophosphamide + cisplatin + vincristine)	5-year EFS: 82.3%	5-year OS: 88.0%
COG-ACNS0331 [4]	2004–2016	464 (comparing PF vs. involved field boost), and 226 comparing CSI dosing	3–21 years	Randomized trial evaluating (1) reduction in boost volume (54 Gy) from PF to limited tumor bed/ involved field (IFRT) across all patients, and (2) reduction in CSI dosing from standard dose 23.4 Gy (SDCSI) to low-dose 18 Gy (LDCSI) for patients aged 3–7 years; all patients received concurrent weekly vincristine during irradiation, followed by maintenance chemotherapy containing 6 CCNU + cisplatin + vincristine cycles and 3 cyclophosphamide + vincristine cycles	5-year EFS: Boost volume: 82.5% (IFRT) vs. 80.5% (PF) (p = 0.44) CSI dosing: 71.4% (LDCSI) vs. 82.9% (SDCSI) (p = 0.028*)	5-year OS: Boost volume: 84.6% (IFRT) vs. 85.2% (PF) (p = 0.44) CSI dosing: 77.5% (LDCSI) vs. 85.6% (SDCSI) (p = 0.049*)

Table 2.

Key contemporary clinical trials for pediatric and adolescent patients (> 3–5 years of age) with newly diagnosed high-risk medulloblastoma

Study	Years conducted	Number of evaluable patients	Age at diagnosis	Treatment investigated	Event-free survival (EFS)	Overall survival (OS)
SJMB96 [7] High-risk arm	1996–2003	48	3–21 years	Non-randomized trial, in which high-risk patients received 6 weeks of pre-radiotherapy topotecan (31 patients, others proceeded to CSI), followed by 36–39.6 Gy CSI (36 Gy for M0-1, 39.6 Gy for M2-3) + 55.8–59.4 PF boost, followed by shortened, dose-intensive chemotherapy with autologous stem cell rescue (4 cycles: high-dose cyclophosphamide + cisplatin + vincristine)	5-year EFS: 70%	5-year OS: 70%
HIT2000 [6]	2001–2007	123	4–21 years	Non-randomized trial, in which metastatic patients received 2 cycles of induction chemotherapy (systemic cyclophosphamide + vincristine + methotrexate + carboplatin + etoposide, and intraventricular methotrexate), followed by hyperfractionated CSI (40 Gy CSI + 20 Gy PF + 8–12 Gy tumor site), followed by 4 cycles of maintenance chemotherapy (cisplatin + CCNU + vincristine)	5-year EFS: 62% Comparison to HIT’91(40) patients of same age group: 5-year EFS: 47% (p = 0.067)	5-year OS: 73% Comparison to HIT’91(40) patients of same age group: 5-year EFS: 55% (p = 0.007*)
SJMB03 [8] High-risk arm	2003–2013	103	3–21 years	Non-randomized trial, in which high-risk patients received 36–39.6 Gy CSI (36 Gy for M0-1, 39.6 Gy for M2-3) + 55.8–59.4 PF boost, followed by shortened, dose-intensive chemotherapy with autologous stem cell rescue (4 cycles: high-dose cyclophosphamide + cisplatin + vincristine)	5-year EFS: 56.7%	5-year OS: 69.5%
COG- ACNS0332 [5]	2007–2018	261	3–21 years	Randomized trial evaluating (1) carboplatin given concurrently with radiotherapy and (2) isotretinoin given for 12 cycles during and following maintenance therapy; all patients received 36 Gy CSI + 55.8 Gy PF boost with concurrent weekly vincristine; followed by 6 cycles of cisplatin + cyclophosphamide + vincristine	5-year EFS: Carboplatin randomization: 66.4% (carboplatin) vs. 59.2% (No carboplatin) (p = 0.11) Group 3: 73.2% (carboplatin) vs. 53.7% (no carboplatin) (p = 0.047*) Isotretinoin randomization: 68.6% (isotretinoin) vs. 67.8% (no isotretinoin) (discontinued at interim futility analysis)	5-year OS: Carboplatin randomization: 77.6% (carboplatin) vs. 68.8% (no carboplatin) (p = 0.28) Group 3: 82.8% (carboplatin) vs. 63.7% (No carboplatin) (p = 0.06)

Current treatment for medulloblastoma in the pediatric and adolescent patient population (> 3–5 years of age at diagnosis) uniformly includes the multimodal combination of maximal safe surgical resection, CSI, and chemotherapy. Radiotherapy is delivered to the entire neuroaxis for local control, given medulloblastoma’s predilection for leptomeningeal dissemination, even without metastatic disease detected on spine imaging or by cerebrospinal fluid (CSF) cytology, as well as relatively high radiosensitivity [32]. Therefore, post-operative CSI (36 Gy) with a cumulative dose of 54–55.8 Gy to the posterior fossa has been a mainstay of treatment beginning several decades ago [33]; gross spinal disease requires 45 Gy and diffuse (radiographically visible) leptomeningeal disease requires 39.6 Gy. Early studies demonstrated improved outcomes with the addition of adjuvant or neoadjuvant chemotherapy [34–36]. In the International Society of Paediatric Oncology (SIOP) PNET-3 trial, patients with non-metastatic medulloblastoma randomized to pre-radiotherapy chemotherapy had significantly increased 5-year EFS, compared to patients who received radiotherapy alone (74% vs. 60%) [36], facilitating subsequent studies investigating dose-reduced CSI with chemotherapy in the average-risk setting.

Average-Risk Patients, Radiotherapy De-intensification (Table 1)

With the goal of mitigating late sequelae, Packer et al. investigated reduction in CSI dosing from 36 to 23.4 Gy for patients with localized medulloblastoma, with excellent outcomes preserved [20, 37]. In the Children’s Oncology Group (COG) A9961 trial, all patients with average-risk medulloblastoma received 23.4-Gy CSI, with a 55.8-Gy posterior fossa boost, followed by maintenance chemotherapy, with randomization to CCNU (lomustine)- or cyclophosphamide-containing regimens (both combined with cisplatin and vincristine) [20]. Five-year EFS for the entire cohort was > 80% (with no significant difference observed between CCNU- and cyclophosphamide-based adjuvant treatments), supporting feasibility of reduced-dose radiotherapy in this population without compromising outcomes.

Further attempts at decreasing both the dosing and volume of radiotherapy were studied in the COG ACNS0331 trial [4], in which patients with average-risk medulloblastoma (non-metastatic, ≤ 1.5 cm² of residual tumor post-operatively, and no anaplastic histology) were randomized to receive a radiation boost (54 Gy) to either the entire posterior fossa versus the more limited tumor bed/involved field [4]. For patients aged 3–7 years, there was an additional randomization investigating CSI dosing of 23.4 Gy vs. 18 Gy. Results demonstrated safety and efficacy of reduced radiotherapy boost volume, without inferior outcomes (and without increased posterior fossa failures) among patients treated with a tumor bed/involved field boost; however, reduction in CSI dosing to 18 Gy was associated with poorer EFS and OS when assessed across the entire cohort. Retrospective methylation and genomic profiling corroborated findings from other studies of inferior outcomes in group 3 patients and superior survival in the WNT subgroup [4, 9, 10]. Notably, failures of reduced dose CSI were driven by group 4 patients, with too few patients with WNT-activated disease to draw conclusions.

Importantly, retrospective central review of neuro-imaging and histopathology of patients enrolled on A9961 and ACNS0331 revealed a relatively high rate (approximately 10% in each) of discordant staging and/or insufficient imaging, inappropriately allowing patients with high-risk disease features (including metastatic disease) to enroll on therapy reduction trials [4, 20]; current and future prospective studies at COG mandate central radiographic, histologic, and molecular review to confirm eligibility.

The St Jude Medulloblastoma (SJMB)-96 and -03 protocols investigated risk-adapted radiotherapy, with non-randomized 23.4-Gy CSI dosing for all average-risk patients (and 36–39.6 Gy for high-risk patients, described below) [7, 8], followed by short, dose-intensive alkylator-based chemotherapy; the same average-risk treatment backbone was studied in both SJMB96 and SJMB03, with biology- and molecular-focused objectives incorporated in the latter. Among average-risk patients in each of the studies, 5-year EFS was 83% and 82.3%, respectively, again demonstrating efficacy of 23.4 Gy CSI dosing, as well as similarly favorable outcomes achieved (compared to the above contemporary trials) with a shortened schedule of dose-intense chemotherapy with autologous stem cell rescue, including higher cyclophosphamide, yet decreased cumulative platinum and vincristine dosing [7, 8]. Methylation-based molecular subgrouping was prognostic, including among average-risk patients, with superior survival in WNT medulloblastoma, poorer outcomes in group 3 (especially with concurrent MYC amplification), and dichotomous results among SHH patients; ERBB2 status was not predictive of survival [8].

High-Risk Patients, Therapy Intensification (Table 2)

Recent trials have evaluated therapy intensification for high-risk medulloblastoma (traditionally defined by the presence of metastases, post-operative residual tumor > 1.5 cm², and/or diffuse anaplasia) in efforts to improve survival in this patient population. ACNS0332 was a randomized phase 3 trial assessing (1) use of carboplatin administered concurrently with CSI (36 Gy), based on preclinical and clinical data suggesting radiosensitization, especially in metastatic medulloblastoma [38] and (2) incorporation of the pro-apoptotic agent isotretinoin during and following maintenance chemotherapy [5]. While results did not demonstrate a clear benefit of concurrent carboplatin with CSI when evaluated across the entire cohort, molecularly stratified analyses revealed significantly greater EFS and OS (both by approximately 20%) among high-risk group 3 patients who received chemoradiotherapy with concurrent carboplatin [5]. Isotretinoin did not improve outcomes; this randomization was discontinued following an interim futility analysis.

In the prospective multicenter HIT (Therapieprotokoll für Kinder und Jugendliche mit Hirntumoren [Treatment Protocol for Children and Adolescents With Brain Tumors] 2000 trial [6], patients with metastatic medulloblastoma were treated with an intensified regimen, expanding upon the precursor HIT’91 “sandwich strategy” backbone, consisting of post-operative chemotherapy followed by CSI [39, 40]. Specifically, HIT2000 investigated augmentation of the HIT’91 regimen through intensified neoadjuvant chemotherapy (both systemic and intraventricular [HIT-SKK]), hyperfractionated CSI (40 Gy), and addition of four adjuvant maintenance chemotherapy cycles, with results demonstrating improved 5-year EFS when compared to HIT’91 (62% vs. 47%), and comparable outcomes to other high-risk medulloblastoma contemporary trials [6]. The authors hypothesize that the enhanced efficacy of HIT2000 compared to HIT’91 is less likely due to hyperfractionation, given that hyperfractionated radiotherapy was not shown to be superior to conventional fractionated radiotherapy in a large randomized European trial among average-risk medulloblastoma patients, HIT-SIOP PNET 4 [41], and therefore may instead be explained by intensified neoadjuvant and adjuvant chemotherapy, as well as improved supportive care.

SJMB96- and SJMB03-treated patients with high-risk medulloblastoma with 36- to 39.6-Gy CSI (36 Gy for M0-1 patients, 39.6 Gy for M2-3 patients), followed by the cyclophosphamide-based dose-intensive adjuvant chemotherapy regimen described above; in SJMB96, 31 of 48 high-risk patients also received neoadjuvant topotecan, with no significant improvement in EFS [7, 8]. Five-year EFS outcomes for high-risk medulloblastoma patients within the range of 60–70% were demonstrated in both SJMB trials, similar to ACNS0332 and HIT2000, suggesting comparable efficacy across all modern regimens. Importantly, SJMB03, ACNS0332, and HIT2000 all performed biology-based analyses, confirming the prognostic relevance of molecular subgrouping as well as further subtype-specific genomic signatures when assessed across medulloblastoma patients treated uniformly [5, 6, 8], which will inform risk stratification in future clinical trials, discussed further below.

Molecular Landscape of Medulloblastoma

Over the past decade, landmark methylomic and transcriptomic analyses using high-throughput genomic and proteomic methods have transformed our understanding of medulloblastoma from a single entity to a biologically heterogeneous disease [9–14]. These studies, performed in large cohorts and validated by independent investigators across the world, led to the current consensus that medulloblastoma comprises four distinct molecular subgroups: WNT-activated, SHH-activated, group 3, and group 4 [9–14], with even more granular subtypes, defined by characteristic DNA methylation profiles, copy number variation, somatic alterations, and cytogenetic abnormalities [12, 42–47]. Molecular stratification of medulloblastoma has correspondingly been incorporated into the recently updated WHO 2021 CNS tumor classification schema, with diagnosis requiring integrated histopathologic and molecular data [48]. Efforts to improve outcomes must tailor therapy to the unique clinical, prognostic, and biological features within each medulloblastoma subgroup and potentially subtype [1, 3].

WNT-Activated Medulloblastoma

Clinical and Histopathologic Features

The WNT subgroup comprises approximately 10% of medulloblastoma cases, usually present in older children (median age 10 years, rarely in infants), and is often characterized by classic histology and localized, non-metastatic disease [10]. WNT-medulloblastomas are commonly midline tumors, in proximity to or infiltrating the brainstem, in accord with preclinical studies demonstrating that their cell of origin is located in the lower rhombic lip and dorsal brainstem [49]. Across several retrospective and prospective studies, pediatric patients with WNT medulloblastoma, including those historically classified as high-risk, have consistently exhibited superior outcomes, with 5-year survival greater than 90% [4, 6–8, 10, 38]. However, adults with WNT medulloblastoma experience poorer survival, as detailed below [26–28].

Genomics

Canonical WNT-signaling pathway activation occurs due to nuclear accumulation of beta-catenin, usually as a result of somatic mutations in CTNNB1, and/or other genes (SMARCA4, CREBBP, KMT2D) encoding proteins that interact with nuclear beta-catenin [8, 13, 50–52]. Somatic alterations of DDX3X and TP53, as well as germline pathogenic variants in APC, have also been described in WNT medulloblastoma [8, 13, 50, 51]. Additionally, the chromosomal aberration monosomy 6 is identified in most, but not all, WNT-activated tumors [50, 52].

Therapeutic Approaches

Given the excellent prognosis observed for children in this subgroup with multimodal therapy [4, 6–8, 10, 38], current efforts are focused on therapy reduction to mitigate treatment-related morbidity. Several ongoing prospective trials across consortia are studying reduction in CSI dosing (to 15–18 Gy) and/or decreased chemotherapy intensity for pediatric patients newly diagnosed with average-risk WNT medulloblastoma (ACNS1422 [NCT02724579]), SIOP-PNET 5 [NCT02066220] [53], and SJMB12 [NCT01878617]). Due to the aforementioned potentially devastating consequences of administering inadequate therapy to patients with higher risk and/or more unfavorable molecular subgroup disease, these trials are relying on real-time central neuro-radiology, pathology, and molecular review to ensure eligibility.

SHH-Activated Medulloblastoma

Clinical and Histopathologic Features

Approximately 25% of medulloblastoma fall into the SHH-activated subgroup, which has a heterogeneous age distribution, presenting most commonly in infants and young children as well as older adolescents and adults [10]. Nodular/desmoplastic histology is enriched in the SHH subgroup, though other histopathologic types can be seen [10, 52]. SHH tumors commonly arise from the cerebellar hemispheres, in accordance with cerebellar granule neuron precursors being the presumed cell of origin [49]. Outcomes are variable and dependent on SHH subtype-specific genomic profiles [44, 54, 55].

Genomics

The molecular hallmark of SHH medulloblastoma is aberrant activation of the SHH-signaling cascade, with somatic mutations in the tumor suppressor gene PTCH1 (which encodes the protein PTC1, the receptor for SHH), representing the most prevalent alteration identified in this subgroup and occurring across the age continuum of this disease [44, 54]. Most other alterations are age group specific, with SUFU mutations occurring in infant SHH patients, including with germline inheritance (Gorlin syndrome) [54, 56], whereas SMO and TERT promoter mutations are nearly exclusively seen in adult SHH medulloblastoma [54]. Genetic aberrations of the PI3K/AKT/mTOR signaling pathway (e.g., PIK3CA, PTEN, and PIK3C2G) have also been reported in SHH medulloblastoma, predominantly in adult patients [54]. Childhood SHH tumors are characterized by germline TP3 mutations (Li Fraumeni predisposition syndrome) as well as co-occurring somatic amplifications of GLI2 and MYCN, likely due to chromothripsis (chromosome shattering) [12, 54, 55]; these tumors also commonly have diffuse anaplasia histologically. Across retrospective and prospective studies, TP53 mutations, GLI2 and/or MYCN amplifications, and chromosome 14q loss have been consistently identified as adverse prognostic factors within SHH medulloblastoma [4–6, 8, 10, 26, 47, 57, 58].

Therapeutic Approaches

Infant SHH medulloblastoma patients experience favorable outcomes and often can be cured with radiation-sparing approaches involving intraventricular or myeloablative chemotherapy, with future efforts focused on de-intensification as detailed below [23–25, 59–61]. Conversely, even with high-dose CSI and chemotherapy in contemporary trials, prognosis remains dismal for children with TP53-mutant, GLI2- and MYCN- amplified, 14q-deleted SHH tumors, a population for whom novel, effective agents are desperately needed [8, 12, 54, 57]. SMO inhibitors such as vismodegib have been studied in SHH patients, with some encouraging initial responses in adults with recurrent disease [62], motivating incorporation of vismodegib as post-chemotherapy maintenance treatment in newly diagnosed SHH medulloblastoma in SJMB12 (NCT01878617). However, responses to vismodegib monotherapy in the recurrent setting were transient, due to acquired resistance (development of somatic SMO point mutation resulting in 100-fold decreased drug binding) [63]. Furthermore, preclinical and clinical evidence suggests only tumors with upstream SHH-pathway mutations (SMO, PTCH1) will have sensitivity to SMO inhibitors like vismodegib, whereas SHH medulloblastomas with downstream SHH-pathway alterations (SUFU, GLI2) will likely be resistant [54]. However, mass spectrometry-based phosphoproteomics have identified the protein kinase CK2 as responsible for regulating the SHH pathway, specifically affecting the terminal-most signaling components [64], therefore representing a reasonable therapeutic target to circumvent aforementioned challenges with resistance and downstream alterations. Based on encouraging preclinical data, an ongoing clinical trial is investigating CX-4945, a selective CK2 inhibitor, in patients with recurrent SHH-activated medulloblastoma (NCT03904862). Additionally, an association between SHH pathway inhibition and bone-related toxicities in prepubertal patients has been demonstrated, with reports of persistent growth plate fusion and corresponding short stature in young patients treated with vismodegib and sonidegib (another SMO inhibitor) [65, 66]; careful growth plate monitoring in skeletally immature children receiving agents targeting the SHH pathway is therefore necessary. Finally, given aforementioned activation of the PI3K/AKT/mTOR pathway in adult SHH medulloblastoma tumors, there may be a role for PI3K and/or mTOR inhibitors, potentially in combination with other agents, including SMO inhibitors to overcome resistance [67, 68].

Group 3 Medulloblastoma

Clinical and Histopathologic Features

Group 3 disease accounts for about 25% of medulloblastoma cases, primarily present in infants and young children, and occurs more commonly in males than females [10]. Outcomes are inferior to other subgroups, likely influenced by higher prevalence of metastatic disease at diagnosis (40–45%), large cell/ anaplastic histology (40%), and/or MYC amplification [9, 10, 69].

Genomics

In contrast to WNT and SHH medulloblastoma, group 3 is not defined by activation of a specific signaling pathway or recurrent somatic mutations, but rather by a distinct transcriptional profile as well as genomic instability with widespread chromosomal structural alterations (gains of 1q, 7, and 17q; deletions of 10q, 11, 16q, and 17p) [12, 14, 69]. These structural variants, which result in aberrant activation of growth factor-independent 1 family proto-oncogenes GFI1 and GFI1B, have been implicated as drivers of group 3 (and group 4) medulloblastoma [70]. MYC amplification and isochromosome 17q are poor prognostic biomarkers, defining high-risk subtypes of group 3 medulloblastoma [9, 10, 12, 14, 46, 47, 69, 71]. Mutations of SMARCA4 and within the histone lysine-specific demethylase (KDM) gene family, as well as amplicons of OTX2, have also been described in group 3 [14, 71, 72]. Furthermore, multi-omic analyses incorporating methylation profiling and transcriptomic data have revealed even greater heterogeneity within group 3 (and group 4) medulloblastoma, with identification of eight distinct subtypes across the biological spectrum of group 3/4 disease [43, 45].

Therapeutic Approaches

Given consistently worse survival (< 60%) experienced by group 3 medulloblastoma patients compared to other subgroups with conventional multimodal therapy [4–6, 8–10], there is a critical need to thoughtfully intensify and augment treatment regimens with biologically guided agents. Concurrent administration of carboplatin during CSI is recommended for non-infant patients with high-risk (especially metastatic) group 3 disease, given this subgroup-specific benefit demonstrated in ACNS0332 [5]. Development of group 3 medulloblastoma mice models has facilitated preclinical efforts to elucidate therapeutic targets [73, 74]. High-throughput drug screening identified the promising combination of pemetrexed and gemcitabine for group 3, MYC-amplified disease [75], providing rationale for addition of these agents to a standard chemotherapy backbone for non-WNT, non-SHH tumors in SJMB12 (NCT01878617). The bromodomain and extraterminal domain (BET) family represents an additional target in group 3 medulloblastoma, given suppression of MYC-associated transcriptional activity with BET-bromodomain inhibition [76, 77]. Cyclin-dependent kinase (CDK) inhibitors (CDK4/6, CDK2, and pan-CDK) have exhibited preclinical efficacy in group 3 medulloblastoma [78, 79], including in combination with BET-bromodomain inhibitors [80], and remain an area of continued research. Preclinical sensitivity of GFI1/GFI1B-activated medulloblastoma to Lsd1 inhibition has also been shown and deserves further study [81].

There is additional ongoing investigation of epigenetic-targeted therapies such as histone deacetylase (HDAC) inhibitors for group 3 medulloblastoma, given prevalence of deregulated chromatin modifiers (such as SMARCA4) as well as genetic alterations of histone lysine methylation (inactivating KDM family mutations) [14, 71, 72]. HDACs were identified as potential therapeutic targets in medulloblastoma from drug screening assays, with preclinical models showing HDAC-inhibitor induced apoptosis, including in group 3, MYC-amplified disease, as well as synergy with cytotoxic chemotherapy [82–88]. There is also preclinical rationale for HDAC inhibition across other medulloblastoma subgroups, with HDAC2 overexpression observed in group 3, group 4, and SHH disease [88, 89]. PBTC-026 prospectively studied incorporation of the HDAC inhibitor vorinostat into an intensive chemotherapy backbone for infants with newly diagnosed high-risk CNS embryonal tumors, many with group 3 medulloblastoma [90]. Results demonstrated feasibility of this combination, motivating development of an upcoming frontline COG trial for molecularly stratified, high-risk pediatric medulloblastoma which proposes to use vorinostat concurrently with radiotherapy and maintenance chemotherapy.

Finally, despite an improved understanding of biologically distinct subtypes within group 3 (as well as group 4) medulloblastoma defined by methylomic and transcriptomic profiles [43, 45], translation of findings into subtype-specific clinical treatment has lagged; future efforts should therefore seek to identify and target therapeutic vulnerabilities within these divergent group 3/4 subtypes.

Group 4 Medulloblastoma

Clinical and Histopathologic Features

Group 4 is the most common medulloblastoma subgroup, comprising approximately 35% of cases, has male predominance, and presents across the age spectrum [10]. Most group 4 tumors have classic histology, though diffuse anaplasia can be rarely seen [10]. Overall prognosis of group 4 patients with conventional multimodal therapy is intermediate compared to other subgroups [9, 10, 14].

Genomics

Group 4 medulloblastoma exhibits the closest molecular overlap with group 3 disease, yet is distinct and remains the least well biologically understood. Group 4 tumors similarly lack recurrent defining somatic mutations or ubiquitous pathway activation, and instead often harbor chromosomal structural variants, which in a subset of cases result in GFI1/GFI1B activation by enhancing “hijacking” [12, 14, 58, 69, 70]. Isochromosome 17q is the most prevalent cytogenetic aberration identified in group 4 medulloblastoma (approximately 80%), but is not independently predictive of survival within this subgroup, in contrast to group 3 disease [9, 12, 71]. Whole chromosome 11 loss has consistently been shown to represent a favorably prognostic feature within group 4 medulloblastoma [4, 12, 58, 91]. MYCN amplification and overexpression are commonly observed in group 4, though is not associated with poorer outcomes, unlike SHH counterparts [9, 12, 71]. Mutations of KDM family members (including of KDM6A) as well as amplifications of OTX2, SNCAIP, and CDK6 have also been reported in group 4 tumors [14, 92, 93].

Therapeutic Approaches

Modern therapy cures a high proportion of patients with group 4 disease, though further research is necessary to improve understanding of the molecular drivers and prognostic factors within this subgroup, such that future treatment can be appropriately tailored [1, 3]. Given superior outcomes observed in group 4 medulloblastoma with chromosome 11 loss, reduced CSI dosing (18 Gy) will be studied in this patient subpopulation in an upcoming COG trial, ACNS2031. Additionally, as the molecular landscape of group 4 is most similar to group 3, many of the aforementioned agents and strategies being researched in group 3 medulloblastoma (inhibitors of BET, CDK, Lsd1, and HDAC, as well as pemetrexed and gemcitabine) also warrant exploration in patients with high-risk group 4 disease, some already being investigated.

Integrated Clinical and Molecular Risk Stratification

As introduced above, risk stratification in prospective medulloblastoma trials must combine both clinical and molecular features with established prognostic relevance. Figure 1 presents an example integrated risk stratification schema for newly diagnosed pediatric (non-infant, non-adult) patients, in accord with proposed eligibility to current and upcoming COG frontline studies. The independent significance of large cell anaplastic histology in the absence of other high-risk clinical or molecular features remains unknown [5].

Fig. 1.

Integrated risk stratification schema for newly diagnosed pediatric (non-infant, non-adult) patients

Unique Medulloblastoma Patient Populations, with Contemporary Clinical Trial and Molecular Updates

Infant and Young Child Medulloblastoma

Due to the greater vulnerability of the developing young child brain to cranial irradiation, treatments for patients diagnosed with medulloblastoma under 3–4 years of age have focused on radiotherapy-sparing or -delaying approaches to preserve neurocognitive function [21, 22]. Regimens that have been investigated in this patient population include systemic chemotherapy alone or in combination with high-dose myeloablative chemotherapy, intraventricular methotrexate, and/or focal primary site irradiation [60, 61, 94–97]. Contemporary clinical trials for these young children have incorporated histology- and biology-based stratification, given emerging knowledge that SHH-activated infant nodular desmoplastic medulloblastoma or medulloblastoma with extensive nodularity (MBEN) carries a substantially more favorable prognosis [59–61].

Myeloablative chemotherapy regimens have been studied in North American trials for infant medulloblastoma, based on encouraging survival results from the Children’s Cancer Group (CCG) 99703 trial, consisting of three induction chemotherapy cycles followed by three consolidation cycles with dose-intensive thiotepa and carboplatin with autologous stem cell rescue [60, 97]; 5-year progression-free (PFS) survival of 70% was observed across the infant medulloblastoma cohort, with superior outcomes among patients with SHH medulloblastoma (86%) compared to group 3 disease (49%) [97]. In the successor COG ACNS0334 phase 3 trial, infants with non-desmoplastic medulloblastoma received the 99703 chemotherapy backbone, with randomization of added high-dose systemic methotrexate to induction (with post-chemotherapy radiation at the treating oncologist’s discretion) [23, 24]. The addition of methotrexate resulted in improved objective response rates (63% vs. 30%) and 2-year EFS (68% vs. 46%), with this benefit primarily driven by group 3 patients, as SHH patients experienced 100% survival irrespective of methotrexate [21, 23, 24]. Feasibility and efficacy of marrow-ablative chemotherapy as a radiation-sparing approach in early childhood medulloblastoma have similarly been demonstrated across the sequential HeadStart trials, which studied 3–5 induction chemotherapy cycles (containing high-dose methotrexate) followed by one consolidation cycle of high-dose carboplatin, thiotepa, and etoposide with autologous stem cell rescue for non-metastatic and/or SHH patients [95, 98]; randomization of single versus triplet consolidation cycles (following three induction cycles) for high-risk patients is being investigated in the currently ongoing HeadStart IV study.

The European approach to treating infant medulloblastoma patients has focused on incorporation of intraventricular and systemic methotrexate, without high-dose chemotherapy. In the prospective HIT2000-BIS4 study, children diagnosed with non-metastatic medulloblastoma under 4 years of age received systemic chemotherapy with intraventricular methotrexate (HIT-SKK) as well as risk-adapted focal radiotherapy (54 Gy, in patients without nodular desmoplastic histology and/or with incomplete remission) [25]. Outcomes were excellent for patients with nodular desmoplastic medulloblastoma or MBEN, all confirmed as SHH-activated by methylation profiling, with 5-year PFS and OS of 93% and 100%, respectively [25]. Despite local radiotherapy and intraventricular methotrexate, inferior outcomes were observed for patients with classic or large cell anaplastic histology (5-year PFS and OS: 37% and 62%), with a higher rate of distant relapses; findings were similar to contemporary studies demonstrating poorer prognoses in non-desmoplastic infant medulloblastoma, most with group 3 disease [95–97, 99].

Recent efforts have investigated therapy reduction among infant nodular desmoplastic (SHH) medulloblastoma patients, given their consistently favorable outcomes, and in order to mitigate risk of leukoencephalopathy associated with intraventricular methotrexate [100], as well as risks of sepsis, ototoxicity, and infertility with myeloablative chemotherapy [60, 98]. In the prospective single-arm trial ACNS1221, patients with nodular desmoplastic medulloblastoma or MBEN (all confirmed as SHH) received modified HIT SKK chemotherapy without intraventricular methotrexate [101]. However, this study closed prematurely, given its failure to maintain 2-year PFS ≥ 90%, with a worrisome pattern of disseminated relapses observed, suggesting that intraventricular or otherwise augmented chemotherapy is necessary for adequate disease control in most cases of infant SHH medulloblastoma [101]. Risk-adapted multimodal therapy for young children newly diagnosed with medulloblastoma was explored in the SJYC07 phase II trial, in which low-risk patients (non-metastatic with gross- or near-totally resected disease) received systemic chemotherapy alone, without intraventricular or high-dose treatment; despite low toxicity, outcomes were inferior to aforementioned intensified regimens [99]. Both ACNS1221 and SJYC07 identified a methylation-defined subtype of infant SHH patients (SHH-II [SHH-γ]) with improved survival (compared to SHH-I [SHH-β]) with therapy reduction, further elucidating the biological heterogeneity of this disease and supporting a potential role for cautious treatment de-escalation in the former with careful molecular stratification [22, 44, 99, 101].

Compared to their pediatric and adolescent counterparts, infants with medulloblastoma currently have fewer clinical trial options, in part due to reluctance to investigate novel agents and approaches in upfront regimens for these young patients. Future prospective studies, incorporating clinical and molecular risk stratification, will aim to de-intensify (but not omit) myeloablative or intraventricular chemotherapy for infants with low-risk SHH medulloblastoma. Research is critically needed to appropriately and thoughtfully augment treatment for young children with group 3 disease, who experience continued poor outcomes even with intensive myeloablative and/or intraventricular therapy, especially in the absence of craniospinal irradiation.

Adult Medulloblastoma

A standard treatment regimen for adult medulloblastoma is currently lacking due to the relative rarity of medulloblastoma within adult neuro-oncology (comprising < 1% of brain tumor diagnoses among adults [2, 102]), frequent exclusion of adults from pediatric cooperative group medulloblastoma trials, and paucity of prospective studies for this specific patient population [26–28, 103–105]. General management approach has consisted of maximal safe surgical resection and post-operative CSI, with or without adjuvant chemotherapy, for which there is substantial variation in regimens [26, 27, 103, 105]. Studies assessing the prognostic impact of adjuvant chemotherapy for adult medulloblastoma patients have yielded inconsistent results, with some demonstrating associated improvement in PFS and OS, yet others without obvious benefit [106–113]; tolerability of pediatric chemotherapy protocols has likely influenced these findings, given high rates of treatment-related toxicity among adult medulloblastoma patients, often requiring dose reductions and interruptions [26, 103, 105, 107, 112, 114].

Adult and pediatric medulloblastoma are clinically and molecularly distinct. Medulloblastoma in adults is rarely metastatic at diagnosis, usually arises from the cerebellar hemispheres, and has propensity for late and extraneural (bone, bone marrow) relapses [10, 26, 27, 103, 108, 115, 116]. By methylation profiling, SHH represents the most prevalent molecular subgroup among adult medulloblastomas (55–65%), followed by group 4 and WNT; group 3 is rare among adults [10, 26–28, 110, 117]. In contrast to childhood medulloblastoma, molecular subgrouping among adults is not consistently predictive of survival, and adults with WNT-activated or group 4 disease experience inferior outcomes compared to their pediatric counterparts [26–28]. Cytogenetic subtype-specific signatures are prognostic, with worse outcomes observed among adult SHH medulloblastomas with chromosome 3p or 10q loss and improved survival in adult group 4 medulloblastoma with chromosome 8 loss, but not chromosome 11 loss [26]. Mutational analyses have revealed enrichment of both TERT promoter alterations within adult SHH medulloblastoma as well as TP53 mutations in adult WNT medulloblastoma; additionally, mutations of the tumor suppressor KMT2C, identified across adult molecular subgroups, are associated with poorer survival [27].

To improve outcomes for this historically understudied medulloblastoma patient population, continued research will be essential—including through prospective trials of standardized multimodal therapy, tailoring treatment according to anticipated tolerance as well as extraneural recurrence risk, stratification based on appropriate prognostic factors, and incorporation of minimally toxic biologically targeted agents.

Relapsed Medulloblastoma

Despite aforementioned clinical and molecular advances in the treatment of medulloblastoma, approximately 30% of patients experience disease recurrence. Prognosis remains dismal for patients with relapsed or refractory medulloblastoma after receipt of upfront CSI, with limited therapeutic options and 5-year OS less than 10% [29–31]. Young children with recurrent medulloblastoma after prior radiotherapy-sparing approaches can often be salvaged with high-dose CSI [118, 119], but there are no proven curative regimens for other relapsed medulloblastoma populations. Surgical resection, re-irradiation, and/or systemic therapy can often prolong survival at recurrence. Improved survival after medulloblastoma relapse has been demonstrated with re-irradiation (administered focally or in rare instances to the entire neuroaxis), most effective in cases of lower pre-treatment disease burden, and with an important need to balance potential benefit with risk of neurotoxicity, especially CNS radiation necrosis [29, 120–122]. Several chemotherapy agents, such as temozolomide and irinotecan, have previously been studied in recurrent medulloblastoma, with evidence of safety, feasibility, and preliminary anti-tumor activity [123–125]. Based on this data, as well as the angiogenic profiles of medulloblastoma with elevated VEGF and VEGF receptor expression [126], ACNS0821 randomized patients with recurrent disease to the combination of temozolomide and irinotecan, with or without bevacizumab [127]. The addition of bevacizumab to temozolomide and irinotecan resulted in significantly improved EFS and OS, and was well tolerated in a heavily pretreated cohort [127], representing a reasonable first-line systemic regimen at recurrence (and soon to be incorporated following upfront maintenance therapy in a prospective trial for newly diagnosed high-risk patients through COG). Myeloablative chemotherapy with autologous stem cell rescue has been investigated in relapsed medulloblastoma, with significant toxicity risk and lack of clear efficacy, thus is not recommended [128–130].

Key biology analyses of patients with recurrent disease following treatment with SJMB03 and SJYC07 [29], as well as within large discovery cohorts from Toronto, the UK, and the International Cancer Genome Consortium [131, 132], have facilitated an improved understanding of the molecular landscape of relapsed medulloblastoma. Among patients from the St Jude studies with paired diagnostic and recurrent tumor samples, molecular subgroup and subtype were conserved in 96% and 80% of cases, respectively, with rare potential for subgroup divergence from group 4 to group 3, as well as for acquisition of MYC or MYCN amplifications in relapsed disease [29]. Nearly ubiquitous conservation of molecular subgroup at medulloblastoma relapse was similarly demonstrated by Ramaswamy et al. and Richardson et al. [131, 132]. Furthermore, relapsed medulloblastoma is characterized by selective maintenance of established driver mutations (e.g., alterations of SHH or WNT pathways, TP53) in most cases, yet overall genetic divergence from diagnosis, with emergence of subgroup-specific novel pathways and events [132]. Acquired TP53 mutations and copy number variants are enriched in relapsed non-infant SHH medulloblastoma, while CDK amplifications and USH2A mutations were newly identified in relapsed group 4 patients [132]. Anatomic location of relapse varied by molecular subgroup, with local recurrences identified in SHH medulloblastoma, whereas disseminated (distant and/or local) relapse was more commonly observed in group 3/4 [29, 131]. Group 4 patients experienced longer time to relapse and post-relapse survival [29, 131].

In addition to above ongoing efforts to decrease recurrence risk among high-risk patients, there remains a critical need to develop novel effective treatments for relapsed medulloblastoma, taking advantage of molecular subgroup- and subtype-specific therapeutic targets.

Innovative Treatment Modalities, Strategies for Toxicity Mitigation, and Early Detection Methods of Disease Response or Recurrence

Targeted Radionuclide Therapy

Targeted delivery of beta radiation based on the presence of medulloblastoma-specific cell surface proteins is currently under investigation. Lutetium (¹⁷⁷Lu)-DOTATATE, an intravenous radionuclide therapy which binds type-2A somatostatin receptors (SST2A) and delivers local radiation via beta particle emission, has gained FDA approval for the treatment of adult patients with gastroenteropancreatic neuroendocrine tumors [133, 134]. Given emerging evidence of consistently high membranous SST2A expression in medulloblastoma [135–142], a phase I/II trial of ¹⁷⁷Lu-DOTATATE in recurrent medulloblastoma is underway (NCT05278208). Iodine (¹³¹I)-Omburtamab is a similar radionuclide therapy targeting B7H3, a highly prevalent tumor cell surface antigen in embryonal tumors including medulloblastoma [143]. Based on safety, feasibility, and preliminary efficacy of intraventricular administration of radioimmunotherapies [144], intraventricular Omburtamab is being studied in ongoing clinical trials in relapsed medulloblastoma (NCT04743661; NCT04167618). Notably, these radionuclide therapies cause cell death by release of beta radiation with a path length of the therapeutic beta particle of < 2 mm in soft tissue [145]; even with strong uptake in medulloblastoma tumor cells, the extent of surrounding normal brain tissue that will receive significant radiation exposure will be very small, sparing most healthy brain tissue and allowing use of this modality irrespective of previous radiotherapy.

Immunotherapy

Given exciting clinical efficacy in other malignancies, immunotherapeutic approaches remain a topic of continued research in medulloblastoma. However, immunotherapy progress in medulloblastoma has been limited to date by lack of known immunogenic antigens, low tumor mutational burden, minimal to absent PD-L1 expression, few infiltrating T lymphocytes in the tumor microenvironment, immuno-resistant properties of medulloblastoma cells, and inadequate blood–brain barrier penetration [146–150]. To overcome some of these challenges, Donovan et al. studied CSF-directed locoregional delivery of CAR T cells targeting medulloblastoma cell surface proteins (EPHA2, HER2, and interleukein 13 receptor α2), with demonstrated efficacy in metastatic group 3 mice models [151], motivating ongoing clinical trials for recurrent patients (NCT03500991; NCT04661384). Additionally, oncolytic viral therapies based on cytomegalovirus, measles virus, and herpes simplex virus are currently being investigated in early phase trials in relapsed medulloblastoma (NCT05096481; NCT03299309; NCT02962167; NCT03911388). Success of immunotherapy in medulloblastoma will likely require combinatorial approaches that improve tumor immunogenicity, enhance drug delivery, and target medulloblastoma-specific antigens; moreover, treatment needs to be tailored to the diverse immunologic profiles of medulloblastoma subgroups [146–150].

Techniques to Decrease Radiotherapy-Related Toxicity

The historical standard of care for radiation therapy delivery for patients with medulloblastoma is photon-based CSI. The vast majority of prior medulloblastoma clinical trials used 3D conformal radiation or intensity-modulated radiation therapy (IMRT), a more conformal photon technique which allows for greater sparing of normal tissue [152]. However, with these conventional techniques, patients with medulloblastoma remain at significant risk of radiation-related acute and late toxicities, including dermatitis, alopecia, hearing loss, myelosuppression, neurologic sequelae, cognitive decline, endocrine abnormalities, and secondary malignancies (4% at 10 years) [4, 153, 154]. With photon-based spinal irradiation, there is also low and intermediate radiation exposure to surrounding normal tissues including organs of the chest and abdominal cavities, leading to risk of acute esophagitis and late cardiac, pulmonary, and gastrointestinal toxicity.

With both IMRT and smaller tumor boost volumes for average-risk patients (based on ACNS0331 results), there is a greater sparing of the temporal lobe/hippocampus and middle ear/cochlea, which may decrease radiation-related neurocognitive and ototoxicity. In ACNS0331, patients who received the more limited involved field tumor boost (compared to entire posterior fossa boost) had significantly higher intelligence quotient (IQ) scores at the second neurocognitive assessment timepoint (approximately 30 months post-diagnosis), but this difference did not remain significant at the third timepoint (60 months post-diagnosis) [4]. Hearing outcomes, evaluated after completion of therapy, were not significantly impacted by radiotherapy boost volume, possibly due to high rates of adjuvant cisplatin-related sensorineural hearing loss [4].

With expanded access to proton centers worldwide, there is increasing use of proton-based CSI, which represents another technique to decrease radiotherapy-related toxicity. Proton-beam irradiation uses a charged particle, which provides superior radiation dosimetry, allowing adequate tumor coverage with a steep dose falloff after dose deposition in tumor tissue, ensuring surrounding normal tissues receive no radiation dose beyond the treatment target. In a prospective phase II trial of proton-based CSI in pediatric patients with medulloblastoma, no late cardiac, pulmonary, or gastrointestinal toxicity was observed [155]. In addition, superior long-term intellectual outcomes (global IQ, perceptual reasoning, and working memory) have also been demonstrated with proton- compared to photon-based CSI, without compromised disease control [156]. Among adults with medulloblastoma, compared to photon CSI, proton CSI was associated with significantly lower rates of acute adverse events, including weight loss (1.2% vs. 5.8%, p = 0.004), nausea or vomiting (16% vs. 64%, p = 0.004), esophagitis (5% vs. 57%, p < 0.001), and hematologic toxicity (due to a significant reduction in mean vertebral body dose) [157]. Additionally, lower rates of central and primary hypothyroidism have been observed with proton vs. photon CSI among pediatric medulloblastoma patients, likely due to reduced radiation doses to the pituitary and thyroid glands [158, 159]. Ongoing research efforts seek to further improve proton CSI techniques, including through intensity-modulated proton therapy (IMPT), enabling better conformality and offering greater sparing of critical brain regions, such as the hippocampus, which may also aid in preserving cognitive function. Another area of intense study is FLASH proton therapy, which delivers proton radiation at extremely high dose rates which may further mitigate normal tissue toxicity, although clinical use of this technique remains limited [160].

Finally, given the NMDA pathway’s role in acute and chronic radiation-induced neuroinflammation and associated brain injury, memantine, a competitive NMDA receptor antagonist with demonstrated evidence of neurocognitive protection in adults receiving whole brain radiotherapy [161], is being studied in pediatric CNS tumor patients undergoing proton or photon cranial irradiation in an ongoing COG trial, ACCL2031 (NCT04939597).

Sodium Thiosulfate for Ototoxicity Prevention

A large proportion of medulloblastoma survivors suffer from clinically significant and permanent cisplatin-induced sensorineural hearing loss [18, 162], which can negatively impact speech, language, and social development. Sodium thiosulfate (STS) inactivates platinum agents and scavenges reactive oxygen species believed to mediate ototoxicity in the cochlear hair cells. The otoprotective potential of STS has been demonstrated across several in vitro and in vivo preclinical studies [163–165], as well as in clinical trials in hepatoblastoma [166, 167]; however, questions remain about possible tumor protective properties, given worse outcomes observed in a heterogeneous patient population with disseminated disease who received STS [167]. The upcoming COG study ACNS2031 will prospectively evaluate the effect of STS on cisplatin-induced ototoxicity and survival in a carefully (clinically and molecularly) defined cohort of patients with average- and low-risk medulloblastoma.

CSF “Liquid Biopsy”

Disease evaluation and response assessment in medulloblastoma have historically relied upon neuro-imaging (MRI), clinical status, and CSF cytology [168]. Recent research has aimed to identify minimally invasive biomarkers of therapy response or recurrence by analyzing tumor-associated copy number variations (CNVs) and/or somatic alterations in CSF-derived cell-free DNA (cfDNA) [169–171]; this strategy represents an innovative “liquid biopsy” tool to measure and serially monitor minimal residual disease (MRD) in medulloblastoma, for which therapy intensification may be indicated. Liu et al. describe the superior sensitivity of CSF liquid biopsy cfDNA methodology compared to traditional CSF cytology, with more than half of cytologic-negative specimens harboring tumor-derived cfDNA [169]. Additionally, persistent MRD positivity by CSF-derived cfDNA during and at the completion of chemotherapy was predictive of subsequent radiographic progression, therefore, representing a clinically meaningful biomarker to guide treatment intensification [169]. Prospective and longitudinal evaluation of CSF-derived cfDNA will be incorporated as critical biology correlatives in upcoming medulloblastoma clinical trials.

Conclusion

Impressive progress has been made over the past two decades in elucidating the clinical, biological, and molecular heterogeneity of medulloblastoma, for which treatment must be appropriately tailored. With an improved understanding of medulloblastoma risk stratification integrating both clinical and molecular features, ongoing research is focused on (1) careful therapy reduction and interventions to mitigate toxicity in patients with average-low risk disease and (2) biologically guided identification and treatment intensification with novel agents for high-risk patients. These important efforts will hopefully continue to decrease treatment- and tumor-related morbidity and mortality across the age spectrum of medulloblastoma.

Supplementary Information

Below is the link to the electronic supplementary material.