Clinical manifestations

Presentation and course

Brain tumors can often be localized based on the symptoms they provoke; however, this is not universally true of diffuse pediatric high-grade gliomas. Depending on the subtype, these tumors may be vastly infiltrative or relatively well-circumscribed. The infant type of diffuse high-grade glioma is the most likely subtype to be circumscribed, but the diffuse pediatric-type high-grade glioma, H3-wildtype and IDH-wildtype can also be circumscribed (17). Headaches are a leading symptom when the tumor causes a decreased flow of cerebral spinal fluid, leading to increased intracranial pressure (36; 46; 04) and may be part of the classic triad of increased intracranial pressure, lethargy, and vomiting. The headaches are usually worse when waking up and improve during the day. School-aged children more commonly report vague, intermittent headaches and fatigue. They may have a declining academic performance and exhibit personality changes. Infants may present with irritability, failure to thrive, anorexia, developmental delay, or regression (42). On physical examination, raised intracranial pressure is best appreciated as papilledema on fundoscopic evaluation; however, this may present only as optic pallor in infants with open cranial sutures.

Although brain tumors are found in 1% to 3% of children with new-onset seizures, seizures occur in approximately 20% of patients with supratentorial malignant gliomas (36). Supratentorial tumors were more commonly associated with seizures than infratentorial tumors. Delayed seizure onset, lesser seizure frequency with focal weakness, and signs of increased intracranial pressure were noted more in malignant tumors (50). Hypothalamic tumors may be associated with neuroendocrine abnormalities, growth hormone deficiency, diabetes insipidus, and precocious pubertal development. These tumors may also impinge on the optic chiasm, resulting in optic atrophy and visual deficits. Patients with diencephalic tumors may present with the classic diencephalic syndrome (emesis, emaciation, and unusual euphoria), but the syndrome is rare in children older than 3 years.

Prognosis and complications

In most children with pediatric-type diffuse high-grade glioma, outcomes are poor. Most members of this group are considered to have aggressive tumors with very poor overall survival (17).

Diffuse hemispheric glioma, H3 G34-mutant. In a study of 135 patients diagnosed with H3 G34-mutant diffuse hemispheric glioma between 2012 and 2021, the median overall survival was 17.3 months; those patients with gross total or near total resection or older age had improved survival (11). Surprisingly, approximately half of the patients in this study were able to have a gross or near total resection, suggesting that this type of glioma may not be as infiltrative. A better prognosis is observed in those tumors with MGMT promoter methylation (17).

Diffuse pediatric-type high-grade glioma, H3-wildtype and IDH-wildtype. Within this subgroup, those tumors with MYCN amplification and pontine location have the worst prognosis, with a median survival of fewer than 2 months (17). The RTK1 subgroup is often seen in patients who have had prior radiation, and this portends a worse prognosis in this setting (17).

Infant-type hemispheric glioma. Infant-type hemispheric gliomas have a much better prognosis than other pediatric-type high-grade gliomas, with a 60% to 70% overall survival (12).

Given that this is a new tumor classification, the overall and event-free survival has yet to be accurately established for each subgroup. Using previous classification schemas, the consensus was that as a group, supratentorial malignant gliomas have a less than 30% 5-year survival rate (52), and this likely translates to pediatric-type diffuse high-grade gliomas as well. The one exception to this poor prognosis is the infant-type hemispheric glioma in which specific fusions may respond to targeted therapy and, thus, have a better overall prognosis (17).

A comprehensive review of Surveillance, Epidemiology, and End Results (SEER) data consisted of 6212 patients younger than 20 years of age at diagnosis of glioma (1973 to 2005) in four age categories: less than 1 year, 1 to 3 years, 3 to 5 years, and 5 to 20 years (41). Tumor grade emerged as the most significant independent prognostic factor in all age groups except the youngest, in which extent of resection was most significant. A more recent SEER data review on 302 pediatric patients with glioblastoma reveals an overall median survival of 20 months and 2-year survival rate of 46.9%. When broken down into age categories, those aged 0 to 4 years had the highest median survival of 59 months; 5 to 9 years was 12 months (24.5% 2-year survival rate), 10 to 14 years was 19 months (35.8% 2-year survival rate), and 15 to 19 years was 23 months (48% 2-year survival rate). Those with supratentorial tumors had better survival outcomes compared to infratentorial tumors (median overall survival 22 months, 45.4% 2-year survival rate, vs. 14 months, 13.5% 2-year survival rate).

Clinical vignette

A 12-year-old male presented to the emergency room with several months of worsening coordination that left him unable to play his favorite sport of soccer, followed by progressive right-sided weakness and slurred speech. An MRI showed a diffuse, expansile T2 hyperintense signal that involved the corpus callosum and extended into the right cerebral hemisphere and cortical spinal tracts. The differential diagnosis at that time was a primary glial neoplasm and, less likely, lymphoma, given a lack of significant enhancement. He underwent needle biopsy, and pathology showed an infiltrating glioma with brisk mitotic activity, which, according to the most recent WHO classification, fell under the subtype of a diffuse pediatric high-grade glioma, H3-wildtype and IDH-wildtype. Given the high grade of this tumor, he immediately started intensity-modulated photon radiation therapy to the areas of the brain that contained tumor for a total of 60 Gy. He was also concurrently treated with oral temozolomide as per the Children’s Oncology Group study ACNS0423. Molecular testing on the tumor returned with PDGFRA N468S-subclonal (but only a 1% variant allele frequency), PIK3R1 E462_R465del (45.8% variant allele frequency), ATR 1774fs*3, STAG2 L813, TERT promoter -146C>T. Unfortunately, no successful targeted therapies were available at that time.

At the completion of radiation, MRI showed a slight increase in the T2/FLAIR hyperintensity in the subcortical white matter of the frontal lobes. Four weeks after the completion of radiation, he started maintenance chemotherapy with 1 day of lomustine and 5 days of temozolomide, with plans to repeat this every 6 weeks. He remained steroid-dependent, so a dose of bevacizumab was given in an attempt to allow for more rapid weaning of steroids. Unfortunately, he then developed right-sided paralysis and refractory seizures, requiring admission to the hospital and multiple antiepileptics. MRI obtained on admission did not show any infarct or hemorrhage at that time, and the overall tumor burden was stable.

Over the course of the next month, the tumor showed increased infiltration into the left cerebral hemisphere with a midline shift to 11 mm. A shunt was placed, which did help with some symptoms. Previous chemotherapy was discontinued, and the patient was started on oral metronomic chemotherapy with celecoxib, fenofibrate, valproic acid, and alternating cyclophosphamide and etoposide. After 4 months of metronomic chemotherapy, there was an improvement in the mass effect caused by the tumor, and after 8 months of this therapy, there was a significant decrease in enhancement along the corpus callosum, left peritrigonal white matter, and the centrum semiovale. There was still a T2 hyperintense signal throughout the cerebral hemisphere, corpus callosum, and right centrum semiovale but almost complete resolution of any midline shift.

At the last reported follow-up 13 months after diagnosis and 9 months after starting metronomic therapy, the patient remained alive. He had significant deficits, including residual right hemiparesis, apraxia, dysarthria, dysphonia, and cognitive impairment.

Biological basis

Etiology and pathogenesis

When discussing the etiology and pathogenesis of pediatric-type diffuse high-grade gliomas, it is best to separate them into their subtypes.

Diffuse hemispheric glioma, H3 G34-mutant. The H3 G34-mutant diffuse hemispheric glioma is triggered by a point mutation in the histone H3.3 gene at codon 34, which results in a glycine being replaced most commonly by an arginine, but it can also be replaced by a valine (17). It is hypothesized that the tumors are neuronal in origin (40). The majority of these tumors will have mutations in TP53 and a loss of ATRX, and they will lack OLIG2 expression (23).

Diffuse pediatric-type high-grade glioma, H3-wildtype and IDH-wildtype. This particular subgroup of pediatric-type diffuse high-grade gliomas is not as cohesive. To fall into this category, a tumor must be high grade as assessed by the pathologist, not have any histone mutations on molecular testing, and fall into one of three subgroups when methylation testing is performed (17). The three methylation subgroups are RTK1, RTK2, and MYCN (17). They commonly have mutations in MYCN, EGFR, NF1, TP53, and PDGFRA (39). When tumors with these characteristics appear in patients with cancer predisposition syndromes, such as Li Fraumeni syndrome, they should be viewed as completely different tumors because they often behave differently in that specific patient population (40).

Infant-type hemispheric glioma. Infant-type hemispheric gliomas originate from astrocytes and are not associated with histone mutations (17). These tumors that occur in the youngest of patients frequently express fusions that involve the following receptor tyrosine kinases: MET, NTRK, ALK, and ROS1 (39). The diagnosis is made based on a characteristic fusion or specific methylation clustering (17). There is a small subset of infant-type hemispheric gliomas that do not have high-grade features (17).

Epidemiology

Pediatric-type diffuse high-grade gliomas as a whole (including diffuse midline gliomas) are responsible for between 15% and 20% of all pediatric brain tumors and, unfortunately, are the top contributor to mortality from malignancy in children (39). Diffuse hemispheric glioma, H3 G34-mutant is most common in teenagers and makes up approximately one sixth of parietal and temporal tumors in children (17). Fewer than 15% of pediatric high-grade gliomas occur in children 4 years and younger (12). Diffuse pediatric-type high-grade glioma, H3-wildtype and IDH-wildtype tumors have been seen after radiation for other malignancies, including acute lymphoblastic leukemia and medulloblastoma (17). Epidemiologic studies investigating parental and patient occupational exposure, environmental exposure, and nutritional intake have failed to identify linkages with the development of supratentorial malignant gliomas (10).

Prevention

There is no known prevention of pediatric-type diffuse high-grade glioma.

Differential diagnosis

Confusing conditions

Diagnostic workup

Imaging studies. Head CT or preferably MRI with and without contrast must be performed in all patients suspected of having a brain tumor. On CT, malignant gliomas typically appear as irregularly shaped lesions with partial contrast enhancement and peritumoral edema with or without mass effect (29). MRI is also useful by better demonstrating the anatomic origin and extent of tumors. MRI is the imaging modality of choice for detecting disseminated spinal cord lesions and should be performed in all patients with suspected pediatric-type diffuse high-grade glioma. A postoperative MRI or CT is required to measure the extent of surgical resection and detect residual disease. Postoperative MRI or CT evaluation must be performed within 72 hours of surgery to delineate residual tumor from the postsurgical inflammatory changes that are visualized at this time.

CSF cytological examination. Lumbar puncture may be useful in malignant gliomas for detecting microscopic leptomeningeal dissemination if clinically or radiographically suspected. CT imaging or MRI must be performed before the lumbar puncture to rule out the presence of hydrocephaly in those patients suspected of having a brain tumor. Hydrocephaly places the patient at risk for herniation as a consequence of the procedure. In general, the lumbar puncture is deferred for up to 2 weeks postoperatively to avoid identifying contaminating tumor cells that may have disseminated due to surgery.

Biopsy. Biopsy, whether a stereotactic needle biopsy or open biopsy, depending on the recommendations of the neurosurgeon, should be obtained if safely possible. Enough tissue should be procured to allow for morphologic diagnosis but also molecular testing and oftentimes methylation profiling as well. The characteristics that define each of the subtypes of pediatric-type diffuse high-grade glioma have been discussed previously.

Management

Surgery. In pediatric-type diffuse high-grade gliomas, surgery alone is not curative but is necessary for diagnosis. Enough tissue must be obtained to send for molecular testing and, oftentimes, methylation profiling. Debulking can be considered especially for symptomatic relief but should be done by a pediatric-trained neurosurgeon.

Radiotherapy. Radiation therapy is often used for the upfront treatment of diffuse hemispheric glioma, H3 G34-mutant tumors, and diffuse pediatric-type high-grade glioma, H3-wildtype and IDH-wildtype. Radiation is not used in infant-type hemispheric glioma given the young age of these patients and the desire to avoid radiation in the early stages of brain development (typically 3 years and younger) (12). Postoperative local (2 to 4 cm margin around the area of edema defined by imaging) or wide-field irradiation to 50 to 60 Gy is the mainstay of therapy for supratentorial malignant gliomas (26). If the tumor is disseminated, then craniospinal radiation should be used. The radiation modality should be discussed with an experienced pediatric radiation oncologist as each modality has its pros and cons.

Chemotherapy. Because the entity pediatric-type diffuse high-grade gliomas was only established in 2021, much of the data available on the use of chemotherapy concerns the broader category of pediatric malignant gliomas. Children with malignant gliomas appear to benefit from the addition of adjuvant chemotherapy to postoperative radiotherapy, although the extent of this benefit remains unclear. Early phase II and III trials with single-agent and multiagent chemotherapy demonstrated little or no impact on the overall survival despite showing objective response rates as high as 45% (30).

A Children’s Cancer Group trial with prednisone, lomustine (CCNU), and vincristine following postoperative radiotherapy reported a 5-year progression-free survival of 46% versus 18% with radiotherapy alone (47). However, a subsequent trial comparing pCV to the “8-drugs-in-1-day” regimen showed no significant differences in survival between the two groups, and the progression-free survival associated with pCV (26%) was substantially inferior to that noted in the original study (14). Furthermore, a review of the histology from the initial study with pCV suggested that a significant proportion of patients (69/250) did not fit the central consensus definition of malignant glioma. The Pediatric Oncology Group completed a study with adjuvant cisplatin and carmustine versus cyclophosphamide and vincristine following postoperative radiotherapy. The 5-year overall survival between the groups was 27% (± 9%) and 7% (± 7%) respectively.

Multiple trials utilizing neoadjuvant chemotherapy with a variety of regimens, including cisplatin and cytosine arabinoside; carboplatin, ifosfamide, or cyclophosphamide with etoposide; and procarbazine, have been uniformly disappointing (44). The report from the Children’s Oncology Group of a phase II trial (CCG 9933) randomizing between three separate high-dose chemotherapy regimens consisting of carboplatin/etoposide, ifosfamide/etoposide, or cyclophosphamide/etoposide prior to radiation failed to show any significant improvement in overall survival (24% +/- 5% at 5 years), and outcome did not differ between groups. The 5-year, event-free survival rate for all patients was 8% +/- 3% (35). In a report from the German study HIT-GBM-C utilizing radiation and simultaneous intensive multiagent chemotherapy, the 5-year overall survival rate for patients with complete resection was 63% +/- 12% SE, compared with 17% +/- 10% SE for the historical control group (P = .003), suggesting for the first time a therapeutic advantage with the addition of chemotherapy (53).

High-dose chemotherapy with autologous bone marrow or stem cell rescue has also been attempted in this population (13; 16; 21). In a study, objective responses were observed in 4 of 14 assessable patients treated with thiotepa and etoposide followed by autologous bone marrow rescue for malignant glioma (16). All of these patients (28%) remained free of disease for at least 3 years after autologous bone marrow rescue. A study using carmustine, thiotepa, and etoposide followed by autologous bone marrow rescue for 11 patients with malignant glioma demonstrated a 46% 2-year survival rate (21). In a Children’s Oncology Group study, myeloablative chemotherapy with autologous marrow rescue produced durable remissions in children with recurrent malignant gliomas in the setting of minimal residual tumor burden achieved surgically and was superior to conventional chemotherapy (15).

Despite the promising response rates reported, the toxicity associated with these regimens has been excessively high (5% to 15% toxic death rate due to treatment). In an effort to reduce toxicity, investigations have used multiple cycles of somewhat lower doses of chemotherapy followed by peripheral blood stem cell support (25). To a great extent, this has decreased transplant-related complications; however, the data relating to the maintenance of treatment efficacy are still premature, and the role of high-dose chemotherapy in the treatment of primary malignant gliomas remains undefined.

The use of temozolomide alone in pediatric patients has been less promising (31; 06). Because the use of concomitant radiation and temozolomide showed an overall survival benefit in newly diagnosed adult glioblastoma patients (49), several studies have been done to see if the same results are seen in pediatric high-grade glioma but have thus far failed to demonstrate a benefit on overall survival (09; 07). Another Phase II study by the Children's Oncology Group of children with high-grade gliomas showed that the addition of lomustine to temozolomide as adjuvant therapy improved outcomes compared to patients treated with temozolomide alone (09; 24); however, the study lacked treatment randomization and there was a substantial increase in hematologic toxicities (22).

An international phase 2 randomized controlled trial (HERBY) evaluated the use of concomitant and adjuvant temozolomide with/without bevacizumab (an anti-VEGF antibody) in pediatric high-grade glioma, with results showing that the addition of bevacizumab did not increase event-free survival and was associated with increased toxicity (20; 22). Currently, clinical trials are still underway to determine the most effective regimen.

Targeted therapy. The NTRK inhibitor larotrectinib has been successful in some patients with infantile high-grade glioma with NTRK fusions and was not associated with serious adverse side effects (12). Attempts have been made to use EGFR inhibitors in EGFR-amplified high-grade gliomas, but, unfortunately, this has not been successful at stopping tumor growth. However, newer EGFR inhibitors, such as osimertinib, may have better luck penetrating the blood-brain barrier (08). In a subset of six patients with PDGFRA-altered pediatric high-grade gliomas, targeting PDGFRA with dasatinib and also using an mTOR inhibitor seemed to prolong survival, but clinical trials are needed (38).

Other biological-based clinical trials are also currently being investigated. Histone deacetylase inhibitors, such as valproic acid, are well tolerated and have an encouraging response rate in heavily pretreated patients (37). Other histone deacetylase inhibitors currently being investigated include vorinostat and panobinosat, as well as BET bromodomain inhibitors such as BMS-986158 (22).

Immunotherapy. One of the mechanisms of cancer cells’ proliferative growth is their ability to evade the body’s immune detection system. The advent of immunotherapy in melanoma, lung, and other cancers has significantly impacted the treatment of these cancers. It has been theorized that tumors with high mutational burden respond better to immunotherapy; however, pediatric cancers generally harbor a lower mutational burden. One of the few exceptions is the acquired hypermutator phenotype associated with constitutional mismatch repair deficiency, which is present in less than 5% of patients with high-grade glioma. Immune checkpoint inhibitors such as programmed cell death-1 receptor monoclonal antibody (nivolumab) have been shown to have a promising role in this subset and are being studied in children with newly diagnosed or recurrent high-grade glioma with and without ipilimumab, a human cytotoxic T-lymphocyte antigen (05).

Autologous dendritic cell vaccination has shown promise in adult glioblastoma and is currently undergoing clinical trials (01). This process involves isolating monocytes from patients, maturing them with cytokines into dendritic cells, loading dendritic cells with appropriate tumor antigens in vitro, and infusing them back into patients to activate their immune system. According to clinicaltrials.gov, a clinical trial using a CMV dendritic cell vaccine given with GMCSF after priming with tetanus toxoid was conducted in 11 children, and nine of those children completed the trial (NCT03615404). Full results are pending.

Viral infection injected directly into tumor cells such as Herpes simplex virus 1716 (18), cytomegalovirus-specific cytotoxic T cells (34), or parvovirus H-1 (27) are being investigated. New viruses being studied in pediatric gliomas are in the early phases, including recombinant polio/rhinovirus delivered by convection-enhanced delivery. Intravenous reovirus as well as Newcastle disease virus are being evaluated. Other immunotherapies that are being studied for pediatric high-grade glioma include immunomodulatory agents such as thalidomide and pomalidomide. Chimeric antigen receptor T cells, which are T lymphocytes genetically engineered to target specific proteins with greater affinity, are currently in phase 1 trials, including children with recurrent high-grade glioma (22).

Finally, inhibitors against key survival pathways such as Akt and Notch signaling are also being investigated as potential therapeutic targets.

Tumor-treating fields. Tumor-treating fields are intermediate-frequency alternating current electric fields delivered to the tumor site via an array of electrodes applied directly to the scalp, acting as cytotoxic modalities by interfering with key components of cell division, specifically disrupting the spatial orientation of highly polarized molecules required for mitosis (03). The EF-14 trial showed that the addition of tumor-treating fields to standard-of-care adjuvant temozolomide following concurrent radiation and chemotherapy improved the 5-year overall survival in newly diagnosed glioblastoma patients (13% vs. 5%, p=.004) (48). In a feasibility study applying tumor-treating fields to five pediatric patients with high-grade glioma, all five patients tolerated tumor-treating fields well without treatment-limiting toxicities (19).

Outcomes

Patients who survive pediatric-type diffuse high-grade gliomas are often left with significant neurocognitive, neuroendocrinologic, and other neurologic deficits as a result of the tumor’s location within the central nervous system. Furthermore, secondary effects can result from surgery, irradiation, and chemotherapy (51). Complications that may arise as a result of either direct tumor effect or the treatment administered include obstructive hydrocephaly, neurologic impairment, radiation-induced effects (neurocognitive decline, endocrinologic dysfunction, mineralizing microangiopathy with ischemia or infarct, secondary CNS malignancies, transient headaches, fatigue, nausea, vomiting, and anorexia), and chemotherapy-induced effects (myelosuppression, infection, nausea, vomiting, anorexia, renal damage, hepatic damage, hearing damage, neurotoxicity, and secondary malignancies). Infertility and impairment of growth may also be long-term sequelae of therapy (18).

Special considerations

Pregnancy

It is advisable that women of childbearing age take measures to avoid pregnancy during treatment for pediatric-type diffuse high-grade glioma, due to the known teratogenic effects of chemotherapy and radiotherapy. It is not known what the effects are on children conceived by men undergoing therapy for malignant glioma. There is a risk of sterility for both men and women following treatment for malignant glioma.

Anesthesia

In general, all types of anesthesia are safe, and medication does not need to be changed or discontinued in anticipation of a planned surgical procedure. However, it should be kept in mind that mental status and seizure threshold may be compromised by any anesthesia and or sedation medication administered to patients with pediatric-type diffuse high-grade glioma.