Study
Number of patients, age at transplant
Pre-transplant status
Conditioning
Posttransplant therapies
Outcome
Kalifa et al. (1992)
6, age 8 months to 16 years
All patients had residual disease
Busulfan, thiotepa
Local XRT in 2 patients
2/6 with EFS at 24 months
Mahoney et al. (1996)
8, age 2.5–15 years
95 % of patients had residual disease
Cyclophosphamide, melphalan
N/A
0 patients with EFS
Finlay et al. (1996)
9, median age 8 years (8 months to 36 years)
4 patients had bulky disease at transplant
Thiotepa, etoposide
N/A
0 patients with EFS
Graham et al. (1997)
19, age 12 months to 27 years
10/19 NED at transplant
Melphalan, cyclophosphamide in 16 patients
N/A
3/19 with EFS at >24 months posttransplant
Guruangan et al. (1998)
5, median age 2.9 years (0.7–5.9 years)
3 NED; 2 patients with minimal residual disease
Carboplatin, thiotepa, etoposide
Reduced-dose craniospinal XRT and local boost
3/5 with EFS at 10–30 months posttransplant
Did not receive pre-transplant XRT
Fagioll et al. (2004)
8, median age 11 years
2/8 NED
Thiotepa, etoposide
No XRT posttransplant given
2/8 with EFS at 7 and 16 months posttransplant
Sung et al. (2007)
7, age 3–17 years
3/7 NED
3 single-transplant patients, 4 double-transplant patientsa
No XRT posttransplant
3/7 with EFS at 9–52 months posttransplant
Ridola et al. (2007)
39 patients with localized relapse, median age 3 years (1–7 years)
9/39 NED
Busulfan, thiotepa
9 patients had second surgery posttransplant, all had posterior fossa XRT
5-year EFS = 61.5 %
Shih et al. (2008)
12, median age 6.7 years (1.1–18.8 years)
3/12 NED
Different conditioning regimens
5/12 patients received radiation as part of salvage therapy
3/12 EFS 20–56 months posttransplant
Dunkel et al. (2010)
25, median age 13.8 years (7.6–45 years), all received previous XRT
N/A
Carboplatin, thiotepa, etoposide
3/25 received XRT pre- or posttransplant
6/25 (24 %) EFS at 151 months posttransplant
Bode et al. (2014)
27, median age 12.6 years (0.4–28.8 years)
9 NED, 18 residual disease
Carboplatin, thiotepa, etoposide
Craniospinal XRT was a part of salvage therapy in children >4 years of age who have not received XRT previously
3-year EFS 10 %
High-dose chemotherapy with peripheral stem cell rescue showed more promising results in young patients with newly diagnosed medulloblastoma, including metastatic disease, than in patients with recurrent disease.
Mason et al. evaluated high-dose chemotherapy with autologous stem cell rescue in 13 children less than 6 years of age with newly diagnosed medulloblastoma (Head Start I protocol). Patients received five cycles of induction chemotherapy with vincristine, etoposide, cisplatin, and cyclophosphamide, 3 weeks apart, followed by consolidation chemotherapy with a single myeloablative cycle of thiotepa, carboplatin, and etoposide. Irradiation was used only for residual tumor at consolidation or for progressive disease. Two-year overall survival was 62 % (95 % CI 35–89 %) with a 2-year progression-free survival of 38 % (95 % CI 11–65 %) (Mason et al. 1998b). In the follow-up study known as Head Start II, 21 children (<10 years of age) with newly diagnosed, high-risk disseminated medulloblastoma underwent induction with high-dose intravenous methotrexate in addition to the four drugs used in Head Start I protocol. Consolidation was the same as in the Head Start I study. The 3-year EFS of those high-risk patients was 49 % (95 % CI = 27–72 %), and overall survival was 60 % (95 % CI = 36–84 %) (Chi et al. 2004).
Strother et al. used full-dose craniospinal radiation (36 Gy) and posterior fossa boost, followed by four cycles of intensive chemotherapy with cisplatin, cyclophosphamide, and vincristine and stem cell rescue in 53 patients older than 3 years with newly diagnosed average- and high-risk medulloblastoma or supratentorial PNET. The 2-year progression-free survival rate was 93 % in the average-risk group and 74 % in the high-risk group (Strother et al. 2001). A similar approach was used by St. Jude Children’s Research Hospital; however, they used risk-adapted craniospinal radiotherapy (23.4 Gy for average-risk disease and 36–39.6 Gy for high-risk disease) in children older than 3 years with newly diagnosed medulloblastoma, followed by four cycles of intensive chemotherapy with cisplatin, cyclophosphamide, and vincristine. Autologous stem cell infusion was used in order to maintain dose intensity in patients who received previous craniospinal radiation. The 5-year overall survival was 85 % (95 % CI = 75–94 %) in the average-risk group and 70 % (95 % CI = 54–84 %) in the high-risk group. Five-year EFS was 83 % (95 % CI = 73–93 %) and 70 % (95 % CI = 55–85 %), respectively (Gajjar et al. 2006). COG 99703 study, conducted between 1998 and 2004, enrolled 92 infants <3 years of age with malignant brain tumors including medulloblastoma, glial tumors, and AT/RT tumors. This study used three rounds of intensive induction chemotherapy with cyclophosphamide, etoposide, vincristine, and cisplatin, followed by three rounds of high-dose chemotherapy with etoposide and thiotepa followed by autologous stem cell rescue after each round. Although the results of this study have not been published yet, the preliminary report indicates this approach to be superior to previously used five rounds of chemotherapy. This approach has become the backbone for future COG studies as well as a common clinical approach to treatment of young children <3 years of age with malignant brain tumors.
15.5.3 Gliomas
High-dose chemotherapy with autologous stem cell rescue has been studied in patients with recurrent as well as newly diagnosed high-grade gliomas. Two clinical trials using high-dose therapy following radiation for patients with diffuse pontine glioma showed no impact on survival, in comparison to historical controls (Bouffet et al. 2000). Heideman et al. treated 11 patients with newly diagnosed and recurrent high-grade astrocytoma with thiotepa and cyclophosphamide following surgery or biopsy. Radiation was given to patients who showed response or stable disease following high-dose chemotherapy. Although one complete response and two partial responses were observed, median progression-free survival remained disappointingly low at 9 months (Heideman et al. 1993). The Children’s Cancer Group reported on 18 patients with recurrent malignant astrocytoma/glioblastoma multiforme treated with thiotepa and etoposide and autologous stem cell rescue, five of whom had progression-free survival ranging from 39 to 59 months (Finlay et al. 1996). A follow-up phase II pilot study conducted by the Children’s Cancer Group added carmustine to thiotepa and etoposide, followed by radiation therapy for newly diagnosed glioblastoma multiforme. Although 2-year progression-free survival was promising at 46 %, accrual was closed early because of unacceptable pulmonary and neurologic toxicities (Grovas et al. 1999). In a subsequent study, 27 children <21 years of age with malignant astrocytomas were treated with myeloablative chemotherapy following initial tumor progression (Finlay et al. 2008). The conditioning regimens included thiotepa and etoposide, or thiotepa and etoposide, preceded by carmustine or carboplatin. Treatment-related death continued to be high in this group of patients (18.5 %); however, five patients (18 %) remained alive at a median of 11.1 years after transplant.
Papadakis et al. published a large series of children with newly diagnosed malignant gliomas, who were treated with high-dose carmustine, thiotepa, and etoposide and autologous bone marrow rescue. This treatment was given following surgery and local radiation. Out of 29 patients with gliomas, four died from toxicity, and three (10 %) were alive without evidence of disease or with stable disease at 64–86 months posttransplant (Papadakis et al. 2000). High-dose chemotherapy was used also in patients with newly diagnosed high-grade gliomas. Massimino et al. treated 21 pediatric patients with a combination of induction chemotherapy with cisplatin, etoposide, cyclophosphamide, and high-dose methotrexate, followed by high-dose thiotepa with stem cell rescue. The myeloablative cycle was given a second time if patients had residual disease after the first round. After high-dose chemotherapy, patients received radiation and 27 weeks of maintenance therapy with vincristine and lomustine. With this approach, at a median follow-up of 57 months, progression-free survival was 46 %, and overall survival was 43 % (Massimino et al. 2005).
Although high-dose chemotherapy may be promising in high-grade glial tumors, due to the inconsistency of results from studies published so far and reported high treatment-related mortality, use of high-dose chemotherapy with autologous stem cell rescue should be considered experimental for patients with glial tumors and used only in the context of clinical research.
15.5.4 Other Tumor Types
High-dose chemotherapy is not superior to standard chemotherapy in patients with recurrent or newly diagnosed ependymoma as confirmed in Children’s Oncology Study and Head Start I–III studies (Mason et al. 1998b; Zacharoulis et al. 2007).
In a study of recurrent, noncerebellar PNETs, which included pineoblastoma (n = 8), Broniscer et al. describe 17 patients (age 0.9–31.4 years) who were treated with high-dose chemotherapy of carboplatin and etoposide and autologous stem cell rescue. Eleven percent of patients died of toxicity. EFS was 29 % (surviving patients were followed for 40–123 months) (Broniscer et al. 2004). Sung et al. described seven patients with supratentorial PNET (three newly diagnosed and four with recurrent tumors) who were treated with a double-transplant approach using cyclophosphamide and melphalan for the first transplant and carboplatin, thiotepa, and etoposide for the second transplant. Two out of three patients with newly diagnosed PNET and one out of four with recurrent disease remained disease-free (14–31 months follow-up) (Sung et al. 2007). A number of smaller studies that included up to six patients with recurrent supratentorial PNET indicated salvage rates of approximately 25 % with high-dose chemotherapy. However, reported length of follow-up was frequently short (Kalifa et al. 1992; Mahoney et al. 1996; Busca et al. 1997; Graham et al. 1997; Fleischhack et al. 1998; Mikaeloff et al. 1998).
High-dose chemotherapy used in the first-line treatment of pineoblastoma showed quite promising results. Gururangan et al. reported on 12 patients with newly diagnosed pineoblastoma (age 0.3–43.7 years), who were treated with surgery, radiation (given to all but two patients), and high-dose chemotherapy with cyclophosphamide, melphalan, and busulfan. Four-year progression-free survival was 69 % (Gururangan et al. 2003).
Forty-three children with newly diagnosed supratentorial PNET were treated on the Head Start I and II studies using five rounds of induction followed by high-dose consolidation therapy with etoposide, carboplatin, and thiotepa. Five-year EFS was 39 % and overall survival was 49 %. Nonpineal supratentorial PNET (sPNET) patients fared significantly better than patients with pineal sPNETs. Twelve of 20 survivors never received radiation therapy (Fangusaro et al. 2008). Finally, children less than 3 years of age with sPNET were included in the CCG 99703 study, which used three rounds of induction therapy followed by three rounds of consolidation with carboplatin and thiotepa.
Similar to its use in medulloblastoma, high-dose chemotherapy has efficacy in patients with supratentorial PNET. In the studies published so far, outcomes were best when high-dose chemotherapy was used as a first-line therapy and in combination with radiation.
AT/RT is another embryonal malignant brain tumor that is typically sensitive to chemotherapy. Due to its poor prognosis in children, this tumor is often included in infant brain studies and treated with up-front high-dose chemotherapy. A report from Head Start III protocol (Zaky et al. 2014) described outcomes of 19 children younger than 3 years of age, with newly diagnosed CNS AT/RT tumors treated with five cycles of induction chemotherapy with cyclophosphamide, vincristine, etoposide, cisplatin, and high-dose methotrexate and consolidation with a single cycle of thiotepa, etoposide, and carboplatin. Eleven out of 19 children progressed early, at a median time of 4.1 months, and only four out of 19 children were without evidence of disease at 40–79 months from the diagnosis. The authors concluded that more effective approaches are required in young children with AT/RT. Similar to these findings, European data suggested that radiation therapy significantly increased the mean survival time in infants with AT/RT treated with high-dose chemotherapy and methotrexate (Seeringer et al. 2014). Thus high-dose chemotherapy with stem cell rescue is not recommended as a sole modality of treatment in young children with AT/RT (Seeringer et al. 2014). Studies of high-dose chemotherapy in CNS germ cell tumors are rare as this tumor responds well to standard chemotherapy and radiation therapy. However, if tumor recurs following both modalities, high-dose chemotherapy can be effective. A group of Japanese investigators reported a small series of six patients with intracranial nongerminomatous germ cell tumors treated with myeloablative chemotherapy alone, all of whom survived without tumor recurrence at follow-up of 1–7 years (Tada et al. 1999). Another study described 21 patients with CNS germ cell tumors that progressed following initial chemotherapy and radiation. These patients were treated with a thiotepa-based conditioning regimen. The response was very good in patients with germinoma as 7/9 patients survived disease-free with a median follow-up of 48 months. However, in the nongerminomatous germ cell tumor category, only 33 % patients survived without disease progression (Modak et al. 2004).
Although the toxicity of stem cell transplant has significantly decreased over the last 10 years, some studies of transplant in children with brain tumors still report significant treatment-related mortality (11–19 %) (Sung et al. 2007; Dhall et al. 2008). Over the last 20 years, high-dose chemotherapy has been established as an important modality for treating recurrent, chemosensitive brain tumors as well as first-line treatment for malignant brain tumors in young children.
15.6 New Strategies
15.6.1 Radiation Sensitizers
Numerous investigational chemotherapy treatment strategies are designed to maximize the known benefit of radiation therapy in CNS tumors. Multiple potentiating effects of platinum agents on radiation have been described. Hypoxic cells (typically radiation resistant) are more sensitive to radiation following exposure to platinum agents in vitro (Skov and MacPhail 1991). Platinum agents may also have a role in preventing the development of radiation-resistant clones by inhibiting “potentially lethal damage recovery,” a mechanism by which tumor cells are able to repair what would otherwise be lethal or sublethal DNA damage following radiation exposure (Wilkins et al. 1993). A phase 2 study piloting the use of daily carboplatin in combination with weekly vincristine and radiation in patients with high-risk medulloblastoma reported promising outcomes with overall and progression-free survival of 83 and 71 %, suggesting that this is a meaningful strategy (Jakacki et al. 2012). An ongoing phase 3 trial conducted by the Children’s Oncology Group is investigating the role of combination carboplatin, vincristine, and radiation in comparison with vincristine and radiation alone in high-risk medulloblastoma (clinicaltrials.gov; NCT00392327).
Inhibitors of the enzyme histone deacetylase (HDAC) have also been shown to have radiation-sensitizing effects in preclinical models of malignant glioma (Chinnaiyan et al. 2008). The combination of radiation with vorinostat is currently under investigation in clinical trials for malignant glioma and diffuse pontine glioma (clinicaltrials.gov; NCT01189266).
15.6.2 Targeting Drug Resistance
Many brain tumors are resistant to conventional chemotherapeutic agents. Investigators have identified mechanisms of drug resistance amenable to pharmacologic treatment that would render tumor cells more sensitive to chemotherapy.
15.6.2.1 P-Glycoprotein Pump
The P-glycoprotein pump (PGP) is the protein product of the multidrug resistance gene (MDR–1), which is amplified in many resistant and refractory tumors, including glioblastoma, medulloblastoma, and ependymoma (Chou et al. 1995, 1996; von Bossanyi et al. 1997; Decleves et al. 2002). The PGP serves as an efflux pump, allowing the cell to transport specific toxins, including chemotherapeutic agents (Sikic et al. 1997). In mice, capillary endothelial cells composing the BBB have high concentrations of PGP (Schinkel et al. 1994). Cyclosporine A is a potent inhibitor of PGP and effectively sensitizes high PGP-expressing cells in vitro (Sikic et al. 1997). Clinical trials using cyclosporine A in combination with chemotherapy, largely in the setting of adult myeloid leukemia, have shown high toxicity with unclear therapeutic benefit (Chauncey 2001). A phase I clinical trial in pediatric CNS tumors consisted of intravenous cyclosporine A in combination with oral etoposide, intravenous vincristine, and radiation therapy for patients with intrinsic pontine glioma. The trial was halted early due to excessive neurotoxicity, and there was no survival advantage for the few evaluable patients (Greenberg et al. 2005).
15.6.2.2 Alkylguanine-DNA-Alkyltransferase
Alkylguanine-DNA-alkyltransferase (AGT) is a DNA repair enzyme that plays an important role in tumor resistance to alkylnitrosoureas and temozolomide. AGT reverses DNA methylation and chloroethylation (induced by chemotherapy) at the O6 position of guanine, thus rescuing the cell from lethal injury. Many brain tumors have high levels of AGT, and these high levels are associated with poor survival in clinical trials in adults with malignant glioma (Wiestler et al. 1984; Pegg 1990; Pegg and Byers 1992; Hongeng et al. 1997). A recent trial in adult patients with malignant glioma showed that patients with methylated AGT gene promoters (and thus decreased AGT expression) had a better response to chemotherapy with the alkylating agent temozolomide (Hegi et al. 2005). Experiments in brain tumor cell lines as well as tumor xenografts have shown that depletion of AGT with O6-benzylguanine (acting as an alternate substrate for AGT) increases tumor cell sensitivity to chemotherapy (Jaeckle et al. 1998). A phase I clinical trial in malignant glioma of O6-benzylguanine administered preoperatively as a single agent showed reduced levels of AGT in the resected tumor, suggesting that combination treatment with O6-benzylguanine and nitrosourea or temozolomide would increase tumor cell sensitivity (Friedman et al. 1998). A phase I clinical trial of temozolomide in combination with O6-benzylguanine for recurrent or refractory pediatric brain tumors was recently completed through the Pediatric Brain Tumor Consortium. This study demonstrated that this regimen was well tolerated and also established modest activity, with three patients with recurrent glioma having partial responses (Broniscer et al. 2007).
15.6.2.3 PARP Inhibition
Poly (ADP-ribose) polymerases (PARPs) are enzymes with wide-ranging cellular functions, including regulation of DNA transcription, cell cycling, and DNA repair. Cancer cell resistance to DNA-damaging agents is thought to be in part related to enhanced DNA repair mediated by PARP. PARP overexpression has been observed in a number of pediatric brain tumor subtypes including DIPG and malignant glioma, leading to the hypothesis that inhibition of PARP may be a therapeutic strategy for chemotherapy-resistant tumors. The PARP inhibitor veliparib was shown to be tolerable in combination with temozolomide in patients with recurrent CNS tumors in a phase 1 study (Su et al. 2014). Although no objective responses were observed, four patients had stable disease greater than 6 months. The efficacy of this approach is under investigation in a phase 2 study of pediatric patients with diffuse pontine glioma (clinicaltrials.gov NCT01514201).
15.6.3 Molecular Targets and Signal Transduction Inhibition
The enormous expansion in the understanding of the molecular basis of oncogenesis has led to the development of a new class of agents designed to inhibit specific intracellular biochemical pathways. The majority of these agents function by inhibiting receptor tyrosine kinases, a class of cellular proteins that bind a specific ligand through their extracellular domains. Activation of the intracellular tyrosine kinase catalytic domain of the receptor after ligand binding subsequently triggers a cascade of biochemical signals. This ligand-dependent tyrosine kinase activation mediates a host of cellular properties, including proliferation, survival, and differentiation. In normal cellular homeostasis, these functions are tightly regulated. In tumors, unregulated, ligand-independent kinase phosphorylation and subsequent receptor activation is a common event and is likely a key mechanism in maintaining the malignant phenotype. Receptor tyrosine kinases known to be important in CNS tumors include the epidermal growth factor receptor EGFR, platelet-derived growth factor receptor (PDGFR), and vascular endothelial growth factor receptor (VEGFR). Important downstream effector kinases are also relevant therapeutic targets. Of particular interest in pediatric tumors are drugs targeting BRAF or its downstream effector MEK, as a significant proportion of pediatric astrocytomas have aberrant signaling of this pathway related to genetic alterations in BRAF or NF1. Also of interest in pediatric astrocytoma is inhibition of mTOR, as preclinical studies demonstrate that this pathway is also activated in some tumors. The biology of these pathways are discussed in greater detail in Chap. 16.
The most notable clinical success of therapeutic molecular targeting of brain tumor is for the low-grade glioma variant subependymal giant cell astrocytoma (SEGA) in the setting of tuberous sclerosis (TS). Tuberous sclerosis is caused by genetic mutations in the TSC1/TSC2 gene complex and results in a predisposition to subependymal giant cell astrocytoma. Loss of TSC1/TSC2 function results in constitutive activation of mTOR signaling (Franz et al. 2012). There is strong clinical evidence demonstrating that inhibition of mTOR signaling with everolimus results in objective tumor response, and everolimus is FDA approved for the treatment of SEGA (Krueger et al. 2010). This experience provides proof of principle that CNS tumors can be treated with biologically targeted agents.
Genetic aberrations of BRAF are frequently observed in pediatric high-grade and low-grade gliomas that result in activation of the Ras–MAP signaling kinase pathway, an event which drives neoplastic cell behavior (see Chap. 1). Objective responses have been observed in tumors bearing a BRAF V600E mutation treated with the BRAF inhibitors vemurafenib and dabrafenib (Bautista et al. 2014; Shih et al. 2014; Usubalieva et al. 2015). Targeting these abnormalities with pharmacologic inhibitors of BRAF and its downstream effector MEK is currently being evaluated in clinical trials (clinicaltrials.gov NCT01748149, NCT01089101).
A subset of medulloblastoma have alterations in the sonic hedgehog (SHH) signaling pathway. Both upstream and downstream mutations in the pathway have been described in multiple genes in the pathway (Samkari et al. 2015). Aberrant signaling of one of the members of this pathway, SMO, can be targeted with vismodegib, and responses have been described in medulloblastoma patients treated with this agent (Robinson et al. 2015). The heterogeneous nature of the SHH pathway aberrations in medulloblastoma, however, suggests that the target tumor population who may benefit from this therapy needs to be well defined with genetic testing, as patients with activating mutation downstream of the target may not be susceptible to SMO inhibitors. A clinical trial of vismodegib in combination with conventional chemoradiotherapy is ongoing (clinicaltrials.gov NCT01878617).
15.6.4 Angiogenesis Inhibitors
The role of angiogenesis in supporting tumor cell proliferation and survival has been extensively investigated since the hypothesis of “angiogenesis dependency” of tumors was first proposed by Judah Folkmann (1971) and Balis et al. (2002). The angiogenesis hypothesis proposes that tumor-induced proliferation of blood vessels is necessary to support ongoing proliferation and survival. This deceptively straightforward statement must be tempered by the fact that tumor cell-induced angiogenesis appears to have multiple mechanisms, some of which are redundant. Tumor cells produce proangiogenic cytokines, including acidic and basic fibroblastic growth factor (aFGF, bFGF), angiogenin, vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF), and interleukin-8 (IL-8). Additionally, tumor cells produce matrix metalloproteinases (MMP), which can induce breakdown in extracellular matrix, again allowing for release of proangiogenic peptides. Finally, tumor cells recruit inflammatory cells that subsequently produce proangiogenic cytokines (Bicknell and Harris 1996; Pluda 1997; Paku 1998).
A number of therapeutic agents have been shown to inhibit neovascularization in vitro and in vivo. Thalidomide, initially developed as a sedative and subsequently found to be a potent teratogen in humans, has antiangiogenic properties (D’Amato et al. 1994). It has been shown to inhibit bFGF-induced corneal vascularization in animals (Kenyon et al. 1997). In a single-agent clinical trial of recurrent gliomas, objective response rates (partial response and stable disease) of up to 45 % were reported (Fine et al. 2000). Lenalidomide is an analog of thalidomide that has shown considerable efficacy in multiple myeloma and myelodysplasia. A phase 1 study in children demonstrated that lenalidomide was well tolerated in patients with recurrent, refractory, or progressive CNS tumors (Fine et al. 2007). Additionally, objective responses and prolonged stable disease were observed in patients with low-grade glioma (Warren et al. 2011). These findings are under further investigation in a phase 2 study of lenalidomide for children with pilocytic astrocytoma and optic pathway glioma (clincialtrials.gov NCT01553149).
The role of cyclooxygenase inhibitors as an antiangiogenic treatment strategy is also under investigation. Cyclooxygenase 2 (COX-2) induces vascular proliferation following trauma or stimulation with growth factor and is highly expressed in some human tumors, including high-grade glioma (Joki et al. 2000). Treatment of glioma cell cultures with a specific COX-2 inhibitor was found to produce diminished proliferation and invasion and increased apoptosis (Joki et al. 2000). Celecoxib is a specific inhibitor of COX-2, with a highly favorable toxicity profile. While celecoxib is not currently under investigation in large-scale pediatric clinical trials, its ease of use and low toxicity make it an appealing agent to pursue in combination with other agents.
The most successful strategy in adults thus far has been with bevacizumab, a monoclonal antibody targeting VEGF, in combination with conventional cytotoxic chemotherapy such as irinotecan (Vredenburgh et al. 2007) or temozolomide (Hofland et al. 2014). This combination is active in pediatric low-grade glioma and is reported to have similar progression-free survival to more conventional carboplatin- or nitrosourea-based regimens in a phase 2 study in tumors that had failed prior treatment (Gururangan et al. 2014). Experience with this regimen in pediatric high-grade glioma, diffuse pontine glioma, and medulloblastoma, however, has been disappointing, without an appreciable improvement in survival. Interestingly, bevacizumab may be an alternate treatment for tumor-associated edema, as transient symptomatic improvement has been observed after its use in patients with high-grade glioma. Toxicities of bevacizumab include fatigue, proteinuria, and hypertension. Less common but of concern is an increased risk of hemorrhage. Finally, the use of conventional cytotoxic agents given in low-dose, metronomic regimens is being piloted. These regimens are based on the principle that while endothelial cell proliferation appears to be sensitive to chronic but low-dose exposure, it has ample recovery time during the recovery phase of dose-intensive therapy schedules. Evidence from pilot clinical trials supports this hypothesis (Einhorn 1991; Ashley et al. 1996; Kushner et al. 1999; Klement et al. 2000). Kieran et al. described the use of metronomic chemotherapy with daily oral thalidomide and celecoxib, in addition to alternating cycles of daily oral etoposide and oral cyclophosphamide in the treatment of 20 children with recurrent or progressive cancer. The regimen was well-tolerated and prolonged progression-free survival in a number of children with CNS malignancies (Kieran et al. 2005).
15.6.5 Overcoming the Blood–Brain Barrier
The BBB is composed of tight endothelial cell junctions that exclude most large molecules and is freely permeable only to small molecules that are highly lipophilic. This barrier limits the ability of many systemically administered chemotherapeutic agents to penetrate the CNS. Radiographic evidence based on heterogeneous uptake of gadolinium on magnetic resonance imaging suggests, however, that the BBB is only partially intact in many patients with CNS tumors. Further support of tumor degradation of the BBB lies in the responsiveness of tumors to large, water-soluble molecules such as the platinum agents. Nevertheless, resistance of CNS tumors to therapy may partially lay in the infiltrative, nonenhancing portions of tumor that presumably have an intact BBB and thus are able to escape cytotoxicity of systemically administered agents. This is supported by the propensity of many tumors to recur locally, at the infiltrating edge of the tumor. A number of strategies are under investigation with the intent to disrupt or bypass the BBB.
15.6.5.1 Blood–Brain Barrier Disruption
Mannitol was one of the first agents used to attempt disruption of the BBB. Increased osmotic pressure transiently opens the BBB, allowing entry of molecules, otherwise unable to penetrate the CNS (Neuwelt et al. 1983). Increased disease response has been reported following BBB disruption, largely in adult patients with non-AIDS CNS lymphoma (Neuwelt et al. 1981). The Children’s Cancer Group reported on the only pediatric experience with this strategy, using mannitol in combination with etoposide for recurrent or refractory CNS tumors. They were unable to document a clear benefit (Kobrinsky et al. 1999).
RMP-7 is a bradykinin analog, which, on binding to specific B2 bradykinin receptors on the surface of endothelial cells, transiently increases permeability of the BBB. While increased concentration of carboplatin when administered with this agent has been documented in animals (Dean et al. 1999), a randomized, placebo-controlled phase II trial in adults with malignant glioma showed no survival benefit from the addition of RMP-7 to carboplatin alone (Prados et al. 2003). A phase II pediatric trial was conducted by the Children’s Oncology Group for patients with recurrent or refractory brain tumors (Warren et al. 2006) and did not show any activity in children with brainstem gliomas or high-grade gliomas.
15.6.5.2 Intra-arterial Delivery
Intra-arterial delivery of chemotherapy, often delivered in conjunction with mannitol, may improve delivery of drug and minimize systemic toxicity, possibly allowing for the use of lower doses delivered directly to the tumor. A number of studies have investigated the use in intra-arterial carmustine, cisplatin, and carboplatin. While modest responses have been reported, neurologic toxicities are substantial, including irreversible encephalopathy and vision loss (Bashir et al. 1988; Mahaley et al. 1989; Newton et al. 1989).
15.6.5.3 Intratumoral Drug Delivery
A variety of novel techniques to deliver drug directly to the tumor or resection cavity are under investigation. The use of carmustine-impregnated “wafers” in adults with malignant glioma has been reported. This strategy allows for the passive diffusion of high concentrations of carmustine from wafers surgically implanted in the resection cavity to surrounding tumor cells, with minimal systemic exposure. Modest improvements in survival have been shown with this intervention (Brem et al. 1995; Valtonen et al. 1997). A multi-institutional pediatric trial designed to investigate the impact of this treatment in children with recurrent high-grade glioma closed early due to poor accrual, suggesting that this strategy may have feasibility concerns in this patient population (clinicaltrials.gov NCT0004572).
Passive diffusion, however, is limited by minimal ability of the drug to penetrate beyond the margin of the tumor resection cavity. Convection-enhanced delivery (CED) is a novel delivery strategy that overcomes this barrier and allows for delivery of larger molecules. CED requires the surgical placement of catheters intra- or peritumorally, through which a therapeutic agent is infused under positive pressure (Bobo et al. 1994). This allows for a substantially larger area of the brain to be treated. Multiple agents have been tested in early-phase trials for adult patients with newly diagnosed or recurrent malignant glioma using this technique, including conventional chemotherapy drugs such as carboplatin, cell surface-targeted cytotoxin-ligand conjugates, monoclonal antibodies with or without conjugated radioisotopes (Bigner et al. 1998), antisense oligonucleotides, and liposomal vectors to deliver gene therapy. While some of these strategies have resulted in preliminary signals of efficacy, none have shown superiority over conventional treatment in a randomized setting. In addition, there are significant technical barriers associated with catheter placement (Vogelbaum and Aghi 2015).
15.6.6 Differentiation of Neoplastic Cells
Agents that induce differentiation of tumor cells, thereby suppressing neoplastic proliferation, may have a role in the management of brain tumors. Experiments in cell culture using both retinoic acid and phenylacetate show both differentiation and inhibition of proliferation of astrocytoma-derived and medulloblastoma-derived cell lines (Mukherjee and Das 1990; Rodts and Black 1994). Treatment of adult patients with malignant glioma with single-agent 13-cis-retinoic acid showed a modest partial response plus a stable disease rate of 46 %, with tolerable toxicity (Yung et al. 1996). A recent phase II study investigated the combination of temozolomide with 13-cis-retinoic acid for recurrent malignant glioma in adults. A slight improvement in a 6-month progression-free survival was observed over historical controls, suggesting that the combination is active in recurrent malignant glioma (Jaeckle et al. 2003). Another recent phase II trial used 13-cis-retinoic acid as maintenance therapy for adult patients with high-grade glioma after first-line multimodal therapy (Wismeth et al. 2004). This approach was well tolerated and resulted in a median survival of 74 weeks. In a randomized trial adding retinoic acid to combination therapy for high-risk neuroblastoma, patients treated with retinoic acid had better outcomes (Matthay et al. 1999). Preclinical data in medulloblastoma mouse models supports this strategy, and it is being tested in a phase 3 trial for newly diagnosed high-risk medulloblastoma patients, in which patients are randomized to receive 13-cis-retinoic acid after standard treatment with surgery, radiation, and chemotherapy (Spiller et al. 2008; clinicaltrials.gov NCT00392327).
15.6.7 Immunotherapy
The goal of immunologically directed antitumor therapy is to eradicate tumor cells either by stimulating host immunologic antitumor reactions or by blocking tumor-related local immunosuppression, and recent progress has been made with immunotherapy approaches in childhood leukemia and melanoma. Chimeric antigen receptor (CAR) T-cell therapy has resulted in durable remissions in CD19-positive B lineage leukemia (Lee et al. 2015). Inhibition of the T-cell antigen PD-1 along with inhibition of CTLA-4 has resulted in remarkable improvements in survival in metastatic melanoma (Larkin et al. 2015). The role of immunotherapy for CNS tumors, however, remains under investigation.
Antigen-targeted tumor vaccination is showing early promise in CNS tumors. Pollack et al. reported immunologic and radiographic responses with the use of a multiple glioma-associated antigen vaccination in combination with radiation and chemotherapy in children newly diagnosed with brainstem glioma and high-grade glioma (Pollack et al. 2014). Alternative vaccine approaches are under investigation in adults with glioblastoma. Heat shock proteins act as molecular chaperones inside cells and have been found to bind a unique “fingerprint” of peptides that are tumor specific. These HSPCCs have been found to be antigenic in human trials with a variety of tumors and function by eliciting T-cell-mediated cytotoxicity. A phase 2 study of a tumor-derived heat shock protein/protein complex (HSPPC) vaccine in adults with surgically resectable, recurrent GBM demonstrates 6-month PFS of approximately 20 %.
Dendritic cell therapy involves the collection of antigen-presenting cells followed by an ex vivo pulse with autologous tumor lysate or more specific tumor peptides. These cells are then administered intradermally or intratumorally and stimulate host cytotoxic T cells to attack the tumor. Numerous dendritic cell vaccination approaches have been tested in clinical trials. Tumor lysate-pulsed DC vaccinations have been shown to be safe, and preliminary phase 1 and phase 2 studies suggest that newly diagnosed, adult GBM patients have equivalent or improved survival compared to historical controls, and phase 3 trials of this strategy are ongoing in adults. Glioma-associated antigen-pulsed dendritic cell vaccines are also under investigation (Antonios et al. 2015). Although relatively few immunotherapies have entered pediatric clinical trials, ongoing research in this field carries future promise.
15.7 Conclusions
While prognosis for malignant brain tumors in children has improved somewhat in the past several decades, 5-year survival for all tumors except low-grade astrocytoma remains suboptimal at 60 %. The incorporation of chemotherapy into pediatric brain tumor management has allowed for advances in survival and reduction of morbidity and is now the standard of care for many childhood brain tumors. Critical to the further improvement of prognosis and long-term outcome is the continued effort of multi-institutional, cooperative group clinical trials. The largest clinical trial group conducting pediatric brain tumor trials is the National Cancer Institute (NCI)-sponsored Children’s Oncology Group. The Children’s Oncology Group includes the majority of pediatric cancer treatment centers in the United States and incorporates programs in Canada, Europe, and Australia. Research activities include clinical trials for the majority of newly diagnosed and recurrent brain tumors, as well as studies of new agents. The Pediatric Brain Tumor Consortium is a smaller, NCI-sponsored consortium with a mission to expedite the development of new agents’ high-risk pediatric brain tumors by bringing novel agents and translational research to the pediatric brain tumor community in a multi-institutional setting. The Pacific Pediatric Neuro-oncology Consortium (http://www.pnoc.us) is another multi-institutional clinical trials group dedicated to investigation of new therapies for childhood brain tumor. Information on the Children’s Oncology Group and the Pediatric Brain Tumor Consortium can be found on the World Wide Web at http://www.childrensoncologygroup.org and http://www.pbtc.org.
References
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
