Chapter 46: Pediatric Cancers

Fortunately, pediatric cancers are rare, with only 1 in 300 children being diagnosed before 18 years of age. And, with overall survival (OS) approaching 80%, there is hope for most of these children and for their families. However, just as in adults, if their cancer is metastatic at diagnosis, if the cancer does not respond to standard therapies, or if they suffer a relapse, the prognosis is universally grim. Despite intensified therapies, the survival for most children with relapsed cancer has not improved in decades. In large part, this is due to the toxicities of such regimens as we have likely reached the tolerable limit with most chemotherapeutic agents. Thus, as is the case for adults with cancer, there has been a focus on understanding the underlying biology of the disease to find targetable lesions and to develop novel agents and treatment regimens to improve survival in children with relapse. However, there are multiple challenges we face. First, the spectrum of pediatric cancers is distinctly different from adults (Fig. 46-1), with acute lymphoblastic leukemia (ALL), medulloblastoma and gliomas, neuroblastoma, Hodgkin and non-Hodgkin lymphomas, and sarcomas the most common cancers in children as opposed to the most common carcinomas of the prostate, breast, lung, colon, and so on in adults. Second, the relative rarity of these cancers results in less preclinical research to find and develop potential therapeutic targets and limited patient numbers to enroll and test novel therapeutic strategies.

One of our approaches has been to study the pediatric cancers in parallel with similar adult tumors and to enroll patients on early adult trials to give children access to promising agents and to provide some data on safety and efficacy to support the development of pediatric-specific trials. In this chapter, we present a few examples of tumor types seen commonly in children and describe some of the therapeutic advances and promising strategies for treating children with relapsed cancers.

Acute Lymphoblastic Leukemia

Acute lymphoblastic leukemia is the most common cancer diagnosis in children, accounting for 25% of cancer diagnoses in children 15 years old or less. An estimated 3,000 new cases of childhood ALL are diagnosed yearly in the United States. After a peak incidence of 90 cases per million per year at age 2 to 3 years, ALL incidence rates decrease steadily into adolescence. Initial complete remission (CR) rates are 95%, and survival in childhood ALL is approaching 90% through the application of reliable prognostic factors that permit use of risk-oriented treatment protocols. However, relapse occurs in approximately 20%, with higher rates of relapse in adolescents and young adults as well as children less than 1 year of age (ie, “infants”). Despite excellent outcomes overall, relapsed patients with ALL outnumber nearly all other childhood malignancies. With traditional intensive combination chemotherapy and allogeneic hematopoietic stem cell transplantation, 30% to 40% of all children with relapsed ALL can be cured. The factors that effect salvage rate are timing of relapse (18-36 months from diagnosis), site of relapse (bone marrow, central nervous system [CNS]/testicular, combined), and immunophenotype (T cell vs B cell). Unfortunately, most children still die from relapsed ALL despite aggressive chemoradiotherapy approaches, including transplantation, and novel salvage regimens are needed (¹).

For children with first relapse of childhood ALL, the mitoxantrone-based Medical Research Council (MRC) ALL R3 relapsed protocol has relatively better outcomes than other common regimens and is currently our standard reinduction regimen (²). Promising CR2 rates have also been seen with the addition of the proteasome inhibitor bortezomib to the MRC ALL R3 backbone, with 80% (16/20) CR/CRi in B-ALL patients (³), as well as the addition of bortezomib to the more traditional four-drug reinduction regimen in COG AALL07P1, with 64% (63/99) of B-ALL patients achieving CR2. Surprisingly, AALL07P1 showed similar CR2 rates of 68% (15/22) in T-ALL patients (⁴). The current first relapse regimen under study in the COG is testing the substitution of the CD19-targeting bispecific engaging antibody blinatumomab to the MRC ALL R3 relapsed protocol.

With attainment of CR2 in 70% to 85% (<1% minimal residual disease by flow cytometry), we offer transplant to most children with relapsed ALL, with very late relapses and isolated CNS relapses as typical exceptions. We offer a variety of transplant trials, including cord blood and haploidentical, as well as addition of anti-CD19 chimeric antigen receptor (CAR) T cells, natural killer (NK) cells, and so on. Failure to respond adequately to first-relapse regimens leads to a variety of second-line regimens. We commonly offer clofarabine/cyclophosphamide/etoposide (CR2 rates of 44% ^[5]), mitoxantrone/cytarabine (CR2 rate of 57% ^[6]) or one of several targeted agents with the institutional adult ALL regimen hyper-CVAD (⁷), or promising single agents when available. With recently published complete responses in children (⁸), we are testing the combination of anti-CD22 immunotoxin inotuzumab with hyper-CVAD (without anthracycline), with some dramatic responses in children. In addition, the combination of the mTOR inhibitor everolimus and Hyper-CVAD is being tested. For T-cell ALL specifically, we have used nelarabine in combination with hyper-CVAD (⁹), a trial testing a 5-day continuous infusion of nelarabine (NCT01094860), and a single agent Notch-inhibiting gamma secretase inhibitor BMS-906024 phase I trial (NCT01363817), based on multiple responses in adults (¹⁰).

Targeting ALL blasts with antibodies, both unconjugated and conjugated to toxins, has been shown to be an effective therapeutic strategy. Monoclonal antibodies to surface antigens such as CD19, CD20, CD22, and CD52 have been used in unconjugated form (eg, rituximab and epratuzumab), conjugated to immunotoxins or chemotherapeutic drugs (eg, moxetumomab, inotuzumab ozogamicin), or in the form of bispecific antibodies (eg, blinatumomab). The incorporation of rituximab (CD20) into the hyper-CVAD regimen showed improved outcome in adults younger than 60 years with more than 5% CD20-positive, BCR-ABL1–negative B lymphoblastic leukemia (¹¹). Recent significant success with the anti-CD19–anti-CD3 bispecific T-cell engaging (BiTE) antibody blinatumomab has led to responses in about 50% of patients with relapsed pediatric B-ALL.

Recent therapeutic advances in ALL have also included the use of adoptive immunotherapies. Anti-CD19 CAR-expressing T cells have shown dramatic success in patients with relapsed B-ALL, with up to 90% remission rates reported (¹²). We have developed a unique nonretroviral approach utilizing the Sleeping Beauty transposon system to transfect T cells with CAR constructs, opening this technology to rapid translation of targetable tumor markers (¹³).

Patients with BCR-ABL1–positive ALL had poor prognoses. When tyrosine kinase inhibitors targeting ABL1 are added to multidrug chemotherapy, CR rates are more than 90%, and event-free survival (EFS) is superior to that in historical controls. Recent work has discovered that a subset (~10%) of children with BCR-ABL1–negative ALL have gene expression patterns similar to those with BCR-ABL1–positive ALL. Most children and adults with this “Ph-like” ALL harbor a range of novel kinase translocations (including ABL1, ABL2, PDGFRB, EPOR, JAK2, JAK1, etc.), some of which are targetable clinically. Thus, we and others have begun to screen for these lesions and to target these cases with appropriate kinase inhibitors (dasatinib, ruxolitinib) and chemotherapy in the clinical trial setting (^14,15).

Hodgkin Lymphoma

Hodgkin lymphoma (HL) is one of the first malignancies cured by a combination of chemotherapy and radiation therapy (RT). With appropriate staging and evaluation of treatment response using positron emission tomographic (PET) scanning, current treatment for HL achieves a 5-year EFS of 80% and 5-year OS of more than 90% (¹⁶). The standard of care for patients with relapsed or refractory HL is salvage chemotherapy followed by autologous stem cell transplantation (SCT), which can induce long-term remissions in approximately 50% of patients. The malignant Hodgkin’s Reed-Sternberg cells of classical HL are characterized by the expression of CD30, a member of the tumor necrosis factor T cells, and eosinophils; it represents an ideal target for monoclonal antibody therapy. Brentuximab vedotin (SGN-35) is an antibody-drug conjugate (ADC) comprising an anti-CD30 antibody conjugated by a protease-cleavable linker to the potent antimicrotubule agent monomethyl auristatin E (MMAE). Binding of the ADC to CD30 on the cell surface initiates internalization of the ADC-CD30 complex, which then disrupts the microtubule network, induces cell cycle arrest, and results in apoptotic death of the CD30-expressing tumor cell. Brentuximab vedotin is currently approved by the Food and Drug Administration (FDA) for HL after failure of autologous SCT or failure of two chemotherapy regimens with patients who are not candidates for autologous SCT (¹⁷). Use of brentuximab vedotin at an earlier treatment time is under investigation with multiple clinical trials. Phase I study of brentuximab vedotin with standard chemotherapy of doxorubicin, bleomycin, vincristine, and decarbazine (ABVD) and modified standard chemotherapy with doxorubicin, vincrisitine, and decarbazine (AVD) as first-line treatment for advanced-stage HL showed significantly increased pulmonary toxicity associated with the combination of brentuximab vedotin and bleomycin. This study confirmed the dose of brentuximab vedotin was safely escalated to 1.2 mg/kg and showed a remarkable response rate of 95% to 96% to the therapy (¹⁸).

Osteosarcoma

Osteosarcoma is the most common malignant bone tumor of childhood. Osteosarcoma typically affects pubertal adolescents, with a peak incidence in childhood of 12 years. Risk stratification can be made using tumor stage, location, and response to therapy. Survival of patients with nonmetastatic disease at the time of diagnosis has improved dramatically over the past 30 years due to advances in chemotherapy and surgery, with 60% to 65% of patients surviving more than 10 years. However, patients who present with metastatic disease at the diagnosis or those who have recurrent disease have a poor prognosis, with OS rates of less than 20% (¹⁹).

Treatment of metastatic or relapsed osteosarcoma is challenging, although multiple novel approaches are being tested. One approach under investigation is targeting the bone-forming behavior of this tumor through treatment with radioactive samarium 153–EDTMP or more recently radium 223 dichloride, which concentrated in bone-producing osteosarcoma (²⁰). As another approach, we have introduced immunemodulators into osteosarcoma therapy. Based on its potent immune-stimulatory properties, liposomal muramyl tripeptide phosphatidylethanolamine (L-MTP-PE) was added to standard osteosarcoma chemotherapy. Although the statistical benefit to survival was not clear in the largest US trial, the European Medicines Evaluation Agency approved this therapy for patients with pulmonary metastases of osteosarcoma; however, it is not currently available in the United States. More recently, we have used aerosolized interleukin (IL) 2 to enhance local immune responses against pulmonary metastases. Preclinical evidence demonstrates that aerosolized gemcitabine upregulates Fas expression in pulmonary metastases, leading to immune sensitivity (²¹).

Desmoplastic Small Round Cell Tumor

Desmoplastic small round cell tumor (DSRCT) is an extremely rare undifferentiated mesechymal tumor that affects about 100 children, adolescents, and young adults annually in the United States. It is an aggressive tumor that may show a response to multimodal frontline therapy, although OS rates remain below 30%. The t(11;22) EWS-WT1 translocation is seen in the majority of cases, confirming the unique pathobiology of this disease.

For patients with relapsed DSRCT, few chemotherapeutics have shown activity, and targeted agents have had limited success as well. However, the insulinlike growth factor (IGF) 1R–inhibiting antibody ganitumab showed one partial response, and there were 3 patients with prolonged stable disease (>24 weeks) of 16 patients with DSRCT. Individual partial responses have been seen with other agents (eg, the multikinase inhibitor pazopanib, the novel agent tasisulam, etc.). Unfortunately, DSRCT frequently disseminates throughout the peritoneal cavity. For these cases, we have used the approach of combining extensive surgical resection followed by hyperthermic intraperitoneal chemotherapy (HIPEC), which has shown promise (²²).

Medulloblastoma

Medulloblastoma is the most common pediatric malignant CNS tumor, with a peak age of incidence at 5 years and with 80% of medulloblastomas occurring before the age of 15 years. Medulloblastoma is known to be associated with familial cancer syndrome in less than 1% of patients; Gorlin (PTCH mutation, SHH receptor) and Turcot (mismatch repair genes) syndromes are the most common. Age less than 3 years at diagnosis, residual tumor after resection, anaplastic histology, MYC amplification, 17p loss, and metastatic disease may predict poor outcome. Children with localized disease have a greater than 80% 5-year EFS comparing to those with disseminated disease, who have less than 40% EFS. Treatment typically is multidisciplinary, including surgery, chemotherapy, and RT in newly diagnosed children more than 3 years old.

Recent genomic studies have identified four subtypes of medulloblastoma with distinct cellular origins, namely, WNT, SHH, group 3, and group 4, with the WNT and group 3 subtypes having the best and worst prognosis, respectively. Recurrent mutations in CTNNB1, PTCH1, MLL2, SMARCA4, DDX3X, CTDNEP1, KDM6A, and TBR1 have been identified (²³). The pathways of current clinical interest for medulloblastoma include VEGF, SHH, WNT, Notch, and ERBB with therapeutic implications (²⁴).

For standard-risk medulloblastoma, Packer et al reported that 421 patients between 3 and 21 years of age with nondisseminated medulloblastoma were prospectively randomly assigned to treatment with 23.4 Gy of craniospinal RT, 55.8 Gy of posterior fossa RT, plus one of two adjuvant chemotherapy regimens: lomustine (CCNU), cisplatin, and vincristine or cyclophosphamide, cisplatin, and vincristine. The results of this study disclosed an 81% 5-year EFS rate for children older than 3 years of age with nondisseminated disease (²⁵). Proton beam–based radiation is currently being investigated as an alternative to conventional irradiation (NCT01063114).

For high-risk medulloblastoma in children under the age of 3, a German group conducted a study for new diagnoses of medulloblastoma treated with postoperative chemotherapy alone, consisting of cyclophosphamide, methotrexate, vincristine, carboplatin, and etoposide. Forty-three children were treated according to the protocol. In children who had complete resection (17 patients), residual tumor (14), and macroscopic metastases (12), the 5-year progression-free survival (PFS) and OS rates (+/- SE) were 82% +/- 9% and 93% +/- 6%, 50% +/- 13% and 56 +/- 14 %, and 33% +/- 14% and 38% +/- 15%, respectively (²⁶).

An alternate treatment approach for patients less than 3 years old is to avoid radiation and use high-dose chemotherapy and autologous stem cell rescue. Dhall et al reported the result for 21 patients with nonmetastatic disease. The 5-year EFS and OS rates (+/- SE) for all patients, patients with gross total resection, and patients with residual tumor were 52% +/- 11% and 70% +/- 10%, 64% +/- 13% and 79% +/- 11%, and 29% +/- 17% and 57% +/- 19%, respectively. The 5-year EFS and OS (+/-SE) for patients with desmoplastic and classical medulloblastoma were 67% +/- 16% and 78% +/- 14%, and 42% +/- 14% and 67% +/- 14%, respectively (²⁷). There were four treatment-related deaths. The majority of survivors (71%) avoided irradiation completely. Mean intellectual functioning and quality of life for children surviving without irradiation was within average range for a majority of survivors tested.

For relapsed medulloblastoma, the combination of irinotecan and temozolomide showed some efficacy, with an objective response rate 33.3%; 68.3% experienced clinical benefit, and median survival was 16.7 months (²⁸).

In MD Anderson, our focus is to profile novel therapy for relapsed brain tumors. Focusing on locoregional and immunotherapy, we have begun to infuse methotrexate into a catheter placed in the fourth ventricle to target posterior fossa tumors (medulloblastoma, ependymoma, and ATRT; NCT02458339) (²⁹). Extending this approach of targeting the posterior fossa tumors, we now infuse autologous ex vivo expanded NK cells (NCT02271711). Another area of interest is target therapy; we are testing an SHH inhibitor for relapsed medulloblastoma.

Low-Grade Gliomas

Approximately half of pediatric CNS tumors are gliomas and are associated with two cancer predisposition syndromes: neurofibromatosis 1 (NF1 mutations) and tuberous sclerosis syndrome (TSC mutations). Gliomas occur in both children and adults; however, the majority of pediatric gliomas are low-grade gliomas (LGGs), with indolent pilocytic astrocytomas the most common subtype. Surgical resection, if feasible, is the most effective therapeutic option, with complete resection of an LGG leading to greater than 90% survival, while less than total resection leads to an about 60% survival. The management of pediatric LGGs that cannot be completely resected has evolved considerably during the last two decades. Although radiation used to be the standard treatment for incompletely resected or unresectable LGGs, chemotherapy and observation have now progressively become the most commonly used options after initial diagnosis, depending on several factors, including tumor location, amount of residual tumor, age, or association with NF1.

The most widely used first-line chemotherapy is a combination of vincristine and carboplatin. In a study of children with newly diagnosed LGGs with evidence of progression, 56% had an objective response to vincristine/carboplatin, and PFS was 68% +/- 7% at 3 years (³⁰). Multiple chemotherapies are proven to be effective as treatment of LGGs, including vinblastine, temozolomide, irinotecan, and bevacizumab (³¹). However, some of the short-term and long-term adverse effects are not negligible: severe myelosuppression that requires transfusion and growth factors, hearing loss, or infertility. The COG study showed that a procarbazine, thioguanine, lomustine, and vincristine combination had similar EFS as a combination of vincristine and carboplatin (³²). Any of the chemotherapies mentioned can be used as relapse treatment if not used as frontline treatment. Although there is no single standard of care, tolerability and long-term adverse effects are considered to be major factors to determine the treatment plan.

Importantly, translocations of BRAF occur in 70% of pilocytic astrocytomas (³³), and the BRAF V600E-activating mutation occurs in 5% to 10% of pediatric pilocytic astrocytomas. Although the BRAF V600E mutation is not frequently seen with pediatric gliomas, there are case reports that gliomas with BRAF V600E mutation showed a dramatic response to vemurafenib (³⁴). We offer screening of the mutation and have been experiencing cases with great responses. Because adult melanoma studies showed that most of the patients develop resistance to vemurabenib after a certain period, development of resistance to vemurafenib is also concerning with pediatric gliomas. To overcome this resistance, mTOR inhibitor (everolimus) has been combined with vemurafinib in a phase I trial at MD Anderson (NCT01596140). Alternatively, MEK1 inhibitor, when combined with BRAF inhibitor in melanoma patients, showed improvement of EFS and overcame single-drug resistance (³⁵).

Diffuse Intrinsic Pontine Glioma

Diffuse intrinsic pontine glioma (DIPG) comprises 70% to 80% of brainstem tumors in children, with an annual incidence of nearly 300 in the United States. Despite the collaborative efforts and advancement in the multimodality management of brain tumors, the prognosis of these tumors has remained dismal over the last two decades and poses therapeutic challenges. The median OS has remained 9 to 12 months. These tumors occur commonly between the ages of 5 and 10 years, arising from the pons and causing its diffuse enlargement. Stereotactic biopsy of these tumors was first reported in 1978. There have been arguments against biopsy as this is thought to have poor yield and biology of the limited sample would not be truly representative of the entire tumor. Treatment strategies based on clinical trials in adults with high-grade gliomas (HGGs) did not translate into increased OS of DIPG. Today, focal RT given over 6 weeks remains the standard of care for newly diagnosed DIPG (³⁶).

Recently, whole-genome, whole-exome, or transcriptome sequencing identified recurrent somatic mutations in ACVR1 exclusively in DIPGs (32%), in addition to previously reported frequent somatic mutations in histone H3 genes TP53 and ATRX (³⁷). Structural variants generating fusion genes were found in 47% of DIPGs, with recurrent fusions involving the neurotrophin receptor genes NTRK1 and NTRK2. Mutations targeting receptor tyrosine kinase–RAS-PI3K signaling, histone modification or chromatin remodeling, and cell cycle regulation were found in 68%, 73%, and 59% of pediatric HGGs, respectively, including in DIPGs. There have been multiple trials with target therapies, imatinib, sunitinib; however, none of them showed increased OS. There is also always concern of drug penetration to the tumor in the pons area. Currently, a COG phase I/II study is investigating use of suberoylanilide hydroxamic acid (SAHA) along with RT and maintenance therapy with SAHA for newly diagnosed DIPG. For relapsed DIPG, the tumors that respond to radiation once and show progression after a certain period, multiple target therapies such as SAHA and everolimus are under consideration for clinical trials. Re-irradiation is also an option for patients who maintain stable disease for a certain period.

High-Grade Gliomas

The outcome for children with HGGs remains poor despite the use of multimodal therapy with surgery, RT, and chemotherapy. Although RT does prolong time to progression slightly, adjuvant chemotherapy has had little impact on survival in children with HGGs. In the first study by the Children’s Cancer Group (CCG 943), RT with chemotherapy consisting of chloroethyl-cyclohexyl nitrosourea (CCNU), vincristine, and prednisone following surgical therapy showed a 5-year EFS of 46%, compared to 18% in patients with RT only. Then, the CCG 945 study showed no improvement in survival for patients treated with the 8-in-1drug regimen compared to the CCNU/vincristine/prednisone regimen. Five-year PFS was 19% +/- 3%, while those who did not have a GTR had a 5-year PFS of 11% +/- 4% (³⁸). The COG ACNS0126 study used temozolomide as a radiosensitizer followed by 10 cycles of temozolomide at 200 mg/m²/d for 5 days of every 28-day cycle. The 3-year EFS and OS were 11% +/- 3% and 22% +/- 5%, respectively (³⁹). The results with temozolomide given during RT and as an adjuvant therapy were similar to CCG 945 (P = .98). ACNS0126 demonstrated comparable survival with less toxicity than in studies utilizing prior nitrosourea-based regimens, which made temozolomide a de facto standard treatment.

Recently, the COG study conducted a randomized “pick-the-winner” approach to determine if either of the two experimental treatment arms (bevacizumab or vorinostat during chemoradiotherapy) had a higher nominal 1-year EFS than the standard treatment arm (temozolomide during chemoradiotherapy). A preliminary report showed that there was no significant benefit from choosing bevacizumab or vorinostat over temozolomide (⁴⁰). From adult studies and small pediatric studies, using bevacizumab along with temozolomide as adjuvant therapy is tolerable therapy, and this was used as a backbone neoadjuvant therapy in this study (⁴¹).

For relapse HGGs, adult studies for glioblastoma multiforme (GBM) suggest bevacizumab and protracted temozolomide for recurrent/progressive disease even after prior temozolomide exposure (⁴²). Multiple adult studies with recurrent HGGs showed activity of bevazizumab with or without irinotecan (⁴³). Bevacizumab in patients with recurrent GBM is approved by the FDA. To date, robust identification and correlation of dysfunctional genes with the tumorigenesis have not been performed—a likely reason for continuing therapeutic failure. There is a dire need to treat these patients as they come for therapy after having exhausted treatment options. To understand the genetic landscape of relapsed brain tumors, we enroll our patients in the CM50 study, which is a sequencing study of 50 genes commonly altered in cancers, with at least 50% of those genes pertinent to pediatric brain cancers. We are currently opening a phase I targeted therapy trial for pediatric HGG. The trial is using a combination of dasatinib (PDGFR and SRC inhibitor) and temsirolimus (mTOR inhibitor) and metronomic cyclophosphamide (antiangiogenesis) for targeting the most common driving pathways in pediatric HGG: AKT and angiogenesis (NCT02389309) (⁶¹).

Although OS for children with cancer is excellent, children with relapsed disease still face dismal outcomes. Several recent successes (Table 46-1) have allowed us to make significant progress for specific subsets of patients, although much more work must be done. We still face the challenges of increasing knowledge of the biology of these tumors, and through increasing access to novel therapeutic agents, we are hopeful that we will be able to make a brighter future for our young patients.