Review

Science and health for all children with cancer

See allHide authors and affiliations

Science  15 Mar 2019:
Vol. 363, Issue 6432, pp. 1182-1186
DOI: 10.1126/science.aaw4892

Abstract

Each year ~429,000 children and adolescents aged 0 to 19 years are expected to develop cancer. Five-year survival rates exceed 80% for the 45,000 children with cancer in high-income countries (HICs) but are less than 30% for the 384,000 children in lower-middle-income countries (LMICs). Improved survival rates in HICs have been achieved through multidisciplinary care and research, with treatment regimens using mostly generic medicines and optimized risk stratification. Children’s outcomes in LMICs can be improved through global collaborative partnerships that help local leaders adapt effective treatments to local resources and clinical needs, as well as address common problems such as delayed diagnosis and treatment abandonment. Together, these approaches may bring within reach the global survival target recently set by the World Health Organization: 60% survival for all children with cancer by 2030.

In recent years, 5-year survival rates for children with cancer (typically equating to cure) have risen to ~80% in most high-income countries (HICs) (14). This progress reflects partly the optimized use of conventional therapies (e.g., cytotoxic drugs) through better risk stratification of patients. For example, on the basis of molecular prognostic markers, treatments are intensified for patients identified as high risk and deescalated for patients identified as lower risk to reduce the likelihood of immediate and long-term side effects. Expanding portfolios of new drugs that target the biological mechanisms driving the growth of pediatric cancers are also starting to contribute to improved cure rates in HICs (5). The situation is more bleak for children with cancer in lower-middle-income countries (LMICs) (6). In LMICs taken as a whole, the 5-year survival rate is only ~30% (Table 1). Even considering only geographic sites with adequate resources to support population-based cancer registration, recent global data show that the 5-year survival rate for children can be up to 45% higher in HICs than in LMICs for acute lymphoblastic leukemia and up to 51% higher for children with brain tumors (1).

Table 1 Estimated incidence of pediatric cancer each year and estimated cure rates in countries within different World Bank income categories.

The incidence gap is defined as the percentage of patients expected to develop cancer each year divided by the mean of the incidence reported in HICs. Some variation is expected on the basis of epidemiologic variation; however, lack of diagnosis is a likely cause of gaps that exceed 2 standard deviations from the HIC mean (107). UMICs, upper-middle-income countries.

View this table:

In 2018, the World Health Organization (WHO) set a global survival target of 60% for all children with cancer, with the goal of saving a million more lives by the year 2030 (7). This new WHO target may appear daunting, but it cannot be ignored. About 89% of the world’s children (aged 0 to 19 years) live in LMICs, and they account for 95% of the mortality from cancer in this age group worldwide (Table 1) (4, 6). Furthermore, the incidence rates of many childhood cancers are increasing by ~1% per annum in many countries with population-based cancer registration (8). In this Review, we aim to demonstrate that this target can be brought within reach by building on the long-established work of the global childhood cancer community. Success will also require new partnerships and increased cooperation between stakeholders, including not only health care professionals but also parents, patients, civil society communities, industry, academia, and governments (9).

Global incidence of childhood cancers

The major cancer types affecting children younger than 15 years and adolescents (aged 15 to 19 years) differ from those affecting adults, which are typically epithelial in origin. The most common cancers in children include acute leukemias, brain tumors, lymphomas, bone and soft tissue sarcomas, and germ cell tumors. The typical “embryonal tumors” (neuroblastoma, renal tumors, and retinoblastoma) are confined largely to younger children, whereas cancers in adolescents include more epithelial tumors (such as thyroid carcinoma) and melanoma. An estimated 429,000 new cases of childhood cancer are expected globally each year (Table 1). Globally, the age-standardized reported incidence rates are 141 per million person-years (children) and 185 per million person-years (adolescents) (8). Age-specific incidence rates vary by geography and ethnicity. Notably, most of the world’s children are not covered by population-based cancer registries (10). Furthermore, the 170 per million incidence in HICs is double that in low-income countries (LICs) and LMICs (Table 1 and Fig. 1). However, if one extrapolates incidence from HICs to the sub-Saharan population, it is apparent that fewer than half the expected cases of acute leukemia, brain tumors, neuroblastoma, and bone tumors are diagnosed (Fig. 1). By contrast, there is less of an incidence deficit for cancers that present with more obvious clinical symptoms, such as Burkitt’s lymphoma, renal (Wilms) tumor, and retinoblastoma (3). Similar conclusions can be drawn when comparing incidence rates in India with those in Japan and South Korea (8).

Fig. 1 Childhood cancer incidence deficit: Percentage of expected cancer cases diagnosed in each country.

The map shows the percentage of children expected to develop cancer each year who are diagnosed. The incidence ratio is calculated by dividing the number of cases reported in a country by the number expected if incidences were the same as those in HICs, where nondiagnosis is rare. The incidence deficit is the incidence ratio subtracted from 100%. An incidence ratio of 80% or higher (yellow) is consistent with diagnosis of all cases, as it is within the observed variability across HICs.

Within each of the main cancer types, there are subcategories that can be treated with generic medicines and regimens that are readily adapted to resource-limited settings. These include Burkitt’s lymphoma, Wilms tumor, acute lymphoblastic leukemia, Hodgkin’s lymphoma, certain brain tumors, germ cell tumors, and low-risk or anatomically localized cases of neuroblastoma, sarcomas, and retinoblastoma. About 50% of all cancers in the 0- to 19-year age group would be in this “favorable prognosis” category if diagnosed and treated appropriately.

Unfortunately, treatment failure is common in LMICs because of many factors, including failure to diagnose, misdiagnosis, unaffordable or abandoned treatment, toxic (treatment-related) death, and excess relapse (Fig. 2) (9, 11, 12). Delayed presentation or diagnosis, drug shortages, intermittent adherence to treatment, and treatment regimens that are of reduced intensity to facilitate tolerability also contribute to treatment failure. As discussed below, many causes of treatment failure are preventable. Although children with cancer can be managed in very different settings in LMICs, we will use the term “cancer center” to characterize dedicated wards or units for children with cancer within a hospital, as well as stand-alone hospital facilities for patients (solely for or including children) with cancer.

Fig. 2 Causes of treatment failure for children with cancer by World Bank income categories for all children with cancer and selected Burkitt’s lymphoma studies in LICs, LMICs, and HICs.

Although the causes of treatment failure in HICs differ from those in upper-middle-income countries (UMICs), LMICs, and LICs, a variety of published studies of Burkitt’s lymphoma in each World Bank income group illustrate that even within the same World Bank group, the relative contribution of each cause of treatment failure differs in different settings, even for the same cancer. In the Burkitt’s lymphoma studies, data for nondiagnosis, misdiagnosis, and no treatment are estimates, as these are not reported in the clinical trials themselves. Studies specifically addressing misdiagnosis in sub-Saharan Africa have found rates as high as 18 to 35% (26). AHOPCA, Association of Pediatric Hematology Oncology of Central America.

(ILLUSTRATION): S. C. HOWARD

The challenge of late diagnosis

Patients whose cancer diagnosis is delayed often present with more advanced disease that is inherently harder to cure and necessitates more intensive therapy. Such patients often have comorbidities, such as malnutrition and infection, and are at increased risk of tumor lysis syndrome and treatment-related death (9, 1319). When delayed diagnosis is associated with more refractory disease and excess relapse, this increases the costs and morbidity of treatment and in turn increases treatment abandonment. As demonstrated in HICs, community-based public and professional awareness-raising campaigns can be effectively coupled with diagnosis-specific referral guidelines; in the United Kingdom, such a national strategy reduced the median time to diagnose brain tumors in children from 14 weeks to just under 7 weeks (20). In Honduras, awareness materials on retinoblastoma, the most common eye cancer in young children, were integrated into national immunization platforms across health centers and resulted in a significant decrease in the proportion of patients presenting with advanced retinoblastoma (21). Other resources to increase timely diagnosis of childhood cancer have been developed for various settings, including educational resources from the Pan American Health Organization, as well as locally adapted and international signs-and-symptoms campaigns, with ongoing opportunities to demonstrate effectiveness, especially in LMICs (2224).

Reducing diagnostic failure

Reliable cancer diagnosis can be hampered by a lack of an appropriately trained workforce and imaging and laboratory equipment, as well as a lack of access to more specialized techniques, such as immunohistochemistry and immunophenotyping, that are routine in HIC pathology departments. These deficits may affect individual treatment centers or entire geographic regions. Even Burkitt’s lymphoma, which has a characteristic histologic appearance and is commonly diagnosed in sub-Saharan Africa, has been shown to be misdiagnosed in 18 to 35% of cases in single- and multicenter analyses (Fig. 2) (2527). Discrepancies in pathology diagnosis reached 64% in one study from Uganda that included all subtypes of suspected non-Hodgkin’s lymphoma (NHL) (26, 27). In addition to diagnostic difficulties, staging and risk stratification are suboptimal when patients lack access to imaging methods and assays that measure prognostic biomarkers (e.g., amplification of the MYCN oncogene for neuroblastoma) (28). Barriers to access can be due to the unavailability of services or the inability of patients to pay for these services in health care systems that depend primarily on families’ out-of-pocket contributions (29). Strategies to facilitate earlier and more accurate cancer diagnosis in LMICs include the use of telemedicine with local, cross-regional, or international expert groups to complement on-site continuing education; both clinical referral networks and professional networks for local input or virtual input from international experts have been successfully implemented in LMICs (3, 9, 3035). An example of such long-term collaboration is outlined later in this Review.

Targeting toxic (treatment-related) death

Toxic death is a leading contributor to treatment failure for children with cancer in LMICs (9, 32). The occurrence rate can be as high as 24 to 30% in higher-risk patients during the first month of therapy, with risk depending on the cancer type, regimen used, and supportive care available (14, 3638). Appropriate supportive care especially at the start of treatment is one of the biggest challenges in LMICs (39). Nutritional support, management of infections, and aggressive hydration (for hematologic cancers and bulky tumors) to prevent tumor lysis syndrome can effectively reduce early toxic death (40). Resource-sensitive tools and interventions to address patients’ nutritional needs in LMICs have been developed as key components of supportive care (41). In some areas of Africa, intestinal parasite therapy is routinely given before the start of chemotherapy, to help protect children from common infections that can become overwhelming and potentially fatal when the child’s immune system is weakened by malignancy as well as the treatment. Likewise, during treatment, other forms of antimicrobial prophylaxis and therapy have been adapted to localities where the incidence and types of infection differ from those in HICs (17). To help staff respond quickly to infections and other causes of rapid clinical decline in children with cancer, an early warning system has been developed to facilitate identification, team communication, and management for children with cancer who are deteriorating in LMICs (42). In collaboration with St. Jude Children’s Research Hospital, this early warning system was validated in Guatemala, where the system effectively reduced the number of children who deteriorated and decreased the need for patients to be transferred to an intensive care unit (43). Strategies to adapt the treatment regimen to local resources and needs, with the intent to reduce toxic deaths, are further described below.

Adapting treatment regimens in LMICs

So-called “adapted treatment regimens” are widely used in LMICs. These are typically lower-intensity regimens that can be associated with higher overall survival by causing fewer toxic deaths (44). For instance, patients with Burkitt’s lymphoma who have bulky disease and frequent morbidity after initial treatment may be managed with an additional few weeks of reduced-intensity treatment before continuing standard therapy (45). To address the challenges of bed shortages and other treatment delays (potentially compounding patients’ risks of not tolerating treatment well, as their condition may worsen during the wait), some providers have incorporated inexpensive medicines that can be administered on an outpatient basis (e.g., hydroxyurea for patients with acute myeloid leukemia). Other providers, to mitigate known chemotherapy toxicity in settings with less available supportive care, have studied the use of reduced chemotherapy doses, as well as low-cost modifications to facilitate monitoring when measurement of drug levels is not possible (4548). As adapted treatment regimens may also increase the risk of relapse due to reduced-intensity treatment, local data for both outcomes (toxic death and relapse) need to be continuously monitored (49, 50). In LMICs, many factors unrelated to the treatment regimen also increase relapse risk, including the unavailability or unaffordability of medicines, lack of adherence to treatment, and lack of infrastructure and support to families to enable treatment completion (35, 5155). Ideally, these factors should be considered in the planning of adapted treatment regimens—for instance, by planning medicines that can be substituted for commonly unavailable or unaffordable medicines or including funding for a clinical coordinator or patient navigator to help reinforce adherence to the treatment regimen. The International Pediatric Oncology Society (SIOP) has established procedures to develop and deploy such regimens (28, 41, 5658).

Reducing treatment abandonment

Treatment abandonment, defined as 4 weeks or more of missed appointments during therapy, is a major contributor to treatment failure for children with cancer in most cancer centers in LMICs but is rare in HICs (12, 35, 38, 46, 5963). In some settings, such as rural Zambia, abandonment rates approach 50%; after war or during civil unrest or natural disasters, abandonment rates increase sharply (64). In Côte d’Ivoire, nearly half of the children with Burkitt’s lymphoma abandoned treatment shortly after the first admission, resulting in only a 6% cure rate (Fig. 2) (65). Risk factors for abandonment include poverty, the local cost of treatment, low educational attainment of parents, distance from the cancer center, cancer type, and in some cases, patient gender; one study showed that the 12-month cumulative incidence of abandonment was 22% in females versus 7% in males (13). Government support matters; in a study from Kenya where more than 70% of children with cancer lacked health insurance at diagnosis, these children had a risk of treatment failure (most commonly treatment abandonment) three times as high as those with insurance (66). To address treatment abandonment, various strategies have been deployed successfully. In Central America, only 6.5% of patients with anaplastic large-cell lymphoma abandoned therapy, a rate somewhat lower than that typically experienced with other tumors in the same settings, perhaps because of relatively short treatment duration (compared with that for acute lymphoblastic leukemia) and government coverage of all chemotherapy costs (6769). However, in Sierra Leone, despite a shortened duration of inpatient treatment and the provision of free treatment, meals, and transportation, treatment abandonment persisted in rural areas, a finding replicated in other countries with long travel times and no established referral network (38, 61, 70). In El Salvador, in addition to free treatment, implementation of a tracking protocol with community-based interventions for missed appointments successfully prevented abandonment in almost all patients, whereas in Recife, Brazil, a comprehensive social support and educational program reduced abandonment from 16 to 0% (38, 71). To further explore local creative strategies to address treatment abandonment and related topics, the global SIOP community has organized active working groups and launched a podcast related to these issues (59).

Stimulating development of comprehensive health services for children with cancer

Many of the causes of preventable treatment failure in LMICs noted above are rooted in fragile or insufficient health care infrastructures (including facilities) and workforces. Childhood cancer can provide a lens to examine and improve the performance of health care systems more generally, with potential benefits beyond children with cancer. Management of childhood cancer in LMICs is ideally facilitated by being part of a national cancer control program that coordinates the identification of priorities and resource allocation and supports the delivery of goods and services, including needs for the workforce, essential medicines and technologies, and information systems and policies (72). Additional work should be done locally to understand and address the root causes of diagnostic, referral, and treatment delays and inform data-driven solutions to improve care coordination and optimize resource allocation (73). For example, improvements in the laboratory and pathology infrastructure and in team communication should lead to faster diagnosis and response to patients’ critical electrolyte derangements or life-threatening infections; this would benefit all patients, not only children with cancer. Similarly, the prioritization of hand hygiene and isolation rooms for infection prevention and control should decrease morbidity and mortality for children with cancer and all other patients in the same hospital. The design of strategies to improve care should integrate the input of partners across the health system—not only clinicians, but also others, such as policy-makers responsible for budget allocations, legal and regulatory bodies overseeing the approval of medicines academic bodies overseeing the training and accreditation of providers, and civil society organizations providing support for patients and families. Successful sustained partnerships have stimulated the engagement of other local governmental and nongovernmental partners to invest in services benefiting children (73).

National policies should support the ongoing collection and analysis of local data for monitoring and evaluation, starting with allocating resources for infrastructure and trained personnel to register all newly diagnosed patients and to document core patient outcomes, such as survival, abandonment, and relapse (71, 74, 75). Beyond increasing survival alone, coordinated policy efforts for children with cancer can improve patients’ quality of life and reduce suffering by ensuring access to morphine for pain, as well as palliative care and psychosocial support (76). As more children survive cancer, policies can also stipulate resources to support the distinct needs of children and families affected by a chronic illness such as cancer while strengthening the local capacity to prevent, monitor, and manage late effects of treatment among survivors (7678).

Access to essential medicines

An additional challenge affecting the care of children with cancer in LMICs is access to essential medicines. WHO has provided global guidance recommendations in the form of Model Lists of Essential Medicines (EMLs) across health conditions; these have included common cancers for adolescents and adults since 1977 and those for children (up to age 12) since 2007 (79). Recent analyses suggest an ongoing need for implementation research within countries to facilitate access to the recommended medicines, particularly in resource-limited settings (54, 80).

A further complexity is guaranteeing the supply of high-quality medicines. The experience of Brazil with asparaginase is instructive. An essential medicine for the most common childhood cancer, acute lymphoblastic leukemia, asparaginase was approved by the U.S. Food and Drug Administration in 1978 and has been recognized in the WHO EMLs since 1993 (79, 81). However, native asparaginase (derived from the bacterium Escherichia coli) has been only intermittently accessible in many settings (82). In 2017, anticipating a shortage of native asparaginase, the Ministry of Health in Brazil changed the national supplier to a foreign manufacturer that offered a new, lower-cost generic product (83). Concerned by the lack of published data on this new product, investigators in Brazil compared the drug’s properties with those of the native asparaginase used previously. They found that the new drug was less bioactive and contained contaminating proteins that increased the risk for immune-related side effects (83, 84). Although the nationwide distribution of this new product in Brazil was halted, this drug continues to be manufactured and distributed to other countries around the world (83). Moving forward, WHO’s 2018 global initiative will hopefully engage additional partners to increase economies of scale and uphold accountability mechanisms for product availability and quality consistency worldwide. Lessons learned from addressing social and financial barriers to increase access to medicines could then also be leveraged for other critical technologies (7). High-quality data and research will be critical in directing how resources can be harnessed to reach WHO’s global goal.

Science and drug development for all

Most children with cancer in HICs are enrolled in research and multicenter clinical trials, and this has been described as a key factor contributing to increased survival rates (2, 4, 85). However, there are several obstacles to further improvements in survival rates: Pediatric cancer is rare, the histologic subtypes of pediatric cancers differ substantially from those in adults, and the high cure rate for many pediatric cancers in HICs means that only small numbers of children with relapsed or refractory disease are available to enroll in early-phase trials of new drugs. Hence, effective drug development in pediatric oncology can benefit from engaging LMIC partners. Anti-CD20 antibody therapies are one example. NHL affects 70,000 adults each year in the United States, and several anti-CD20 antibodies are available or in development for CD20-positive disease (86, 87). However, in the United States, only 400 children are diagnosed annually with CD20-positive NHL across more than 200 pediatric oncology centers; fewer than 80 relapse, and even fewer relapse a second time and would be eligible for clinical trials for refractory mature B cell NHL. This means that pediatric trials of new CD20 antibodies are limited to a small number of eligible patients treated in a large number of centers (8). Furthermore, under the current paradigm, multiple pharmaceutical companies may compete for the same patient population to test different products with the same mechanism of action. It is not surprising that many drug development programs fail despite prolonged and costly efforts to meet accrual targets.

To address these issues in a HIC setting, the ACCELERATE Platform was launched in Europe and has been expanded to include partners worldwide (88). Strategic aims include ensuring that drugs for child and adolescent cancers are developed on the basis of the pharmaceutical mechanism of action rather than adult indications alone and facilitating international collaboration among all stakeholders—including researchers, pharma, and governmental regulators in conjunction with influential parent and patient advocates and resourcing partners (8890). Although currently only European and North American cancer centers are involved, there is the potential to involve well-structured cancer centers from LMICs, and this may be essential for timely progress. Comparative analysis of survival rates between HICs and LMICs is limited by many factors, including different proportions of high-risk patients and environments of care; thus, continuous research in LMICs remains vital to improve outcomes (14, 37, 91). Furthermore, only 11% of children with cancer live in HICs, so there is an untapped opportunity to help the remaining 89% of children who develop cancer in LMICs by providing them access to new therapies through clinical trials while accelerating scientific progress that can benefit all children (8).

The “twinning program” between the pediatric hematology-oncology department at King Hussein Cancer Center (KHCC) in Amman, Jordan, and the neuro-oncology section at the Hospital for Sick Children, Toronto, Canada, exemplifies the benefit of collaborative research between institutions in LMICs and HICs to advance science as well as patient outcomes. Stemming from collaborations since 2004, monthly multidisciplinary video teleconferences were held to discuss the management of KHCC patients. On average, four or five patients were discussed during each 1-hour session. The benefits of this interaction quickly became evident for all participants (9296). The group conducted a series of research projects together and found that 17 (39%) of 44 children with high-grade malignant brain tumor cases at KHCC over 10 years had defective mismatch repair genes, compared with less than 4% in North America (97, 98). This research led to the development of a larger collaborative network, the mismatch repair deficiency (MMRD) consortium, which in turn facilitated analyses of tumor and blood samples from individuals with MMRD-associated tumors, which are ultrahypermutated and responsive to immune checkpoint inhibitors. Sustained remissions in MMRD patients treated with the checkpoint inhibitors nivolumab and ipilimumab have been reported (99, 100), and continued efforts are now warranted to initiate immunotherapy trials and studies of adapted treatment regimens in LMICs. Clinical trials involving LMIC partners can integrate local epidemiology and resource considerations into the trial design and address new implementation research questions while galvanizing partnerships to sustain progress (34, 101103).

Conclusion

Childhood cancer demonstrates how a relatively small group of stakeholders investing in a relatively uncommon cluster of diseases has the potential to provide systemic benefits for science and health. As we have discussed, nearly 40% of the world’s children expected to have cancer are undiagnosed. We need to address this problem of children not being diagnosed, while continuing to improve treatment for those diagnosed. We need to invest in more prospective research and clinical trials to improve care for children in LMICs, with the expectation that lessons learned can translate to global improvements. Achieving health, defined by WHO as “not merely the absence of disease,” requires cross-cutting investments to address underlying determinants of health and needs across the health system (104, 105). We advocate for health for all children with cancer, as summarized in the Erice Statement: “The long-term goal of the cure and care of a child with cancer is that he/she becomes a resilient and autonomous adult with optimal health-related quality of life, accepted in society at the same level as his/her age peers” (106). There is no health for all unless there is also science for all.

References and Notes

Acknowledgments: Funding: C.G.L. is funded at St. Jude by the American Lebanese Syrian Associated Charities. E.B. is receiving funding from Bristol Myers Squibb for an investigator-initiated study of immunotherapy (NCT02992964). E.B. also received funding from Roche for an investigator-initiated study (NCT02840409). K.P.-J. is funded in part by the National Institute for Health Research Biomedical Research Centre at Great Ormond Street Hospital. Competing interests: The authors declare no competing interests.
View Abstract

Subjects

Navigate This Article