Outcomes of Wilms tumor therapy in Lilongwe, Malawi, 2016–2021: successes and ongoing research priorities

Abstract

Introduction:

Wilms tumor therapy in low- and middle-income countries (LMICs) relies on treatment protocols adapted to resource limitations, but these protocols have rarely been evaluated in real-world settings. Such evaluations are necessary to identify high-impact research priorities for clinical and implementation trials in LMICs. The purpose of this study was to identify highest priority targets for future clinical and implementation trials in sub-Saharan Africa by assessing outcomes of a resource-adapted treatment protocol in Malawi.

Methods:

We conducted a retrospective cohort study of children treated for Wilms tumor with an adapted SIOP-backbone protocol in Lilongwe, Malawi between 2016–2021. Survival analysis assessed variables associated with poor outcome with high potential for future research and intervention.

Results:

We identified 136 patients, most commonly with stage III (n=35; 25.7%) or IV disease (n=35; 25.7%). Two-year event-free survival (EFS) was 54% for Stage I/II, 51% for Stage III, and 13% for Stage IV. A single patient with Stage V disease survived to one year. Treatment abandonment occurred in 36 (26.5%) patients. Radiotherapy was indicated for 55 (40.4%), among whom three received it. Of these 55 patients, 2-year EFS was 31%. Of 14 patients with persistent metastatic pulmonary disease at time-of-nephrectomy, none survived to two years. Notable variables independently associated with survival were severe acute malnutrition (hazard ratio, HR, 1.9), increasing tumor stage (HR, 1.5), and vena cava involvement (HR, 3.1).

Conclusion

High-impact targets for clinical and implementation trials in low-resource settings include treatment abandonment, late presentation, and approaches optimized for healthcare systems with persistently unavailable radiotherapy.

Introduction

Wilms tumor is the most common non-hematologic malignancy diagnosed in African children.^1–3 In high-income countries, advances in therapy have achieved long-term survival in >90% of children.^4–6 This success is attributable to early diagnosis, multidisciplinary risk-adapted therapy, and well-resourced supportive care infrastructure.^{4, 7, 8} Similar outcomes have not been realized in low-and-middle income countries (LMICs) where most pediatric cancer occurs, particularly in sub-Saharan Africa where long-term survival remains dismal.^9–12 A systematic review of Wilms tumor outcomes in the LMICs of sub-Saharan Africa from 2000–2019 found that overall survival was consistently between 20–50%.¹³

Such disparities are the result of limited capacity to deliver pediatric cancer care within the context of resource-constrained public health systems. Capacity limitations manifest as delayed diagnosis, treatment abandonment, personnel shortages, absent therapeutics, and limited supportive care infrastructure.^1,14 To overcome these challenges, pediatric cancer centers in LMICs have implemented resource-adapted strategies that contextually adapt protocols to local capacity constraints, such as by lowering chemotherapy intensity to limit blood product requirements.^15–17

In contrast to cancer protocols in high-resource settings, resource-adapted protocols are rarely evaluated in the LMICs for which they are designed.¹⁸ Rigorous clinical trials evaluating specific resource-adaptations are rarer still. As front-line clinical trials in high-income countries focus on patients with high-risk and relapsed disease, the unique challenges inherent to treating childhood cancer in LMICs require dedicated clinical and implementation trials specific to our setting. As pediatric cancer programs in LMICs continue to grow rapidly, so to do opportunities to evaluate these protocols in multicenter trials.

The purpose of this study was to report the outcomes of six years’ experience treating Wilms tumor with a resource-adapted protocol in Malawi to identify high-impact targets and research questions for next-step clinical and implementation trials. Such trials will evaluate and improve upon resource-adapted Wilms tumor protocols designed for LMICs, particularly those for the extremely resource limited setting of sub-Saharan Africa.

Materials and Methods

Study design and setting

We conducted a retrospective cohort study of patients aged ≤18 years with Wilms tumor between January 2016 and December 2021 at Kamuzu Central Hospital (KCH) in Lilongwe, Malawi, a referral center serving a population of nine million people spanning Central and Northern Malawi. The Pediatric Oncology Unit contains 30 beds with access to chemotherapy, pathology services, oncology nursing, pediatric hematologist-oncologists, and pediatric surgeons. Notable resource-limitations are limited pediatric intensive care (e.g., mechanical ventilation), inconsistent platelet transfusion availability, high patient-to-nurse ratios, and absent radiotherapy.

Patient characteristics and diagnosis

Data were extracted from paper medical records. The diagnostic protocol utilized during the entire study period included imaging (abdomen and chest), local surgical staging (e.g., capsular involvement), and histological grouping. Tumor stage and histologic group were determined according to locally adapted International Society of Paediatric Oncology (SIOP)-based guidelines by reviewing histological and surgical reports.¹⁹ The staging system used locally largely mirrored published SIOP guidelines but differed notably in that 1) Percent necrosis at resection margins and lymph node sinuses were not consistently considered; and 2) Percent cellular sub-type of viable tissue was not considered (Supplementary table 1). In accordance with SIOP guidelines, patients were identified by characteristic abdominal imaging. These guidelines implement a post-surgical staging system; thus, histological and surgical evaluation are required to differentiate local stages I-III. Patients with Stage IV and V disease were identified with imaging; otherwise, patients without surgical or histopathological data available (e.g., due to presurgical death) were categorized as un-staged. Malnutrition was diagnosed and treated by nutritionists according to World Health Organization guidelines.²⁰

Treatment received

The treatment protocol at use in Lilongwe since 2016 was built upon a SIOP backbone incorporating neoadjuvant pre-operative vincristine (1.5 mg/m²/dose) and dactinomycin (0.045 mg/kg/dose) for non-metastatic disease with the addition of doxorubicin (50 mg/m²/dose) for metastatic or bilateral disease (Supplemental figure 1).²⁰ Post-operative chemotherapy regimens were selected according to tumor stage and tumor histology (Supplemental table 2). Surgical complications were defined as any of the following during or up to 30 days post-operatively: death, re-operation, surgical site infection, or wound dehiscence.

Patients with persistent metastatic disease at the time of nephrectomy and patients with Stage V disease were given escalated treatment intensity with cyclophosphamide (2200 mg/m²/cycle) and etoposide (500 mg/m²/cycle) added to vincristine (2 mg/m²/dose) and doxorubicin (30 mg/m²/dose) (i.e., Children’s Oncology Group, COG, “Regimen M”).⁵ If patients were deemed unlikely to tolerate escalated treatment intensity, the pre-operative regimen was continued. While radiotherapy was not available in Malawi during the study period, need for radiation was determined utilizing SIOP guidelines.²¹

The protocol was modified to accommodate resource limitations consistent with SIOP published guidelines for LMICs.¹⁶ Specifically, pre-operative chemotherapy was extended for up to eight weeks (12 weeks for Stage V) to minimize risk of perioperative morbidity/mortality and need of radiotherapy. Two additional site-specific modifications were made: 1) Patients with Stage I disease and low or intermediate risk histology were given five cycles of dactinomycin and vincristine-based therapy rather than zero or one due to limited relapse salvage options in Malawi; and 2) Selected patients with Stage III tumors who tolerated pre-operative chemotherapy were chosen to receive Regimen M post-operatively to account for radiotherapy unavailability at the center. Finally, upon frequent stockouts of dactinomycin, dactinomycin-containing regimens were substituted during these periods with doxorubicin (50 mg/m²/cycle) and cyclophosphamide (1200 mg/m²/cycle). Any patient with concurrent HIV infection was referred to the hospital’s pediatric HIV service for consultation, but no modifications to the Wilms tumor therapy protocol were made.

Statistical analysis

Overall survival was determined by vital status at patients’ last encounter. A treatment-related death was defined as death during therapy unattributable to disease progression. Event-free survival was calculated using the earliest of death, disease progression, relapse, or treatment abandonment. Treatment abandonment was defined as an unplanned hiatus of ≥4 weeks in the scheduled curative-intent therapy plan.²² Patients were right-censored at the earlier of their last date of follow-up or March 2022.

Kaplan-Meier survival curves and multivariable Cox proportional hazards models were used to identify risk factors associated with survival. Variables assessed were age, sex, use of traditional medicine, acute malnutrition at diagnosis, tumor stage, time-to-surgery, and inferior vena cava involvement. If individual variables were found to be significantly associated with outcomes on unadjusted analysis, we conducted adjusted analysis for each significant variable informed by causal diagrams.²³ As treatment abandonment was a censoring event in the OS analysis and was an outcome in the EFS analysis, a hazard ratio (HR) for treatment abandonment could not be generated. Median follow-up time was calculated using the reverse Kaplan-Meier method.²⁴

Analyses were performed in R.^25,26 Data collection occurred from January-March 2022 with analysis through May 2022 in accordance with Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.²⁷ The study was approved by institutional review boards in Malawi and the United States.

Results

Study cohort

Wilms tumor was the most common non-hematologic cancer diagnosis at the center during the study period, initially identified in 141 patients. Five patients (3.5%) with non-Wilms histology were excluded to create a study cohort of 136 patients (Table 1). Non-Wilms diagnoses included neuroblastoma, renal cell carcinoma, and mesoblastic nephroma. The median age was 3.8 years (interquartile range, IQR, 2.4–5.3). A single patient had concurrent HIV infection.

Table 1:

Characteristics of 136 patients with Wilms tumor within the study cohort

Age in years (Median, IQR)	3.8 (2.4, 5.3)
Male sex	68 (50.0%)
Reported duration of symptoms in weeks (median, IQR)	6.0 (4.0, 12.0)
Moderate acute malnutrition at diagnosis	20 (14.7%)
Severe acute malnutrition at diagnosis	36 (26.5%)
Overall stage
Stage I	30 (22.1%)
Stage II	7 (5.1%)
Stage III	35 (25.7%)
Stage IV	35 (25.7%)
Stage V	5 (3.7%)
Un-staged	24 (17.6%)
Wilms histological subtype ^
Low risk	8 (11%)
Intermediate risk	56 (77%)
High risk	9 (12%)

IQR, interquartile range. Data are n (%) unless otherwise specified.

^{^}Histologic subtypes are 1) Low risk: 100% necrotic; 2) Intermediate risk: regressive type, mixed type, epithelial type, stromal type, or focal anaplasia; and 3) High risk: blastemal type or diffuse anaplasia. Histological group data represent 73 patients with post-surgical histological group data available.

Radiological diagnoses were made by ultrasound in 123 patients (90.4%) and by computed tomography (CT) with/without ultrasound in 63 (46.3%). Chest imaging was documented by X-ray (n=98) and/or CT (n=76) in 126 patients (92.7%). The median reported duration of symptoms prior to presentation was six weeks (IQR 4–12). Moderate acute malnutrition was identified in 20 (14.7%) patients and severe acute malnutrition in 36 (26.5%) (Table 1).

Tumor characteristics and staging

Most patients had either Stage III (25.7%) or Stage IV (25.7%, Table 1) disease. Staging was unavailable for 24 (17.7%) patients, most commonly due to pre-operative death or treatment abandonment. Bilateral disease occurred in five patients (3.7%). Among those with metastatic disease, 16 (11.8%) had pulmonary metastases, 10 (7.4%) with hepatic metastases, and 7 (5.1%) with both. Thirteen (9.6%) patients had tumor within the inferior vena cava (IVC). Post-operative histology reports were available for 74 patients (80% of patients undergoing nephrectomy). Among these, 8 (10.8%) were SIOP low risk tumors, 56 (75.7%) intermediate risk, and 9 (12.2%) high risk.

Therapy received

Pre-operative chemotherapy was started in 129 (94.9%) patients (Figure 1). Two patients had up-front resection for suspected mesoblastic nephroma, and five patients were discharged on palliative care. The median duration of pre-operative chemotherapy (i.e., time-to-nephrectomy) was seven weeks (IQR, 5–11). Ninety-two patients (68%) underwent nephrectomy and 44 (32%) failed to reach nephrectomy. The most common reasons for failure to reach nephrectomy were pre-operative death (n = 22, 50%) and treatment abandonment (n = 17, 39%).

Figure 1:

Study profile for Wilms tumor Patients treated at KCH from 2016–2021. T

The number of abandoned patients depicted does not include those who abandoned but returned to continue therapy. TRM = treatment related mortality, DRM = disease related mortality, VA = vincristine and dactinomycin, VAD = vincristine, dactinomycin, and doxorubicin, VDC = vincristine, doxorubicin, and cyclophosphamide

Eighty-nine patients (65.4%) received postoperative chemotherapy for a median duration of 15 weeks (IQR, 10–22 weeks). In total, fifty-nine patients (43.4%) completed therapy. The most common reasons for failure to complete therapy were death and treatment abandonment. In total, thirty-six patients (26.5%) abandoned therapy, fourteen (10.3%) post-operatively and seventeen (12.5%) pre-operatively. The median time to abandonment was eight weeks (IQR, 1–17 weeks).Thirty-three patients (24%) died post-operatively, four (12%) due to treatment related mortality, 22 (67%) due to disease related mortality, and 7 (21%) due to unknown causes at home.

Radiotherapy was indicated for 55 patients (40.4%). Among these, three received it abroad due to its unavailability in Malawi.

Surgical outcomes

Operative or postoperative complications occurred in 12 patients (13% of nephrectomies), most often due to infection (n=6, 6.5%). The 30-day re-operation rate was 4%. There were four cases of post-operative intussusception requiring laparotomy and reduction, and one case of small bowel stricture requiring laparotomy with bowel resection and anastomosis. A single intra-operative death occurred due to suspected hemorrhage versus air embolus. There were no other 30-day perioperative deaths. The median post-operative length of stay was 7 days (IQR, 6–8 days).

Survival outcomes

Median follow-up time was 21 months (IQR, 13–31). Overall survival (OS) at one and two years was 59% (95%CI, 51–69) and 48% (95%CI, 39–60), respectively. Event-free survival (EFS) was 34% (95%CI, 27–43) at both one and two years (Figure 2). Among those who completed all therapy, two-year OS was 80% (95%CI, 69–93) and EFS was 69% (95%CI, 59–82). Two-year EFS ranged from 60% (95% CI, 45–80) in those with Stage I disease to 13% in those with Stage IV disease to (Table 4). Of five patients with Stage V disease, one was alive at 12 months. Among those for whom radiotherapy was indicated, OS was 45% (95%CI, 31–65) and EFS was 31% (95%CI, 21–48).

Figure 2 – Event-free and overall survival of Wilms tumor patients in Malawi.

Panels C & D represent 104 patients for whom staging workup was completed. Stages I & II have been combined due to low sample size among patients with Stage II disease (n=7).

Of 24 patients presenting with lung metastases, 16 (67%) had follow-up imaging. Fourteen patients (88%) had persistent lung involvement, among whom none were alive two years from diagnosis. Both patients with resolution of lung metastases were alive at two years from diagnosis. Of ten patients with liver metastases at diagnosis, a single patient survived to two years. Of patients with IVC involvement, two of thirteen survived.

Within the cohort, 39 (28.7%) patients died from disease progression, 10 (7.4%) from treatment-related causes (e.g., sepsis), and 12 (8.8%) of unknown causes at home (Table 4). Among eight patients returning to care after abandonment, two were alive two years from diagnosis. Of all deaths, 25% occurred within fifty days of diagnosis.

Thirteen children (9.6%) had documented relapse. All were started on a salvage chemotherapeutic regimen with dose-reduced ifosfamide, carboplatin, and etoposide, and four patients had repeat surgery for debulking. The median time to relapse was 15 months from diagnosis (IQR, 13–21 months). Two relapsed patients are alive at 3.6 and 4.7 years from diagnosis, one of whom received radiotherapy.

In addition to patients with persistent metastatic disease, thirteen patients with Stage III disease received escalated therapy with Regimen M post-operatively rather than standard therapy due to absent radiotherapy. Of these, four patients (31%) were alive at 12 months, four (31%) abandoned therapy, four (31%) died, and one (8%) was lost to follow-up prior to 12 months. In comparison, of ten patients receiving standard therapy with follow-up data available to one year, nine were alive, one died, and none abandoned.

Factors Influencing Survival

Factors significantly associated with OS on unadjusted analysis were severe acute malnutrition (hazard radio, HR, 1.9; 95%CI, 1.1–3.2), increasing tumor stage (HR, 1.5/stage; 95%CI, 1.2–1.8), and vena caval tumor involvement (HR, 2.7; 95%CI, 1.4–5.4). Each variable remained significantly associated with mortality on adjusted analysis with similar effect sizes.

Discussion

Wilms tumor is highly curable with well-established surgical techniques, minimal-to-moderately myelosuppressive chemotherapy, and external beam radiation that can be given safely to young children. Its backbone chemotherapy is relatively inexpensive, widely available, and typically well-tolerated with minimal toxicity. Wilms tumor is thus ideally suited to be cured even in resource-limited settings.

Our evaluation of outcomes of a resource-adapted Wilms tumor treatment protocol in Malawi demonstrates the potential for high rates of cure for patients with low stage disease, but also the enduring barriers to broadly realizing these cure rates in resource-limited settings. Beyond establishing pediatric medical and surgical oncology programs in LMICs – these are no small feats – ever more difficult barriers must be overcome to achieve rates of cure reported in high-income countries.

Between 2016–2021, two-year OS among patients in this study was 48%, consistent with contemporary outcomes in the low-resource settings of sub-Saharan Africa.¹³ During this period, regional OS ranged from 0–53% at approximately three years, averaging 44% and 27% in East and West Africa, respectively. Survival reported here is similar to that reported by Queen Elizabeth Central Hospital in Blantyre, Malawi where 2-year OS and EFS was 46% using a similar protocol.¹⁰ By comparison, OS in relatively high-income South Africa averaged 76% during the same period.¹³ In the most recent SIOP and COG reports from high-income countries, OS was >80% in aggregate, approaching 100% in low-risk disease.^5,28,29

With the goal of improving these outcomes, these data highlight the highest impact barriers to address with research and capacity building programs: treatment abandonment, late presentation, and absent radiotherapy.

Treatment abandonment

Treatment abandonment occurred in 27% of patients and had a large impact on event-free survival (EFS) within the cohort. In regional studies, Wilms tumor therapy abandonment ranges from 3% in South Africa to as high as 35% in Sudan.^{11, 30, 31} A recent review of all-cancer abandonment across five African pediatric cancer centers reported that contemporary rates remain consistently above 10%.³²

While high overall survival (OS) among low-stage disease reflects progress within the cancer center, EFS reflects the true burden of mortality at the center. For example, OS for stage I disease was 78%, but EFS was reduced to 60% due primarily to abandonment. Only two patients who abandoned therapy survived to two years after completing treatment upon return, thus abandonment is likely to be fatal in all but rare cases where treatment can be completed. Given its effect on prognosis, addressing abandonment must be as central to resource-adapted protocols as diagnostic protocols and risk-stratification.

Treatment abandonment is a multifaceted phenomenon caused by complex structural and socioeconomic barriers that are often beyond the control of medical providers and families.^32–36 Financial pressures are common and include unaffordable therapy, absent transportation, and loss of parental income. Developing strategies to improve treatment retention is a top priority in sub-Saharan Africa. For example, regional networks such as the Collaborative African Network for Childhood Cancer and Research have decreased treatment abandonment and improved outcomes for Wilms tumor through follow-up programs, robust contact networks, and direct financial support to families.^18,32

Despite providing funding for all transportation and utilizing a program which re-established contact with families following abandonment, abandonment rates remained high at the center. Potential contributors include a large catchment area with travel times frequently over 12 hours, food insecurity at home, and the possibility that financial support for transportation did not adequately address all financial challenges experienced by a population facing extreme poverty. For example, anecdotal experience from the center shows that when faced with food insecurity, families may use transportation money to purchase food.

Further study into protocolized, evidence-based implementation strategies that are generalizable to the heterogenous array of logistical and financial challenges facing families in LMICs is essential. As abandonment is fundamentally preventable, special focus should identify specific, family-level socioeconomic burdens that can be alleviated by LMIC cancer programs.

Advanced disease at diagnosis

Another major contributor to low survival in this study was the high proportion of patients with advanced disease at presentation. Two-thirds of staged patients had either stage III or stage IV disease, a distribution that is consistent with other African studies and is the reciprocal of the stage distribution in high-income countries where 40–60% of patients present at stages I or II.^37,38

Addressing the negative prognostic influence of late-stage cancer presentation in low-resource settings is essential, particularly as capacity-building efforts in LMICs continue to improve outcomes among those who are able to access care at these centers. At 78%, the two-year OS among patients in the cohort with stage I disease was relatively favorable compared to historical experience in the region, albeit limited by the general inability to salvage relapsed disease.¹³ By contrast, patients with stage IV disease, presenting in a quarter of patients, had an OS of 23%. Notably, patients with stage II disease had a 2-year EFS of 0%, driven down by very low number (n=7) and a high rate of abandonment.

Strategies to reduce late presentation in LMICs must be pursued as a comprehensive approach to treating Wilms tumor in our setting. Public health education programs and media engagement strategies to increase cancer awareness can be locally effective, but penetration into LMICs with vast, poorly connected rural regions without access to broadcast media is challenging.^39,40

Engaging traditional healers in regional cancer care programs is another potentially high-impact target given their prominent ad hoc role in LMIC healthcare.⁴¹ When specifically queried, over half of patients in this study endorsed seeking care from traditional healers. While facilitators and barriers to parents’ use of traditional medicine in LMICs is well documented, future research directions may focus on identifying the facilitators and barriers among traditional healers themselves to integration within national pediatric cancer referral pathways.^42,43

Absent radiotherapy

Radiotherapy is prescribed in COG and SIOP protocols for treating high-risk and relapsed Wilms tumor, but it is unavailable in over half of African countries.⁴⁴ While front-line Wilms tumor trials in high-income countries have reduced the regularity and dose of radiotherapy in Wilms tumor therapy, it remains a central feature of modern protocols.

Strategies to eliminate radiotherapy from pediatric cancer protocols are challenging and typically rely on intensifying myelosuppressive therapy.⁴⁵ However, increasing front-line chemotherapy intensity is limited in many LMICs by inadequate supportive care infrastructure, particularly regular availability of blood products. Although a larger study with sufficiently powered hypothesis tests is required to draw conclusions, increasing post-operative therapy intensity to Regimen M among patients with Stage III disease had no apparent effect on survival in the cohort. Furthermore, abandonment was exclusive to patients receiving higher-intensity therapy in this sub-cohort, highlighting the potential risk of treatment intensification leading to higher rates of abandonment.

Beyond continuing to advocate and strategize for better access to radiotherapy in LMICs, patients with persistent pulmonary metastases at time-of-nephrectomy are a potentially high-impact target for novel approaches to Wilms therapy in LMICs. Given the poor prognosis of this sub-group, LMIC treatment guidelines suggest palliation in these cases, supported by data from this study where mortality was 100% in this sub-cohort despite therapy intensification.^{16 15}In centers with pediatric surgical oncology programs, excision of residual, resectable metastatic lesions following neoadjuvant chemotherapy, as implemented in the SIOP 93–01 trial, may allow for cure in subset of patients with refractory pulmonary metastases without requiring radiotherapy.⁴⁵

In SIOP 93–01, patients with persistent, unresectable pulmonary metastases were given intensified therapy after partial metastasectomies and only received radiation if metastatic lesions remained. This regimen achieved EFS of 46% in children with residual pulmonary metastases at time-of-nephrectomy without the need of pulmonary radiotherapy, a significant improvement upon outcomes reported here.

Studies evaluating the safety and effectiveness of more aggressive surgical and medical therapy for patients with pulmonary metastases in LMICs are likely feasible. The overall post-operative surgical complication rate was 13% in this study’s resource-limited setting, the majority being infections treated effectively with antibiotics, thus approximating complication rates of high-resource settings.⁴⁶ Treatment-related mortality occurred in 16% of patients in total and^{20, 29, 45} in only 2/25 patients receiving Regimen M intensification, figures typical of regional experience.^{19, 28, 45}

Strengths and Limitations

This study leveraged high-quality clinical data documenting six years’ experience treating Wilms tumor in a resource-limited setting in sub-Saharan Africa to create among the largest datasets of clinical outcomes in such a setting yet published. In doing so, the study achieves stage-specific insights into the realities of treating Wilms tumor in a low-income country, and it provides support for the priorities to be pursued with next-step funding, research, and intervention. As with all retrospective studies, the study was limited by bias introduced by incomplete data collection or documentation (e.g., chest imaging). Inconsistent drug availability created heterogeneity in the approach to treating individual risk groups and limited conclusions regarding risk-group-specific treatment approaches. Protocol deviations due to stockouts are common in pediatric cancer care in sub-Saharan Africa, thus the protocol’s pragmatic, adaptable approach is generalizable to the broader region.^47,48 Regarding surgical staging, a rigorous staging protocol was applied during the study period, but given the very large size of tumors at diagnosis, sampling error may have occasionally resulted in under-staging of patients. However, given the low proportion of Stage I and II patients in the cohort, this is not expected to have had a large effect on the study’s findings. Likewise, sampling may have under-identified areas of anaplasia in large tumors. Finally, while the median follow-up time within this study of 21 months is expected to have captured the majority of Wilms relapses which typically occur within two years, long-term follow-up after therapy was not maintained for many patients; thus, rates of relapse may be underreported in this study. ²⁸Finally, while the median follow-up time within this study of 21 months is expected to have captured the majority of Wilms relapses which typically occur within two years, long-term follow-up after therapy was not maintained for many patients; thus, rates of relapse may be underreported in this study.²⁸

Conclusions

Survival rates of children with Wilms tumor in this low-resource setting are a stark reminder that the greatest prognostic factor for Wilms tumor survival is local treatment capacity rather than tumor biology. Capacity-building programs in sub-Saharan Africa have resulted in favorable outcomes among children with low-stage disease and those who complete therapy, but factors such as advanced disease at diagnosis, concomitant comorbidities, high abandonment rates, and lack of available radiotherapy represent the important scientific, public health, and capacity-building efforts yet to be sufficiently addressed.