Introduction of treatment for children with acute lymphoblastic leukemia in a resource-limited hospital in Cambodia Open Access

In resource-rich countries, improvements in diagnostic classification, subgroup treatment allocation, and selective therapy intensification have allowed for better relapse-free survival (RFS) of children with acute lymphoblastic leukemias (ALL), exceeding 90% in many cases.¹ However, many of these improvements are not available in resource-limited settings. Reasons for this include limitations in funding, lack of necessary equipment, limited laboratory testing options, less extensive hospital facilities, fewer fully trained caregivers, lower health competency of families, drug accessibility, and difficulties accessing health care facilities.^2-5 Despite these significant challenges, many low- and middle-income countries (LMICs) have been able to implement lower intensity regimens for ALL treatment, albeit with varying levels of survival.^6-9 

Angkor Hospital for Children (AHC) is a philanthropic-supported hospital in Siem Reap, Cambodia that cares for children aged ≤16 years, regardless of ability to pay. Since 2007, AHC has been a part of the American Society of Hematology (ASH)’s global volunteer collaboration with Health Volunteers Overseas (HVO). Through the ASH-HVO partnership, volunteer physicians and nurses travel to AHC throughout the year, providing in-person mentoring. In efforts to provide ongoing support to the physicians treating patients at AHC with hematologic and oncologic diagnoses, a weekly case conference was established, during which the AHC team presents patients for discussion with a group of volunteer US-based physicians, including pediatric hematologist/oncologists and pathologists. In addition, there was regular email correspondence for ongoing mentorship and case discussions for times when volunteers were not physically present at AHC. Educational sessions were provided to support staff including nurses and pharmacists on oncology fundamentals including chemotherapy administration and supportive care. There are an estimated 3.5 million children aged <16 years in Cambodia, with ∼100 children per year diagnosed with ALL in the country. Siem Reap is the second largest city in Cambodia, and AHC serves patients and families from many remote villages throughout Cambodia, especially the northwest region of the country. AHC has a total of 90 inpatient beds available, with 12 of them dedicated to the care of oncology patients. The oncology team at AHC is made up of 3 pediatricians with additional training in hematology and oncology, as well as a team of nurses with advanced training in chemotherapy delivery and pharmacy support. The oncology team also has access to the AHC pediatric intensive care unit (PICU), surgery, and microbiology teams of infectious disease specialists. AHC has a laboratory on site that can complete blood counts and metabolic panels, with a general turnaround time of <12 hours. The hematology-oncology program at AHC began to treat children with selected diagnoses in 2013 and expanded to treat ALL in 2018. The ALL treatment strategy follows a protocol with curative intent, which is designed for a resource-limited setting, with modifications particular to the local resources and capacity at AHC, and with the support of US laboratories to assist with diagnosis and response assessment because local pathology is not available at AHC. After the initial implementation of a very low-intensity regimen, modifications in November 2019 were made to intensify the regimen with the increasing experience of the AHC oncology team. Here, we report the results of the first 5 years after the implementation of ALL treatment with patients treated on the regimens from 2018 to 2023.

Diagnosis of ALL was based on both bone marrow and peripheral blood morphology. Images from peripheral blood and bone marrow aspirate were sent by email for initial morphologic review. The ALL diagnosis and immunophenotype were confirmed by flow cytometric analysis of the bone marrow aspirate or peripheral blood through a partnership with Hematologics, Inc (Seattle, WA). In cases in which flow could not be performed, the diagnosis was established by immunohistochemical studies (CD3, CD20, PAX5, CD10, CD34, TDT, and myeloperoxidase) performed on the bone marrow core biopsies sent to Children’s Wisconsin (CW; Milwaukee, WI). These diagnostic tests were performed at no charge to AHC as part of the partnerships established in developing the ALL program in Cambodia. Karyotypes and molecular studies were not performed. Bone marrow and/or peripheral blood samples were shipped to the United States, and the turnaround time from bone marrow assessment to return of results was generally <10 days. Central nervous system (CNS) status was assessed at diagnosis in 79 of 84 patients (94.0%) before initiating therapy. In 5 patients (6%), the sample was inadequate for assessment. CNS 1 status was assigned to patients with <5 white blood cells (WBC) with no blasts, CNS 2 was <5 WBC with blasts, and CNS 3 as WBC ≥5 with blasts. CNS assessment was performed by morphologic evaluation of the CSF by the laboratory personnel and physicians at AHC. Samples with uncertain morphology were reviewed at CW using provided images of suspicious cells from cerebrospinal fluid (CSF) cytospin preparation slides or, in some cases, slides were sent to CW for review and assessment. Complete remission (CR) was defined as <5% marrow blasts by morphology and no evidence of leukemia elsewhere at the end of induction (EOI). Measurable residual disease (MRD) status in the marrow was assessed by flow cytometry by Hematologics, Inc at EOI. Although the level of sensitivity of the assay is <0.01%, there were cases where sample viability limited this level of sensitivity; thus, MRD negativity was defined as no detectable blasts in the sample and ranged from a level of sensitivity of <0.01% to 0.3%. Patients who had MRD detected at the EOI underwent a second assessment at the end of either the intensification phase or reinduction, depending on their treatment regimen, to determine whether MRD-negative status had been achieved before proceeding with further therapy.

The treatment regimen (Table 1) used at AHC was designed based on limited resource availability at AHC, including limited financing and availability of some chemotherapeutic agents, as well as limitations around supportive care requiring modifications to minimize risk of neutropenia, infections, transfusion, and renal toxicity related to the treatment regimen. Using these guiding principles, significant changes to the established ALL regimens were made to create a very low-intensity treatment regimen that could be augmented over time. The regimen started with a 7-day prednisone prophase to assist with cytoreduction and allow for confirmation of the diagnosis in cases in which ALL was suspected based on morphology review of slide images but had not yet been confirmed by flow cytometry or immunohistochemistry. The prophase was followed by a 3-drug induction with IT methotrexate. Due to practices at other hospitals in the area, a substantial proportion of children had been pretreated with steroids for variable times in the 2 months before treatment initiation at AHC. Thus, in 2020, a decision was made to give the prophase only to non–steroid-pretreated patients. After induction treatment, remission was assessed by morphologic examination of the bone marrow and confirmed by measurement of MRD with flow cytometry in 88 of 92 patients (95.7%) who were alive at the EOI. Intensification was then begun using Capizzi escalating methotrexate which was followed by either continuation or a reinduction phase. In 2019, a reinduction phase identical to induction was introduced for all patients in efforts to further improve outcomes from the initial protocol after an interim analysis demonstrating its general safety and feasibility. Patients determined to be MRD negative by flow cytometry at the end of either intensification or reinduction proceeded to continuation therapy. Due to lack of other treatment options, patients who had persistent MRD were permitted to continue treatment, although their persistent MRD was considered positive for disease in determination of RFS. Treatment was continued for 2 years after the end of the intensification phase.

Consecutive children with ALL who presented or were referred to AHC from March 2018 to December 2023 were included in this study. The diagnosis and treatment recommendations were discussed with the families, and written informed consent before starting chemotherapy was obtained in Khmer or English.

Standard supportive care guidelines were introduced, which included the initiation of allopurinol for all patients at the time of suspected leukemia and IV hydration, and both interventions were continued through the start of chemotherapy. Patients generally received all chemotherapy and supportive care through a peripheral IV. Transfusion support was generally available with packed red blood cells and platelets as indicated by clinical status, although intermittent shortages in supply did occur. Prophylactic sulfamethoxazole-trimethoprim was given to all patients after the start of treatment until 3 months after the completion of treatment. Patients with a fever before diagnosis and/or during induction, as well as in periods of neutropenia at any point throughout therapy, were started on empiric meropenem, which was selected in consultation with the local infectious disease physicians and based on hospital susceptibility patterns. The first cohort of patients treated on this regimen for the first ∼6 months after implementation were required to remain hospitalized throughout induction. All subsequent patients were hospitalized based on clinical needs and whether outpatients were required to stay within close proximity to the hospital until the start of continuation with lodging support provided through social services at AHC.

Continuous variables were described using median and range, whereas categorical variables were described using frequency and percentage. RFS was defined as the time from the start of chemotherapy until either the date of relapse or death in the absence of relapse. RFS was estimated using the Kaplan-Meier method, and any tests comparing RFS between groups were done using the log-rank test to compare the entire survival curves.¹⁰ Relapse risk (RR) was defined as the time from the start of chemotherapy until relapse from primary disease. Treatment-related mortality (TRM) was defined as the time from the start of chemotherapy until death in the absence of relapse. RR and TRM were computed using the cumulative incidence estimator, and any tests comparing RR or TRM between groups were done using Gray’s test to compare the entire incidence curves. Relapse was the competing event for TRM, and death in the absence of relapse was the competing event for RR. All patient outcomes were censored at the time of last follow-up. Overall survival was not assessed because most families did not communicate with their providers after relapse/initiation of palliative therapy and thus were lost to follow-up. However, because relapse treatment was not available at AHC or other centers in Cambodia, it can be assumed that most children who relapsed have died.

Between March 2018 and December 2023, a total of 98 consecutive patients with ALL (B-cell ALL [B-ALL], n = 88; T-cell ALL [T-ALL], n = 10) were treated on the ALL regimen (Table 2). The median age of patients was 4.6 years (range, 0.2-15.6). Patient characteristics are summarized in Table 2. Because many patients had sought prior medical care, and steroids are commonly prescribed for numerous presumed conditions, we also analyzed patients by whether they had received steroid pretreatment, defined as at least 1 week of steroid treatment in the 2 months preceding the initiation of therapy at AHC. In 18 patients (18.4%), steroid pretreatment had occurred.

EOI MRD was evaluated in 88 of 98 patients (89.8%) in the cohort. Among the 10 patients for whom EOI was not evaluated, most (n = 9) were due to induction deaths. There was 1 patient for whom a sample was not sent because of COVID lockdowns. Among patients with MRD assessments, 22 (25%) were MRD positive at the EOI (Table 2). Among patients with B-ALL, the overall EOI MRD-negative rate was 70%. For patients with T-ALL with MRD assessments (n = 8), 4 (50%) were MRD-negative at the EOI. Because protocols for relapsed/refractory ALL were not available at AHC, patients with detectable disease proceeded with intensification (and reinduction after its introduction as a treatment regimen modification). Among the patients with positive MRD at the EOI, 17 patients (77%) underwent subsequent end of reinduction evaluation. For the 12 patients with detectable blasts in the CNS (CNS 2, n = 3; CNS 3, n = 9) at the time of diagnosis, 11 were alive at the EOI and evaluable for CNS response, and 7 (63.6%) had cleared CNS disease at the EOI.

The 2-year RFS for all patients was 66.3% (95% confidence interval [CI], 56.2-78.2; Figure 1). Survival analysis according to immunophenotype demonstrated a 2-year RFS of 69.6% (95% CI, 58.9-82.1) for B-ALL (supplemental Table 1). Analysis outcomes for T-ALL and infant ALL were limited by the small number of patients in each cohort; thus, a descriptive analysis is included in supplemental Table 1. Given the rate of steroid pretreatment in our cohort and concerns from prior studies that this can negatively affect outcome, we analyzed outcomes according to steroid pretreatment in B-ALL. Steroid pretreated patients (n = 18) experienced a 2-year RFS of 79.3% (95% CI, 47.4-93.1) vs 66.9% (95% CI, 51.6-78.3) for nonpretreated patients (n = 65; P = .5; Figure 2A). Analysis according to EOI MRD status demonstrated a significantly superior outcome for patients who were MRD negative at the EOI vs MRD positive, with the marked divergence occurring mostly after the 30-month mark (P = .013; Figure 2B).

The 1-year RR for the cohort overall was 9.9% (95% CI, 4.6-17.7) and at 2-years was 21.0% (95% CI, 12.3-31.3; Figure 3A). Analysis according to MRD status at the EOI demonstrated a 2-year RR of 31.9% (95% CI, 12.4-53.5) in MRD-positive patients vs 20.2% (95% CI, 9.7-33.4) in MRD-negative patients (P = .002; Figure 3B). There were 3 patients who had persistent MRD detected at the end of reinduction, and all 3 patients experienced progression to fulminant relapse while receiving ongoing therapy. The 2-year RR among patients with B-ALL was 16.7% (95% CI, 8.4-27.5), and although infant and T-ALL cohorts were too small to directly compare, descriptive analysis of RR is reported in supplemental Table 2. Consistent with a relatively more aggressive phenotype, the 1-year RR was 56.5% for the T-ALL cohort (n = 10), with all but 1 reported relapse occurring in the first 12 months of therapy. Among the sites of relapse for the cohort overall, medullary relapses occurred in 18 (75%) and CNS relapses in 3 (12.5%), with ocular involvement specifically being reported in additional 3 relapses (12.5%), 2 of whom had T-ALL.

The highest incidence of TRM events was observed during induction, with an induction TRM of 7.2% (95% CI, 3.2-13.6). Among the causes of induction death, infection was the most common in 91% patients (n = 10) and included bacteremia (n = 10) and concomitant fungal infection (n = 5). There was 1 patient who died in induction due to complications of leukostasis. Deaths occurred throughout induction, with the median number of days from the start of chemotherapy to death of 18 (range, 4-33). The rate of TRM at 1 year was 11.0% and at 2 years was 12.7%; thus, most of the observed nonrelapse deaths in our cohort occurred very early in therapy.

A total of 3 patients (3.1%) had treatment abandonment (TA) after their families declined further treatment (at 6, 7, and 21 months after starting therapy). All were considered to be in remission at the time of TA and were censored for analysis at this time in remission. The reasons for TA included family complications (n = 2) and pursuing alternative therapy (n = 1).

Here, we report the first results of using a uniform regimen for ALL treatment for children at AHC in Cambodia. This relatively low-intensity regimen was initiated as a part of a long-standing collaboration between ASH, HVO, US international physician and nurse volunteers, and the leadership, physicians, nurses, and associated personnel at AHC in Siem Reap, Cambodia. The 2-year RFS resulting from this protocol was 66.3%, with a corresponding 2-year RR of 21% and TRM of 12.7%. This represents a substantial improvement compared to the survival for children with ALL managed at AHC before protocol implementation in 2018, which included palliation only and all children died of their disease.

We highlight several notable factors contributing to the success that has been achieved in implementing an ALL protocol in Cambodia. The first factor is that ALL treatment was initiated in the context of an ongoing collaboration between ASH-associated physicians and nurses and the medical team at AHC. This collaboration includes >15 years of international physician volunteers traveling to Cambodia, physician-trainees from Cambodia traveling to the United States regularly for ongoing training and to attend annual ASH meetings, and the growing experience with chemotherapy at AHC since 2013. The second factor is that ALL care at AHC is performed according to a uniform treatment regimen, enabling consistent delivery of care for every patient. The third factor is that ALL diagnostics were able to occur in collaboration and support of US physicians, facilitating accuracy in diagnosis and staging, without any cost to AHC or the patients and families it serves. We recognize that accurate and timely diagnostic capabilities are a significant challenge in many LMICs; improving accessibility and costs to these tools is critical for improving oncology care in resource-limited settings. The fourth factor is that AHC has robust supportive care capacities for the region, including an expert infectious disease team, dedicated social work and oncology nurses, and blood product transfusion capabilities.

Several factors contributed to the lower cure rate achieved at AHC than that observed in higher-resource settings.^11-14 The first is likely the less-intensive ALL treatment regimen used in patients at AHC compared to what is used in high-income countries or those with significant experience in treating ALL. AHC’s ALL protocol does not include some agents that are used in high resource countries (eg, uniform anthracyclines, high-dose methotrexate, targeted therapies, and immunotherapies such as blinatumomab) that are not readily available or are cost prohibitive in Cambodia. Thus, the protocol was designed being cognizant that access to AHC by either public or private transportation is challenging for many patients. For this reason, less-intensive treatment was favored because this would likely be associated with fewer complications in the outpatient setting, where local care facilities are limited. Finally, although we were not able to assess for adherence to continuation therapy, we hypothesize that given more limited understanding of factors affecting cancer occurrence, diagnosis, and treatment, adherence to full protocol therapy was likely less than what is commonly observed in more developed settings. Importantly, we observed a lower TA rate than what has generally been reported in LMICs.^15,16 We think this may be due to the robust social and educational support provided to families at diagnosis and throughout treatment at AHC, as well as the fact that AHC does not charge families more than they can afford to pay, thereby lowering the incentive to abandon treatment due to financial burden.

A recent publication from Küpfer et al reported on 110 consecutive patients with ALL treated at another hospital in Cambodia using a more intensive 4-drug induction and intensive consolidation approach.¹⁷ Although 65% of patients achieved a CR on this protocol, 26% died during induction, and an additional 24.5% died during consolidation, resulting in a 3-year event-free survival of only 35%. Thus, by using a lower-intensity regimen with comparatively lower TRM in induction and overall, we report a comparatively more favorable RFS than the more intense strategy reported by Kupfer. The protocol at AHC was specifically designed to avoid higher-intensity treatment strategies that typically result in longer periods of neutropenia until a track record with lower-intensity treatment and management of toxicities had been established.¹⁸ In addition, due to limited renal supportive care resources, we used a prednisone prophase that may have helped gently reduce disease burden because there were many patients who presented with elevated WBC, and we had no patients who experienced severe tumor lysis syndrome.

Similar to results from higher-intensity regimens in resource-rich settings, MRD at any time during treatment remains a poor prognostic factor.^19-21 In our series, we observed significantly higher rates of relapse at 1 and 2 years in the MRD-positive vs MRD-negative cohorts. In our analysis, patients with B-ALL made up most of the cohort. It is also important to note the survival according to MRD appeared to diverge at ∼30 months from diagnosis. This suggests that for patients with persistent MRD at the EOI, this relatively low-intensity chemotherapy regimen was able to achieve control of the disease but is not likely to provide durable remissions and cure. This provides a rationale to further modify our protocol to intensify therapy, especially for those patients who have postinduction MRD. Based on the tenets of chemotherapy protocol development in resource-poor countries put forth by Hunger et al, the success in implementing this first ALL protocol at AHC provides a strong rationale for careful intensification of the treatment regimen to improve disease control while maintaining patient safety.¹⁸ We acknowledge that our results have relatively short follow-up for some patients, and as expected with the kinetics of B-ALL relapse, we will continue to see relapses occur after the 2-year time point. Even among the MRD negative cohort, survival continued to decrease after the 2-year mark in patients with available longer follow up. Thus, further follow-up would be required to determine the definitive long-term survival in this cohort. With the initial experience gained in treating patients with ALL, the team at AHC in collaboration with US partners is now developing a modified protocol, specifically for patients with persistent MRD, and toxicity and TRM rates will be closely monitored.

Although flow cytometry for diagnosis and MRD evaluation was available for all patients, it is important to note that no other prognostic factors were used in risk stratification or analyzed for impact on outcome. Risk-adapted regimens have been successfully used in LMICs.⁷ This is a current focus of AHC as they incorporate anthracycline for patients with persistent MRD and utilization of intermediate- or high-dose methotrexate for selected patients. Additionally, cytogenetic data were not available in any patients, and this has been shown to be highly predictive in children with ALL at identifying cohorts of children for whom lower-intensity therapy is adequate to achieve cure in most cases and those for whom intensified therapy is required.^22,23 Future efforts aimed at enhanced risk stratification may help identify patients for whom lower-intensity therapy can be given without compromising survival, as well as patients who require an intensified approach, while balancing the higher TRM associated with the increased treatment intensity and supportive care challenges.

In conclusion, we describe the first results of an ALL protocol implemented by the ASH-HVO-AHC collaboration in Siem Reap, Cambodia, which has led to dramatic improvement, from a uniformly fatal disease with palliative care to a 2-year RFS of 66.3%. In efforts to continue to improve outcomes for these patients, further refinements in diagnostic techniques with a focus on using more local resources, additional supportive care resources, ensuring drug accessibility, education of families and the referring community, and protocol intensification will be needed.

The authors thank Hematologics, Inc and the Department of Pathology at the Medical College of Wisconsin for donating their time and resources to support diagnostic efforts at Angkor Hospital for Children (AHC). The authors also thank the American Society of Hematology and Health Volunteers Overseas for their support of the education opportunities for physicians at AHC.

Contribution: K.T., B.C., L.S.K., F.G.K., S.L., and B.S. conceived of and initiated the study; K.T., B.C., and L.S.K. wrote the initial draft of the manuscript; F.G.K., M.J.B., and B.S. edited the manuscript with input from all other authors; S.L., B.S., S.H., T.B., H.S., and N.C. oversaw the treatment and contributed to data collection; M.R.L., L.E.B., C.A.H., V.L., A.H., and J.J. contributed to pathologic analyses; and K.H. and D.S.N. contributed to data analyses.

Conflict-of-interest disclosure: M.R.L. has equity and ownership in Hematologics, Inc. L.E.B. had equity and ownership in Hematologics, Inc during the period of the study. The remaining authors declare no competing financial interests.

Correspondence: Katherine Tarlock, Seattle Children’s Hospital, 4800 Sand Point Way NE, Seattle, WA 98105; email: katherine.tarlock@seattlechildrens.org.