Assessment of immunophenotype and DNA content of leukemia is important for diagnosis and risk assessment, but requires costly equipment, reagents and technical expertise. Samples from 1118 children in Central America were sent for flow cytometric examination to a reference laboratory in Guatemala City. Consultation with St Jude Children's Research Hospital facilitated quality control. Acute leukemia was diagnosed in 908 cases. Consultation was required in 20% of cases during year 2 and only 3% during year 5. Cost per sample was $240 vs $720 if performed at St Jude Children's Research Hospital. Immunophenotyping in resource-poor areas proved feasible and cost effective.

Immunophenotyping of leukemias is necessary for diagnosis and risk-directed therapy.1 Otherwise, misdiagnosis (eg confusion of acute lymphoblastic leukemia (ALL) with acute myeloid leukemia (AML)) may occur in as many as 10% of cases.2,3,4,5 Flow cytometry is widely used for immunophenotyping because it allows quantification of antigen expression, rapid analysis of large numbers of cells, and determination of DNA content,6,7 an important prognostic indicator.8,9 One-fourth of children with B-lineage ALL have a leukemic cell DNA index (ratio of DNA content to that of normal diploid cells) ⩾1.16, a favorable prognostic factor allowing use of less intensive chemotherapy.

Immunophenotyping is often unavailable in countries with limited resources. The resulting diagnostic uncertainty and inability to stratify precisely patients based on relapse risk prevents the use of risk-directed therapy and compromises outcomes. Several pediatric oncology programs in Central America previously shipped diagnostic specimens to St Jude Children's Research Hospital (SJCRH) for flow cytometric analysis. However, shipping and processing of samples were expensive, delayed diagnosis, and adversely affected the sample quality. To solve these problems, the SJCRH International Outreach Program collaborated with a pathologist (RL) to establish a flow cytometry laboratory in Guatemala City to serve cancer centers in Guatemala, El Salvador, and Honduras. We analyzed data from the first 1148 samples to evaluate feasibility and cost of referral of specimens to a regional center for immunophenotyping of childhood leukemia.

From April 30, 1998 to December 31, 2002, specimens from pediatric patients with suspected leukemia at the Hospital Escuela (Tegucigalpa, Honduras), the Hospital Benjamin Bloom (San Salvador, El Salvador), and the Unidad Nacional de Oncología Pediátrica (Guatemala City, Guatemala) were analyzed by the Pathology and Flow Cytometry Laboratory in Guatemala City. Specimens from El Salvador and Honduras were shipped by overnight express and processed on the day of arrival.

Flow cytometric analysis was performed by standard methods using commercially available antibodies.3 Sample processing, antibody staining, and flow cytometric analysis were performed by a pathologist (RL) who had attended Becton Dickinson's standard training course plus an additional week of training at SJCRH in Memphis, TN, USA. Web-based consultation with SJCRH, as described previously, facilitated quality control.10

A total of 1148 diagnostic specimens (1089 bone marrow, 59 blood) from 1118 patients were analyzed. Of these, 30 were recollected to replace specimens that could not be evaluated because of nonviable cells due to delayed shipping or processing (n=11), bacterial contamination (n=3), dilution of bone marrow with blood (n=6), or inconclusive findings (n=10). In these cases, a diagnosis was subsequently made from the recollected sample. All but 11 specimens arrived within 72 h of sampling; results were returned by facsimile or e-mail within 48 h.

Of the 1118 evaluable samples, 908 indicated acute leukemia: 650 B-lineage ALL (positive for CD19, CD22, and/or CD79a), 62 T-lineage ALL (positive for CD7 and cytoplasmic CD3), three mixed-lineage ALL (positive for both B- and T-cell markers), 182 AML (positive for CD13, CD33, and/or myeloperoxidase, and negative for lymphoid-associated markers), and 11 unclassifiable (cells expressed lymphoid and myeloid markers or showed no conclusive lineage differentiation). A diagnosis of acute leukemia was ruled out for 210 patients, 11 of whom were subsequently found to have other cancers: chronic myeloid leukemia (8), non-Hodgkin's lymphoma (2), and myelodysplastic syndrome (1).

During year 2 of the program, an on-line system was developed that allowed direct transfer of raw flow cytometry data from the cytometer in Guatemala City to the cytometer at SJCRH. This innovation allowed consultation to clarify the diagnosis or the technical adequacy of results. Consultation was requested in 20% of cases during year 2, but in only 3% during year 5 (Table 1 ). Consultation led to a change in or refinement of the diagnosis in 4% of all cases in year 2 compared to only 1% in year 5.

Table 1 Requests for on-line flow cytometry consultation and the outcome of consultation by year
Full size table

The average cost of analysis per sample, including shipping, antibodies, technician salaries, professional fees, and overhead, was $240 as compared to $720 if the samples were shipped to SJCRH. This difference reflects the lower cost of shipping regionally and the lower salaries paid in Guatemala.

Accurate diagnosis and risk assessment are essential for optimal treatment of childhood leukemia. Immunophenotype and DNA content provide important diagnostic and prognostic information. Risk stratification also ensures cost-effective use of treatment resources.11 While it is impractical to establish a diagnostic laboratory in each pediatric oncology treatment facility due to the limited number of patients, we developed a regional reference laboratory that provides high-quality testing at relatively low cost. Centralization of testing reduced costs of shipping and laboratory overhead. Ongoing training and quality assurance were accomplished by our innovative use of a web-based system for rapid exchange of flow cytometric data between centers, which proved highly effective.10 Importantly, in the fifth year of the program, after analysis of hundreds of cases, the reference laboratory rarely required consultation with SJCRH (3% of cases) and the consultations refined or altered the diagnosis in only 1% of all cases (two of eight consulted cases). We consider this program to be a paradigm of technology transfer from developed countries to those with limited resources in a way that builds local capacity and fosters independence – characteristics essential for sustainable development.12,13,14,15,16 We have previously suggested that low-cost infection control programs, improved supportive care, and prevention of abandonment of therapy are important factors for improving outcomes of children with ALL in countries with limited resources.17,18 As correct diagnosis and risk classification are essential for appropriate and cost-effective treatment of children with acute leukemia, we would add flow cytometry to this list.