Non-Hodgkin lymphoma in children: a 20-year population-based epidemiologic study in western Sweden

The aim of this study was to investigate incidence, clinicopathologic features, prognostic risk factors, and long-term survival in non-Hodgkin lymphoma (NHL) in a 20-year population-based study of children using Swedish health care organizations and their central registry for childhood malignancies.

The hospital registry, the Cause of Death Registry, and the two established Swedish registries for malignancy (the Swedish Cancer Registry and the National Registry for Solid Tumours in Childhood) were searched for children in western Sweden with NHL diagnosed from 1975 to 1994. The clinical files of all children with NHL were collected and abstracted for information regarding age at diagnosis, gender, disease characteristics, treatment, and outcome of treatment. All sections from paraffin-embedded blocks of tumors with a diagnosis of malignant lymphoma were collected and reexamined histopathologically and immunohistochemically. To guarantee that no patients with NHL were misdiagnosed, a reexamination of other childhood malignancies collected from these registries was also performed. Median follow-up duration of surviving patients is 10 years.

The annual incidence of NHL in children younger than 15 years of age was 9/million children, representing 6% of all childhood malignancies during the investigation time. The male-female ratio was 4.1:1.0. Immunologic marker studies were available for 64 of the 77 NHLs: 41 patients had B-cell, 17 had T-cell, and 6 had Ki-1-positive anaplastic large cell lymphoma (ALCL). Two patients with Ki-1-positive ALCL were originally thought to have malignant histiocytosis and Langerhans cell histiocytosis (LCH), respectively. Treatment was the most significant prognostic factor; event-free survival (EFS) was 19% in the preprotocol era (1975 to 1979) and 74% from 1980 to 1994. Other than treatment, stage was the most significant prognostic factor; EFS was 86% for patients (1980 to 1994) with stage I or II disease and 64% for patients with stage III or IV disease, with a dismal prognosis for children with initial involvement of the bone marrow or central nervous system (EFS was 38% and 20%, respectively). Bulky disease and performance state at diagnosis were independent prognostic factors. The patterns of relapse, including early recurrence of the B-cell lymphomas, are in accordance with previous experience.

The incidence of NHL was found to be somewhat higher than reported in our previous Nordic study. The higher incidence found in this study might be the result of the thorough data collection (based on hospital registry and cross-checked with all registries for malignant diseases in Sweden) or because reexamination of the tissue material was performed. A more pronounced male predominance than found in previous investigations was observed. The immunophenotypic distribution and the stage distribution is in accordance with earlier investigations. Treatment was the most important factor affecting outcome. A dramatic improvement of survival was seen with the introduction of intensive therapy; treatment success can be expected in 86% of children with localized disease and 64% of children with extensive disease. The absence of improvement in survival despite further treatment stratification with the introduction of the BFM protocol for B-cell-NHL is surprising. LSA2L2-like protocols seem to be as effective. Future studies on treatment of NHL must also concentrate on reducing the intensity of therapy in patients with lower risk disease to minimize late toxic effects.