children with acute myeloid leukaemia (AML) were treated in the MRC AML 10 trial. Three risk groups were identified based on cytogenetics and response to treatment. One hundred and twenty-five children relapsed -103 in the bone marrow only, 12 in the bone marrow combined with other sites, and six had isolated extramedullary relapses (site was not known in four cases). Eighty-seven children received further combination chemotherapy, one all-trans retinoic acid for acute promyelocytic leukaemia, and one a matched unrelated donor allograft in relapse, and 61 achieved a second remission. One patient with no details on reinduction therapy also achieved second remission. Treatment in second remission varied -44 children received a BMT (22 autografts, 12 matched unrelated donor allografts, 10 family donor allografts), and 17 were treated with chemotherapy alone. The overall survival rate for all children (treated and untreated) was 24% at 3 years, with a disease-free survival of 44% for those achieving a second remission. Length of first remission was the most important factor affecting response rates -children with a first remission of less than 1 year fared poorly (second remission rate 36%, 3 year survival 11%), whereas those with longer first remissions had a higher response rate (second remission rate 75%, 3 year survival 49%, P Ͻ 0.0001).

Clinical, morphologic, and cytogenetic features were examined in a group of 68 children with myelodysplasia (MDS) referred to a single institution between 1971–1991. The morphologic French-American-British (FAB) system of classification proved of limited value in this group of patients because 50% of the cases were categorized as chronic myelomonocytic leukemia and three patients with eosinophilia and MDS were unclassifiable. Cytogenetic analysis was performed in 63 cases and clonal abnormalities were detected in 55%; the most common chromosome involved was number 7. Modification of the FAB system to incorporate additional diagnostic features such as pretreatment fetal hemoglobin (Hb F) and cytogenetics allowed incorporation of the categories of juvenile chronic myeloid leukemia (JCML) and infantile monosomy 7 syndrome (IMo7). The resulting groups of patients had highly significant differences in survival (P = .00009). The overall 5-year survival for the patients was 31.9% (95% CI 21.7 to 44.1) and factors influencing prognosis included: modified FAB type, platelet count, Hb F level, and cytogenetic complexity. We developed a scoring system (“FPC”) where each of the following findings at diagnosis scored one point: HbF greater than 10%, platelets < or = 40 x 10(9)/L, and complex karyotypic changes (two or more clonal structural/numerical abnormalities), which produced groups with highly significant differences, patients with a score of 0 having a 5-year survival of 61.6% (CI 33% to 84%), whereas those with a score of two or three all died within 4 years of diagnosis. The revised classification and scoring system may prove helpful in making treatment choices in pediatric MDS and now needs to be tested prospectively in large scale population-based studies.

Radiotherapy can be replaced by long-term intrathecal therapy. Intravenous methotrexate gives some additional benefit by reducing non-CNS relapses.

Results of three consecutive completed UK trials (1980)(1981)(1982)(1983)(1984)(1985)(1986)(1987)(1988)(1989)(1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997) for childhood lymphoblastic leukaemia are presented. National accrual has progressively increased so that over 90% of all the country's ALL cases were treated on the latest trial reported, UKALLXI. From 1980 to 1990, event-free and overall survival progressively improved, following adoption of an American therapy template and use of two post-remission intensification modules. Since 1990 despite demonstration of the benefit of a third intensification module overall event-free survival (EFS) has not improved further. Survival remains high due to a good retrieval rate especially for those relapsing off treatment after receipt of two intensification pulses. Possible reasons for the plateau in event-free survival (including type and dose of induction steroid, dropping of induction anthracycline, type and dose of asparaginase, gaps early in therapy following intensification, and overall lack of compliance in maintenance) are being explored in the latest protocol ALL '97. Cranial irradiation had been successfully replaced by a long course of intrathecal methotrexate injections for the majority of patients. Age (Ͻ1 year Ͼ10 years) sex (male) and white count Ͼ50 × 10 9 /l plus slow initial bone marrow clearance were consistently the most important independent prognostic indicators during this time period. Rome/NCI criteria accurately predict standard and highrisk groups for B cell lineage, but not consistently for T cell disease. This international collaborative venture might help us to define those truly at highest risk, and how we can optimise therapy for specific subgroups including T-ALL and those with unfavourable cytogenetics. Leukemia (2000) 14, 2307-2320.