Prognostic Factors in Childhood Anaplastic Large Cell Lymphoma (ALCL): Results of the European Intergroup Study.

Purpose and methods

In order to study prognostic factors of disease progression and relapse, the data of 235 children enrolled for an ALCL between 1987 and 1997 in the BFM (93 patients, pts), SFOP (82 pts) and UKCCSG (60 pts) studies have been merged in 1998. All these children have been treated according to protocols designed to treat childhood ALCL with a short and intensive chemotherapy treatment, similar to those used for B-cell. Slides had been reviewed by the national pathologist panel for 226 patients (96%). Hodgkin-like ALCL were excluded from this series. Follow-up has been updated in April 2006.

Results

Among the 235 pts included in the study, 206 pts (88%) achieved a CR. With a median follow-up of 7.4 years, 77 events were observed (13 early progressions, 55 relapses, 3 secondary leukaemias, 6 toxic deaths) and 46 patients died. The 5-year overall survival and event-free survival of the whole population is 81% [95%CI, 76–86%] and 69% [63–75%] respectively. Event-free survival differs significantly according to the treatment (p=.01) : the risks of failure of children treated in the SFOP studies and in the UKCCSG studies are multiplied by 1.8 [1.05–3.2] and 2.3 [1.3–4.1] respectively, as compared to pts treated in the BFM studies.

Several factors seem to be prognostic of the risk of early progression or relapse in univariate analysis: B-symptoms, mediastinal involvement, skin lesions, visceral involvement, St Jude stage 3–4 or Ann Arbor stage 3–4 or elevated LDH. Patients with bone lesions seem to have a better prognosis than others. Soft tissue masses do not worsen the prognosis.

In multivariate analysis, three factors are found to be associated to the risk of progression/relapse :

• * mediastinal involvement (hazard ratio of progression/relapse, HR=2.2 [1.3–3.6], p=.003),

• * visceral involvement defined as lung, liver or spleen involvement (HR=2.1 [1.3–3.5], p=.005),

• * skin lesions (HR = 1.9 [1.2–3.2], p=.01).

Based on the results of this Cox model, it is possible to define:

• * a good prognosis group (85 patients without skin, mediastinal or visceral involvement) with 89% pts [82–96%] free of progression/relapse at 5 year

• * a poor prognosis group (150 patients with skin and/or mediastinal and/or visceral involvement) with 61% pts [53–69%] free of progression/relapse at 5 year

This risk group classification appears significant in each of the three treatment groups.

Conclusion :

Results obtained with BFM-studies are significantly better than those obtained with the French and British pediatric studies. Pts with mediastinal involvement, visceral involvement and/or skin lesions are at a higher risk of disease failure. These results was the basis of ALCL99 aiming to improve EFS of high risk pts by adding vinblastine to BFM protocol.