Introduction Current protocols use radiotherapy (craniospinal irradiation) as the treatment of choice to cure isolated central nervous system (CNS) acute lymphoblastic leukemia (ALL) relapses. The severe toxicity of this treatment encouraged us to develop a new CNS-ALL protocol without radiotherapy.

Patients and methods From Jan 1987 till Aug 2004, 13 children were diagnosed in our centre with an isolated CNS relapse after initial treatment according to standard DCOG-ALL protocols. Treatment of CNS relapse consisted of induction by weekly intrathecal Methotrexate (MTX). At remission, an Ommaya reservoir was implanted and CNS-directed intraventricular sandwich therapy, consisting of MTX day 1; ARA-C day 2 and MTX day 3 (dosage according to age) was given every four weeks for one year. At the same time systemic treatment, based on the ALL-6 protocol (

), was started in which at week 23, 44 and 65 intensification courses of 6 weeks duration with Teniposide, HD-ARA-C and HD-MTX were inserted. The total duration of treatment is 95 weeks. Six of 13 patients could be assessed for behavior, intelligence, memory, visual-spatial and visual-motor skills before and after treatment using the CBCL and WISC-RN tests among others.

Results All 13 patients, 3 girls and 10 boys aged 2.3 till 14.8 years (9 precursor B-ALL and 4 T-ALL), had an early isolated CNS relapse after a median first remission duration of 16 months (range 2–30 months). Nine of them were high risk according to BFM relapse criteria (male, age < 6 years, T-ALL phenotype, relapse < 18 months from diagnosis). At present, eight patients are alive in 2nd complete remission (CR) with a median follow up of 82 months (range 7–189 months). Five patients relapsed, all high risk, of which three died. One died after a secondary AML, one after a bone marrow (BM) relapse in 2nd CR due to fungal sepsis and one after a combined BM and CNS relapse due to streptococcal meningitis/encephalitis during neutropenia. The fourth patient had a second CNS relapse after 42 months in second remission. He is still in 3rd CR for 86 months after an autologous BM infusion. The fifth patient had recently an isolated BM relapse after 11 months in 2nd CR and started systemic reinduction therapy. The 5 years EFS of this study is 57% ± 15% and the 5 years OS 73% ± 14%. Before start of chemotherapy no significant differences in psychological testing were found in comparison with the normal population. After stop chemotherapy significant lower scores were obtained on the domains of perceptual organization and behavior similar to those found in other patients treated for cancer. Furthermore, our treatment protocol has no significant effect on neurocognitive functioning in comparison with craniospinal radiotherapy.

Conclusion Sandwich intraventricular therapy together with systemic anti-leukemia therapy without radiotherapy seems to be an effective treatment with minimal neurocognitive disfunctioning for isolated CNS-ALL relapse. Further investigations in a larger group of patients are essential with special emphasis on comparing late effects of this therapy with radiotherapy.

Author notes

Corresponding author