Preventive CNS Irradiation with 12 Gy Compared to 18 Gy: Results of Studies AML-BFM 98 and 2004.

To achieve a power of 80% for non-inferiority (range 11%) 240 patients per group were required, which could only be achieved by randomization over two study periods (7/1998 - 4/2009).

Out of a total number of 1,202 study patients, 471 were not eligible for randomization (126 patients with Down syndrome, 138 patients who did not achieve remission or died/relapsed during the first 140 days before irradiation, 97 patients with initial CNS involvement and 110 patients who were assigned to stem cell transplantation [SCT]). Out of the 731 eligible patients, 241 refused randomization. Thus, a total of 490 children and adolescents <18 years with de novo AML were randomized to receive 12 Gy (n = 248) or 18 Gy (patients between 15 and 24 months: 15 Gy) (n = 242). Eighteen of the randomized patients did not receive radiotherapy. Four of the patients randomized to 12 Gy were treated with 18 Gy, and 15 patients randomized to 12 Gy received 18 Gy. Since this was a non-inferiority study, the analysis was performed for patients as treated (12 Gy n=251, 18 Gy n=221).

Treatment regimens of study AML-BFM 98 (see Creutzig, J Clin Oncol 24, 2006) and AML-BFM 2004 were largely similar. AML-BFM 2004: AIE (cytarabine, idarubicin and etoposide) or ADxE (Dx = liposomal daunorubicine) followed by HAM (high-dose cytarabine and mitoxantrone) and subsequently, two short therapy cycles with medium and high-dose cytarabine and anthracyclines. Intensification and maintenance were as in study -98. Allogeneic SCT from a family donor was restricted to high-risk patients. Intrathecal cytarabine was given 12 times in addition to CNS-RT.

The patient characteristics were similar in both randomized groups. Five-year survival (OS), event-free survival (EFS) and relapse rate (cumulative incidence) were similar in patients who received 12 Gy or 18 Gy (80%±3% vs. 77%±3%, 68%±3% vs. 60%±4%, 30%±3% vs. 36%±4%, respectively. The lower limit of the one-sided confidence interval for the difference in 5-years pEFS (7%) was 0.004. There were 7 relapses with CNS involvement (1 in the 12 Gy group, incidence 0.4% and 6 in the 18Gy group, incidence 3%, p=0.04). Secondary leukemias occurred in 1 and 3 patients, respectively. The analysis for patients treated as randomized (12 Gy n=229, 18 Gy n=216) was comparable.

Our results demonstrate that there is no disadvantage for patients irradiated with a reduced CNS dose of 12 Gy regarding OS, EFS and rate of relapse. Due to intensification of chemotherapy, results in the AML-BFM studies improved considerably since study -87. Since the results of other international pediatric AML studies indicate that prophylactic CNS-RT does not improve outcome in the context of current chemotherapy regimen and because intensified chemotherapy elements with improved CNS efficiency such as high-dose cytarabine and liposomal daunorubicin have been introduced in our treatment schedule after study -87, we intend to replace CNS-RT by inclusion of triple intrathecal therapy in the next trial in order to reduce irradiation related long-term sequelae.

No relevant conflicts of interest to declare.