The Incidence of Central Nervous System Involvement in AML: A SAL-Tor Vergata Analysis

Acute myeloid leukemia (AML) involves infrequently the central nervous system (CNS). However, most studies examined paediatric patients. Thus little is known regarding CNS-involvement in adult AML. Therefore, we analyzed the data of patients with AML treated in SAL study group and Tor Vergata study group for CNS involvement.

In this retrospective analysis in adult AML patients included in the prospective AML96- (NCT00180115), AML2003- (NCT00180102), AML60+ trials (NCT00180167) of the SAL study group and patients that were admitted to the University Hospital of Rome Tor Vergata from 2005 to 2011, 3863 patients were evaluated for CNS involvement of AML. The median age of the patients was 49 years (range, 18 –74 years). A total of 30 patients had CNS involvement at time of initial presentation and 30 patients had CNS involvement at first, second or third relapse. CNS involvement was proven depending on morphology and/ or flow cytometry of the Cerebrospinal fluid (CSF). Clinical variables of statistical significance were compared using the χ²-test for dichotomized variables and the Mann-Whitney U-test for continuous variables. The method of Kaplan Meier was used to estimate OS and event-free survival (EFS). Survival distributions were compared using the log-rank test.

21 of 1862 patients (1%) in the AML96-trial, 26 patients of 1182 (2.2%) in the AML2003 trial, 3 of 486 patients (0.6%) in the AML60+ trial, and 10 of 270 patients (3.7%) in the University of Rome Tor Vergata studies had CNS involvement resulting to a total of 60 patients with a combined overall incidence of 1.6%. 30 patients had CNS involvement at the initial presentation and 30 patients at relapse of AML. The majority of patients (n=53, 88%) had de novo AML. Data from the Tor Vergata study group need further analysis therefore in this abstract patients from SAL study group were included. Extramedullary AML other than CNS involvement was observed in 21 patients (42 %) as compared to 246 (7%) in patients without CNS involvement, p < 0.001. Complex aberrant karyotype was not significantly higher in patients with CNS involvement as compared to those without, whereas patients with CNS involvement had a higher tendency for trisomy of chromosome 8 (n=8, 16%) in comparison with patients without CNS involvement (n= 277, 8%), p=0.03. AML FAB M5 occurred more frequently in patients with CNS involvement (n= 15, 30%) as compared to those without (n=459, 13%), p=0.01. Patients with CNS involvement at diagnosis had significantly higher lactate dehydrogenase (LDH) levels (1318 vs. 665 IE/l, p< 0.001) and higher white blood cell count (WBC) (64 vs. 35 Gpt/l, p<0.001). Only 7 of 30 patients (23%) with CNS involvement at initial diagnosis developed relapse and none of them had CNS involvement at the time of relapse, whereas almost all patients with CNS involvement at relapse developed again systemic and CNS-relapse. On the other hand, there was no significant difference in the number of patients who achieved complete remission between these two groups. Comparing patients with CNS involvement with patients without CNS involvement at the initial diagnosis no significant difference in overall survival at 2 years (40% vs. 41%) and event free survival at 2 years (20% vs. 26%) was observed.

CNS involvement in AML is a rare entity and it is accompanied with higher incidence of other extramedullary AML. This represents the largest analysis of AML patients with CNS involvement studied in the literature. Factors associated with an increased risk of CNS involvment include: trisomy of chromosome 8, FAB M5, higher LDH levels and higher WBC count at diagnosis. Patients with CNS disease at initial diagnosis of AML have similar survival outcomes as compared to those without CNS disease if treated by intrathecal therapy, whereas extramedullary relapse within the CNS is associated with poor outcome.

No relevant conflicts of interest to declare.