Abstract 1407

Background:

Aberrant expression of MLL5, BAALC, ID1, and WT1 genes is frequently associated with inferior outcome in cytogenetically normal acute myeloid leukemia patients (Damm et al. Blood 2011; 117(17):4561–8). The expression levels of these genes vary in patients with acute promyelocytic leukemia (APL), but the clinical significance of these findings remains unclear.

Objective:

(1) to determine if the gene expression levels of MLL5, BAALC, ID1, and WT1 are associated with clinical outcome of APL patients treated with ATRA and anthracycline-based chemotherapy, (2) to generate an integrative score (IS) based on these potential prognostic factors and clinical parameters and (3) to use this score for outcome prediction in APL.

Design and Methods:

One hundred and fifty APL patients (age, 15–73y) from seven different Brazilian institutions and treated according to the IC-APL protocol were included. The treatment schedule was identical to the PETHEMA-LPA 2005, except for the replacement of idarubicin by daunorubicin; ATRA treatment was initiated immediately in all cases in which the diagnosis of APL was suspected based on morphology. Gene expression profile was analyzed by Real-time PCR. Integer weights for the IS were derived from Cox proportional hazard model, using overall survival (OS) as outcome parameter. Hazard ratios (HR) for OS were calculated for each variable separately (Table 1). Variables with P<0.05 in univariate analyses were included in the model. Variables considered for the model inclusion consisted in 2 clinical (WBC counts, albumin levels) and 5 molecular markers (FLT3-ITD status and gene expression levels of MLL5, BAALC, ID1, and WT1). Other candidates, such as age, platelet count, gender, ECOG performance status, PML breakpoint and FAB subtype were not significant and not included in the score. The HR were converted to integer weights according to the following: variables with HR < 1 were excluded from analyses; variables with HR 3 1 and < 1.5 were assigned a weight of 1; variables with HR 3 1.5 and < 2.5 were assigned a weight of 2; variables with HR 3 2.5 were assigned a weight of 3. The final score was the sum of these integer weights. Based on maximally selected rank statistics, the scores were grouped into 3 risk-groups: 0–5 (low-IS), 6–9 (intermediate-IS), and > 9 (high-IS).

Results:

The integrative weights of variables analyzed are summarized in Table 1. The IS was modeled in 137 patients (median score: 6; range, 1–17). According to PETHEMA-GIMEMA relapse risk criteria, 22%, 23% and 70% of patients assigned in the low-IS (n=46), intermediate-IS (n=57) and high-IS (n=34) groups were deemed high-risk of relapse (P<0.001). Overall, 118 (86%) patients achieved CR; the remaining 19 patients (14%) experienced early death due to hemorrhage (n=12), therapy-related infection (n=6) and differentiation syndrome (n=1). Induction mortality was significantly higher in the high-IS group (low: 2%; intermediate: 15%; high: 26%) (P=0.001). CR was achieved in the low-, intermediate-, and high-IS group in 98%, 84%, and 73% of the patients, respectively (P=0.007). With a follow-up of 24 months among survivors, patients assigned in the high-IS group had a lower 2-y OS rate (63%) compared with those in the intermediate- (80%) and low-IS groups (97%; P<0.001). Eight relapses were recorded. The IS was not predictive of relapses (P=0.351).

Conclusions:

Our results suggest that MLL5, BAALC, ID1, and WT1 expression levels are associated with clinical outcome and that the IS may become a useful tool for outcome prediction in APL.

Table 1.

Variables used for the determination of the Integrated Score (IS).

Log-rank test, p-valueHR (95%CI); p-valueInteger weight**
WBC counts (<=10×109/L vs >10×109/L) p = 0.027 2.32 (1.1-5.02); 0.032 2 
Albumin levels (<= 3.5 vs >3.5) p = 0.003 0.24 (0.09-0.65); 0.005 3 
FLT3-ITD status (FLT3-ITDnegativevs FLT3-ITD positive) p = 0.013 2.72 (1.19-6.24); 0.017 3 
MLL5 gene expression (> 25th percentile*) p = 0.019 0.42 (0.14-0.72); 0.006 2 
ID1 gene expression (>75th percentile*) p = 0.005 3.22 (1.34-7.76); 0.009 3 
BAALC gene expression (>50th percentile*) p = 0.019 3.13 (1.13-8.62); 0.027 3 
WT1 gene expression (>25th percentile*) p = 0.033 6.91 (0.92-51.6); 0.05 3 
Log-rank test, p-valueHR (95%CI); p-valueInteger weight**
WBC counts (<=10×109/L vs >10×109/L) p = 0.027 2.32 (1.1-5.02); 0.032 2 
Albumin levels (<= 3.5 vs >3.5) p = 0.003 0.24 (0.09-0.65); 0.005 3 
FLT3-ITD status (FLT3-ITDnegativevs FLT3-ITD positive) p = 0.013 2.72 (1.19-6.24); 0.017 3 
MLL5 gene expression (> 25th percentile*) p = 0.019 0.42 (0.14-0.72); 0.006 2 
ID1 gene expression (>75th percentile*) p = 0.005 3.22 (1.34-7.76); 0.009 3 
BAALC gene expression (>50th percentile*) p = 0.019 3.13 (1.13-8.62); 0.027 3 
WT1 gene expression (>25th percentile*) p = 0.033 6.91 (0.92-51.6); 0.05 3 
*

Calculated using the values obtained in whole cohort of patients;

**

Integer weights for the risk score were derived from Cox proportional hazard model, using overall survival (OS) as outcome parameter.

View Large
Disclosures:

Lo-Coco:Cephalon: Speakers Bureau; Boehringer Ingelheim: Membership on an entity's Board of Directors or advisory committees. Löwenberg:Skyline Diagnostics: Membership on an entity's Board of Directors or advisory committees.

Topics:
acute promyelocytic leukemia, anthracycline antibiotics, chemotherapy regimen, tretinoin, wt1 gene, impedance threshold device, ms-like tyrosine kinase 3, albumins, biological markers, daunorubicin

This icon denotes a clinically relevant abstract

Author notes

*

Asterisk with author names denotes non-ASH members.

© 2012 by The American Society of Hematology
2012
Sign in via your Institution