Abstract 567FN2

The successful incorporation of risk-associated factors in contemporary treatment regimens has remarkably improved the treatment outcome of pediatric ALL. Currently, the most commonly used prognostic factors such as the age and white blood cell (WBC) count which are tightly correlated with certain cytogenetic/molecular subtypes, are mainly related to leukemic factors. Beyond that, host pharmacogenetic factors, i.e., how the host genetic makeup influences the metabolism of chemotherapeutic agents, are expected to have additional but smaller impact on treatment outcome.

Modern, successful multi-agent chemotherapy incorporates more than 8 different chemotherapeutic agents to treat pediatric ALL. Variability in individual patients in handling these drugs arises from their pharmacogenetic factors. However, such pharmacogenetic variants are rarely evaluated in existing prognostication schema. We studied both tumor- and host pharmacogenetic factors to determine the impact of germline variations on the treatment outcome in the Malaysia-Singapore (Ma-Spore) ALL 2003 study essentially guided by the MRD-risk stratification.

A total of 463 patients were included, with a median follow-up of 3.8 years (0-8.2). Genotyping was carried out for 20 germline polymorphisms in 11 genes (i.e., GSTM1, GSTT1, GSTP1, NQO1, MTHFR, MTHFD1, SLC19A1, ABCB1, TYMS, CCR5, and IL15). The genotypic influence on the event-free survival (EFS) probability was estimated by Cox regression, adjusted for patients' characteristics (including the ethnicity, sex, age, lineage, and initial WBC count) and cytogenetic subtypes. An event was defined as any of: induction failure (N=20), relapse (N=32, any site) or death (N=25, any cause).

Only cytogenetic subtypes showed significant impact on the EFS (P=0.015). After adjusting for host-related and leukemic factors, ABCB1 3435C>T (rs1045642), CCR5 246A>G (rs1799987), and IL15 67276493G>C (rs17015014) were also found to significantly and independently affect the survival probabilities. Specifically, ABCB1 3435T/T and CCR5 246G/A predicted lower EFS probabilities compared to 3435C/C and 246G/G, respectively (P=0.012, HR=2.496, 95%CI=1.223-5.092, Fig.1a and P=0.033, HR=1.820, 95%CI=1.051-3.153, Fig.1b), while IL15 67276493G/C was associated with a higher survival probability compared to 67276493G/G (P=0.030, HR=0.545, 95%CI=0.315-0.942, Fig.1c). In addition, the risk increased with increasing number of risk genotypes (i.e., ABCB1 3435T/T, CCR5 246G/A, and IL15 67276493G/G, which were associated with the worst survival respectively; Fig.1d), implicating a cumulative adverse effect (P<0.001).

Further analyses revealed that compared to the reference genotypes respectively, ABCB1 3435T/T particularly correlated with the highest risk of relapse (P=0.011); CCR5 246G/A was associated with a higher risk of failure of induction therapy (including induction failure and death during induction; P=0.022); and IL15 67276493G/C conferred a protection from septic death (P=0.041). We subsequently investigated whether these polymorphisms could still influence the outcome in cytogenetic groups with known prognosis. ABCB1 3435C/T and T/T in combination were predictive of a poorer survival in 184 patients with favorable subtypes including t(12;21), t(1;19), and hyperdiploidy (P=0.027, HR=5.271, 95%CI=1.211-22.944); CCR5 246G/A appeared to retain its significance in 29 patients with unfavorable subtypes including t(11q23)/MLL rearrangements and t(9;22) (P=0.052, HR=7.708, 95%CI=0.983-60.471); and in 229 patients without common chromosomal abnormalities, IL15 67276493C/C was associated with the best survival (P=0.027, HR=0.291, 95%CI=0.098–0.868).

Our results suggest that risk assignment for contemporary ALL treatment can be refined by incorporating the host pharmacogenetic profile into existing strategy of risk stratification that uses mainly leukemic factors.
Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

This icon denotes a clinically relevant abstract

Sign in via your Institution