The prognostic significance of a molecular risk score in APL patients

In 2015, Anna Hecht from the University Hospital Mannheim, University of Heidelberg, et al. reported  findings study which showed that high gene expression of Brain and Acute Leukemia, Cytoplasmic gene (BAALC), Ets Related Gene (ERG) and low and high expression of Wilms Tumor 1 (WT1) gene were negative prognostic risk factors for high relapse in patients with Acute Promyelocytic Leukemia (APL). Additionally, integration of these three genes to create a molecular risk score revealed differences in survival and relapse rates in APL patients with low and high scores.¹

In an article published in Clinical Lymphoma, Myeloma & Leukemia on 14^th August 2017, Hecht et al. aimed to validate the prognostic value of this integrative molecular risk score formed by the combination of BAALC, ERG and WT1 genes in independent cohort of APL patients.²

In this validation study, two cohorts of patients were analyzed. The first cohort (original cohort) included 79 newly diagnosed APL patients (median age = 46 years) who were treated in two consecutive German AML Cooperative Group (AMLCG) trials. The other cohort (validation cohort) consisting of 76 patients (median age = 50 years) who were treated in the AIDA2000 SAL study. An integrative risk score combining the expression levels of BAALC, ERG and WT1 were calculated for patients in the original and validation cohorts. Patients were scored either 0 (absence of BAALC, ERG and WT1), 1 (presence of one risk factor), 2 (presence of two risk factors) or 3 (BAALC^high, ERG^high and WT1^{low or high}) points.

The key results of the study were:

• In both the original and the validation cohorts, patients with a risk score of 0 points had very good prognosis in terms of survival and relapse
• Overall Survival (OS) and Relapse Free Survival (RFS) in patients with integrative risk score of 3 points in the original and the validation cohorts were very poor compared to patients with 0 points.
• Integrative risk score was a significant predictor for OS, RFS, and Cumulative Incidence of Relapse (CIR)

In summary, the findings of the study showed that an integrative risk score combining BAALC, WT1 and ERG was able to “reliably distinguish between patients at very low and very high risk of relapse and inferior survival” thus indicating that the presence of “concurrent molecular risk factor” is a very “strong prognostic factor” in APL patients.

Overall, Hecth et al. concluded by stating that the “prognostic impact of an integrative risk score combining BAALC, ERG and WT1 expression on APL patient outcome has now been validated in two independent APL cohorts in an ATRA plus chemotherapy setting”. They further proposed that the prognostic impact of integrative risk score should be investigated in the All Trans Retinoic Acid (ATRA) plus arsenic trioxide setting.

Abstract

Introduction

Although treatment of acute promyelocytic leukemia (APL) has evolved dramatically during the past decades, especially with the introduction of all-trans retinoic acid, risk stratification remains an important issue. To date, relapse risk can be predicted by leukocyte and platelet counts only. In the present report, we present a validation study on 3 candidate genes and a newly developed molecular risk score for APL in 2 independent patient cohorts.

Patients and Methods

An integrative risk score combining the expression levels of BAALC, ERG, and WT1 was calculated for 79 de novo APL patients from the original cohort and 76 de novo APL patients from a validation cohort. Gene expression analysis was executed the same for both cohorts, and the results regarding the effect on patient outcomes were compared.

Results

The expression levels of BAALC, ERG, and WT1 were similar in both cohorts compared with the healthy controls. The relapse and survival rates were not different between the low- and high-risk patients according to the Sanz score. However, application of the molecular risk score on the validation cohort distinctly discriminated patients according to their risk of relapse and death just as in the original APL cohort, although single gene analyses could not reproduce the negative prognostic impact.

Conclusion

The analysis clearly validated the prognostic effect of the integrative risk score on the outcome in APL patients. The value was further empowered because the single gene analyses did not show similar results. Whether the integrative risk score retains its prognostic power in the chemotherapy-free setting should be investigated further.