MRD Can't Predict Response in Early Days of Intensive AML Therapy, Study Says

Molecular minimal residual disease (MRD) testing during the first 4 days of induction therapy does not differentiate responders and nonresponders and should not be used in predicting clinical response for patients with acute myeloid leukemia (AML), according to the results from a recent study.

Over recent years, MRD has become an essential prognostic tool for AML, with highly predictive capabilities. Patients with detectable MRD are usually nonresponders with poorer progression free and overall survival while patients with undetectable MRD generally have clinical response and much longer survival rates. The prognostic value of MRD has been seen at key treatment timepoints such as after 1 to 2 cycles of induction therapy; however, its use has not yet been seen during initial induction therapy. In a recent study, investigators determined whether MRD quantification can predict clinical response during the first 4 days of intensive cytotoxic therapy.

Blood was collected daily from days 1 through 5 from 45 adult patients with AML, 10 of whom had relapsed or refractory disease. Two different molecular techniques were used for AML MRD quantification: WT1 expression by quantitative polymerase chain reaction (qPCR) and targeted DNA sequencing for common myelodysplastic syndrome/AML variants. Of the 45 patients, 19 of them achieved a complete remission (CR) after cytotoxic treatment (42%).

Of the 45 patients in this study, 34 of them had enough RNA for a qPCR analysis. On day 1, 91% of the patients had overexpressed WT1. By day 4, the percentage dropped to 68%. Of the 13 patients who achieved a CR, 7 of the patients had a 4-fold reduction in WT1 expression. Of the 21 patients who were nonresponders, 7 of them also had at least a 4-fold reduction in WT1 expression. Based on the scattered results, investigators concluded that WT1 expression between day 1 and 4 appeared to be uninformative of clinical response.

All 45 patients had enough DNA for sequencing and a total of 163 variants were found in 43 patients (2 patients had no variants for disease tracking). The most frequent mutations were seen at DNMT3A and ASXL1. No differences in the number of coding variants at baseline in CR versus nonresponders were observed. The only 2 mutations that had predictive significance were TP53 and NPM1, both already used with the current prognostic risk classification.

All 9 patients with TP53 mutations did not achieve remission and all 5 patients with NPM1 mutations achieved CR.

Early DNA sequencing also produced varied results in predicting post-remission relapse. In 1 patient, there were no changes in wild type to mutated sequence ratios during the first 4 days of therapy and mutation levels remained negligible. For the other patient, KRAS decreased during the first 4 days of therapy and DNMT3A remained stable. Once the patient was in remission, both mutations became undetectable and reappeared after relapse, with the addition of a second KRAS mutation.

Aside from the clinical utility of monitoring NPM1 and TP53 variants, investigators concluded that molecular testing during the first few days of induction therapy does not appear to be predictive of response in patients with AML. MRD should instead be used during follow-up evaluations rather than early AML treatment.

Reference

Wong HY, Sung AD, Lindblad KE, et al. Molecular measurable residual disease testing of blood during AML cytotoxic therapy for early prediction of clinical response. Front Oncol. 2019;8:669. doi: 10.3389/fonc.2018.00669.