Time Points, Dynamics and Clinical Impact of Minimal Residual Disease (MRD) Monitoring in AML with NPM1 mutation: A Study on 428 Patients

Introduction: Investigation of minimal residual disease (MRD) using NPM1 as a target has been proven to be of importance in AML. Guidelines for best schedules and implication on clinical use need to be defined.

Aims: To better define the clinical impact and to suggest strategies for MRD monitoring in AML with NPM1 mutation.

Patients and Methods: Between 2005 and 2015 we investigated 428 AML patients (pts) with NPM1 mutation at diagnosis and at a minimum of 2 follow-up time points. All pts had to achieve at least once a complete molecular remission (CMR) to be considered for this study. Sensitivity for MRD detection was at least 1:10,000. The median age of the cohort was 57 years (range: 18-85 yrs) and comprised of 198 males and 230 females. 3,039 samples (median number of samples per pts: 7, range: 2-35) were studied during course of disease. Molecular techniques applied included gene scan, sequencing and quantitative real-time PCR at diagnosis and quantitative real-time PCR during follow-up. Median time between 2 investigations was 2.8 months (mo; range: 0.3-71.0 mo). All pts were treated with standard protocols according to genotype and age. Allogeneic bone marrow or stem cell transplantation was performed in 136 pts (31.8%).

Results:NPM1 type A mutation was the most frequent mutation type (317/428, 74.1%), followed by type B and D (36/428, 8.4% and 23/428, 5.4%), respectively. 25 other NPM1 types occurred at frequencies between 0.2 and 3.7%, in total demonstrating the expected distribution of NPM1 mutation types in an adult AML cohort. Subgroups of these pts were analyzed for FLT3-ITD (n=421) and mutations in DNMT3A (n=236).

122/421 (29%) pts showed a FLT3-ITD. In 96/236 (41%) DNMT3A was mutated. Further in 33/235 (14%) both genes were mutated. 103/235 (44%) screened for all three genes had a sole NPM1 mutation.

All sole NPM1 mutated study pts achieved the CMR after a median of 4.1 mo (range: 1.0-8.6 mo). The presence of an additional DNMT3A mutation (CMR after a median of 4.4 mo, range 1.0-8.7) or a FLT3-ITD (CMR after a median of 2.7 mo, range 1.0-8.7) or both mutations (CMR after a median of 4.1 mo, range 1.1-7.9 mo) had no influence on time to achieve CMR.

After achievement of CMR an increase of NPM1 ratio was detected in 185/428 (43%) pts. The median time to loss of CMR was 5.1 mo (range: 0.4-88 mo). In more detail, 42/185 of these patients also had FLT3-ITD, 53/109 had DNMT3A mutations and 13/109 had mutations in both genes. Patients with a DNMT3A mutation showed more often loss of CMR (40/60, 67%), while FLT3-ITD and FLT3-ITD/DNMT3A mutated patients showed no significant influence on loss of CMR ratio (46% and 48%, respectively) maybe due to number of cases.

In 152/185 molecular relapses further follow up samples after loss of CMR were available. The median time between detected loss of CMR and the next follow-up sample was 2.0 mo. Due to treatment intervention 46/152 patients achieved a second CMR and 27/152 a decrease in NPM1 ratio. However, in 79/152 a further increase leading to clinical relapse was observed. The increase after loss of CMR was in median 13-fold between first and second sample after CMR was lost.

Importantly, keeping periods between two MRD samplings at an interval of 3 mo allowed the detection of nearly all cases of first relapse at the molecular level.

Addressing the sensitivity levels of the assays applied to bone marrow (BM) versus peripheral blood (pB) samples showed a 1.6 fold higher sensitivity for BM samples (median copies of reference gene, 14,628 vs 9,363). Due to the comparable sensitivities pB can be investigated until a first increase on the molecular level is detectable, followed by BM sampling for confirmation 4 weeks later.

Conclusions: 1) NPM1 has proven to be a good marker for MRD monitoring in AML. 2) Time to CMR is short with a median of 4.1 mo. 3) An increase of NPM1 in all cases is followed by relapse after a median of 5.1 mo, if no treatment intervention has been initiated before. 4) Time intervals for MRD should be no longer than 3 mo, pB can be used. 5) Transplantation should already be planned after first molecular increase is detected.