Patients with MRD- CR at transplant had better outcomes than patients with MRD + CR or patients with AD at transplant. In patients where the transplant indication was triggered by relapse or persistent disease post or during CT, outcome after alloHSCT was best when relapse was detected on a molecular level after completion of CT, demonstrating the importance of continuous MRD-monitoring via high sensitive RT-PCR. Patients with molecular or hematologic relapse after completing CT responded well to salvage CT in contrast to patients relapsing while still on CT. Patients who achieved a 2nd MRD- CR pre alloHSCT had a very good chance for long-term remission after alloHSCT. Patients with early hematologic relapse during CT had poor response to salvage CT and did not achieve MRD- CR before transplant. Fludarabine, amsacrine and cytarabine (FLAMSA) based sequential conditioning did not improve outcome in patients with MRD + CR or AD at transplant. Late relapses or relapses detected on a molecular level after alloHSCT is associated with a good chance of still achieving a long-term disease control after relapse, which explains the divergence between OS and PFS in the analyses shown here. Effective relapse strategies are available here, with particular focus on the occurrence of mild or moderate graft versus host disease (GvHD) probably reflecting an immunological graft versus leukemia (GvL) effect.
Dillon et al. showed in their study with 107 patients, that factors like high levels of MRD (above 200 copies per 105 ABL in the PB or 1000 copies in the BM) and the presence of a FLT3-ITD mutation at diagnosis were associated with adverse outcome.16 Kayser et al identified the same threshold of 1000 copies per 105 ABL in the bone marrow in their study of 39 patients with NPM1-mutated AML.15 Finally, Bill et al. reported a significant difference in outcome according to the molecular MRD status in 51 patients with a lower threshold equivalent to 10 copies per 105 ABL according to the technical characteristics of the digital droplet PCR platform employed.24 In contrast to the aforementioned studies, we could not define a stringent MRD cut-off value, because as part of clinical routine we had results from different laboratories, who modified their high sensitive RT-PCR methods over the years.
Nevertheless, unlike these studies we were able to define two prognostic subgroups in the 2nd line group by time of relapse and response to S-CT. In our study, 10 of 15 patients (67%) with hematological or molecular relapse responded to salvage CT with 9 of 10 patients (90%) responding when relapse occurred after completing front-line conventional chemotherapy. Six of these patients (60%) achieved a 2nd MRD- CR and had an excellent outcome after alloHSCT. These results confirm data, reported by Dillon et al, where response to salvage therapy was associated with favorable survival after alloHSCT.16 In contrast, only 1 of 5 patients (20%) relapsing during front-line therapy responded to salvage therapy, showing an unmet medical need for clinical studies with new drugs in these patients.25 Here, the direct use of hypomethylating agents combined with Venetoclax may also be an effective alternative, especially because NPM1mut AML (especially in combination with an IDH2-Mutation) reflects and specifically sensitive subgroup for this therapy.26 Tiong et al. also showed the clinical impact of NPM1mut molecular persistence after front line conventional chemotherapy.27 Patients with NPM1mut MRD positivity after completing front-line intensive chemotherapy had a variable course, with a substantial fraction (42%) remaining relapse-free at 1 year and 30% achieving MRD negativity. Preemptive salvage therapy prior to morphologic relapse, resulted in a significantly prolonged relapse free survival compared to those not receiving preemptive therapy.27 Since in our cohort molecular relapse was defined by a dynamic MRD increase in two consecutive samples detected by RT-PCR, we did not assume the option of spontaneous MRD clearance, but impending hematologic relapse in these patients. The selection of conditioning regimens particularly for patients with AD or MRD positive CR remains controversial, and studies provide conflicting results.16,19,22,28−30 Some studies suggest that myeloablative or sequential conditioning regimens should be preferred in patients who are MRD positive at transplant. As our data is retrospective and non-randomized, we could not see any additional benefit of a FLAMSA based sequential or myeloablative conditioning in these patients. Further improvement of current conditioning regimes may include the use of new agents like Venetoclax during conditioning rather than increasing dose intensity.31
In patients with late or molecular relapses after alloHSCT effective relapse treatments were available and accompanying occurrence of controlled GvHD was associated with long-term disease control. Concomitantly, Dillon et al. observed a strong association between the use of T-depletion and adverse outcome of alloHSCT in patients with NPM1mut AML.16 These observations argue for less intensive post-transplant immunosuppression and immune interventions like preemptive donor lymphocyte infusions in the absence of GvHD in high-risk patients with MRD + CR or even AD at transplantation.
To decrease the rate of failure of conventional chemotherapy in NPM1mut AML the identification of additional risk factors and the impact of common concomitant mutations like DNMT3A mutations is crucial.32 Since the data from our patients regarding an NGS panel that goes beyond cytogenetics and detection of NPM1 and FLT3 mutations, are incomplete, we did not want to discuss this further in our cohort.
Controversially to the ELN 2017 classification, the presence of unfavorable genetic alterations or mutations highly specific for secondary AML in the good risk group, identifies patients with a higher risk for relapse. These individuals could be seen as prime candidates for alloHSCT in 1st CR in the future.8,33 Therefore, detailed complementary next generation sequencing analysis in addition to conventional standard PCR and cytogenetic analysis should always be performed. In addition, the recently published ELN 2022 guidelines must also be taken into account in the future, in which an additional FLT3-ITD already represents an intermediate risk, regardless of the ratio, which is reasoned here on the basis of the technical aspects of the determination of the ratio.2
In conclusion, according to our data outcome post alloHSCT in NPM1mut AML depends on pre-transplant remission and molecular MRD status. Later relapses respond well to salvage CT and patients who achieve a 2nd MRD- CR before transplant have excellent long term survival. Furthermore, survival is superior if relapse is diagnosed as molecular relapse, which underlines the importance of intensive MRD monitoring. Patients who experience hematologic relapse during therapy represent a clinical challenge as they rarely achieve remission with conventional chemotherapy. Studies evaluating therapy augmentations such as the additional application of Venetoclax during salvage therapy or conditioning may be considered in the future.