Patients and Methods. In order to find some markers, which reliably enable to predict clinical relapse in AML patients, we primarily focused on patients with the known fusion transcript (CBFB-MYH, AML1-ETO, or involving the MLL gene) and correlated this values with the WT1 for estimation of the true value of WT1 monitoring of disease behavior in a given patients. The value of different compartments for minimal residual disease (MDR) monitoring (PB, BM or CD34+ BM cells) was also analyzed. The study was prospective; in the case of MDR dynamics, the patients were actively called for earlier visit. This strategy could estimate the real time interval from MRD level increase to hematological relapse detection. In the interventional part of the study, the patients with already known MRD dynamics were treated at the time of molecular relapse.

Results and Discussion. In 67 AML patients and 3 healthy volunteers, 2184 BM or PB samples were examined, including 240 samples from CD34+ BM cells. Follow-up was 31–252 weeks (median: 88 w). The correlation between the fusion transcripts levels in BM and PB was excellent (r=0.9676). The correlation between WT1 PB and BM levels was far less satisfactory. Since the WT1 values were frequently >0 even if the level of fusion transcript =0, we wanted to find some “normal” value for WT1. Using the ROC curves, however, we were not able to find any WT1 level being a confidential marker of molecular remission in either compartment (PB, BM or CD34+). Molecular relapse was defined as a reappearance of the fusion transcript detection or its 10-fold increase, repeatedly detected. The time from molecular to hematological relapse was 8–79 days (median: 25 d). In the cases of subsequent development of hematological relapses, the levels of fusion transcript in CD34+ BM cells were one order of magnitude higher than in the BM or PB, even in the case of CD34− blasts. Eight patients were treated for 13 molecular relapses with following results: chemotherapy, CR=2, PR=2; gemtuzumab ozogamicin, CR=3, PR=1, NR=3; IL-2±DLI, CR=2 (PR was defined as a decrease in fusion transcript level at least 10-fold). Patients with CD33− blast at diagnosis did not respond to gemtuzumab ozogamicin. Non-responsiveness to one treatment option did not mean non-responsiveness to another treatment.

Conclusion: Fusion transcript monitoring enables reliable detection of molecular relapse in AML and high values in CD34+ BM cells signalize imminent hematological relapse (even in the case of CD34− blasts). PB is a suitable compartment for frequent monitoring. However, in some cases, relapse are fulminate, hardly allowing any intervention. WT1 does not seem to be a reliable marker for exact molecular relapse detection. AML at the stage of molecular relapse behaves similarly to AML at the frank hematological relapse: there are CRs, PRs, or NRs when using chemotherapy or gemtuzumab ozogamicin. AML with CD33− blasts at diagnosis does not seem to respond to gemtuzumab ozogamicin at the stage of molecular relapse. Success of AML therapy in the future seems to be dependant on efficient targeting the leukemia stem cell.

Disclosure: No relevant conflicts of interest to declare.

Author notes

*

Corresponding author

Sign in via your Institution