There is limited data on the optimal timing and predictive value of minimal residual disease (MRD) detection in newly diagnosed patients with acute promyelocytic leukemia (APL) treated with arsenic trioxide (ATO). Between January 1998 and December 2007, 120 patients with newly diagnosed APL were treated with a single agent ATO based regimen (Mathews V et al. Blood 2006). Sample for MRD assessment was taken at the end of induction, end of consolidation, once in 3 months during maintenance and subsequently once in 3 months for two years followed by once in 6 months for the next 3 years. Qualitative nested RT-PCR was performed prospectively for all patients using recommendations of BIOMED-1 Concerted Action (1999). Real-time quantitative RT-PCR (RQ-PCR) was done retrospectively on archived MRD samples 3–6 months prior to documented relapse, using ABI Prism 7000 sequence detection system based on the EAC protocol. PML-RARa copy number was normalized to the expression of the Abelson gene and normalized copy number (NCN) given as ratio of PML-RARa CN/ABL CN X 100. The NCN 30.001 was considered as positive RQ-PCR value. During this period there were 18 haematological relapses (HRel), 2 transient molecular relapses (MRel). Prospective RT-PCR and retrospective RQ-PCR from the relapsed cases prior to either HRel or MRel were analyzed and compared with 18 patients (included two cases who had transient MRel) in continuous complete remission (CCR).

Among 105 patients who achieved hematological remission, an RT-PCR at the end of induction was positive in 71/105 (68%) and at the end of consolidation in 2/105 (1.9%). One of these post consolidation positive cases subsequently had a HRel. Retrospective RQ-PCR analysis at the end of consolidation was positive in 6/18 (33%) cases who had a HRel. RQ-PCR was also positive at the same time point in 2/18 (11%) of the non relapsed controls who are in continuous complete remission (CCR) at 44 and 35 months respectively.

Eight of the 18 cases with HRel were RT-PCR positive in the 3–6 month period prior to HRel and progressed from MRel to HRel at a median of 3 months (range: 1–4 months). Ten cases had HRel without a prior documented MRel. A positive RQ-PCR result was detected in 2 of these 10 cases which were not picked up by RT-PCR. Among cases in CCR, there were 2 patients who had a transient RQ-PCR positivity but did not progress to a HRel with a follow up period of 30.3 and 39.5 months respectively. Among the 10 MRel cases an RQ-PCR was positive 3 months prior to detection of MRel by RT-PCR in 5 cases (PML-RARa NCN range: 0.02–3.9), all these subsequently had a HRel. This analysis suggests that with the use of ATO in induction and consolidation, RT-PCR for MRD at the end of consolidation would not predict relapse in the majority of cases. Also post consolidation, the current RT-PCR based MRD assessment would not predict the majority of HRel and if RQ-PCR was utilized at the same time intervals it would have a higher positive predictive value of 87.5% but would continue to have a low sensitivity (44%). More frequent MRD monitoring than the currently used time points may be required to improve the sensitivity.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author