Minimal Residual Disease (MRD) and Risk of Relapse in Acute Promyelocytic Leukemia (APL): Insights from the North American Intergroup Phase III Trial C9710.

MRD results during and following treatment (Rx) of APL have become an important predictor of clinical outcome and are beginning to be used to guide Rx for this highly curable leukemia, but important questions remain. Some studies suggest that low/intermediate (LI)-risk patients (pts), as defined by the Sanz criteria, who achieve a negative PCR result after intensive consolidation Rx have a very low risk of relapse and may not require maintenance Rx, while those with MRD might benefit from hematopoietic stem cell transplantation. However, these data are controversial. We studied whether a single quantitative MRD test obtained immediately after completion of C9710 consolidation Rx (with all-trans retinoic acid, daunorubicin, with or without arsenic trioxide) predicted relapse and could be used to guide further Rx. Real-time quantitative RT-PCR( RQ-PCR) was used to measure PML-RARα in bone marrow (BM) and blood (B) of 43 pts treated on C9710. All 43 pts were in first complete remission after completion of consolidation Rx when MRD was measured, but all of them later relapsed during, or after completion of maintenance Rx. By Sanz criteria, 14/43 were high risk and 29/43 LI-risk at diagnosis. 29 had S-isoform, 12 had L-isoform, and 2 had V-isoform. Median time from MRD assessment to relapse was 253 days (range: 44–1231). RQ-PCR results were expressed as a normalized quotient (NQ) of PML-RARα/GAPDH or ABL control. The sensitivity of the assay was 1 in 10^4–5. A NQ of >10⁻⁵ was found in the prior intergroup study, INT0129, to be strongly associated with risk of relapse and was set as the threshold for considering a result “positive”. MRD with NQ >10⁻⁵ was detected in only 4 of 43 (9%) pts after completion of consolidation Rx. 8 others had detectable PML-RARα transcript but had NQs < 10⁻⁵. The remaining 31 pts had no evidence of MRD. There was excellent concordance between MRD measurements in 26 paired B and BM NQ values and a strong correlation between NQ values in B and BM using either GAPDH or ABL controls (p = 0.0001 and p <0.0001, respectively). To determine whether a more sensitive assay might identify a higher percentage of pts with MRD, we also evaluated the post-consolidation samples in 19 of these pts using a qualitative nested RT-PCR assay with a sensitivity of detection of 1 in 10⁶. We found 13 of 19 (68%) pts had detectable MRD. In conclusion, RQ-PCR analysis of MRD using a threshold NQ of >10⁻⁵ at a single post-consolidation Rx timepoint did not predict relapse of APL in pts on C9710. Setting a lower NQ threshold for the RQ-PCR assay or using a more sensitive, but qualitative PCR assay improved the detection rate; however, neither approach identified all pts destined to relapse, nor has the prognostic value of a positive result using the more sensitive assay been determined. Based on these results, we caution against the use of a single MRD test at the end of consolidation Rx to determine further Rx. Ongoing analysis of serial samples obtained during and following completion of Rx might provide further insights into optimal timing and frequency of MRD measurements needed to identify APL pts at high risk of relapse.