ORLANDO – Minimal residual disease (MRD) measured before and after allogeneic hematopoietic stem cell transplantation (HCT) is a powerful predictor of survival in children with acute lymphoblastic leukemia (ALL), according to a review of hundreds of cases from around the world.

The findings could have implications for using minimal residual disease measures to guide posttransplant interventions, Michael A. Pulsipher, MD, reported at the combined annual meetings of the Center for International Blood & Marrow Transplant Research and the American Society for Blood and Marrow Transplantation.

“MRD pretransplant was a very powerful predictor of outcomes. MRD posttransplant highlights individual patients at risk,” Dr. Pulsipher said. Results comparing reverse transcriptase–polymerase chain reaction with flow cytometry require validation by direct comparison in the same patient cohort, but “the new risk scores ... very nicely predict outcomes both pre- and post-transplant and can guide study planning and patient counseling.”

A total of 2,960 bone marrow MRD measurements were performed in the 747 patients included in the study. MRD was assessed prior to HCT and on or near days 30, 60, 100, 180, and 365 and beyond after HCT.

Patients were grouped for analysis according to MRD level: Group 1 had no detectable MRD, group 2 had low detectable MRD levels (less than 10E-4, or 0.01% by flow cytometry), and group 3 had high detectable MRD levels (10E-4 or higher). A second analysis compared findings in those tested by flow cytometry and those tested by real-time quantitative PCR (RQ-PCR), said Dr. Pulsipher of Children’s Hospital Los Angeles.

In 648 patients with pre-HCT MRD measurements available, the 4-year probability of event-free survival was 62%, 67%, and 37% for groups 1, 2, and 3, respectively. Group 3 – the high MRD level group – had 2.47 times the increased hazard ratio for relapse and 1.67 times the increased risk of treatment-related mortality, Dr. Pulsipher said, adding that pre-HCT MRD and remission status both significantly influenced survival, while age, sex, relapse site, cytogenetics, donor type, and stem cell source did not influence outcome.

Post-HCT MRD values were analyzed as time-dependent covariates.

“As time went by more and more, any detectable level of MRD led to a very poor prognosis, whereas patients arriving at day 365 with no detectable MRD had exceptional prognosis with survival approaching 90%,” he said.

Specifically, the 4-year probability of event-free survival for groups 1, 2, and 3, respectively, were 59%, 65%, and 43% at day 30; 64%, 47%, and 36% at day 60; 65%, 69%, and 44% at day 90; 79%, 40%, and 12% at day 180; and 87%, 36%, and 25% at day 365.

Of note, a very significant interaction was seen between acute graft-versus-host disease (GVHD) and MRD, Dr. Pulsipher said, explaining that patients who were MRD positive and had developed GVHD had a significant decrease in the cumulative incidence of relapse, compared with those with no GVHD.

“This translated into improved event-free survival with patients post transplant, who were MRD-positive [and] developing GVHD, still having a reasonable chance of survival, whereas patients post transplant who had MRD measured who did not develop GVHD had a very poor chance of survival,” he added.

Additionally, based on detailed multivariate analysis including a number of clinical factors, risk predictive scores were developed for event-free survival risk at 18 months or cumulative incidence of relapse at 18 months. Multiple scores were developed for each, but, as an example of factors that had an important effect on outcomes, patients with very early pretransplant relapse (those who went into remission but relapsed within 18 months) or with greater than 2nd relapse had a high risk for poor event-free survival. Mismatched donors and unrelated cord-blood stem cell transplant recipients also had high risk, he said, noting that, “of course, MRD had a significant effect” and was the most important factor prior to transplant.

These patients, who had a 4-point or greater risk score, were the poorest risk group, with survival that was less than 50%, as opposed to better risk groups that exceeded 90%, he said.

“A score of greater than 5 could identify 80% of patients who were going to relapse after transplant, and of course, event-free and overall survival in those patients were very poor,” he added.

As time went by, the early risk of GVHD diminished somewhat, as did the risk of mismatched donors.

“Most of the risk was associated with any MRD detection,” he said.

Flow cytometry and RQ-PCR levels of at least 10-4 were highly predictive of relapse at all pre- and post-HCT time points; however, RQ-PCR values between 10-4 and 10-3, in cases where adequate numbers were available for comparison, better predicted relapse as compared with flow cytometry results.

For example, before HCT, hazard ratios were 1.26 and 2.41 with flow cytometry vs. RQ-PCR. At day 30, the hazard ratios were 1.33 and 2.53, and at day 365, they were 3.54 and 31.84, Dr. Pulsipher reported.

The findings provide important information for improving outcomes in children with high-risk ALL undergoing HCT, he said.

“Older prognostic models for relapsed and refractory high-risk ALL have focused on timing and location of relapse, as well as disease phenotype. But it is clear that, in order to treat children with very high risk ALL with transplantation, MRD has become the most important thing to look at in the pretreatment setting. The challenges that we face in assessing MRD, however, have been hampered by the fact that we have differing MRD measurements,” he said, noting that RQ-PCR is often used in Europe, while flow cytometry is more often used in the United States. As such, direct comparisons are lacking, as are T-cell and posttransplant data.

The current study represents a “tremendous effort” by international collaborators to address these shortcoming, he said.

“This is a great opportunity, as our goal, of course, is to avoid futility in transplantation, but, more importantly, to find opportunities to identify groups for which we can improve our outcomes,” he added.

Patients included in the study were treated in Europe, North America, and Australia and were transplanted during Sept. 1999-May 2016. Most were in first or second remission, and most (586) had pre-B ALL. A notable 145 had T-cell ALL – “more than ever has been analyzed previously” – and 16 had B-lineage or biphenotypic ALL. About half were under age 10 years, 62% were boys, and stem cell sources were typical, although 20% received a cord blood transplant.

Dr. Pulsipher reported serving as an advisor and/or consultant for Chimerix, Novartis, Jazz Pharmaceutical, and receiving housing support from Medac Pharma for an educational meeting.

sworcester@frontlinemedcom.com

Pre- and post-HCT MRD levels predict ALL survival