Study Revises Definition of Induction Failure in Pediatric ALL

The use of minimal residual disease provided a more objective measure of induction failure in patients with pediatric acute lymphoblastic leukemia than did morphology.

The use of minimal residual disease (MRD) provided a more objective measure of induction failure in patients with pediatric acute lymphoblastic leukemia (ALL) than did morphology, according to the results of a study published recently in the Journal of Clinical Oncology.

In fact, combining those patients who failed to achieve morphologic remission with those patients with high MRD identified almost 4% of trial participants, suggesting “that induction failure is more common than previously reported,” according to David O’Connor, MD, of Great Ormond Street Hospital for Children NHS Foundation Trust in London, and colleagues.

“MRD has now been established as the most powerful factor in predicting outcome in pediatric ALL,” O’Connor and colleagues wrote. “Our study adds to this evidence by demonstrating the value of MRD assessment in the context of morphologic induction failure.”

Prior research has shown that MRD is more accurate for defining complete remission than morphology; however, studies of the role of MRD in stratifying induction failure in pediatric ALL are lacking. In this study, O’Connor and colleagues studied all cases of morphologic induction failure from the large, multicenter Medical Research Council UK ALL 2003 trial, to determine whether end of induction MRD should be used in defining induction failure. The study included 3,113 patients treated between 2003 and 2011. MRD was measured using real-time PCR. Median follow-up was 5 years and 9 months.

Of the included patients, 59 had induction failure (1.9%); 44 patients had M2 marrow, and 15 patients had M3 marrow. Patients with induction failure had a 5-year event-free survival of 50.7% and a 5-year overall survival of 57.7%.

MRD data were available for 45 of the 59 patients with induction failure. Of these patients, a small proportion of patients with M2 marrow (6 of 44) and a low end-of-induction MRD (< 0.01%) had a 5-year event-free survival of 100%.

“Without central review, the accuracy of M2 status cannot be assured in these six cases and may explain the nonconcordance of morphologic and MRD levels,” the researchers wrote. “Although this only affects a small number of cases, integration of MRD results could potentially spare these patients the toxicity of more intensive therapy.”

In contrast, of patients with morphologic remission, 2.3% had a high MRD (5% or greater) and a 5-year event-free survival of 47%, a rate similar to those patients with morphologic induction failure.

“It is important to note that the real-time quantitative PCR method that was used for MRD measurement is designed to optimally detect disease between 0.01% and 1% and may therefore not give an exact measurement of disease at higher levels,” the researchers wrote. “However, our data clearly show that those patients with morphologic remission, but high end-of-induction MRD, have outcomes similar to patients with morphologic induction failure. Overall, this suggests that MRD measurement provides a more reliable assessment of disease response than morphology, as previously described in pediatric acute myeloid leukemia.”

When the researchers redefined induction failure to also include morphologic induction failure and/or MRD of 5% or greater, they identified 3.9% of patients in the trial with a 5-year event-free survival of 48%. With this definition, induction failure occurred most frequently in T-ALL (10.1%) and B-other ALL (5.6%).

“In light of our results, within the United Kingdom group, we have amended the current protocol to redefine induction failure as end-of-induction MRD ≥ 5% and/or M3 marrow,” the researchers wrote. “Those patients with M2 marrow without an end-of-induction MRD result will have morphology assessed by central review.”

Finally, the researchers also examined whether a targeted genetic screening strategy could identify patients with lesions that could be treated with novel agents. This testing showed the presence of PDGFRB gene fusions within the B-other group, particularly EBF1-PDGFRB in about one-third of B-other ALL cases.