A disease risk index is now available for pediatric patients with acute myeloid leukemia or acute lymphoblastic leukemia who undergo allogeneic hematopoietic stem cell transplantation.
The model, which was developed and validated using data from more than 2,000 patients, stratifies probabilities of leukemia-free survival (LFS) into four risk groups for acute myeloid leukemia (AML) and three risk groups for acute lymphoblastic leukemia (ALL), reported lead author Muna Qayed, MD, of Emory University, Atlanta, who presented findings as part of the American Society of Clinical Oncology virtual scientific program.
“The outcome of stem cell transplantation for hematologic malignancy is influenced by disease type, cytogenetics, and disease status at transplantation,” Dr. Qayed said. “In adults, these attributes were used to develop the disease risk index, or DRI, that can stratify patients for overall survival for purposes such as prognostication or clinical trial entry.”
But no such model exists for pediatric patients, Dr. Qayed said, noting that the adult DRI was found to be inaccurate when applied to children.
“[T]he [adult] DRI did not differentiate [pediatric] patients by overall survival,” Dr. Qayed said. “Therefore, knowing that pediatric AML and ALL differ biologically from adult leukemia, and further, treatment strategies differ between adults and children, we aimed to develop a pediatric-specific DRI.”
This involved analysis of data from 1,135 children with AML and 1,228 children with ALL who underwent transplantation between 2008 and 2017. All patients had myeloablative conditioning, and 75% received an unrelated donor graft. Haploidentical transplants were excluded because of small sample size.
Analyses were conducted in AML and ALL cohorts, with patients in each population randomized to training and validation subgroups in a 1:1 ratio. The primary outcome was LFS. Cox regression models were used to identify significant characteristics, which were then integrated into a prognostic scoring system for the training groups. These scoring systems were then tested in the validation subgroups. Maximum likelihood was used to identify age cutoffs, which were 3 years for AML and 2 years for ALL.
In both cohorts, disease status at transplantation was characterized by complete remission and minimal residual disease status.
In the AML cohort, approximately one-third of patients were in first complete remission with negative minimal residual disease. Risk was stratified into four groups, including good, intermediate, high, and very high risk, with respective 5-year LFS probabilities of 81%, 56%, 44%, and 21%. Independent predictors of poorer outcome included unfavorable cytogenetics, first or second complete remission with minimal residual disease positivity, relapse at transplantation, and age less than 3 years.
In the ALL cohort, risk was stratified into three risk tiers: good, intermediate, and high, with 5-year LFS probabilities of 68%, 50%, and 15%, respectively. Independent predictors of poorer outcome included age less than 2 years, relapse at transplantation, and second complete remission regardless of minimal residual disease status.
The models for each disease also predicted overall survival.
For AML, hazard ratios, ascending from good to very-high-risk tiers, were 1.00, 3.52, 4.67, and 8.62. For ALL risk tiers, ascending hazard ratios were 1.00, 2.16, and 3.86.
“In summary, the pediatric disease risk index validated for leukemia-free survival and overall survival successfully stratifies children with acute leukemia at the time of transplantation,” Dr. Qayed said.
She concluded her presentation by highlighting the practicality and relevance of the new scoring system.
“The components included in the scoring system used information that is readily available pretransplantation, lending support to the deliverability of the prognostic scoring system,” Dr. Qayed said. “It can further be used for improved interpretation of multicenter data and in clinical trials for risk stratification.”
In a virtual presentation, invited discussant Nirali N. Shah, MD, of the National Cancer Institute, Bethesda, Md., first emphasized the clinical importance of an accurate disease risk index for pediatric patients.
“When going into transplant, the No. 1 question that all parents will ask is: ‘Will my child be cured?’ ” she said.
According to Dr. Shah, the risk model developed by Dr. Qayed and colleagues is built on a strong foundation, including adequate sample size, comprehensive disease characterization, exclusion of patients that did not undergo myeloablative conditioning, and use of minimal residual disease status.
Still, more work is needed, Dr. Shah said.
“This DRI will need to be prospectively tested and compared to other established risk factors. For instance, minimal residual disease alone can be further stratified and has a significant role in establishing risk for posttransplant relapse. And the development of acute graft-versus-host disease also plays an important role in posttransplant relapse.”
Dr. Shah went on to outline potential areas of improvement.
“[F]uture directions for this study could include incorporation of early posttransplant events like graft-versus-host disease, potential stratification of the minimal residual disease results among those patients in complete remission, and potential application of this DRI to the adolescent and young adult population, which may have slight variation even from the adult DRI.”The study was funded by the National Institutes of Health. The investigators disclosed no conflicts of interest
SOURCE: Qayed M et al. ASCO 2020, Abstract 7503.