Contributions
Abstract: EP366
Type: E-Poster Presentation
Session title: Acute lymphoblastic leukemia - Clinical
Background
Treatment response measures for R/R aALL have evolved with improved understanding of ALL characteristics, which has led to different definitions in clinical trials.
Aims
This study aimed to examine the comparability of response measurements used in trials of current standard of care (SoC) treatments for R/R aALL, with a focus on complete remission (CR) and CR with incomplete hematologic recovery (CRi).
Methods
An SLR was conducted to identify clinical trials reporting treatment response with SoC for R/R aALL. Searches of the MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials databases as well as conference proceedings/trial registries were conducted in November 2020. Inclusion criteria were applied to restrict to trials of licensed and recommended treatments for R/R aALL, i.e. blinatumomab, inotuzumab ozogamicin, tyrosine kinase inhibitors (imatinib, dasatinib, ponatinib), and chemotherapies (fludarabine-, cytarabine-, alkylator-containing regimens). Additionally, criteria were further applied to exclude clinical trials with no full-text publications that reported CR alone or both CR and CRi by search completion date.
Results
The SLR identified 15 treatment arms in 12 studies which reported CR, and 14 treatment arms in 11 studies which reported both CR and CRi. Across the 15 treatment arms reporting CR, the definitions were comparable, i.e., an absolute neutrophil count (ANC) response ≥1 x 109/l and platelet count ≥100 x 109/l along with ≤5% bone marrow blasts. For the 14 treatment arms in 11 studies with both CR and CRi, the CRi definitions varied in hematologic criteria. Many of the 11 studies were heterogenous in terms of specific ANC and platelet thresholds employed for CRi. Moreover, some studies did not require hematologic recovery. Specifically, six studies included a hematological recovery requirement in ANC and/or platelets for CRi with variant thresholds. Three studies did not have a minimum requirement for ANC or platelets, and two studies did not explicitly describe the hematological criteria. Among all studies with CRi hematological recovery thresholds, the most stringent criteria were used in the TOWER trial, defined as ANC ≥1 x 109/l or platelet count ≥100 x 109/l. The differences in CRi definitions were clinically meaningful and led to significantly different response rates.
Conclusion
Whereas CR definitions were comparable across trials, CRi definitions were very heterogenous in terms of hematologic recovery criteria, with older definitions being less stringent or requiring no hematological recovery. The lack of standardized CRi definitions make it challenging for cross-trial efficacy comparisons with composite response endpoints such as CR+CRi due to the different levels of stringency in criteria for CRi. This suggests that treatment response outcome measures for historical control studies should be based on CR alone. Outcomes with comparable definitions across trials such as CR and standard efficacy measures such as overall survival are likely to provide more robust and clinically relevant cross-trial efficacy comparisons.
Keyword(s): Acute lymphoblastic leukemia, Complete remission, Refractory, Relapsed acute lymphoblastic leukemia
Abstract: EP366
Type: E-Poster Presentation
Session title: Acute lymphoblastic leukemia - Clinical
Background
Treatment response measures for R/R aALL have evolved with improved understanding of ALL characteristics, which has led to different definitions in clinical trials.
Aims
This study aimed to examine the comparability of response measurements used in trials of current standard of care (SoC) treatments for R/R aALL, with a focus on complete remission (CR) and CR with incomplete hematologic recovery (CRi).
Methods
An SLR was conducted to identify clinical trials reporting treatment response with SoC for R/R aALL. Searches of the MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials databases as well as conference proceedings/trial registries were conducted in November 2020. Inclusion criteria were applied to restrict to trials of licensed and recommended treatments for R/R aALL, i.e. blinatumomab, inotuzumab ozogamicin, tyrosine kinase inhibitors (imatinib, dasatinib, ponatinib), and chemotherapies (fludarabine-, cytarabine-, alkylator-containing regimens). Additionally, criteria were further applied to exclude clinical trials with no full-text publications that reported CR alone or both CR and CRi by search completion date.
Results
The SLR identified 15 treatment arms in 12 studies which reported CR, and 14 treatment arms in 11 studies which reported both CR and CRi. Across the 15 treatment arms reporting CR, the definitions were comparable, i.e., an absolute neutrophil count (ANC) response ≥1 x 109/l and platelet count ≥100 x 109/l along with ≤5% bone marrow blasts. For the 14 treatment arms in 11 studies with both CR and CRi, the CRi definitions varied in hematologic criteria. Many of the 11 studies were heterogenous in terms of specific ANC and platelet thresholds employed for CRi. Moreover, some studies did not require hematologic recovery. Specifically, six studies included a hematological recovery requirement in ANC and/or platelets for CRi with variant thresholds. Three studies did not have a minimum requirement for ANC or platelets, and two studies did not explicitly describe the hematological criteria. Among all studies with CRi hematological recovery thresholds, the most stringent criteria were used in the TOWER trial, defined as ANC ≥1 x 109/l or platelet count ≥100 x 109/l. The differences in CRi definitions were clinically meaningful and led to significantly different response rates.
Conclusion
Whereas CR definitions were comparable across trials, CRi definitions were very heterogenous in terms of hematologic recovery criteria, with older definitions being less stringent or requiring no hematological recovery. The lack of standardized CRi definitions make it challenging for cross-trial efficacy comparisons with composite response endpoints such as CR+CRi due to the different levels of stringency in criteria for CRi. This suggests that treatment response outcome measures for historical control studies should be based on CR alone. Outcomes with comparable definitions across trials such as CR and standard efficacy measures such as overall survival are likely to provide more robust and clinically relevant cross-trial efficacy comparisons.
Keyword(s): Acute lymphoblastic leukemia, Complete remission, Refractory, Relapsed acute lymphoblastic leukemia