Novel Indicators of Transplant Outcomes for PhALL: Current Molecular-Relapse-Free Survival

Molecular relapse after allogeneic hematopoietic cell transplantation (allo-HCT) has been thought to predict clinical relapse in patients with Philadelphia chromosome-positive acute lymphoblastic leukemia (PhALL). Tyrosine kinase inhibitor (TKI) administration after allo-HCT may dynamically change the status from molecular relapse to molecular remission, but these state changes cannot be accurately represented by conventional survival indicators such as relapse-free survival, where events are usually considered irreversible. We aimed to develop novel indicators of transplant outcomes for allo-HCT recipients with PhALL and to visualize current molecular-relapse-free survival (CMRFS) and current on-TKI status (CTKI), treating molecular relapse or TKI administration after allo-HCT as a reversible event. We retrospectively analyzed 286 patients with PhALL who received allo-HCT between 2000 and 2016 in order to develop the indicators. CMRFS was defined as the probability of molecular remission without clinical relapse or death at any time after allo-HCT. Similarly, CTKI was defined as the probability of TKI administration without clinical relapse or death at any time after allo-HCT. The 1- and 5-year CMRFS rates were 67% and 59%, respectively, whereas the 1- and 5-year conventional molecular relapse-free survival rates were 42% and 37%. The 1- and 5-year CTKI rates were 14% and 8%, respectively. In a post hoc analysis focusing on patients who had achieved a molecular complete remission within 6 weeks (n = 201), the 5-year CMRFS rate (71%) was similar to the 5-year conventional molecular relapse-free survival (molRFS) rate (70%) in the non-TKI group. On the other hand, the 5-year CMRFS rate in the TKI group was 61%, whereas the 5-year conventional molRFS rate was only 38%. CMRFS and CTKI might become useful indicators of transplant success in terms of survival, leukemia-free status, and treatment-free status at any time point. Future extension of these survival models to other clinical situations is warranted.