Response: Unrelated donor transplantation for adults with Ph⁻ ALL in first complete remission

In response to our paper¹  about unrelated donor (URD) allografts for first remission acute lymphoblastic leukemia (ALL), Dr Mehta raises issues of how selection of hematopoietic cell transplantation (HCT) candidates (who need to stay in remission long enough to receive the HCT), often termed selection bias, can limit the clinical value of such analyses in informing patients of their likely outcome and guiding therapeutic decisions.

Approximately 90% of adults with Philadelphia-negative (Ph⁻) ALL will attain remission (possibly less in this high risk group). A proportion will relapse or suffer serious complications before an URD allograft can be performed. However, the high-risk patient population in our study had a low chance of cure with chemotherapy and is typical of patients for whom URD HCT is considered.

Certain patient subsets (high white blood cell count, adverse cytogenetics) clearly do poorly with standard chemotherapy; more than 90% of our reported cohort had 1 or more high-risk features. It is unlikely that a randomized study will ever directly study these high-risk Ph⁻ patients and we must rely on lesser levels of evidence. In the next United Kingdom/United States trial, patients with high-risk features will proceed to URD HCT if a donor is identified and available within a defined interval, permitting donor (related or URD) versus no-donor analyses of outcome. However, these statistical techniques may underestimate the value of the allografts because not all patients with a donor proceed to HCT.

Dr Mehta proposes subtracting 20% from the observed survival of URD HCT series based on a reported average 80% 6-month overall survival for adult ALL patients. However, this proposed 20% subtraction approach does not fully consider and adjust for the differing survival probability from CR1 to HCT based on each individual patient's characteristics, including the time until donor availability and the initiation of either HCT or intensified nontransplantation therapies. This oversimplified approach is likely misleading as it does not distinguish reasons for treatment failure: both early relapse and serious postinduction complications prevent delivery of both intensive chemotherapy and HCT. Importantly, our study analyzed patients from the time of transplantation, not from CR1. To estimate the survival probability since CR1 would require a complete file on the available patients considered for HCT as they achieve CR1; information not available in data reported to the CIBMTR.

Despite the recently published but somewhat dated (and selected) 20+-year series of autografts he cites,²  the recently published nearly 2000-patient UKALL XII/ECOG prospective randomized trial clearly showed that chemotherapy is superior to autografting for adult ALL patients in CR1.³  In addition, the recent CIBMTR long-term follow-up report⁴  noting similarity in outcome in adult ALL between selected (and somewhat lower risk) autografts and earlier era (1989-1998) URD allografts also does not favor a broader application of autotransplantation. Outside of a clinical trial, autografts currently have no clearly defined utility in the management of adult ALL patients.