Unrelated Transplants for Poor Prognosis Adult Acute Lymphoblastic Leukemia: Long-Term Comparative Analysis Based on the Hematopoietic Progenitor Source

Background: Adults with high risk ALL features at diagnosis, slow responders or with recurrent disease have a poor outcome with standard chemotherapy and are considered for unrelated transplants in most centers if a matched sibling donor is not available. Unrelated cord blood (UCB) has emerged as an option for unrelated transplant in adult patients. Aim of the study: to compare the outcome of adult patients with unrelated transplant for poor prognosis ALL based on the hematopoietic source used for transplant. Patients and Methods: One hundred and forty- nine adult patients (median 29 years [15-59], 90M/59F) with poor prognosis ALL received an unrelated transplant in 13 Spanish institutions from 2000 to 2007. Patients in 1^st CR at transplant met at least one adverse prognostic factor (adverse cytogenetics, hyperleukocytosis or a slow-responder). ALL was of precursor B lineage in 111 (74%), T-cell lineage in 28 (19%) and undetermined lineage in 10 (7%) patients. ALL was in 1^st CR in 81 (54%) patients, in 2^nd CR in 37 (25%), in 3^rd CR in 11 (7%) and overt disease in 20 (13 %) patients. Both groups were comparable for the main clinical and biologic ALL features except for patients with more overt disease in UCB transplant group. Conditioning therapy consisted on TBI-CY in 68 (46%), BU-CY in 9 (6%), Thiotepa-BU-CY/FLU in 60 (40%) patients (all of them UCB transplants) and other regimens in 12 (8%) patients (6 of them were non-myeloablative). The source of hematopoietic progenitors was UCB in 62 (41%), mobilized PB in 41 (28%) and BM in 46 (31%) patients. HLA compatibility requirements for selecting unrelated donors (BM and mobilized PB) were 7-8 out of 8 allelic identities, and for UCB were 4-6 out of 6 A / B antigenic and DR allelic identities. Results: Acute GVHD was most frequent after PBSCT and BMT thanin UCBT. There was no significant difference for limited and extensive chronic GVHD between UCBT and PBSCT/BMT. The median follow-up was 16 months (0.3-101.4). DFS estimated at 5 years for patients transplanted with any source was not significantly better in transplants in 1^st CR versus 2^nd CR or more advanced disease (median 12 [2-22] vs 5 [2-9], p=0.106) . There was no statistical difference in OS or DFS at 5 years between UCB and PBSCT/BMT). TRM was significantly lower in UCB transplants (p=0.021). The relapse probability was 17% for the PBSCT/BMT versus 29% for UCBT (p=0.088). The use of TBI as conditioning therapy was not associated with a lower relapse rate.

	5y-OS (95%CI)	Median OS (month, 95%CI)	5y-DFS (95%CI)	Median DFS (month, 95%CI)	5y-TRM (95%CI)
Whole series	26% (17-35)	10 (4-17)	21% (13-29)	7 (3-11)	54% (44-64)
PBSCT/BMT	22% (11-33)	9 (2-17)	21% (11-31)	5 (2-9)	63% (50-76)*
UCBT	33% (18-48)	15 (7-22)	22% (8-36)	9  (1-17)	39% (25-53)*
1st CR ALL	30% (16-44)	12 (0-29)	24% (11-37)	12 (2-22)	57% (43-71)
PBSCT/BMT	27% (9-45)	27 (2-51)	24% (8-40)	12 (2-22)	63% (44-82)
UCBT	35% (14-56)	10 (0-23)	24% (4-44)	10 (0-26)	50% (31-69)

   *p=0.021
GVHD was the main cause of death for unrelated PBSCT/BMT in contrast with infection and relapse for UCBT (p=0.001). Conclusions: UCB transplant and unrelated PBSCT/BMT are equivalent options for poor prognosis adult ALL patients without a sibling donor. However, all modalities are associated with high transplant related mortality.
Supported in part with grants PEF-07 from the Jose Carreras Leukaemia Foundation and RD 06/0020/1014 from RETICS.