Haplo-Cord UCB SCT with Low Cell Dose, Well Matched UCB Units. a Prospective Multicenter Study

The outcome of umbilical cord blood transplantation is governed by UCB cell dose and matching. But stringent cell dose requirements limit access to well matched units. In haplo –cord transplant, early engraftment is determined by the adult graft. Elsewhere we have shown that excessive doses of haplo-identical cells resulted in failure of the UCB graft. Here we investigated the effect of a lower UCB cell dose threshold on ability to identify well matching units. We also prospectively evaluated the effects of UCB cell dose and matching (combined with a fixed dose haplo-graft ) on cord engraftment, platelet recovery, neutrophil recovery and acute GVHD, PFS and OS.

We enrolled three consecutive cohorts with increasingly lower UCB cell dose thresholds (1) UCB cell dose >2x 10^ NC/kg (2)≥1x 10^ NC/kg and (3)≥.5x 10^ NC/kg. Additional selection criteria included (1) best HLA match (initially based on 6 HLA, as of mid 2013 based on 8 HR antigens (2)UCB viability, (3) avoidance of DSA antibodies and (4) NIMA compatibility. We continued enrollment on each cohort until at least 10 patients were evaluable for day 100 chimerism. If more than 7 of ten evaluable patients in a cohort achieved >40% UCB chimerism by d 100, the next cohort was started. This rate of success was achieved in all three cohorts, so we have continued accrual in the lowest cell dose. Sixty patients have been accrued at two institutions over 19 months. 43/60 pts had AML or MDS. The majority had ASBMT advanced and intermediate disease. Nearly half of the pts were of minority descent. Conditioning in all pts consisted of fludarabine-melphalan (three pts also received TBI 400) and GVHD prophylaxis of tacrolimus and mycophenolate. Haplo identical CD34 selected cells were co-administered at a dose of 3-5 x10^6 CD34 cells/kg. Pt characteristics and preliminary outcomes are summarized in the table below.

The average HLA matching was significantly better (P<.001) in groups 2 and 3 (66% with >6/8 match), than in group 1 (6% with >6/8 match). UCB cell doses were significantly lower in groups 2 and 3 (55% with UCB cell dose <2x10^7/kg vs none in group 1). There were no significant difference between the groups in hematologic recovery, OS, PFS or UCB engraftment.

Seven of 50 evaluable pts had failure of the UCB graft (<40% UCB chimerism on d 100). 4 of them had 100% haplo chimerism, 2 had autologous recovery and 1 had partial <40% UCB chimerism. The incidence of UCB failure increased with decreasing UCB cell dose (P<0.001 U test), with five of seven graft failures occurring in patients receiving UCB doses <1.2 x10^7/kg.

This novel cord blood selection algorithm allows identification of well-matched UCB cells (at least 6/8 HR) in two thirds of patients. In approximately one third of cases the best matching UCB unit has a cell dose <1.5 x10^7/kg. Very low UCB cell doses appear to be associated with somewhat increased risk of UCB graft failure-delay. Long-term follow up will establish the impact of this selection algorithm on OS, PFS, chronic GVHD.