TNC/Kg and CD34+/Kg Cell Dose Impact on Patient Engraftment after Allogeneic Transplant with Conventional Bone Marrow; Better Predictability of Graft Content Needed

Introduction: Bone marrow (BM) harvested from a related (RD) or unrelated donor (URD) remains a viable graft source for patients (pts) receiving an allogeneic hematopoietic stem cell transplant (allo-HCT). However guidance on minimum TNC/kg & CD34+/kg thresholds to optimize engraftment & a method to best predict graft content are needed.

Methods: All BM products (prdcts) harvested between 1/1/2016 to 6/30/2023 for adult & pediatric (ped) pts were studied (n= 261). Pt, donor, & BM prdct characteristics were collected. For each BM prdct, pre-processing (pre-proc) data was obtained to determine a median pre-proc TNC, TNC/mL, CD34+/mL, and CD34+/TNC % of all BM prdcts & infused TNC & CD34+ content was also recorded. Pt neutrophil (Neutr) & platelet (Plt) engraftment data was collected with Neutr engraftment defined as the first of 3 consecutive days of an absolute neutr count ≥ 500 x 10^3/μl. Plt engraftment was defined as ≥ 20K/μl plts & transfusion independence for ≥ 7 days. Univariable & multivariable analyses (MVAs) using the Fine and Gray model were performed to assess impact of infused TNC/kg (MVA1), CD34+/kg (MVA2), and CD34+/TNC % (MVA3) of conventional BM prdcts on Neutr (n eval pts = 239) & Plt engraftment (n eval pts = 238).

The National Marrow Donor Program's (NMDP) provided estimate for median TNC/mL of BM (0.183 x 10^8) was used w/ donor weight (wt), the NMDP maximum (max) allowable amount of BM (20mL per donor kg), & a standard max BM volume request (1500mL) to calculate an expected pre-proc TNC for all BM prdcts. The expected TNC for small ped pts was limited to (15 * pt wt) to ensure a TNC/kg cell dose of ≤ 15 x 10^8. An expected pre-proc CD34+ for each BM prdct was then calculated by expected TNC x CD34+/TNC % of all BM prdcts on study. An expectation accuracy (EA) % for both TNC & CD34+ was calculated for each BM prdct (expected / actual pre-proc values observed by our center). TNC & CD34+ EA % of BM prdcts collected by other domestic & international centers were analyzed.

Results: 261 BM prdcts from 259 donors for 258 pts were studied. The BM prdcts had the following pre-proc medians: TNC/mL of 0.136 x 10^8, CD34+/mL of 0.124 x 10^6, & CD34+/TNC % of 0.96. MSKCC harvested 141 (54%) of the BM prdcts and the rest were facilitated by the NMDP: 71 (27%) from 26 domestic centers and 49 (19%) from 28 international centers. The cumulative incidence of Neutr engraftment was 96% (95% CI 92, 98%) by days +30 and +100 after allo-HCT and for Plts it was 48% (95% CI 42,54%) and 91% (95% CI 87,94%) respectively.

Table 1 shows the MVA results testing TNC/kg (MVA1), CD34+/kg (MVA2), CD34+/TNC % (MVA3) & engraftment after allo-HCT w/ conventional BM, indicating Neutr & Plt engraftment is impacted by TNC/kg and CD34+/kg cell dose but not by CD34+/TNC %.

In all 3 MVA models, associations were observed for pt diagnosis [SAA w/ better Plt engraftment vs. ref grp], donor type (≤ 7/8 URD or haploidentical RD w/ worse Neutr & Plt engraftment vs. ref grp), donor ABO incompatibility (Minor w/ better Plt engraftment vs. ref grp), and infusion type [cryopreserved (cryo'ed)/thawed w/ worse Neutr engraftment vs. fresh].

The expectation accuracy (EA) analysis (Wilcoxon rank-sum test) of pre-proc TNC & CD34+ is seen in Table 2. The median expected pre-proc TNC (274) & CD34+ (264) of BM prdcts received from domestic centers & international centers were the same. However BM prdcts from international centers had a lower median actual pre-proc TNC (192) & CD34+ (162), a lower median TNC EA of 74% (vs. 89% for dom centers, p=0.008), & a lower median CD34+ EA of 63% (vs. 93% for dom centers, p<0.001). All BM prdcts combined (n = 120) had a median TNC EA of 83% & a median CD34+ EA of 76%.

Conclusions:

This large analysis suggests BM products with a TNC/kg > 2.0 x 10^8 and a CD34+/kg > 2.0 x10^6 were associated with better Neutr & Plt engraftment, as was allo-HCT with an HLA-matched donor. Cryo'ed products were associated with worse Neutr but not Plt engraftment, while SAA & having a BM graft from a donor with a minor ABO incompatibility were associated with better Plt engraftment.

The expectation accuracy analysis suggests BM graft content is often not meeting expectations, highlighting a need for updated guidance on expectations of BM graft content. An update to NMDP's guidance on estimated TNC/mL of BM or analysis and provision of additional data (e.g., median pre-proc CD34+/mL or CD34+/TNC %) could allow centers to better anticipate TNC & CD34+ BM graft content and help guide donor & graft selection.