Donor Chimerism Monitorization for Relapse Prediction in Ex-Vivo CD34-Selected Allogenic Hematopoietic Stem Cell Transplant

Introduction. Allogenic hematopoietic stem cell transplant with CD34 selection (CD34S-HSCT) is an effective strategy for reducing acute graft versus host disease (GVHD) at the expense of increasing the risk of infections. Loss of donor complete chimerism (CC) or the presence of a mixed population chimerism (MC) is a widely described risk factor for relapse. However, information about chimerism behavior in CD34S-HSCT and its association with disease recurrence is lacking. We aimed to assess the impact of MC on CD34S-HSCT outcomes and describe donor chimerism dynamics in these patients.

Methods. We conducted a retrospective study including patients who received a CD34S-HSCT from January 2016 to December 2023. We analyzed baseline characteristics, relapse, survival outcomes and related complications according to the presence of MC in T cells (TL) and granulocytes (GY) at 1, 3, 6, 9 and 12 months post-infusion. Chimerism was measured in peripheral blood and MC was defined as <95% of donor cells (determination by analysis of 15 polymorphic markers “Short Tandem Repeat” type and 1 gender marker). Impact of MC on disease relapse, graft failure, infections and GVHD rates was analyzed. As well as a subgroup analysis based on chimerism dynamics in TL (persistent CC, persistent MC, and fluctuant chimerism).

Results. We included 98 patients with acute myeloblastic leukemia, myelodysplastic syndrome or acute lymphoblastic leukemia diagnosis (n= 60, 20 and 18, respectively). Median follow-up for the full data set was 37.8 months (95% CI 31.5-50.4).

Median age at HSCT was 52 years (18-71). All patients presented a high risk based on Disease Risk Index and median Specific Comorbidity Index was 2. Most frequent donor type was matched unrelated (47%) and half received myeloablative conditioning. Median CD34 and CD3 cells infused were 5.6E6/kg (IQR 4.6-6.2) and 0.02·E5/kg (IQR 0.01-0.03).

For the full cohort, 2-years relapse-free survival (RFS) and overall survival (OS) was 55% (95% CI 46 - 66) and 63% (95% CI 53 - 74), respectively. Twenty-nine patients relapsed at a median time of 5.4 months (IQR 3.6 – 9.8). Regarding HSCT complications, 24% of the patients presented acute GVHD (median maximum grade 2, [range 1-3]) and 3 patients presented chronic GVHD (median maximum grade 2, [range 2-3]. One patient presented primary graft failure and poor engraftment rate was 6.1%. Most frequent cause of death was relapse (n= 20, 20.4%) and non-relapse mortality was 18.3%, predominantly due to infections.

MC in TL rate at 1, 3, 6, 9 and 12 months was 37.5% (n=18/48), 79.7% (n=51/64), 83.4% (n=52/62), 90.4% (n=38/42) and 87.1% (n=34/39), respectively. At each point, there were no differences in acute or chronic GVHD incidence, CMV reactivation, other infections, poor engraftment or relapse rate. RFS and OS were similar in TL-MC and CC subgroups. Baseline characteristics (including CD3 cells infused, intensity conditioning, HLA mismatch, CMV serology mismatch and letermovir prophylaxis) in both groups were comparable.

In contrast, CC in GY prevailed at all time points (n=85/94; n=74/88; n=66/73; n=38/42; n=42/46). No differences regarding baseline characteristics were observed among patients with CC or MC in GY, except for a tendency to higher MC for reduced-intensity conditioning (RIC). More cases of poor engraftment were recorded in the GY-MC group at 9 months, with no differences in infections, GVHD and relapse rates. RFS and OS were also comparable.

Finally, we carried out a subgroup analysis according to chimerism dynamics in TL during the follow up: persistent CC (n=39, 50%), persistent MC (n=17, 21.8%) and fluctuant chimerism (n=22, 28.2%). RIC correlates with persistent MC and fluctuant chimerism (51% and 55% vs 18%), There were no differences in poor engraftment, infection or relapse rate. Incidence and severity of acute GVHD was similar, however CC group presented all recorded cases of chronic GVHD (p=0.045). Persistent MC and fluctuant chimerism were not associated with a worst RFS or OS.

Conclusions. In our cohort, persistence of MC in TL along the post CD34S-HSCT long period does not seem to confer a negative impact on disease relapse or survival. Early intervention based only on MC behavior would not be necessary.