Introduction:

Allogeneic stem cell transplantation (allo-SCT) is a curative option in treating acute myelogenous leukemia (AML), yet relapse remains a major cause of treatment failure. Subsequent therapy for relapsed patients is limited and outcomes remain poor, underscoring the need to identify early signs of relapse for timely intervention. Early donor cell chimerism (DCC) is a key prognostic indicator of relapse, monitoring donor cell engraftment and quantifying donor cells in peripheral blood (CD3) or bone marrow (CD34). High lymphoid CD3 DCC may suggest rapid immune reconstitution and serve as a surrogate marker for early treatment response.

Methods:

This single center retrospective study included AML patients who received bone marrow or peripheral blood allo-SCT at our institution from 2018 to 2022. CD3 chimerism testing was performed at weeks 4, 8, 12, and periodically post-transplant as part of routine care. Complete chimerism (CC) was defined as ≥97% CD3 donor cells at all times, mixed chimerism (MC) as <97% CD3 donor cells, CC to MC as ≥97% CD3 donor cells followed by a decrease to <97%, and MC to CC as <97% CD3 donor cells followed by an increase to ≥97%. Chimerism testing was not done after the time of documented relapse. Primary endpoints were the association of CD3 chimerism trends with overall survival (OS) and relapse-free survival (RFS). Survival differences were estimated using the log-rank test.

Results:

At data cut-off (July 1, 2024), 156 AML patients underwent bone marrow (21.8%) or peripheral blood (78.2%) allo-HSCT, with 26.3% having HLA haplo-matched donors, 12.8% related donors, and 60.9% with unrelated donors. Demographics were as follows: 56.4% male, 83.3% White, 10.3% Black, 6.4% Asian, and 84.6% had English as their primary language. Median age at diagnosis and allo-HSCT was 58 years (range 22-77) and 58 years (range 22-78), respectively. Of these patients, 38.5% received myeloablative conditioning with either fludarabine and busulfan (FluBu4) (72.4%) or fludarabine and 150cGy total body irradiation (FluTBI) (27.6%); 17.9% received non-myeloablative conditioning with fludarabine, cyclophosphamide, and 200cGy TBI (FluCyTBI) (67.9%) or fludarabine and 300cGy TBI (FluTBI) (32.1%); and 43.6% received reduced-intensity conditioning with FluBu2 (27.9%), fludarabine and melphalan (FluMel) (35.3%), or fludarabine, melphalan, and 200cGy TBI (FluMelTBI) (36.8%). Post-transplant graft-versus-host disease (GvHD) immunosuppression regimens included tacrolimus/methotrexate (59%), tacrolimus/mycophenolate (35.9%), or cyclosporine/sirolimus/mycophenolate (5.1%). Of these patients, 48.7% received rabbit anti-thymocyte globulin (rATG) and 22.4% received abatacept. At a median follow-up of 21.73 months (range 1.97-63.94), 55% of patients were alive and 25% experienced relapse.

Based on the CD3 donor cells, patients were categorized in the CC (42.3%), MC (18.0%), CC to MC (5.8%), and MC to CC (34.0%) groups. Using Kaplan-Meier curves, the combined MC and CC to MC groups showed a strong association with worse OS (p<0.001) and RFS (p<0.001) compared to the CC and MC to CC groups. This association remained significant when stratified by conditioning regimens (myeloablative, OS p<0.001, RFS p<0.001; non-myeloablative, OS p=0.001, RFS p<0.001; reduced-intensity, OS p<0.001, RFS p<0.001). In multivariate analysis, MC and CC to MC CD3 chimerism levels were significantly associated with relapse (odds ratio [OR] 9.7 [95% CI 4.03-24.86], p<0.001), 1-year survival (OR 0.22 [95% CI 0.09-0.51], p<0.001), RFS (hazard ratio [HR] 3.36 [95% CI 2.03-5.56], p<0.001), OS (HR 3.12 [95% CI 1.88 - 5.18], p<0.001). Non-English primary language was associated with worse OS (HR 1.83 [95% CI 1.03-3.25], p=0.039). When included in the multivariate analysis, age of transplant, conditioning regimen, ABO match status, CMV status, GvHD regimen were not statistically significant.

Discussion:

Our study found that lower or decreasing CD3 DCC levels are significantly associated with relapse and mortality in AML patients after allo-HSCT. The transplant conditioning and GvHD regimen may impact the predictive value of DCC for relapse risk, and the relationship between DCC and graft-versus-leukemia (GvL) effect remains unknown. These results warrant further investigation of its use as a predictive and prognostic marker for managing early relapse after allo-HSCT in AML patients.

Disclosures

Vesole:Sanofi: Speakers Bureau; Karyopharm: Speakers Bureau; Janssen: Speakers Bureau; BMS: Speakers Bureau; Takeda: Speakers Bureau; Amgen: Speakers Bureau.

This content is only available as a PDF.
2024
Sign in via your Institution