The predictive value of T-cell chimerism for disease relapse after allogeneic hematopoietic stem cell transplantation

Introduction Chimerism is closely correlated with disease relapse after allogeneic hematopoietic stem cell transplantation (allo-HSCT). However, chimerism rate is dynamic changes, and the sensitivity of different chimerism requires further research. Methods To investigate the predictive value of distinct chimerism for relapse, we measured bone marrow (BM), peripheral blood (PB), and T-cell (isolated from BM) chimerism in 178 patients after allo-HSCT. Results Receiver operating characteristic (ROC) curve showed that T-cell chimerism was more suitable to predict relapse after allo-HSCT compared with PB and BM chimerism. The cutoff value of T-cell chimerism for predicting relapse was 99.45%. Leukemia and myelodysplastic syndrome (MDS) relapse patients’ T-cell chimerism was a gradual decline from 2 months to 9 months after allo-HSCT. Higher risk of relapse and death within 1 year after allo-HSCT. The T-cell chimerism rates in remission and relapse patients were 99.43% and 94.28% at 3 months after allo-HSCT (P = 0.009), 99.31% and 95.27% at 6 months after allo-HSCT (P = 0.013), and 99.26% and 91.32% at 9 months after allo-HSCT (P = 0.024), respectively. There was a significant difference (P = 0.036) for T-cell chimerism between early relapse (relapse within 9 months after allo-HSCT) and late relapse (relapse after 9 months after allo-HSCT) at 2 months after allo-HSCT. Every 1% increase in T-cell chimerism, the hazard ratio for disease relapse was 0.967 (95% CI: 0.948–0.987, P<0.001). Discussion We recommend constant monitoring T-cell chimerism at 2, 3, 6, and 9 months after allo-HSCT to predict relapse.


Introduction
Allogeneic hematopoietic stem cell transplantation (allo-HSCT) provides the most effective therapy in hematological malignancies (HM) and is used to treat leukemia and myelodysplasia (MDS) (1,2).Despite the enormous benefit of allo-HSCT and improved patient overall survival, disease relapse remains the primary obstacle to successful allo-HSCT.Although the majority of patients with leukemia/MDS can be cured by allo-HSCT, almost all high-risk patients will eventually relapse (3).Relapse represents the leading cause of failure in patients after allo-HSCT and the clinical prognosis for HM patients with allo-HSCT failure is poor (4).Relapse is the most frequent cause of death in transplant patients (5).Notably, deaths due to relapse account for nearly 30% of all deaths within 100 days post-transplant (6).As a consequence, an ideal strategy would be to monitor patients who are at risk of relapses early and then institute effective therapies.
Multiple studies have demonstrated that chimerism monitoring might be associated with HM relapse after allo-HSCT (6,7).Chimerism can broadly be defined as being able to quantify donor and recipient hematopoiesis and can be conducted in T cells, peripheral blood (PB) and bone marrow (BM) (6,8,9).The complete donor chimerism relies on allo-HSCT and the complete reconstruction immune system in HM patients (10).Generally, posttransplant chimerism is a dynamic phenomenon and the mixed chimerism (MC) may be indicative of relapse (11).Given the relapse and poor outcomes in the disease of leukemia/ MDS, it is necessary to verify the predictive value of donor chimerism.In the present study, we compare and contrast the dynamic range difference between T-cell chimerism, PB chimerism, and BM chimerism and put forward the most appropriate testing method.

Patients
After approval of the Ethical Committee of the 960th Hospital of the People's Liberation Army, a cohort of 178 allo-HSCT patients who underwent chimerism determination was retrospectively analyzed.Chimerism determination was performed at least twice.

Magnetic bead sorting
Magnetic bead sorting separates T cells by binding antibodies on magnetic beads to antigens on T cell surfaces.BM is collected in a centrifuge tube, an appropriate amount of antibody is added and shaken well, followed by the addition of pre-prepared magnetic beads, shaken again, and allowed to stand.Place the tube in the magnetic pole, and the magnetic beads will adhere to the tube's wall, removing the unbound cell solution while binding the T cells to the magnetic beads.PBS was added and washed twice, resulting in purified T cells with over 90% purity.

Definitions
HM relapse was defined by the reappearance of leukemic blasts in PB or relapse of BM blasts ≥5% or the development of extramedullary disease after allo-HSCT (12, 13).T-cell, PB, and BM donor chimerism were monitored at different time points after transplantation.Complete chimerism (CC) was broadly defined as >95% donor cells in the PB and BM samples and MC was defined as 5 to 95% donor cells (12, 14).

Statistical analysis
Descriptive statistics were used to summarize the clinical characteristics of patients in the study.The quantitative data were expressed as mean ± SD (standard deviation).Shapiro-Wilk and Levene test were used to analyze normality and homogeneity of variance of all experimental data.Data with normally and homogeneously distributed were analyzed using SPSS 26.0 by one-way analysis of variance (ANOVA).When all variables in the distribution failed to meet the assumptions of normality and homogeneity, Kruskal-Wallis one-way ANOVA by ranks was used for data evaluation.Significant differences were defined as pvalues less than 0.05.Diagnostic accuracy of chimerism on relapse was compared using receiver operating characteristic (ROC) curves.The influence of chimerism at different periods of relapse was evaluated using the Cox proportional hazards model.

Patients
Between May 2017 and June 2023, 178 patients underwent allo-HSCT for leukemia/MDS in our institution.In case of multiple relapse, only the first relapse was considered.One hundred fiftynine patients underwent T-cell chimerism testing.In addition, 170 patients underwent BM chimerism testing, while 144 underwent PB chimerism testing.Baseline characteristics of patients are detailed in Table 1.

The number of patients and chimerism status
In total, 178 leukemia/MDS allo-HSCT patients were entered into the study cohort.The total number of remission in leukemia/ MDS was 106 and 34, respectively, whereas the number of relapses was 32 and 6 (Figure 1A).In the ROC curve, T-cell chimerism showed the highest area under the curve (AUC) value (Figure 1B).The cutoff value of the T-cell chimerism was 99.45%, the sensitivity was 88.5%, and the specificity was 86.7% (AUC = 0.879, P < 0.001).The cutoff value of the BM chimerism was 98.12%, the sensitivity was 51.9%, and the specificity was 100% (AUC = 0.707, P = 0.01).The cutoff value of the PB chimerism was 98.18%, the sensitivity was 45.5%, and the specificity was 100% (AUC = 0.739, P = 0.083).

Comparison of donor-host chimerism in relapse patients at different time intervals after allo-HSCT
There was a high risk of relapse and mortality in relapsed patients within 12 months after allo-HSCT and three chimerism displayed a downward trend as a whole (Figure 2).Chimerism reached its highest point at 12 months after allo-HSCT.T-cell chimerism was statistically significant for 9 months and 12 months after allo-HSCT (P = 0.001).Detailed values are shown in Supplementary Table S1.

Comparison of donor-host chimerism in remission and relapse patients at different time intervals after allo-HSCT
T-cell chimerism decreased progressively from 2 months after allo-HSCT and dropped to the lowest point at 9 months (Figure 3A).T-cell chimerism of remission and relapsed patients at 3 months was 99.43% and 94.28% (P = 0.009), respectively.T-cell  chimerism of remission and relapsed patients at 6 months was 99.31% and 95.27% (P = 0.013), respectively.T-cell chimerism of remission and relapsed patients at 9 months was 99.26% and 91.32% (P = 0.024), respectively.BM chimerism decreased progressively from 3 months after allo-HSCT and dropped to the lowest point at 9 months (Figure 3B).BM chimerism of remission and relapsed patients at 6 months was 96.38% and 89.82% (P = 0.044), respectively.BM chimerism of remission and relapsed patients at 9 months was 99.33% and 82.79% (P = 0.198), respectively.PB chimerism decreased progressively from 6 months after allo-HSCT and dropped to the lowest point at 9 months (Figure 3C).PB chimerism of remission and relapsed patients at 9 months was 99.58% and 85.78% (P = 0.531), respectively.Detailed values are shown in Supplementary Table S2.
Comparison of donor-host chimerism in early relapse and late relapse patients at different time intervals after allo-HSCT Among 38 relapse patients, T-cell chimerism of early relapse (relapse within 9 months after allo-HSCT) and late relapse (relapse after 9 months after allo-HSCT) was 99.16% and 99.76% (P = 0.036) at 2 months after allo-HSCT, respectively (Figure 4A).BM chimerism of early relapse patients showed a decreasing trend at 3 months after allo-HSCT, while the PB chimerism decreasing tendency was preserved at 6 months after allo-HSCT (Figures 4B, C).Detailed values are shown in Supplementary Table S3.

Relapse risk
We created a univariate Cox model for T-cell chimerism predicting relapse.

Discussion
Due to failure to avoid relapse of disease, relapse patients after allo-HSCT had a poor prognosis and limited therapies (15).HM relapse remains the primary cause of the dismal prognosis after allo-HSCT (16, 17).Thus, precision predicting the relapse of disease after allo-HSCT to guide treatment decisions is particularly important.To predict the relapse of high-risk individuals early and take specific intervention measures in time, it is helpful for clinicians to reduce the poor outcomes of patients.Related studies have confirmed that the chimerism status was strongly associated with the risk of disease relapse (7,(18)(19)(20).Chimerism is a dynamic process and ever-changing, and MC is responsible for allo-HSCT failure (21).The most appropriate chimerism analysis is based on disease and allo-HSCT type (22).However, there is a lack of comparative analysis of different chimerism of dynamic changes.This report details the evolution of dynamic changes in T-cell chimerism, BM chimerism, and PB chimerism at different points after allo-HSCT.By examining these chimerism statuses, one can identify the chimerism that is most strongly linked to relapse, which offers valuable reference information for clinical practice.
The ROC curve is universally used as a method for assessing the accuracy of potential biomarkers by calculating the area under the ROC curve (AUC) (23)(24)(25).As shown in the ROC curve, compared with BM chimerism and PB chimerism, T-cell chimerism had the highest AUC and was more suitable for the prediction of disease relapse.Through successive tests can be able to effectively identify immunophenotypic markers, chimerism detection can be used to supplement them.The feasibility of combining the two in clinical research has been confirmed in related studies.
In conclusion, our research clarifies that, in contrast to BM and PB chimerism, continuous T-cell chimerism monitoring predicting relapse is more accurate.Therefore, we recommend that the T-cell chimerism rate should be closely monitored for 2, 3, 6, and 9 months after allo-HSCT.This would be helpful for early detection of relapse after allo-HSCT and timely interventions.

TABLE 1
Baseline characteristics of patients in chimerism analyses.