Chimerism-Based Tolerance to Kidney Allografts in Humans: Novel Insights and Future Perspectives

Chronic rejection and immunosuppression-related toxicity severely affect long-term outcomes of kidney transplantation. The induction of transplantation tolerance – the lack of destructive immune responses to a transplanted organ in the absence of immunosuppression – could potentially overcome these limitations. Immune tolerance to kidney allografts from living donors has been successfully achieved in humans through clinical protocols based on chimerism induction with hematopoietic cell transplantation after non-myeloablative conditioning. Notably, two of these protocols have led to immune tolerance in a significant fraction of HLA-mismatched donor-recipient combinations, which represent the large majority of cases in clinical practice. Studies in mice and large animals have been critical in dissecting tolerance mechanisms and in selecting the most promising approaches for human translation. However, there are several key differences in tolerance induction between these models and humans, including the rate of success and stability of donor chimerism, as well as the relative contribution of different mechanisms in inducing donor-specific unresponsiveness. Kidney allograft tolerance achieved through durable full-donor chimerism may be due to central deletion of graft-reactive donor T cells, even though mechanistic data from patient series are lacking. On the other hand, immune tolerance attained with transient mixed chimerism-based protocols initially relies on Treg-mediated suppression, followed by peripheral deletion of donor-reactive recipient T-cell clones under antigenic pressure from the graft. These conclusions were supported by data deriving from novel high-throughput T-cell receptor sequencing approaches that allowed tracking of alloreactive repertoires over time. In this review, we summarize the most important mechanistic studies on tolerance induction with combined kidney-bone marrow transplantation in humans, discussing open issues that still need to be addressed and focusing on techniques developed in recent years to efficiently monitor the alloresponse in tolerance trials. These cutting-edge methods will be instrumental for the development of immune tolerance protocols with improved efficacy and to identify patients amenable to safe immunosuppression withdrawal.

PREVENTION AND MANAGEMENT OF RELAPSE OF ACUTE LEUKEMIA AND MYELODYSPLASTIC SYNDROME AFTER ALLOGENEIC HEMATOPOIETIC CELL TRANSPLANTATION IN PEDIATRIC PATIENTS

The best way to manage acute leukemia relapse after HCT is to prevent it, buying time for GVL with immunomodulation and, if no GVHD between days +60 and + 90, prophylactic DLI can be indicate for very high or high risk patients. Short-term low dose of cyclosporine or methotrexate can add safety to pro-DLI, particularly after mismatched or unrelated transplantation. Maintenance with imatinib or dastinib, recommended for Ph-positive ALL, with sorafenib, for FLT3-ITD AML, or azacitidine, for myelodysplastic syndrome patients, can be effective in reducing relapse rates. However, target agent maintenance can add toxicity, depends on patient adherence and demands physician experience to know when is safe to start, how adjust the dose according individual tolerance after transplant and to detect undesirable drug interactions. The second step to avoid hematological relapse is preemptive approach guided by measurable residual disease or mixed chimerism. In patients off immunosuppression, chemotherapy followed by DLI is a useful strategy, and if no response, interferon alpha can be associated to enhance GVL. Target-specific agents can be start at this point either. After relapse, antigen-directed therapy with blinatumumab for CD19 ALL, inotuzumab for CD22 ALL are excellent options to induce MRD negativity and facilitate HCT. Disadvantages of new immunotherapies are: high incidence of VOD with inotuzumab and gemtuzumab; lower response in patients with high leukemia burden or concurrent extramedullary relapse; necessity of consolidation with HCT after a bridging therapy with BiTE and probably with CAR-T cell therapy also. It is important to realize that if remission after chemotherapy is associated with the development of GVHD, then there may be limited benefit (and possibly harm) in consolidating with any kind of cellular therapy. However, for patients who achieved remission without GVHD, either DLI or second transplant can be recommend. Further studies are necessary to determine at which point each strategy might yield the best results.

Alterations of Peripheral Blood T Cell Subsets following Donor Lymphocyte Infusion in Patients after Allogeneic Stem Cell Transplantation

Donor lymphocyte infusion (DLI) after allogeneic stem cell transplantation (alloSCT) is an established method to enhance the Graft-versus-Leukemia (GvL) effect. However, alterations of cellular subsets in the peripheral blood of DLI recipients have not been studied. We investigated the changes in lymphocyte subpopulations in 16 patients receiving DLI after successful alloSCT. Up to three DLIs were applied in escalating doses, prophylactically for relapse prevention in high-risk disease (n = 5), preemptively for mixed chimerism and/or a molecular relapse/persistence (n = 8), or as part of treatment for hematological relapse (n = 3). We used immunophenotyping to measure the absolute numbers of CD4+, CD8+, NK, and CD56+ T cells and their respective subsets in patients’ peripheral blood one day before DLI (d-1) and compared the results at day + 1 and + 7 post DLI to the values before DLI. After the administration of 1 × 106 CD3+ cells/kg body weight, we observed an overall increase in the CD8+ and CD56+ T cell counts. We determined significant changes between day − 1 compared to day + 1 and day + 7 in memory and activated CD8+ subsets and CD56+ T cells. Applying a higher dose of DLI (5 × 106 CD3+ cells/kg) led to a significant increase in the overall counts and subsets of CD8+, CD4+, and NK cells. In conclusion, serial immune phenotyping in the peripheral blood of DLI recipients revealed significant changes in immune effector cells, in particular for various CD8+ T cell subtypes, indicating proliferation and differentiation.

Spatial distribution of conspecific genotypes within chimeras of the branching coral Stylophora pistillata

Chimerism is a coalescence of conspecific genotypes. Although common in nature, fundamental knowledge, such as the spatial distribution of the genotypes within chimeras, is lacking. Hence, we investigated the spatial distribution of conspecific genotypes within the brooding coral Stylophora pistillata, a common species throughout the Indo-Pacific and Red Sea. From eight gravid colonies, we collected planula larvae that settled in aggregates, forming 2–3 partner chimeras. Coral chimeras grew in situ for up to 25 months. Nine chimeras (8 kin, 1 non-related genotypes) were sectioned into 7–17 fragments (6–26 polyps/fragment), and genotyped using eight microsatellite loci. The discrimination power of each microsatellite-locus was evaluated with 330 ‘artificial chimeras,’ made by mixing DNA from three different S. pistillata genotypes in pairwise combinations. In 68% of ‘artificial chimeras,’ the second genotype was detected if it constituted 5–30% of the chimera. Analyses of S. pistillata chimeras revealed that: (a) chimerism is a long-term state; (b) conspecifics were intermixed (not separate from one another); (c) disproportionate distribution of the conspecifics occurred; (d) cryptic chimerism (chimerism not detected via a given microsatellite) existed, alluding to the underestimation of chimerism in nature. Mixed chimerism may affect ecological/physiological outcomes for a chimera, especially in clonal organisms, and challenges the concept of individuality, affecting our understanding of the unit of selection.

Long-term results and GvHD after prophylactic and preemptive donor lymphocyte infusion after allogeneic stem cell transplantation for acute leukemia

We report on 318 patients with acute leukemia, receiving donor lymphocyte infusion (DLI) in complete hematologic remission (CHR) after allogeneic stem cell transplantation (alloSCT). DLI were applied preemptively (preDLI) for minimal residual disease (MRD, n = 23) or mixed chimerism (MC, n = 169), or as prophylaxis in high-risk patients with complete chimerism and molecular remission (proDLI, n = 126). Median interval from alloSCT to DLI1 was 176 days, median follow-up was 7.0 years. Five-year cumulative relapse incidence (CRI), non-relapse mortality (NRM), leukemia-free and overall survival (LFS/OS) of the entire cohort were 29.1%, 12.7%, 58.2%, and 64.3%. Cumulative incidences of acute graft-versus-host disease (aGvHD) grade II–IV°/chronic GvHD were 11.9%/31%. Nineteen patients (6%) died from DLI-induced GvHD. Age ≥60 years (p = 0.046), advanced stage at transplantation (p = 0.003), shorter interval from transplantation (p = 0.018), and prior aGvHD ≥II° (p = 0.036) were risk factors for DLI-induced GvHD. GvHD did not influence CRI, but was associated with NRM and lower LFS/OS. Efficacy of preDLI was demonstrated by decreasing MRD/increasing blood counts in 71%, and increasing chimerism in 70%. Five-year OS after preDLI for MRD/MC was 51%/68% among responders, and 37% among non-responders. The study describes response and outcome of DLI in CHR and helps to identify candidates without increased risk of severe GvHD.

Complex Clonal Evolution Can Occur Following Transplantation for Transfusion Dependent Thalassaemia in the Context of Mixed Myeloid Chimerism and Reduced Conditioning Regimens

Haemopoietic stem cell transplantation (HSCT) is a well-established treatment modality for the cure of transfusion dependent thalassaemia (TDT) and sickle cell disease (SCD). Clonal evolution has recently been identified as a concerning event in the setting of mixed chimerism and/or ineffective haemopoiesis following conventional bone marrow transplantation and gene therapy for haemoglobinopathies. This has so far been restricted to SCD only, with the presumption that despite both conditions sharing an ineffective erythropoietic marrow compartment, there may be inflammatory and hypoxic differences enabling clonal evolution, in addition to the different exposure to hydroxycarbamide as a therapeutic agent. However, there is a need to investigate whether this may also be an occurrence in TDT. From 2011 to 2021, over a ten-year period, sixty-five consecutive paediatric patients received a sibling HSCT (n=55) or a haploidentical HSCT (n=10) for TDT in our institution. Conditioning intensity was minimised at the start of this cohort in order to limit toxicity and late effects, abandoning the use of Bu/Cy, which resulted in approximately 50% of the patients having stable mixed chimerism long-term. Sibling HSCT was conditioned with fludarabine 160 mg/m 2, treosulfan 42 g/m 2, thiotepa 10 mg/kg and ATG (Thymoglobulin) or alemtuzumab, and received GvHD prophylaxis with ciclosporin and MMF. Haploidentical HSCT was conditioned with fludarabine 150 mg/m 2, cyclophosphamide 30 mg/kg, TBI 2 or 4 Gy, and ATG 4.5 mg/kg (thiotepa 10 mg/kg added if TBI 2 Gy only) with GvHD prophylaxis provided by two doses of post-transplantation cyclophosphamide 50 mg/kg, sirolimus and MMF. All patients had pre-transplantation endogenous haemopoiesis was suppressed pre-transplantation with hypertransfusions for a minimum of 8 weeks, and/or the use of hydroxycarbamide and azathioprine. GvHD prophylaxis was provided with ciclosporin and MMF. All patients were Pesaro class I or II at the time of transplantation. The median age was 5 years (2 - 19). The median survival was 26.8 months (2.6-101.8). The OS was 93.8% and DFS was 89.2%. Three patients in this cohort developed clonal evolution in the context of myeloid mixed chimerism identified due to the development of cytopenias and transfusion dependence. All patients had a complex karyotype and it involved deletion of chromosome 7: Patient 1 had a sibling BMT at 3 years and 2 months. Day +28 chimerism was >95% in whole blood and 95% in T cells. The myeloid fraction had a progressive reduction from day +60 onwards. At 15 months post-transplantation clonal evolution was identified [18% ring sideroblasts, 46;XY, del (7) (q22 q34) [5]/46;XY [5], chimerism 13% donor in whole marrow and 41% marrow T cells. Two months later he became red cell transfusion dependence. A second BMT with busulfan based conditioning resulted in long-term cure. Patient 2 had a sibling BMT at 3 years and 8 months of age. Day +28 chimerism was 99% donor in whole blood and 99% donor in T cells. Day +161 post-transplantation he started requiring erythropoietin support to maintain him transfusion independent and on day +217 we first identified in the bone marrow the appearance of myeloid mixed chimerism (75% donor in whole marrow and 91% marrow T cells) and complex clonal evolution involving -7 and FISH identified deletion of one copy of KMT2E (7q22) and MET (7q31) detected [16/200]. The patient is under monitoring at present with a progressive reduction of the size of the abnormal clones. Patient 3 had a haplo BMT at 5 years and 8 months. Day +28 chimerism was 97% donor in whole blood and 99% donor in T cells. She developed mixed myeloid chimerism following cessation of immunosuppression on day +206: 86% donor whole blood and 99% donor in T cells. At 22 months post-BMT she started to require transfusion and a complex clonal evolution involving deletion 7 in her bone marrow when the chimerism was 42% donor whole blood and 81% donor marrow T cells. FISH identified deletion of one copy of KMT2E(7q22) and MET(7q31) detected[22/100]. She is being prepared for a second bone marrow transplant. In conclusion, complex clonal evolution also occurs post HSCT in TDT, at least in the context of reduction of conditioning intensity and development of mixed myeloid chimerism. This finding warrants further investigation and may have significant implications for the design of both conventional HSCT and gene therapy strategies. Disclosures No relevant conflicts of interest to declare.

Peri-Transplant Alemtuzumab Levels Predict Risk of Secondary Graft Failure and Inversely Impact CXCL9 Levels after RIC HCT (A Correlative Biology Study to BMT-CTN 1204 RICHI)

Introduction The BMT-CTN 1204 study for Hemophagocytic Syndromes or Selected Primary Immune Deficiencies (NCT01998633) (RICHI) was a single arm study testing safety and efficacy of reduced intensity conditioning (RIC) with alemtuzumab (1mg/kg), fludarabine (150 mg/m2) and melphalan (140 mg/m2). Survival was favorable compared to historical studies, but patients experienced high rates of mixed chimerism (MC) and ultimate secondary graft failure (GF). Mechanisms for GF are not known. Expansion of recipient T cells and interferon-gamma pathway activation have been proposed as drivers for GF. However, high peri-transplant alemtuzumab levels have been associated with higher risk of MC and eventual secondary GF, suggesting an inverse relationship between GF and immune activation in the context of RIC. In order to elucidate mechanisms of GF for patients on the RICHI study, we systematically evaluated cytokine patterns and alemtuzumab levels and their association with durable engraftment. Methods Serial blood samples were collected, processed, and stored for consenting patients at day -14 (window: day -28 to -14), day -7 (+/- 1 day), day -1 (+/- 1), day +1 (+1 to +3), day +14 (+/- 2), day +28 (+/- 2), day +42 (+/- 3), day +70 (+/- 10), and day +100 (+/- 10). Alemtuzumab levels were measured using a flow cytometric assay as previously described. Comprehensive cytokine analysis was performed for over 100 analytes using the MagPix platform. Primary GF was defined as donor chimerism <5% prior to day +42. Secondary GF was defined as donor chimerism <5% after initial engraftment and/or requirement of donor lymphocyte infusion (DLI) or second HCT (with or without conditioning) to manage MC or graft loss. Mixed chimerism (MC) was defined as donor chimerism <95% on at least one occasion. Results Thirty-three patients were included in this study with HLH (n=25), CAEBV (n=3), CGD (n=2), HIGM (n=2), and IPEX (n=1). All patients received bone marrow grafts and 27 (82%) patients had unrelated donors. Twenty-one grafts were 8/8 or 6/6 HLA-matched (64%) and 12 grafts were 7/8 HLA-matched (36%). Among all patients, 1 patient (3%) developed primary GF, 22 (67%) developed mixed chimerism (MC), and 11 patients (33%) developed secondary GF. Survival with sustained engraftment without DLI or second HCT was 40.0%. We first evaluated peripheral blood levels of 100+ cytokines. Analysis revealed significant differences between patients with and without GF as shown in Figure 1A. Notably, on day +14 and +28, patients with secondary GF had significantly lower CXCL9 levels than those without GF. We then estimated the cumulative incidence (CI) of secondary GF among patients with CXCL9 levels above and below the day +14 median level of 2394pg/mL. The CI of secondary GF in patients with a day +14 CXCL9 level ≤2394pg/mL was 73.6% vs 0% in patients with a level >2394pg/mL (p=0.002). The CI of secondary GF in patients with a day +28 CXCL9 level ≤2867pg/mL (day +28 median) was 64.3%, vs 0% in patients with levels >2867pg/mL (p=0.004). We then sought to correlate CXCL9 levels with alemtuzumab exposure, as high alemtuzumab levels would result in more efficient T cell depletion of donor grafts that could lead to lower CXCL9 levels. Indeed, CXCL9 levels inversely correlated with day 0 alemtuzumab levels. Patients with day 0 alemtuzumab levels >0.32µg/mL had lower CXCL9 levels compared to patients with levels ≤0.32µg/mL (Figure 1B). Finally, we examined the impact of alemtuzumab levels on MC and secondary GF. Patients with day 0 alemtuzumab levels ≤0.32µg/mL had a lower CI of MC compared to patients with levels >0.32µg/mL, 14.3% vs 90.9%, respectively (p=0.03). The CI of secondary GF was 0% in patients with day 0 alemtuzumab levels ≤0.32µg/mL compared to 54.3% in patients with levels >0.32µg/mL (p=0.08). Conclusions This study demonstrates a strong relationship between alemtuzumab levels and durable engraftment. Further, interferon gamma activity, as reflected by CXCL9, inversely correlates with peri-transplant alemtuzumab levels in this prospective cohort treated with RIC. Our findings support the paradigm that higher alemtuzumab levels drive efficient T cell depletion of the stem cell product which increases the risk of MC and secondary GF, suggesting that donor T cells promote engraftment via a graft versus hematopoiesis function. Precision alemtuzumab dosing strategies may offer an opportunity to improve outcomes for patients who undergo RIC HCT. Figure 1 Figure 1. Disclosures Pulsipher: Adaptive: Research Funding; Equillium: Membership on an entity's Board of Directors or advisory committees; Jasper Therapeutics: Honoraria. Bollard: Neximmune: Current equity holder in publicly-traded company; Catamaran Bio: Membership on an entity's Board of Directors or advisory committees; Cabaletta Bio: Current equity holder in publicly-traded company, Membership on an entity's Board of Directors or advisory committees; Mana Therapeutics: Current equity holder in publicly-traded company, Membership on an entity's Board of Directors or advisory committees, Patents & Royalties; Cellectis: Honoraria, Membership on an entity's Board of Directors or advisory committees; Repertoire Immune Medicines: Current equity holder in publicly-traded company; ROCHE: Consultancy, Honoraria; SOBI: Honoraria, Membership on an entity's Board of Directors or advisory committees. Kean: Regeneron: Research Funding; Bristol Myers Squibb: Patents & Royalties: From clinical trial data, Research Funding; Bluebird Bio: Research Funding; Gilead: Research Funding; Vertex: Consultancy; Novartis: Consultancy; EMD Serono: Consultancy. Jordan: Sobi: Consultancy. Allen: Sobi: Consultancy. OffLabel Disclosure: Alemtuzumab, humanized monoclonal antibody against CD52, used as part of allogeneic HCT conditioning

Engraftment Kinetics and Recipient Chimerism Increase to Predict Leukemia Relapse By Ptcy and Non-Ptcy Transplant

Recipient chimerism increase has been used to predict leukemia relapse in post-hematopoietic cell transplant (HCT) patients with conventional GVHD prophylaxis. However, the value of recipient chimerism increase in patients with post-transplant cyclophosphamide (PTCy) is not clear. We compared PTCy to conventional GVHD prophylaxis (non-PTCy) patients regarding engraftment kinetics and the clinical significance of the 2 chimerism parameters, increasing mixed chimerism (IMC) and degree of recipient chimerism increase at the first event (Δ increase). We studied both total and T-cell-specific chimerism. While leukemia relapse is manifested by an increase in total cell recipient chimerism, an increase in T-cell-specific recipient chimerism may be more impactful in predicting relapse because of the effect of increased T-cell-specific chimerism on the graft-versus-leukemia effect. A total of 220 patients (PTCy: 44, non-PTCy: 176) with AML, MDS, and ALL underwent HCT at our institution from January 2014 to September 2020 and were included in this study (Table). Chimerism was tested at least monthly for the first 3 months, followed by every 3 months up until 1-year post-HCT, and then every 6-12 months thereafter. Short tandem repeat or quantitative PCR were used when percent recipient chimerism was ≥5% and <5%, respectively. Cumulative incidence of competing events and Gray's test were applied for engraftment analysis. Relapse and non-relapse mortality were considered as competing risks for engraftment. Mantel-Byar test and Simon-Makuch plot with landmark analysis were used to visualize disease-free survival (DFS) curves. The Cox proportional hazards model with time-dependent covariates was performed to identify the factors affecting relapse. PTCy patients achieved complete donor chimerism (CC) in total cells earlier at a deeper level (>99%) as compared to non-PTCy patients. Deeper total cell CC (>99%) was achieved in 79.5% of PTCy vs. 51% of non-PTCy patients at day 250, while CC (>95%) was achieved in almost 90% of patients in both groups within 100 days (Figure 1A and B). In comparison, the percentage of PTCy patients achieving T-cell-specific CC was significantly higher at day 250 post-HCT: CC (>95%/>99%) was 79.7%/68.4% in PTCy patients vs. 56.1%/37.5% in non-PTCy patients (Figure 1C and D). To evaluate their impact in predicting relapse, IMC was stratified into no IMC, 1 IMC (≥1 nonconsecutive IMC), and 2 IMC (≥2 consecutive IMC), and degree of recipient chimerism increase at the first event (Δ increase) was stratified into <0.1% (no Δ increase), 0.1-1%, and ≥1%. Two IMC (total), 1 IMC (T-cell), and 2 IMC (T-cell) groups were associated with shorter DFS in non-PTCy patients but not in PTCy patients (Figure 2). One and 2 IMC groups (both total and T-cell) were associated with relapse risk in non-PTCy patients. Furthermore, 1 IMC (T-cell) in non-PTCy patients showed a strong association in relapse risk (HR 7.0 (95%CI 2.83-17.8) p<0.0001). Δ increase ≥1% (total and T-cell) and Δ increase ≥0.1% (T-cell) were associated with shorter DFS in non-PTCy patients, while only Δ increase ≥1% (T-cell) only showed a trend towards shorter DFS in PTCy patients (Figure 3). The Cox regression model showed Δ increase ≥1% in both total, and T-cell chimerism was associated with relapse risk in non-PTCy patients (HR 6.4 (95%CI 2.9-14.2) p<0.0001 and HR 7.2 (95%CI 2.9-18.1) p<0.0001, respectively). Δ increase ≥0.1% (T-cell) in non-PTCy patients was also associated with relapse risk (HR 7.2: 95%CI 2.5-20.4, p<0.0001). In comparison, no association was found between Δ increase and relapse risk in PTCy patients. This is one of the most extensive studies investigating engraftment kinetics and the association of total and T-cell recipient chimerism increase to predict leukemia relapse in PTCy and non-PTCy HCT recipients. We found that PTCy HCT recipients achieved deeper engraftment earlier as compared to non-PTCy recipients. In addition, the two chimerism parameters (IMC and Δ increase) are less reliable in predicting relapse in PTCy than non-PTCy recipients. However, other factors, such as disease type, conditioning regimen, and donor HLA disparity, may have affected engraftment kinetics and the significance of chimerism parameters. Further investigations are warranted to elucidate the impact of the engraftment kinetics and recipient chimerism increase to predict relapse, especially in the PTCy setting. Figure 1 Figure 1. Disclosures Rakszawski: SeaGen: Speakers Bureau. Naik: Takeda: Other: Virtual Advisory Board Member ; Sanofi: Other: Virtual Advisory Board Member ; Kite: Other: Virtual Advisory Board Member. Rybka: Spark Therapeutics: Consultancy; Merck: Consultancy. Claxton: Astellas: Other: Clinical Trial; Novartis: Research Funding; Astex: Research Funding; Cyclacel: Research Funding; Daiichi Sankyo: Research Funding; Incyte: Research Funding.

Chimerism of Blood T-Cells and Leukemia Lineage Cells Following Myeloablative Hematopoietic Cell Transplantation Is a Risk Factor for Relapse and Chronic Graft-Versus-Host Disease

Background: Mixed chimerism of blood leukemia lineage cells has been reported to be highly predictive of relapse (Mattsson J et al. Leukemia 2001), but the utility of T-cell chimerism is controversial. Methods: Chimerism of T-cells (CD3 +) and leukemic lineage cells (CD13 +/CD33 + for myeloid malignancies and CD19 + for B-lymphoid malignancies) was measured in the peripheral blood for 600 hematopoietic cell transplant (HCT) recipients at 3-months post-transplant. Conditioning was myeloablative (fludarabine+busulfan+4GyTBI) and GVHD prophylaxis was with ATG+CsA+MTX. Results: Mixed (<95% donor) chimerism was present in 6% of patients in the leukemic lineage and 16% of patients in T-cells. Compared to patients with complete chimerism (≥95%), mixed chimerism predicted a significantly greater incidence of relapse (52% vs. 27%, P=0.044), and surprisingly, also a greater incidence of cGVHD (43% vs. 23%, P=0.028). Patients with mixed leukemic lineage chimerism also had poorer cGRFS (7% vs. 39%, P<0.001) and OS (29% vs. 55%, P<0.001). Mixed T-cell chimerism predicted a significantly greater incidence of relapse (46% vs. 24%, P<0.001) and lower cGVHD incidence (9% vs. 31%), with no difference in cGRFS or OS. The sensitivity/specificity/positive predictive value (PPV)/negative predictive value (NPV) of mixed leukemic lineage and T-cell chimerism for relapse were 65%/55%/35%/81% and 66%/57%/60%/79%, respectively. Sensitivity/specificity/PPV/NPV was similarly poor for cGVHD. In patients with complete leukemic lineage chimerism at 3-months, a ≥5% drop in donor chimerism at any future timepoint had sensitivity/specificity/PPV/NPV of 82%/79%/79%/82% for subsequent relapse. Conclusion: In the setting of myeloablative conditioning and ATG-based GVHD prophylaxis, mixed chimerism at 3-months post-transplant is a risk factor for subsequent relapse. However, the utility of these measurements in guiding medical interventions may be limited by insufficient predictive values. Nevertheless, patients with mixed chimerism may be candidates for more intensive surveillance. Disclosures Storek: Atara Biotherapeutics: Other: Site PI, Research Funding.