Peripheral Blood CD34+ Donor Chimerism Is Superior to CD3+ Donor Chimerism for Predicting Relapse Following Allogeneic Stem Cell Transplantation for Myeloid Malignancies

Background: Disease relapse remains the main cause of mortality after allogeneic stem cell transplantation (allo-SCT) for patients with myeloid malignancies. Loss of donor chimerism (DC) is commonly used as a biomarker of impending relapse, which allows the initiation of pre-emptive therapies such as withdrawal of immunosuppression or donor lymphocyte infusion (DLI). Surprisingly, there are few if any direct comparisons of peripheral blood CD34+ and CD3+ DC as biomarkers to predict relapse. We hypothesized that loss of CD34+ DC may be a more direct measure of impending relapse given most myeloid malignancies express CD34. Methods: We prospectively measured peripheral blood CD34+ and CD3+ DC on days 30, 60, 90, 120 and 180 following allo-SCT for patients with AML (n=113) or MDS (n=23) transplanted at a single centre between July 2011 and November 2019. Chimerism analysis was performed using purified cell subsets isolated from 60 mL peripheral blood using PCR-based amplification of short tandem repeats (STRs). The goal of this retrospective analysis was to compare the value of CD3+ and CD34+ DC for predicting relapse. Institutional practice for CD34+ DC below 80% included a bone marrow biopsy to identify morphologic relapse and donor lymphocyte infusion. Statistical analysis was performed with R 3.5.2 (The R project for Statistical computing) or GraphPad (v8.2.0). Results: Overall, 41 of 136 (30%) patients had morphologic relapse at a median time of 153 days after allo-SCT (range 50-1742). The sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of CD3+ and CD34+ donor chimerism for morphologic relapse is shown in Figure 1A. CD34+ DC outperformed CD3+ DC in all criteria, irrespective of the percentage chosen. Furthermore, a concurrent reduction in CD34+ DC was seen in almost all patients (14/16 at the 80% level) that had a fall in CD3+ DC. A fall in CD34+ DC below 80% was highly predictive for relapse, with a 5 year relapse free survival of only 17% compared with 80% for those that maintained DC > 80% for the first 180 days (Figure 1B). To determine the clinical utility of CD34+ DC, we measured the time to relapse in those patients that had a fall in DC without morphologic relapse at the time of DC measurement (Figure 1A). Based upon our institutional trigger of CD34+ DC < 80%, the median time to relapse was 49 days in 22 patients. In contrast, a CD34+ DC < 90% had a much longer time to relapse (71 days). However, this longer lead time would come at the price of unnecessary intervention in almost half of all patients as the PPV for CD34+ DC < 90% was only 55%. Loss of CD3+ DC had the longest lead time (> 70 days), however it was only useful up to 10 (24%) of all relapses. Finally, DLI administered for CD34+ DC < 80% maintained durable remission (> 12 months) in only 2 of 19 patients. Conclusion: This is the largest comparison of peripheral blood CD34+ and CD3+ DC following allo-SCT. Our results show that monthly monitoring of CD34+ DC in the first 6 months after allo-SCT is a more useful biomarker than CD3+ DC for predicting relapse in patients allografted for AML or MDS. The level of CD34+ DC chosen to trigger intervention should be guided by characteristics of the intervention such as toxicity and expected response time. Given the relative ineffectiveness of DLI for CD34+ < 80%, we suggest that 90% may provide greater time for immunologic responses. Figure Legend (A) Characteristics of different levels of CD3 and CD34 DC at any time in the first 180 days post-allo-SCT. Number of patients that fulfil the level in the total cohort of 136 patients. Number of patients before relapse indicates the number that do not have morphologic relapse at the time of DC measurement. Median days before relapse is shown for those patients. (B) Relapse free survival of patients according to CD34+ DC > 80% (blue line) or < 80% (yellow line) in the first 180 days. Disclosures Spencer: AbbVie, Celgene, Haemalogix, Janssen, Sanofi, SecuraBio, Specialised Therapeutics Australia, Servier and Takeda: Consultancy; Celgene, Janssen and Takeda: Speakers Bureau; Amgen, Celgene, Haemalogix, Janssen, Servier and Takeda: Research Funding; AbbVie, Amgen, Celgene, Haemalogix, Janssen, Sanofi, SecuraBio, Specialised Therapeutics Australia, Servier and Takeda: Honoraria. Wei:Celgene: Consultancy, Honoraria, Research Funding; Janssen: Consultancy, Honoraria; Astellas: Consultancy, Honoraria; Novartis: Consultancy, Honoraria, Research Funding; Amgen: Consultancy, Honoraria, Research Funding; MacroGenics: Consultancy, Honoraria; Servier: Consultancy, Honoraria, Research Funding; Walter and Eliza Hall Institute: Other: former employee and receives a fraction of its royalty stream related to venetoclax; Pfizer: Honoraria; Genentech: Honoraria; Astra Zeneca: Honoraria, Research Funding; AbbVie Inc.: Consultancy, Honoraria, Research Funding.