Otteson, Gregory E.

Abstract Background: Flow cytometry (FC) is widely utilized for the identification of measurable residual disease (MRD) in multiple myeloma (MM), providing an essential laboratory indicator for prognostic assessment and therapeutic management. Unfortunately, the high level of expertise required, long manual analysis time, and complexity of FC data files currently limit the offering of this testing to selected reference laboratories. Methods: We developed an artificial intelligence (AI) pipeline to automatically enhance unaltered FC files, minimizing the expertise and manual analysis time needed to detect MM MRD using the 2-tube Euroflow Consortium Panel (Cytognos). Raw FC files corresponding to 150 MRD-positive bone marrow samples (MRD < 5%, median = 0.0022%), and 39 MRD-negative samples were processed through our CCADDAS (Clustering and Classification of All events, Dimensionality reduction, Downsampling and Aberrancy Scaling) pipeline, deployed on a custom clinical-grade cloud environment (Biolegend). Automated processing steps included elimination of acquisition errors (FlowCut), state-of-the-art clustering (PARC), cluster-informed and density-based downsampling with preservation of low-event subsets, dimensionality reduction (UMAP), and cluster-based anomaly detection compared to negative controls (29 bone marrow aspirates). In addition, a deep neural network (DNN) trained on expert-defined subpopulations from negative controls was included for automatic gating of normal subsets (Tensorflow v2). AI-enhanced data files were analyzed using a general-purpose flow cytometry analysis software (Kaluza v 2.3, Beckman Coulter), and %MRD estimates were compared to reported results based on traditional expert analysis of original FC files. Results: Cluster-informed downsampling reduced the number of cells to be manually analyzed from 4.1 million to 196,761 cells per tube, on average (95% cellularity reduction); resulting in a smaller FC data file (from 203.2 MB to 20.5 MB; 90% data reduction). Importantly, low-level MRD events were adequately preserved after downsampling (median 88% and 77% retention, for 0.01-0.001% MRD and < 0.001% MRD, respectively). True number of events were accurately estimated on gated subsets using an “upsampling factor” parameter [observed # events x mean (upsampling factor)]. Gating of normal subsets was completely automated using a DNN classifier parameter. In 96% of positive cases, MRD was rapidly identified as immunoglobulin light chain-restricted plasma cells on a single pre-calculated UMAP plot (tube 2), after dual coloring for kappa and lambda. On the remainder 4% of positive cases, MRD was easily segregated from benign plasma cells using the AI-generated “aberrancy scale” parameter or other strategies. The use of AI-enhanced files reduced manual analysis time from 15.05 minutes (SD = 6.84) to 1.29 minutes (SD = 0.90) per case, on average (p < 0.0001) (91% reduction of manual analysis time). MM MRD was detected in all positive cases above limit of detection for our assay (≥0.0002%), with excellent quantitative correlation with conventional analysis on original FC files (linear regression slope = 0.9841, P < 0.0001). Conclusion: We introduce a largely unsupervised AI pipeline that transforms raw MM MRD FC data from the Euroflow Consortium panel into a markedly smaller, AI-annotated and software-agnostic FC file, including comparison to normal controls. This CCADDAS pipeline simplifies and accelerates detection of MM MRD in clinical diagnostics, reducing the number of cells analyzed by 95% and manual analysis time by 91%, without impacting test performance. Moreover, CCADDAS can be utilized for any laboratory-developed MM MRD assay, and its small-sized export is compatible with any clinical FC software, computing platform, and analysis strategy. Adoption of CCADDAS is likely to facilitate the implementation of MM MRD FC analysis by more clinical laboratories.

Abstract Measurable residual disease (MRD) assessment by flow cytometry (FC) plays an essential role in prognosis and therapy escalation of B-cell acute lymphoblastic leukemia (B-ALL). However, the high degree of expertise and manual analysis time required limits the availability of this assay. To overcome this limitation, we developed a data-enhancing artificial intelligence (AI) pipeline that accelerates and simplifies MRD analysis. Unaltered FC files from 171 B-ALL MRD-positive and 89 MRD-negative cases were processed through an AI pipeline trained with 31 expert-gated negative controls. Cluster-informed downsampling reduced FC files from 1.2 million to 155 884 cells per case, on average, (87% cellularity reduction), whereas preserving small MRD populations (median, 100% retention for MRD of <1%) and allowing for true percentage MRD estimates using a correction factor. A deep neural network cell classifier automatically identified normal hematopoietic subsets (macro-averaged F1 score of 0.86); and an AI measure of anomaly discriminated B-ALL from benign mononuclear (area under the curve [AUC] of 0.98) or B-lymphoid cells (AUC of 0.94). Manual analysis of AI-enhanced files was completed in only 1.01 minutes per case, on average (standard deviation of ±0.57); with 100% positive agreement with conventional analysis (for MRD of ≥0.01%), 100% negative agreement, and excellent quantitative correlation (R2 = 0.92). Our cloud-based AI-enhancement solution accelerates B-ALL MRD identification without compromising test performance and has the potential of facilitating B-ALL MRD analysis by more clinical laboratories.

Abstract Monoclonal antibodies targeting CD38 are a therapeutic mainstay in multiple myeloma (MM). Although they have contributed to improved outcomes, most patients still experience disease relapse, and little is known about tumor-intrinsic mechanisms of resistance to these drugs. Antigen escape has been implicated as a mechanism of tumor-cell evasion in immunotherapy. Yet, it is unknown whether MM cells can develop permanent resistance to anti-CD38 antibodies by acquiring genomic events leading to biallelic disruption of the CD38 gene locus. Here, we analyzed whole-genome and whole-exome sequencing data from patients 701 newly diagnosed MM, 67 patients at relapse with naivety to anti-CD38 antibodies, and 50 patients collected at relapse after anti-CD38 antibodies. We report a loss of CD38 in 10 of 50 patients (20%) after CD38 therapy, 3 of whom exhibited a loss of both copies. Two of these cases showed convergent evolution in which distinct subclones independently acquired similar advantageous variants. Functional studies on missense mutations involved in biallelic CD38 events revealed that 2 variants, L153H and C275Y, decreased binding affinity and antibody-dependent cellular cytotoxicity of the commercial antibodies daratumumab and isatuximab. However, a third mutation, R140G, conferred selective resistance to daratumumab, while retaining sensitivity to isatuximab. Clinically, patients with MM are often rechallenged with CD38 antibodies after disease progression and these data suggest that next-generation sequencing may play a role in subsequent treatment selection for a subset of patients.

Background: Flow cytometric detection of minimal residual disease (MRD) in chronic lymphocytic leukemia (CLL) is complex, time-consuming, and subject to inter-operator variability. Deep neural networks (DNNs) offer potential for standardization and efficiency improvement, but require rigorous validation and monitoring for safe clinical implementation. Methods: We evaluated a DNN-assisted human-in-the-loop approach for CLL MRD detection. Initial validation included method comparison against manual analysis (n = 240), precision studies, and analytical sensitivity verification. Post-implementation monitoring comprised four components: daily electronic quality control, input data drift detection, error analysis, and attribute acceptance sampling. Laboratory efficiency was assessed through a timing study of 161 cases analyzed by five technologists. Results: Method comparison demonstrated 97.5% concordance with manual analysis for qualitative classification (sensitivity 100%, specificity 95%) and excellent correlation for quantitative assessment (r = 0.99, Deming slope = 0.99). Precision studies confirmed high repeatability and within-laboratory precision across multiple operators. Analytical sensitivity was verified at 0.002% MRD. Post-implementation monitoring identified 2.97% of cases (26/874) with input data drift, primarily high-burden CLL and non-CLL neoplasms. Error analysis showed the DNN alone achieved 97% sensitivity compared to human-in-the-loop-reviewed results, with 13 missed cases (1.5%) showing atypical immunophenotypes. Attribute acceptance sampling confirmed 98.8% of reported negative cases were true negatives. The DNN-assisted workflow reduced average analysis time by 60.3% compared to manual analysis (4.2 ± 2.3 vs. 10.5 ± 5.8 min). Conclusions: The implementation of a DNN-assisted approach for CLL MRD detection in a clinical laboratory provides diagnostic performance equivalent to expert manual analysis while substantially reducing analysis time. Comprehensive performance monitoring ensures ongoing safety and effectiveness in routine clinical practice. This approach provides a model for responsible AI integration in clinical laboratories, balancing automation benefits with expert oversight.

Background/Objectives: Acute myeloid leukemia (AML) is a biologically heterogeneous malignancy with a variable prognosis. Despite many patients achieving complete remission, relapse remains common, underscoring the need for effective prognostic markers. Measurable residual disease (MRD) has emerged as a critical prognostic indicator, associated with higher relapse risk and shorter survival. This study reports on our initial experience of MRD detection by flow cytometry in 385 bone marrow samples from 126 AML patients. Methods: The flow cytometry MRD assay, validated according to stringent consensus recommendations, consists of a 3-tube, 10-color panel incorporating a broad spectrum of lineage differentiation markers. Analytical specificity, sensitivity, precision, and reproducibility were evaluated, demonstrating the assay’s robustness. Results: The results reveal distinct immunophenotypic aberrancies in all AML cases, with consistent identification of aberrant immunophenotypes in follow-up specimens. AML MRD was detected in 32 out of 126 patients (25%) and in 77 out of 385 analyses (20%), with a median aberrant blast percentage of 1.87% (range, 0.01–12). A change in immunophenotype was documented in 21% of the MRD-positive cases. MRD positivity detected in the first sample studied was associated with reduced overall survival (HR: 5.153; p < 0.0001). Conclusions: Our findings support the integration of flow cytometric MRD analysis into routine clinical practice to enhance risk stratification and treatment planning for AML patients, as currently recommended by professional guidelines.

Introduction: Myeloma response to treatment can be biochemical, pathological, or radiological, with pathological assessments being invasive. Next-generation blood flow (PBF) for assessment of circulating tumor plasma cells (CTCs) has shown promising results as a prognostic in Multiple Myeloma (MM) patients. CTC detection via PBF has a potential to serve as non-invasive method for assessing pathological burden of disease and response to treatment. Comparing CTC responses to the International Myeloma Working Group (IMWG) criteria will further help to understand PBF's utility in predicting response to therapy. Patients and Methods: The study conducted is a retrospective observational study. We included patients with MM treated at Mayo Clinic from 01/01/2011 to 12/31/2019 with CTC records available at the time of best response (BR) to frontline therapy. The CTC values were expressed as the number of plasma cell events per 150,000 Peripheral Blood Mononuclear Cells (CD38, CD138, CD19, CD45, Kappa, Lambda). The CTC levels were compared across all response(IMWG Criteria) groups at BR, and at baseline. The CTC response was expressed as % CTC decrease at BR from baseline. Time-to-event analyses utilized the Kaplan Meier method and Cox-proportional hazards models, with comparisons made using the log-rank test. Results: We identified 621 patients with CTC data at the time of BR to frontline therapy. Of these, 148 (24%) had baseline CTC data from diagnosis. The median age at diagnosis was 62.3 (27.2-83.7). The median follow-up was 66.15 months; the distribution of patients by treatment class is as follows- Alkylator: 0.8%, PI(Proteasome Inhibitor)+Cyclophosphamide: 27.4%, IMid: 12.7%, PI+IMid(Immunomodulator): 55.9%, PI+IMid+antiCD38: 3.2%; 378 patients (78.8%) underwent an autologous transplant. BR to initial therapy was sCR/stringent Complete Response (10, 1.6%), CR (230, 36.9%), VGPR/Very Good Partial Response (224, 36.1%), PR (123, 19.8%), MR/Minimal Response (12, 1.9%), SD/Stable Disease (14, 2.3%), PD/Progressive Disease (8, 1.3%). In the study, 64 patients (10.31%) had detectable CTCs at their BR. The distribution of CTCs across the different response categories was as follows: none (0%) in sCR, 3 patients (1.3%) in CR, 18 patients (8%) in VGPR, 24 patients (19.5%) in PR, 8 patients (66.7%) in MR, 6 patients (42.9%) in SD, and 5 patients (62.5%) in PD. The median CTC counts and their interquartile ranges for these categories showed significant variability: CR had a median of 46 (IQR: 28.5, 923), VGPR had 39.5 (IQR: 15.5, 111.25), PR had 47 (IQR: 22.75, 82.5), MR had 55 (IQR: 28.5, 85), SD had 96.5 (IQR: 22.25, 218.75), and PD had 317 (IQR: 57, 415). This Analysis of Variance (ANOVA) for response-wise distribution of CTC was &lt;0.01. Response was dichotomized as PR and above for sCR, CR, VGPR, PR and less than PR for MR, SD and PD. Median Overall Survival(OS) for PR and above was 127.8 months (vs less than PR 56.7 months, p=0.01). Median OS for CTC+ at BR was 56.7 months (vs 134.8 for CTC negative, p=0.01). On multivariate analysis, presence of CTC at best response (Hazards Ratio = 2.9, n=1.1-2.9), RISS staging at diagnosis (Hazards Ratio = 2.35, n=1.5-3.05), less than PR (Hazards Ratio =1.8, 1.1-2.9) were independent predictors of inferior OS. In patients with data for CTC available at both baseline and at BR to frontline therapy (n=148), 110 had CTC detectable at diagnosis. Among patients in various response categories, a complete disappearance of CTCs was observed in sCR (n=3, 75.0%), CR (n=36, 69.2%), VGPR (n=38, 64.4%), and PR (n=14, 60.9%). CTC responses greater than 90% were noted in VGPR (n=3, 5.1%), PR (n=2, 8.7%), and SD (n=2, 40.0%). No CTC response was reported in sCR (1, 25.0%), CR (16, 30.8%), VGPR (2, 3.4% and 13, 22.0%), PR (3, 13.0% and both 2, 8.7%), SD (1, 20.0% and both 2, 40.0%), MR (1, 50.0%), and PD (1, 33.3% and 1, 50.0%, respectively). The median OS for those with a CTC response was 112.4 months vs 28.2 months for no response. (p=0.01). Conclusion: The presence of CTC at best response is an independent prognostic marker for inferior OS. Depth of response to frontline therapy correspond to a high rate of clearance of CTCs. Presence of CTC at best response provides independent prognostic value beyond the depth of response. A complete clearance of CTC at best response is associated with improved OS. CTC levels have not been corrected for WBC counts and may not be applicable in all settings.

Background: Flow cytometry (FC) is widely utilized for the identification of minimal residual disease (MRD) in B acute lymphoblastic leukemia (BALL), providing an essential laboratory indicator for prognostic assessment and therapeutic management. Unfortunately, the high level of expertise required, long manual analysis time, and complexity of FC data files limit the offering of this testing to selected reference laboratories. Methods: We developed an artificial intelligence (AI) pipeline to automatically enhance unaltered FC files, minimizing the expertise and manual analysis time needed to detect BALL-MRD on a single-tube 10-color panel (CD10, CD19, CD20, CD22, CD24, CD34, CD38, CD45, CD58, CD66c). Raw FC files corresponding to peripheral blood or bone marrow specimens from 171 BALL MRD positive cases (MRD &lt; 5%, median = 0.11%, 107 patients), and 89 MRD-negative cases (63 patients) were processed through our CCADDAS (Clustering and Classification of All events, Dimensionality reduction, Downsampling and Aberrancy Scaling) pipeline on a cloud environment (Google Vertex AI). Automated processing steps included elimination of acquisition errors (FlowCut), state-of-the-art clustering (PARC), dimensionality reduction (UMAP), cluster-based anomaly detection compared to negative controls (24 bone marrow aspirates and 7 peripheral bloods), and cluster-informed downsampling with preservation of low-event subsets. In addition, a deep neural network (DNN) trained on expert-defined subpopulations from negative controls was included for automatic gating of normal subsets (Tensorflow v2). AI-enhanced data files were analyzed using a general purpose flow cytometry analysis software (Kaluza v 3.5, Beckman Coulter), and %MRD estimates compared to reported results based on expert analysis of original FC files. Results: Cluster-informed downsampling reduced the number of cells per case to be manually analyzed from 1.2 million to 155,884 cells on average (87% cellularity reduction); resulting in a smaller FC data file (from 64.5 MB to 13.7 MB; 79% data reduction). Importantly, low-level MRD events were adequately preserved after downsampling (median 100% retention for both MRD&lt;0.01% and MRD 0.01-1%). True number of events were accurately estimated on gated subsets using an “upsampling factor” parameter [observed # events x mean (upsampling factor)]. Gating of normal subsets was completely automated using a DNN classifier parameter. Identification of MRD was aided by an AI-generated “aberrancy scale” parameter that discriminated BALL (median: 991) from both benign mononuclear cells (median: 486, p &lt; 0.0001) and background B-lymphoid cells (median: 379, p &lt; 0.0001), with a performance (AUC: 0.98 and 0.94 for mononuclear and B-lymphoid cells, respectively) superior to CD20 (0.54 and 0.69), CD38 (0.60 and 0.81), CD45 (0.75 and 0.74), CD58 (0.52 and 0.89), CD66c (0.56 and 0.72) and all other antigens studied. The use of AI-enhanced files reduced manual analysis time from 14.16 minutes (SD = 3.7) to 1.05 minutes (SD = 0.6) per case, on average (p &lt; 0.0001) (93% reduction). BALL MRD was detected in all positive cases above clinically relevant threshold (≥0.01%, as recommended by NCCN guidelines), and in 28 of 32 (88%) cases below this threshold, with excellent quantitative correlation with conventional analysis on original FC files (linear regression slope = 1.18, intercept = 0.04%, R squared = 0.92, P &lt; 0.0001). Conclusion: We introduce a largely unsupervised AI pipeline that transforms raw BALL-MRD FC data into a markedly smaller, AI-annotated and software-agnostic FC file, including comparison to normal controls. This CCADDAS pipeline simplifies and accelerates detection of BALL-MRD in clinical diagnostics, reducing number of cells analyzed by 87% and manual analysis time by 93%, without impacting test performance. Moreover, CCADDAS can be utilized for any labroatory-developed BALL-MRD assay, and its small-sized export is compatible with any clinical FC software, computing platform and analysis strategy. Adoption of CCADDAS is likely to facilitate the implementation of BALL-MRD FC analysis by more clinical laboratories.

Introduction: The current definition of plasma cell leukemia (PCL) is based on peripheral blood smear (PBS) demonstrating ≥5% plasma cells. However, PCL cases are enriched in translocation t(11;14), which often has a lympho-plasmacytoid appearance. Detection and quantification of circulating tumor plasma cells (CTC) through peripheral blood flow cytometry (PBF) has emerged as a promising approach for assessing disease burden and prognosis in multiple myeloma (MM) patients. While PBF is more sensitive, the potential threshold values on PBF that correspond with the PBS detection of circulating plasma cells (PBS-PC) remain undetermined. It is important to understand this comparison better to categorize PCL based on PBF. Patients and Methods: This is a retrospective cross-sectional diagnostic accuracy study. We reviewed results of patients diagnosed with plasma cell dyscrasia at the Mayo Clinic from January 2011 to December 2020. The presence of circulating tumor plasma cells (CTC) in encounters was evaluated using peripheral blood flow cytometry (PBF) and peripheral blood smear (PBS) for plasma cells (PBS-PC). PBS-PC was measured as a percentage of plasma cells (PC) in a 100-cell manual differential smear count, and PBF was expressed as the number of plasma cell events per 150,000 peripheral blood mononuclear cells (PBMCs) (Ficoll-isolated peripheral blood mononuclear cells using antibodies to CD38, CD138, CD19, CD45, kappa, lambda, sensitivity (Sn)=6.7x10-5). The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of PBS with respect to PBF were calculated. For threshold analysis, four groups were defined: A = absent PBS-PC, C = ≥1% and &lt;5%, D = ≥5%. Sensitivity and specificity of detecting PBS-PC (y-axis) for groups A vs. B, B vs. C, and C vs. D were plotted against CTC levels on PBF (x-axis) to determine the best cut-off. Area Under the Curve (AUC) was calculated to determine accuracy. Agreement analysis was done for internal validation and reliability assessment. Results: 6066 test records of CTC were reviewed. Circulating tumor plasma cells (CTC) by peripheral blood flow cytometry (PBF) were reported in 1610 (26.5%) instances, and peripheral blood smear for plasma cells (PBS-PC) in 226 (3.7%) tests. Peripheral blood smear (PBS) demonstrated a sensitivity of 9.3%, specificity of 99.5%, positive predictive value (PPV) of 86.7%, and negative predictive value (NPV) of 75.3% when compared to peripheral blood flow cytometry (PBF). The correlation between PBS-PC% and CTC was weakly positive (Spearman's rho=0.27, p&lt;0.01). The median (Q1, Q3) of circulating tumor plasma cells (CTC) for group A (n=5840) was 0 (0, 0), group B (n=82) was 780 (97.75, 4012.75), group C (n=67) was 1801 (61.50, 8723.50), and group D (n=77) was 20518 (2296, 41349) (p-value &lt;0.001). On post-hoc analysis, all pairs had significant differences in CTCs. To distinguish between A and B, a threshold of 32.5 for CTC demonstrated a sensitivity of 81.8% (AUC=87.8%), and to distinguish B and C, 1125 demonstrated a sensitivity of 59.4% (AUC=63.2%). To distinguish between C and D, 5875 for CTC demonstrated a sensitivity of 68.8% (AUC=74.6%). Agreement Analysis: The Bland-Altman plot demonstrates a slight positive bias between the Clonal Event % and %PBS-PC values (mean of difference = 0.28). The variability of differences increases with higher mean values, indicating a potential proportional bias. All predicted thresholds lie within the limits of agreement (LoA) (upper: 6.13, lower: -5.58) and are therefore reliable. The precision range (upper: 0.35, lower: 0.2) is narrower for the predicted thresholds. The methods should not be compared beyond a circulating tumor plasma cell (CTC) count of 9,195. Conclusion: Detection of circulating tumor plasma cells (CTC) by peripheral blood smear (PBS) is very specific but not sensitive and can be improved by peripheral blood flow cytometry (PBF) detection of CTC. Important clinical cutoffs by morphology (presence and ≥5%) have their counterparts by PBF (32 and 5875 out of 150,000 peripheral blood mononuclear cells (PBMC), respectively). The predicted thresholds are reliable. The study is retrospective and may not be generalizable across all demographics and disease severity. Circulating tumor plasma cell (CTC) values aren't absolute counts and may not be generalizable for low white blood cell (WBC) counts.

Introduction Monoclonal antibodies (MoAbs) directed against CD38 are a mainstay in multiple myeloma (MM). Downregulation of CD38 expression is a resistance mechanism of antigenic escape, however, cell surface expression may return. While this tumor cell plasticity is likely driven by epigenetic mechanisms, it is unknown if MM cells can develop permanent resistance to anti-CD38 MoAb by acquiring genomic events leading to biallelic disruption of CD38 thus precluding re-treatment. Here, we report biallelic loss of CD38 as a recurrent mechanism of resistance, often via convergent evolution where distinct subclones are advantaged towards the same selective pressure. Methods We interrogated two sets of patients (Mayo Clinic; n = 31, Calgary University; n =19) with relapse following anti-CD38 MoAbs. Two additional cases of interest were selected from Heidelberg University. Whole genome sequencing (WGS, 60-100x) or whole exome sequencing (WXS), flow cytometry (flow), and bulk RNAseq were used to validate the effect of genomic events on CD38. Missense mutations predicted to affect CD38 MoAb binding were validated with site-directed mutagenesis of a CD38 plasmid, transduction into K562 MM cells, and CD38 MoAb binding assays. Results The prevalence of biallelic inactivation of CD38 at anti-CD38 MoAb relapse was 6% (3/50). An additional selected biallelic event was noted in the Heidelberg set. Each case is described in detail: MM-19 (Calgary) received daratumumab (Dara), lenalidomide, and dexamethasone (Dex), and then had a second exposure to Dara, following relapse. At this second relapse, WGS was performed on two independent extramedullary disease localizations in the liver. Both samples had unique, large deletions (&gt;3 Mb) on chromosome 4 coupled with additional unique focal deletions involving CD38. Flow confirmed the presence of CD138+/CD38- plasma cell populations in both sites. A second case of convergent evolution was seen in a patient (Mayo) after seventh-line Dara/Len/Dex. WGS revealed a common deletion encompassing CD38. On the remaining allele, one subclone had a structural variation-mediated loss of CD38(cancer cell fraction; CCF 40%) while another harbored an L153H missense mutation (CCF 60%). Flow confirmed presence of two CD138+ clones; one with diminished CD38 expression. L153H was functionally validated to abrogate Dara binding, while maintaining CD38 expression. The third sample was collected after fifth-line Dara and sixth-line Dara/Bortezomib/Dex (Heidelberg) and represents another case of convergent evolution with a common loss of chromosome 4 and two subclones detected by WGS (P98Lfs*12, CCF 39% resulting in biallelic inactivation; R140G, CCF 61%). Bulk RNAseq revealed, in comparison to NDMM from CoMMpass, CD38 expression below the first decile. Notably, the R140G variant reduced binding affinity to Dara but not to Isatuximab (Isa). A fourth case (Mayo) relapsing after 6 months of Dara-based therapy similarly had a focal deletion of CD38 with an L18Sfs*16 frameshift deletion on the remaining allele associated with a 4.35-fold reduction in CD38 expression by RNAseq compared to pre-treatment. Interestingly, in 3/4 cases the CD38-negative clone emerged after a second exposure to anti-CD38 MoAb, suggesting emergence with the first exposure and then selection during the second. One other monoallelic event (C275Y; Heidelberg) was detected post-Isa. Copy number of the alternate allele is not known given targeted sequencing methodology, but the variant was seen to reduce binding affinity to both Dara and Isa. The prevalence of monoallelic loss of CD38 in newly diagnosed MM from MMRF CoMMpass was 7% (50/701) and two patients had mutations in CD38, but there were no cases of biallelic loss suggesting these events are driven by treatment pressure. Conclusion Distinction of the mode of CD38 loss is valuable for predicting outcome of re-challenge to anti-CD38 MoAbs. We see here an estimated 5-10% of anti-CD38 MoAb relapses are characterized by genomic antigen escape. Results here are in line with antigen escape to BCMA- and GPRC5D-directed immunotherapies (Lee et al, Nat Med 2023), suggesting that acquired antigen escape resulting from prolonged exposure to the selective pressures of immunotherapy is a recurrent mechanism of resistance in MM. Moreover, in select cases, non-overlapping resistance mutations provide rationale for switching agents at second exposure.