Evaluation of a machine-learning model based on laboratory parameters for the prediction of acute leukaemia subtypes: a multicentre model development and validation study in France.

BACKGROUND: Acute leukaemias are life-threatening haematological cancers characterised by the infiltration of transformed immature haematopoietic cells in the blood and bone marrow. Prompt and accurate diagnosis of the three main acute leukaemia subtypes (ie acute lymphocytic leukaemia [ALL], acute myeloid leukaemia [AML], and acute promyelocytic leukaemia [APL]) is of utmost importance to guide initial treatment and prevent early mortality but requires cytological expertise that is not always available. We aimed to benchmark different machine-learning strategies using a custom variable selection algorithm to propose an extreme gradient boosting model to predict leukaemia subtypes on the basis of routine laboratory parameters. METHODS: This multicentre model development and validation study was conducted with data from six independent French university hospital databases. Patients aged 18 years or older diagnosed with AML, APL, or ALL in any one of these six hospital databases between March 1, 2012, and Dec 31, 2021, were recruited. 22 routine parameters were collected at the time of initial disease evaluation; variables with more than 25% of missing values in two datasets were not used for model training, leading to the final inclusion of 19 parameters. The performances of the final model were evaluated on internal testing and external validation sets with area under the receiver operating characteristic curves (AUCs), and clinically relevant cutoffs were chosen to guide clinical decision making. The final tool, Artificial Intelligence Prediction of Acute Leukemia (AI-PAL), was developed from this model. FINDINGS: 1410 patients diagnosed with AML, APL, or ALL were included. Data quality control showed few missing values for each cohort, with the exception of uric acid and lactate dehydrogenase for the cohort from Hôpital Cochin. 679 patients from Hôpital Lyon Sud and Centre Hospitalier Universitaire de Clermont-Ferrand were split into the training (n=477) and internal testing (n=202) sets. 731 patients from the four other cohorts were used for external validation. Overall AUCs across all validation cohorts were 0·97 (95% CI 0·95-0·99) for APL, 0·90 (0·83-0·97) for ALL, and 0·89 (0·82-0·95) for AML. Cutoffs were then established on the overall cohort of 1410 patients to guide clinical decisions. Confident cutoffs showed two (0·14%) wrong predictions for ALL, four (0·28%) wrong predictions for APL, and three (0·21%) wrong predictions for AML. Use of the overall cutoff greatly reduced the number of missing predictions; diagnosis was proposed for 1375 (97·5%) of 1410 patients for each category, with only a slight increase in wrong predictions. The final model evaluation across both the internal testing and external validation sets showed accuracy of 99·5% for ALL diagnosis, 98·8% for AML diagnosis, and 99·7% for APL diagnosis in the confident model and accuracy of 87·9% for ALL diagnosis, 86·3% for AML diagnosis, and 96·1% for APL diagnosis in the overall model. INTERPRETATION: AI-PAL allowed for accurate diagnosis of the three main acute leukaemia subtypes. Based on ten simple laboratory parameters, its broad availability could help guide initial therapies in a context where cytological expertise is lacking, such as in low-income countries. FUNDING: None.

Real life evaluation of AlphaMissense predictions in hematological malignancies.

High-throughput sequencing plays a pivotal role in hematological malignancy diagnostics, but interpreting missense mutations remains challenging. In this study, we used the newly available AlphaMissense database to assess the efficacy of machine learning to predict missense mutation effects and its impact to improve our ability to interpret them. Based on the analysis of 2073 variants from 686 patients analyzed for clinical purpose, we confirmed the very high accuracy of AlphaMissense predictions in a large real-life data set of missense mutations (AUC of ROC curve 0.95), and provided a comprehensive analysis of the discrepancies between AlphaMissense predictions and state of the art clinical interpretation.

First-line second generation tyrosine kinase inhibitors in patients with newly diagnosed accelerated phase chronic myeloid leukemia.

Accelerated phase (AP) CML at onset and have poorer prognosis than CP-CML. We hypothesize that off-license use of second generation TKI (TKI2) as front-line therapy might counterbalance this poor prognosis, with limited toxicity. In "real-life" conditions, newly diagnosed patients meeting the ELN cytological criteria for AP-CML or harboring ACA and treated with first-line TKI2 were included in this retrospective multicenter observational study. We enrolled 69 patients [69.5 % male, median age 49.5 years, median follow-up 43.5 months], segregated into hematologic AP [HEM-AP (n = 32)] and cytogenetically defined AP [ACA-AP (n = 37)]. Hematologic parameters were worse in HEM-AP [spleen size (p = 0.014), PB basophils (p < .001), PB blasts (p < .001), PB blasts+promyelocytes (p < .001), low hemoglobin levels (p < .001)]. Dasatinib was initiated in 56 % patients in HEM-AP and in 27 % in ACA-AP, nilotinib in 44 % and 73 % respectively. Response and survival do not differ, regardless of the TKI2: 81 % vs 84.3 % patients achieved CHR, 88 % vs 84 % CCyR, 73 % vs 75 % MMR respectively. The estimated 5-year PFS 91.5 % (95%CI: 84.51-99.06 %) and 5-year OS 96.84 % (95%CI: 92.61-100 %). Only BM blasts (p < 0.001) and BM blasts+promyelocytes (p < 0.001) at diagnosis negatively influenced OS. TKI2 as front-line therapy in newly diagnosed AP-CML induce excellent responses and survival, and counterbalance the negative impact of advanced disease phase.

Role of Hyperthermic Intraperitoneal Chemotherapy Combined with Cytoreductive Surgery as Consolidation Therapy for Advanced Epithelial Ovarian Cancer.

BACKGROUND: Patients with advanced epithelial ovarian cancer who undergo incomplete surgery followed by six cycles of chemotherapy could benefit from second-look or consolidation cytoreductive surgery (CCRS). The primary goal of this study was to evaluate the overall survival (OS) in patients undergoing complete CCRS and the factors affecting survival. The secondary goal was to study the benefit of hyperthermic intraperitoneal chemotherapy (HIPEC) in these patients. METHODS: This was a retrospective analysis of 173 patients with CCRS with (n = 118) or without (n = 55) HIPEC treated at 12 French centers. Only patients having a completeness of cytoreduction (CC) 0/1 resection and a minimum of 5 years of follow-up were included. HIPEC was performed systematically for all patients except those treated at the four centers that did not perform HIPEC. RESULTS: The median Peritoneal Cancer Index was 6 (range 0-33). Closed HIPEC was performed in 59 (34.1%) patients and open HIPEC was performed in 56 (32.3%) patients. Grade 3-4 complications occurred in 64 (36.9%) patients. The median OS was 35.67 months (95% confidence interval [CI] 29.8-46.1) and was significantly longer for CCRS + HIPEC (31.4 months without HIPEC and 42.5 months with HIPEC; p = 0.022). On multivariate analysis, closed HIPEC (hazard ratio [HR] 0.46, 95% CI 0.29-0.73; p < 0.001) resulted in a longer OS, and age > 65 years (HR 2.17, 95% CI 1.14-4.11; p = 0.018) and bowel resection (HR 1.98, 95% CI 1.27-3.08; p = 0.020) led to a shorter OS. On multivariate logistic regression analysis, closed HIPEC (odds ratio 0.18; p = 0.001) was associated with a lower risk of dying at 5 years. CONCLUSIONS: CCRS was performed with an acceptable morbidity and resulted in good overall survival. The role of HIPEC in addition to CCRS should be evaluated in prospective, randomized studies and the closed technique prospectively compared with the open technique.

Distinct Immune Reconstitution Profiles Captured by Immune Functional Assays at 6 Months Post Allogeneic Hematopoietic Stem Cell Transplantation.

Immune reconstitution after allogeneic-hematopoietic-stem-cell transplantation (allo-HSCT) is a complex and individual process. In this cross-sectional study, whole-blood (WB) immune functional assay (IFA) was used to characterize immune function by assessing immune-related gene/pathway alterations. The usefulness of this tool in the context of infection, 6 months after transplantation, was evaluated. Sixty allo-HSCT recipients at 6 months after transplantation and 10 healthy volunteers (HV) were included. WB was stimulated in standardized TruCulture tubes using lipopolysaccharides and Staphylococcal enterotoxin B. Gene expression was quantified using a custom 144-gene panel using NanoString nCounter technology and analyzed using Ingenuity Pathway Analysis. The relationships between immune function and clinical characteristics, immune cell counts, and post-transplantation infections were assessed. Allo-HSCT recipients were able to activate similar networks of the innate and adaptive immune response compared to HV, with, nevertheless, a lower intensity. A reduced number and a lower expression of genes associated with immunoregulatory and inflammatory processes were observed in allo-HSCT recipients. The use of immunosuppressive treatments was associated with a protracted immune reconstitution revealed by transcriptomic immunoprofiling. No difference in immune cell counts was observed among patients receiving or not receiving immunosuppressive treatments using a large immunophenotyping panel. Moreover, the expression of a set of genes, including CCL3/CCL4, was significantly lower in patients with Herpesviridae reactivation (32%, 19/60), which once again was not identified using classical immune cell counts. Transcriptional IFA revealed the heterogeneity among allo-HSCT recipients with a reduced immune function, a result that could not be captured by circulating immune cell counts. This highlights the potential added value of this tool for the personalized care of immunocompromised patients.

A Simple-to-Perform ifn-γ mRNA Gene Expression Assay on Whole Blood Accurately Appraises Varicella Zoster Virus-Specific Cell-Mediated Immunity After Allogeneic Hematopoietic Stem Cell Transplantation.

Herpes zoster, which is due to the reactivation of Varicella zoster virus (VZV), is a leading cause of morbidity after allogeneic hematopoietic stem cell transplantation (allo-HSCT). While cell-mediated immunity (CMI) is critical to inhibiting VZV reactivation, CMI is not routinely assessed due to a lack of reliable tests. In this study, we aimed to evaluate VZV-specific CMI among allo-HSCT recipients (n = 60) and healthy individuals (HI, n = 17) through a panel of three immune functional assays after ex vivo stimulation by VZV antigen: quantification of (i) IFN-γ release in the supernatants, (ii) T-cell proliferation after a 7-day stimulation of peripheral blood mononuclear cells (PBMC), and (iii) measurement of the ifn-γ mRNA gene expression level after 24 h of stimulation of a whole-blood sample. VZV responsiveness was defined according to IFN-γ release from VZV-stimulated PBMC. Upon VZV stimulation, we found that allo-HSCT recipients at a median time of 6 [5-8] months post-transplant had lower IFN-γ release (median [IQR], 0.34 [0.12-8.56] vs. 409.5 [143.9-910.2] pg/ml, P <.0001) and fewer proliferating T cells (0.05 [0.01-0.57] % vs. 8.74 [3.12-15.05] %, P <.0001) than HI. A subset of allo-HSCT recipients (VZV-responders, n = 15/57, 26%) distinguished themselves from VZV-non-responders (n = 42/57, 74%; missing data, n = 3) by higher IFN-γ release (80.45 [54.3-312.8] vs. 0.22 [0.12-0.42] pg/ml, P <.0001) and T-cell proliferation (2.22 [1.18-7.56] % vs. 0.002 [0.001-0.11] %, P <.0001), suggesting recovery of VZV-specific CMI. Interestingly, VZV responders had a significant fold increase in ifn-γ gene expression, whereas ifn-γ mRNA was not detected in whole blood of VZV-non-responders (P <.0001). This study is the first to suggest that measurement of ifn-γ gene expression in 24-h-stimulated whole blood could be an accurate test of VZV-specific CMI. The routine use of this immune functional assay to guide antiviral prophylaxis at an individual level remains to be evaluated.

Chimeric antigen receptor T-cells targeting IL-1RAP: a promising new cellular immunotherapy to treat acute myeloid leukemia.

BACKGROUND: Acute myeloid leukemia (AML) remains a very difficult disease to cure due to the persistence of leukemic stem cells (LSCs), which are resistant to different lines of chemotherapy and are the basis of refractory/relapsed (R/R) disease in 80% of patients with AML not receiving allogeneic transplantation. METHODS: In this study, we showed that the interleukin-1 receptor accessory protein (IL-1RAP) protein is overexpressed on the cell surface of LSCs in all subtypes of AML and confirmed it as an interesting and promising target in AML compared with the most common potential AML targets, since it is not expressed by the normal hematopoietic stem cell. After establishing the proof of concept for the efficacy of chimeric antigen receptor (CAR) T-cells targeting IL-1RAP in chronic myeloid leukemia, we hypothesized that third-generation IL-1RAP CAR T-cells could eliminate AML LSCs, where the medical need is not covered. RESULTS: We first demonstrated that IL-1RAP CAR T-cells can be produced from AML T-cells at the time of diagnosis and at relapse. In vitro and in vivo, we showed the effectiveness of IL-1RAP CAR T-cells against AML cell lines expressing different levels of IL-1RAP and the cytotoxicity of autologous IL-1RAP CAR T-cells against primary cells from patients with AML at diagnosis or at relapse. In patient-derived relapsed AML xenograft models, we confirmed that IL-1RAP CAR T-cells are able to circulate in peripheral blood and to migrate in the bone marrow and spleen, are cytotoxic against primary AML cells and increased overall survival. CONCLUSION: In conclusion, our preclinical results suggest that IL-1RAP CAR T-based adoptive therapy could be a promising strategy in AML treatment and it warrants the clinical investigation of this CAR T-cell therapy.

HERVs characterize normal and leukemia stem cells and represent a source of shared epitopes for cancer immunotherapy.

Human endogenous retroviruses (HERVs) represent 8% of the human genome. The expression of HERVs and their immune impact have not been extensively studied in Acute Myeloid Leukemia (AML). In this study, we used a reference of 14 968 HERV functional units to provide a thorough analysis of HERV expression in normal and AML bone marrow cells. We show that the HERV retrotranscriptome accurately characterizes normal and leukemic cell subpopulations, including leukemia stem cells, in line with different epigenetic profiles. We then show that HERV expression delineates AML subtypes with different prognoses. We finally propose a method to select and prioritize CD8+ T cell epitopes derived from AML-specific HERVs and we show that lymphocytes infiltrating patient bone marrow at diagnosis contain naturally occurring CD8+ T cells against these HERV epitopes. We also provide in vitro data supporting the functionality of HERV-specific CD8+ T-cells against AML cells. These results show that HERVs represent an important source of genetic information that can help enhancing disease stratification or biomarker identification and an important reservoir of alternative tumor-specific T cell epitopes relevant for cancer immunotherapy.

Tumour burden and antigen-specific T cell magnitude represent major parameters for clinical response to cancer vaccine and TCR-engineered T cell therapy.

BACKGROUND: Cancer vaccines and T-cell receptor (TCR) engineered T cells (Tg-T cell) represent two different therapeutic strategies that can target the same tumour epitopes. The first approach requires the induction of a specific immune response in patients, while the second relies on the efficacy of adoptively transferred T cells. Because the ratio of antigen-specific T cells to tumour cells engaged by these strategies may influence the clinical outcome, we evaluated the efficacy of these two therapeutic approaches in solid tumours according to the tumour burden. METHODS: We performed a meta-analysis restricted to the therapeutic vaccine and Tg-T cell trials, presenting annotated individual clinical data. We adapted a previously published mathematical model for tumour immune dynamics to estimate the clinical impact of the number of specific T cells in regard to the tumour burden. RESULTS: A focused analysis of Tg-T cell studies revealed that clinical responses were mostly observed with the highest doses of infused T cells, suggesting that exceeding a threshold of effector T cells may be required for clinical efficacy. In silico modelling of cancer vaccine and Tg-T cell therapies starting at different tumour burdens showed that therapeutic vaccines control low or moderate tumour burdens, whereas increasing the amount of infused Tg-T cells succeeds in controlling high tumour masses. CONCLUSION: We propose that therapeutic vaccines should be considered in the context of low or moderate tumour burden, whereas Tg-T cell strategies may be more adapted for the treatment of advanced metastatic diseases.

Panel Informativity Optimizer: An R Package to Improve Cancer Next-Generation Sequencing Panel Informativity.

Mutation detection by next-generation sequencing is routinely used for cancer diagnosis. Selecting an optimal set of genes for a given cancer is not trivial as it has to optimize informativity (ie, the number of patients with at least one mutation in the panel), while minimizing panel length to reduce sequencing costs and increase sensitivity. We propose herein Panel Informativity Optimizer (PIO), an open-source software developed as an R package with a user-friendly graphical interface to help optimize cancer next-generation sequencing panel informativity. Using patient-level mutational data from either private data sets or preloaded data set of 91 independent cohorts from 31 different cancer types, PIO selects an optimal set of genomic intervals to maximize informativity and panel size in a given cancer type. Different options are offered, such as the definition of genomic intervals at the gene or exon level and the use of optimization strategy at the patient or patient per kilobase level. PIO can also propose an optimal set of genomic intervals to increase informativity of custom panels. A panel tester function is also available for panel benchmarking. Using public databases, as well as data from real-life settings, we demonstrate that PIO allows panel size reduction of up to 1000 kb, and accurately predicts the performance of custom or commercial panels.

Gene signature of circulating platelet-bound neutrophils is associated with poor prognosis in cancer patients.

Beyond their critical role in hemostasis, platelets physically interact with neutrophils to form neutrophil-platelet aggregates (NPAs), enhancing neutrophil effector functions during inflammation. NPAs may also promote disease worsening in various inflammatory diseases. However, characterization of NPAs in cancer remains totally unexplored. Using ImageStreamX (ISX) imaging flow cytometer, we were not only allowed able to detect CD15+ CD14- CD36+ ITGA2B+ NPAs in both healthy donors' (HDs) and cancer patients' bloods, but we also showed that NPAs result from the binding of platelets preferentially to low-density neutrophils (LDNs) as opposed to normal-density neutrophils (NDNs). By reanalyzing two independent public scRNAseq data of whole blood leukocytes from cancer patients and HDs, we could identify a subset of neutrophils with high platelet gene expression that may correspond to NPAs. Moreover, we showed that cancer patients' derived NPAs possessed a distinct molecular signature compared to the other neutrophil subsets, independently of platelet genes. Gene ontology (GO) term enrichment analysis of this NPAs-associated neutrophil transcriptomic signature revealed a significant enrichment of neutrophil degranulation, chemotaxis and trans-endothelial migration GO terms. Lastly, using The Cancer Genome Atlas (TCGA), we could show by multivariate Cox analysis that the NPAs-associated neutrophil transcriptomic signature was associated with a worse patient prognosis in several cancer types. These results suggest that neutrophils from NPAs are systemically primed by platelets empowering them with cancer progression capacities once at tumor site. NPAs may therefore hold clinical utility as novel noninvasive blood prognostic biomarker in cancer patients with solid tumors.

Class I/Class II HLA Evolutionary Divergence Ratio Is an Independent Marker Associated With Disease-Free and Overall Survival After Allogeneic Hematopoietic Stem Cell Transplantation for Acute Myeloid Leukemia.

Class I Human Leukocyte Antigen (HLA) evolutionary divergence (HED) is a metric which reflects immunopeptidome diversity and has been associated with immune checkpoint inhibitor responses in solid tumors. Its impact and interest in allogeneic hematopoietic stem cell transplantation (HCT) have not yet been thoroughly studied. This study analyzed the clinical and immune impact of class I and II HED in 492 acute myeloid leukemia (AML) recipients undergoing HCT. The overall cohort was divided into a training (n=338) and a testing (n=132) set. Univariate cox screening found a positive impact of a high class I HED and a negative impact of a high class II HED on both disease-free (DFS) and overall survival (OS). These results were combined in a unique marker, class I/class II HED ratio, and assessed in the testing cohort. The final multivariate cox model confirmed the positive impact of a high versus low class I/class II HED ratio on both DFS (Hazard Ratio (HR) 0.41 [95% CI 0.2-0.83]; p=0.01) and OS (HR 0.34 [0.19-0.59]; p<0.001), independently of HLA matching and other HCT parameters. No significant association was found between the ratio and graft-versus-host disease (GvHD) nor with neutrophil and platelet recovery. A high class I HED was associated with a tendency for an increase in NK, CD8 T-cell, and B cell recovery at 12 months. These results introduce HED as an original and independent prognosis marker reflecting immunopeptidome diversity and alloreactivity after HCT.

Identification of shared tumor epitopes from endogenous retroviruses inducing high-avidity cytotoxic T cells for cancer immunotherapy.

Human endogenous retroviruses (HERVs) represent 8% of the human genome. HERV products may represent tumor antigens relevant for cancer immunotherapy. We developed a bioinformatic approach to identify shared CD8+ T cell epitopes derived from cancer-associated HERVs in solid tumors. Six candidates among the most commonly shared HLA-A2 epitopes with evidence of translation were selected for immunological evaluation. In vitro priming assays confirmed the immunogenicity of these epitopes, which induced high-avidity CD8+ T cell clones. These T cells specifically recognize and kill HLA-A2+ tumor cells presenting HERV epitopes on HLA molecules, as demonstrated by mass spectrometry. Furthermore, epitope-specific CD8+ T cells were identified by dextramer staining among tumor-infiltrating lymphocytes from HLA-A2+ patients with breast cancer. Last, we showed that HERV-specific T cells lyse patient-derived organoids. These shared virus-like epitopes are of major interest for the development of cancer vaccines or T cell-based immunotherapies, especially in tumors with low/intermediate mutational burden.

[Development of allogeneic CAR T-cells]. / Développement des CAR-T allogéniques.

Chimeric antigen receptor (CAR) T-cell therapy represents a major breakthrough in the field of hematology. "Off-the-shelf" allogeneic CAR T-cells from donors have many potential advantages over autologous approaches, such as the immediate availability of cryopreserved batches, possible standardization of the cell product, time for multiple cell modifications, redosing and decreased cost. However, allogeneic T-cells possess foreign immunological identities that can lead to graft-versus-host disease (GvHD) and their rejection by the host immune system. In this review, we describe the different approaches to produce allogeneic CAR T-cells with limited potential for GvHD and that can persist in the recipient. The preliminary clinical results obtained with the first generation of allogeneic CAR T-cells are presented as well as the perspectives in hematological malignancies and solid tumors.

How to Obtain a High Quality ctDNA in Lymphoma Patients: Preanalytical Tips and Tricks.

The analysis of circulating tumor DNA (ctDNA) released by tumor cells holds great promise for patients with lymphoma, to refine the diagnostic procedure, clarify the prognosis, monitor the response to treatment, and detect relapses earlier. One of the main challenges of the coming years is to adapt techniques from highly specialized translational teams to routine laboratories as this requires a careful technical and clinical validation, and we have to achieve this as fast as possible to transform a promising biomarker into a routine analysis to have a direct consequence on patient care. Whatever the analytical technology used, the prerequisite is to obtain high yields of ctDNA of optimal quality. In this review, we propose a step-by-step description of the preanalytical process to obtain high-quality ctDNA, emphasizing the technical choices that need to be made and the experimental data that can support these choices.