Comparison of characteristics and laboratory tests of COVID-19 hematological patients from France and Brazil during the pre-vaccination period: identification of prognostic profiles for survival

ABSTRACT Introduction: COVID-19 disease presentation is heterogeneous, from asymptomatic up to severe life-threatening forms. Getting further insights into patients with specific diseases is of particular interest. We aimed to identify profiles of hematology patients hospitalized with COVID-19 that would be associated with survival and to assess the differences between cohorts Methods: A binational cohort of 263 patients with COVID-19 and hematological disease was studied in Paris, France and São Paulo, Brazil. Patient profiles were based on age, comorbidities, biological measurements, COVID-19 symptoms and hematological disease characteristics. A semi-supervised learning method with a survival endpoint was first used, following which, a classifier was identified to allow the classification of patients using only baseline information Main results: Two profiles of patients were identified, one being young patients with few comorbidities and low C-reactive protein (CRP), D-dimers, lactate dehydrogenase (LDH) and creatinine levels, and the other, older patients, with several comorbidities and high levels of the 4 biology markers. The profiles were strongly associated with survival (p < 0.0001), even after adjusting for age (p = 0.0002). The 30-day survival rate was 77.1% in the first profiles, versus 46.7% in the second. The Brazilian analysis emphasized the importance of age, while the French focused on the comorbidities Conclusion: This analysis showed the importance of CRP, LHD and creatinine in the COVID-19 presentation and prognosis, whatever the geographic origin of the patients.

Comparison of characteristics and laboratory tests of COVID-19 hematological patients from France and Brazil during the pre-vaccination period: identification of prognostic profiles for survival.

INTRODUCTION: COVID-19 disease presentation is heterogeneous, from asymptomatic up to severe life-threatening forms. Getting further insights into patients with specific diseases is of particular interest. We aimed to identify profiles of hematology patients hospitalized with COVID-19 that would be associated with survival and to assess the differences between cohorts METHODS: A binational cohort of 263 patients with COVID-19 and hematological disease was studied in Paris, France and São Paulo, Brazil. Patient profiles were based on age, comorbidities, biological measurements, COVID-19 symptoms and hematological disease characteristics. A semi-supervised learning method with a survival endpoint was first used, following which, a classifier was identified to allow the classification of patients using only baseline information MAIN RESULTS: Two profiles of patients were identified, one being young patients with few comorbidities and low C-reactive protein (CRP), D-dimers, lactate dehydrogenase (LDH) and creatinine levels, and the other, older patients, with several comorbidities and high levels of the 4 biology markers. The profiles were strongly associated with survival (p < 0.0001), even after adjusting for age (p = 0.0002). The 30-day survival rate was 77.1% in the first profiles, versus 46.7% in the second. The Brazilian analysis emphasized the importance of age, while the French focused on the comorbidities CONCLUSION: This analysis showed the importance of CRP, LHD and creatinine in the COVID-19 presentation and prognosis, whatever the geographic origin of the patients.

Definition of a novel breast tumor-specific classifier based on secretome analysis.

BACKGROUND: During cancer development, the normal tissue microenvironment is shaped by tumorigenic events. Inflammatory mediators and immune cells play a key role during this process. However, which molecular features most specifically characterize the malignant tissue remains poorly explored. METHODS: Within our institutional tumor microenvironment global analysis (T-MEGA) program, we set a prospective cohort of 422 untreated breast cancer patients. We established a dedicated pipeline to generate supernatants from tumor and juxta-tumor tissue explants and quantify 55 soluble molecules using Luminex or MSD. Those analytes belonged to five molecular families: chemokines, cytokines, growth factors, metalloproteinases, and adipokines. RESULTS: When looking at tissue specificity, our dataset revealed some breast tumor-specific characteristics, as IL-16, as well as some juxta-tumor-specific secreted molecules, as IL-33. Unsupervised clustering analysis identified groups of molecules that were specific to the breast tumor tissue and displayed a similar secretion behavior. We identified a tumor-specific cluster composed of nine molecules that were secreted fourteen times more in the tumor supernatants than the corresponding juxta-tumor supernatants. This cluster contained, among others, CCL17, CCL22, and CXCL9 and TGF-ß1, 2, and 3. The systematic comparison of tumor and juxta-tumor secretome data allowed us to mathematically formalize a novel breast cancer signature composed of 14 molecules that segregated tumors from juxta-tumors, with a sensitivity of 96.8% and a specificity of 96%. CONCLUSIONS: Our study provides the first breast tumor-specific classifier computed on breast tissue-derived secretome data. Moreover, our T-MEGA cohort dataset is a freely accessible resource to the biomedical community to help advancing scientific knowledge on breast cancer.

PD-L1 and ICOSL discriminate human Secretory and Helper dendritic cells in cancer, allergy and autoimmunity.

Dendritic cells (DC) are traditionally classified according to their ontogeny and their ability to induce T cell response to antigens, however, the phenotypic and functional state of these cells in cancer does not necessarily align to the conventional categories. Here we show, by using 16 different stimuli in vitro that activated DCs in human blood are phenotypically and functionally dichotomous, and pure cultures of type 2 conventional dendritic cells acquire these states (termed Secretory and Helper) upon appropriate stimuli. PD-L1highICOSLlow Secretory DCs produce large amounts of inflammatory cytokines and chemokines but induce very low levels of T helper (Th) cytokines following co-culturing with T cells. Conversely, PD-L1lowICOSLhigh Helper DCs produce low levels of secreted factors but induce high levels and a broad range of Th cytokines. Secretory DCs bear a single-cell transcriptomic signature indicative of mature migratory LAMP3+ DCs associated with cancer and inflammation. Secretory DCs are linked to good prognosis in head and neck squamous cell carcinoma, and to response to checkpoint blockade in Melanoma. Hence, the functional dichotomy of DCs we describe has both fundamental and translational implications in inflammation and immunotherapy.

Negative Relationship between Post-Treatment Stromal Tumor-Infiltrating Lymphocyte (TIL) and Survival in Triple-Negative Breast Cancer Patients Treated with Dose-Dense Dose-Intense NeoAdjuvant Chemotherapy.

Background: Patients with triple-negative breast cancers (TNBC) have a poor prognosis unless a pathological complete response (pCR) is achieved after neoadjuvant chemotherapy (NAC). Few studies have analyzed changes in TIL levels following dose-dense dose-intense (dd-di) NAC. Patients and methods: From 2009 to 2018, 117 patients with TNBC received dd-di NAC at our institution. We aimed to identify factors associated with pre- and post-NAC TIL levels, and oncological outcomes relapse-free survival (RFS), and overall survival (OS). Results: Median pre-NAC and post-NAC TIL levels were 15% and 3%, respectively. Change in TIL levels with treatment was significantly correlated with metabolic response (SUV) and pCR. High post-NAC TIL levels were associated with a weak metabolic response after two cycles of NAC, with the presence of residual disease and nodal involvement at NAC completion. In multivariate analyses, high post-NAC TIL levels independently predicted poor RFS and poor OS (HR = 1.4 per 10% increment, 95%CI (1.1; 1.9) p = 0.014 and HR = 1.8 per 10% increment 95%CI (1.3−2.3), p < 0.0001, respectively). Conclusion: Our results suggest that TNBC patients with TIL enrichment after NAC are at higher risk of relapse. These patients are potential candidates for adjuvant treatment, such as immunotherapy, in clinical trials.

A multivariate modeling framework to quantify immune checkpoint context-dependent stimulation on T cells.

Cells receive, and adjust to, various stimuli, which function as part of complex microenvironments forming their "context". The possibility that a given context impacts the response to a given stimulus defines "context-dependency" and it explains large parts of the functional variability of physiopathological and pharmacological stimuli. Currently, there is no framework to analyze and quantify context-dependency over multiple contexts and cellular response outputs. We established an experimental system including a stimulus of interest, applied to an immune cell type in several contexts. We studied the function of OX40 ligand (OX40L) on T helper (Th) cell differentiation, in 4 molecular (Th0, Th1, Th2, and Th17) and 11 dendritic cell (DC) contexts (monocyte-derived DC and cDC2 conditions). We measured 17 Th output cytokines in 302 observations, and developed a statistical modeling strategy to quantify OX40L context-dependency. This revealed highly variable context-dependency, depending on the output cytokine and context type itself. Among molecular contexts, Th2 was the most influential on OX40L function. Among DC contexts, the DC type rather than the activating stimuli was dominant in controlling OX40L context-dependency. This work mathematically formalizes the complex determinants of OX40L functionality, and provides a unique framework to decipher and quantify the context-dependent variability of any biomolecule or drug function.

Multiobjective semisupervised learning with a right-censored endpoint adapted to the multiple imputation framework.

Semisupervised learning aims to use additional knowledge in the search for data structure. In clinical applications, including predictive information in the construction of a data-driven classification is of major importance. This work was motivated by a study that aimed to identify different patterns of immune parameters that would be associated with relapse-free survival in a cohort of breast cancer patients. Supervised and unsupervised objectives can be concomitantly optimized using multiobjective optimization. We propose such a procedure that addresses two challenges in the semisupervised approach, that is, missing data and additional knowledge based on survival time. The former was handled by using multiple imputation and consensus clustering. Survival information was incorporated in the supervised objective through the estimation of a cross-validation error of a Cox regression. A simulation study was performed to assess the performance of the proposed procedure. On complete datasets, the performances were compared to those of an existing modified multiobjective semisupervised learning method. The added value of including the survival data in the learning process was assessed by comparing the procedure to unsupervised learning. The proposed procedure showed better performance than the existing method, notably in the selection of the number of clusters. On incomplete datasets, the procedure showed little sensitivity to most of its parameters, even though a high number of imputations and partition initialization seeds improved the performance. The performance was degraded with a high proportion of missing data (40%) and with more ambiguous data structures. Simulation results and application on real data support the conclusion that our procedure enables the construction of a classification associated with a right-censored endpoint on a possibly incomplete dataset.

Clustering with missing and left-censored data: A simulation study comparing multiple-imputation-based procedures.

Cluster analysis, commonly used to explore large biomedical datasets, can be challenging, notably due to missing data or left-censored data induced by the sensitivity limits of the biochemical measurement method. Usually, complete-case analysis, simple imputation, or stochastic simple imputation are applied before clustering. More recently, consensus methods following multiple imputation have been proposed. However, they ignore left-censoring and do not allow the number of clusters to vary across the partitions of each imputed dataset. Here, we developed a consensus-based clustering algorithm in which left-censored data are taken into account using a modified multiple imputation method and the number of clusters is estimated for each imputed dataset. A simulation study was conducted to assess the performance in terms of the number of clusters, the percentage of unclassified observations, and the adjusted Rand index. The simulation results showed that the investigated method works well compared to several alternative approaches. A real-world application in breast cancer patients showed that the proposed method may reveal novel clusters of patients.