NGFR regulates stromal cell activation in germinal centers

Nerve growth factor receptor (NGFR) is expressed by follicular dendritic cells (FDCs). However, the role of NGFR in the humoral response is not well defined. Here, we study the effect of Ngfr loss on lymph node organization and function, demonstrating that Ngfr depletion leads to spontaneous germinal center (GC) formation and an expansion of the GC B cell compartment. In accordance with this effect, stromal cells are altered in Ngfr-/- mice with a higher frequency of FDCs, characterized by CD21/35, MAdCAM-1, and VCAM-1 overexpression. GCs are located ectopically in Ngfr-/- mice, with lost polarization together with impaired high-affinity antibody production and an increase in circulating autoantibodies. We observe higher levels of autoantibodies in Bcl2 Tg/Ngfr-/- mice, concomitant with a higher incidence of autoimmunity and lower overall survival. Our work shows that NGFR is involved in maintaining GC structure and function, participating in GC activation, antibody production, and immune tolerance.

Unique repetitive nucleic acid structures mirror switch regions in the human IgH locus

Immunoglobulin (Ig) genes carry the unique ability to be reshaped in peripheral B lymphocytes after these cells encounter a specific antigen. B cells can then further improve their affinity, acquire new functions as memory cells and eventually end up as antibody-secreting cells. Ig class switching is an important change that occurs in this context, thanks to local DNA lesions initiated by the enzyme activation-induced deaminase (AID). Several cis-acting elements of the Ig heavy (IgH) chain locus make it accessible to the AID-mediated lesions that promote class switch recombination (CSR). DNA repeats, with a non-template strand rich in G-quadruplexes (G4)-DNA, are prominent cis-targets of AID and define the so-called "switch" (S) regions specifically targeted for CSR. By analyzing the structure of the human IgH locus, we uncover that abundant DNA repeats, some with a putative G4-rich template strand, are additionally present in downstream portions of the IgH coding genes. These like-S (LS) regions stand as 3' mirror-images of S regions and also show analogies to some previously reported repeats associated with the IgH locus 3' super-enhancer. A regulatory role of LS repeats is strongly suggested by their specific localization close to exons encoding the membrane form of Ig molecules, and by their conservation during mammalian evolution.

The negative influence of baseline cell-free DNA on long-term survival in DLBCL depends on frontline treatment intensity

PURPOSE

In this study, we address the question of the effect of first-line treatment intensity in de novo DLBCL on the impact of baseline cfDNA on long-term survival.

EXPERIMENTAL DESIGN

The GOELAMS 075 randomized clinical trial compared R-CHOP with high-dose R-chemotherapy plus autologous stem cell transplantation (R-HDT) for patients aged ≤60. An interim PET assessment was used to refer patients for salvage therapy. With a median follow-up of more than 5.8 years, we analyzed the effects of the treatment arm, salvage therapy, and cfDNA level at diagnosis on overall survival (OS).

RESULTS

In a representative group of 123 patients, a high cfDNA concentration (>55ng/ml) at diagnosis was associated with poor clinical prognostic factors and constituted a prognostic marker, independently of the age-adjusted International Prognostic Index. A cfDNA level above a threshold value of 55 ng/ml at diagnosis was associated with significantly worse OS. In an intention-to-treat analysis, high-cfDNA R-CHOP patients (but not high-cfDNA R-HDT patients) had worse OS (hazard ratio [95% confidence interval]: 3.99 [1.98-10.74], P=0.006). In patients with high cfDNA levels, salvage therapy and transplantation were associated with a significantly higher OS rate. Among 50 patients with complete response 6 months after the end of treatment, for 11 out of 24 R-CHOP patients, the cfDNA did not fall back to normal values.

CONCLUSIONS

In this randomized clinical trial, intensive regimens mitigated the negative influence of high cfDNA levels in de novo DLBCL, relative to R-CHOP.

Regulatory B Cells Contribute to the Clinical Response After Bone Marrow-Derived Mesenchymal Stromal Cell Infusion in Patients With Systemic Sclerosis

Mesenchymal stromal cells (MSCs) have recently emerged as an interesting therapeutic approach for patients with progressive systemic sclerosis (SSc), a rare and life-threatening orphan autoimmune disease. Whereas MSC immunomodulatory potential is considered as a central mechanism for their clinical benefit, very few data are available on the impact of MSCs on immune cell subsets in vivo. In the current extended study of a phase I/II clinical trial exploring the injection of a single dose of allogeneic bone marrow-MSCs (alloBM-MSCs) in patients with severe SSc (NCT02213705), we performed a longitudinal in-depth characterization of circulating immune cells in 19 MSC-treated patients, including 14 responders and 5 non-responders. By a combination of flow cytometry and transcriptomic analyses, we highlighted an increase in circulating CD24hiCD27posCD38lo/neg memory B cells, the main IL-10-producing regulatory B cell (Breg) subset, and an upregulation of IL10 expression in ex-vivo purified B cells, specifically in responder patients, early after the alloBM-MSC infusion. In addition, a deeper alteration of the B-cell compartment before alloBM-MSC treatment, including a higher expression of profibrotic cytokines IL6 and TGFβ by sorted B cells was associated with a non-responder clinical status. Finally, BM-MSCs were able to directly upregulate IL-10 production in activated B cells in vitro. These data suggest that cytokine-producing B cells, in particular Breg, are pivotal effectors of BM-MSC therapeutic activity in SSc. Their quantification as activity biomarkers in MSC potency assays and patient selection criteria may be considered to reach optimal clinical benefit when designing MSC-based clinical trials.

RNA exosome drives early B cell development via noncoding RNA processing mechanisms

B cell development is linked to successful V(D)J recombination, allowing B cell receptor expression and ultimately antibody secretion for adaptive immunity. Germline noncoding RNAs (ncRNAs) are produced at immunoglobulin (Ig) loci during V(D)J recombination, but their function and posttranscriptional regulation are incompletely understood. Patients with trichohepatoenteric syndrome, characterized by RNA exosome pathway component mutations, exhibit lymphopenia, thus demonstrating the importance of ncRNA surveillance in B cell development in humans. To understand the role of RNA exosome in early B cell development in greater detail, we generated mouse models harboring a B cell-specific cre allele (Mb1^cre), coupled to conditional inversion-deletion alleles of one RNA exosome core component (Exosc3) or RNase catalytic subunits (Exosc10 or Dis3). We noticed increased expression of RNA exosome subunits during V(D)J recombination, whereas a B cell developmental blockade at the pro-B cell stage was observed in the different knockout mice, overlapping with a lack of productive rearrangements of VDJ genes at the Ig heavy chain (Igh). This unsuccessful recombination prevented differentiation into pre-B cells, with accumulation of ncRNAs and up-regulation of the p53 pathway. Introduction of a prearranged Igh VDJ allele partly rescued the pre-B cell population in Dis3-deficient cells, although V-J recombination defects were observed at Ig light chain kappa (Igκ), preventing subsequent B cell development. These observations demonstrated that the RNA exosome complex is important for Igh and Igκ recombination and establish the relevance of RNA processing for optimal diversification at these loci during B cell development.

Committed Human CD23-Negative Light-Zone Germinal Center B Cells Delineate Transcriptional Program Supporting Plasma Cell Differentiation

B cell affinity maturation occurs in the germinal center (GC). Light-zone (LZ) GC B cells (B_GC-cells) interact with follicular dendritic cells (FDCs) and compete for the limited, sequential help from T follicular helper cells needed to escape from apoptosis and complete their differentiation. The highest-affinity LZ B_GC-cells enter the cell cycle and differentiate into PCs, following a dramatic epigenetic reorganization that induces transcriptome changes in general and the expression of the PRDM1 gene in particular. Human PC precursors are characterized by the loss of IL-4/STAT6 signaling and the absence of CD23 expression. Here, we studied the fate of human LZ B_GC-cells as a function of their CD23 expression. We first showed that CD23 expression was restricted to the GC LZ, where it was primarily expressed by FDCs; less than 10% of tonsil LZ B_GC-cells were positive. Sorted LZ B_GC-cells left in culture and stimulated upregulated CD23 expression but were unable to differentiate into PCs – in contrast to cells that did not upregulate CD23 expression. An in-depth analysis (including single-cell gene expression) showed that stimulated CD23-negative LZ B_GC-cells differentiated into plasmablasts and time course of gene expression changes delineates the transcriptional program that sustains PC differentiation. In particular, we identified a B cell proliferation signature supported by a transient MYC gene expression. Overall, the CD23 marker might be of value in answering questions about the differentiation of normal B_GC-cells and allowed us to propose an instructive LZ B_GC-cells maturation and fate model.

Nonclassical Monocytes Are Prone to Migrate Into Tumor in Diffuse Large B-Cell Lymphoma

Absolute count of circulating monocytes has been proposed as an independent prognostic factor in diffuse large B-cell lymphoma (DLBCL). However, monocyte nomenclature includes various subsets with pro-, anti-inflammatory, or suppressive functions, and their clinical relevance in DLBCL has been poorly explored. Herein, we broadly assessed circulating monocyte heterogeneity in 91 DLBCL patients. Classical- (cMO, CD14pos CD16neg) and intermediate- (iMO, CD14pos CD16pos) monocytes accumulated in DLBCL peripheral blood and exhibited an inflammatory phenotype. On the opposite, nonclassical monocytes (ncMOSlanpos, CD14low CD16pos Slanneg and ncMOSlanneg, CD14low CD16pos, Slanneg) were decreased in peripheral blood. Tumor-conditioned monocytes presented similarities with ncMO phenotype from DLBCL and were prone to migrate in response to CCL5 and CXCL12, and presented similarities with DLBCL-infiltrated myeloid cells, as defined by mass cytometry. Finally, we demonstrated the adverse value of an accumulation of nonclassical monocytes in 2 independent cohorts of DLBCL.

SARS-CoV-2-Induced ARDS Associates with MDSC Expansion, Lymphocyte Dysfunction, and Arginine Shortage

Purpose

The SARS-CoV-2 infection can lead to a severe acute respiratory distress syndrome (ARDS) with prolonged mechanical ventilation and high mortality rate. Interestingly, COVID-19-associated ARDS share biological and clinical features with sepsis-associated immunosuppression since lymphopenia and acquired infections associated with late mortality are frequently encountered. Mechanisms responsible for COVID-19-associated lymphopenia need to be explored since they could be responsible for delayed virus clearance and increased mortality rate among intensive care unit (ICU) patients.

Methods

A series of 26 clinically annotated COVID-19 patients were analyzed by thorough phenotypic and functional investigations at days 0, 4, and 7 after ICU admission.

Results

We revealed that, in the absence of any difference in demographic parameters nor medical history between the two groups, ARDS patients presented with an increased number of myeloid-derived suppressor cells (MDSC) and a decreased number of CD8^pos effector memory cell compared to patients hospitalized for COVID-19 moderate pneumonia. Interestingly, COVID-19-related MDSC expansion was directly correlated to lymphopenia and enhanced arginase activity. Lastly, T cell proliferative capacity in vitro was significantly reduced among COVID-19 patients and could be restored through arginine supplementation.

Conclusions

The present study reports a critical role for MDSC in COVID-19-associated ARDS. Our findings open the possibility of arginine supplementation as an adjuvant therapy for these ICU patients, aiming to reduce immunosuppression and help virus clearance, thereby decreasing the duration of mechanical ventilation, nosocomial infection acquisition, and mortality.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10875-020-00920-5.

Integrative Analysis of Cell Crosstalk within Follicular Lymphoma Cell Niche: Towards a Definition of the FL Supportive Synapse

Simple Summary

Follicular lymphoma, the most frequent indolent non-Hodgkin’s B cell lymphoma, arises from a germinal center B cell proliferation supported by a multidirectional crosstalk with the tumor microenvironment, in particular with follicular helper T cells and mesenchymal stromal cells. Here, we explored this complex network, starting from a comparative analysis of the molecular signatures of B cells, T cells, and stromal cells obtained from normal versus lymphoma tissues, and focusing on deregulated genes reflecting the crosstalk between these three cell subsets organizing the lymphoma cell niche. This helps us to point out new lymphoma-specific pathways, related to transcriptomic and functional specific features of T and stromal cells, and contributing to tumor B cell support directly or through the recruitment and/or activation of other pro-tumoral cell components. In the future, targeting these cell interactions with specific drugs in the FL niche could represent an attractive option for novel therapeutic strategies.

Abstract

Follicular lymphoma (FL), the most frequent indolent non-Hodgkin’s B cell lymphoma, is considered as a prototypical centrocyte-derived lymphoma, dependent on a specific microenvironment mimicking the normal germinal center (GC). In agreement, several FL genetic alterations affect the crosstalk between malignant B cells and surrounding cells, including stromal cells and follicular helper T cells (Tfh). In our study, we sought to deconvolute this complex FL supportive synapse by comparing the transcriptomic profiles of GC B cells, Tfh, and stromal cells, isolated from normal versus FL tissues, in order to identify tumor-specific pathways. In particular, we highlighted a high expression of IL-6 and IL-7 in FL B cells that could favor the activation of FL Tfh overexpressing IFNG, able in turn to stimulate FL B cells without triggering MHC (major histocompatibility) class II expression. Moreover, the glycoprotein clusterin was found up-regulated in FL stromal cells and could promote FL B cell adhesion. Finally, besides its expression on Tfh, CD200 was found overexpressed on tumor B cells and could contribute to the induction of the immunosuppressive enzyme indoleamine-2,3 dioxygenase by CD200R-expressing dendritic cells. Altogether our findings led us to outline the contribution of major signals provided by the FL microenvironment and their interactions with malignant FL B cells.

Soluble BTN2A1 as a potential predictive biomarker of immune checkpoint inhibitor efficacy in advanced non-small cell lung cancer (NSCLC)

9561

Background: Medical treatment of lung cancer has irreversibly changed since the development of immune checkpoint inhibitors (ICI). However, immune biomarkers of efficacy are still lacking. Preliminary data in leukemia and pancreatic cancer showed that soluble immune checkpoints are associated with a reduced overall survival (OS). This led us to explore the prognostic and predictive value of soluble immune checkpoints in non-small cell lung cancer (NSCLC) patients treated with chemotherapy or ICI. Methods: We analyzed 90 advanced NSCLC patients. The pilot cohort (Rennes University Hospital, France), included 48 patients treated with platinum doublets (n = 33) or ICI (n = 15) (LOC/11-16 protocol). The confirmation cohort (Paoli-Calmettes Institute) included 42 patients treated with ICI (nivolumab or pembrolizumab) in a longitudinal prospective setting (Immunosup trial, NCT03595813). In both cohorts, enzyme-linked immunosorbent assays (ELISA) were performed in baseline plasma samples for soluble forms of PD-1, PD-L1, global BTN3, BTLA, BTN3-A1 and BTN2A1. Soluble ICI levels were linked to clinical data using Kaplan-Meier, log-rank and Cox proportional-hazards models. Cut-points were determined using maxstat package for survival, R software R 3.6.2. Results: Five soluble immune checkpoints correlated and clustered together in unsupervised analysis (PD-1, PD-L1, global BTN3, BTLA, BTN3-A1), but were not associated with ICI efficacy. In patients treated with ICI, in the pilot and confirmation cohort, a high baseline plasmatic concentration of soluble BTN2A1 was significantly associated with an improved OS (confirmation cohort with a BTN2A1 cut-point of 3.55 ng/ml: HR = 0.30, 95%CI = 0.12-0.74, p = 0.0057, median OS in BTN2A1low = 7.6 months and median OS in BTN2A1hi = 19.5 months). On the contrary, in patients treated with chemotherapy, soluble BTN2A1 concentration was not associated with overall survival. Conclusions: In advanced NSCLC patients, a high baseline plasmatic concentration of soluble BTN2A1 was correlated with improved outcomes for ICI, but not for chemotherapy, suggesting that baseline soluble BTN2A1 level is a potential predictive biomarker of ICI efficacy. Additional studies are ongoing to confirm this finding.

Frequent expression of PD-L1 on circulating breast cancer cells

Immune checkpoint regulators such as PD-L1 have become exciting new therapeutic targets leading to long lasting remissions in patients with advanced malignancies. However, in view of the remarkable costs and the toxicity profiles of these therapies, predictive biomarkers able to discriminate responders from non-responders are urgently needed. In the present paper, we provide evidence that PD-L1 is frequently expressed on metastatic cells circulating in the blood of hormone receptor-positive, HER2-negative breast cancer patients. We performed western blot, flow cytometry and immunocytochemical analyses to demonstrate the specificity of the PDL1 antibody used in our study and established immunoscores for PDL1 expression on single tumor cells. We then selected sixteen patients with circulating tumor cells (CTCs) using the CellSearch(®) system and found PD-L1((+)) CTCs in 11 patients (68.8%). The fraction of PD-L1((+)) CTCs varied from 0.2 to 100% in individual patients. This is the first report demonstrating the expression of PD-L1 on CTCs. The established CTC/PD-L1 assay can be used for liquid biopsy in future clinical trials for stratification and monitoring of cancer patients undergoing immune checkpoint blockade.

Lymphoma occurring in patients over 90 years of age: characteristics, outcomes, and prognostic factors. A retrospective analysis of 234 cases from the LYSA

BACKGROUND

Lymphoma occurring in patients aged 90 or older is not uncommon, and its incidence is expected to increase over time. Management of these patients is difficult given their underlying fragility and the lack of information regarding this population.

PATIENTS AND METHODS

We retrospectively analyzed 234 patients diagnosed with lymphoma at the age of 90 years or older (90+) between 1990 and 2012 to describe their characteristics, management, outcomes and prognostic factors.

RESULTS

The median age was 92 years; 88% were B-cell lymphomas consisting mainly in diffuse large B-cell lymphoma. The median overall survival (OS) was 7.2 months (range, 0-92 months) for the 227 patients with non-Hodgkin Lymphoma (NHL), with a significant difference between aggressive and indolent NHL (5.2 months versus 19.4 months, respectively). We further analyzed 166 NHL patients for whom detailed characteristics were available. Among these patients, 63.5% received a treatment, either local (7.5%) or systemic (56%). Lymphoma was reported as the main cause of death (40%). Treatment administration was associated with improved OS in patients with aggressive (P < 0.001) but not indolent NHL (P = 0.96). In patients with aggressive NHL, hypoalbuminemia appeared as a strong and independent negative prognostic factor.

CONCLUSIONS

The median OS is short in 90+ patients diagnosed with lymphoma but some patients experience prolonged survival. Lymphoma represents the main cause of death in these patients. Treatment may improve survival of selected patients with aggressive but not indolent NHL. Management of these patients may be guided by prognostic factors identified in this study, notably serum albumin.

IL-2 requirement for human plasma cell generation: coupling differentiation and proliferation by enhancing MAPK-ERK signaling

Mature B cell differentiation involves a well-established transcription factor cascade. However, the temporal dynamics of cell signaling pathways regulating transcription factor network and coordinating cell proliferation and differentiation remain poorly defined. To gain insight into the molecular processes and extrinsic cues required for B cell differentiation, we set up a controlled primary culture system to differentiate human naive B cells into plasma cells (PCs). We identified T cell-produced IL-2 to be critically involved in ERK1/2-triggered PC differentiation. IL-2 drove activated B cell differentiation toward PC independently of its proliferation and survival functions. Indeed, IL-2 potentiated ERK activation and subsequent BACH2 and IRF8 downregulation, sustaining BLIMP1 expression, the master regulator for PC differentiation. Inhibition of the MAPK-ERK pathway, unlike STAT5 signaling, impaired IL-2-induced PC differentiation and rescued the expression profile of BACH2 and IRF8. These results identify IL-2 as a crucial early input in mature B cell fate commitment.

Comparison of Bayesian network and decision tree methods for predicting access to the renal transplant waiting list

The study compares the effectiveness of Bayesian networks versus Decision Trees for predicting access to renal transplant waiting list in a French healthcare network. The data set consisted in 809 patients starting renal replacement therapy. The data were randomly divided into a training set (90%) and a validation set (10%). Bayesian network and CART decision tree were built on the training set. Their predictive performances were compared on the validation set. The age variable was found to be the most important factor in both models. Both models were highly sensitive and specific: sensitivity 90.0% (95%CI: 76.8-100), specificity 96.7% (95%CI: 92.2-100). Moreover, the models were complementary since the Bayesian network provided a global view of the variables' associations while the decision tree was more easily interpretable by physicians. These approaches provide insights on the current care process. This knowledge could be used for optimizing the healthcare process.

Comparing the APGAR score representation in HL7 and OpenEHR formalisms

Semantic interoperability, a prerequisite to eHealth projects, relies on sharing both information and knowledge models between information systems. Two of the standards of information models are HL7 v3 and the European norm, EN13606/OpenEHR. The paper compares both standards on a fragment of the prenatal medical record, the APGAR score. Two factors are compared: the formal representation of both information models, and the binding to knowledge models. The HL7v3 perinatality DMIM specification and the OpenEHR APGAR archetype were used. HL7v3 appears to be more formal than OpenEHR and able to represent in an easier way the clinical context. For both standards, the binding to reference terminologies such as LOINC is poor. We provide recommendations to improve the standards.

Managing an emergency department by analysing HIS medical data: a focus on elderly patient clinical pathways

INTRODUCTION

The objective of this paper is to present complementary views of the activity of the emergency department for a specific group of patients. Once validated, these views will be used as decision support tools for better managing the department and providing better care delivery for this population. The views are produced from the data stored in Healthcare Information Systems that correspond potentially to a vast source of information for supporting decisions on management or public health issues.

METHOD

The study focuses on two groups of patients: the elderly population (over 75-years-old) and the under 75-year-old patients, at the Rennes hospital. The validation of the views is performed by comparing results for the two distinct groups. Relevant data were extracted from the Emergency Department database. Several analysis (like cusum chart) and representation tools (Graphviz) were used to study the patients' pathways, the dynamics of arrivals and the patients' characteristics.

RESULTS

The representations provided a synthetic, global and comprehensive view of the department activities, to the satisfaction of the clinicians. The study showed that ICD-10 coding, assigned at the patient's departure from the emergency department hence from all available known clinical data, is not appropriate for the elderly population as these patients are mainly diagnosed by "symptoms" and several solutions are proposed. Finally, it is stressed out that a proper delivery of care to elderly patients should require some level of scheduling in the emergency department that is by essence characterized by its non scheduled activity.

Integrating clinical, gene expression, protein expression and preanalytical data for in silico cancer research

We present the phase I development of an integrative platform for the analysis of clinical, gene expression, protein expression and pre-analytical data. The platform is aimed at providing transparent access and analysis tools to researchers investigating new biomarkers and prognosis factors in the particular field of lymphoma diseases. In this article, we report on the data integration phase. The platform's principal advantage is its completeness as it integrates in a single environment clinical, genomic and proteomic data, allowing for their combined analysis. The architecture consists in a data warehouse including data on patients, clinical trials and array platforms and a DeMilitarized Zone for data exchange. A secure web-based platform allows any collaborative team to request the data warehouse and access basic statistics on integrated data. The presented system is currently in use.

Towards a decision support system for optimising clinical pathways of elderly patients in an emergency department

Data stored in Healthcare Information Systems correspond potentially to a vast source of information for supporting decisions in management or public health issues. The presented study illustrates clinical data valuation, through the analysis of clinical pathways of elderly patients at the Emergency Department (ED) of Rennes hospital.

METHOD

Relevant data were extracted from the Emergency Department database. Several analysis (e.g., cusum method) and representation tools (e.g., Graphviz) were used to study the patients' pathways, the dynamics of flows and the patients' characteristics.

RESULTS

4951 admissions were analysed and visualized. The representations provided a synthetic, global and comprehensive view of the department activities, to the satisfaction of the clinicians. Limitations of the ICD-10 coding of the diagnoses at the ED were pointed out as well as syntax and semantic interoperability issues. A solution is proposed for automating and scaling the Decision Support System.

A model for a regional health information network sharing clinical information between professionals in Britanny

The purpose of this paper is to present a generic model of information system supporting healthcare networks for health professionals in Brittany. This model is aimed to develop cooperation between hospital professionals, primary care and practitioners whatever the specific pathology, by providing tools for exchanging and sharing medical-related data. The model is designed based on the heterogeneity factors revealed by a prospective survey. It includes secured exchange of nominative medical data, all other information and activities being accessible by a Web portal. Other associated tools are a synchronous collaborative platform, and a e-learning module. The first implementation, currently in use by the professionals, is presented for the existing neurology healthcare networks.

Modelling a decision-support system for oncology using rule-based and case-based reasoning methodologies

In most hospital medical units, multidisciplinary committees meet weekly to discuss their patients' cases. The medical experts base their decisions on three sources of information. First, they check if their patient complies with existing guidelines. Failing these, the medical experts will base their therapeutic decisions on the cases of similar patients that they have treated in the past. We propose a multi-modal reasoning decision-support system based on both guideline and case series, which will automatically compare the patient's case to the corresponding guideline, then to other cases, and retrieve similar cases. The general structure of the system is presented here, the domain of application being oncology. As the patients' records are not currently stored in a database in a format which is directly accessible, an object-oriented model is proposed, which includes prognosis factors currently tested in clinical trials, well-established ones, and a description of the illness episodes. The system is designed to be a data warehouse. Such a system does not exist in the literature. Future work will be needed to define the similarity measures, and to connect the system to the current database.

Modelling of a case-based retrieval system for oncology

When a new patient does not comply with the guidelines, experts base their therapeutic decisions on similar patients' cases they have treated in the past. The case-based reasoning decision-support system we propose will automatically compare the patient's case to the structured guideline, then to other cases, and retrieve similar cases. The general structure of the system is presented here, the domain of application being oncology. As the patients' records are not stored in the current database in a format directly exploitable, an object-oriented model is proposed, that includes prognosis factors currently tested in clinical trials, well-established ones, and a description of the illness' episodes. Such system does not exist in the literature. It can be viewed as a data warehouse. Future work consists in defining the similarity measures, and connecting the system to the current database.