Machine learning-based classification of diffuse large B-cell lymphoma patients by eight gene expression profiles

Application of machine learning in the management of lymphoma: Current practice and future prospects

In the past decade, digitization of medical records and multiomics data analysis in lymphoma has led to the accessibility of high-dimensional records. The digitization of medical records, the visualization of extensive volume data extracted from medical images, and the integration of multiomics methods into clinical decision-making have produced many datasets. As a promising auxiliary tool, machine learning (ML) intends to extract homologous features in large-scale data sets and encode them into various patterns to complete complicated tasks. At present, artificial intelligence and digital mining have shown promising prospects in the field of lymphoma pathological image analysis. The paradigm shift from qualitative analysis to quantitative analysis makes the pathological diagnosis more intelligent and the results more accurate and objective. ML can promote accurate lymphoma diagnosis and provide patients with prognostic information and more individualized treatment options. Based on the above, this comprehensive review of the general workflow of ML highlights recent advances in ML techniques in the diagnosis, treatment, and prognosis of lymphoma, and clarifies the boundedness and future orientation of the ML technique in the clinical practice of lymphoma.

Diagnosis of acute myeloid leukaemia on microarray gene expression data using categorical gradient boosted trees

We define an iterative method for dimensionality reduction using categorical gradient boosted trees and Shapley values and created four machine learning models which potentially could be used as diagnostic tests for acute myeloid leukaemia (AML). For the final Catboost model we use a dataset of 2177 individuals using as features 16 probe sets and the age in order to classify if someone has AML or is healthy. The dataset is multicentric and consists of data from 27 organizations, 25 cities, 15 countries and 4 continents. The performance of our last model is specificity: 0.9909, sensitivity: 0.9985, F1-score: 0.9976 and its ROC-AUC: 0.9962 using ten fold cross validation. On an inference dataset the perormance is: specificity: 0.9909, sensitivity: 0.9969, F1-score: 0.9969 and its ROC-AUC: 0.9939. To the best of our knowledge the performance of our model is the best one in the literature, as regards the diagnosis of AML using similar or not data. Moreover, there has not been any bibliographic reference which associates AML or any other type of cancer with the 16 probe sets we used as features in our final model.

Recent advances of Phosphodiesterase 4B in cancer

INTRODUCTION

Phosphodiesterase 4B (PDE4B) is a crucial enzyme in the phosphodiesterases (PDEs), acting as a regulator of cyclic adenosine monophosphate (cAMP). It is involved in cancer process through PDE4B/cAMP signaling pathway. Cancer occurs and develops with the regulation of PDE4B in the body, suggesting that PDE4B is a promising therapeutic target.

AREAS COVERED

This review covereed the function and mechanism of PDE4B in cancer. We summarized the possible clinical applications of PDE4B, and highlighted the possible ways to develop clinical applications of PDE4B inhibitors. We also discussed some common PDEs inhibitors, and expected the development of combined targeting PDE4B and other PDEs drugs in the future.

EXPERT OPINION

The existing research and clinical data can strongly prove the role of PDE4B in cancer. PDE4B inhibition can effectively increase cell apoptosis, inhibit cell proliferation, transformation, migration, etc., indicating that PDE4B inhibition can effectively inhibit the development of cancer. Other PDEs may antagonize or coordinate this effect. As for the further study on the relationship between PDE4B and other PDEs in cancer, it is still a challenge to develop multi-targeted PDEs inhibitors.

Multiple-model machine learning identifies potential functional genes in dilated cardiomyopathy

Introduction

Machine learning (ML) has gained intensive popularity in various fields, such as disease diagnosis in healthcare. However, it has limitation for single algorithm to explore the diagnosing value of dilated cardiomyopathy (DCM). We aim to develop a novel overall normalized sum weight of multiple-model MLs to assess the diagnosing value in DCM.

Methods

Gene expression data were selected from previously published databases (six sets of eligible microarrays, 386 samples) with eligible criteria. Two sets of microarrays were used as training; the others were studied in the testing sets (ratio 5:1). Totally, we identified 20 differently expressed genes (DEGs) between DCM and control individuals (7 upregulated and 13 down-regulated).

Results

We developed six classification ML methods to identify potential candidate genes based on their overall weights. Three genes, serine proteinase inhibitor A3 (SERPINA3), frizzled-related proteins (FRPs) 3 (FRZB), and ficolin 3 (FCN3) were finally identified as the receiver operating characteristic (ROC). Interestingly, we found all three genes correlated considerably with plasma cells. Importantly, not only in training sets but also testing sets, the areas under the curve (AUCs) for SERPINA3, FRZB, and FCN3 were greater than 0.88. The ROC of SERPINA3 was significantly high (0.940 in training and 0.918 in testing sets), indicating it is a potentially functional gene in DCM. Especially, the plasma levels in DCM patients of SERPINA3, FCN, and FRZB were significant compared with healthy control.

Discussion

SERPINA3, FRZB, and FCN3 might be potential diagnosis targets for DCM, Further verification work could be implemented.

The regulatory role of PDE4B in the progression of inflammatory function study

Inflammation is a response of the body to external stimuli (eg. chemical irritants, bacteria, viruses, etc.), and when the stimuli are persistent, they tend to trigger chronic inflammation. The presence of chronic inflammation is an important component of the tumor microenvironment produced by a variety of inflammatory cells (eg. macrophages, neutrophils, leukocytes, etc.). The relationship between chronic inflammation and cancer development has been widely accepted, and chronic inflammation has been associated with the development of many cancers, including chronic bronchitis and lung cancer, cystitis inducing bladder cancer. Moreover, chronic colorectitis is more likely to develop into colorectal cancer. Therefore, the specific relationship and cellular mechanisms between inflammation and cancer are a hot topic of research. Recent studies have identified phosphodiesterase 4B (PDE4B), a member of the phosphodiesterase (PDEs) protein family, as a major cyclic AMP (cAMP) metabolizing enzyme in inflammatory cells, and the therapeutic role of PDE4B as chronic inflammation, cancer. In this review, we will present the tumors associated with chronic inflammation, and PDE4B potential clinical application.

Immunohistochemical algorithms and gene expression profiling in primary cutaneous B-cell lymphoma

OBJECTIVE

to assess whether immunohistochemical (IHC) algorithms used to classify the cell of origin (COO) of nodal Diffuse Large B-cell lymphoma (nDLBCL) in Germinal Center type (GCB) and non-GCB subtypes may be applied to Primary Cutaneous B-cell lymphoma (PCBCL) too, and which of these algorithms performs better on PCBCL.

DESIGN

Retrospective case control study.

SETTING

Pathology Department of the University Hospital "San Giovanni di Dio e Ruggi d'Aragona" Salerno, Italy.

PARTICIPANTS

Fourteen PCBCL, including Primary Cutaneous follicle centre lymphoma (PCFCL) and primary cutaneous diffuse large B-cell lymphoma, Leg type (PCDLBCL-LT) and 14 nDLBCL were evaluated for 7-year period (January 2011 to December 2017). Primary cutaneous marginal zone cell lymphoma (PCMZL) cases were not included in the present study.

INTERVENTION

Evaluation of immunohistochemical CD10, BCL6, MUM1/IRF4, BCL2, MYC and Ki-67 expression and classification according to three different algorithms. Gene expression profiling (GEP) was performed on the same series using Lymph2Cx assay (Nanostring). The data obtained were compared and analysed.

RESULTS

All the IHC algorithms showed 13 GCB and 15 non-GCB. GEP showed 12 GCB, 12 activated B cell-type and 4 unclassified.

CONCLUSIONS

The PCBCL were classifiable as GCB and non-GCB like the nDLBCL as IHC algorithms were concordant to GEP and produced the same results.

Machine Learning Applications in the Diagnosis of Benign and Malignant Hematological Diseases

The use of machine learning (ML) and deep learning (DL) methods in hematology includes diagnostic, prognostic, and therapeutic applications. This increase is due to the improved access to ML and DL tools and the expansion of medical data. The utilization of ML remains limited in clinical practice, with some disciplines further along in their adoption, such as radiology and histopathology. In this review, we discuss the current uses of ML in diagnosis in the field of hematology, including image-recognition, laboratory, and genomics-based diagnosis. Additionally, we provide an introduction to the fields of ML and DL, highlighting current trends, limitations, and possible areas of improvement.

IQGAP2 acts as an independent prognostic factor and is related to immunosuppression in DLBCL

Background

Almost one-third of patients with diffuse large B-cell lymphoma (DLBCL) cannot be cured with initial therapy and will eventually succumb to the disease. Further elaboration of prognostic markers of DLBCL will provide therapeutic targets. IQ motif-containing GTPase activating protein 2 (IQGAP2) acts as a tumour suppressor in hepatocellular, prostate, and gastric cancers. However, the role of IQGAP2 in DLBCL remains unclear.

Methods

We collected mRNA expression data from 614 samples and the corresponding clinical information. The survival time of patients was compared between groups according to the mRNA expression level of IQGAP2. Survival analyses were performed in different subgroups when considering the effect of age, tumour stage, serum lactate dehydrogenase (LDH) concentration, performance status, and the number of extra nodal disease sites. The biological processes associated with IQGAP2-associated mRNAs were analysed to predict the function of IQGAP2. The correlation of IQGAP2 mRNA with immunosuppressive genes and leukocyte infiltration were analysed.

Results

The overall survival of patients with increased IQGAP2 mRNA levels was reduced even after aggressive treatment independent of age, tumour stage, serum LDH concentration, performance status, and the number of extra nodal disease sites. Furthermore, the biological processes of IQGAP2-associated mRNAs were mainly immune processes. IQGAP2 mRNA expression was correlated with the expression of immunosuppressive genes and leukocyte infiltration.

Conclusion

IQGAP2 mRNA is an independent prognostic factor and is related to immunosuppression in DLBCL. This discovery may provide a promising target for further development of therapy.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-021-08086-y.

Translational Applications of Artificial Intelligence and Machine Learning for Diagnostic Pathology in Lymphoid Neoplasms: A Comprehensive and Evolutive Analysis

Genomic analysis and digitalization of medical records have led to a big data scenario within hematopathology. Artificial intelligence and machine learning tools are increasingly used to integrate clinical, histopathological, and genomic data in lymphoid neoplasms. In this study, we identified global trends, cognitive, and social framework of this field from 1990 to 2020. Metadata were obtained from the Clarivate Analytics Web of Science database in January 2021. A total of 525 documents were assessed by document type, research areas, source titles, organizations, and countries. SciMAT and VOSviewer package were used to perform scientific mapping analysis. Geographical distribution showed the USA and People's Republic of China as the most productive countries, reporting up to 190 (36.19%) of all documents. A third-degree polynomic equation predicts that future global production in this area will be three-fold the current number, near 2031. Thematically, current research is focused on the integration of digital image analysis and genomic sequencing in Non-Hodgkin lymphomas, prediction of chemotherapy response and validation of new prognostic models. These findings can serve pathology departments to depict future clinical and research avenues, but also, public institutions and administrations to promote synergies and optimize funding allocation.

Role of Roflumilast Combined with ESHAP Chemotherapy in Relapsed/Refractory Patients with Diffuse Large B-Cell Lymphoma

Purpose

There are unmet needs associated with the current treatment strategies for relapsed/refractory diffuse large B-cell lymphoma (DLBCL) due to the poor treatment outcomes of these strategies. Roflumilast, a selective phosphodiesterase-4 inhibitor used for treating chronic obstructive pulmonary disease, is effective against B-cell malignancy via phosphoinositide 3-kinase (PI3K)–activity suppression. We analyzed the effects of roflumilast combined with ESHAP (etoposide, cisplatin, methylprednisolone, and cytarabine) chemotherapy in experimental and clinical settings.

Materials and Methods

An in vitro study using lymphoma cell lines and a pilot study on relapsed/refractory DLBCL patients were conducted to investigate the effects and mechanism of the combination of roflumilast and chemotherapy. The complete response (CR), overall response rate (ORR), and 1-year progression-free survival (PFS) were analyzed.

Results

We found that roflumilast is efficient when combined with other chemotherapy drugs, especially cytarabine. Synergistic effects between these two drugs influence the translation of mammalian target of rapamycin and myeloid cell leukemia 1, resulting in apoptosis and inhibition of B-cell lymphoma proliferation. In clinical setting, the roflumilast group showed better rates of CR (46.2% vs. 34.6%), ORR (76.9% vs. 53.8%), and 1-year PFS (50.0% vs. 25.9%) compared with the control group, though not statistically significant. The roflumilast group showed a higher incidence of asthenia and gastrointestinal adverse events. However, grade 3 or 4 adverse events were similar in both groups.

Conclusion

We found that roflumilast, when combined with ESHAP chemotherapy, for relapsed/refractory DLBCL was clinically active and well tolerated. This combined treatment was able to suppress PI3K activity, which is correlated with the degree of clinical response.

TIPE Regulates DcR3 Expression and Function by Activating the PI3K/AKT Signaling Pathway in CRC

Tumor necrosis factor-induced protein-8 (TIPE) is highly expressed in colorectal cancer (CRC). Decoy receptor 3 (DcR3) is a soluble secreted protein that can antagonize Fas ligand (FasL)-induced apoptosis and promote tumorigenesis. It remains unclear whether TIPE can regulate DcR3 expression. In this study, we examined this question by analyzing the relationship between these factors in CRC. Bioinformatics and tissue microarrays were used to determine the expression of TIPE and DcR3 and their correlation in CRC. The expression of TIPE and DcR3 in colon cancer cells was detected. Plasma samples were collected from CRC patients, and DcR3 secretion was measured. Then, dual-luciferase reporter gene analysis was performed to assess the interaction between TIPE and DcR3. We exogenously altered TIPE expression and analyzed its function and influence on DcR3 secretion. Lipopolysaccharide (LPS) was used to stimulate TIPE-overexpressing HCT116 cells, and alterations in signaling pathways were detected. Additionally, inhibitors were used to confirm molecular mechanisms. We found that TIPE and DcR3 were highly expressed in CRC patients and that their expression levels were positively correlated. DcR3 was highly expressed in the plasma of cancer patients. We confirmed that TIPE and DcR3 were highly expressed in HCT116 cells. TIPE overexpression enhanced the transcriptional activity of the DcR3 promoter. TIPE activated the PI3K/AKT signaling pathway to regulate the expression of DcR3, thereby promoting cell proliferation and migration and inhibiting apoptosis. In summary, TIPE and DcR3 are highly expressed in CRC, and both proteins are associated with poor prognosis. TIPE regulates DcR3 expression by activating the PI3K/AKT signaling pathway in CRC, thus promoting cell proliferation and migration and inhibiting apoptosis. These findings may have clinical significance and promise for applications in the treatment or prognostication of CRC.

Artificial intelligence: A new budding star in gastric cancer

The pursuit of health has always been the driving force for the advancement of human society, and social development will be profoundly affected by every breakthrough in the medical industry. With the arrival of the information technology revolution era, artificial intelligence (AI) technology has been rapidly developed. AI has been combined with medicine but it has been less studied with gastric cancer (GC). AI is a new budding star in GC, and its contribution to GC is mainly focused on diagnosis and treatment. For early GC, AI’s impact is not only reflected in its high accuracy but also its ability to quickly train primary doctors, improve the diagnosis rate of early GC, and reduce missed cases. At the same time, it will also reduce the possibility of missed diagnosis of advanced GC in cardia. Furthermore, it is used to assist imaging doctors to determine the location of lymph nodes and, more importantly, it can more effectively judge the lymph node metastasis of GC, which is conducive to the prognosis of patients. In surgical treatment of GC, it also has great potential. Robotic surgery is the latest technology in GC surgery. It is a bright star for minimally invasive treatment of GC, and together with laparoscopic surgery, it has become a common treatment for GC. Through machine learning, robotic systems can reduce operator errors and trauma of patients, and can predict the prognosis of GC patients. Throughout the centuries of development of surgery, the history gradually changes from traumatic to minimally invasive. In the future, AI will help GC patients reduce surgical trauma and further improve the efficiency of minimally invasive treatment of GC.

Classification of Current Procedural Terminology Codes from Electronic Health Record Data Using Machine Learning

BACKGROUND

Accurate anesthesiology procedure code data are essential to quality improvement, research, and reimbursement tasks within anesthesiology practices. Advanced data science techniques, including machine learning and natural language processing, offer opportunities to develop classification tools for Current Procedural Terminology codes across anesthesia procedures.

METHODS

Models were created using a Train/Test dataset including 1,164,343 procedures from 16 academic and private hospitals. Five supervised machine learning models were created to classify anesthesiology Current Procedural Terminology codes, with accuracy defined as first choice classification matching the institutional-assigned code existing in the perioperative database. The two best performing models were further refined and tested on a Holdout dataset from a single institution distinct from Train/Test. A tunable confidence parameter was created to identify cases for which models were highly accurate, with the goal of at least 95% accuracy, above the reported 2018 Centers for Medicare and Medicaid Services (Baltimore, Maryland) fee-for-service accuracy. Actual submitted claim data from billing specialists were used as a reference standard.

RESULTS

Support vector machine and neural network label-embedding attentive models were the best performing models, respectively, demonstrating overall accuracies of 87.9% and 84.2% (single best code), and 96.8% and 94.0% (within top three). Classification accuracy was 96.4% in 47.0% of cases using support vector machine and 94.4% in 62.2% of cases using label-embedding attentive model within the Train/Test dataset. In the Holdout dataset, respective classification accuracies were 93.1% in 58.0% of cases and 95.0% among 62.0%. The most important feature in model training was procedure text.

CONCLUSIONS

Through application of machine learning and natural language processing techniques, highly accurate real-time models were created for anesthesiology Current Procedural Terminology code classification. The increased processing speed and a priori targeted accuracy of this classification approach may provide performance optimization and cost reduction for quality improvement, research, and reimbursement tasks reliant on anesthesiology procedure codes.

Artificial Intelligence and Digital Microscopy Applications in Diagnostic Hematopathology

Digital Pathology is the process of converting histology glass slides to digital images using sophisticated computerized technology to facilitate acquisition, evaluation, storage, and portability of histologic information. By its nature, digitization of analog histology data renders it amenable to analysis using deep learning/artificial intelligence (DL/AI) techniques. The application of DL/AI to digital pathology data holds promise, even if the scope of use cases and regulatory framework for deploying such applications in the clinical environment remains in the early stages. Recent studies using whole-slide images and DL/AI to detect histologic abnormalities in general and cancer in particular have shown encouraging results. In this review, we focus on these emerging technologies intended for use in diagnostic hematology and the evaluation of lymphoproliferative diseases.

Cell-of-origin determined by both gene expression profiling and immunohistochemistry is the strongest predictor of survival in patients with diffuse large B-cell lymphoma

The tumor cells in diffuse large B‐cell lymphomas (DLBCL) are considered to originate from germinal center derived B‐cells (GCB) or activated B‐cells (ABC). Gene expression profiling (GEP) is preferably used to determine the cell of origin (COO). However, GEP is not widely applied in clinical practice and consequently, several algorithms based on immunohistochemistry (IHC) have been developed. Our aim was to evaluate the concordance of COO assignment between the Lymph2Cx GEP assay and the IHC‐based Hans algorithm, to decide which model is the best survival predictor. Both GEP and IHC were performed in 359 homogenously treated Swedish and Danish DLBCL patients, in a retrospective multicenter cohort. The overall concordance between GEP and IHC algorithm was 72%; GEP classified 85% of cases assigned as GCB by IHC, as GCB, while 58% classified as non‐GCB by IHC, were categorized as ABC by GEP. There were significant survival differences (overall survival and progression‐free survival) if cases were classified by GEP, whereas if cases were categorized by IHC only progression‐free survival differed significantly. Importantly, patients assigned as non‐GCB/ABC both by IHC and GEP had the worst prognosis, which was also significant in multivariate analyses. Double expression of MYC and BCL2 was more common in ABC cases and was associated with a dismal outcome. In conclusion, to determine COO both by IHC and GEP is the strongest outcome predictor to identify DLBCL patients with the worst outcome.

Deep neural network-assisted computed tomography diagnosis of metastatic lymph nodes from gastric cancer

BACKGROUND

Artificial intelligence-assisted image recognition technology is currently able to detect the target area of an image and fetch information to make classifications according to target features. This study aimed to use deep neural networks for computed tomography (CT) diagnosis of perigastric metastatic lymph nodes (PGMLNs) to simulate the recognition of lymph nodes by radiologists, and to acquire more accurate identification results.

METHODS

A total of 1371 images of suspected lymph node metastasis from enhanced abdominal CT scans were identified and labeled by radiologists and were used with 18,780 original images for faster region-based convolutional neural networks (FR-CNN) deep learning. The identification results of 6000 random CT images from 100 gastric cancer patients by the FR-CNN were compared with results obtained from radiologists in terms of their identification accuracy. Similarly, 1004 CT images with metastatic lymph nodes that had been post-operatively confirmed by pathological examination and 11,340 original images were used in the identification and learning processes described above. The same 6000 gastric cancer CT images were used for the verification, according to which the diagnosis results were analyzed.

RESULTS

In the initial group, precision-recall curves were generated based on the precision rates, the recall rates of nodule classes of the training set and the validation set; the mean average precision (mAP) value was 0.5019. To verify the results of the initial learning group, the receiver operating characteristic curves was generated, and the corresponding area under the curve (AUC) value was calculated as 0.8995. After the second phase of precise learning, all the indicators were improved, and the mAP and AUC values were 0.7801 and 0.9541, respectively.

CONCLUSION

Through deep learning, FR-CNN achieved high judgment effectiveness and recognition accuracy for CT diagnosis of PGMLNs.

TRIAL REGISTRATION

Chinese Clinical Trial Registry, No. ChiCTR1800016787; http://www.chictr.org.cn/showproj.aspx?proj=28515.

Three inflammation-related genes could predict risk in prognosis and metastasis of patients with breast cancer

BACKGROUND

Current predictive model is not developed by inflammation-related genes to evaluate clinical outcome of breast cancer patients.

METHODS

With mRNA expression profiling, we identified 3 mRNAs with significant expression between 15 normal samples and 669 breast cancer patients. Using 7 cell lines and 150 paraffin-embedded specimens, we verified the expression pattern by bio-experiments. Then, we constructed a three-mRNA model by Cox regression method and approved its predictive accuracy in both training set (n = 1095) and 4 testing sets (n = 703).

RESULTS

We developed a three-mRNA (TBX21, TGIF2, and CYCS) model to stratify patients into high- and low-risk subgroup with significantly different prognosis. In training set, 5-year OS rate was 84.5% (78.8%-90.5%) vs 73.1% (65.9%-81.2%) for the low- and high-risk group (HR = 1.573 (1.090-2.271); P = 0.016). The predictive value was similar in four independent testing sets (HR>1.600; P < 0.05). This model could assess survival independently with better predictive power compared with single clinicopathological risk factors and any of the three mRNAs. Patients with both low-risk values and any poor prognostic factors had more favorable survival from nonmetastatic status (HR = 1.740 (1.028-2.945), P = 0.039). We established two nomograms for clinical application that integrated this model and another three significant risk factors to forecast survival rates precisely in patients with or without metastasis.

CONCLUSIONS

This model is a dependable tool to predict the disease recurrence precisely and could improve the predictive accuracy of survival probability for breast cancer patients with or without metastasis.

TBX21 predicts prognosis of patients and drives cancer stem cell maintenance via the TBX21-IL-4 pathway in lung adenocarcinoma

Background

The Th1 cell-specific transcription factor TBX21 functions as a regulator of expression of a Th1 cytokine, interferon gamma (IFN-γ). However, the specific function of TBX21 correlated with cancer stemness remains unclear.

Methods

Using univariate and multivariate survival analysis, TBX21was identified as an independent predictive factor and was associated with poor prognosis in 1389 patients with lung adenocarcinoma (LUAD). Its mechanism in the prognosis was explored by functional enrichment analysis and validated in bioexperiments.

Results

In the training and test sets, TBX21 could classify 1389 LUAD patients into high and low-risk groups with significantly different prognosis (P < 0.01). Its prognostic power was independent of other clinical factors including stage, age, gender and smoking status. Functional studies indicated that downregulating TBX21 in lung cancer cells decreased the fraction of cancer stem cells and their sphere and tumor initiation frequency. Furthermore, the study showed that TBX21 activation transduced a TBX21–IL-4 signaling cascade to promote tumor initiation, tumor growth and expression of stemness markers.

Conclusions

These data demonstrated a key role of TBX21 in the maintenance of cancer stemness and that the TBX21–IL-4 pathway is a crucial factor contributing to lung carcinogenesis.

Graphical abstract

TBX21 prognostic model correlated with cancer stemness via TBX21-IL-4 pathway in LUAD patients

Electronic supplementary material

The online version of this article (10.1186/s13287-018-0820-6) contains supplementary material, which is available to authorized users.

Molecular heterogeneity in diffuse large B-cell lymphoma and its implications in clinical diagnosis and treatment

Over half of patients with diffuse large B-cell lymphoma (DLBCL) can be cured by standard R-CHOP treatment (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone). However, the remaining patients are refractory and ultimately succumb to progressive or relapsed disease. During the past decade, there has been significant progress in the understanding of molecular mechanisms in DLBCL, largely owing to collaborative efforts in large-scale gene expression profiling and deep sequencing, which have identified genetic alterations critical in lymphomagenesis through activation of key signaling transduction pathways in DLBCL. These discoveries have not only led to the development of targeted therapies, including several currently in clinical trials, but also laid a solid foundation for the future identification of more effective therapies for patients not curable by R-CHOP. This review summarizes the recent advances in our understanding of the molecular characterization and pathogenesis of DLBCL and new treatment directions.

Data Science in the Research Domain Criteria Era: Relevance of Machine Learning to the Study of Stress Pathology, Recovery, and Resilience

Diverse environmental and biological systems interact to influence individual differences in response to environmental stress. Understanding the nature of these complex relationships can enhance the development of methods to: (1) identify risk, (2) classify individuals as healthy or ill, (3) understand mechanisms of change, and (4) develop effective treatments. The Research Domain Criteria (RDoC) initiative provides a theoretical framework to understand health and illness as the product of multiple inter-related systems but does not provide a framework to characterize or statistically evaluate such complex relationships. Characterizing and statistically evaluating models that integrate multiple levels (e.g. synapses, genes, environmental factors) as they relate to outcomes that a free from prior diagnostic benchmarks represents a challenge requiring new computational tools that are capable to capture complex relationships and identify clinically relevant populations. In the current review, we will summarize machine learning methods that can achieve these goals.

Biology Informs Treatment Choices in Diffuse Large B Cell Lymphoma

The effective deployment of rationally developed therapies for diffuse large B cell lymphoma (DLBCL) requires rapid assimilation of new biological data. Within this framework, here we address topical issues at the intersection of DLBCL biology and the clinic. We discuss targeting of B cell receptor (BCR) signaling, with emphasis on identifying patients who may benefit from this maneuver and how to best achieve it. We address strategies to modulate the DLBCL microenvironment, including the use of immune checkpoint inhibitors in selected DLBCL subsets, and the potential activity of alternative antiangiogenic therapies. Lastly, we highlight the emerging recognition of MYC and BCL2 coexpression as the most robust predictor of DLBCL outcome, and discuss rationally conceived experimental approaches to treat these high-risk patients.

Toll-like receptor 9 stimulation can induce IκBζ expression and IgM secretion in chronic lymphocytic leukemia cells

Chronic lymphocytic leukemia cells strongly depend on external stimuli for their survival. Both antigen receptor and co-stimulatory receptors, including Toll-like receptors, can modulate viability and proliferation of leukemic cells. Toll-like receptor ligands, and particularly the TLR9 ligand CpG, mediate heterogeneous responses in patients' samples reflecting the clinical course of the subjects. However, the molecular framework of the key signaling events underlying such heterogeneity is undefined. We focused our studies on a subset of chronic lymphocytic leukemia cases characterized by expression of CD38 and unmutated immunoglobulin genes, who respond to CpG with enhanced metabolic cell activity. We report that, while CpG induces NFKBIZ mRNA in all the samples analyzed, it induces the IκBζ protein in a selected group of cases, through an unanticipated post-transcriptional mechanism. Interestingly, IκBζ plays a causal role in sustaining CpG-induced cell viability and chemoresistance, and CpG stimulation can unleash immunoglobulin secretion by IκBζ-positive malignant cells. These results identify and characterize IκBζ as a marker and effector molecule of distinct key pathways in chronic lymphocytic leukemia.

O-Linked N-Acetylglucosamine (O-GlcNAc) Expression Levels Epigenetically Regulate Colon Cancer Tumorigenesis by Affecting the Cancer Stem Cell Compartment via Modulating Expression of Transcriptional Factor MYBL1

To study the regulation of colorectal adenocarcinoma progression by O-GlcNAc, we have focused on the O-GlcNAc-mediated epigenetic regulation of human colon cancer stem cells (CCSC). Xenograft tumors from colon tumor cells with O-linked N-acetylglucosamine transferase (OGT) knockdown grew significantly slower than those formed from control cells, indicating a reduced proliferation of tumor cells due to inhibition of OGT expression. Significant reduction of the CCSC population was observed in the tumor cells after OGT knockdown, whereas tumor cells treated with the O-GlcNAcase inhibitor showed an increased CCSC population, indicating that O-GlcNAc levels regulated the CCSC compartment. When grown in suspension, tumor cells with OGT knockdown showed a reduced ability to form tumorspheres, indicating a reduced self-renewal of CCSC due to reduced levels of O-GlcNAc. ChIP-sequencing experiments using an anti-O-GlcNAc antibody revealed significant chromatin enrichment of O-GlcNAc-modified proteins at the promoter of the transcription factor MYBL1, which was also characterized by the presence of H3K27me3. RNA-sequencing analysis showed an increased expression of MYBL1 in tumor cells with OGT knockdown. Forced overexpression of MYBL1 led to a reduced population of CCSC and tumor growth in vivo, similar to the effects of OGT silencing. Moreover, two CpG islands near the transcription start site of MYBL1 were identified, and O-GlcNAc levels regulated their methylation status. These results strongly argue that O-GlcNAc epigenetically regulates MYBL1, functioning similarly to H3K27me3. The aberrant CCSC compartment observed after modulating O-GlcNAc levels is therefore likely to result, at least in part, from the epigenetic regulation of MYBL1 expression by O-GlcNAc, thereby significantly affecting tumor progression.

Phosphodiesterase 4 inhibitors have wide-ranging activity in B-cell malignancies

Phosphodiesterase 4 (PDE4) inhibition restores the suppressive effects of 3',5'-cyclic adenosine monophosphate in lymphocytes. In this concise review, we detail how PDE4 inhibition downmodulates the B-cell receptor (BCR)-related kinases spleen tyrosine kinase and phosphatidylinositol 3-kinase and inhibits vascular endothelial growth factor A secretion by tumor cells, inducing cancer cell apoptosis and blocking angiogenesis in the microenvironment. We describe the successful clinical repurposing of PDE4 inhibitors in B-cell malignancies, and propose that given their anti-inflammatory/immunomodulatory activity, these agents will suppress BCR signals without the toxicity associated with other targeted biological doublets.

Safety and Pharmacodynamics of the PDE4 Inhibitor Roflumilast in Advanced B-cell Malignancies

Purpose: In this study, we aimed to validate our extensive preclinical data on phosphodiesterase 4 (PDE4) as actionable target in B-cell malignancies. Our specific objectives were to determine the safety, pharmacokinetics, and pharmacodynamics (PI3K/AKT activity), as well as to capture any potential antitumor activity of the PDE4 inhibitor roflumilast in combination with prednisone in patients with advanced B-cell malignancies.Experimental Design: Single-center, exploratory phase Ib open-label, nonrandomized study. Roflumilast (500 mcg PO) was given daily for 21 days with prednisone on days 8 to 14. Additional 21-day cycles were started if patients tolerated cycle 1 and had at least stable disease.Results: Ten patients, median age 65 years with an average of three prior therapies, were enrolled. The median number of cycles administered was 4 (range, 1-13). Treatment was well tolerated; the most common ≥grade 2 treatment-related adverse events were fatigue, anorexia (≥25%), and transient ≥ grade 2 neutropenia (30%). Treatment with roflumilast as a single agent significantly suppressed PI3K activity in the 77% of patients evaluated; on average, patients with PI3K/AKT suppression stayed in trial for 156 days (49-315) versus 91 days (28-139 days) for those without this biomarker response. Six of the nine evaluable patients (66%) had partial response or stable disease. The median number of days in trial was 105 days (range, 28-315).Conclusions: Repurposing the PDE4 inhibitor roflumilast for treatment of B-cell malignancies is safe, suppresses the oncogenic PI3K/AKT kinases, and may be clinically active. Clin Cancer Res; 23(5); 1186-92. ©2016 AACR.

Machine learning-based classification of diffuse large B-cell lymphoma patients by eight gene expression profiles

Gene expression profiling (GEP) had divided the diffuse large B‐cell lymphoma (DLBCL) into molecular subgroups: germinal center B‐cell like (GCB), activated B‐cell like (ABC), and unclassified (UC) subtype. However, this classification with prognostic significance was not applied into clinical practice since there were more than 1000 genes to detect and interpreting was difficult. To classify cancer samples validly, eight significant genes (MYBL1, LMO2, BCL6, MME, IRF4, NFKBIZ, PDE4B, and SLA) were selected in 414 patients treated with CHOP/R‐CHOP chemotherapy from Gene Expression Omnibus (GEO) data sets. Cutoffs for each gene were obtained using receiver–operating characteristic curves (ROC) new model based on the support vector machine (SVM) estimated the probability of membership into one of two subgroups: GCB and Non‐GCB (ABC and UC). Furtherly, multivariate analysis validated the model in another two cohorts including 855 cases in all. As a result, patients in the training and validated cohorts were stratified into two subgroups with 94.0%, 91.0%, and 94.4% concordance with GEP, respectively. Patients with Non‐GCB subtype had significantly poorer outcomes than that with GCB subtype, which agreed with the prognostic power of GEP classification. Moreover, the similar prognosis received in the low (0–2) and high (3–5) IPI scores group demonstrated that the new model was independent of IPI as well as GEP method. In conclusion, our new model could stratify DLBCL patients with CHOP/R‐CHOP regimen matching GEP subtypes effectively.