Systematic analysis of ChatGPT, Google search and Llama 2 for clinical decision support tasks.

It is likely that individuals are turning to Large Language Models (LLMs) to seek health advice, much like searching for diagnoses on Google. We evaluate clinical accuracy of GPT-3·5 and GPT-4 for suggesting initial diagnosis, examination steps and treatment of 110 medical cases across diverse clinical disciplines. Moreover, two model configurations of the Llama 2 open source LLMs are assessed in a sub-study. For benchmarking the diagnostic task, we conduct a naïve Google search for comparison. Overall, GPT-4 performed best with superior performances over GPT-3·5 considering diagnosis and examination and superior performance over Google for diagnosis. Except for treatment, better performance on frequent vs rare diseases is evident for all three approaches. The sub-study indicates slightly lower performances for Llama models. In conclusion, the commercial LLMs show growing potential for medical question answering in two successive major releases. However, some weaknesses underscore the need for robust and regulated AI models in health care. Open source LLMs can be a viable option to address specific needs regarding data privacy and transparency of training.

GEFAAR: a generic framework for the analysis of antimicrobial resistance providing statistics and cluster analyses.

Easy access to antimicrobial resistance data and meaningful visualization is essential to guide the empirical antimicrobial treatment and to promote the rational use of antimicrobial agents. Currently available solutions are commonly externally hosted, centralized systems. However, there is a need for close monitoring by local analysis tools. To fill this gap, we developed GEFAAR-a generic framework for the analysis of antimicrobial resistance data. Following the example of the German Robert Koch Institute (RKI), an interactive web-application is provided to determine basic pathogen and resistance statistics. In addition to the RKI's externally maintained database, our application provides a generic framework to import tabular data and to analyze them safely in a local environment. Moreover, our application offers an intuitive web-based user interface to visualize resistance trend analysis as well as advanced cluster analyses on species- or clinic/unit level to generate alerts of potential transmission events.

Knockout of the Cardiac Transcription Factor NKX2-5 Results in Stem Cell-Derived Cardiac Cells with Typical Purkinje Cell-like Signal Transduction and Extracellular Matrix Formation.

The human heart controls blood flow, and therewith enables the adequate supply of oxygen and nutrients to the body. The correct function of the heart is coordinated by the interplay of different cardiac cell types. Thereby, one can distinguish between cells of the working myocardium, the pace-making cells in the sinoatrial node (SAN) and the conduction system cells in the AV-node, the His-bundle or the Purkinje fibres. Tissue-engineering approaches aim to generate hiPSC-derived cardiac tissues for disease modelling and therapeutic usage with a significant improvement in the differentiation quality of myocardium and pace-making cells. The differentiation of cells with cardiac conduction system properties is still challenging, and the produced cell mass and quality is poor. Here, we describe the generation of cardiac cells with properties of the cardiac conduction system, called conduction system-like cells (CSLC). As a primary approach, we introduced a CrispR-Cas9-directed knockout of the NKX2-5 gene in hiPSC. NKX2-5-deficient hiPSC showed altered connexin expression patterns characteristic for the cardiac conduction system with strong connexin 40 and connexin 43 expression and suppressed connexin 45 expression. Application of differentiation protocols for ventricular- or SAN-like cells could not reverse this connexin expression pattern, indicating a stable regulation by NKX2-5 on connexin expression. The contraction behaviour of the hiPSC-derived CSLCs was compared to hiPSC-derived ventricular- and SAN-like cells. We found that the contraction speed of CSLCs resembled the expected contraction rate of human conduction system cells. Overall contraction was reduced in differentiated cells derived from NKX2-5 knockout hiPSC. Comparative transcriptomic data suggest a specification of the cardiac subtype of CSLC that is distinctly different from ventricular or pacemaker-like cells with reduced myocardial gene expression and enhanced extracellular matrix formation for improved electrical insulation. In summary, knockout of NKX2-5 in hiPSC leads to enhanced differentiation of cells with cardiac conduction system features, including connexin expression and contraction behaviour.

Immune gene co-expression signatures implicated in occurence and persistence of cognitive dysfunction in depression.

Cognitive dysfunction contributes significantly to the burden caused by Major Depressive Disorder (MDD). Yet, while compelling evidence suggests that different biological processes play a part in both MDD aetiology and the development of cognitive decline more generally, we only begin to understand the molecular underpinnings of depression-related cognitive impairment. Developments in psychometric assessments, molecular high-throughput methods and systems biology derived analysis strategies advance this endeavour. Here, we aim to identify gene expression signatures associated with cognitive dysfunction and cognitive improvement following therapy using RNA sequencing to analyze the whole blood-derived transcriptome of altogether 101 MDD patients who enrolled in the CERT-D study. The mRNA(Nova)Seq based transcriptome was analyzed from whole blood taken at baseline assessment, and patients' cognitive performance was measured twice at baseline and following eight weeks of therapy by means of the THINC integrated tool. Thirty-six patients showed comparatively low cognitive performance at baseline assessment, and 32 patients showed comparatively strong cognitive improvement following therapy. Differential gene expression analysis was performed using limma to a significance threshold of 0.05 and a logFC cutoff of |1.2|. Although we observed some indications for expression differences related to low cognitive performance and cognitive therapy response, signals did not withstand adjustment for multiple testing. Applying WGCNA, we retrieved altogether 25 modules of co-expressed genes and we used a combination of correlational and linear analyses to identify modules related to baseline cognitive performance and cognitive improvement following therapy. Three immune modules reflected distinct but interrelated immune processes (the yellow module: neutrophil-mediated immunity, the darkorange module: interferon signaling, the tan module: platelet activation), and higher expression of the yellow (r = -0.21, p < .05), the dark orange (r = 0.2, p < .05), and the tan (r = -0.23, p < .05) module correlated significantly negatively with patients' cognitive baseline performance. Patients' cognitive baseline performance was a significant predictor of the darkorange module (b = -0.039, p < .05) and the tan module's expression (b = 0.02, p < .05) and was close to becoming a significant predictor of the yellow module's expression (b = -0.02, p = .05). Furthermore, patients characterized by comparatively low cognitive performance at baseline showed significantly higher expression of the tan module when compared to all other patients F(1,97) = 4.32, p < .05, η= 0.04. Following eight weeks of treatment, we observed altogether significant improvement in patients' cognitive performance (b = 0.30, p < .001), and patients with comparatively high cognitive gain showed noticeably lower, but not significantly lower F(1,98) = 3.76, p = .058, expression of a dark turquoise module, which reflects complement and B-cell-associated immune processes. Noteworthy, the relation between cognitive performance and module expression remained observable after controlling for symptom severity and BMI, which partly accounted for variance in module expression. As such, our findings provide further evidence for the involvement of immune processes in MDD related cognitive dysfunction and they suggest that different immune processes contribute to the development and long-term persistence of cognitive dysfunction in the context of depression.

OGRE: calculate, visualize, and analyze overlap between genomic input regions and public annotations.

BACKGROUND: Modern genome sequencing leads to an ever-growing collection of genomic annotations. Combining these elements with a set of input regions (e.g. genes) would yield new insights in genomic associations, such as those involved in gene regulation. The required data are scattered across different databases making a manual approach tiresome, unpractical, and prone to error. Semi-automatic approaches require programming skills in data parsing, processing, overlap calculation, and visualization, which most biomedical researchers lack. Our aim was to develop an automated tool providing all necessary algorithms, benefiting both bioinformaticians and researchers without bioinformatic training. RESULTS: We developed overlapping annotated genomic regions (OGRE) as a comprehensive tool to associate and visualize input regions with genomic annotations. It does so by parsing regions of interest, mining publicly available annotations, and calculating possible overlaps between them. The user can thus identify location, type, and number of associated regulatory elements. Results are presented as easy to understand visualizations and result tables. We applied OGRE to recent studies and could show high reproducibility and potential new insights. To demonstrate OGRE's performance in terms of running time and output, we have conducted a benchmark and compared its features with similar tools. CONCLUSIONS: OGRE's functions and built-in annotations can be applied as a downstream overlap association step, which is compatible with most genomic sequencing outputs, and can thus enrich pre-existing analyses pipelines. Compared to similar tools, OGRE shows competitive performance, offers additional features, and has been successfully applied to two recent studies. Overall, OGRE addresses the lack of tools for automatic analysis, local genomic overlap calculation, and visualization by providing an easy to use, end-to-end solution for both biologists and computational scientists.

Mitochondrial DNA mutations in Medulloblastoma.

To date, several studies on genomic events underlying medulloblastoma (MB) biology have expanded our understanding of this tumour entity and led to its division into four groups-WNT, SHH, group 3 (G3) and group 4 (G4). However, there is little information about the relevance of pathogenic mitochondrial DNA (mtDNA) mutations and their consequences across these. In this report, we describe the case of a female patient with MB and a mitochondriopathy, followed by a study of mtDNA variants in MB groups. After being diagnosed with G4 MB, the index patient was treated in line with the HIT 2000 protocol with no indications of relapse after five years. Long-term side effects of treatment were complemented by additional neurological symptoms and elevated lactate levels ten years later, resulting in suspected mitochondrial disease. This was confirmed by identifying a mutation in the MT-TS1 gene which appeared homoplasmic in patient tissue and heteroplasmic in the patient's mother. Motivated by this case, we explored mtDNA mutations across 444 patients from ICGC and HIT cohorts. While there was no statistically significant enrichment of mutations in one MB group, both cohorts encompassed a small group of patients harbouring potentially deleterious mtDNA variants. The case presented here highlights the possible similarities between sequelae caused by MB treatment and neurological symptoms of mitochondrial dysfunction, which may apply to patients across all MB groups. In the context of the current advances in characterising and interpreting mtDNA aberrations, recognising affected patients could enhance our future knowledge regarding the mutations' impact on carcinogenesis and cancer treatment.

Group-specific cellular metabolism in Medulloblastoma.

BACKGROUND: Cancer metabolism influences multiple aspects of tumorigenesis and causes diversity across malignancies. Although comprehensive research has extended our knowledge of molecular subgroups in medulloblastoma (MB), discrete analysis of metabolic heterogeneity is currently lacking. This study seeks to improve our understanding of metabolic phenotypes in MB and their impact on patients' outcomes. METHODS: Data from four independent MB cohorts encompassing 1,288 patients were analysed. We explored metabolic characteristics of 902 patients (ICGC and MAGIC cohorts) on bulk RNA level. Moreover, data from 491 patients (ICGC cohort) were searched for DNA alterations in genes regulating cell metabolism. To determine the role of intratumoral metabolic differences, we examined single-cell RNA-sequencing (scRNA-seq) data from 34 additional patients. Findings on metabolic heterogeneity were correlated to clinical data. RESULTS: Established MB groups exhibit substantial differences in metabolic gene expression. By employing unsupervised analyses, we identified three clusters of group 3 and 4 samples with distinct metabolic features in ICGC and MAGIC cohorts. Analysis of scRNA-seq data confirmed our results of intertumoral heterogeneity underlying the according differences in metabolic gene expression. On DNA level, we discovered clear associations between altered regulatory genes involved in MB development and lipid metabolism. Additionally, we determined the prognostic value of metabolic gene expression in MB and showed that expression of genes involved in metabolism of inositol phosphates and nucleotides correlates with patient survival. CONCLUSION: Our research underlines the biological and clinical relevance of metabolic alterations in MB. Thus, distinct metabolic signatures presented here might be the first step towards future metabolism-targeted therapeutic options.

Pharyngeal Communities and Antimicrobial Resistance in Pangolins in Gabon.

Wildlife can be a reservoir and source of zoonotic pathogens for humans. For instance, pangolins were considered one of the potential animal reservoirs of SARS-CoV-2. The aim of this study was to assess the prevalence of antimicrobial-resistant species (e.g., extended-spectrum ß-lactamase [ESBL]-producing Enterobacterales) and Staphylococcus aureus-related complex and to describe the bacterial community in wild Gabonese pangolins. The pharyngeal colonization of pangolins sold in Gabon (n = 89, 2021 to 2022) was analyzed using culture media selective for ESBL-producing Enterobacterales, S. aureus-related complex, Gram-positive bacteria and nonfermenters. Phylogenetic analyses of ESBL-producing Enterobacterales was done using core-genome multilocus sequence typing (cgMLST) and compared with publicly available genomes. Patterns of cooccurring species were detected by network analysis. Of the 439 bacterial isolates, the majority of species belonged to the genus Pseudomonas (n = 170), followed by Stenotrophomonas (n = 113) and Achromobacter (n = 37). Three Klebsiella pneumoniae isolates and one Escherichia coli isolate were ESBL-producers, which clustered with human isolates from Nigeria (MLST sequence type 1788 [ST1788]) and Gabon (ST38), respectively. Network analysis revealed a frequent cooccurrence of Stenotrophomonas maltophilia with Pseudomonas putida and Pseudomonas aeruginosa. In conclusion, pangolins can be colonized with human-related ESBL-producing K. pneumoniae and E. coli. Unlike in other African wildlife, S. aureus-related complex was not detected in pangolins. IMPORTANCE There is an ongoing debate if pangolins are a relevant reservoir for viruses such as SARS-CoV-2. Here, we wanted to know if African pangolins are colonized with bacteria that are relevant for human health. A wildlife reservoir of antimicrobial resistance would be of medical relevance in regions were consumption of so-called bushmeat is common. In 89 pangolins, we found three ESBL-producing Klebsiella pneumoniae strains and one ESBL-producing Escherichia coli strains, which were closely related to isolates from humans in Africa. This points toward either a transmission between pangolins and humans or a common source from which both humans and pangolins became colonized.

Research article network analysis of polymicrobial chronic wound infections in Masanga, Sierra Leone.

BACKGROUND: Chronic wounds are frequently colonized or infected with multiple bacterial or fungal species, which can both promote or inhibit each other. Network analyses are helpful to understand the interplay of these species in polymicrobial infections. Our aim was to analyse the network of bacterial and fungal species in chronic wounds. METHODS: Swabs (n = 163) from chronic wound infections (Masanga, Sierra Leone, 2019-2020) were screened for bacterial and fungal species using non-selective agars. Some of these wounds were suspected but not confirmed Buruli ulcer. Species identification was done with MALDI-TOF mass spectrometry. Network analysis was performed to investigate co-occurrence of different species within one patient. All species with n ≥ 10 isolates were taken into account. RESULTS: Of the 163 patients, 156 had a positive wound culture (median of three different species per patient; range 1-7). Pseudomonas aeruginosa (n = 75) was the dominating species with frequent co-detections of Klebsiella pneumoniae (21 cases; OR = 1.36, 95%CI: 0.63-2.96, p = 0.47), Staphylococcus aureus (14 cases; OR = 1.06, 95%CI: 0.44-2.55, p = 1) and Proteus mirabilis (13 cases; OR = 0.84, 95%CI: 0.35-1.99, p = 0.69). CONCLUSION: The culturome of chronic wounds in Sierra Leonean patients is highly diverse and characterized by the co-occurrence of P. aeruginosa, K. pneumoniae and S. aureus.

Reconstructing Clonal Evolution-A Systematic Evaluation of Current Bioinformatics Approaches.

The accurate reconstruction of clonal evolution, including the identification of newly developing, highly aggressive subclones, is essential for the application of precision medicine in cancer treatment. Reconstruction, aiming for correct variant clustering and clonal evolution tree reconstruction, is commonly performed by tedious manual work. While there is a plethora of tools to automatically generate reconstruction, their reliability, especially reasons for unreliability, are not systematically assessed. We developed clevRsim-an approach to simulate clonal evolution data, including single-nucleotide variants as well as (overlapping) copy number variants. From this, we generated 88 data sets and performed a systematic evaluation of the tools for the reconstruction of clonal evolution. The results indicate a major negative influence of a high number of clones on both clustering and tree reconstruction. Low coverage as well as an extreme number of time points usually leads to poor clustering results. An underlying branched independent evolution hampers correct tree reconstruction. A further major decline in performance could be observed for large deletions and duplications overlapping single-nucleotide variants. In summary, to explore the full potential of reconstructing clonal evolution, improved algorithms that can properly handle the identified limitations are greatly needed.

The Role of Clonal Evolution on Progression, Blood Parameters, and Response to Therapy in Multiple Myeloma.

Introduction: A variety of biomarkers are considered for diagnosis (e.g., ß2-microgobulin, albumin, or LDH) and prognosis [e.g., cytogenetic aberrations detected by fluorescence in situ hybridization (FISH)] of multiple myeloma (MM). More recently, clonal evolution has been established as key. Little is known on the clinical implications of clonal evolution. Methods: We performed in-depth analyses of 25 patients with newly diagnosed MM with respect to detailed clinical information analyzing blood samples collected at several time points during follow-up (median follow-up: 3.26 years since first diagnosis). We split our cohort into two subgroups: with and without new FISH clones developing in the course of disease. Results: Each subgroup showed a characteristic chromosomal profile. Forty-three percent of patients had evidence of appearing new clones. The patients with new clones showed an increased number of translocations affecting chromosomes 14 (78% vs. 33%; p = 0.0805) and 11, and alterations in chromosome 4 (amplifications and translocations). New clones, on the contrary, were characterized by alterations affecting chromosome 17. Subsequent to the development of the new clone, 6 out of 9 patients experienced disease progression compared to 3 out of 12 for patients without new clones. Duration of the therapy applied for the longest time was significantly shorter within the group of patients developing new clones (median: 273 vs. 406.5 days; p = 0.0465). Discussion: We demonstrated that the development of new clones, carrying large-scale alterations, was associated with inferior disease course and shorter response to therapy, possibly affecting progression-free survival and overall survival as well. Further studies evaluating larger cohorts are necessary for the validation of our results.

Clonal Evolution at First Sight: A Combined Visualization of Diverse Diagnostic Methods Improves Understanding of Leukemic Progression.

Patients with myeloid neoplasia are classified by the WHO classification systems. Besides clinical and hematological criteria, cytogenetic and molecular genetic alterations highly impact treatment stratification. In routine diagnostics, a combination of methods is used to decipher different types of genetic variants. Eight patients were comprehensively analyzed using karyotyping, fluorescence in situ hybridization, array-CGH and a custom NGS panel. Clonal evolution was reconstructed manually, integrating all mutational information on single nucleotide variants (SNVs), insertions and deletions (indels), structural variants and copy number variants (CNVs). To allow a correct integration, we differentiate between three scenarios: 1) CNV occurring prior to the SNV/indel, but in the same cells. 2) SNV/indel occurring prior to the CNV, but in the same cells. 3) SNV/indel and CNV existing in parallel, independent of each other. Applying this bioinformatics approach, we reconstructed clonal evolution for all patients. This generalizable approach offers the possibility to integrate various data to analyze identification of driver and passenger mutations as well as possible targets for personalized medicine approaches. Furthermore, this model can be used to identify markers to assess the minimal residual disease.

Clinical relevance of molecular characteristics in Burkitt lymphoma differs according to age.

While survival has improved for Burkitt lymphoma patients, potential differences in outcome between pediatric and adult patients remain unclear. In both age groups, survival remains poor at relapse. Therefore, we conducted a comparative study in a large pediatric cohort, including 191 cases and 97 samples from adults. While TP53 and CCND3 mutation frequencies are not age related, samples from pediatric patients showed a higher frequency of mutations in ID3, DDX3X, ARID1A and SMARCA4, while several genes such as BCL2 and YY1AP1 are almost exclusively mutated in adult patients. An unbiased analysis reveals a transition of the mutational profile between 25 and 40 years of age. Survival analysis in the pediatric cohort confirms that TP53 mutations are significantly associated with higher incidence of relapse (25 ± 4% versus 6 ± 2%, p-value 0.0002). This identifies a promising molecular marker for relapse incidence in pediatric BL which will be used in future clinical trials.

Machine learning in the detection and management of atrial fibrillation.

Machine learning has immense novel but also disruptive potential for medicine. Numerous applications have already been suggested and evaluated concerning cardiovascular diseases. One important aspect is the detection and management of potentially thrombogenic arrhythmias such as atrial fibrillation. While atrial fibrillation is the most common arrhythmia with a lifetime risk of one in three persons and an increased risk of thromboembolic complications such as stroke, many atrial fibrillation episodes are asymptomatic and a first diagnosis is oftentimes only reached after an embolic event. Therefore, screening for atrial fibrillation represents an important part of clinical practice. Novel technologies such as machine learning have the potential to substantially improve patient care and clinical outcomes. Additionally, machine learning applications may aid cardiologists in the management of patients with already diagnosed atrial fibrillation, for example, by identifying patients at a high risk of recurrence after catheter ablation. We summarize the current state of evidence concerning machine learning and, in particular, artificial neural networks in the detection and management of atrial fibrillation and describe possible future areas of development as well as pitfalls. Typical data flow in machine learning applications for atrial fibrillation detection.

Design of a targeted next-generation DNA sequencing panel for pediatric T-cell lymphoblastic lymphoma to unravel biology and optimize treatment.

Low incidence and molecular heterogeneity of pediatric T-cell lymphoblastic lymphoma (T-LBL) require an international, large-scale effort to identify novel clinical biomarkers. The ongoing international clinical trial LBL2018 (NCT04043494) represents an ideal opportunity to implement a common analytic approach. Targeted next-generation sequencing is well-suited for this purpose; however, selection of relevant target genes for T-LBL remains subject of ongoing debates. Our group has recently designed and evaluated a first target panel of 80 candidate genes for T-LBL. The present study aimed at developing a novel optimized gene panel for large-scale application and to promote an international agreement on a common core panel. Small sequence variants obtained from our former study were systematically analyzed and classified with regards to pathogenic relevance, to prioritize candidate genes. Additional genes were curated from literature and online databases for a more comprehensive analysis of relevant functions and signaling pathways. The new target panel TGP-T-LBL entails 84 candidate genes which are key actors in NOTCH, PI3K-AKT, JAK-STAT, RAS signaling, epigenetic regulation, transcription, DNA repair, cell cycle regulation, and ribosomal function. From our former gene panel, 35 out of 80 candidate genes were selected for the novel panel. Forty-six out of 84 genes are currently being analyzed in the ongoing international trial LBL2018. Exploratory analysis of prognostic relevance on mutation-level suggested a potential association of PIK3CA variants c.1624G>A(p.Glu542Lys) and c.1633G>A(p.Glu545Lys) to occurrence of relapse, emphasizing particular relevance of mutation analysis in PI3K-AKT signaling. Our approach promotes comprehensive and clinically relevant mutational profiling of pediatric T-LBL.

Exploring Current Challenges and Perspectives for Automatic Reconstruction of Clonal Evolution.

BACKGROUND/AIM: In the field of cancer research, reconstructing clonal evolution is of major interest. The technique provides new insights for analysis and prediction of tumor development. However, reconstruction based on mutational data is characterized by several challenges. MATERIALS AND METHODS: By performing extensive literature research, we identified 51 currently available tools for reconstructing clonal evolution. By analyzing two cancer data sets (n=21), we investigated the applicability and performance of each tool. RESULTS: Seventeen out of 51 tools could be applied to our data. Correct clustering of variants can be observed for 4 patients in the presence of ≤3 clusters and ≥5 time points. Correct phylogenetic trees are determined for 10 patients. Accurate visualization is possible, by applying adjustments to the original algorithms. CONCLUSION: Despite bearing considerable potential, automatic reconstruction of clonal evolution remains challenging. To replace tedious manual reconstruction, further research including systematic error analyses using simulation tools needs to be conducted.

clevRvis: visualization techniques for clonal evolution.

BACKGROUND: A thorough analysis of clonal evolution commonly requires integration of diverse sources of data (e.g., karyotyping, next-generation sequencing, and clinical information). Subsequent to actual reconstruction of clonal evolution, detailed analysis and interpretation of the results are essential. Often, however, only few tumor samples per patient are available. Thus, information on clonal development and therapy effect may be incomplete. Furthermore, analysis of biallelic events-considered of high relevance with respect to disease course-can commonly only be realized by time-consuming analysis of the raw results and even raw sequencing data. RESULTS: We developed clevRvis, an R/Bioconductor package providing an extensive set of visualization techniques for clonal evolution. In addition to common approaches for visualization, clevRvis offers a unique option for allele-aware representation: plaice plots. Biallelic events may be visualized and inspected at a glance. Analyzing 4 public datasets, we show that plaice plots help to gain new insights into tumor development and investigate hypotheses on disease progression and therapy resistance. In addition to a graphical user interface, automatic phylogeny-aware color coding of the plots, and an approach to explore alternative trees, clevRvis provides 2 algorithms for fully automatic time point interpolation and therapy effect estimation. Analyzing 2 public datasets, we show that both approaches allow for valid approximation of a tumor's development in between measured time points. CONCLUSIONS: clevRvis represents a novel option for user-friendly analysis of clonal evolution, contributing to gaining new insights into tumor development.

Divergent Effects of EZH1 and EZH2 Protein Expression on the Prognosis of Patients with T-Cell Lymphomas.

T-cell lymphomas are highly heterogeneous and their prognosis is poor under the currently available therapies. Enhancers of zeste homologue 1 and 2 (EZH1/2) are histone H3 lysine-27 trimethyltransferases (H3K27me3). Despite the rapid development of new drugs inhibiting EZH2 and/or EZH1, the molecular interplay of these proteins and the impact on disease progression and prognosis of patients with T-cell lymphomas remains insufficiently understood. In this study, EZH1/2 mutation status was evaluated in 33 monomorphic epitheliotropic intestinal T-cell lymphomas by next generation sequencing and EZH1/2 and H3K27me3 protein expression levels were detected by immunohistochemistry in 46 T-cell lymphomas. Correlations with clinicopathologic features were analyzed and survival curves generated. No EZH1 mutations and one (3%) EZH2 missense mutation were identified. In univariable analysis, high EZH1 expression was associated with an improved overall survival (OS) and progression-free survival (PFS) whereas high EZH2 and H3K27me3 expression were associated with poorer OS and PFS. Multivariable analysis revealed EZH1 (hazard ratio (HR) = 0.183; 95% confidence interval (CI): 0.044-0.767; p = 0.020;) and EZH2 (HR = 8.245; 95% CI: 1.898-35.826; p = 0.005) to be independent, divergent prognostic markers for OS. In conclusion, EZH1/2 protein expression had opposing effects on the prognosis of T-cell lymphoma patients.