FHBF: Federated hybrid boosted forests with dropout rates for supervised learning tasks across highly imbalanced clinical datasets.

Although several studies have deployed gradient boosting trees (GBT) as a robust classifier for federated learning tasks (federated GBT [FGBT]), even with dropout rates (federated gradient boosting trees with dropout rate [FDART]), none of them have investigated the overfitting effects of FGBT across heterogeneous and highly imbalanced datasets within federated environments nor the effect of dropouts in the loss function. In this work, we present the federated hybrid boosted forests (FHBF) algorithm, which incorporates a hybrid weight update approach to overcome ill-posed problems that arise from overfitting effects during the training across highly imbalanced datasets in the cloud. Eight case studies were conducted to stress the performance of FHBF against existing algorithms toward the development of robust AI models for lymphoma development across 18 European federated databases. Our results highlight the robustness of FHBF, yielding an average loss of 0.527 compared with FGBT (0.611) and FDART (0.584) with increased classification performance (0.938 sensitivity, 0.732 specificity).

Boosting the performance of MALT lymphoma classification in patients with primary Sjögren's Syndrome through data augmentation: a case study.

Sjögren's Syndrome (SS) patients with mucosa associated lymphoid tissue lymphomas (MALTLs) and diffuse large B-cell lymphomas (DLBCLs) have 10-year survival rates of 80% and 40%, respectively. This highlights the unique biologic burden of the two histologic forms, as well as, the need for early detection and thorough monitoring of these patients. The lack of MALTL patients and the fact that most studies are single cohort and combine patients with different lymphoma subtypes narrow the understanding of MALTL progression. Here, we propose a data augmentation pipeline that utilizes an advanced synthetic data generator which is trained on a Pan European data hub with primary SS (pSS) patients to yield a high-quality synthetic data pool. The latter is used for the development of an enhanced MALTL classification model. Four scenarios were defined to assess the reliability of augmentation. Our results revealed an overall improvement in the accuracy, sensitivity, specificity, and AUC by 7%, 6.3%, 9%, and 6.3%, respectively. This is the first case study that utilizes data augmentation to reflect the progression of MALTL in pSS.

Computational and Functional Insights of Protein Misfolding in Neurodegeneration.

Protein folding is the process by which a polypeptide chain self-assembles into the correct three-dimensional structure, so that it ends up in the biologically active, native state. Under conditions of proteotoxic stress, mutations, or cellular aging, proteins can begin to aggregate into non-native structures such as ordered amyloid fibrils and plaques. Many neurodegenerative diseases involve the misfolding and aggregation of specific proteins into abnormal, toxic species. Experimental approaches including crystallography and AFM (atomic force microscopy)-based force spectroscopy are used to exploit the folding and structural characterization of protein molecules. At the same time, computational techniques through molecular dynamics, fold recognition, and structure prediction are widely applied in this direction. Benchmarking analysis for combining and comparing computational methodologies with functional studies can decisively unravel robust interactions between the side groups of the amino acid sequence and monitor alterations in intrinsic protein dynamics with high precision as well as adequately determine potent conformations of the folded patterns formed in the polypeptide structure.

A Mobile Application for Supporting and Monitoring Elderly Population to Perform the Interventions of the FINGER Study.

Alzheimer's disease is a progressive disease that is caused by the destruction of brain neurons. It seems it affects a large group of the world's population that is estimated around 47 million and is expected to triple by 2050. Slowly but surely, the patient's condition is deteriorating, due to the increase of symptom severity, rendering him/her in need of special care. A great percentage of these cases can be attributed to some common modifiable risk factors such as hypertension, obesity, a lack of exercise, alcohol misuse, smoking, unhealthy diet, and a low level of education. The Finnish Geriatric Intervention Study (FINGER Study) proves that some interventions focused on the abovementioned risk factors of the individual's daily life can contribute to delay the occurrence of Alzheimer's disease. Concurrently, the rapid development of smart devices encourages the use of health applications that provide guiding tools and suggestions based on the user's status. The outcome of this paper is the development of a mobile application, to implement and monitor the interventions proposed by the FINGER Study. Based on the user's profile, it offers the ability to evaluate the likelihood of cognitive decline, monitor the process, and help delay the disease's occurrence.

Signature-Based Computational Drug Repurposing for Amyotrophic Lateral Sclerosis.

Amyotrophic lateral sclerosis (ALS) is a late-onset fatal neurodegenerative disease characterized by progressive loss of the upper and lower motor neurons. There are currently limited approved drugs for the disorder, and for this reason the strategy of repositioning already approved therapeutics could exhibit a successful outcome. Herein, we used CMAP and L1000CDS2 databases which include gene expression profiles datasets (genomic signatures) to identify potent compounds and classes of compounds which reverse disease's signature which could in turn reverse its phenotype. ALS signature was obtained by comparing gene expression of muscle biopsy specimens between diseased and healthy individuals. Statistical analysis was conducted to explore differentially transcripts in patients' samples. Then, the list of upregulated and downregulated genes was used to query both databases in order to determine molecules which downregulate the genes which are upregulated by ALS and vice versa. These compounds, based on their chemical structure along with known treatments, were clustered to reveal drugs with novel and potentially more effective mode of action with most of them predicted to affect pathways heavily involved in ALS. This evidence suggests that these compounds are strong candidates for moving to the next phase of the drug repurposing pipeline which is in vitro and in vivo experimental evaluation.

Addressing the clinical unmet needs in primary Sjögren's Syndrome through the sharing, harmonization and federated analysis of 21 European cohorts.

For many decades, the clinical unmet needs of primary Sjögren's Syndrome (pSS) have been left unresolved due to the rareness of the disease and the complexity of the underlying pathogenic mechanisms, including the pSS-associated lymphomagenesis process. Here, we present the HarmonicSS cloud-computing exemplar which offers beyond the state-of-the-art data analytics services to address the pSS clinical unmet needs, including the development of lymphoma classification models and the identification of biomarkers for lymphomagenesis. The users of the platform have been able to successfully interlink, curate, and harmonize 21 regional, national, and international European cohorts of 7,551 pSS patients with respect to the ethical and legal issues for data sharing. Federated AI algorithms were trained across the harmonized databases, with reduced execution time complexity, yielding robust lymphoma classification models with 85% accuracy, 81.25% sensitivity, 85.4% specificity along with 5 biomarkers for lymphoma development. To our knowledge, this is the first GDPR compliant platform that provides federated AI services to address the pSS clinical unmet needs.

A federated AI strategy for the classification of patients with Mucosa Associated Lymphoma Tissue (MALT) lymphoma across multiple harmonized cohorts.

Mucosa Associated Lymphoma Tissue (MALT) type is an extremely rare type of lymphoma which occurs in less than 3% of patients with primary Sjögren's Syndrome (pSS). No reported studies so far have been able to investigate risk factors for MALT development across multiple cohort databases with sufficient statistical power. Here, we present a generalized, federated AI (artificial intelligence) strategy which enables the training of AI algorithms across multiple harmonized databases. A case study is conducted towards the development of MALT classification models across 17 databases on pSS. Advanced AI algorithms were developed, including federated Multinomial Naïve Bayes (FMNB), federated gradient boosting trees (FGBT), FGBT with dropouts (FDART), and the federated Multilayer Perceptron (FMLP). The FDART with dropout rate 0.3 achieved the best performance with sensitivity 0.812, and specificity 0.829, yielding 8 biomarkers as prominent for MALT development.

Combined seronegativity in Sjögren's syndrome.

OBJECTIVES: To describe the clinical spectrum of Sjögren's syndrome (SS) patients with combined seronegativity. METHODS: From a multicentre study population of consecutive SS patients fulfilling the 2016 ACR-EULAR classification criteria, patients with triple seronegativity [anti-Ro/SSA(-), anti-La/SSB(-), RF(-) and ANA(+)] and quadruple seronegativity [anti-Ro/SSA(-), anti-La/SSB(-), RF(-) and ANA(-)] were identified retrospectively. Both groups were matched in an 1:1 ratio with 2 distinct control SS groups: i) classic anti-Ro/SSA seropositive patients [SS(+)] and ii) classic anti-Ro/SSA seropositive patients with negative rheumatoid factor [SS(+)/RF(-)] to explore their effect on disease expression. Clinical, laboratory and, histologic features were compared. A comparison between triple and quadruple seronegative SS patients was also performed. REESULTS: One hundred thirty-five SS patients (8.6%) were identified as triple seronegative patients and 72 (4.5%) as quadruple. Triple seronegative patients had lower frequency of peripheral nervous involvement (0% vs. 7.2% p=0.002) compared to SS(+) controls and lower frequency of interstitial renal disease and higher prevalence of dry mouth than SS(+)/RF(-) controls. Quadruple seronegative patients presented less frequently with persistent lymphadenopathy (1.5% vs. 16.9 p=0.004) and lymphoma (0% vs. 9.8% p=0.006) compared to SS(+) controls and with lower prevalence of persistent lymphadenopathy (1.5% vs. 15.3% p=0.008) and higher frequency of dry eyes (98.6% vs. 87.5% p=0.01) and autoimmune thyroiditis (44.1% vs. 17.1% p=0.02) compared to SS(+)/RF(-) SS controls. Study groups comparative analysis revealed that triple seronegative patients had higher frequency of persistent lymphadenopathy and lymphoma, higher focus score and later age of SS diagnosis compared to quadruple seronegative patients. CONCLUSIONS: Combined seronegativity accounts for almost 9% of total SS population and is associated with a milder clinical phenotype, partly attributed to the absence of rheumatoid factor.

Applied machine learning in cancer research: A systematic review for patient diagnosis, classification and prognosis.

Artificial Intelligence (AI) has recently altered the landscape of cancer research and medical oncology using traditional Machine Learning (ML) algorithms and cutting-edge Deep Learning (DL) architectures. In this review article we focus on the ML aspect of AI applications in cancer research and present the most indicative studies with respect to the ML algorithms and data used. The PubMed and dblp databases were considered to obtain the most relevant research works of the last five years. Based on a comparison of the proposed studies and their research clinical outcomes concerning the medical ML application in cancer research, three main clinical scenarios were identified. We give an overview of the well-known DL and Reinforcement Learning (RL) methodologies, as well as their application in clinical practice, and we briefly discuss Systems Biology in cancer research. We also provide a thorough examination of the clinical scenarios with respect to disease diagnosis, patient classification and cancer prognosis and survival. The most relevant studies identified in the preceding year are presented along with their primary findings. Furthermore, we examine the effective implementation and the main points that need to be addressed in the direction of robustness, explainability and transparency of predictive models. Finally, we summarize the most recent advances in the field of AI/ML applications in cancer research and medical oncology, as well as some of the challenges and open issues that need to be addressed before data-driven models can be implemented in healthcare systems to assist physicians in their daily practice.

Machine Learning Approaches on High Throughput NGS Data to Unveil Mechanisms of Function in Biology and Disease.

In this review, the fundamental basis of machine learning (ML) and data mining (DM) are summarized together with the techniques for distilling knowledge from state-of-the-art omics experiments. This includes an introduction to the basic mathematical principles of unsupervised/supervised learning methods, dimensionality reduction techniques, deep neural networks architectures and the applications of these in bioinformatics. Several case studies under evaluation mainly involve next generation sequencing (NGS) experiments, like deciphering gene expression from total and single cell (scRNA-seq) analysis; for the latter, a description of all recent artificial intelligence (AI) methods for the investigation of cell sub-types, biomarkers and imputation techniques are described. Other areas of interest where various ML schemes have been investigated are for providing information regarding transcription factors (TF) binding sites, chromatin organization patterns and RNA binding proteins (RBPs), while analyses on RNA sequence and structure as well as 3D dimensional protein structure predictions with the use of ML are described. Furthermore, we summarize the recent methods of using ML in clinical oncology, when taking into consideration the current omics data with pharmacogenomics to determine personalized treatments. With this review we wish to provide the scientific community with a thorough investigation of main novel ML applications which take into consideration the latest achievements in genomics, thus, unraveling the fundamental mechanisms of biology towards the understanding and cure of diseases.

A biomarker for lymphoma development in Sjogren's syndrome: Salivary gland focus score.

The aim of this study is to explore the role of labial minor salivary gland (LMSG) focus score (FS) in stratifying Sjögren's Syndrome (SS) patients, lymphoma development prediction and to facilitate early lymphoma diagnosis. Ιn an integrated cohort of 1997 patients, 618 patients with FS ≥ 1 and at least one-year elapsing time interval from SS diagnosis to lymphoma diagnosis or last follow up were identified. Clinical, laboratory and serological features were recorded. A data driven logistic regression model was applied to identify independent lymphoma associated risk factors. Furthermore, a FS threshold maximizing the difference of time interval from SS until lymphoma diagnosis between high and low FS lymphoma subgroups was investigated, to develop a follow up strategy for early lymphoma diagnosis. Of the 618 patients, 560 were non-lymphoma SS patients while the other 58 had SS and lymphoma. FS, cryoglobulinemia and salivary gland enlargement (SGE) were proven to be independent lymphoma associated risk factors. Lymphoma patients with FS ≥ 4 had a statistically significant shorter time interval from SS to lymphoma diagnosis, compared to those with FS < 4 (4 vs 9 years, respectively, p = 0,008). SS patients with FS ≥ 4 had more frequently B cell originated manifestations and lymphoma, while in patients with FS < 4, autoimmune thyroiditis was more prevalent. In the latter group SGE was the only lymphoma independent risk factor. A second LMSG biopsy is patients with a FS ≥ 4, 4 years after SS diagnosis and in those with FS < 4 and a history of SGE, at 9-years, may contribute to an early lymphoma diagnosis. Based on our results we conclude that LMSG FS, evaluated at the time of SS diagnosis, is an independent lymphoma associated risk factor and may serve as a predictive biomarker for the early diagnosis of SS-associated lymphomas.

A DSS for Predicting Lymphoma in Primary Sjogren's Syndrome Patients.

Primary Sjogren's syndrome (pSS) is a chronic autoimmune disease followed by exocrine gland dysfunction. In this work, a web application was developed as a screening test based on a machine learning model that was trained on clinical data and is used to predict lymphoma outcomes in pSS patient. The results of the final model reveal a sensitivity of 100%, accuracy of 82%, and area under the curve of 98% and confirms the importance of C4 value, lymphadenopathy, and rheumatoid factor as prominent lymphoma predictors.

Decision Support System for Breast Cancer Detection Using Biomarker Indicators.

Breast cancer is the second most common type of cancer among women in the USA, and it is very common to appear in its invasive form. Detecting its presence in the early stages can potentially aid in the mortality rate depletion since at that point large tumours are highly unlikely to have developed. Technological advances of the last decades have provided advanced tools that employ machine learning for early detection. Common techniques include tumour imaging using special equipment that in most cases is not widely accessible. In order to overcome this limitation, new techniques that employ blood-based biomarkers are being explored. In the current work machine learning algorithms are exploited for the development of a decision support system for breast cancer using easily obtainable user information, age, body mass index, glucose and resistin. The explored algorithms include Logistic Regression, Naive Bayes, Support Vector Machine and Gradient Boosting Classification, all of which are used for the classification of new patients based on a dataset that includes information from previous breast cancer incidents. The results depict that the optimal algorithm based on the current methodology and implementation is the Gradient Boosting Classification which exhibits the highest prediction scores. In order to ensure wide accessibility, a mobile application is developed. The user can easily provide the required information for the prediction to the application and obtain the results rapidly.

Hepatocellular Carcinoma Detection Using Machine Learning Techniques.

Hepatocellular carcinoma (HCC) is a form of primary cancer appearing in the liver. In this work used the hepatocellular carcinoma dataset from the UCI machine learning repository and tested different techniques for feature selection and classification. The following algorithms were used: decision trees, random forests, SVMs, k-NN classifiers, AdaBoost, and gradient boost. The best results were obtained using gradient boost with 84% accuracy and 93% precision. Finally, we deployed the model to a web application as a decision support system for clinicians.

Machine Learning for Autistic Spectrum Disorder Risk Screening.

In the modern world of rapidly increasing autistic spectrum disorder case rates, medical costs, societal impact, and long-waiting times from initial screening, there is a need for an easy, early screening of autistic spectrum disorder risk in children. In this paper, a mobile application was developed with these requirements, using machine learning algorithms achieving high performance compared to other applications that use simple rule approaches. A recently published autistic spectrum disorder dataset is used to train the model, containing hundreds of screening data from children in the ages 4 to 11.

Primary Sjögren's Syndrome of Early and Late Onset: Distinct Clinical Phenotypes and Lymphoma Development.

Objectives: To study the clinical, serological and histologic features of primary Sjögren's syndrome (pSS) patients with early (young ≤35 years) or late (old ≥65 years) onset and to explore the differential effect on lymphoma development. Methods: From a multicentre study population of 1997 consecutive pSS patients, those with early or late disease onset, were matched and compared with pSS control patients of middle age onset. Data driven analysis was applied to identify the independent variables associated with lymphoma in both age groups. Results: Young pSS patients (19%, n = 379) had higher frequency of salivary gland enlargement (SGE, lymphadenopathy, Raynaud's phenomenon, autoantibodies, C4 hypocomplementemia, hypergammaglobulinemia, leukopenia, and lymphoma (10.3% vs. 5.7%, p = 0.030, OR = 1.91, 95% CI: 1.11-3.27), while old pSS patients (15%, n = 293) had more frequently dry mouth, interstitial lung disease, and lymphoma (6.8% vs. 2.1%, p = 0.011, OR = 3.40, 95% CI: 1.34-8.17) compared to their middle-aged pSS controls, respectively. In young pSS patients, cryoglobulinemia, C4 hypocomplementemia, lymphadenopathy, and SGE were identified as independent lymphoma associated factors, as opposed to old pSS patients in whom SGE, C4 hypocomplementemia and male gender were the independent lymphoma associated factors. Early onset pSS patients displayed two incidence peaks of lymphoma within 3 years of onset and after 10 years, while in late onset pSS patients, lymphoma occurred within the first 6 years. Conclusion: Patients with early and late disease onset constitute a significant proportion of pSS population with distinct clinical phenotypes. They possess a higher prevalence of lymphoma, with different predisposing factors and lymphoma distribution across time.

Sjögren's Syndrome: The Clinical Spectrum of Male Patients.

BACKGROUND: To compare the clinical, serological and histologic features between male and female patients with Sjögren's syndrome (SS) and explore the potential effect of gender on lymphoma development. METHODS: From a multicenter population (Universities of Udine, Pisa and Athens, Harokopion and Ioannina (UPAHI)) consisting of consecutive SS patients fulfilling the 2016 ACR/EULAR criteria, male patients were identified, matched and compared with female controls. Data-driven multivariable logistic regression analysis was applied to identify independent lymphoma-associated factors. RESULTS: From 1987 consecutive SS patients, 96 males and 192 matched female controls were identified and compared. Males had a higher frequency of lymphoma compared to females (18% vs. 5.2%, OR = 3.89, 95% CI: 1.66 to 8.67; p = 0.0014) and an increased prevalence of serum anti-La/SSB antibodies (50% vs. 34%, OR = 1.953, 95% CI: 1.19 to 3.25; p = 0.0128). No differences were observed in the frequencies of lymphoma predictors between the two genders. Data-driven multivariable logistic regression analysis revealed negative association of the female gender with lymphoma and positive association with lymphadenopathy. CONCLUSION: Male SS patients carry an increased risk of lymphoma development. Although statistics showed no difference in classical lymphoma predictors compared to females, data-driven analysis revealed gender and lymphadenopathy as independent lymphoma-associated features.

Overcoming the Barriers That Obscure the Interlinking and Analysis of Clinical Data Through Harmonization and Incremental Learning.

Goal: To present a framework for data sharing, curation, harmonization and federated data analytics to solve open issues in healthcare, such as, the development of robust disease prediction models. Methods: Data curation is applied to remove data inconsistencies. Lexical and semantic matching methods are used to align the structure of the heterogeneous, curated cohort data along with incremental learning algorithms including class imbalance handling and hyperparameter optimization to enable the development of disease prediction models. Results: The applicability of the framework is demonstrated in a case study of primary Sjögren's Syndrome, yielding harmonized data with increased quality and more than 85% agreement, along with lymphoma prediction models with more than 80% sensitivity and specificity. Conclusions: The framework provides data quality, harmonization and analytics workflows that can enhance the statistical power of heterogeneous clinical data and enables the development of robust models for disease prediction.

Predicting Lymphoma Development by Exploiting Genetic Variants and Clinical Findings in a Machine Learning-Based Methodology With Ensemble Classifiers in a Cohort of Sjögren's Syndrome Patients.

Lymphoma development constitutes one of the most serious clinico-pathological manifestations of patients with Sjögren's Syndrome (SS). Over the last decades the risk for lymphomagenesis in SS patients has been studied aiming to identify novel biomarkers and risk factors predicting lymphoma development in this patient population. Objective: The current study aims to explore whether genetic susceptibility profiles of SS patients along with known clinical, serological and histological risk factors enhance the accuracy of predicting lymphoma development in this patient population. Methods: The potential predicting role of both genetic variants, clinical and laboratory risk factors were investigated through a Machine Learning-based (ML) framework which encapsulates ensemble classifiers. Results: Ensemble methods empower the classification accuracy with approaches which are sensitive to minor perturbations in the training phase. The evaluation of the proposed methodology based on a 10-fold stratified cross validation procedure yielded considerable results in terms of balanced accuracy (GB: 0.7780 ± 0.1514, RF Gini: 0.7626 ± 0.1787, RF Entropy: 0.7590 ± 0.1837). Conclusions: The initial clinical, serological, histological and genetic findings at an early diagnosis have been exploited in an attempt to establish predictive tools in clinical practice and further enhance our understanding towards lymphoma development in SS.

Sjögren's syndrome towards precision medicine: the challenge of harmonisation and integration of cohorts.

Primary Sjögren's syndrome (pSS) is a chronic, systemic autoimmune disease with diverse clinical picture and outcome. The disease affects primarily middle-aged females and involves the exocrine glands leading to dry mouth and eyes. When the disease extends beyond the exocrine glands (systemic form), certain extraglandular manifestations involving liver, kidney, lungs, peripheral nervous system and the skin may occur. Primary SS is considered the crossroad between autoimmunity and lymphoproliferation, since approximately 5% of patients develop NHL associated lymphomas. As with every chronic disease with complex aetiopathogenesis and clinical heterogeneity, pSS has certain unmet needs that have to be addressed: a) classification and stratification of patients; b) understanding the distinct pathogenetic mechanisms and clinical phenotypes; c) defining and interpreting the real needs of patients regarding the contemporary diagnostic and therapeutic approaches; d) physician and patients' training regarding the wide spectrum of the disease; e) creating common policies across European countries to evaluate and manage SS patients. To achieve these goals, an intense effort is being currently undertaken by the HarmonicSS consortium in order to harmonise and integrate the largest European cohorts of pSS patients. In this review, we present an overview of our perception and vision, as well as new issues arising from this project such as harmonisation protocols and procedures, data sharing principles and various ethical and legal issues originating from these approaches.