A Bayesian nonparametric approach to correct for underreporting in count data.

We propose a nonparametric compound Poisson model for underreported count data that introduces a latent clustering structure for the reporting probabilities. The latter are estimated with the model's parameters based on experts' opinion and exploiting a proxy for the reporting process. The proposed model is used to estimate the prevalence of chronic kidney disease in Apulia, Italy, based on a unique statistical database covering information on m = 258 municipalities obtained by integrating multisource register information. Accurate prevalence estimates are needed for monitoring, surveillance, and management purposes; yet, counts are deemed to be considerably underreported, especially in some areas of Apulia, one of the most deprived and heterogeneous regions in Italy. Our results agree with previous findings and highlight interesting geographical patterns of the disease. We compare our model to existing approaches in the literature using simulated as well as real data on early neonatal mortality risk in Brazil, described in previous research: the proposed approach proves to be accurate and particularly suitable when partial information about data quality is available.

Updating the Clinical Application of Blood Biomarkers and Their Algorithms in the Diagnosis and Surveillance of Hepatocellular Carcinoma: A Critical Review.

The most common primary liver cancer is hepatocellular carcinoma (HCC), and its mortality rate is increasing globally. The overall 5-year survival of patients with liver cancer is currently 10-20%. Moreover, because early diagnosis can significantly improve prognosis, which is highly correlated with tumor stage, early detection of HCC is critical. International guidelines advise using α-FP biomarker with/without ultrasonography for HCC surveillance in patients with advanced liver disease. However, traditional biomarkers are sub-optimal for risk stratification of HCC development in high-risk populations, early diagnosis, prognostication, and treatment response prediction. Since about 20% of HCCs do not produce α-FP due to its biological diversity, combining α-FP with novel biomarkers can enhance HCC detection sensitivity. There is a chance to offer promising cancer management methods in high-risk populations by utilizing HCC screening strategies derived from new tumor biomarkers and prognostic scores created by combining biomarkers with distinct clinical parameters. Despite numerous efforts to identify molecules as potential biomarkers, there is no single ideal marker in HCC. When combined with other clinical parameters, the detection of some biomarkers has higher sensitivity and specificity in comparison with a single biomarker. Therefore, newer biomarkers and models, such as the Lens culinaris agglutinin-reactive fraction of Alpha-fetoprotein (α-FP), α-FP-L3, Des-γ-carboxy-prothrombin (DCP or PIVKA-II), and the GALAD score, are being used more frequently in the diagnosis and prognosis of HCC. Notably, the GALAD algorithm was effective in HCC prevention, particularly for cirrhotic patients, regardless of the cause of their liver disease. Although the role of these biomarkers in surveillance is still being researched, they may provide a more practical alternative to traditional imaging-based surveillance. Finally, looking for new diagnostic/surveillance tools may help improve patients' survival. This review discusses the current roles of the most used biomarkers and prognostic scores that may aid in the clinical management of HCC patients.

NIAPU: network-informed adaptive positive-unlabeled learning for disease gene identification.

MOTIVATION: Gene-disease associations are fundamental for understanding disease etiology and developing effective interventions and treatments. Identifying genes not yet associated with a disease due to a lack of studies is a challenging task in which prioritization based on prior knowledge is an important element. The computational search for new candidate disease genes may be eased by positive-unlabeled learning, the machine learning (ML) setting in which only a subset of instances are labeled as positive while the rest of the dataset is unlabeled. In this work, we propose a set of effective network-based features to be used in a novel Markov diffusion-based multi-class labeling strategy for putative disease gene discovery. RESULTS: The performances of the new labeling algorithm and the effectiveness of the proposed features have been tested on 10 different disease datasets using three ML algorithms. The new features have been compared against classical topological and functional/ontological features and a set of network- and biological-derived features already used in gene discovery tasks. The predictive power of the integrated methodology in searching for new disease genes has been found to be competitive against state-of-the-art algorithms. AVAILABILITY AND IMPLEMENTATION: The source code of NIAPU can be accessed at https://github.com/AndMastro/NIAPU. The source data used in this study are available online on the respective websites. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Economic impact of industry-sponsored clinical trials in inflammatory bowel diseases: Results from the national institute of gastroenterology "Saverio de Bellis".

Introduction: The majority of the money spent on possible new medications' clinical trials is accounted for by the innovative pharmaceutical sector, which also stimulates the economy of a nation. The objective of this study was to evaluate the impact of pharmaceutical industry-sponsored clinical trials (ISCTs) in inflammatory bowel diseases (IBDs) towards the national health service (NHS) in terms of avoided costs and leverage effect. Methodology: The research was conducted at National Institute of Gastroenterology, "Saverio De Bellis", Castellana Grotte (Apulia, Italy) collecting data from profit ISCTs of pharmaceutical products conducted over the time period 2018-2020 with focus on inflammatory bowel diseases. After the quantification of health services and drug costs from the latter studies, avoided costs and leverage effects were then estimated. Results: The results on the avoided costs for healthcare facilities deriving from the conduct of clinical studies show that, in relation to the sample of five drug companies participating in our 2018-2020 analysis, out of a total of 235,102.46 €, identified as direct investment, 628,158.21 € of avoided costs for the NHS were measured, with an additional saving (leverage effect) for the NHS of 3.67 € for each € invested by the companies promoting clinical trials. Conclusion: Conducting profit clinical trials has practical benefits and a favourable macroeconomic impact that, by completing its limited resources, helps to sustain one country NHS thanks to the avoided costs while also contributing to locational and industrial policy while guaranteeing novel therapeutics and health services for the patients enrolled.

EdgeSHAPer: Bond-centric Shapley value-based explanation method for graph neural networks.

Graph neural networks (GNNs) recursively propagate signals along the edges of an input graph, integrate node feature information with graph structure, and learn object representations. Like other deep neural network models, GNNs have notorious black box character. For GNNs, only few approaches are available to rationalize model decisions. We introduce EdgeSHAPer, a generally applicable method for explaining GNN-based models. The approach is devised to assess edge importance for predictions. Therefore, EdgeSHAPer makes use of the Shapley value concept from game theory. For proof-of-concept, EdgeSHAPer is applied to compound activity prediction, a central task in drug discovery. EdgeSHAPer's edge centricity is relevant for molecular graphs where edges represent chemical bonds. Combined with feature mapping, EdgeSHAPer produces intuitive explanations for compound activity predictions. Compared to a popular node-centric and another edge-centric GNN explanation method, EdgeSHAPer reveals higher resolution in differentiating features determining predictions and identifies minimal pertinent positive feature sets.

Network Proximity-Based Drug Repurposing Strategy for Early and Late Stages of Primary Biliary Cholangitis.

Primary biliary cholangitis (PBC) is a chronic, cholestatic, immune-mediated, and progressive liver disorder. Treatment to preventing the disease from advancing into later and irreversible stages is still an unmet clinical need. Accordingly, we set up a drug repurposing framework to find potential therapeutic agents targeting relevant pathways derived from an expanded pool of genes involved in different stages of PBC. Starting with updated human protein-protein interaction data and genes specifically involved in the early and late stages of PBC, a network medicine approach was used to provide a PBC "proximity" or "involvement" gene ranking using network diffusion algorithms and machine learning models. The top genes in the proximity ranking, when combined with the original PBC-related genes, resulted in a final dataset of the genes most involved in PBC disease. Finally, a drug repurposing strategy was implemented by mining and utilizing dedicated drug-gene interaction and druggable genome information knowledge bases (e.g., the DrugBank repository). We identified several potential drug candidates interacting with PBC pathways after performing an over-representation analysis on our initial 1121-seed gene list and the resulting disease-associated (algorithm-obtained) genes. The mechanism and potential therapeutic applications of such drugs were then thoroughly discussed, with a particular emphasis on different stages of PBC disease. We found that interleukin/EGFR/TNF-alpha inhibitors, branched-chain amino acids, geldanamycin, tauroursodeoxycholic acid, genistein, antioestrogens, curcumin, antineovascularisation agents, enzyme/protease inhibitors, and antirheumatic agents are promising drugs targeting distinct stages of PBC. We developed robust and transparent selection mechanisms for prioritizing already approved medicinal products or investigational products for repurposing based on recognized unmet medical needs in PBC, as well as solid preliminary data to achieve this goal.

"The Disease Awareness Innovation Network" for chronic kidney disease identification in general practice.

BACKGROUND: The "awareness gap" and the under-recognition of chronic kidney disease (CKD) by general practitioners (GPs) is well documented. We set a framework to evaluate the impact in primary care of targeted training and networking with nephrologists with regard to CKD awareness in terms of potential increase of the proportion of patients classified according to KDIGO in the general population and in patients with diabetes, hypertension and heart failure. METHODS: Data were extracted from the Millewin Digital Platform in use by the GPs (N = 17) at baseline (T0, N = 17,854) and after 6 months (T6, N = 18,662) of networking (education, instant messaging and selected joint visits) with nephrologists (N = 2). The following variables were extracted: age, sex, eGFR (estimated glomerular filtration rate), ACR (urinary albumin-to-creatinine ratio), presence of type 2 diabetes, hypertension and heart failure. The proportion of patients detected having an eGFR below 60 mL/min/1.73m2 was also reported as deemed clinically relevant. RESULTS: We observed an increase in the use of ACR and eGFR tests in the entire cohort (+ 121% and + 73%, respectively) and in patients with comorbidities. The proportion of patients with eGFR < 60 mL/min/1.73m2 significantly increased from 2.2% to 3.8% in the entire cohort,  from 6.3% to 12.7% in patients with diabetes, and from 5.6% to 9.9% in those with hypertension and finally from 10.8% to 23.7% in patients with heart failure. CONCLUSIONS: Training and network support to GPs by nephrologists can improve CKD awareness and increase its identification in the general population and, even more, in categories at risk.

Role of plasma phosphorylated neurofilament heavy chain (pNfH) in amyotrophic lateral sclerosis.

The phosphorylated neurofilament heavy chain (pNfH) is a promising biomarker in amyotrophic lateral sclerosis (ALS). We examined plasma pNfH concentrations in order to corroborate its role as a diagnostic and prognostic biomarker in ALS. Incident ALS cases enrolled in a population-based registry were retrospectively selected and matched by sex and age with a cohort of healthy volunteers. Plasma pNfH levels were measured by an ELISA kit and correlated with clinical parameters. Discrimination ability of pNfH was tested using receiving operating characteristic (ROC) curves. Kaplan-Meier (KM) analysis and Cox proportional hazard models were used for survival analysis. Plasma pNfH was significantly higher in patients compared to controls. An optimal cut-off of 39.74 pg/ml discriminated cases from controls with an elevated sensitivity and specificity. Bulbar-onset cases had higher plasma pNfH compared to spinal onset (p = 0.0033). Furthermore, plasma pNfH positively correlated with disease progression rate (r = 0.19, p = 0.031). Baseline plasma pNfH did not influence survival in our cohort. Our findings confirmed the potential utility of plasma pNfH as a diagnostic biomarker in ALS. However, further studies with longitudinal data are needed to corroborate its prognostic value.

Early versus delayed antiretroviral therapy based on genotypic resistance test: Results from a large retrospective cohort study.

Rapid start of antiretroviral therapy (ART) pending genotypic resistance test (GRT) has been recently proposed, but the effectiveness of this strategy is still debated. The rate of virological success (VS), defined as HIV-RNA < 50 copies/ml, with and without GRT was compared in drug-naïve individuals enrolled in the Italian ARCA cohort who started ART between 2015 and 2018. 521 individuals started ART: 397 without GRT (pre-GRT group) and 124 following GRT (post-GRT group). Overall, 398 (76%) were males and 30 (6%) were diagnosed with AIDS. In the pre-GRT group, baseline CD4+ cell counts were lower (p < 0.001), and viral load was higher (p < 0.001) than in the post-GRT group. The estimated probability of VS in pre-GRT versus post-GRT group was 72.54% (CI95 : 67.78-76.60) versus 66.94% (CI95 : 57.53-74.26) at Week 24 and 92.40% (CI95 : 89.26-94.62) versus 92.92% (CI95 : 86.35-96.33) at Week 48, respectively (p = 0.434). At Week 48, VS was less frequent among individuals with baseline CD4+ cell counts <200 versus >500 (90.33% vs. 97.33%), log viral load <5.00 versus >5.70 log10 cps/ml (97.17% vs 78.16%; p < 0.001), and those treated with protease inhibitors or non-nucleoside reverse transcriptase inhibitors versus those treated with integrase strand transfer inhibitors (p < 0.001). The rate of VS does not seem to be affected by an early ART initiation pending GRT results, but it could be influenced by the composition of the ART regimen, as well as immuno-virological parameters.

Identification of glomerulosclerosis using IBM Watson and shallow neural networks.

BACKGROUND: Advanced stages of different renal diseases feature glomerular sclerosis at a histological level which is observed by light microscopy on tissue samples obtained by performing a kidney biopsy. Computer-aided diagnosis (CAD) systems leverage the potential of artificial intelligence (AI) in healthcare to support physicians in the diagnostic process. METHODS: We propose a novel CAD system that processes histological images and discriminates between sclerotic and non-sclerotic glomeruli. To this goal, we designed, tested, and compared two artificial neural network (ANN) classifiers. The former implements a shallow ANN classifying hand-crafted features extracted from Regions of Interest (ROIs) by means of image-processing procedures. The latter, instead, employs the IBM Watson Visual Recognition System, which uses a deep artificial neural network making decisions taking the images as input, without the need to design any procedure for describing images with features. The input dataset consisted of 428 sclerotic glomeruli and 2344 non-sclerotic glomeruli derived from images of kidney biopsies scanned by the Aperio ScanScope System. RESULTS: Both AI approaches allowed to very accurately distinguish (mean MCC 0.95 and mean Accuracy 0.99) between sclerotic and non-sclerotic glomeruli. Although the systems may seem interchangeable, the approach based on feature extraction and classification would allow clinicians to gain information on the most discriminating features. In fact, further procedures could explain the classifier's decision by analysing which subset of features impacted the most on the final decision. CONCLUSIONS: We developed a customizable support system that can facilitate the work of renal pathologists both in clinical and research settings.

Protocol to explain graph neural network predictions using an edge-centric Shapley value-based approach.

Here we present EdgeSHAPer, a workflow for explaining graph neural networks by approximating Shapley values using Monte Carlo sampling. In this protocol, we describe steps to execute Python scripts for a chemical dataset from the original publication; however, this approach is also applicable to any user-provided dataset. We also detail steps encompassing neural network training, an explanation phase, and analysis via feature mapping. For complete details on the use and execution of this protocol, please refer to Mastropietro et al. (2022).1.

Predictors of Virological Failure Among People Living with HIV Switching from an Effective First-Line Antiretroviral Regimen.

Aim of this study was to assess the predictors of virological failure (VF) among patients living with HIV (PLWHIV) switching from an effective first-line antiretroviral therapy (ART) regimen, and to evaluate the emergence of resistance-associated mutations. All adult patients enrolled in the Antiviral Response Cohort Analysis cohort who started ART after 2010, with at least 6 months of virological suppression (VS) before ART switch and with an available genotypic resistance test (GRT) at baseline were included. Thirty-two patients out of the 607 PLWHIV included (5.3%) experienced VF after a median of 11 months from ART switch. Younger age (adjusted Hazard Ratio [aHR] 0.96, 95% confidence interval [CI] 0.92-0.99, p = .023), being male who have sex with male (aHR 0.15, 95% CI 0.03-0.69, p = .014), and longer time from VS to ART switch (aHR 0.97, 95% CI 0.95-1.00, p = .021) resulted protective toward VF, while receiving a first-line regimen containing a backbone other than ABC/3TC or TXF/FTC (aHR 3.61, 95% CI 1.00-13.1, p = .050) and a boosted protease inhibitor as anchor drug (aHR 3.34, 95% CI 1.20-9.28, p = .021) were associated with higher risk of VF. GRT at the moment of VF was available only for 13 patients (40.6%). ART switch in patients with stable control of HIV infection is a safe practice, even if particular attention should be paid in certain cases of patients switching from regimens containing low-performance backbones or protease inhibitors.

The Role of Age on Beta-Amyloid1-42 Plasma Levels in Healthy Subjects.

Beta-amyloid (Aß) plaques have been observed in the brain of healthy elderlies with frequencies strongly influenced by age. The aim of the study is to evaluate the role of age and other biochemical and hematological parameters on Aß1-42 plasma levels in cognitively and neurologically normal individuals. Two-hundred and seventy-five normal subjects stratified by age groups (<35 years, 35-65 years, and >65 years) were included in the study. Aß1-42 plasma levels significantly correlated with age (rs = 0.27; p < 0.0001) in the whole sample, inversely correlated with age in the first age group (rs = -0.25, p = 0.01), positively correlated in the second group (rs = 0.22, p = 0.03), while there was no significant correlation in the older group (rs = 0.02, p = 0.86). Both age (ß-estimate = 0.08; p < 0.001) and cholesterol (ß-estimate = 0.03; p = 0.009) were significantly associated with Aß1-42 plasma level in multivariable analysis. However, only the association with age survived post hoc adjustment for multiple comparisons. The different effects of age on the Aß level across age groups should be explored in further studies to better understand the age-dependent variability. This could better define the value of plasma Aß as a biomarker of the Alzheimer neuropathology.

A non-stationary Markov model for economic evaluation of grass pollen allergoid immunotherapy.

INTRODUCTION: Allergic rhino-conjunctivitis (ARC) is an IgE-mediated disease that occurs after exposure to indoor or outdoor allergens, or to non-specific triggers. Effective treatment options for seasonal ARC are available, but the economic aspects and burden of these therapies are not of secondary importance, also considered that the prevalence of ARC has been estimated at 23% in Europe. For these reasons, we propose a novel flexible cost-effectiveness analysis (CEA) model, intended to provide healthcare professionals and policymakers with useful information aimed at cost-effective interventions for grass-pollen induced allergic rhino-conjunctivitis (ARC). METHODS: Treatments compared are: 1. no AIT, first-line symptomatic drug-therapy with no allergoid immunotherapy (AIT). 2. SCIT, subcutaneous immunotherapy. 3. SLIT, sublingual immunotherapy. The proposed model is a non-stationary Markovian model, that is flexible enough to reflect those treatment-related problems often encountered in real-life and clinical practice, but that cannot be adequately represented in randomized clinical trials (RCTs). At the same time, we described in detail all the structural elements of the model as well as its input parameters, in order to minimize any issue of transparency and facilitate the reproducibility and circulation of the results among researchers. RESULTS: Using the no AIT strategy as a comparator, and the Incremental Cost Effectiveness Ratio (ICER) as a statistic to summarize the cost-effectiveness of a health care intervention, we could conclude that: SCIT systematically outperforms SLIT, except when a full societal perspective is considered. For example, for T = 9 and a pollen season of 60 days, we have ICER = €16,729 for SCIT vs. ICER = €15,116 for SLIT (in the full societal perspective).For longer pollen seasons or longer follow-up duration the ICER decreases, because each patient experiences a greater clinical benefit over a larger time span, and Quality-adjusted Life Year (QALYs) gained per cycle increase accordingly.Assuming that no clinical benefit is achieved after premature discontinuation, and that at least three years of immunotherapy are required to improve clinical manifestations and perceiving a better quality of life, ICERs become far greater than €30,000.If the immunotherapy is effective only at the peak of the pollen season, the relative ICERs rise sharply. For example, in the scenario where no clinical benefit is present after premature discontinuation of immunotherapy, we have ICER = €74,770 for SCIT vs. ICER = €152,110 for SLIT.The distance between SCIT and SLIT strongly depends on under which model the interventions are meta-analyzed. CONCLUSIONS: Even though there is a considerable evidence that SCIT outperforms SLIT, we could not state that both SCIT and SLIT (or only one of these two) can be considered cost-effective for ARC, as a reliable threshold value for cost-effectiveness set by national regulatory agencies for pharmaceutical products is missing. Moreover, the impact of model input parameters uncertainty on the reliability of our conclusions needs to be investigated further.