SEARCH · PubMed Health

Results for “artificial intelligence (AI)”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 163 records · Page 9Linked to original sources

PubMed2005

Improving end of life care: an information systems approach to reducing medical errors.

Chronic and terminally ill patients are disproportionately affected by medical errors. In addition, the elderly suffer more preventable adverse events than younger patients. Targeting system wide "error-reducing" reforms to vulnerable populations can significantly reduce the incidence and prevalence of human error in medical practice. Recent developments in health informatics, particularly the application of artificial intelligence (AI) techniques such as data mining, neural networks, and case-based reasoning (CBR), presents tremendous opportunities for mitigating error in disease diagnosis and patient management. Additionally, the ubiquity of the Internet creates the possibility of an almost ideal network for the dissemination of medical information. We explore the capacity and limitations of web-based palliative information systems (IS) to transform the delivery of care, streamline processes and improve the efficiency and appropriateness of medical treatment. As a result, medical error(s) that occur with patients dealing with severe, chronic illness and the frail elderly can be reduced.The palliative model grew out of the need for pain relief and comfort measures for patients diagnosed with cancer. Applied definitions of palliative care extend this convention, but there is no widely accepted definition. This research will discuss the development life cycle of two palliative information systems: the CONFER QOLP management information system (MIS), currently used by a community-based palliative care program in Brooklyn, New York, and the CAREN case-based reasoning prototype. CONFER is a web platform based on the idea of "eCare". CONFER uses XML (extensible mark-up language), a W3C-endorced standard mark up to define systems data. The second system, CAREN, is a CBR prototype designed for palliative care patients in the cancer trajectory. CBR is a technique, which tries to exploit the similarities of two situations and match decision-making to the best-known precedent cases. The prototype uses the opensource CASPIAN shell developed by the University of Aberystwyth, Wales and is available by anonymous FTP. We will discuss and analyze the preliminary results we have obtained using this CBR tool. Our research suggests that automated information systems can be used to improve the quality of care at the end of life and disseminate expert level 'know how' to palliative care clinicians. We will present how our CBR prototype can be successfully deployed, capable of securely transferring information using a Secure File Transfer Protocol (SFTP) and using a JAVA CBR engine.

Artificial Intelligence↗

PubMed2026

Artificial intelligence-based tumour infiltrating lymphocyte quantification in patients with triple-negative breast cancer: an independent validation study.

BACKGROUND: Tumour-infiltrating lymphocytes (TILs) are a robust prognostic marker in patients with triple-negative breast cancer. Artificial intelligence (AI)-derived computational tools assessing TILs could improve efficiency, but require independent validation against clinical outcomes. We aimed to compare the prognostic performance of AI-derived TIL scores with pathologist-scored TILs in a large, prospectively collected dataset pooled from randomised controlled trials. METHODS: CATALINA was an independent, external validation study using prospectively collected long-term clinical outcome data pooled from seven randomised clinical trials conducted at multiple sites. We independently evaluated two previously validated AI pipelines that generate five computationally assessed tumour-infiltrating lymphocyte (cTIL) scores by masked, independent deployment of locked models. cTIL scores were correlated with the mean of the pathologist-scored stromal TILs (sTILs) in 220 digitised haematoxylin and eosin whole slide images in a cohort of patients with early-stage triple-negative or HER-2 positive breast cancer, previously scored by trained pathologists in a TIL-reproducibility study. Prognostic performance was assessed in a separate cohort of patients with early triple-negative breast cancer pooled from seven prospective, randomised adjuvant trials. Multivariable Cox regression models adjusted for clinicopathological factors and study heterogeneity assessed associations of cTIL score and sTIL score with invasive disease-free survival, distant disease-free survival, and overall survival. 5-year discrimination was estimated using time-dependent area under the receiver operating characteristic curve (AUC). FINDINGS: Individual data were collated from 1759 patients, of whom 1356 had complete clinicopathological data, pathologist sTIL scores, and cTIL scores available. Modest correlation (r 0·375-0·473) was observed between cTIL scores and the mean pathologist sTIL score. Both sTIL and cTIL were independently associated with 5-year invasive disease-free survival, distant disease-free survival, and overall survival after adjustment for clinicopathological factors (hazard ratio for invasive disease-free survival was 0·73 [95% CI 0·66-0·82]; q<0·0001, distant disease-free survival was 0·70 [0·61-0·79]; q<0·0001, and overall survival was 0·72 [0·63-0·82]; q<0·0001 for sTIL scores and 0·80 [0·73-0·89]; q<0·0001, 0·77 [0·69-0·86]; q<0·0001, and 0·79 [0·70-0·88]; q=0·0002, respectively, for percentage_lymphocyte scores). In models adjusted for clinicopathological variables and sTIL score, cTIL score did not maintain a statistically significant prognostic association. Both sTIL and cTIL scores improved the 5-year AUC over clinicopathological variables alone, while cTIL score did not significantly further improve AUC when combined with clinicopathological variables and sTIL score. INTERPRETATION: Two cTIL models deployed entirely without retraining or modification provided statistically significant prognostic information and improved risk discrimination compared with clinicopathological variables alone in this large, platform-based, independent validation study. Although cTIL score did not incrementally improve prognostication compared with models combining clinicopathological variables with sTIL score, these findings support the application of cTILs as a reproducible prognostic biomarker, particularly in settings where routine or widespread pathologist assessment is unavailable. FUNDING: Breast Cancer Research Foundation (USA).

Humans↗

PubMed2006

Interpretation of surface EMGs in children with cerebral palsy: An initial study using a fuzzy expert system.

Surface EMG detected simultaneously at different muscles has become an important tool for analysing the gait of children with cerebral palsy (CP), as it offers essential information about muscular coordination. However, the interpretation of surface EMG is a difficult task that assumes extensive knowledge and experience. As such, this noninvasive procedure is not frequently used in the general clinical routine. An Artificial Intelligence (AI) system for interpreting surface EMG signals and the resulting muscular coordination patterns could overcome these limitations. To support such interpretation, an expert system based on fuzzy inference methodology was developed. The knowledge-base of the system implemented 15 rules, from which the fuzzy inference methodology performs a prediction of the effectiveness of the muscular coordination during gait. Our aim was to assess the feasibility and value of such an expert system in clinical applications. Surface EMG signals were recorded from the tibialis anterior, soleus muscle, and gastrocnemius muscles of children with CP to assess muscular coordination patterns of ankle movement during gait. Nineteen children underwent 114 surface EMG measurements. Simultaneously, the gait cycles of each patient were determined using foot switches and videotapes. From the EMG signals, the effectiveness of the ankle movement was predicted by the expert system, and predictions were classified using a three-point ordinal scale. In 91 cases (80%), the clinical findings matched the predictions of the expert system. In 23 cases (20%) the predictions of the expert system differed from the clinical findings with 12 cases revealing worse and 11 cases revealing better results in comparison to the clinical findings. As this study is a first attempt to verify the feasibility and correctness of this expert system, the results are promising. Further study is required to assess the correlation with the kinematic data and to include the whole leg.

Cerebral Palsy↗

PubMed2003

Approaches towards the automated interpretation and prediction of electrospray tandem mass spectra of non-peptidic combinatorial compounds.

Combinatorial chemistry is widely used within the pharmaceutical industry as a means of rapid identification of potential drugs. With the growth of combinatorial libraries, mass spectrometry (MS) became the key analytical technique because of its speed of analysis, sensitivity, accuracy and ability to be coupled with other analytical techniques. In the majority of cases, electrospray mass spectrometry (ES-MS) has become the default ionisation technique. However, due to the absence of fragment ions in the resulting spectra, tandem mass spectrometry (MS/MS) is required to provide structural information for the identification of an unknown analyte. This work discusses the first steps of an investigation into the fragmentation pathways taking place in electrospray tandem mass spectrometry. The ultimate goal for this project is to set general fragmentation rules for non-peptidic, pharmaceutical, combinatorial compounds. As an aid, an artificial intelligence (AI) software package is used to facilitate interpretation of the spectra. This initial study has focused on determining the fragmentation rules for some classes of compound types that fit the remit as outlined above. Based on studies carried out on several combinatorial libraries of these compounds, it was established that different classes of drug molecules follow unique fragmentation pathways. In addition to these general observations, the specific ionisation processes and the fragmentation pathways involved in the electrospray mass spectra of these systems were explored. The ultimate goal will be to incorporate our findings into the computer program and allow identification of an unknown, non-peptidic compound following insertion of its ES-MS/MS spectrum into the AI package. The work herein demonstrates the potential benefit of such an approach in addressing the issue of high-throughput, automated MS/MS data interpretation.

Automation↗

PubMed2026

Reassessing Semen Analysis: Clinical Insights Beyond Sperm Count and Motility.

BACKGROUND: Semen analysis (SA), recognized by the World Health Organization (WHO) as the cornerstone of male infertility evaluation, remains indispensable in reproductive medicine. However, advances in assisted reproductive technology (ART) and artificial intelligence (AI) have highlighted the limitations of relying solely on conventional semen parameters. OBJECTIVE: To critically review the evolving clinical role of SA by integrating conventional assessment with emerging functional, molecular, and computational approaches that improve diagnostic accuracy and individualized patient care. METHODS: A narrative review of contemporary evidence was conducted, focusing on conventional semen parameters, biofunctional sperm testing, omics technologies, AI-assisted analysis, and broader clinical applications of SA. RESULTS: Conventional parameters, including sperm concentration, motility, and morphology, remain essential but inadequately reflect fertilizing capacity. Adjunctive assessments, including oxidative stress biomarkers and sperm DNA fragmentation, provide valuable insights into sperm function and reproductive potential. Omics technologies, including genomics, transcriptomics, proteomics, and metabolomics, deepen mechanistic understanding, while AI enhances diagnostic precision, reproducibility, and standardization. Beyond infertility evaluation, SA also supports male contraceptive assessment, natural conception, ART, and patient counseling. CONCLUSIONS: Integrating conventional SA with functional, molecular, and AI-driven diagnostics provides a comprehensive framework for evaluating male fertility, advancing precision reproductive medicine and personalized clinical management.

cryopreservation↗

PubMed1999

Reliability of neural-network functional electrical stimulation gait-control system.

Functional electrical stimulation (FES) has been used for restoring walking in spinal-cord injured (SCI) persons. Using artificial intelligence (AI), FES controllers have been developed that allow the automatic phasing of stimulation, to replace the function of hand or heel switches. However, there has been no study to evaluate the reliability of these AI systems. Neural networks were used to construct FES controllers to control the timing of stimulation. Different numbers of sensors in the sensor set and different numbers of data points from each sensor were used. Two incomplete-SCI subjects were recruited, and each was tested on three separate occasions. The results show the neural-network controllers can maintain a high accuracy (around 90% for the two- and three-sensor groups and 80% for the one-sensor group) over a period of six months. Two or three sensors were sufficient to provide enough information to construct a reliable FES control system, and the number of data points did not have any effect on the reliability of the system.

Adult↗

PubMed2025

Foliar disease resistance phenomics of fungal pathogens: image-based approaches for mapping quantitative resistance in cereal germplasm.

Host plant resistance is the most effective and environmentally sustainable means of reducing yield losses caused by fungal foliar pathogens of cereal species. Cereal genebank collections hold diverse pools of potentially underutilized disease resistance alleles, and cereal genomic resources are well advanced due to large-scale sequencing and genotyping efforts. Genome-Wide Association Studies (GWAS) have emerged as the predominant association genetics technique to initially discover novel disease resistance loci or alleles in these diverse collections. Traditional disease resistance phenotyping methods are reliant on visual estimation of disease symptom severity and have successfully supported genetic mapping studies either via GWAS or QTL mapping in biparental populations facilitating both marker development and gene cloning efforts. Due to foliar pathogens having a high capacity to evolve, there is a need to pyramid disease resistance genes with diverse mechanisms for durable control. Resistance expressed as a quantitative trait, known as quantitative resistance (QR), is hypothesized to be more durable, unlike major R-gene resistance that is race-specific and can be vulnerable to breaking down without gene stewardship. However, assessing QR visually is challenging, particularly when complicated by complex genotype × environment (G × E) effects in the field. High-throughput image-based phenotyping provides accurate and unbiased data that can support foliar disease resistance screening efforts of genebank collections using GWAS. In this review, we discuss image-based disease phenotyping based on macroscopic (visible symptoms) and microscopic features during the host-pathogen interaction. Quantitative image analysis approaches using conventional and artificial intelligence (AI) algorithms are also discussed.

Disease Resistance↗

PubMed2025

Haplotype stacking to improve stability of stripe rust resistance in wheat.

Genotype-by-environment interaction analysis and haplotype-level characterisation provide novel insights into the stability of stripe rust resistance. Breeding selection strategies are proposed to achieve rapid and stable genetic gains across environments. This study investigated stripe/yellow rust (YR) responses in the Vavilov wheat diversity panel evaluated across 11 field experiments conducted in Australia and Ethiopia during 2014-2021. Genotype-by-environment interaction (GEI) was analysed using a factor analytic (FA) model. Genotype-level selection was performed with overall performance (OP) and root-mean-square deviation (RMSD), which reflected average performance and stability of YR resistance across environments, respectively. Genomic estimated breeding values (GEBV) for these traits were calculated and compared with those from a multi-trait GBLUP model with average performance represented by the mean GEBV across environments and stability by the standard deviation of GEBV across environments. The FA-based and multi-trait GBLUP GEBV had high correlations. Haplotypes with large effects on OP and RMSD were identified using the local GEBV method. Favourable haplotypes were then used for stacking in breeding simulations, using the Vavilov collection as a base. Compared to truncation selection, optimal haplotype selection (OHS) using an artificial intelligence (AI)-based algorithm achieved longer-term genetic gains for both OP and RMSD (after many generations) by initially selecting founder parents that maximised favourable haplotypes. Simulations using YR responses from diverse environments that mimicked fluctuating environmental conditions across seasons were conducted to evaluate strategies for selection of YR resistance that is stable across years. Strategies which gave most weight to OP, but some weight to RMSD were optimal in these conditions, and substantially reduced variation of performance across years. This study provides useful information for breeding cultivars with both high YR resistance and high stability of resistance across environments.

Triticum↗

PubMedDate not supplied

Reasoning methods in medical consultation systems: artificial intelligence approaches.

It has been argued that the problem of medical diagnosis is fundamentally ill-structured, particularly during the early stages when the number of possible explanations for presenting complaints can be immense. This paper discusses the process of clinical hypothesis evocation, contrasts it with the structured decision making approaches used in traditional computer-based diagnostic systems, and briefly surveys the more open-ended reasoning methods that have been used in medical artificial intelligence (AI) programs. The additional complexity introduced when an advice system is designed to suggest management instead of (or in addition to) diagnosis is also emphasized. Example systems are discussed to illustrate the key concepts.

Diagnosis, Computer-Assisted↗

PubMedDate not supplied

Approaching the millennium: perinatal problems and software solutions.

Strategic planning for rational development of perinatal computing capabilities for the year 2000 should be driven by anticipated trends in (1) the health care business, (2) computer technology and (3) medicine, as well as (4) the needs of perinatal practitioners. In the USA, health care is the fastest growing segment of the economy. This will produce increasing attention from hardware and software developers, and vendors, and will lead to a proliferation of computing platforms, operating systems and specific medical application software. Desktop computers, already capable of 20 million instructions per second (MIPS) with massive storage capacities, will continue to evolve and fall in price. Increasingly, perinatologists will develop software packages to facilitate patient care in their own environments. All of these trends will lead to severe fragmentation in medical computing. Simultaneously, however, the need for integrated institutional computer-based data access for quality assurance and fiscal and operations management will increase. Perinatal care will be more regionalized, complex and rigorous with new clinical trial- and effectiveness research-based interventions, as well as molecular diagnosis and therapy. To practice appropriately, clinicians will need to be familiar with computer capabilities. Having been exposed to computer-aided instruction (CAI) at the undergraduate and postgraduate levels, they will except on-line access to detailed and accurate patient information with linkage to laboratory, radiology and other medical databases, as well as to reference databases, such as Medlines and the Oxford Database of Perinatal Trials. Artificial intelligence (AI) software may support perinatal decision making; computerized professional and facility billing will be available.

Forecasting↗

PubMed2005

Improving outcomes in radiology: bringing computer-based decision support and education to the point of care.

Many computer applications have been developed in radiology and other medical disciplines to help physicians make decisions. Artificial intelligence (AI)--an approach to computer-based manipulation of symbols to simulate human reasoning--forms the basis of many of these systems. This article's goals are to: acquaint the reader with the motivations and opportunities for computer-based medical decision support systems; identify AI techniques and applications in radiology decision making; assess the impact of these technologies; and consider new directions and opportunities for AI in radiology. Among the exciting new directions is the use of AI to integrate radiology reporting, online decision support, and just-in-time learning to provide useful information and continuing education that is embedded within a radiologist's daily workflow.

Decision Making, Computer-Assisted↗

PubMed2026

Insecticidal peptides as sustainable tools for future agriculture.

The increasing global human population and the intensification of agriculture present unprecedented challenges for pest control. The escalating resistance of pests to conventional synthetic insecticides, coupled with ecological and health concerns, underscores the urgent need for innovative and sustainable management approaches. Insecticidal peptides, due to their structural diversity, molecular specificity, and biodegradability, are emerging as promising candidates for the development of next-generation bioinsecticides. This strategic roadmap synthesizes recent advances in peptide architectures, ranging from pore-forming scaffolds to designs targeting enzyme inhibition and mimicking neuroendocrine actions, with a focus on the molecular mechanisms underpinning their selectivity and efficacy. By integrating structure-function insights with translational frameworks, we identify critical knowledge gaps and propose a pathway toward biotechnological tools, including bioinspired synthesis, artificial intelligence (AI)-guided peptide engineering, and nanodelivery systems for controlled release. Our analysis positions peptide-based insecticides at the forefront of sustainable agriculture, with the potential to minimize off-target effects, reduce environmental impact, and enhance crop resilience in the face of global change.

Agricultural biotechnology↗

PubMed2026

Molecular Landscape and Advanced Diagnostic Technologies for BRAF Mutations in Cancer: From Quantitative PCR and ddPCR to CRISPR-Based Platforms.

BRAF mutations are key oncogenic alterations across multiple malignancies, including melanoma, thyroid carcinoma, colorectal cancer, non-small cell lung cancer, glioma, and hairy cell leukemia. The most prevalent variant, BRAF-V600E, induces constitutive activation of the MAPK signaling pathway, promoting tumor progression and influencing therapeutic responsiveness. Accurate detection of BRAF alterations is therefore essential for molecular classification, prognostic assessment, treatment selection, and resistance surveillance. This review summarizes the molecular heterogeneity of BRAF mutations and critically evaluates current diagnostic methodologies. Conventional approaches such as allele-specific PCR and Sanger sequencing are compared with advanced quantitative platforms, including high-resolution melting analysis, droplet digital PCR, and next-generation sequencing, with emphasis on analytical sensitivity, mutation coverage, and clinical applicability. Emerging technologies such as CRISPR-based assays, rolling circle amplification systems, and nanoparticle-based biosensors and point-of-care diagnostic platforms are also discussed for their potential to enhance ultra-sensitive detection, particularly in liquid biopsy settings. These emerging tools are highlighted for their potential to enable ultra-sensitive, rapid, and decentralized mutation detection, particularly in liquid biopsy settings. Key challenges, including intratumoral heterogeneity, low allele-frequency variants, FFPE-associated artifacts, and clonal evolution under therapeutic pressure, are examined within a translational framework. In addition, we examine critical barriers to clinical implementation, including standardization, cost, and global accessibility of molecular diagnostics, and outline potential solutions through scalable technologies and decentralized testing strategies. We propose that optimal BRAF testing requires a mutation subclass-informed and clinically integrated strategy combining comprehensive baseline profiling with longitudinal molecular monitoring. Future diagnostic paradigms will likely integrate multi-omics data and artificial intelligence (AI)-assisted interpretation to refine precision oncology implementation. Looking forward, we propose that optimal BRAF testing will require integration of multi-omics profiling with AI-assisted interpretation, enabling automated variant classification, real-time clinical decision support, and improved prediction of therapeutic response and resistance.

Humans↗

PubMed2026

Artificial intelligence agents and agentic artificial intelligence applied to precision medicine.

Precision medicine seeks to individualise care by integrating multimodal biomedical data, yet most deployed clinical artificial intelligence (AI) remains assistive, providing predictions without managing workflows or adapting autonomously. Agentic AI, built on large language models (LLMs), has emerged as a paradigm characterised by autonomy, goal-directed reasoning, memory, planning and tool use. This review synthesises evidence on agentic AI and LLMs applied to precision medicine, encompassing drug discovery, genomics, oncology, rare disease diagnostics and clinical pharmacology. This review also examines architectural components, recent validation milestones and emerging challenges, including hallucination, sociodemographic bias and evolving regulatory frameworks across the FDA, the EU AI Act and the WHO.

agentic AI↗

PubMed2026

Digital pathology, image analysis, and artificial intelligence in liver disease.

Advances in digital pathology, image analysis, and artificial intelligence (AI) are rapidly transforming how pathologists and researchers interact with tissue samples and enable the development of diagnostic tools that harness high-resolution whole-slide images; these advances are in turn creating new opportunities for research, education, and routine clinical care globally. Liver disease is no exception, and digital pathology and AI have many applications in the diagnosis of liver cancer and liver diseases and in the assessment and management of transplantation. Although quantitative image analysis techniques have been applied to liver disease in research settings for over 50 years, recent improvements in image resolution, data storage, and the availability of advanced AI methods such as deep learning have driven multiple exciting developments. In this Review, we summarise the advancements in digital pathology, image analysis, and AI in liver disease. Key challenges such as access to and the logistics of using digital solutions, quality issues, and appropriate guidance in research and clinical use are reviewed, along with potential solutions to these challenges in the context of liver pathology and liver disease. Digital technologies are well established in liver pathology research, and access in clinical practice is increasing, with potential to address current laboratory challenges. Further evaluation is required to assess real-world effectiveness, clinical safety, and implementation of AI tools in liver pathology.

Journal Article↗

PubMed2025

ChemPerturb-seq screen identifies a small molecule cocktail enhancing human beta cell survival after subcutaneous transplantation.

Traditional chemical screens have focused on a single assay per screen, making them labor intensive and costly. Here, we combined a chemical screen with single-cell RNA sequencing (scRNA-seq) to perform Chemical Perturb-seq (ChemPerturb-seq), enabling a systematic analysis of the molecular changes of human beta cells upon individual small molecule treatments. Using this platform, we performed an in vivo barcoded screen and discovered a small molecule cocktail, including beta-lipotropin 61-91, insulin growth factor-1, and prostaglandin E2, with which preconditioning human beta cells and primary islets significantly enhanced function and survival when transplanted subcutaneously to female, but not to male, mice. We identified two additional molecules, serotonin and histamine, that promote islet function when transplanted subcutaneously to male mice using ChemPerturb-seq. Such small molecule cocktails could be applied to improve the current FDA-approved islet transplantation procedure. Finally, we developed an artificial intelligence (AI)-powered website, ChemPerturbDB, which provides user-friendly open access analysis of the extensive ChemPerturb-seq dataset.

Humans↗

PubMed2025

Nanobioreactor detection of space-associated hematopoietic stem and progenitor cell aging.

Human hematopoietic stem and progenitor cell (HSPC) fitness declines following exposure to stressors that reduce survival, dormancy, telomere maintenance, and self-renewal, thereby accelerating aging. While previous National Aeronautics and Space Administration (NASA) research revealed immune dysfunction in low-earth orbit (LEO), the impact of spaceflight on human HSPC aging had not been studied. To study HSPC aging, our NASA-supported Integrated Space Stem Cell Orbital Research (ISSCOR) team developed bone marrow niche nanobioreactors with lentiviral bicistronic fluorescent, ubiquitination-based cell-cycle indicator (FUCCI2BL) reporter for real-time HSPC tracking in artificial intelligence (AI)-driven CubeLabs. In month-long International Space Station (ISS) missions (SpX-24, SpX-25, SpX-26, and SpX-27) compared with ground controls, FUCCI2BL reporter, whole-genome and transcriptome sequencing, and cytokine arrays demonstrated cell-cycle, inflammatory cytokine, mitochondrial gene, human repetitive element, and apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like 3 (APOBEC3) deregulation together with clonal hematopoietic mutations. Furthermore, HSPC functionally organized multi-omics aging (HSPC-FOMA) analyses revealed reduced telomere maintenance, adenosine deaminase acting on RNA1 (ADAR1) p150 self-renewal gene expression, and replating capacity indicative of space-associated HSPC aging that may limit long-duration spaceflight.

Humans↗

PubMedDate not supplied

Comparative evaluation of the use of artificial neural networks for modelling the epidemiology of schistosomiasis mansoni.

There has been a marked increase in the application of approaches based on artificial intelligence (AI) in the field of computer science and medical diagnosis, but AI is still relatively unused in epidemiological settings. In this study we report results of the application of neural networks (NN; a special category of AI) to schistosomiasis. It was possible to design an NN structure which can process and fit epidemiological data collected from 251 schoolchildren in Egypt using the first year's data to predict second and third years' infection rates. Data collected over 3 years included age, gender, exposure to canal water and agricultural activities, medical history and examination, and stool and urine parasitology. Schistosoma mansoni infection rates were 50%, 78% and 66% at the baseline and the 2 follow-up periods, respectively. NN modelling was based on the standard back-propagation algorithm, in which we built a suitable configuration of the network, using the first year's data, that optimized performance. It was implemented on an IBM compatible computer using commercially available software. The performance of the NN model in the first year compared favourably with logistic regression (NN sensitivity = 83% (95% confidence interval [CI] 78-88%) and positive predictive value (PPV) = 63% (95% CI 57-69%); logistic regression sensitivity = 66% (95% CI 60%-72%) and PPV = 59% (95% CI 53%-65%). The NN model generalized the criteria for predicting infection over time better than logistic regression and showed more stability over time, as it retained its sensitivity and specificity and had better false positive and negative profiles than logistic regression. The potential of NN-based models to analyse and predict wide-scale control programme data, which are inevitably based on unstable egg excretion rates and insensitive laboratory techniques, is promising but still untapped.

Adolescent↗