Celiac disease remains underdiagnosed despite an established diagnostic pathway, reflecting phenotypic heterogeneity, delayed recognition, and dependence on resource-intensive testing. This narrative review synthesizes evidence on artificial intelligence and phenomics in celiac disease (CD) across four operational layers: electronic health record phenotyping, Human Phenotype Ontology-based semantic encoding, machine-learning pre-screening from routine clinical data, and deep-learning-assisted histopathology. A targeted literature search of PubMed/MEDLINE, Embase, and Google Scholar covered publications from January 2010 through December 2025, using combinations of CD/coeliac disease with artificial intelligence, machine learning, deep learning, phenomics, Human Phenotype Ontology, computable phenotype, electronic health records, and natural language processing, supplemented by targeted searches and citation chaining. We present a curated minimum viable CD phenome and discuss clinical actionability, age-specific considerations, and current pediatric validation gaps. Selected models have demonstrated promising CD detection or pre-screening performance in specific datasets, but prospective, multicenter evidence with external validation remains insufficient to establish clinical utility. Artificial intelligence should augment expert clinical care rather than replace it.

Anal canal melanoma (ACM) is a rare, aggressive malignancy with an ominous prognosis. However, its associated factors are largely unknown. Therefore, we investigated potential independent associations of sex and self-reported race with ACM and physiological anal canal pigmentation.

Two case-control cohorts from a single institution in Miami, Florida, were evaluated. The biopsy cohort comprised 117 sequential patients evaluated prospectively for melanocytic cells/pigment and anal intraepithelial neoplasia (AIN)/squamous intraepithelial lesion (SIL) in anal transitional zone biopsies between January 2021 and August 2022. The melanoma cohort consisted of 28 patients diagnosed with ACM between January 2003 and August 2021 and 116 of the patients in the pigmentation cohort. Multivariable logistic regression analysis was conducted to examine independent factor associations.

In the biopsy cohort, anal transitional melanocytic cells/pigment were identified in 48% (26/54) of Black patients compared to 17% (10/59) of White patients. Multivariable analysis demonstrated that White race was inversely associated with mucosal pigmentation (odds ratio (OR) = 0.19, 95% confidence interval (CI): 0.08–0.48, P < 0.001). Conversely, White race was associated with ACM (OR = 6.25, 95% CI: 2.07–18.81, P = 0.001). Male sex was also inversely associated with ACM (OR = 0.18, 95% CI: 0.07–0.47, P < 0.001). No significant correlations were observed between mucosal pigmentation and AIN/SIL or human immunodeficiency virus status.

Anal transitional zone pigmentation is more prevalent in self-reported Black individuals but is not associated with neoplastic squamous cell precursors. White (versus Black) individuals and female (versus male) individuals both have a significant association with ACM.

Hepatitis B virus (HBV) vertical transmission remains a major health challenge, and the characteristics of the placental immune microenvironment in chronic infection remain unclear. This study aimed to use mass cytometry to comprehensively analyze phenotypic changes in placental immune cell subsets.

We collected placental tissues from 20 pregnant women (6 healthy controls and 14 with chronic HBV infection). CD45+ leukocytes were detected using a 35-marker antibody panel. Unsupervised clustering identified 14 clusters, of which 13 immune clusters were analyzed.

HBV mainly caused functional remodeling of placental immune cells rather than changes in cell composition (P > 0.05 for all subsets). In natural killer (NK) cells, CD38 (P = 0.015) and CD16 (P = 0.0020) were notably upregulated and strongly correlated (ρ = 0.872, P < 0.001), suggesting potentially enhanced ADCC function. T cells showed upregulation of CD27, CD38, and CXCR5. Basophils exhibited the most pronounced changes: 9 differential markers (including chemokine receptors CCR4 and CCR7) were consistently downregulated. Classical monocytes showed an M1 polarization tendency (CD38↑, CD163↓). All P-values were raw; no markers remained significant after false discovery rate correction at an FDR threshold of < 0.1, indicating nominal significance. In addition, the coordinated expression patterns of markers within multiple cell subsets were weakened after infection.

These exploratory findings suggest that HBV may reshape placental immunity through enhanced NK cell activation, broad basophil suppression, and disrupted marker coordination. The sample size was limited (n = 20), so the results need to be validated in larger cohorts. These findings provide new perspectives for understanding the mechanisms of mother-to-child transmission.

Inter-individual variability in drug efficacy and toxicity remains a major obstacle to precision therapeutics. Candidate-gene pharmacogenomics and star-allele-based guidelines have established clinically useful examples, but they cannot capture the full spectrum of mechanisms that shape drug response. Pharmacogenomic genome-wide association studies (pharmacoGWAS) extend this framework by enabling discovery beyond known pharmacogenes and can identify human genetic variants associated with efficacy, adverse drug reactions, dose requirements, pharmacokinetics, and pharmacodynamics. This mini-review aims to summarize practical principles for human pharmacoGWAS, with emphasis on study design, phenotype and exposure definition, reproducible bioinformatics pipelines, gene-expression-based functional interpretation, and clinical translation. This review discusses randomized trials, prospective cohorts, biobanks, electronic health records, claims databases, and rare adverse-event designs, highlighting the specific biases that arise because drug response is defined among exposed individuals. It then outlines core analytical steps, including genotype quality control, imputation, ancestry-aware association testing, mixed models, survival and longitudinal analyses, rare-variant aggregation, replication, and meta-analysis. Particular attention is given to expression quantitative trait loci, splicing quantitative trait loci, and protein quantitative trait loci, tissue prioritization informed by the Genotype-Tissue Expression project, transcriptome-wide association studies, and colocalization as tools for prioritizing candidate genes and plausible mechanisms. Finally, we propose a translation framework connecting discovery to clinical validity, guideline development, electronic health record decision support, and equitable implementation across diverse populations. When combined with rigorous epidemiology and functional genomics, pharmacoGWAS may help translate genome-wide signals into safer and more effective prescribing.

Machine-learning approaches that combine predictive performance with model interpretability may improve lung cancer status classification. This study aimed to develop and internally evaluate an Explainable Precision Screening Framework for lung cancer risk classification using demographic, behavioral, and symptom-based variables from a publicly available dataset.

This retrospective cross-sectional study analyzed a publicly available Kaggle dataset containing 309 records, including 270 labeled as lung cancer and 39 as non-cancer. Six models—logistic regression, support vector machine (SVM), random forest, LightGBM, XGBoost, and a stacking ensemble—were compared using a stratified hold-out test set and repeated stratified five-fold cross-validation. Performance was assessed using classification metrics with bootstrap 95% confidence intervals (CIs). A separate Shapley additive explanations (SHAP) analysis was applied to the standalone SVM model for exploratory feature attribution.

Across all models, accuracy ranged from 85% to 92% (ROC-AUC: 0.93–0.95). The stacking ensemble achieved 0.92 accuracy (95% CI: 0.85–0.98), 0.94 sensitivity (95% CI: 0.88–1.00), 0.75 specificity (95% CI: 0.40–1.00), 0.96 precision (95% CI: 0.89–1.00), 0.95 F1 score (95% CI: 0.91–0.99), and 0.95 ROC-AUC (95% CI: 0.89–0.99). Its PR-AUC was 0.993 (95% CI: 0.981–0.999), and its Brier score was 0.074 (95% CI: 0.037–0.121). SHAP analysis identified smoking, yellow fingers, coughing, chest pain, wheezing, shortness of breath, and age as the features contributing most strongly to its predictions.

Within this public retrospective dataset, the stacking ensemble was among the highest-performing models, whereas separate SHAP analysis of the standalone SVM model identified the features contributing most strongly to its predictions. Given the marked class imbalance, the absence of a reported clinical reference standard for the source labels, the framework requires independent validation in clinically verified multicenter cohorts before clinical use.

Metabolic dysfunction-associated steatotic liver disease affects roughly 38% of adults, yet approved agents do not correct the upstream redox-metabolic perturbations—a depressed nicotinamide adenine dinucleotide (NAD+/NADH) ratio, saturated lipid excess, and endoplasmic reticulum (ER) stress—that drive hepatocyte injury. Cytochrome b5 reductase 3 (CYB5R3) couples NADH oxidation to fatty acid desaturation, nuclear factor erythroid 2-related factor 2 (NRF2)-linked antioxidant and cholesterol-handling pathways, NAD+/sirtuin signaling, and ER-phagy, and is the sole electron input to mitochondrial amidoxime-reducing component 1 (mARC1). This review grades every link in the axis across the steatosis–cirrhosis–hepatocellular carcinoma spectrum, reporting effect estimates with sample sizes and test statistics alongside a study-level appraisal of clinical relevance. The common MTARC1 p.A165T variant protects against all-cause cirrhosis (odds ratio, 0.91; 95% CI, 0.89–0.94; P = 2.3 × 10−11; 12,361 cases, 790,095 controls), with lower hepatic fat, liver enzyme levels, and low-density lipoprotein cholesterol levels. Germline mARC1 deletion reduces picrosirius red fibrosis area by 24–50% depending on diet, without altering histological disease activity, whereas partial protein reduction confers no protection. Critically, therapeutic hepatocyte-directed knockdown loses its anti-fibrotic effect when started at higher disease burden and in the choline-deficient model: efficacy depends on the depth, compartment, and timing of inhibition. Deep, hepatocyte-restricted mARC1 inhibition by GalNAc-conjugated oligonucleotides—which also avoids a male-predominant cardiac liability—is therefore the most credible near-term strategy, only in pre-cirrhotic F2–F3 disease. CYB5R3 activation, and its combination with mARC1 inhibition, remain unproven, and no clinical trial of this axis has been reported. Falsifiable in vivo and pharmacodynamic biomarker roadmaps are proposed, together with the efficacy, delivery, and safety uncertainties specific to advanced cirrhosis and a non-invasive pharmacodynamic framework.