This narrative review aimed to (1) map the global research landscape of artificial intelligence (AI) in dementia care, (2) characterize domestic Korean research, and (3) propose priority directions for nurse-led AI research. Methods: International literature was retrieved from PubMed (January 2015 to December 2025; n=5,710) and analyzed using VOSviewer with the Leiden algorithm. Korean literature from RISS (n=265) was synthesized narratively. Results: Publications increased substantially since 2015, with acceleration after 2024 driven by large language models. Five thematic clusters emerged: (1) care, daily living, and digital health; (2) aging and brain structure; (3) risk prediction and clinical data; (4) early screening and behavioral biomarkers; and (5) Alzheimer’s diagnosis and deep learning. Four domains were most relevant to nursing: monitoring technologies, digital biomarkers, AI chatbots for caregivers, and clinical decision support systems. Korean research concentrated on emotional support robots with methodological limitations. Conclusion: Rigorous nurse-led trials with multidimensional outcomes are needed to establish AI’s clinical value in dementia nursing.

Objectives: To develop a Loneliness and Social Isolation scale (LSIS) that can measure both social isolation and loneliness in order to understand the degree of social isolation in Korea. Exploratory and confirmatory factor analysis was used to examine the factorial validity of the scale. Methods: The subjects of the study were 300 adults aged 19 or older who visited Samsung Medical Center and voluntarily expressed their willingness to participate in this research. Exploratory factor analysis (n=150) and confirmatory factor analysis (n=150) were conducted to construct the factorial structure model and to determine the model fit. Results: Exploratory factor analysis showed a three-factor structure with a total variance of 65.8%; factor 1 consisted of social support, factor 2 of social networks, and factor 3 of items representing loneliness. After conducting confirmatory factor analysis on the three-factor models, a three-factor model consisting of 8 items (LSIS-8) and a three-factor model consisting of 6 items (LSIS-6) showed significant goodness-of-fit. Internal consistency for all items was good (Cronbach’s α=0.774), and correlations with existing social isolation and loneliness measures were significant. Conclusion This study is meaningful as provides a tool that comprehensively measures social support, social networks, and loneliness. We believe that the application of such tools that are relatively easy to apply in communities will aid understanding of the current state of social isolation and loneliness in Korea.

Background: Pheochromocytoma and paragangliomas (PPGL) are known as tumors with the highest level of heritability, approximately 30% of all cases. Clinical practice guidelines of PPGL recommend genetic testing for germline variants in all patients. In this study, we used whole exome sequencing to identify novel causative variants associated with PPGL to improve the detection of rare genetic variants in our cohort. Methods: Thirty-six tested negative for pathogenic variants in previous Sanger sequencing or targeted gene panel testing for PPGL underwent whole exome sequencing. Whole exome sequencing was performed using DNA samples enriched using TruSeq Custom Enrichment Kit and sequenced with MiSeq (Illumina Inc.). Sequencing alignment and variant calling were performed using SAMtools. Results: Among previously mutation undetected 36 patients, two likely pathogenic variants and 13 variants of uncertain significance (VUS) were detected in 32 pheochromocytoma-related genes. SDHA c.778G>A (p.Gly260Arg) was detected in a patient with head and neck paraganglioma, and KIF1B c.2787-2A>C in a patient with a bladder paraganglioma. Additionally, a likely pathogenic variant in BRCA2, VUS in TP53, and VUS in NFU1 were detected. Conclusion Exome sequencing further identified genetic alterations by 5.6% in previously mutation undetected patients in PPGL. Implementation of targeted gene sequencing consisted of extended genes of PPGL in routine clinical screening can support the level of comprehensive patient assessment.