Purpose: To estimate the sample size required for developing artificial intelligence (AI) that can predict soft-tissue and alveolar bone changes following orthodontic treatment. Materials and Methods: From the original data sets with N=887, consisting of 132 input and 88 output variables used to create AI models for predicting treatment changes following orthodontic treatment, six subsets of the data (n=75, 150, 300, 450, 600, and 750) were generated through random resampling procedures. The process was repeated four times, resulting in 24 different data subsets. Each data subset was used to create a total of 24 AI models using the TabNet deep neural network algorithm. The clinically acceptable prediction accuracy was defined as a less than 1.5 mm prediction error on the lower lip. The prediction errors from each AI model were compared according to sample sizes and analyzed to estimate the optimal sample size. Results: The prediction error decreased with increasing sample sizes. A training sample size greater than approximately 1650 was estimated to develop an AI model with less than 1.5 mm of prediction errors at the lower lip area. Conclusion From a statistical and research design perspective, a considerable amount of training data appears necessary to develop an AI prediction model with clinically acceptable accuracy.

Background: With growing elderly populations, management of patients with acute stroke is increasingly important. In South Korea, the Acute Stroke Quality Assessment Program (ASQAP) has contributed to improving the quality of stroke care and practice behavior in healthcare institutions. While the mortality of hemorrhagic stroke remains high, there are only a few assessment indices associated with hemorrhagic stroke. Considering the need to develop assessment indices to improve the actual quality of care in the field of acute stroke treatment, this study aims to investigate the current status of human resources and practices related to the treatment of patients with acute stroke through a nationwide survey. Methods: For the healthcare institutions included in the Ninth ASQAP of the Health Insurance Review and Assessment Service (HIRA), data from January 2022 to December 2022 were collected through a survey on the current status and practice of healthcare providers related to the treatment of patients with acute stroke. The questionnaire consisted of 19 items, including six items on healthcare providers involved in stroke care and 10 items on the care of patients with acute stroke. Results: In the treatment of patients with hemorrhagic stroke among patients with acute stroke, neurosurgeons were the most common providers. The contribution of neurosurgeons in the treatment of ischemic stroke has also been found to be equivalent to that of neurologists. However, a number of institutions were found to be devoid of healthcare providers who perform definitive treatments, such as intra-arterial thrombectomy for patients with ischemic stroke or cerebral aneurysm clipping for subarachnoid hemorrhage. The intensity of the workload of healthcare providers involved in the care of patients with acute stroke, especially those involved in definitive treatment, was also found to be quite high. Conclusion Currently, there are almost no assessment indices specific to hemorrhagic stroke in the ASQAP for acute stroke. Furthermore, it does not reflect the reality of the healthcare providers and practices that provide definitive treatment for acute stroke. The findings of this study suggest the need for the development of appropriate assessment indices that reflect the realities of acute stroke care.

Background: An ideal lung allocation system should reduce waiting list deaths, improve transplant survival, and ensure equitable organ allocation. This study aimed to develop a novel lung allocation score (LAS) system, the MaxBenefit LAS, to maximize transplant benefits. Methods: This study retrospectively analyzed data from the Korean Network for Organ Sharing database, including 1,599 lung transplant candidates between September 2009 and December 2020. We developed the MaxBenefit LAS, combining a waitlist mortality model and a post-transplant survival model using elastic-net Cox regression, was assessed using area under the curve (AUC) values and Uno’s C-index. Its performance was compared to the US LAS in an independent cohort. Results: The waitlist mortality model showed strong predictive performance with AUC values of 0.834 and 0.818 in the training and validation cohorts, respectively. The post-transplant survival model also demonstrated good predictive ability (AUC: 0.708 and 0.685). The MaxBenefit LAS effectively stratified patients by risk, with higher scores correlating with increased waitlist mortality and decreased post-transplant mortality. The MaxBenefit LAS outperformed the conventional LAS in predicting waitlist death and identifying candidates with higher transplant benefits. Conclusion The MaxBenefit LAS offers a promising approach to optimizing lung allocation by balancing the urgency of candidates with their likelihood of survival post-transplant. This novel system has the potential to improve outcomes for lung transplant recipients and reduce waitlist mortality, providing a more equitable allocation of donor lungs.

Inclisiran and bempedoic acid (BA) are non-statin lipid-lowering agents that have been approved for use in the US and Europe. Inclisiran, a subcutaneously administered small interfering RNA targeting proprotein convertase subtilisin/kexin type 9 messenger RNA, is effectively delivered to the liver via lipid nanoparticles and conjugation. In several phase 3 trials, it has successfully reduced low-density lipoprotein cholesterol (LDL-C) by 50% and has an acceptable safety profile. Currently, the results of clinical outcome studies are awaited. While it is indicated for both primary and secondary cardiovascular prevention, it is selectively recommended after statin-based regimens. BA, an oral inhibitor of adenosine triphosphate-citrate lyase, decreases cholesterol production and enhances LDL uptake by hepatocytes. This enzyme is absent in muscle cells, and BA has fewer muscle-related adverse events. In clinical trials, it lowered LDL-C by 17%–21% compared to placebo and showed a clinical outcome benefit in patients with statin intolerance. This agent modestly increases the incidence of gout and cholelithiasis. For primary and secondary prevention, it may be recommended as a non-first-line agent, either alone or in combination therapy.

Background: To identify the optimal surgical technique for single-level anterior cervical discectomy and fusion (ACDF), this study compared surgical outcomes and incidence of adjacent segment degeneration (ASD) in patients undergoing single-level ACDF using cage alone single-level fusion and plate fixation techniques. Methods: This single-center retrospective study (2003–2018) included patients who underwent single-level ACDF with either plate fixation (PLATE) or cage (CAGE) alone. The radiologic and clinical outcomes between the 2 surgical groups were compared over a 4-year follow-up period. Outcomes of interest included parameters related to range of motion, sagittal alignment, as well as fusion, subsidence, and ASD rates. Clinical outcomes were evaluated using the Neck Disability Index (NDI) and visual analog scale (VAS) for pain. Dysphagia and hoarseness rates were estimated based on medical records. Results: Forty-seven patients were included (n=17 in CAGE group). In the CAGE group, 94.1% of the patients had Bridwell grade 1 or 2, compared to 83.3% in the PLATE group (p = 0.396). Subsidence occurred in 12.5% and 3.6% of the CAGE and PLATE cases, respectively (p = 0.543). Segmental kyphosis progressed in the CAGE group compared to the PLATE group at 12, 24, and 48 months (p < 0.001). Radiographic ASD was observed in 41.2% and 30.0% of patients in the CAGE and PLATE groups, respectively, with a higher incidence in the upper segments for both groups. Preoperative NDI scores were similar between the groups; however, postoperatively, the CAGE group had significantly lower NDI scores (3.50 ± 2.74 vs. 8.00 ± 5.81) at 4 years (p = 0.020). Neck pain VAS scores also showed significant improvement in the CAGE group (2.33 ± 2.94) compared with that in the PLATE group (3.07 ± 2.31) at 4 years (p = 0.045). Both groups showed comparable arm pain VAS scores at 2 and 4 years postoperatively. Postoperative dysphagia occurred in 1 patient in the PLATE group, resolving almost completely by 1 year. Conclusions Single-level ACDF using a cage alone technique demonstrated favorable radiologic and clinical outcomes overall compared to plate-augmented ACDF. However, plate augmentation is recommended for patients with severe cervical kyphosis or those at high risk of subsidence.

Background: and Purpose: This clinical practice guideline provides evidence-based recommendations for treatment of dementia, focusing on cholinesterase inhibitors and N-methyl-D-aspartate (NMDA) receptor antagonists for Alzheimer’s disease (AD) and other types of dementia. Methods: Using the Population, Intervention, Comparison, Outcomes (PICO) framework, we developed key clinical questions and conducted systematic literature reviews. A multidisciplinary panel of experts, organized by the Korean Dementia Association, evaluated randomized controlled trials and observational studies. Recommendations were graded for evidence quality and strength using Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) methodology. Results: Three main recommendations are presented: (1) For AD, cholinesterase inhibitors (donepezil, rivastigmine, galantamine) are strongly recommended for improving cognition and daily function based on moderate evidence; (2) Cholinesterase inhibitors are conditionally recommended for vascular dementia and Parkinson’s disease dementia, with a strong recommendation for Lewy body dementia; (3) For moderate to severe AD, NMDA receptor antagonist (memantine) is strongly recommended, demonstrating significant cognitive and functional improvements. Both drug classes showed favorable safety profiles with manageable side effects. Conclusions This guideline offers standardized, evidence-based pharmacologic recommendations for dementia management, with specific guidance on cholinesterase inhibitors and NMDA receptor antagonists. It aims to support clinical decision-making and improve patient outcomes in dementia care. Further updates will address emerging treatments, including amyloid-targeting therapies, to reflect advances in dementia management.

Although oxidative stress is the main pathophysiology of the development of diabetic neuropathy, oral administration of antioxidants has given disappointing results. Here, we hypothesized that local delivery of antioxidants would provide protective effects on Schwann cells due to the high concentration of local lesions. We prepared alpha-tocopherol (ATF)-loaded liposomes and tested their skin penetration after sonication. An in vitro study using IMS-32 cells was conducted to determine the level of reactive oxygen species (ROS) scavenging effects of ATF-liposomes. ATF reduced ROS in high-glucose-exposed IMS-32 cells in a dosedependent manner. ATF-liposomes also reduced the ROS level in vitro and ultrasound irradiation enhanced delivery to the dermis in porcine ear skin. This study showed that it is feasible to deliver ATF through the skin and can effectively reduce ROS. This model is worthy of development for clinical use.

Purpose: This study aimed to develop a comprehensive practical guide for enteral nutrition (EN) designed to enhance patient safety and reduce complications in Korea. Under the leadership of the Korean Society for Parenteral and Enteral Nutrition (KSPEN), the initiative sought to standardize EN procedures, improve decision-making, and promote effective multidisciplinary communication. Methods: The KSPEN EN committee identified key questions related to EN practices and organized them into seven sections such as prescribing, delivery route selection, formula preparation, administration, and quality management. Twenty-one experts, selected based on their expertise, conducted a thorough literature review to formulate evidence-based recommendations. Drafts underwent peer review both within and across disciplines, with final revisions completed by the KSPEN Guideline Committee. The guide, which will be published in three installments, addresses critical elements of EN therapy and safety protocols. Results: The practical guide recommends that EN orders include detailed elements and advocates the use of electronic medical records for communication. Standardized prescription forms and supplementary safety measures are outlined. Review frequency is adjusted according to patient condition—daily for critically ill or unstable patients and as dictated by institutional protocols for stable patients. Evidence indicates that adherence to these protocols reduces mortality, complications, and prescription errors. Conclusion The KSPEN practical guide offers a robust framework for the safe delivery of EN tailored to Korea’s healthcare context. It emphasizes standardized protocols and interdisciplinary collaboration to improve nutritional outcomes, patient safety, and operational efficiency. Rigorous implementation and monitoring of adherence are critical for its success.

The common data model (CDM) has found widespread application in healthcare studies, but its utilization in cancer research has been limited. This article describes the development and implementation strategy for Cancer Clinical Library Databases (CCLDs), which are standardized cancer-specific databases established under the Korea-Clinical Data Utilization Network for Research Excellence (K-CURE) project by the Korean Ministry of Health and Welfare. Fifteen leading hospitals and fourteen academic associations in Korea are engaged in constructing CCLDs for 10 primary cancer types. For each cancer type-specific CCLD, cancer data experts determine key clinical data items essential for cancer research, standardize these items across cancer types, and create a standardized schema. Comprehensive clinical records covering diagnosis, treatment, and outcomes, with annual updates, are collected for each cancer patient in the target population, and quality control is based on six-sigma standards. To protect patient privacy, CCLDs follow stringent data security guidelines by pseudonymizing personal identification information and operating within a closed analysis environment. Researchers can apply for access to CCLD data through the K-CURE portal, which is subject to Institutional Review Board and Data Review Board approval. The CCLD is considered a pioneering standardized cancer-specific database, significantly representing Korea’s cancer data. It is expected to overcome limitations of previous CDMs and provide a valuable resource for multicenter cancer research in Korea.

Purpose: The molecular classification of breast cancer is crucial for effective treatment. The emergence of digital pathology has ushered in a new era in which weakly supervised learning leveraging whole-slide images has gained prominence in developing deep learning models because this approach alleviates the need for extensive manual annotation. Weakly supervised learning was employed to classify the molecular subtypes of breast cancer. Materials and Methods: Our approach capitalizes on two whole-slide image datasets: one consisting of breast cancer cases from the Korea University Guro Hospital (KG) and the other originating from The Cancer Genomic Atlas dataset (TCGA). Furthermore, we visualized the inferred results using an attention-based heat map and reviewed the histomorphological features of the most attentive patches. Results: The KG+TCGA-trained model achieved an area under the receiver operating characteristics value of 0.749. An inherent challenge lies in the imbalance among subtypes. Additionally, discrepancies between the two datasets resulted in different molecular subtype proportions. To mitigate this imbalance, we merged the two datasets, and the resulting model exhibited improved performance. The attentive patches correlated well with widely recognized histomorphologic features. The triple-negative subtype has a high incidence of high-grade nuclei, tumor necrosis, and intratumoral tumor-infiltrating lymphocytes. The luminal A subtype showed a high incidence of collagen fibers. Conclusion The artificial intelligence (AI) model based on weakly supervised learning showed promising performance. A review of the most attentive patches provided insights into the predictions of the AI model. AI models can become invaluable screening tools that reduce costs and workloads in practice.