Recent advancements in artificial intelligence have led to increased interest in predictive modeling across various domains, including medicine. Although numerous metrics have been established for binary classification, the growing adoption of multi-class systems necessitates robust evaluation measures. However, comprehensive simulation studies comparing the performance of existing multi-class metrics under diverse data conditions remain limited. In this study, we first provide a concise overview of commonly used accuracy metrics for multi-class classification. Then, we report a simulation study that systematically evaluates several diagnostic accuracy measures under a wide range of scenarios, including three- and five-class settings, balanced and imbalanced sample sizes, and different distributional assumptions for predictors. We assessed each metric’s performance in terms of bias and 95% confidence interval coverage. Under balanced conditions, most metrics demonstrated stable and unbiased performance, closely approximating the true values. However, under imbalanced conditions, greater bias was observed, with the M-index and polytomous discrimination index exhibiting comparatively more stable performance across various scenarios. The micro-averaged receiver operating characteristic curve area consistently showed higher bias under class imbalance. Finally, we applied these metrics to a glioma tumor grading task using external datasets. This study provides a systematic comparison of commonly used metrics and offers practical guidance for selecting appropriate measures in multi-class classification tasks.

Objective: Although most gallbladder polyps are benign, some neoplastic polyps may be malignant or may serve as precursors to malignancy. Distinguishing neoplastic and non-neoplastic polyps using imaging examinations remains a major challenge.This meta-analysis aimed to identify the ultrasound (US) features that are significantly associated with neoplastic polyps. Materials and Methods: The MEDLINE, EMBASE, Cochrane, and KoreaMed databases were searched for articles published up to August 31, 2025. Bivariate random-effects models were used to calculate the meta-analytic pooled diagnostic odds ratios (DORs), sensitivities, and specificities, along with their 95% confidence intervals (CIs), for each US imaging feature in the diagnosis of neoplastic polyps. Results: Thirty studies evaluating 8,953 patients, including 1,216 (13.6%) patients with neoplastic polyps, were included.Among the nine evaluated US imaging features, namely, size ≥10 mm, sessile morphology, single polyp, coexisting gallstones, hypoechogenicity, heterogeneous echogenicity, gallbladder wall thickening (GBWT), absence of hyperechoic spot, and vascularity, eight were significantly associated with neoplastic polyps: size ≥10 mm (DOR: 6.23 [95% CI: 1.86– 20.90]), sessile morphology (DOR: 3.54 [1.93–5.97]), single polyp (DOR: 2.21 [1.76–2.74]), coexisting gallstones (DOR:1.86 [1.29–2.60]), hypoechogenicity (DOR: 3.55 [1.47–7.30]), GBWT (DOR: 9.38 [1.47–32.20]), absence of hyperechoic spots (DOR: 4.23 [2.46–6.83]), and vascularity (DOR: 9.72 [5.81–15.30]). Of these, size ≥10 mm demonstrated the highest pooled sensitivity (0.79 [95% CI: 0.68–0.87]), whereas hypoechogenicity showed the highest pooled specificity (0.93 [95% CI: 0.82–0.98]). Conclusion Eight US imaging features (size ≥10 mm, sessile morphology, single polyp, coexisting gallstones, hypoechogenicity, GBWT, absence of hyperechoic spots, and vascularity) were significantly associated with the presence of neoplastic polyps.These features may facilitate the management of gallbladder polyps.

Objective: This study aimed to evaluate the feasibility, safety, and oncologic outcomes of irreversible electroporation (IRE) using a clustered electrode in patients with locally advanced pancreatic cancer (LAPC). Materials and Methods: In this single-center prospective cohort study, 13 patients with LAPC (median age, 60 years; range, 48–78 years) underwent clustered electrode IRE between September 2022 and September 2024. Patient characteristics, procedural details, and clinical outcomes were recorded. Endpoints included technical success, procedure-related complications, overall survival (OS), and progression-free survival (PFS). Results: Tumors were located in the pancreatic head in four patients (30.8%) and in the body/tail in nine (69.2%). The median tumor size was 2.4 cm (1.5–4.0 cm), and vascular invasion was present in all patients. Technical success was achieved in all patients. Intraoperative IRE was performed in 11 (84.6%) patients, and 2 (15.4%) patients underwent percutaneous IRE. Gastrointestinal bleeding events as major complications occurred in two patients (15.4%) and, both were successfully controlled by embolization. No 60-day mortality was observed. At a median follow-up of 24.5 months (range, 9.9–33.4 months) after IRE, median OS and PFS from IRE were 20.1 and 14.5 months, respectively. Conclusion IRE using clustered electrodes for LAPC appears to be a feasible therapeutic approach, offering reliable technical success and acceptable safety. Survival outcomes are encouraging; however, larger, controlled studies are required.

Objective: This study aimed to assess the association between exposure to gadolinium-based contrast agents (GBCAs) and the risk of parkinsonism according to the GBCA type. Materials and Methods: Individuals aged ≥40 years who underwent first-ever magnetic resonance imaging (MRI) examinations between 2011 and 2014 were identified from the Korean nationwide population-based health insurance claims database and followed up until 2022. Individuals were divided into those who underwent at least one GBCA-enhanced MRI, and those who underwent only non-enhanced MRI. GBCA-exposed individuals were further categorized into those exposed only to linear or macrocyclic GBCAs, after excluding those exposed to both types. The primary event of interest was allcause parkinsonism. Secondary events included all-cause parkinsonism requiring medication, Parkinson’s disease (PD), atypical parkinsonism, and secondary parkinsonism. Hazard ratios (HRs) were estimated using multivariable Cox proportional hazard regression models for exposure to linear and macrocyclic GBCAs, with the non-enhanced MRI group serving as a reference. The models were adjusted for age, sex, smoking status, alcohol consumption, regular exercise, body mass index, estimated glomerular filtration rate, and comorbidities. Subgroup analyses were performed according to age, sex, renal function, and history of cancer. Results: A total of 222,977 individuals were included in this study. Among them, 92,230, 48,335, and 82,412 individuals underwent non-enhanced, linear GBCA-enhanced, and macrocyclic GBCA-enhanced MRI, respectively. Exposure to linear GBCAs slightly increased the risk of all-cause parkinsonism (adjusted HR, 1.13 [97.5% confidence interval, 1.08–1.19]), while exposure to macrocyclic GBCAs did not increase the risk (adjusted HR, 1.00 [97.5% confidence interval, 0.95–1.05]).The results were similar for all-cause parkinsonism requiring medication, PD, and secondary parkinsonism, whereas no significant association was observed for atypical parkinsonism. Conclusion Exposure to linear GBCAs may slightly increase the risk of parkinsonism in adults, whereas exposure to macrocyclic GBCAs may not. Caution should be exercised when using linear GBCAs until further evidence emerges.

Objective: Evidence remains limited regarding whether Response Evaluation Criteria in Solid Tumors version 1.1 (RECIST 1.1) or modified RECIST (mRECIST) more reliably assesses treatment response in patients with hepatocellular carcinoma (HCC) receiving atezolizumab plus bevacizumab (Atezo/Bev). This study aimed to evaluate response patterns based on RECIST 1.1 and mRECIST, analyze inter-reader agreement, and assess their prognostic value for overall survival (OS) in patients with HCC receiving first-line Atezo/Bev. Materials and Methods: This retrospective study included patients with HCC treated with first-line Atezo/Bev between June 2020 and December 2022 at a tertiary center. Patients with at least one hypervascular hepatic target lesion were eligible. Two radiologists independently assessed treatment responses using RECIST 1.1 and mRECIST. Inter-reader agreement was evaluated using Cohen’s kappa coefficient. Time-dependent Cox regression analysis was performed, with radiologic response and progression treated as time-varying covariates. Prognostic discrimination was evaluated using Harrell’s concordance index (C-index). Results: A total of 207 patients were included (171 men; median age, 63 years; median follow-up, 10.7 months [range, 0.8– 46.4 months]; median OS, 10.7 months [95% confidence interval, 9.2–12.8 months]). mRECIST identified more responders than RECIST 1.1 (54.6% vs. 16.9%). RECIST 1.1 demonstrated excellent inter-reader agreement, whereas mRECIST showed substantial agreement (weighted kappa, 0.89 vs. 0.79). A significantly higher rate of dissociated responses was observed with mRECIST than with RECIST 1.1 (14.0% vs. 4.3%, P < 0.001). Both RECIST 1.1- and mRECIST-based responses and progression were independently associated with OS. Models incorporating RECIST 1.1 demonstrated slightly higher C-index values than those incorporating mRECIST (RECIST 1.1: 0.68 for response and 0.75 for progression; mRECIST: 0.65 and 0.70, respectively). Conclusion RECIST 1.1 is more reproducible and prognostically valuable for guiding treatment decisions in patients with HCC receiving first-line Atezo/Bev. However, this does not invalidate the use of mRECIST as a biological tumor response marker.

Background: We developed a simplified variant of the ADV score, the AFP-PIVKAII (AP) score for post-transplant hepatocellular carcinoma (HCC) prognosis, which considers only AFP and PIVKA-II levels excluding morphometric tumor size information from the ADV score. This study investigated the prognostic performance of the AP score in predicting HCC recurrence and overall survival (OS) after living donor liver transplantation (LDLT). Methods: We analyzed 843 patients with HCC who underwent LDLT between 2006 and 2015, assessing HCC recurrence and OS in relation to AP score. Results: The median pretransplant AFP and PIVKA-II levels were 12.8 ng/mL and 27 mAU/mL, respectively. The median and mean AP scores were 2.6 log (range: 0.6–9.2 log) and 2.9±1.1 log, respectively. The 5-year time-dependent area under the receiver operating characteristic curve for the AP score in predicting post-transplant HCC recurrence was 0.672 (p<0.001). HCC recurrence and OS curves along AP score intervals of 1.0 log showed statistical differences in accordance with the AP scores (both p<0.001). Using a Youden index J-derived AP score cutoff of 4.0 log, two-tiered groups (ADV <4.0 log vs. ADV ≥4.0 log) showed statistically significant differences in HCC recurrence and OS (both p<0.001). Harrell’s c-indices for AP score with cutoff of 4.0 log and ADV scores with cutoff of 5.0 log regarding HCC recurrence and OS were similar. Conclusion The AP score functions as an integrated surrogate marker for predicting post-transplant outcomes in patients with HCC undergoing LDLT. It may serve as a simplified alternative to the ADV score, particularly in patients with small HCCs.

Purpose: Due to its rarity, treatment guidelines for appendiceal cancer have traditionally followed those established for colorectal cancer, despite showing distinct histologic and clinical features. This study aimed to compare the clinicopathologic characteristics and long-term oncologic outcomes of appendiceal cancer with those of right-sided colon cancers. Methods: We retrospectively reviewed the records of patients with stage I–III appendiceal, cecal, or ascending colon cancer who underwent curative resection between 2010 and 2020 at our center. A 1:3:3 exact matching for age, sex, TNM stage, and adjuvant chemotherapy was performed. Survival outcomes were analyzed using the Kaplan-Meier and Cox regression methods. Results: Overall, 245 patients with appendiceal cancer (n = 35), ascending colon cancer (n = 105), and cecal cancer (n = 105) were analyzed. Appendiceal cancer exhibited a higher proportion of T4 tumors and fewer harvested lymph nodes compared with ascending or cecal cancers. The mean follow-up duration was 9.5 years. The 5- and 10-year overall survival rates were lower in appendiceal cancer (66.2% and 52.9%) than in ascending (91.2% and 78.4%) or cecal cancer (88.5% and 78.3%). Similarly, the 10-year disease-free survival rate was lower in appendiceal cancer (59.2%) compared with ascending (83.1%) and cecal cancers (78.4%). Cox regression analysis identified age (≥65 years), perforation, nodal metastasis, and lymphovascular invasion as independent predictors of poor prognosis. Conclusion Appendiceal cancer exhibited significantly worse long-term survival compared to cecal or ascending colon cancer. Tumor perforation, nodal metastasis, and lymphovascular invasion were adverse prognostic factors for overall and disease-free survival.

Purpose: Neoadjuvant chemotherapy (NAC) significantly revolutionized the management of locally advanced breast cancer, especially in human epidermal growth factor receptor 2 (HER2)-positive and triple-negative subtypes. However, its effectiveness is limited in estrogen receptor (ER)-positive, HER2-negative breast cancer. This study investigates the clinical response and prognosis of ER-positive, HER2-negative breast cancer after NAC. Methods: The clinicopathological characteristics and treatment responses of 149 patients with ER-positive, HER2-negative breast cancer treated with NAC and surgery at The Catholic University of Korea, Seoul St. Mary’s Hospital between 2018 and 2023 were retrospectively analyzed. Pathologic complete response (pCR) was defined as the absence of invasive tumors (ypT0/is, ypN0). Disease-free survival (DFS) and overall survival (OS) were analyzed using Kaplan-Meier methods, stratified by age (≤50 years vs. >50 years). Results: Among 149 patients, 13 (8.7%) achieved pCR, 87 (58.4%) attained partial responses, 40 (26.8%) had stable disease, and 9 (6.0%) experienced progressive disease. RECIST responses differed significantly by age (P = 0.003). DFS (P = 0.011) and OS (P = 0.005) were significantly associated with clinical response in patients aged ≤50 years. Post-NAC Ki-67 was associated with DFS (P = 0.013) but not OS (P = 0.083) in patients aged ≤50 years. Clinical responses and post-NAC Ki-67 were not associated with DFS (P = 0.544) or OS (P = 0.569) in patients aged >50 years. Conclusion In ER-positive, HER2-negative breast cancer, clinical responses and post-NAC Ki-67 were significant prognostic factors in patients aged ≤50 years but not in older patients. These findings highlight the need for tailored therapeutic approaches that consider age-specific prognostic differences.

Lumbar spinal stenosis (LSS) is a common degenerative spinal condition where spinal canal narrowing causes symptoms such as neurogenic claudication, radiculopathy, and lower back pain. While non-operative and surgical approaches yield similar long-term outcomes, surgical intervention—particularly decompression—can provide earlier symptom relief, functional recovery, and fall prevention in selected patients with refractory symptoms. Recent advancements in surgical technologies and image guidance have brought about a paradigm shift in LSS management. Biportal endoscopic spine surgery (BESS) has gained global traction as a minimally invasive alternative to traditional decompression methods, offering superior visualization, less soft tissue damage, shorter hospital stays, and faster recovery. High-quality studies, including randomized controlled trials, have shown promising outcomes for this technique. Furthermore, the integration of navigation systems, robot-assisted instrumentation, and artificial intelligence (AI)-driven diagnostics and surgical planning tools is transforming spinal surgery by enhancing precision in preoperative evaluation and intraoperative execution. These innovations enable accurate targeting, reduce complications, and improve reproducibility across diverse surgical settings. This review provides an updated overview of LSS, covering its pathophysiology, clinical assessment, diagnosis, and treatment. Special emphasis is placed on the growing role of BESS and the transformative impact of digital technologies such as navigation, robotics, and AI in the evolving landscape of spinal stenosis care.

Conventional aryl hydrocarbon receptor (AhR) antagonists, which play a critical role in modulating tumor immune evasion, have shown limited clinical translation due to poor solubility, restricted systemic exposure, and dose-limiting toxicities. To overcome these limitations, we developed SB5794, a phosphate prodrug of the potent AhR antagonist SB2617, designed to improve aqueous solubility and pharmacokinetic properties. SB5794 exhibited markedly enhanced solubility and achieved more than six-fold higher systemic exposure in mice compared with SB2617, while fully retaining its in vitro AhR antagonistic activity. In syngeneic tumor models, SB5794 significantly inhibited tumor growth, and its combination with anti–PD-1 therapy further enhanced antitumor efficacy. However, repeated-dose studies revealed dose-dependent histopathological changes in the gastrointestinal tract, liver, and immune organs. Collectively, these findings demonstrate that SB5794 possesses improved drug-like properties and strong immunomodulatory activity, supporting its potential as a next-generation AhR-targeted immunotherapeutic candidate.