https://jovm.smums.ac.ir/ Journal of Visualized Medicine en daily 1 Sat, 22 Nov 2025 00:00:00 +0330 Sat, 22 Nov 2025 00:00:00 +0330 https://jovm.smums.ac.ir/article_58.html This study uses finite element method (FEM) analysis to assess the impact of marginal bone loss (MBL) and bone quality (BQ) on temporary anchorage device (TAD) screw stability, placed between mandibular teeth 4 and 5 per orthodontic guidelines. A 2 N orthodontic force was applied, with micromotion as the stability metric. Five BQ groups were modeled—very strong, strong, normal, weak, and very weak —with Young’s modulus varied by ±15% per group and Poisson’s ratio fixed at 0.3. Five MBL groups were simulated with cortical bone thickness at 1, 1.5, 2, 2.5, and 3 mm. Simulations revealed that reducing cortical thickness from 3 mm to 1 mm increased micromotion by 40% (from 8 µm to 11.2 µm). Similarly, decreasing Young’s modulus from 19.55 GPa (very strong) to 10.2 GPa (very weak) elevated micromotion by 32% (from 8 µm to 10.56 µm) under identical loads. These findings highlight cortical thickness and BQ as key predictors of TAD stability, guiding orthodontic planning. High-resolution imaging is recommended to optimize TAD placement and mitigate MBL-related complications. This FEM framework elucidates mandibular biomechanical interactions. https://jovm.smums.ac.ir/article_60.html Manual interpretation of echocardiography is a labor-intensive process characterized by significant inter-observer variability. Although Deep Learning (DL) has shown expert-level potential, its clinical integration is often hindered by "domain shift" across different ultrasound vendors and a lack of multi-view analysis capabilities. To address these challenges, we developed EchoAI, a secure, vendor-agnostic web-based Clinical Decision Support System (CDSS) that incorporates a multi-task Unsupervised Domain Adaptation (UDA) engine. This platform enables the simultaneous and automated quantification of Left Ventricular Ejection Fraction (LVEF) and cardiac wall thickness across standard A4C, A2C, and PLAX views. EchoAI utilizes a user-centered design and a "human-in-the-loop" workflow, allowing physicians to verify and edit AI-generated segmentation masks in real-time, thereby fostering diagnostic trust. Clinical validation conducted at Guilan University of Medical Sciences (GUMS) demonstrated high operational efficiency with an average processing time of 1.15 seconds per cardiac cycle. The system achieved a strong correlation with expert manual measurements (r=0.95, P<0.001) and an exceptionally low mean bias of -0.17%. Usability assessments yielded a high satisfaction score of 6.18 out of 7, with 86% of the AI outputs being accepted without modification. By providing an interactive and transparent interface, EchoAI effectively bridges the gap between algorithmic potential and routine clinical practice, offering a scalable solution for enhanced cardiac assessment in diverse healthcare settings.