Objective <p>The purpose of this study was to compare the effectiveness of low and high magnifications in detecting occlusal caries in permanent posterior teeth using the International Caries Detection and Assessment System (ICDAS) II criteria through the ICDAS e-learning program.</p> Materials and methods <p>Forty extracted permanent posterior teeth were used. Two examiners received e-learning training on the ICDAS webpage for detecting occlusal caries before visual examinations. Visual inspections were performed under ×2.5 magnification with a magnifying loupe and ×25 magnification with a dental surgical microscope within a week. The ICDAS scores assigned by both examiners were validated by an ICDAS expert. The visual assessment results were compared with the histological analysis results using Downer’s criteria as the gold standard.</p> Results <p>Inter-examiner reliability for ICDAS-II scoring at low magnification (×2.5 loupe) and high magnification (×25 operating microscope) was moderate. However, the lenient (binary) criteria showed substantial agreement at both magnifications. Specificity was highest for both the low and high treatment thresholds (88.2% and 81.5%, respectively) at low magnification, whereas sensitivity peaked at high magnification for the high treatment threshold criterion (91.7%). Image-based ICDAS scoring by an expert also demonstrated good diagnostic accuracy (76.9%), though it was not superior to ICDAS scoring performed under ×2.5 magnification (82.1%).</p> Conclusion <p>The ICDAS-II score, particularly the lenient criterion, demonstrated a strong correlation with histological depth. The e-learning program effectively equips dentists with diagnostic skills. Extreme magnification resulted in the overestimation of dental caries, whereas low magnification (2.5x) resulted in greater diagnostic accuracy.</p>