The aims of this study were: (1) to evaluate the overall reproducibility and accuracy of the International Caries Detection and Assessment System (ICDAS) for assessing coronal caries lesions, and (2) to investigate the use of systems associated with the ICDAS for activity assessment of coronal caries lesions. Specific search strategies were adopted to identify studies published up to 2016. For the first objective, we selected studies that assessed primary coronal caries lesions using the ICDAS as a reference standard. A total of 54 studies were included. Meta-analyses summarized the results concerning reproducibility and accuracy (correlation with histology, summary ROC curves [SROC], and diagnostic odds ratio [DOR]). The latter 2 were expressed at D1/D3 levels. The heterogeneity of the studies was also assessed. Reproducibility values (pooled) were > 0.65. The ICDAS mostly presented a good overall performance as most areas under SROC were > 0.75 at D1 and > 0.90 at D3; DOR ≥6. For the second objective, we selected studies investigating activity assessment associated with the ICDAS. The meta-analyses pooled results based on the same methodology and parameters as above. Longitudinal findings regarding caries progression were described to estimate the validity of these systems. On average, the systems for activity assessment of caries lesions showed moderate values concerning reproducibility and overall performance. Active caries lesions were more prone to progress than inactive ones after 2 years. In conclusion, the ICDAS presented a substantial level of reproducibility and accuracy for assessing primary coronal caries lesions. Additional systems associated with the ICDAS that classify caries lesion activity can be useful as they are moderately reproducible and accurate.