The aim of this study was to critically appraise the performance of detection methods for non-cavitated carious lesions (NCCLs). A detailed search of Medline (via OVID), the Cochrane Collaboration, Scielo and EMBASE identified 2054 publications. After title and abstract review by three investigators (JG, MT, AI), 124 publications were selected for further review. The final publications evaluated the following methods: Visual (V), Caries Lesion Activity Assessment (CLAA), Laser Fluorescence (LF), Radiographic (R), Fibre-optic Transillumination (FOTI), Electrical Conductance (EC) and Quantitative Light-induced Fluorescence (QLF). All included studies used histological assessment as a gold standard for in vitro studies or clinical/visual validation for the in vivo designs. They reported outcomes measures such as sensitivity (SE), specificity (SP), area under the receiver operating characteristic curve (AUROC) and reliability. Data were extracted from the selected studies independently by two reviewers and checked for errors. The quality of the studies was evaluated as described by Bader et al. (2002). Of the 124 articles, 42 were included that described 85 clinical assessments. Overall, the quality of evidence on detection methods was rated ‘poor’, except for EC that was rated ‘fair’. The SE rates were as follows: V (0.17–0.96), LF or DIAGNOdent (DD) (0.16–0.96), R (0.12–0.84), FOTI (0.21–0.96), EC (0.61–0.92) and QLF (0.82). The SP rates were as follows: V (0.46–1.0), LF (0.25–1.00), R (0.55–0.99), FOTI (0.74-0.88), EC (0.73–1.0) and QLF (0.92). There is a large variation in SE and SP values for methods and a lack of consistency in definition of disease and analytical methods. EC and QLF seem to be promising for detection of early lesions. For both cost and practicality considerations, visual methods should remain the standard for clinical assessment in dental practice.