Aim: This systematic review and meta-analysis aimed to assess the diagnostic accuracy and reliability of commonly used caries detection methods for proximal caries diagnostics. Visual examination (VE), bitewing radiography (BWR), laser fluorescence (LF), and fibre-optic transillumination (FOTI) were considered in detail. Material and methods: PRISMA guidelines for the reporting of systematic reviews and meta-analyses were applied. The mnemonic PIRDS (problem, index test, reference test, diagnostic and study type) concept was used to guide the literature search. Next, studies that met the inclusion criteria were stepwise selected and evaluated for their quality with a risk of bias (RoB) assessment tool. Studies with low/moderate bias and sufficient reporting were considered for meta-analysis. The pooled sensitivity (SE), specificity (SP), diagnostic odds ratio (DOR), and area under the ROC curve (AUC) were calculated. Results: From 129 studies meeting the selection criteria, 31 in vitro studies and five clinical studies were finally included in the meta-analysis. The AUC values for in vitro VE amounted to 0.84 (caries detection) and 0.85 (dentin caries detection). BWR ranged in vitro from 0.55 to 0.82 (caries detection) and 0.81–0.92 (dentin caries detection). LF showed higher AUC values for overall caries detection (0.91) and dentin caries detection (0.83) than did other methods. Clinical data are limited. Conclusion: The number of diagnostic studies with low/moderate RoB was found to be low and indicates a need for high-quality, well-designed caries diagnostic studies. Clinical relevance: BWR and LF showed good diagnostic performance on proximal surfaces. However, because of the low number of includable clinical studies, these data should be interpreted with caution.