*Limitations*

The present review has several limitations. First, only three studies addressed thyroid cancer exclusively, and findings in the other studies were not separated according to the indication for surgery. Hence, there are insufficient data to draw any conclusions regarding the impact of fluorescence imaging on the outcome of thyroid cancer surgery in particular. Further studies should include a more detailed investigation of the observations of a potentially increased benefit of NIR imaging for extended oncological surgery, including neck dissection. Second, the majority of the available and included studies were of limited quality, with only three RCTs (randomized controlled trials). The retrospective nature of some studies [48,49,64] hinders correction for confounders between the study and control groups. However, this aspect could enhance the comparability of both groups by allowing the selection of similar controls for each case. Third, the majority of included studies were small studies with a lack of prospective sample size calculation, which may have affected the observed outcomes.

Another limitation is the variability in NIR imaging use among the reports. In most studies, not all four PGs were visualized. Therefore, the vascularization and viability of the unidentified glands remained unknown, which may have affected the outcome of the surgery. Furthermore, the subjective nature of estimating the fluorescence level in a gland is a limitation, as the available systems only allow qualitative assessments. This limitation leads to subjective interpretation and interobserver differences in scoring the fluorescence intensity of a gland. To eliminate this limitation, future studies should integrate computer-assisted quantitative contrast-enhanced NIR imaging evaluations [12] or pixel/color-analyzing computer programs, as suggested by Fanaropoulou et al. [19].
