The aim of this exploratory double-blinded, randomized, cross-over, in situ study was to compare the effects of various model parameters (‘intervention’, ‘brushing’, ‘position’) on enamel caries lesions in a dose-response model.In each of four experimental legs of four weeks 16 participants wore intraoral mandibular appliances with four ‘plaque-retaining’ and four ‘easily cleanable’ positioned pre-demineralized bovine enamel specimens in the vestibular flanges mimicking proximal and buccal surfaces, respectively (n=512). The four randomly allocated interventions (either application only or brushing) included the following dentifrices: AlF3 1360ppmF−+chlorhexidine 0.05% (Lacalut aktiv, LA1360), NaF 1,450ppmF-(Blend-a-Med ProExpert), NaF 500ppm F− and 0ppm F− as negative control (NC) (both experimental, based on Blend-a-Med ProExpert).Differences in integrated mineral loss (ΔΔZ) and lesion depth (ΔLD) were calculated between values before and after the in situ period using transversal microradiography. Significant differences for ΔΔZ [adjusted mean (95% CI))] were found between NC, NaF500 and LA1360for both ‘plaque-retaining’ [−1830 (−2371;1289); −986 (−1530;442); −2 (−548;544)vol%×μm] as well as ‘easily cleanable’ specimens [−399 (−682; −116); −391 (−672; −110); −16 (−302;270)vol%×μm]. Values for NaF1450 revealed a similar dose-response as LA1360.Values for LA1360 and NaF1450 did not differ significantly (p>0.05; ANCOVA).The design of the present in situ study was able to reveal a fluoride dose-response to hamper further demineralization of enamel specimens for ‘easily cleanable’ and ‘plaque-retaining’ sites being brushed or not. Particularly ‘plaque-retaining’ sites seem to be recommendable for measuring potential anticaries efficacy in situ.