Full records of enrolled lesions included:

Lesions within the same subject were randomised using computer random-sampling digit tables.

Treatment Group: Teeth received Ozone treatment for 10 seconds (HealOzone unit (CurOzone and KaVo, USA and Germany) (2100 ppm O3, 615 ml/min)) and then sprayed by a remineralising mouthwash (Table 6).
Control Group: Teeth sprayed by the remineralising mouthwash. Subjects received preventive advice, a toothbrush (Brilliant, Brilliant Products, UK) and a toothpaste (Natural White, Natural White Inc., UK; 1,100 ppm F-).

Repeatability of the clinical severity, DIAGNOdent and ECM scores were tested at baseline using Weighed Kappa values. Score changes at the recalls were tested using Wilcoxon’s Signed Ranks Test for the pair (recall visit score, baseline score) and between groups for the pair (recall visit score minus baseline score for the treatment group, recall visit score minus baseline score for the control group). For the raw DIAGNOdent and ECM readings, change from baseline was tested using the Student’s t-test (Loge (baseline value / recall value)). Tests between groups were performed using univariate analysis of variance for Loge ECM and DIAGNOdent readings. Fixed factors included subjects' ID (to exclude subject variability) and lesion treatment group were used. The subject ID was a significant factor determining the effect of treatment, and thus proving such variability (p<0.05).

End value readings, displayed by the ECM, were chosen for various calculations 14. For each tooth, 4 end-value readings were taken and the average was used to represent the lesion. This average reflected the resistance of the tooth to the applied electric current in standardised clinical conditions. The tooth resistance is expected to be reduced if the mineral content was reduced by the demineralisation activity of cariogenic microorganisms that continues to increase the porosity of decayed lesion. Ion rich saliva fill these porosities and reduce this resistance, thus ECM end-value is the only measured outcome that has an inverse relation to caries severity and activity (Table 7).

258 Non-cavitated POPFC lesions were included in this study (Table 8).

Repeatability tests were obtained in-vivo from 32 occlusal pit and fissure surfaces in two visits, one week apart. Weighed Kappa values were 0.74 and 0.62 for the clinical severity and DIAGNOdent scores respectively. For the 4 end-value ECM readings taken from the same lesion, repeatability is defined as that amount, which two measurements could be expected to exceed on just 5% of occasions as dictated in the following equation: Repeatability = ± 1.96 SDev.

The range in which 95% of the repeated end-value measurements from fissure lesions within one visit occurred was 4.8 times, and 4.3 times for the pit lesions. When the raw end-value readings were transformed to equivalent ECM scores used in this study, mean weighed Kappa value for the four sets of readings were 0.64 and 0.62 for fissure and pit lesions respectively (p<0.001) (Table 9).

At baseline, and although lesions were randomised into treatment and control groups, there was a tendency for severe lesions to be within the treatment group (p=0.055, Wilcoxon’s Signed Ranks Test). However, this tendency was not supported by the other two diagnostic tools, which showed no severity difference between the two groups (p = 0.240 and 0.130, Wilcoxon’s Signed Ranks Test for baseline DIAGNOdent and ECM scores, respectively) (Table 10).

ECM score 0 was interpreted clinically as a tooth with mineral contents comparable with healthy tooth structure (ECM manual, 2000). This ECM score was obtained from lesions diagnosed by the clinical severity score and DIAGNOdent readings as being carious. This may indicate that previous activity of these lesions might have been halted or reversed previous to the baseline diagnosis session. This may demonstrate that the DIAGNOdent device and the clinical severity score system may detect severity, size and depth of such activity but would be limited in the detection of mineral change accompanying fluctuating carious activity. Nevertheless, correlations of baseline scores obtained by the three diagnostic methods were significant (p<0.05) (Table 11).

Changes in the three diagnostic tools were tested in the one-month recall visit (± SE) (33.6 ± 0.48 days). 4 subjects missed visit (12 teeth, 4.7% of enrolled lesions) as one loss of contact with a subject who anticipated a change in her permanent address and unavailability for examination by the other subjects.

ECM readings and score change analysed showed a significant lesion mineral and thus lesion activity difference between groups (p<0.05). The control group had a significant drop in its ECM readings and thus a significant increase in ECM scores (p<0.05). This difference was enhanced with the treatment group having some increase in ECM readings and a decrease in ECM scores from baseline but that was insignificant (p>0.05) (Table 12).

If tests were repeated between treatment and control groups, after lesions with ECM baseline score 0 were removed, then only 62 pairs of lesions meet this condition. In this subcategory, the treatment group had a mean change that was significantly higher (better) than baseline (p<0.05), whilst the control group had no significant change (p>0.05) (Table 13).

The second subcategory (17-pairs) which had baseline ECM score 0 in both treatment and control groups did not experience ECM changes (p>0.05). The third subcategory (38 pairs) had either the treatment or control lesion with a baseline score 0 and was not analysed for changes.

ECM changes were analysed by co-factors. Those that showed no significant effect on the ECM change included: location of the lesion on the occlusal surface, type of lesion (pit or fissure) and baseline DIAGNOdent and clinical severity scores (p>0.05). The following co-factors had a significant effect; baseline ECM score and tooth number (Figures 1 and 2).

Fine demineralisation and remineralisation difference between treatment and control groups detected by the ECM seemed not to be detected by clinical severity scores change (Z= -1.005, p= 0.315). DIAGNOdent reading changes were less obvious and both groups had a tendency for readings being higher from baseline (p= 0.06,0.086 for the treatment, control group respectively, difference between groups p>0.05).

When validated diagnostic tools are used for monitoring lesion activity, their ability to discriminate between positive and desirable changes accompanying remineralisation, from those parallel, but clinically insignificant changes (such as staining) should be anticipated. Interference from these parallel changes may be inflicted on the tool. Another limitation would be the sensitivity of tool to fine caries activity changes after a short period beyond the range of repeatability and reproducibility errors. Finally, once occurred, the change measured should be perpetual. Using more than one diagnostic tool would help to overcome each tools limitations by quantifying lesions activity using more than one measured feature, as mineral content, microbial by-products and physical changes in active lesions.

Ekstrand's clinical severity score system depends on external manifestations to predict the extent of the tooth decay. Being one of the most validated clinical systems for a single diagnosis, it was of high value to detect lesions as they present to the clinics for the first time. Clinical scores correlated with scores using the DIAGNOdent and ECM at baseline. A similar clinical score system had been used to monitor lesions and a significant change was detected in only after 2 years of extensive oral hygiene measures with fluorides 15. No literature was present to monitor changes in teeth subject to Ozone treatment.

Although DIAGNOdent was recommended for lesion monitoring, no published data monitored occlusal lesions in-vivo. Continuous deposition of proteins (porphyrins) at various amounts and depths within the demineralised tissue was reflected by the increase in lesion fluorescence. Theory is that remineralisation might shuffle the protein distribution by continuous washing of saliva and less protein production by biofilms, then DIAGNOdent readings may significantly decline. Stains or calculus may be part of the change occurring in lesions with slow progression and repeated cycles of demineralisation and remineralisation. DIAGNOdent may pick these parallel changes as a higher reading and might not reflect the reduced protein production as previously proposed. A total significant reduction of DIAGNOdent readings was reported when root lesions were treated with Ozone 12 but these results could not be equated to occlusal caries lesions, as the structures involved in the attack are different as well as the environment to which lesions are subjected, including the microbial environment and composition 12.

ECM is known to be highly sensitive to the mineral content of the tooth. Natural changes, as enamel maturation, were monitored by more than one study 16 17 18. Approximate maturation periods were suggested. Minor differences in the mineral content of the sound mature surfaces, as fine cracks and hypomineralised areas, were also detected using the ECM 19. Furthermore, ECM readings at baseline were suggested as a predictor of future lesion behaviour by calculating survival rates 20.

Lesions incorporated in the study covered the spectrum of lesions’ presentations without frank cavitation. Using inclusion criteria, lesions were enrolled in all age groups within both genders. This might be a reflection that age was no longer a factor deterring any caries attack 21.

Lesions were also distributed between all posterior teeth as well as between various types of surface topographies (pits and fissures) where teeth were in the mouth of subjects for more than a year and had full eruption status. However, this didn’t mean that all teeth were in full occlusion.

Each lesion within the treatment group had a control within the same subject. This served to reduce the number of various factors, which might affect different subjects to various degrees, such as fermentable carbohydrate intake, oral hygiene habits, possible age related factors, salivary bacterial counts, social and racial variables as well as gender. Such variables may explain the mineralisation variability detected when the subject ID was used as fixed factor for group analysis and accounted for by adjusting for any block effect.

A significant mineral activity difference occurred between lesions treated with ozone in comparison with the control group lesions. Although further recall visits would be needed to monitor and follow up this trend, it could be claimed that the mechanism for these early changes is different. As there was no significant interfering factor in the control group, then it might be natural to propose that the change was a steady deterioration of the health status of the lesions if left untreated. However, the changes in the treatment group could only be explained by the effect of Ozone treatment. This should be followed up further and assess the clinical presentations and DIAGNOdent reading accompanying such remineralisation shift.

Further analysis of subgroups divided by ECM baseline score showed that demineralised lesions within the treatment group (Scores 1-3) had significant remineralisation, while this mineral change was not parallel in treatment group lesions that most probably had previous remineralisation before baseline visit (ECM score 0). This would support that the effect of Ozone was directed to demineralised lesions while already remineralised lesions, Ozone might help to stabilise it. The control group lesions that had a baseline ECM score 0 had no significant mineral change which may prove the previous conclusion that remineralised lesions may have a more stable mineral content that could be more resistant to further demineralisation from natural tooth minerals due to the accumulation of stronger mineral enamel crystals 22. Further recalls would be needed to follow up this conclusion.

Ozone treatment of primary occlusal pit and fissure caries produced an early remineralisation shift in demineralised lesions after one month while demineralisation of control lesions continued regardless of lesion type and location within the occlusal surface or tooth sequence. Additional monitoring is planned for this study.