In vitro performance of methods of approximal caries detection in primary molars

Mariana Minatel Braga, DDS, PhD,a Caroline Carvalho Morais, DDS,a Renata Cristina Satiko Nakama, DDS,a Victor Moreira Leamari, DDS,a Walter Luiz Siqueira, DDS, PhD,b and Fausto Medeiros Mendes, DDS, MS, PhD,a So Paulo, Brazil; and London, Canada
UNIVERSIDADE DE SO PAULO AND UNIVERSITY OF WESTERN ONTARIO

Objective. The aim was to compare the performance of different methods in detecting approximal caries lesions primary molars ex vivo. Study design. One hundred thirty-one approximal surfaces were examined by 2 observers with visual inspection (VI) using the International Caries Detection and Assessment System, radiographic interpretation, and clinically using the Diagnodent pen (LFpen). To achieve a reference standard, surfaces were directly examined for the presence of white spots or cavitations, and lesion depth was determined after sectioning. The area under the receiver operating characteristic curve (Az), sensitivity, specicity, and accuracy were calculated, as well as the interexaminer reproducibility. Results. Using the cavitation threshold, all methods presented similar sensitivities. Higher Az values were achieved with VI at white spot threshold, and VI and LFpen had higher Az values at cavitation threshold. VI presented higher accuracy and Az than radiographic and LFpen at both enamel and dentin depth thresholds. Higher reliability values were achieved with VI. Conclusions. VI performs better, but both radiographic and LFpen methods also show good performance in detecting more advanced approximal caries lesions. (Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009;108:e35-e41)

Researchers and clinicians alike are still searching for an efcient, cost-effective, and quantitative method for approximal-surface caries detection that provides high validity and reliability. Although the clinical visual inspection (VI) method demonstrates high specicity, its sensitivity and reproducibility are inferior to other methods.1 In addition, it is not a quantitative method. The radiographic method also is not quantitative, and it has the disadvantage of exposing the patient to the hazard of ionizing radiation.1,2 The low reliability observed in these methods is related to their dependence on clinical interpretation. To improve the validity and reliability of VI, a new visual index for caries diagnosis, the International Caries Detection and Assessment
Supported by the Conselho Nacional de Desenvolvimento Cientco e Tecnolgico (CNPq; process nos. 476372/2006-2 and 471952/ 2008-7), Pr-Reitoria de Pesquisa da USP, Canadian Institutes of Health Research (CIHR), and the Natural Sciences and Engineering Research Council of Canada (NSERC). a Department of Pediatric Dentistry, Faculdade de Odontologia da Universidade de So Paulo. b Schulich Dentistry and Department of Biochemistry, Schulich School of Medicine and Dentistry, University of Western Ontario. Received for publication Mar 17, 2009; returned for revision May 25, 2009; accepted for publication May 26, 2009. 1079-2104/$ - see front matter 2009 Published by Mosby, Inc. doi:10.1016/j.tripleo.2009.05.017

System (ICDAS) has been developed.3 However, this new visual scoring system has not been validated in detecting approximal caries lesions. Quantitative methods could improve reliability, because they would give a metric and the dentist would interpret this value using a predetermined cut-off point scale independent from clinician opinion.4 One quantitative method for occlusal and approximal caries detection is the newly introduced pen-type laser uorescence device (LFpen).5-7 The physical principle of this device is based on detection of uorescence emitted from the organic content of the caries lesions after excitation with a red light emitted from a diode laser. Lussi et al.5 found the performance of the LFpen device to be better than radiography in detecting approximal caries lesions in permanent teeth; however, Novaes et al.8 showed that the device had similar performance to the radiographic method in primary molars. However, in later study, the reference standard was the surface status after temporary separation using orthodontic rubber rings for 7 days, and therefore, the methods were not validated according to the lesion depth. An in vitro study could use this type of validation and investigate the relationship of lesion depth and surface condition. The aim of the present ex vivo study was to compare the performance of VI using ICDAS, radiographic, and LFpen techniques in detecting approximal caries lee35

e36

Braga et al.

OOOOE October 2009

sions in extracted primary molars. We also investigated the relationship between lesion depth and surface status. MATERIALS AND METHODS Sample preparation After obtaining approval from the local Committee for Ethics in Research, 84 primary molars from children in So Paulo, Brazil (0.7 mg/L uoride in water supply), which were recently exfoliated or extracted for orthodontic purposes, were available. Before extraction, informed consent was obtained from the childrens parents or guardians. After extraction, samples were positioned in plastic containers and frozen at 20C until their use. A wet cotton roll was placed at the bottom of each container to guarantee 100% humidity, with no contact between the tooth and the wet cotton roll. Before starting the experiment, the selected primary molars were defrosted at room temperature for 4 hours and then cleaned with a rotating brush and pumice/ water slurry. This procedure has not signicantly changed the uorescence in experiments performed with LF methods on extracted teeth.9 To simulate the contact points, the teeth were placed in arch models in the following sequence: a primary canine, a rst primary molar, a second primary molar, and a rst permanent molar. Care was taken to simulate the contact points as best as possible, and the presence of contact points was conrmed using dental oss. The evaluations were performed only on the approximal surfaces of primary molars. The surfaces of canines and rst permanent molars were not evaluated. Exclusion criteria for sample surfaces were presence of approximal restorations, presence of hypoplastic pits, frank approximal cavitation (absence of marginal ridge), presence of large carious lesions on smooth or occlusal surfaces, and surfaces with difculty to simulate the contact point. The nal sample comprised 131 approximal surfaces. Examination methods Caries lesions in the selected approximal surfaces were assessed by 3 methods: VI, radiographic assessment, and LFpen. Two examiners carried out all examinations (R.C.S.N. and C.C.M.). One benchmark examiner (M.M.B.) trained the others using 15 surfaces, but no calibration procedure was performed. These teeth were not included in the sample. The examiners were orientated to analyze each site independently, and they were unaware of each others results. Before the examinations, teeth were defrosted as described above. After a series of examinations, samples were frozen at 20C again.

For the VI method, specimens were positioned about 30 cm from the examiners eyes, with no magnication and with the aid of a light reector. The examiners used a mouth mirror and a ballpoint probe. The teeth were rst examined wet, and then they were dried for 5 seconds with compressed air. The examiners used the ICDAS-II method to perform VI.3 The scores were: 0: 1: 2: 3: 4: 5: 6: Sound tooth. First visual change in enamel. Distinct visual change in enamel. Localized enamel breakdown. Underlying dark shadow from dentin. Distinct cavity with visible dentin. Extensive distinct cavity with visible dentin.

For the radiographic method, bitewing radiographs were taken from each series of teeth, using bitewing holders (Han-Shin PF 682; Jon Ind., So Paulo, Brazil). The x-ray machine (Spectro 70 X; Dabi Atlante, Ribeiro Preto, Brazil) was set to 70 kV, 8 mA, and the exposure time was 0.3 second. Kodak Insight radiographic lms (22 35 mm, Eastman Kodak, Rochester, NY) were used, the focus-to-lm distance was 40 cm, and the lms were manually developed using standard processing times. The radiographs were examined on a backlit screen at 2 magnication. The criteria used to indicate enamel or dentine caries lesions was adapted from the criteria previously described:10 0: No radiolucency visible. 1: Radiolucency visible in enamel. 2: Radiolucency visible in dentin, but restricted to the outer one-third of dentin. 3: Radiolucency extending to the middle one-third of dentin. 4: Radiolucency in the inner one-third of dentin. The third method used was LFpen. This method used a Diagnodent pen device (Kavo, Biberach, Germany) attached to Probe tip 1 (for approximal surfaces). The laser device was rst calibrated against the porcelain reference object. On every tooth, the device was also calibrated on a sound smooth surface. This laser uorescence reading was electronically subtracted from the readings of the approximal site under examination. Teeth were dried by air for 5 seconds with the dental units built-in 3-in-1 spray syringe. The tip was introduced underneath the contact area, rst from the facial side and then from the oral side. The highest value from the 2 measurements of each surface was recorded. Reference standard methods After all examinations, 2 reference standard methods were applied: surface condition evaluated through di-

OOOOE Volume 108, Number 4

Braga et al. e37

rect VI after removal from the arch models; and lesion depth obtained by histopathologic analysis. Different examiners performed the reference standard methods in joint sessions until they reached a consensus. These examiners were unaware of the other examinations. First, the teeth were removed from the arch models, dried for 5 seconds with compressed air, and the surfaces directly examined by the examiners with the aid of a light reector and a ballpoint probe. The examiners classied the surfaces as: Sound: No change in enamel translucency after air drying and absence of surface discontinuity. White spot lesion: White or brown discoloration in wet or dried tooth with no enamel discontinuity. Cavitation: Loss of integrity of the surface visually detected or by ballpoint probe. Subsequently, teeth were embedded in resin blocks and serial sections cut (approximately 250 m thick) using a 0.3-mm-thick diamond saw mounted in a microtome (Labcut 1010; Extec Co., Eneld, CT, USA). All sections were examined by 2 examiners in a joint session using a stereomicroscope at 16-40 magnication and reected light (SZPT; Olympus, Tokyo, Japan). The sites were classied in a 5-point scale according to the lesion depth: D0: No caries. D1: Caries lesion limited to the outer one-half of the enamel. D2: Caries extending into inner half of the enamel but not to amelodentinal junction. D3: Caries limited to the outer one-half of the dentin. D4: Caries involving the inner one-half of the dentin. Statistical analysis The statistical unit in this study was the approximal surface. The correlation between surface status evaluated by direct VI and the lesion depth obtained by the histopathologic analysis (the reference standard methods) was carried out using a Spearman correlation test. Then the Spearman correlation coefcient (Rs) and 95% condence interval (CI) were calculated. The sensitivity, specicity, and accuracy values were calculated for each method for each reference standard method. Accuracy was dened as the percentage of correct diagnosis in all samples (sound and decayed surfaces). Data were analyzed separately for each examiner. First, we used the results obtained by the direct VI of the surface status. For this analysis, we considered 2 thresholds: presence of white spot and presence of cavitation. The cut-off points for VI were score 1 from ICDAS-II at white spot threshold and score 3 at cavitation threshold. For the radiographic method,

scores 1 and 2 were considered to be cut-off points at white spot and cavitation threshold, respectively. For the LFpen method, the best cut-off points at both thresholds were obtained when the highest-accuracy value was achieved. The cut-off points obtained in the present study for the LFpen method were: 0-4 sound; 4.1-38 white spot caries lesion; 38 cavitation. A receiver operating characteristic (ROC) analysis was also performed for each method at both thresholds, and the area under the ROC curve (Az) was calculated. We next used the lesion depth as reference standard method. Then all of the analysis described above was performed considering 2 thresholds: initial enamel caries lesions (D1) and initial dentin caries lesions (D3). The cut-off points for VI were scores 1 and 3 from ICDAS-II at D1 and D3 thresholds, respectively. For the radiographic method, scores 1 and 2 were the cutoff points at D1 and D3 thresholds. For the LFpen, the best cut-off points obtained by the highest accuracy values in the present study were: 0-8 sound teeth; 8.1-30 enamel caries lesions; 30 dentin caries lesions. The McNemar change test was applied to compare the sensitivity, specicity, and accuracy values of the different methods, and a nonparametric approach was used to compare the Az values. The interexaminer reproducibility was initially calculated considering all scores of VI and radiographic methods or values of LFpen readings using intraclass correlation coefcient (ICC) and 95% CI. Then, the interexaminer reliability was calculated using the Cohen kappa test (95% CI) after collapsing the results into 2 categories (sound vs. decayed) considering white spot, cavitation, D1, and D3 thresholds. All analyses were carried out using statistical software (MedCalc 9.3.0.0; MedCalc, Mariakerke, Belgium), and the level of signicance was P .05. RESULTS After direct VI of approximal surfaces, we found 38 sound surfaces, 61 white spot lesions, and 32 surfaces with cavitations (Table I). With the histopathologic examination, 49 surfaces were sound (D0), 20 were initial enamel lesions (D1), 28 surfaces had advanced enamel caries lesions (D2), 24 surfaces had initial dentine caries lesions (D3), and 10 had advanced dentin caries lesions (D4) (Table I). There was a signicant correlation between surface condition and lesion depth (Rs 0.725; 95% CI 0.632-0.797; P .0001). Most surfaces without changes (sound surfaces) presented no caries lesions after histopathologic examination. The majority of caries lesions restricted to the enamel (D1 and D2) showed white spot caries lesions, and all advanced dentin caries lesions (D4) presented cavitation (Table I). Nevertheless, 50% of initial dentin caries

e38

Braga et al.

OOOOE October 2009

Table I. Relationship between lesion depth obtained by histologic examination and surface status evaluated by direct visual inspection of approximal surfaces in primary molars
Histologic scores D0 D1 D2 D3 D4 Total Surface status Sound 32 6 0 0 0 38 (29.0%) White spot 16 11 22 12 0 61 (46.6%) Cavitation 1 3 6 12 10 32 (24.4%) Total 49 (37.4%) 20 (15.3%) 28 (21.4%) 24 (18.3%) 10 (7.6%) 131 (100.0%)

lesions (D3) evaluated through histopathologic analysis presented cavitations and the other 50% showed white spot caries lesions (Table I). When we considered the surface status as reference standard, LFpen method had statistically signicantly higher sensitivities but lower specicities than the other methods in detecting white spot caries lesions. At cavitation threshold, however, similar sensitivities were obtained by all methods. VI showed higher specicity at cavitation threshold than the other methods. Considering the area under the ROC curves, VI presented higher values at white spot threshold, and VI and LFpen methods showed higher values in detecting cavitated caries lesions (Table II). With both examiners, VI and LFpen methods presented higher sensitivity to detect initial enamel caries lesions than the radiographic method, but the LFpen showed statistically signicantly lower specicities at this threshold. VI presented higher accuracy and area under ROC curve than other methods at both D1 and D3 thresholds (Table III). Regarding the reliability of the methods, VI presented higher kappa values in detecting more advanced caries lesions (D3 and cavitation thresholds), but the values of the radiographic method were higher at enamel and white spot caries lesions. Very low reproducibility values were observed in detecting initial caries lesions using the LFpen method (Table IV). DISCUSSION The present study investigated the performance of a novel pen-type laser uorescence device in detecting approximal caries lesions and compared it with other methods, namely, radiographic and VI. Earlier studies have produced controversial results. On the one hand, an in vitro study performed on permanent teeth showed better performance of the LFpen compared with radiographic method.5 On the other hand, an in vivo study in primary molars showed a similarity between the same

methods.8 These disagreements could be explained by different reference methods. Whereas the one study5 used the lesion depth as reference standard, the other8 performed the temporary separation and used the surface status as reference standard method. As a result of this controversy, the present in vitro study was designed to compare the performance of these methods using 2 types of reference standard: status surface evaluated by direct VI; and lesion depth analyzed through histologic examination. In addition, we investigated the relationship between these 2 types of validation. Regarding the relationship between surface status and lesion depth in primary molars, we observed that the majority of visually sound surfaces did not present caries lesions after sectioning. We also veried that two-thirds of surfaces with white spot caries lesions presented enamel caries lesions using histologic validation. Other authors have observed the same pattern in approximal surfaces of primary molars.11 A high number of white spot lesions, however, presented no caries lesions after sectioning. These caries lesions had probably been remineralized, and then these lesions could not have been observed in the histologic examination. Some lesions incorporate minerals into the body of the enamel lesions owing to changes in the disequilibrium among demineralization and remineralization processes.12 In this way, the enamel lesions could not become visible through the histologic examination. In approximal surfaces, where there are no signicantly external forces disturbing the surface, the white spot could have been kept. All lesions reaching the internal one-half of dentin were cavitated, but regarding the initial dentin caries lesions, one-half of them were noncavitated. This is consistent with an earlier study of primary molars.11 Therefore, when a clinician detects an initial dentin caries lesions, using, e.g., the radiographic method, he or she should provide the temporary separation of the tooth to check if the surface is cavitated or not to reach a correct treatment decision. Usually, VI in detecting approximal caries lesions has presented high specicity but low sensitivity.1 Some researchers recently created a visual scoring system, named ICDAS-II, to improve the validity and reliability of visual caries detection methods.3 Although ICDAS-II has shown good performance in occlusal caries detection,13,14 an in vivo study recently performed using the ICDAS-II to detect approximal caries lesions in primary teeth presented lower sensitivity and higher specicity than other methods,8 which agrees with earlier ndings.1 In the present study, however, VI presented the best validity, with higher values of sensitivity and specicity compared with radiographic and LFpen methods. These disagreements between the results from in vivo and in vitro

OOOOE Volume 108, Number 4

Braga et al. e39

Table II. Performance of different methods in detecting approximal caries lesions in primary teeth at white spot (WS) and cavitation (Cav) thresholds evaluated by direct visual inspection of approximal surfaces
Sensitivity Method Examiner 1 Visual inspection Radiographic LFpen* Examiner 2 Visual inspection Radiographic LFpen* WS 0.85a 0.52b 0.91a 0.67a 0.52b 0.96c Cav 0.59a,b 0.44b 0.72a 0.66a 0.47a 0.56a WS 0.95a 0.82a 0.16b 0.82a 0.82a 0.13b Specicity Cav 0.99a 0.89b 0.79b 0.97a 0.83b 0.96a WS 0.88a 0.60b 0.70b 0.71a 0.60a 0.72a Accuracy Cav 0.89a 0.78b 0.77b 0.89a 0.74b 0.86a WS 0.906a 0.662b 0.744b 0.782a 0.658b 0.658b Az Cav 0.887a 0.785b 0.844a,b 0.868a 0.736b 0.793a,b

Az, Area under the receiver operating characteristic curve; LFpen, pen-type laser uorescence device. *Cut-off points: sound 0-4; WS 4.1-38.0; Cav 38. a,bDifferent letters express statistically signicant difference among values within the same column in the examinations performed by the same examiner (P .05).

Table III. Performance of different methods in detecting approximal caries lesions in primary teeth at enamel (D1) and dentin (D3) thresholds evaluated by histologic examinationc
Sensitivity Method Examiner 1 Visual inspection Radiographic LF pen* Examiner 2 Visual inspection Radiographic LF pen* D1 0.89a 0.55b 0.87a 0.72a 0.54b 0.82a D3 0.50a 0.47a 0.77b 0.59a 0.47a 0.59a D1 0.84a 0.80a 0.25b 0.80a 0.78a 0.47b Specicity D3 0.97a 0.91a 0.71b 0.96a 0.84b 0.87b D1 0.87a 0.64b 0.63b 0.75a 0.63b 0.69a,b Accuracy D3 0.85a 0.79a,b 0.67b 0.86a 0.74b 0.79a,b D1 0.904a 0.679b 0.714b 0.801a 0.646b 0.696b Az D3 0.902a 0.771b 0.807a,b 0.895a 0.747b 0.792b

Abbreviations as in Table II. *Cut-off points: D0 0-8; D1 8.1-30.0; D3 30. a,b Different letters express statistically signicant difference among values within the same column in the examinations performed by the same examiner (P .05).

Table IV. Interexaminer reproducibility of visual inspection, radiographic method, and LFpen device in detecting approximal caries lesions in primary teeth
Kappa values, 95% CI Method Visual inspection Radiographic LFpen ICC,* 95% CI 0.769, 0.689-0.830 0.623, 0.508-0.717 0.669, 0.570-0.749 WS 0.474, 0.322-0.626 0.561, 0.417-0.705 0.100, 0.276-0.473 Cav 0.782, 0.635-0.928 0.531, 0.347-0.715 0.453, 0.274-0.633 D1 0.474, 0.322-0.626 0.561, 0.417-0.705 0.196, 0.002-0.411 D3 0.782, 0.635-0.928 0.531, 0.347-0.715 0.505, 0.350-0.661

ICC, Intraclass correlation coefcient; CI, condence interval; D1, enamel lesion; D3, dentin lesion; other abbreviations as in Table II. *Calculated using the values of the scores or of the LFpen device. Calculated after division in sound versus decayed according to cut-off points described in text for different thresholds.

studies could be due to difculty in simulating the approximal contact. Although no differences have been observed under clinical and laboratory conditions in detection of occlusal caries lesions,15 this difculty in simulating the approximal contact could explain better in vitro performance of VI compared with the results

obtained clinically in approximal surfaces of primary molars. Probably, examiners would have more difculty in detecting caries lesions in the approximal spaces or in observing some discoloration through the marginal ridge in the oral environment. Therefore, the good performance of VI obtained in the present study

e40

Braga et al.

OOOOE October 2009

should be interpreted within this limitation of the design. Further studies should be conducted using the ICDAS to check if the better performance is really due to the new visual scoring system or to the limitation of the in vitro design. Regarding the other methods, a similarity was found in accuracy and in the areas under the ROC curve between the radiographic and LFpen methods. This is in agreement with an earlier in vivo study in approximal surfaces of primary molars.8 At the same time, another in vitro study conducted with permanent teeth found higher performance of LFpen compared with the radiographic method.5 These differences could be attributed to the different teeth used, because primary enamel is thinner than permanent enamel,16 the mineral content of primary enamel is lower, porosity is greater, and caries lesions progress faster than in permanent teeth.17 Nevertheless, contrary to an earlier in vivo study which found a similar sensitivity and specicity between LFpen and radiographic methods,8 the present study showed higher sensitivity of LFpen and higher specicity of radiographic method. Slight differences between the performance of the old version of the LF device obtained under laboratory and clinical conditions have been observed in occlusal caries detection.15 Sensitivities obtained in the present study with the radiographic method in detecting more advanced caries lesions (cavitated or dentin caries lesions) were lower than those obtained in earlier studies.1,8 Our results, however, were similar to another study in permanent teeth.5 In fact, radiographic method is more efcient to detect more advanced caries lesions.18,19 Furthermore, no signicant differences have been found under clinical and laboratory conditions using this method.20 Probably, the present results could be due to the fact of the examiners being recent graduate dentists. Moreover, they did not receive extensive training and calibration sessions. This lack of extensive training could be conrmed by low values of interexaminer reproducibility obtained for most methods (except for VI in detecting more advanced caries lesions). Thus, had training lasted longer, performance and reproducibility of the methods may have increased. Earlier studies have demonstrated that the LF or LFpen device performs better in more advanced caries lesions.8,15,21 In fact, the uorescence of the caries lesions induced by diode laser reects changes in the organic content due to bacterial metabolites, probably because of the presence of porphyrins in caries lesions.22,23 Therefore, because more advanced lesions are more infected than noncavitated ones,24 a lower performance of LFpen in detecting initial caries lesions

would be expected. The present study corroborates this assumption. Despite LFpen having presented good results using the histologic validation as reference standard method, the lesion depth was divided into a relative scale. Other authors have claimed that the LFpen could be useful in monitoring caries lesions.5 For this purpose, the device should present a good reliability and a high correlation with mineral loss. Therefore, further studies are necessary correlating the LFpen readings with results of quantitative methods of mineral loss, such as microradiography, microhardness, or polarized light microscopy. In conclusion, VI presents better results in detecting approximal caries lesions in primary molars using an in vitro study design. However, both radiographic and LFpen methods also show good performance, mainly in detecting more advanced caries lesions.
REFERENCES
1. Bader JD, Shugars DA, Bonito AJ. A systematic review of the performance of methods for identifying carious lesions. J Public Health Dent 2002;62:201-13. 2. Shi XQ, Li G. Detection accuracy of approximal caries by black-and-white and color-coded digital radiographs. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009;107:433-6. 3. Ismail AI, Sohn W, Tellez M, Amaya A, Sen A, Hasson H, Pitts NB. The International Caries Detection and Assessment System (ICDAS): an integrated system for measuring dental caries. Community Dent Oral Epidemiol 2007;35:170-8. 4. Tagtekin DA, Ozyoney G, Baseren M, Ando M, Hayran O, Alpar R, et al. Caries detection with DIAGNOdent and ultrasound. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2008;106:729-35. 5. Lussi A, Hack A, Hug I, Heckenberger H, Megert B, Stich H. Detection of approximal caries with a new laser uorescence device. Caries Res 2006;40:97-103. 6. Huth KC, Neuhaus KW, Gygax M, Bucher K, Crispin A, Paschos E, et al. Clinical performance of a new laser uorescence device for detection of occlusal caries lesions in permanent molars. J Dent 2008;36:1033-40. 7. Lussi A, Hellwig E. Performance of a new laser uorescence device for the detection of occlusal caries in vitro. J Dent 2006;34:467-71. 8. Novaes TF, Matos R, Braga MM, Imparato JCP, Raggio DP, Mendes FM. Performance of pen-type laser uorescence device and conventional methods in detecting approximal caries lesions in primary teethin vivo study. Caries Res 2009;43:36-42. 9. Francescut P, Zimmerli B, Lussi A. Inuence of different storage methods on laser uorescence values: a two-year study. Caries Res 2006;40:181-5. 10. Ekstrand KR, Ricketts DNJ, Kidd EAM. Reproducibility and accuracy of three methods for assessment of demineralization depth of the occlusal surface: an in vitro examination. Caries Res 1997;31:224-31. 11. Feldens CA, Tovo MF, Kramer PF, Feldens EG, Ferreira SH, Finkler M. An in vitro study of the correlation between clinical and radiographic examinations of proximal carious lesions in primary molars. J Clin Pediatr Dent 2003;27:143-7. 12. Featherstone JDB. The continuum of dental caries evidence for a dynamic disease process. J Dent Res 2004;83:C39-C42. 13. Ekstrand KR, Martignon S, Ricketts DJ, Qvist V. Detection and

OOOOE Volume 108, Number 4

activity assessment of primary coronal caries lesions: a methodologic study. Oper Dent 2007;32:225-35. Jablonski-Momeni A, Stachniss V, Ricketts DN, Heinzel-Gutenbrunner M, Pieper K. Reproducibility and accuracy of the ICDAS-II for detection of occlusal caries in vitro. Caries Res 2008;42:79-87. Reis A, Mendes FM, Angnes V, Angnes G, Grande RHM, Loguercio AD. Performance of methods of occlusal caries detection in permanent teeth under clinical and laboratory conditions. J Dent 2006;34:89-96. Mortimer KV. The relationship of deciduous enamel structure to dental disease. Caries Res 1970;4:206-23. Shellis RP. Relationship between human enamel structure and the formation of caries-like lesions in vitro. Arch Oral Biol 1984;29:975-81. Kidd EAM, Pitts NB. A reappraisal of the value of the bitewing radiograph in the diagnosis of posterior approximal caries. Br Dent J 1990;169:195-200. Pitts NB, Rimmer PA. An in vivo comparison of radiographic and directly assessed clinical caries status of posterior approximal surfaces in primary and permanent teeth. Caries Res 1992; 26:146-52. Hintze H, Wenze A. Clinical and laboratory radiographic caries

Braga et al. e41

diagnosis. A study of the same teeth. Dentomaxillofac Radiol 1996;25:115-8. Braga MM, Nicolau J, Rodrigues CRMD, Imparato JCP, Mendes FM. Laser uorescence device does not perform well in detection of early caries lesions in primary teeth: an in vitro study. Oral Health Prev Dent 2008;6:165-9. Buchalla W. Comparative uorescence spectroscopy shows differences in noncavitated enamel lesions. Caries Res 2005;39:150-6. Mendes FM, Pinheiro SL, Bengtson AL. Effect of alteration in organic material of the occlusal caries on Diagnodent readings. Braz Oral Res 2004;18:141-4. Kidd EA, Banerjee A, Ferrier S, Longbottom C, Nugent Z. Relationships between a clinical-visual scoring system and two histological techniques: a laboratory study on occlusal and approximal carious lesions. Caries Res 2003;37:125-9.

14.

21.

15.

22. 23.

16. 17.

24.

18.

Reprint requests: Fausto Medeiros Mendes Faculdade de Odontologia da Universidade de So Paulo Av. Lineu Prestes, 2227 So PauloSP Brazil fmmendes@usp.br

19.

20.