Diagnosis of dentin involvement in occiusal caries based on visuai and radiograpiiic examination of tine teetli
Riviane B. Nytun, Magne Raadal and Ivar Espelid Department of Pedodontics, School of Dentistry, tjniversity of Bergen, Bergen, Norway
Nytun RB, Raadal M, Espelid I: Diagnosis of dentin involvement in oeehisal caries based on visual and radt'ographie examination ofthe teeth. Seand J Dent Res 1992: 100: 144-8.
The purpose of the present study was to examine the validity of visual, radiographic and combined visual-radiographic examination of occlusal caries, with special reference to the diagnosis of dentin lesions, using the microscopic diagnosis after sectioning the teeth as validating criterion. Thirty extracted permanent molar teeth with signs of occlusal caries were selected and radiographed using a standard method similar to bitewing. Ten dentists, five from the Department of Cariology and Endodontics and five from thc Department of Pedodontics, were asked to make a visual, radiographic and combined visual-radiographic caries examination with an interval of at least 1 month between different examinations. The teeth were sectioned bucco-lingually and the caries assessed as enamel or dentin lesion in a stereomicroscope. The sensitivity/specificity of thc diagnoses of dentin lesions was 0.72/0.41 for the visual, 0.66/0.50 for the radiographic, and 0.86/0.64 for the combined visual-radiographic examinations. The Receiver Operating Characteristic (ROC) method revealed a statistically significant difference between combined visual-radiographic examination and the two other, but not between the isolated visual or radiographic examination. It is concluded that thc combined use of visual and radiographic examination is better than either visual or radiographic examination alone.
The decline in prevalence of caries in pits and fissures has not been as great as for smooth surface caries, and this fact has drawn attention to prevention of caries on the occlusal surfaces by the use of fissure sealants (1). In teeth with occlusal caries, there is also a trend of selecting the most conservative technique with respect to saving sound tooth substance (2), e.g. preferring non-invasive before invasive techniques (3). The choice of treatment is based on the caries diagnosis, and the quality of diagnosis is of major importance. It is a well known problem that the diagnosis of fissure caries is difficult, and many studies have shown inconsistency between observers (4). Some reports even claim that this is a growing problem related to the increased use of fiuorides and general reduction of caries activity in populations, leading to many clinical cases with large dentin lesion beneath seemingly sound occlusal surfaces (5-8). Despite attempts to find alternative diagnostic methods for occlusal caries (9, 10), mirror, probe, and light are still the tools most extensively used. Bite-
Key words: dental caries, diagnosis; fissure sealing; radiography, oral Magne Raadal, Department of Pedodontics, Schooi of Dentistry, Aarstadveien 17, N-5009 Bergen, Norway Accepted for publication 11 July 1991
wing radiographs are well recognized for diagnosing approximal decay, and even if their importance in diagnosing occlusal lesions is indisputable (7-9), there are few studies dealing with their possible contributing effect in the discrimination between enamel and dentin lesions. The purpose ofthe present study was to examine the validity of visual, radiographic and combined visual-radiographic examination of occlusal caries, with special reference to the diagnosis of dentin lesions, using microscopic diagnosis after sectioning the teeth as a validating criterion. Material and methods Thirty extracted molar teeth with signs of fissure caries in the occlusal surface were selected. Teeth with approximal, buccal, and lingual caries were excluded to make sure that caries-like radiolucencies did not originate from any other part of the tooth. The occlusal surfaces were brushed with pumice,
Diagnosis of occlusal caries cleaned with an airpolishing device (Satelec), and stored in numbered glasses containing saline. Each tooth was radiographed using a standard method similar to bitewing projection. The radiographs were mounted in frames so that only the tooth in question might be viewed through a window, and numbered according to a code, in which the A'-ray numbers were different from the equivalent tooth number to prevent the observers remembering specific tooth numbers. Ten dentists, five from the Department of Cariology and Fndodontics and five from the Department of Pedodontics, participated in the study. They were asked to make visual, radiographic, and combined visual-radiographic caries examinations ofthe occlusal surface of each ofthe 30 teeth. The interval between the different examinations was at least 1 month. The dentists were not informed about the authors' opinion of the teeth's caries status. They were asked to give separate scores for enamel and dentin caries in each tooth, based on the following system: Score 1: Almost definitely not present Score 2: Probably not present . Score 3: Unsure Score 4: Probably present Score 5: Almost definitely present. Visual examination
After being dried with compressed air, the teeth were examined under an operating light. No probe or other instruments were used. The following criteria for caries were given: Enamel caries: Breakdown of enamel of the occlusal fissure, without shadow or opacity underneath the enamel indicating dentin caries. Dentin caries: Breakdown of walls of fissures with obvious dentin involvement, or shadow/opacity indicating dentin caries.
145
iUlicroscopic examination
The roots of each tooth were cut off and the crowns vertically sectioned in a bucco-lingual direction with a circular saw. The sections, about 350 ^m thick, were analyzed independently by two observers (R.B.N., M.R.) in a stereomicroscope under incident light at x 4 magnification. Both sides of each section were observed. In cases of inconsistency between the examiners, the sections were re-analyzed and discussed until a consensus was obtained. For each tooth the diagnosis was based on the following criteria: No caries: All sections caries-free. Enamel caries: Carious destruction of enamel surface, i.e. combination of white demineralizated area and loss of enamel surface. No discoloration or cavity beneath the enamel-dentin junction. Dentin caries: Caries involving enamel with carious area beneath the enamel-dentin junction. The carious area in the dentin may present as a discolored area and/or a cavity. The sensitivity and specificity of the diagnoses of dentin lesions, using the microscopic findings as validating criteria, were calculated using the diagnostic scores 1-3 as negative and scores 4—5 as positive caries diagnosis. The diagnostic quality of the examinations with respect to teeth with dentin lesions, was illustrated by the use of Receiver Operating Characteristic (ROC) curves (11), in which the diagnostic scores were gradually cumulated (scores 5, 5-)-4, 5-1-443, 5-f-4-l-3-F2) to obtain four different decision thresholds of dentin lesion being present, and the corresponding sensitivity and specificity were calculated. The differences between the curves were tested by calculating and comparing the ai'ea beneath the ROC-curves using a computer program (12). Results
Radiographic examination
The radiographs were examined using a magnifying ( X 2) X-ray viewer. The criteria for caries were: Enamel caries: Radioiucency restricted to enamel. Dentin caries: Radioiucency involving dentin. Combined visual-radiographic examination
The teeth and corresponding radiographs were observed together using the same criteria for enamel and dentin caries as before. The decision was made on the basis of a joint visual and radiographic observation.
The microscopic examinations revealed that all of the 30 teeth had caries. In 7 teeth the carious lesion was restricted to the enamel, while 23 teeth showed dentin lesions. In 9 of the 23 teeth with dentin lesions, the lesions were characterized by discolored dentin tissue beneath the enamel lesion, while 14 teeth also showed breakdown of dentin tissue (cavitation). The distributions of under- and over-registrations as well as the correct diagnosis in the different examinations are illustrated in Fig. 1 for teeth with enamel and with dentin lesions. Dentin lesions were correctly diagnosed more frequently than enamel lesions, primarily because of high portions of over-
146
Nytun et al
ENAMEL LESIONS
DENTIN LESIONS
RADIOORAPHIC
^
CORRECT
^
UNDER
C3 OVER
'
^M CORRECT
VI3UAL/RA0I0QRAPHIC
^ ^ UNDER
Fig. 1. Distribution of correct registrations, and under- and over-registrations in teeth with enamel lesions and dentin lesions. Diagnostic scores 1-3 are interpreted as negative caries diagnoses and scores 4-5 as positive.
registrations in teeth witb enamel caries only. The combined use of visual and radiographic examination gave the best results with respect to the frequencies of correct diagnoses, both for teeth with enamel lesions and those with dentin lesions. The frequency of over-registrations in the teeth with enamel lesions was reduced from 59% by visual examination to 36% when the visual examination was combined with X-rays. In teeth with dentin lesions the frequency of under-registrations was correspondingly reduced from 28% to 15%. The sensitivity and specificity of the different examinations with respect to correct diagnosis of dentin caries, are given in Table 1. There were great variations within the group of dentists, as illustrated by the ranges among individual dentists. Visual combined with radiographic examinations gave higher values than isolated visual or radiographic. In Fig. 2 the relationship between the sensitivity and specificity is illustrated by ROC-curves, in which the different eut-off points represent the four different deeision thresholds. The statistical difference between the diagnostic methods, as expressed by the difference between the areas beneath the curves, was significant between visual-radiographic examination and the two other methods (7^ = 0.00) but not between isolated visual or radiographic examination (P>0.05). The individual scores made by the 10 dentists showed great variations, and Fig. 3 shows ROC-
curves from the combined visual-radiographic examination representing the two dentists who scored most differently from the mean. Discussion Since this study was based on the examination of extracted teeth, it is necessary to make some reservations with respect to its clinical relevance. Most of the molars were permanent third molars, with very irregular occlusal fissures, which may be somewhat different from first and second molars. The clinical appearance of caries in extracted teeth may also be different from that in vivo, particularly with respect to the nature of discolorations. Probing was also prohibited, since the same teeth were
.8 -
.5 -
.A -
VISUAL/RAD IOGRAPHIC VISUAL RADIOGRARHIC
Table 1 Sensitivity and specificity of diagnoses of dentin tesions performed by all observers. (The range among individual observers is in parentheses) Specificity Sensitivity Method 0.72(0.39-0.91) 0.41(0.28-0.72) Visual 0.66(0.39-0.96) 0.50(0.14-0.72) Radiographic 0.86 (0.66-1.00) 0.64 (0.14-1.00) Combined visual-radiogr.
.5
.4
.2
SPECIFICITY
Fig. 2. ROC curves indicating quality of diagnosis of dentin caries in different examinations. Cut-off points reflect 4 decision thresholds based on a gradual cumulation of the clinical scores
Diagnosis of occlusal caries
.6 ^
cn
.4 -
LU
.2 -
,OBSERVER 4 I OBSERVER B ALL OBSERVERS
.5
.4
(N=10)
.2
SPECIFICITY
Fig. 3. ROC curves representing the combined visual-radiographic diagnoses made by two dentists who scored the most differently from the mean.
used by many examiners and probing might have affected the clinical appearance from one examination to another. The results of this study support many previous investigations in stating the problems of diagnosing caries in pits and fissures (4). It seems that lesions confined to the enamel only are difficult to diagnose, since only 36% of the recordings in these teeth were correctly assessed by the 10 observers when a joint visual and radiographic examination was done (Fig. 1). The visual diagnosis was characterized by a high portion of over-registrations, while the radiographs alone seemed to be of almost no value. According to the authors who selected the teeth, all of them showed signs of initial caries. Nevertheless, 11% ofthe visual registrations were sound, and this clearly demonstrates the inconsistency among dentists in diagnosing occlusal caries. It may, however, be noted that only seven teeth had such lesions, and the microscopic examination revealed that even if all of them showed demineralized fissure enamel, the loss of substance was minimal in the majority of cases. The diagnosis of dentin lesions is of special importance, since most practitioners prefer invasive treatments for such cases, being concerned in leaving such processes beneath a fissure sealing (3, 13-16). The material comprised 23 teeth with such lesions, while the 7 teeth with enamel lesions constituted the controls. Since 10 different dentists examined them by three different examinations, the material is probably sufficient for drawing some conclusions. It seems obvious that bitewing radio-
147
graphs are of great value in diagnosing dentin lesions in the occlusal surface of molar teeth, assuming that they are accompanied by visual inspection (Table 1, Fig. 1). Even if there were great variations among the different observers (Table 1, Fig. 3), it was a general trend that all of them improved their diagnoses when both visual inspection and radiographs were used. The ROC curves in Fig. 2 are probably the most relevant parameter for measuring this entity, since they are constructed on the basis of the five confidence scores used by the examiners, combining the sensitivity and specificity values. It seems that the isolated visual and radiographic examinations were connected with high levels of inconsistency, while the curve for visual-radiographic examination shows a diagnostic quality comparable to those usually found when dentists diagnose dentin lesions on approximal surfaces on radiographs (17). One ofthe major reasons for initiating this study was the need for information about the diagnostic basis for application of fissure sealants to teeth with signs of initial caries, adopting the hypotheses that sealants may be safely applied to teeth with enamel lesions but not to teeth with dentin lesions. The clinical application of this approach is that over-registrations in teeth with enamel lesions and under-registrations in teeth with dentin lesions, should be avoided. In the present study the most favorable results with respect to this approach were obtained when a joint visual and radiographic examination was used, and the frequencies were 36% and 15% for over- and under-registrations, respectively (Fig. 1). References 1. STAMM JW. Is there a need for dental sealants? Epidemiological indications in the 1980s. J Dent Educ 1984; 48; 9-17. 2. SuRMONT P, MARTENS L, D'HAUWERS R . A decision tree
for the treatment of caries in posterior teeth. Quintessence Int 1990; 21; 239-46. 3. MEIERS CM, JENSEN ME. Management ofthe questionable
carious fissure; invasive vs noninvasive techniques. J Am Dent Assoc 1984; 108; 64-8. 4. ROCK WR The diagnosis of early carious lesions - a review. Pediatr Dent 1987; 3; 1-6. 5. BALL IA. The "Fluoride syndrome"; Occult caries? Br Dent J 1986; 160: 75-6. 6. SAWLE RF, ANDLAW RJ. Has occlusal caries become more
difficult to diagnose? A study comparing clinically undetected lesions in molar teeth of 14-16 year old children in 1974 and 1982. Br Dent J 1988; 164; 209-1!. 7. WEERHEIJM KL, VAN AMERONGEN WE, EGGINK CO. The
clinical diagnosis of occlusal caries; a problem. J Dent Child 1989; 56; 196-200. 8. CREANOR SL, RUSSELL JI, STRANG DM, STEPHEN KW, BUR-
CHELL CK. The prevalence of clinically undetected occlusal dentine caries in Scottish adolescents. Br Denl J 1990; 169; 126-9. 9. MiTROPOULOS CM. A comparison of fibre-optic transillu-
148
Nytun et al.
mination with bitewing radiographs. Br Dent J 1985; 15921-3. 10. ROCK WP, Kii:)i:) EAM. The electronic detection of demineralisation in occlusal rissure.s. Br Dent J 1988; 164: 243-7. 11. Swi'TS JA, PirKi:TT RM. Evaluation of diagnostic systems. Methods from signal detection theory. New York: Academic Press, 1982. 12. CI;NT()R RM, KKTGUTLY J. The ROCamdyzer (version 5.0)
Richmond, VA, 1988. 13. RAADM., M . Follow-up study of sealing and filling with composite resin in the prevention of occlusal caries. Community Dent Oral Epidemiol 1978; 6: 176-80.
14. SiMONSEN, RJ. Preventive resin restorations: Three-year results. .lADA 1980; 100; 535-9. 15. CRAWI'ORD P J M . Sealant restorations (preventive resin restorations). Br Dent J 1988; 165; 250-3. 16. BROWNHILL JW, SHTCOS JC. Treatment selections for fis-
sured grooves of permanent molar teeth. J Dent Child 1990; 57; 274-8. 17. EsptLiD I. Radiographic diagnoses and treatment decisions on approximal caries. Comnninity Dent Oral Epidemiol 1986; 14; 265-70.
Riviane B. Nytun, Magne Raadal and Ivar Espelid Department of Pedodontics, School of Dentistry, tjniversity of Bergen, Bergen, Norway
Nytun RB, Raadal M, Espelid I: Diagnosis of dentin involvement in oeehisal caries based on visual and radt'ographie examination ofthe teeth. Seand J Dent Res 1992: 100: 144-8.
The purpose of the present study was to examine the validity of visual, radiographic and combined visual-radiographic examination of occlusal caries, with special reference to the diagnosis of dentin lesions, using the microscopic diagnosis after sectioning the teeth as validating criterion. Thirty extracted permanent molar teeth with signs of occlusal caries were selected and radiographed using a standard method similar to bitewing. Ten dentists, five from the Department of Cariology and Endodontics and five from thc Department of Pedodontics, were asked to make a visual, radiographic and combined visual-radiographic caries examination with an interval of at least 1 month between different examinations. The teeth were sectioned bucco-lingually and the caries assessed as enamel or dentin lesion in a stereomicroscope. The sensitivity/specificity of thc diagnoses of dentin lesions was 0.72/0.41 for the visual, 0.66/0.50 for the radiographic, and 0.86/0.64 for the combined visual-radiographic examinations. The Receiver Operating Characteristic (ROC) method revealed a statistically significant difference between combined visual-radiographic examination and the two other, but not between the isolated visual or radiographic examination. It is concluded that thc combined use of visual and radiographic examination is better than either visual or radiographic examination alone.
The decline in prevalence of caries in pits and fissures has not been as great as for smooth surface caries, and this fact has drawn attention to prevention of caries on the occlusal surfaces by the use of fissure sealants (1). In teeth with occlusal caries, there is also a trend of selecting the most conservative technique with respect to saving sound tooth substance (2), e.g. preferring non-invasive before invasive techniques (3). The choice of treatment is based on the caries diagnosis, and the quality of diagnosis is of major importance. It is a well known problem that the diagnosis of fissure caries is difficult, and many studies have shown inconsistency between observers (4). Some reports even claim that this is a growing problem related to the increased use of fiuorides and general reduction of caries activity in populations, leading to many clinical cases with large dentin lesion beneath seemingly sound occlusal surfaces (5-8). Despite attempts to find alternative diagnostic methods for occlusal caries (9, 10), mirror, probe, and light are still the tools most extensively used. Bite-
Key words: dental caries, diagnosis; fissure sealing; radiography, oral Magne Raadal, Department of Pedodontics, Schooi of Dentistry, Aarstadveien 17, N-5009 Bergen, Norway Accepted for publication 11 July 1991
wing radiographs are well recognized for diagnosing approximal decay, and even if their importance in diagnosing occlusal lesions is indisputable (7-9), there are few studies dealing with their possible contributing effect in the discrimination between enamel and dentin lesions. The purpose ofthe present study was to examine the validity of visual, radiographic and combined visual-radiographic examination of occlusal caries, with special reference to the diagnosis of dentin lesions, using microscopic diagnosis after sectioning the teeth as a validating criterion. Material and methods Thirty extracted molar teeth with signs of fissure caries in the occlusal surface were selected. Teeth with approximal, buccal, and lingual caries were excluded to make sure that caries-like radiolucencies did not originate from any other part of the tooth. The occlusal surfaces were brushed with pumice,
Diagnosis of occlusal caries cleaned with an airpolishing device (Satelec), and stored in numbered glasses containing saline. Each tooth was radiographed using a standard method similar to bitewing projection. The radiographs were mounted in frames so that only the tooth in question might be viewed through a window, and numbered according to a code, in which the A'-ray numbers were different from the equivalent tooth number to prevent the observers remembering specific tooth numbers. Ten dentists, five from the Department of Cariology and Fndodontics and five from the Department of Pedodontics, participated in the study. They were asked to make visual, radiographic, and combined visual-radiographic caries examinations ofthe occlusal surface of each ofthe 30 teeth. The interval between the different examinations was at least 1 month. The dentists were not informed about the authors' opinion of the teeth's caries status. They were asked to give separate scores for enamel and dentin caries in each tooth, based on the following system: Score 1: Almost definitely not present Score 2: Probably not present . Score 3: Unsure Score 4: Probably present Score 5: Almost definitely present. Visual examination
After being dried with compressed air, the teeth were examined under an operating light. No probe or other instruments were used. The following criteria for caries were given: Enamel caries: Breakdown of enamel of the occlusal fissure, without shadow or opacity underneath the enamel indicating dentin caries. Dentin caries: Breakdown of walls of fissures with obvious dentin involvement, or shadow/opacity indicating dentin caries.
145
iUlicroscopic examination
The roots of each tooth were cut off and the crowns vertically sectioned in a bucco-lingual direction with a circular saw. The sections, about 350 ^m thick, were analyzed independently by two observers (R.B.N., M.R.) in a stereomicroscope under incident light at x 4 magnification. Both sides of each section were observed. In cases of inconsistency between the examiners, the sections were re-analyzed and discussed until a consensus was obtained. For each tooth the diagnosis was based on the following criteria: No caries: All sections caries-free. Enamel caries: Carious destruction of enamel surface, i.e. combination of white demineralizated area and loss of enamel surface. No discoloration or cavity beneath the enamel-dentin junction. Dentin caries: Caries involving enamel with carious area beneath the enamel-dentin junction. The carious area in the dentin may present as a discolored area and/or a cavity. The sensitivity and specificity of the diagnoses of dentin lesions, using the microscopic findings as validating criteria, were calculated using the diagnostic scores 1-3 as negative and scores 4—5 as positive caries diagnosis. The diagnostic quality of the examinations with respect to teeth with dentin lesions, was illustrated by the use of Receiver Operating Characteristic (ROC) curves (11), in which the diagnostic scores were gradually cumulated (scores 5, 5-)-4, 5-1-443, 5-f-4-l-3-F2) to obtain four different decision thresholds of dentin lesion being present, and the corresponding sensitivity and specificity were calculated. The differences between the curves were tested by calculating and comparing the ai'ea beneath the ROC-curves using a computer program (12). Results
Radiographic examination
The radiographs were examined using a magnifying ( X 2) X-ray viewer. The criteria for caries were: Enamel caries: Radioiucency restricted to enamel. Dentin caries: Radioiucency involving dentin. Combined visual-radiographic examination
The teeth and corresponding radiographs were observed together using the same criteria for enamel and dentin caries as before. The decision was made on the basis of a joint visual and radiographic observation.
The microscopic examinations revealed that all of the 30 teeth had caries. In 7 teeth the carious lesion was restricted to the enamel, while 23 teeth showed dentin lesions. In 9 of the 23 teeth with dentin lesions, the lesions were characterized by discolored dentin tissue beneath the enamel lesion, while 14 teeth also showed breakdown of dentin tissue (cavitation). The distributions of under- and over-registrations as well as the correct diagnosis in the different examinations are illustrated in Fig. 1 for teeth with enamel and with dentin lesions. Dentin lesions were correctly diagnosed more frequently than enamel lesions, primarily because of high portions of over-
146
Nytun et al
ENAMEL LESIONS
DENTIN LESIONS
RADIOORAPHIC
^
CORRECT
^
UNDER
C3 OVER
'
^M CORRECT
VI3UAL/RA0I0QRAPHIC
^ ^ UNDER
Fig. 1. Distribution of correct registrations, and under- and over-registrations in teeth with enamel lesions and dentin lesions. Diagnostic scores 1-3 are interpreted as negative caries diagnoses and scores 4-5 as positive.
registrations in teeth witb enamel caries only. The combined use of visual and radiographic examination gave the best results with respect to the frequencies of correct diagnoses, both for teeth with enamel lesions and those with dentin lesions. The frequency of over-registrations in the teeth with enamel lesions was reduced from 59% by visual examination to 36% when the visual examination was combined with X-rays. In teeth with dentin lesions the frequency of under-registrations was correspondingly reduced from 28% to 15%. The sensitivity and specificity of the different examinations with respect to correct diagnosis of dentin caries, are given in Table 1. There were great variations within the group of dentists, as illustrated by the ranges among individual dentists. Visual combined with radiographic examinations gave higher values than isolated visual or radiographic. In Fig. 2 the relationship between the sensitivity and specificity is illustrated by ROC-curves, in which the different eut-off points represent the four different deeision thresholds. The statistical difference between the diagnostic methods, as expressed by the difference between the areas beneath the curves, was significant between visual-radiographic examination and the two other methods (7^ = 0.00) but not between isolated visual or radiographic examination (P>0.05). The individual scores made by the 10 dentists showed great variations, and Fig. 3 shows ROC-
curves from the combined visual-radiographic examination representing the two dentists who scored most differently from the mean. Discussion Since this study was based on the examination of extracted teeth, it is necessary to make some reservations with respect to its clinical relevance. Most of the molars were permanent third molars, with very irregular occlusal fissures, which may be somewhat different from first and second molars. The clinical appearance of caries in extracted teeth may also be different from that in vivo, particularly with respect to the nature of discolorations. Probing was also prohibited, since the same teeth were
.8 -
.5 -
.A -
VISUAL/RAD IOGRAPHIC VISUAL RADIOGRARHIC
Table 1 Sensitivity and specificity of diagnoses of dentin tesions performed by all observers. (The range among individual observers is in parentheses) Specificity Sensitivity Method 0.72(0.39-0.91) 0.41(0.28-0.72) Visual 0.66(0.39-0.96) 0.50(0.14-0.72) Radiographic 0.86 (0.66-1.00) 0.64 (0.14-1.00) Combined visual-radiogr.
.5
.4
.2
SPECIFICITY
Fig. 2. ROC curves indicating quality of diagnosis of dentin caries in different examinations. Cut-off points reflect 4 decision thresholds based on a gradual cumulation of the clinical scores
Diagnosis of occlusal caries
.6 ^
cn
.4 -
LU
.2 -
,OBSERVER 4 I OBSERVER B ALL OBSERVERS
.5
.4
(N=10)
.2
SPECIFICITY
Fig. 3. ROC curves representing the combined visual-radiographic diagnoses made by two dentists who scored the most differently from the mean.
used by many examiners and probing might have affected the clinical appearance from one examination to another. The results of this study support many previous investigations in stating the problems of diagnosing caries in pits and fissures (4). It seems that lesions confined to the enamel only are difficult to diagnose, since only 36% of the recordings in these teeth were correctly assessed by the 10 observers when a joint visual and radiographic examination was done (Fig. 1). The visual diagnosis was characterized by a high portion of over-registrations, while the radiographs alone seemed to be of almost no value. According to the authors who selected the teeth, all of them showed signs of initial caries. Nevertheless, 11% ofthe visual registrations were sound, and this clearly demonstrates the inconsistency among dentists in diagnosing occlusal caries. It may, however, be noted that only seven teeth had such lesions, and the microscopic examination revealed that even if all of them showed demineralized fissure enamel, the loss of substance was minimal in the majority of cases. The diagnosis of dentin lesions is of special importance, since most practitioners prefer invasive treatments for such cases, being concerned in leaving such processes beneath a fissure sealing (3, 13-16). The material comprised 23 teeth with such lesions, while the 7 teeth with enamel lesions constituted the controls. Since 10 different dentists examined them by three different examinations, the material is probably sufficient for drawing some conclusions. It seems obvious that bitewing radio-
147
graphs are of great value in diagnosing dentin lesions in the occlusal surface of molar teeth, assuming that they are accompanied by visual inspection (Table 1, Fig. 1). Even if there were great variations among the different observers (Table 1, Fig. 3), it was a general trend that all of them improved their diagnoses when both visual inspection and radiographs were used. The ROC curves in Fig. 2 are probably the most relevant parameter for measuring this entity, since they are constructed on the basis of the five confidence scores used by the examiners, combining the sensitivity and specificity values. It seems that the isolated visual and radiographic examinations were connected with high levels of inconsistency, while the curve for visual-radiographic examination shows a diagnostic quality comparable to those usually found when dentists diagnose dentin lesions on approximal surfaces on radiographs (17). One ofthe major reasons for initiating this study was the need for information about the diagnostic basis for application of fissure sealants to teeth with signs of initial caries, adopting the hypotheses that sealants may be safely applied to teeth with enamel lesions but not to teeth with dentin lesions. The clinical application of this approach is that over-registrations in teeth with enamel lesions and under-registrations in teeth with dentin lesions, should be avoided. In the present study the most favorable results with respect to this approach were obtained when a joint visual and radiographic examination was used, and the frequencies were 36% and 15% for over- and under-registrations, respectively (Fig. 1). References 1. STAMM JW. Is there a need for dental sealants? Epidemiological indications in the 1980s. J Dent Educ 1984; 48; 9-17. 2. SuRMONT P, MARTENS L, D'HAUWERS R . A decision tree
for the treatment of caries in posterior teeth. Quintessence Int 1990; 21; 239-46. 3. MEIERS CM, JENSEN ME. Management ofthe questionable
carious fissure; invasive vs noninvasive techniques. J Am Dent Assoc 1984; 108; 64-8. 4. ROCK WR The diagnosis of early carious lesions - a review. Pediatr Dent 1987; 3; 1-6. 5. BALL IA. The "Fluoride syndrome"; Occult caries? Br Dent J 1986; 160: 75-6. 6. SAWLE RF, ANDLAW RJ. Has occlusal caries become more
difficult to diagnose? A study comparing clinically undetected lesions in molar teeth of 14-16 year old children in 1974 and 1982. Br Dent J 1988; 164; 209-1!. 7. WEERHEIJM KL, VAN AMERONGEN WE, EGGINK CO. The
clinical diagnosis of occlusal caries; a problem. J Dent Child 1989; 56; 196-200. 8. CREANOR SL, RUSSELL JI, STRANG DM, STEPHEN KW, BUR-
CHELL CK. The prevalence of clinically undetected occlusal dentine caries in Scottish adolescents. Br Denl J 1990; 169; 126-9. 9. MiTROPOULOS CM. A comparison of fibre-optic transillu-
148
Nytun et al.
mination with bitewing radiographs. Br Dent J 1985; 15921-3. 10. ROCK WP, Kii:)i:) EAM. The electronic detection of demineralisation in occlusal rissure.s. Br Dent J 1988; 164: 243-7. 11. Swi'TS JA, PirKi:TT RM. Evaluation of diagnostic systems. Methods from signal detection theory. New York: Academic Press, 1982. 12. CI;NT()R RM, KKTGUTLY J. The ROCamdyzer (version 5.0)
Richmond, VA, 1988. 13. RAADM., M . Follow-up study of sealing and filling with composite resin in the prevention of occlusal caries. Community Dent Oral Epidemiol 1978; 6: 176-80.
14. SiMONSEN, RJ. Preventive resin restorations: Three-year results. .lADA 1980; 100; 535-9. 15. CRAWI'ORD P J M . Sealant restorations (preventive resin restorations). Br Dent J 1988; 165; 250-3. 16. BROWNHILL JW, SHTCOS JC. Treatment selections for fis-
sured grooves of permanent molar teeth. J Dent Child 1990; 57; 274-8. 17. EsptLiD I. Radiographic diagnoses and treatment decisions on approximal caries. Comnninity Dent Oral Epidemiol 1986; 14; 265-70.
