An occlusal restoration is usually considered necessary when a lesion
extends into the dentin. This study aimed to assess the relationship b
etween the clinical appearance of occlusal caries and the lesion depth
. Special attention was paid to differentiate the inner, partly demine
ralitzed layer of carious dentin that should be preserved from the out
er carious layer that should be removed before placing a restoration.
Fissures in extracted premolars and molars were classified clinically
by the authors. The mean pairwise interobserver agreement, weighted Co
hen's kappa (kappa), was 0.73, which indicated 'substantial agreement'
. The depth of zones revealing softness by probing and/or discoloratio
n was assessed on ground surfaces. Of the carious fissures that were s
cored sound by clinical examination (n = 17), 76% revealed enamel lesi
ons on the ground surfaces. Of those occlusal surfaces with a discolor
ation only but no cavitation (n = 57), 13 showed dentin lesions, but o
nly 5 of these revealed softness by probing. Of those lesions with ver
y small cavitations (n = 60), 19 were confined to enamel, 41 showed de
ntin lesions and of these 32 revealed softness by probing. All but 1 o
f the lesions with small cavitations (n = 28) showed dentin involvemen
t and all but 2 of these revealed softness. If cavitation in the occlu
sal lesion was used as a criterion for identifying dentin lesions with
softness, the sensitivity and specificity values were 0.92 and 0.69,
which means a relatively high frequency of overregistration.