The potential for sampling error in incisional biopsies of odontogenic keratocysts

Abstract

Incisional biopsies of large lytic jaw lesions are commonly performed in order to obtain a pretreatment diagnosis. Odontogenic keratocyst (OKC) is a frequent entity in the differential diagnosis of such lesions and mandates a relatively more aggressive surgical approach. Preliminary analysis of incisional biopsies of OKCs revealed that often, in areas of inflammation, samples did not show the classic histopathologic diagnostic features. Instead, the epithelial lining displayed a squamous-type metaplasia that precluded such diagnosis if that was the only area of epithelium sampled. The purpose of this study, therefore, was to determine the likelihood of sampling error resulting in a nondiagnosis of OKC in cases which truly were OKCs. Fifteen cases of totally excised inflamed OKCs were histomorphometrically analyzed to determine the total area of epithelium in each cyst that was diagnostic or nondiagnostic, respectively, utilizing classic histopathologic criteria. The mean for the total area of lining epithelium was 0.098 mm², with nondiagnostic and diagnostic epithelium comprising 0.033 mm² and 0.063 mm² respectively. Standard deviations were 0.025 for the total areas, 0.026 for nondiagnostic areas, and 0.016 for diagnostic areas. Standard errors of the mean were 0.007, 0.007, and 0.004, respectively. Results show that 33.6% of the area sampled when incisionally biopsying an inflamed OKC is likely to be nondiagnostic. We conclude that in order to minimize the risk of sampling cystic lining nondiagnostic for OKC, large lytic lesions should be sampled away from sites associated with inflammation, such as those in proximity to sulcular regions or partially impacted teeth or in areas exposed to the oral cavity. The caveat is also given that the smaller the incisional biopsy, the greater the chance for sampling error.