The Diagnostic Accuracy of Incisional Biopsy in the Oral Cavity

Abstract

To determine the accuracy of incisional biopsy examination to diagnose oral lesions. This retrospective cohort study was performed to determine the concordance rate between incisional biopsy examination and definitive resection diagnosis for different oral lesions. The study sample was derived from the population of patients who presented to the Department of Oral and Maxillofacial Surgery, Massachusetts General Hospital (Boston, MA) from January 2005 through December 2012. Inclusion criteria were the diagnosis of an oral lesion from an incisional biopsy examination, subsequent diagnosis from the definitive resection of the same lesion, and complete clinical and pathologic patient records. The predictor variables were the origin and size of the lesion. The primary outcome variable was concordance between the provisional incisional biopsy diagnosis and definitive pathologic resection diagnosis. The secondary outcome variable was type of biopsy error for the discordant cases. Incisional biopsy errors were assessed and grouped into 5 categories: 1) sampling error; 2) insufficient tissue for diagnosis; 3) presence of inflammation making diagnosis difficult; 4) artifact; and 5) pathologist discordance. A total of 272 patients met the inclusion criteria. The study sample had a mean age of 47.4years and 55.7% were women. Of these cases, 242 (88.9%) were concordant when comparing the biopsy and final resection pathology reports. At histologic evaluation, 60.0% of discordant findings were attributed to sampling error, 23.3% to pathologist discrepancy, 13.3% to insufficient tissue provided in the biopsy specimen, and 3.4% to inflammation obscuring diagnosis. Overall, concordant cases had a larger average biopsy volume (1.53cm(3)) than discordant cases (0.42cm(3)). The data collected indicate an 88.9% diagnostic concordance with final pathologic results for incisional oral biopsy diagnoses. Sixty percent of discordance was attributed to sampling error when sampled tissue was not representative of the lesion in toto. Multiple-site biopsy specimens and larger-volume samples allowed for a more accurate diagnosis.