Reliability of histologic diagnosis of clinically normal intraoral tissue adjacent to clinically suspicious lesions in former upper aerodigestive tract cancer patients.

Oral lesion biopsies often include adjacent clinically normal tissue to assess lesion spread, but the reliability of adjacent tissue histologic diagnoses has not been reported. Sixty-seven patients presenting with subsequent suspicious intraoral lesions underwent tissue staining with tolonium chloride. Biopsies of unstained tissue adjacent to suspicious lesions were reviewed by independent pathologists. The weighted kappa (kappa(w)) for histologic classification was 0.75 (95% confidence interval [CI] 0.64, 0.86) and was lower for: inflamed specimens [kappa(w)=-0.13 (95% CI -0.30, 0.05)] versus noninflamed specimens [kappa(w)=0.81 (95% CI 0.69, 0.92)], wedge biopsies [kappa(w)=0.58 (95% CI 0.37, 0.79)] versus punch biopsies [kappa(w)=0.90 (95% CI 0.83, 0.97)], and current smokers [w=0.59 (95% CI 0.32, 0.86)] versus past and never smokers [kappa(w)=0.82 (95% CI 0.72, 0.91)]. Histologic diagnosis of subsets of clinically normal oral specimens adjacent to suspicious tissue has low reliability. Molecular techniques may improve diagnoses in these subsets.