Diagnostic accuracy in fine-needle aspiration of soft tissue and bone lesions. Influence of clinical history and experience.

Abstract

Interpretation of soft tissue and bone lesions with fine-needle aspiration (FNA) is difficult. We determined whether clinical history or experience improves diagnostic accuracy of FNA interpretation of these lesions. Four cytopathologists with varying degrees of experience retrospectively reviewed 89 soft tissue and bone lesions, initially without knowledge of clinical history and subsequently with knowledge of clinical history. Each time, the pathologist rendered a precise diagnosis (histogenetic classification, e.g., angiosarcoma, lipoma) for each case and classified the lesion into one of 4 categories: benign, probably benign, probably malignant, or malignant. The proportion of correct precise diagnoses was calculated. Also, the numbers of correct and incorrect classifications of malignancy were used in standard relative operating characteristic (ROC) curve analysis. The area under the ROC curve was estimated to give an indicator of diagnostic accuracy. Knowledge of clinical history increased the proportion of correct diagnoses and the accuracy of the classification for all observers. Experienced observers more accurately classified lesions, and knowledge of clinical history particularly improved the diagnostic accuracy of less experienced observers.