Periradicular Lesions in Cancellous Bone Can Be Detected Radiographically.

INTRODUCTION

The purpose of this study was to evaluate whether naturally occurring periradicular lesions confined to cancellous bone can be detected on periapical digital radiography and whether the size of the lesion had any effect on lesion detection.

METHODS

One hundred twenty-nine roots were chosen based on cone-beam computed tomographic imaging and categorized as having no lesion, a lesion confined to cancellous bone, a lesion that encroaches on junctional corticocancellous trabeculae, or a lesion with cortical involvement. The largest buccolingual dimension of the lesions was measured on cone-beam computed tomographic imaging. Two observers separately viewed the corresponding periapical radiographs in their original version as well as in the edge-enhanced setting on MiPACS (LEAD Technologies Inc, Charlotte, NC). Observers were asked to evaluate and interpret the periapical radiographs as having a lesion present, absent, or "unsure." Evaluations of images were conducted at 2 times 1 week apart. Data were analyzed, and the level of significance was set at P = .05.

RESULTS

Lesion size, not the degree of cortical involvement, significantly affected the observers' ability to accurately detect lesions. As the size of the lesion increased, the probability of correctly identifying a lesion increased (P = .0008). Lesions were correctly identified 97.6%, 94.1%, 91.6%, and 89.3% of the time, respectively, when in cortical bone, at the junction of corticocancellous bone, in cancellous bone, and when no lesion was present. Observers were "unsure" whether a lesion was present or absent 10.7% of the time. Only lesions in cortical bone significantly increased observers' certainty in making a diagnosis.

CONCLUSIONS

This study concluded that lesions confined to cancellous bone can be detected radiographically at a high rate. Lesion size was positively correlated with correct lesion identification, whereas the degree of cortical involvement had no significant effect. This suggests that lesion size may be a better predictor for correct lesion identification than lesion location.