The Detection of Pulmonary Metastases with Pathologic Correlation in a Canine Model: Effect of Breathing on the Accuracy of Helical CT

OBJECTIVE. This study was undertaken to determine the effect of breathing on the accuracy of pulmonary nodule detection by helical CT. MATERIALS AND METHODS. Before sacrifice, four anesthetized dogs with metastatic osteosarcoma underwent helical CT with a collimation of 5 mm and a pitch of 2. Helical CT was performed during both induced breath-hold and normal quiet breathing. Images were reconstructed as contiguous 5-mm slices. Macroscopically evident metastases were noted postmortem. Hard-copy CT images were reviewed by 10 radiologists; each circled all suspected metastases. Helical CT images were compared with postmortem results to determine true- and false-positive diagnoses. RESULTS. One hundred thirty-two macroscopically evident pulmonary metastases were identified by pathologic examination. Of these metastases, the 10 radiologists identified an average of 40 metastases on breath-hold helical CT and an average of 36 on non-breath-hold CT. These findings were insignificant when analyzed by logistic regression for repeated measures (p = .8). CONCLUSION. In our animal model, helical CT performed during normal resting breathing resulted in no significant loss of accuracy in the detection of pulmonary metastases.