RESUMO
BACKGROUND: This study aimed to evaluate the utility of 18-fluorodeoxyglucose positron emission/computed tomography (PET/CT) in the management of colorectal cancer (CRC) patients with suspected recurrence. METHODS: Clinical and imaging histories of CRC patients who underwent PET/CT at our institution between 1 April 2007 and 31 August 2008 for evaluation of recurrent disease were retrospectively reviewed. Patients were divided into two groups - (A) patients evaluated for suspected local recurrence (based on conventional imaging) and (B) patients with carcinoembryonic antigen (CEA) levels in excess of 5 ng/mL; in whom conventional imaging was either normal or equivocal. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were derived using either histopathology or follow-up imaging as the standard of reference. RESULTS: In group A (n = 44), the sensitivity, specificity, PPV, NPV and accuracy of PET/CT were 100% (95% confidence interval (CI): 86.7-100), 84.2% (95% CI: 62.4-94.5), 89.3% (95% CI: 72.8-96.3), 100% (95% CI: 80.6-100) and 93.2%, respectively. In group B (n = 18), the sensitivity, specificity, PPV and NPV of PET/CT were 76.9% (95% CI: 49.7-91.8), 60.0%, (95% CI: 23.1-88.2), 83.3% (95% CI: 55.2-95.3) and 50% (95% CI: 18.8-81.2), respectively. CONCLUSION: PET/CT has high accuracy in the assessment of local recurrence, particularly with regard to its NPV. PET/CT is useful for problem solving in cases of unexplained elevated CEA levels.
