Background: The purpose of this study was to analyze the detectability of colorectal neoplasia with fluorine-18-2-fluoro-2-deoxy-D-glucose positron emission tomography/computed tomography (FDG-PET/CT).
Methods: Data for a total of 492 patients who had undergone both PET/CT and colonoscopy were analyzed. After the findings of PET/CT and colonoscopy were determined independently, the results were compared in each of the six colonic sites examined in all patients. The efficacy of PET/CT was determined using colonoscopic examination as the gold standard.
Results: In all, 270 colorectal lesions 5 mm or more in size, including 70 pathologically confirmed malignant lesions, were found in 172 patients by colonoscopy. The sensitivity and specificity of PET/CT for detecting any of the colorectal lesions were 36 and 98%, respectively. For detecting lesions 11 mm or larger, the sensitivity was increased to 85%, with the specificity remaining consistent (97%). Moreover, the sensitivity for tumors 21 mm or larger was 96% (48/50). Tumors with malignant or high-grade pathology were likely to be positive with PET/CT. A size of 10 mm or smaller [odds ratio (OR) 44.14, 95% confidence interval (95% CI) 11.44-221.67] and flat morphology (OR 7.78, 95% CI 1.79-36.25) were significant factors that were associated with false-negative cases on PET/CT.
Conclusion: The sensitivity of PET/CT for detecting colorectal lesions is acceptable, showing size- and pathology-dependence, suggesting, for the most part, that clinically relevant lesions are detectable with PET/CT. However, when considering PET/CT for screening purposes caution must be exercised because there are cases of false-negative results.