PURPOSE: The aim of this study was to evaluate the interpretations of incidental colonic 18F-FDG uptake made by 10 experienced readers and to more clearly identify the pattern of suspicious colonic FDG uptake. The potential contributions of delayed FDG-PET scanning and of immune fecal occult blood testing (FOBT) in making a diagnosis were also analyzed. MATERIALS AND METHODS: Visual interpretations by 10 readers were made for 147 FDG uptake sites from 126 PET scans (cancer, 38 sites; adenoma, 43 sites; and no abnormality, 66 sites) with colonic FDG uptake. Assessments for the early FDG-PET images were (1) FDG uptake pattern, (2) FDG uptake degree, and (3) likelihood of malignancy. For the delayed images, the assessments were (1) change in the FDG uptake position, (2) change in FDG uptake degree, and (3) likelihood of malignancy. The results of FOBT were analyzed independently of the visual interpretations. RESULTS: Interobserver agreement (κ) was 0.501 for assessing FDG uptake patterns, while agreement on assessing changes in uptake degree and changes in uptake position between early and delayed imaging were low (κ = 0.213-0.229). Logistic regression analysis indicated that 'FDG uptake patterns' and 'FDG uptake degree' were significantly related to decide on the suspicion of malignancy (p < 0.001) and the final result (p < 0.001). "Small localized" and "large irregular localized" types had a high probability of a lesion regardless of either (1) FDG uptake degree or (2) variation in the uptake between the early and the delayed image. The delayed image decreased false-positive cases for some FDG uptake patterns, but it had little impact on distinguishing clearly between "cancer or adenoma" and "normal". The addition of FOBT had little impact on the diagnosis. CONCLUSION: There was highest agreement among readers with respect to the recognition of specified colonic FDG uptake patterns, and this pattern recognition had the most influence on the diagnosis. "Small localized" and "large irregular localized" types had a high probability of a lesion. The addition of delayed imaging and of FOBT results to the early imaging did not have much impact on the diagnosis.
PURPOSE: The aim of this study was to evaluate the interpretations of incidental colonic 18F-FDG uptake made by 10 experienced readers and to more clearly identify the pattern of suspicious colonic FDG uptake. The potential contributions of delayed FDG-PET scanning and of immune fecal occult blood testing (FOBT) in making a diagnosis were also analyzed. MATERIALS AND METHODS: Visual interpretations by 10 readers were made for 147 FDG uptake sites from 126 PET scans (cancer, 38 sites; adenoma, 43 sites; and no abnormality, 66 sites) with colonic FDG uptake. Assessments for the early FDG-PET images were (1) FDG uptake pattern, (2) FDG uptake degree, and (3) likelihood of malignancy. For the delayed images, the assessments were (1) change in the FDG uptake position, (2) change in FDG uptake degree, and (3) likelihood of malignancy. The results of FOBT were analyzed independently of the visual interpretations. RESULTS: Interobserver agreement (κ) was 0.501 for assessing FDG uptake patterns, while agreement on assessing changes in uptake degree and changes in uptake position between early and delayed imaging were low (κ = 0.213-0.229). Logistic regression analysis indicated that 'FDG uptake patterns' and 'FDG uptake degree' were significantly related to decide on the suspicion of malignancy (p < 0.001) and the final result (p < 0.001). "Small localized" and "large irregular localized" types had a high probability of a lesion regardless of either (1) FDG uptake degree or (2) variation in the uptake between the early and the delayed image. The delayed image decreased false-positive cases for some FDG uptake patterns, but it had little impact on distinguishing clearly between "cancer or adenoma" and "normal". The addition of FOBT had little impact on the diagnosis. CONCLUSION: There was highest agreement among readers with respect to the recognition of specified colonic FDG uptake patterns, and this pattern recognition had the most influence on the diagnosis. "Small localized" and "large irregular localized" types had a high probability of a lesion. The addition of delayed imaging and of FOBT results to the early imaging did not have much impact on the diagnosis.