Joseph C Lee1, Gemma F Hartnett2, Aravind S Ravi Kumar3. 1. Department of Medical Imaging, The Prince Charles Hospital, Chermside, Herston, Queensland, Australia ; School of Medicine, University of Queensland, Herston, Queensland, Australia. 2. Department of Medical Oncology, Redcliffe General Hospital, Redcliffe, Queensland, Australia. 3. Department of Nuclear Medicine and Specialised PET Services, Royal Brisbane and Women's Hospital, Herston, Queensland, Australia.
Dear Editor,We applaud the efforts of Liu et al.[1] in attempting to make recommendations for specific patients who should undergo colonoscopy following the finding of unexpected colorectal incidentalomas (CIs) avid for fluorodeoxyglucose (FDG) on positron emission tomography with computed tomography (PET/CT). In the largest meta-analysis to date[2] involving almost 90,000 patients, it was concluded that colonoscopy is warranted for all CIs. However, the required time and costs of this recommendation would be huge. Hence, the attempt to prioritize certain patients within a population is commendable. After all, an additional malignant (or premalignant) diagnosis may significantly impact prognosis and treatment.In addition to the specific PET/CT population of interest identified by Liu et al.,[1] we wonder if the selection of patients would be further refined by two factors. First, the statistics are more likely to be meaningful in a population of patients who underwent PET/CT for a clinical indication and then colonoscopy for follow-up incidental findings. Simply matching patients who underwent both investigations would inevitably skew the pretest probabilities of significant findings.Second, the precise colonic segment of interest is important. We certainly benefited from both these factors being central to our own work,[3] finding that proximal colon lesions had a much higher positive predictive value (PPV) for clinically significant lesions (such as malignancies and premalignancies). It was unclear if anatomical location was considered or if whole colon analysis was done in the cases. Whole colon analysis may be difficult to interpret as there remains an uncertainty about the PPV being 100% (i.e., we cannot be sure that the FDG-avid lesion is actually the cancer rather than something else). In short, could segmental analysis further allay concerns about the indiscriminate use of colonoscopy for this purpose?