| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Commentary |
1 Department of Radiology, Unit 57, MD Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030-4009, USA and 2 Department of Radiology, Addenbrooke's Hospital, Cambridge CB2 2QQ, UK
Xiong and colleagues [1] are to be congratulated on an excellent and comprehensive review on the topic of incidental lesions detected in the course of undertaking CT colonography. Their task of summarizing the data should not be underestimated: many of the studies utilized differing patient populations, different CT techniques, and had varying definitions, explicitly or implicitly stated, of what constituted extracolonic findings worthy of reporting. For example, some studies have included simple cysts of the liver and kidneys in their data, while others have not and some considered cardiomegaly an important finding [2]. Their difficulties would have been compounded by inconsistent details in the various papers to enable extraction of these figures.
Nevertheless, a common pattern does appear to emerge, namely, that incidental or "extracolonic" findings in the course of CT colonography are common, averaging 40%. As such, this potentially brings considerable added value to an examination primarily designed to interrogate the colon.
Although the accuracy of CT colonography in the detection of colonic polyps has undergone careful evaluations, the question of the accuracy of the detection of extracolonic findings is less well explored. It is one thing to report or document a finding at CT, but quite another to be right about it. This becomes an increasing consideration in the context of the low dose and unenhanced CT techniques used in the majority of "virtual" colonography studies. This inevitably raises the issue as to how many of the incidental lesions or patients with incidental findings in the studies reported were followed up. It is clearly impossible to evaluate every incidental finding, particularly in those with multiple lesions. In our experience, follow up may occur in less than half the patients, especially in the elderly population. This unfortunately limits validation of the findings, and indeed some findings on further investigation may not be substantiated [3]. It is particularly intriguing that lesions that are considered "important" may not be investigated: patient and physician factors probably have a bearing, but a better understanding of such issues would be enlightening.
It is possible that the false positive rates related to incidental findings may have a substantial impact on the utility of detecting these lesions. A cascade of investigations can ensue following report of such findings, contributing to healthcare costs and patient anxiety. Furthermore, some of these investigations may be invasive and carry their own morbidity. As Xiong and colleagues [1] mention, these consequences need to be carefully evaluated.
Missing or failing to detect lesions (i.e. false negative rates) may also be a consideration. It can be extremely difficult to detect pancreatic, renal, or liver tumours with noisy unenhanced CT images. However, documentation of these failures would likely be extremely difficult to obtain. The category of "important incidental lesions found on follow-up" (Table 2 in Xiong et al [1]) may cast some light on this. Such considerations may be more important in some medical practice environments than in others.
In considering which patient populations might best benefit from the technology, Xiong and colleagues [1] importantly recognize that the study populations that have been reported fall into two broad categories: asymptomatic patients, in which the study is closer to a "screening" examination, and symptomatic patients, in whom the study is more of a "diagnostic" evaluation. A small subset of the latter group consists of studies on frail and elderly patients. One might expect a higher incidence of extracolonic findings in the older population partly because of age alone, but also because the patients in these studies were typically symptomatic, and the extracolonic disease detected might have contributed to, or been the dominant factor in, their presenting symptoms. The latter is quite intriguing and there has been some attempt to explore this [3], but it is admittedly extremely challenging to unravel. Interestingly, contrary to expectations, Xiong and colleagues [1] identify that fewer incidental findings have been reported in the frail and elderly group than in the studies with younger patients. Part of this might be due to varying reporting practices and differing thresholds for what constitutes findings worthy of reporting in these study populations and environments. There may also be an important contribution from the differing CT techniques that are typically used between the two groups; the "screening" studies tending towards a more "virtual", thinner collimation, technique.
It is possible that with ever improving CT techniques, smaller and smaller lesions may be detected, increasing the incidental lesion detection rate and compounding already existing difficulties about what should be done with the findings. This is comparable with the situation of finding incidental pulmonary nodules in the course of CT pulmonary angiography. Whether detection of such incidental findings has any impact on patient outcome is unknown. At least in the context of cancer and abdominal aortic aneurysms, one might intuitively expect early detection to do so.
In evaluating the true independent contribution of the technique, it is important to determine if the findings detected are truly incidental, with no preceding knowledge of the abnormality. Evaluation of the extracolonic findings is only one half of the equation when considering the utility of CT colonography as a whole, and when considering to which patient population the technique should be offered. Assessment of the extracolonic detection rate together with the spectrum and relevance of findings, needs to be combined with the colonic performance of the test. Importantly, Xiong and colleagues [1] identify that the incidental early cancer detection rate (0.9%) is comparable with the early colorectal tumour detection rate by this technique, and indeed may be higher than detection rates in other cancer screening programmes. Furthermore, the overall extracolonic cancer detection rate is even higher (2.7%). Thus this paper reinforces the power of CT over barium enema and colonoscopy for the evaluation of colorectal neoplasia, in that, at the same study it can examine multiple organs outside of the colon. Although, whether this is beneficial remains to be seen. The time is perhaps right for a carefully designed prospective study.
 |
References |
|---|
 Top References |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| BJR | DMFR | IMAGING | ALL BIR JOURNALS |
