Journal of the New Zealand Medical Association, 08-June-2012, Vol 125 No 1356
Computed tomographic colonography (CTC): a retrospective analysis of a single site experience and a review of the literature on the status of CTC
Marcus Ghuman, Ngaire Bates, Helen Moore
Colorectal cancer (CRC) is the second most common cause of cancer death in New Zealand,1 and we have amongst the highest age-standardised rates of the disease in the world. Barium enema and colonoscopy have been the traditional investigations used in the work up of patients presenting with symptoms suggestive of CRC.2 Increasingly, computed tomographic colonography (CTC) is displacing barium enema as a non-invasive rapid imaging technique to investigate these patients, which can be cost effective, and as accurate as colonoscopy for colorectal cancer detection.3
While it is evident that the sensitivity of CTC for detection of polyps over 10mm is generally equivalent to that of colonoscopy, as seen in local4 and international clinical trials,5–8 areas requiring consensus remain. These areas particularly include the reporting, management, and follow up smaller polyps; and also issues surrounding extracolonic findings. Radiation exposure issues will also be discussed.
Study design—This observational study was conducted as a retrospective analysis using as a population all the patients undergoing CT colonography (CTC) for the 3-year period from 9 August 2007–12 August 2010 conducted at a single site; Greenlane Hospital, which is the outpatient hospital for Auckland District Health Board (ADHB), and currently the only ADHB site performing CTC in public.
These patients were identified through Picture Archiving and Communication System (PACS) coding, along with their corresponding National Health Index (NHI) numbers. This dataset also included the date of referral and performance of examination, and the referral source. Each patient’s CTC report was obtained from the Auckland District Health Board software package “Concerto”, where demographic data (sex, ethnicity, age), symptoms leading to referral, findings of the examination (intra and extracolonic), and recommended follow-up were extracted.
Each CTC study was performed on a Phillips 16-slice CT scanner typically using standard full bowel preparation (LoSo Prep, E-Z-M), although some more frail patients had reduced preparation or only fecal tagging. All had fecal tagging (Tagitol) and iv Buscopan unless contraindicated. Supine and prone scans, and occasional supplementary decubitus scans were performed at 120kvP or 90kvp, and mAs 50-150. Colonic distension was primarily using CO2 insufflation (ProtoCO2l, Bracco Diagnostics), or occasionally manual air insufflation.
The CTC reports were coded according to the CT Colonography Reporting and Data System (CRADS), as defined by the Working Group for Virtual Colonoscopy, 2005.9 The vast majority of the scans in this study were reviewed by two consultant radiologists with considerable experience in CT colonography reporting, suggesting that the reporting in this sample was robust.
Statistical analysis—Categorical data were presented as frequency (percentage) and continuous data were presented as mean (standard deviation). Categorical variables were compared using a Chi-squared test or Fisher’s exact test as appropriate. Continuous data were compared using the t-test and the one-way Anova test. All p values reported were two-tailed and a p value <0.05 was considered significant. SAS (version 9.1) statistical software was used for statistical analysis.
This study identified 302 patients undergoing CT colonography (CTC), of whom 184 were female (61%). The mean age of patients in the study population was 66 years (range 16–91 years). New Zealand European (59%), Asian (17%), and Other Europeans (10%) made up the majority of the study population. The demographic data of the study population are recorded in Table 1.
Table 1. Demographic data of the study population
Nearly all of the referrals for CTC came from four sources: General Practice (121 referrals), Gastroenterology (103 referrals), General Surgery (60 referrals), and General Medicine (9 referrals); 57 patients, which accounted for 19% of the study population, were referred for CTC following a failed optical colonoscopy (OC).
The average time from referral to performance of the scan was 43 days, with patients referred from General Medicine and those referred following a failed OC showing a trend toward shorter referral time, though these findings did not reach significance.
During the study period there was no formal priority criteria used to stratify the referrals, however they consisted of symptomatic patients, mostly of low to medium risk category. These findings are summarised in Table 2.
Table 2. Referral base and referral times by referrer for CT colonography
NB: OC is optical colonoscopy; referral time refers to the time from referral for CT colonography to performance of the scan.
CT colonography identified 12 patients as having colorectal cancer (4% of the total study population), 24 as having polyps over 5 mm (8%), and 111 with diverticular disease (37%).
The majority of patients identified as having cancer were of New Zealand European ethnicity, a statistically significant finding (p=0.017). Though there was a trend toward more females than males having cancer, and a higher average age for the cancer cohort as compared to the study population, these findings did not reach significance.
Of the 24 patients identified as having polyps, 13 were male and 11 female. The majority of the polyp patients were also of New Zealand European ethnicity (p=0.0002).
More females than males were found to have diverticular disease (p=0.013), and again patients of New Zealand European ethnicity dominated this cohort (p=0.0001). Table 3 summarises the cancer, polyp, and diverticular disease findings.
Follow-up recommendations in the formal CTC reports of patients discovered to have colorectal cancer varied from no follow-up advice, to recommendation for referral to colorectal multidisciplinary meeting (MDM) and direct visualisation and biopsy via optical colonoscopy.
Follow-up recommendations for patients with polyps varied, but included repeat CTC (with interval duration ranging from 1–5 years) and/or optical colonoscopy.
In total 21 patients (7% of the study population) were referred for optical colonoscopy due to the detection of malignancy, polyps, or for other reasons including poor quality of study, following their CT colonography. In total, 3 patients (1% of the study population) had inadequate bowel imaging at their CTC, as indicated by the C0 notation.
Table 3. Cancer, polyp, and diverticular disease findings
NB: One of the patients identified as having cancer did not have ethnicity recorded.
Of the 57 patients who were referred for CTC following a failed optical colonoscopy, two were found to have cancer, two had polyps over 5 mm, and 25 had diverticular disease.
Although approximately 50% of patients in the study population had some mention of extracolonic findings in their formal CTC report, most were benign incidentals, and only 34 patients (11% of the study population) had a recommendation made in their CTC report for further imaging as a result of the extra-colonic findings. The number of patients who actually went on for further imaging may well be less than this.
Nine patients (3%) were coded as E4 (potentially important extracolonic finding) and were made up of three abdominal aortic aneurysms, three pulmonary lesions suggestive of malignancy, two patients with evidence of disseminated metastatic malignancy, and one porcelain gallbladder.
A further 54 patients (17.8% of the study population) were coded as E3 (likely unimportant extracolonic finding, incompletely characterised, therefore possibly requiring further investigation). These were a heterogeneous collection of pathologies, with some of the more common being renal stones and cysts, small pulmonary nodules, and pelvic/gynaecological cysts.
Of the follow-up recommendations, 3 were for GP follow-up, 7 for specialist follow-up, and 24 for further imaging (comprising 16 USS, 7 CT scans, and 1 MRI scan).
This observational study suggests that, for the patient population undergoing CT colonography, Europeans and Māori were proportionately under-represented, while Asians were over-represented, as compared to Auckland population demographics.
The average time from referral for CT colonography to performance of the exam was just over 6 weeks. This is within acceptable guidelines for the current Northern Cancer Network Regional Prioritisation criteria for priority 2, 3, or surveillance patients.10
This study found that patients of New Zealand European ethnicity had higher rates of colorectal cancer (p=0.0117) and polyp detection (p=0.0002) as compared to other ethnic groups, both significant findings. This finding is consistent with data from the New Zealand Cancer Registry.11
Diverticular disease was shown to be more common in females than males (p=0.0132), and in patients of New Zealand European ethnicity (0.0001). These findings are well established in the literature.8 The prevalence of colorectal cancer in this symptomatic group, at around 4%, is similar to reported rates from other CTC studies in New Zealand.12
A limitation of this study is the lack of formal comparison of CTC findings with those who had colonoscopic assessment; and further work is being undertaken in this regard. Interestingly there is quite limited data available on the verified performance characteristics of both CTC and Optical Colonoscopy for detection of colorectal cancer in NZ. Two studies from Canterbury, assessing the miss rate of Colonoscopy and CTC using the NZ National Cancer Registry as a reference, showed equivalent miss rates of around 5-6%.2,12 The follow-up advice in patients with intermediate sized polyps (6mm–9mm) varied, with the most common advice being repeat CT colonography or direct visualisation via optical colonoscopy. The recommended interval length varied from 1 to 5 years, with a tendency to shorter intervals for larger polyps.
This highlights one of the discussion topics with CT colonography—specifically what size polyp should be referred for OC and polypectomy and/or followed with repeat CTC. Although polyp natural history is complex and not perfectly understood, there is a large amount of evidence from pathological and colonoscopy-based studies showing that apart from actual histology showing the presence of significant dysplasia or villous change, the most dominant predictor of behaviour is polyp size.12
A recent colonoscopic study of 1468 patients found that of 414 polyps smaller than 10mm, only 41 (9.9%) were advanced adenomas, and 1.7% were high-grade neoplasia. None were frankly malignant. Polyp size was the only identified risk factor for the presence of advanced adenoma.13 There is general consensus that patients with polyps greater than 10 mm should be referred for colonoscopy, while the vast majority of diminutive polyps (those 5 mm and smaller) are hyperplastic and do not require compulsory removal. The rate of advanced histology in the diminutive polyp group is particularly low, with high grade neoplasia or malignancy being extremely rare reported from zero to 0.06%.15
Thus the most controversy remains regarding what should be done with small to medium sized (6–9 mm) polyps and how they should be followed up, as there is a lack of consensus across the expert groups involved (gastroenterological, surgical and radiological) regarding decision making in this area.16 The radiologist authors utilised the CRADS consensus, which gives the option of endoscopic polypectomy or follow-up CTC for patients with these polyps of indeterminate size. It also recommends consultation and adaptation in regard to local standards of practice and patient preference.
In our opinion, this approach allows common sense to prevail for individual patient circumstances, and existing data suggest this risk is manageable.15,17 For example an 8mm polyp in a younger or fit person would be actively dealt with compared to a similar polyp in an elderly or frail individual with other comorbidities. Unnecessary polypectomies have repercussions in economic terms and on patient morbidity and mortality.17
Only 7% of patients undergoing CT colonography were referred on for optical colonoscopy in the study, which is similar to rates found elsewhere.3,15 It is important to note that the higher the rate of on-referral to colonoscopy, then the higher the cost of a CT colonography diagnostic approach. Conversely the cost of a primary optical colonoscopy approach is increased by its “fail” rate, with subsequent CTC necessary.
With reference to incidental extracolonic findings discovered at CTC, a key consideration is the need to weigh the benefit of an earlier diagnosis of a potentially important finding against the increased patient anxiety and possible morbidity that necessarily results from a finding, and the cost of further workup.
Although there was some mention of extracolonic findings in close to 50% of the reports, only a minority of patients had potentially important extracolonic findings, and only 34 patients (11% of the study population) had recommendation for some form of follow up for their extracolonic findings, including 24 follow-up scans. Of those 24 recommended follow-up scans, two-thirds of these were for USS, a safe and comparatively cheap investigation. The number of patients who actually went on for further imaging may well be less than the number for whom it was recommended. These results compare favourably with those found elsewhere, with rates of suggested follow-up for extracolonic findings of CT colonography ranging from 11–20%.18,19
Although this study has not generated any data on radiation exposure from CT colonography, this remains an important consideration when considering this investigation as an alternative to optical colonoscopy. Due to technical factors, individual radiation exposure could not be determined in this audit, with only generic data being supplied by the particular CT machine.
Studies attempting to find a correlation between radiation exposure and subsequent cancer risk in later life have relied on modelling rather than direct observation,6 and thus there is no direct data relevant to CT colonography in this area. However, most CT colonography protocols deliver about 3–8 millisieverts of radiation—which is a relatively low dose, and usually less than half that used in a standard CT abdomen.8 The risk from this radiation dose is negligible in a symptomatic adult, relative to their background risk of cancer which is at least 30% lifetime risk.20
The safety profile of CTC is excellent, with no perforations or other complications in this cohort.
This observational study has reported on the experience of CT colonography at Greenlane Hospital over a 3-year period. Though no attempt at direct comparison with optical colonoscopy in regards to efficacy and safety have been made, it has provided important local data on rates of detection of colonic pathology. Key issues such as follow-up advice for small–medium sized polyps, and investigation of extracolonic findings have been discussed.
One of the most important findings for primary care practitioners is the under-representation of Māori and Pacific Island patients referred for bowel symptoms.
Author information: Marcus Ghuman, Radiology Registrar, Radiology Department, Auckland Hospital, Auckland; Ngaire Bates, Consultant Radiologist, Radiology Department, Auckland Hospital, Auckland; Helen Moore, Consultant Radiologist, Radiology Department, Auckland Hospital, Auckland
Correspondence: Dr Marcus Ghuman, 93 Tiki Road, RD2, Te Awamutu 3872, New Zealand. Email: firstname.lastname@example.org
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads)
Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals