Accuracy of minimal-preparation computerized tomography in the investigation of suspected colorectal cancer in frail patients: the experience of Western General Hospital,Edinburgh

Introduction

Colorectal carcinoma (CRC) is the third most common cancer in the world, after lung and breast cancer, with an estimated 1.23 million new cases diagnosed in 2008. It is the third most common cancer in men with 663,000 cases and 570,000 cases in women worldwide. ¹ The number of new cases of colorectal cancer worldwide has been increasing rapidly since 1975. ² In the UK, it is the second most common cause of cancer-related death and the third most common cancer, affecting more than 30,000 people each year with an average five-year survival rate of 40%. ³ It is a disease of advancing age with a greater than 10-fold increase in the incidence beyond 65 years of age compared with that under 65 years of age. ⁴

The investigation of the elderly, frail and immobile patient who presents with signs and symptoms suspicious for a colorectal malignancy is challenging as these patients frequently have difficulty tolerating formal ‘gold standard’ investigations such as barium enema and colonoscopy. The limitations also extend to the examination of the caecal region, which is the most difficult part to reach retrogradely and is prone to residual faecal contamination.^5,6 While computerized tomography (CT) pneumocolon with virtual colonoscopy has been demonstrated to be more comfortable, and thus more acceptable, compared with colonoscopy and double-contrast barium enema, and it still requires bowel preparation, which is distressing to most patients regardless of the method of investigation. ⁷

In recent years, a number of papers have studied the use of minimal-preparation CT (MPCT) for the diagnosis of colorectal cancer in the elderly population.^8–11 MPCT was first described by Day et al. in the early 1990s as an alternative technique for the imaging of the large bowel in frail elderly patients with suspected colorectal pathology who are referred for barium enema. The basic premise is that the detection of colonic cancer is important while the presence of polyps were of less clinical relevance. ⁸ Comparative studies have demonstrated MPCT to be reasonably sensitive (75–96%) and specific (82–94%) in detecting colonic malignancy.^8,9

In this paper, we evaluate our experience using MPCT to determine if the sensitivity and specificity are replicable and in line with current published data. Table 1 discusses the pros and cons of MPCT, CT pneumocolon, barium enema and colonoscopy.

OPEN IN VIEWER

Table 1

Comparison of CT pneumocolon with virtual colonoscopy, MPCT, double-contrast barium enema and colonoscopy

Pros	Cons
CT pneumocolon with virtual colonoscopy
High-sensitivity and high-specificity polyps >1 cm	Need for bowel preparation
Assessment of extracolonic findings	Insufflation can be poorly tolerated
Allows staging at time of diagnosis
Double-contrast barium enema
Full colonic visualization	Need for bowel preparation
Low perforation risk	Poorly tolerated in the elderly
	Sensitivity and specificity less than other colonic investigations
Colonoscopy
Gold standard investigation	Caecal intubation rate failure 5–30%
Ability to obtain histological analysis	Need for sedation
	Need for bowel preparation
	Perforation risk 1 in 500–4000
MPCT
Well tolerated by frail and elderly	Poor accuracy for polyps <1 cm
No bowel preparation
Assessment of extracolonic findings

Methods

Dedicated MPCT reports of all patients between May 2005 and June 2008 were reviewed retrospectively. The results were obtained using APEX i. Lab services (Computer Sciences Corporation, Banbury, UK) and case-notes in order to determine pathological correlation, if any, and subsequent follow-up for any findings for an average period of 22 months (range, 14–24 months). The patient protocol for MPCT used in our institution utilizes 18 mL of meglumine and sodium amidotrizote (Gastrografin; Bayer PLC, Newbury, UK) diluted in 1.5 L water and taken in three divided doses in the 48 hours prior to the CT. A further 20 mL of gastrograffin in 900 mL was administered prior to the CT scan. No bowelcleaning agents or air insufflation was used. Intravenous contrast was also administered via pump injection. Helical CT acquisition of the abdomen and pelvis was then obtained using SOMATOM Sensation 16 (Siemens, Erlangen, Germany) from the domes of the diaphragm to the symphysis pubis utilizing 1.5 mm slice collimation, with 8 mm width and 2 mm reconstruction.

The MPCT findings were categorized into definite colonic malignancy, possible malignancy or no malignancy. A confirmed diagnosis was given following correlation with biopsy-proven colonic malignancy or polyp containing high-grade dysplasia or if there was overwhelming clinical evidence (e.g. colorectal mass with evidence of metastatic disease). For those with normal MPCT, the findings were deemed to be true negatives if no corresponding positive pathology demonstrated in case-note and results review.

False-positives were evaluated in terms of findings and adequacy of bowel preparation in order to determine if this had any bearing on results. The MPCT were also evaluated in terms of extracolonic pathology and the outcomes were reviewed with regards to extracolonic pathology and subsequent follow-up. Results were calculated using a standard two-by-two table with sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) as measures of diagnostic accuracy.

Results

A total of 85 patients (59 women and 26 men) were included in the retrospective review. The age range was from 55 to 99 years of age with an average age of 82 and a median age of 84 years.

Overview

A total of 17 patients were considered to have features consistent with CRC or ‘possible’ CRC and, therefore, were considered positive. A further 68 had no colonic findings and, therefore, were considered negative.

At the end of follow-up, 12 of the 17 positives had features consistent with a definite colorectal carcinoma. The remaining five were identified as having ‘possible’ malignancy; however, at subsequent investigation, they were diagnosed as having no malignancy (i.e. false-positive). Of the 68 patients considered negative on the MPCT, on follow-up, none were found to have colorectal carcinoma. These findings are summarized in Table 2.

OPEN IN VIEWER

Table 2

Tabulated findings of the MPCT study cohort

	CRC positive	CRC negative	Total
Positive MPCT	12 (TP)	5 (FP)	17
Negative MPCT	0 (FN)	68 (TN)	68
Total	12	73	85

FN, false-negative; FP, false-positive; TN, true negative; TP, true positive

The disease prevalence in this cohort was calculated at 0.14. The sensitivity was calculated at 1 (95% confidence interval [CI], 0.69–1) and the specificity 0.93 (9% CI, 0.84–0.97). The PPV was found to be 0.71 (95% CI, 0.44–0.88) and the NPV was 1 (95% CI, 0.93–1).

MPCT findings

Of the 12 positive cases, the majority was found in the caecum (43%), which is the most difficult to evaluate with colonoscopy or barium enema. The remainder of the distribution is summarized in Table 3. Of these, one of the 12 was found to have features of nodal spread. None of these 12 was found to have distant metastasis.

OPEN IN VIEWER

Table 3

Positive MPCTs and outcomes

True-positives
Gender	Age	Location	Diagnosis	Treatment
Male	89 Ascending colon	Adenocarcinoma, Dukes B pT3N0M0	Surgery
Male	91	Caecum	Adenocarcinoma	Palliative care
Male	86	Sigmoid	Adenocarcinoma, Dukes B pT3N0M0	Surgery
Male	87	Transverse colon	Adenocarcinoma	Palliative care
Female	69	Caecum	Adenocarcinoma, Dukes C pT4N1Mx	Surgery
Female	76	Caecum	Adenocarcinoma, Dukes B pT3N1M0	Surgery
Male	78	Caecum	Adenocarcinoma	Imaging follow-up
Male	78	Transverse colon	Adenocarcinoma	Palliative care
Male	80	Caecum	Adenocarcinoma, Dukes B pT3N1M0	Surgery
Female	74	Sigmoid	Adenocarcinoma, Dukes C pT3N2M0	Surgery
Female	99	Ascending colon	Adenocarcinoma, Dukes B pT4N1M0	Surgery
Female	88	Transverse colon	Adenocarcinoma	Palliative care
False-positives

Gender	Age	Location	Further investigation
Female	87	Sigmoid	Diverticular stricture, sigmoidoscopy
Female	84	Ascending colon	Collagenous colitis, colonoscopy and biopsy
Male	76	Caecum	Nil on barium enema
Female	82	Sigmoid	Nil on barium enema
Female	55	Sigmoid	Colonoscopy – intraepithelial colitis

Positive MPCTs were divided into true-positive and false-positive outcomes showing spread of age, sex and eventual diagnosis of the positive MPCT patients

Of the 12 patients, seven underwent curative resection with four patients receiving palliative treatment. The remaining patient was considered too frail for treatment and underwent further imaging follow-up.

Of the false-positives, three were ruled out at colonoscopy/sigmoidoscopy and the remaining two underwent barium enemas. Two were found to have colitis (collagenous and intraepithelial colitis), one had findings due to diverticular disease and the remaining two had no mucosal abnormality and the uncertainty was likely due to poor opacification of the bowel with the oral contrast (Figures 1 and 2).

OPEN IN VIEWER

Figure 1

Coronal (left) and axial (right) MPCT demonstrating ascending colon thickening. The patient subsequently underwent a laparoscopic right hemicolectomy. This was Dukes’ B adenocarcinoma of the colon on the histology. Note the incidental 5.3-cm abdominal aortic aneurysm

OPEN IN VIEWER

Figure 2

Axial MPCT demonstrating transverse colon thickening. By biopsy this was proven to be an adenocarcinoma. The patient subsequently underwent palliative treatment

Extracolonic findings

A total of 26 of the 85 patients studied (30%) were found to have extracolonic pathology, which required further investigation or intervention were identified. This included a total of four extracolonic malignancies, which were pathologically confirmed on subsequent investigations (2 upper gastrointestinal, 1 ovarian and 1 lung adenocarcinoma). The findings are summarized in Table 4 (Figure 3).

OPEN IN VIEWER

Table 4

Extracolonic findings within the study population

Pathology	No. of patients
Malignancy
Upper gastrointestinal	2
Ovarian	1
Lung	1
Aneurysm
AAA	3
Thoracoabdominal	1
Pulmonary embolism	2
Midgut volvulus	1
Hernia
Femoral	1
Spigelian	1
Liver lesion requiring investigation	2
Ovarian pathology	2
Biliary pathology (including gallstones)	8
Bladder calculi	1
Total	26

AAA, abdominal aortic aneurysm

OPEN IN VIEWER

Figure 3

Incidental right lower lobe mass found on a MPCT study. This was subsequently biopsied and histology revealed a lung adenocarcinoma

Discussion

The investigation of the frail, elderly or infirm patient using the previously used ‘gold standard’ investigations of colonoscopy and barium enema is fraught with problems in terms of bowel preparation and patient tolerance. With regard to MPCT versus CT colonoscopy, MPCT does not require air insufflation or bowel preparation, which is better tolerated by the elderly patient. MPCT is a well-tolerated method with a high sensitivity and specificity with the added benefit of identifying extracolonic malignancies or pathology that may contribute to the patient symptoms. ¹² While it is not sensitive enough to detect small colonic lesions or polyps less than 1–2 cm in size compared with CT pneumocolon or colonoscopy, although these may have malignant potential and may require removal in the younger patient cohort, polyps may take an estimated 5–8 years to progress to invasive carcinoma ¹³ and, therefore, is of less clinical relevance in the frail, elderly and infirm population. ⁸

While the study cohort at the Western General Hospital undergoing MPCT has been relatively small compared with studies based in Oxford and Cambridge, our results have yielded a favourable comparison with previously published data. The sensitivity and specificity have been well within the accepted range. The extracolonic malignancy detection rates have also been favourable with four extracolonic tumours detected, including one lung adenocarcinoma (Table 5).

OPEN IN VIEWER

Table 5

Comparison of the results from the Western General Hospital with published studies on MPCT from 1993 to 2007

Studies	No. of patients	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	Extra-colonic malignancy (%)
Day et al. 8	37	88	86	64	96	3.3
Domjan et al. 9	118	75	96	75	96	2.5
Robinson et al. 16	195	100	87	46	100	5.6
Ng et al. 10	1031	85	91	49	98	2
Kealey et al. 11	72	75	87	43	97	1.4
Ganeshan et al. 12	391	95	88	45	99	3.6
Western General Hospital (2008)	85	100	93	71	100	4.7

This demonstrates that the results from the Western General Hospital are comparable with currently published data

While the data with regard to the sensitivity have a wide 95% CI, this may be due to the difficulty in ‘establishing’ missed cancer in the study cohort as confirmatory tests were not generally sought in the aged population with CT-negative examinations. It is interesting to note, however, that while the patients were considered too frail for investigation by colonoscopy or CT colonography, one out of the 12, more than 50%, underwent resection for their colonic malignancy. This may be due to the advent of laparoscopic-assisted resection and enhanced recovery programme, improving perioperative survival.^14,15

With the advent of the bowel screening programme in Lothian that began in May 2008 while MPCT has not incorporated as a second-line investigation following failed colonoscopy, it should be considered as a valid adjunct, especially for the infirm or frail patient within the age 50–74 years who are currently being offered participation in the bowel screening programme.

Conclusion

MPCT is a non-invasive and well-tolerated technique for investigation of CRC in the elderly or frail patient with the added benefit of detecting clinically relevant extracolonic pathology. The sensitivity and specificity of the MPCTs carried out at the Western General Hospital, Edinburgh, have been found to be comparable with the current published data.