Colon: Neoplastic Disease: Radiologic Evaluation of Colon Cancer

Karen M. Horton MD, Anwar Padhani MD, Elliot K. Fishman, MD


Introduction

Colorectal cancer is a common malignancy, resulting in significant morbidity and mortality. It is the second most common cause of cancer death after lung cancer to westernised countries, accounting for 19,000 deaths each year in the UK and 85,000 deaths in the European Community (CRC Fact Sheet-1993). It is estimated that in 1996 there will have been over 130,000 new cases of colorectal cancer in the United States, and nearly 55,000 colorectal cancer deaths (Parker SL-1996). Although initial diagnosis is often made with colonoscopy, the air contrast barium enema continues to be an effective diagnostic tool. Other radiologic imaging modalities such as computed tomography (CT) have come to play a significant role in the staging of adenocarcinoma of the colon and rectum and in the detection of recurrent disease.



Barium Enema

The air-contrast barium enema is a safe, accurate and relatively inexpensive method for the detection of colorectal cancers. Detection rates for colon cancer with barium enema range between 75-95% (Stevenson G-1995, Beggs I-1983, Fork FT-1983). Although colonoscopy is thought be more sensitive than barium enema, detection rates as low as 85% have been reported with colonoscopy (Brady AP-1994, Glick SN-1989). Small polyps (6mm) are often not detected with barium enema, but can also be overlooked with colonoscopy. These polyps however, are usually of little clinical significance. In addition, complication rates are significantly higher with colonoscopy than with barium studies. There continues to be active debate over the use of barium enema vs. colonoscopy for colon cancer screening.

Polyps greater than 1cm are reliably detected on air contrast barium enema and appear as masses. Polyps can often be characterised as sessile or pedunculated . Villous adenomas often have a distinctive appearance as a sessile bulky mass with a fond-like surface coated with barium . Although there are some radiographic criteria to help distinguish benign from malignant polyps, endoscopic removal is typically performed for pathologic diagnosis.

The appearance of colorectal cancer on air contrast barium enema varies. Annular carcinoma is a typical appearance of a primary colon cancer . This abrupt narrowing of the bowel lumen is often referred to as an "apple core" lesion, and results from extensive infiltration of tumour into the bowel wall. Scirrhous carcinoma is considered a variant of annular carcinoma and results in a long segment of circumferential luminal narrowing secondary to an intense desmoplastic reaction. Colon cancers can also appear as polypoid masses or plaques causing very subtle contour or profile abnormalities on barium enema .

Although a well performed air contrast barium enema and colonoscopy are reliable methods for the detection of colorectal cancer, both modalities are limited by the inability to accurately evaluate the extracolonic extension of disease and metastases.Computed Tomography is currently the study of choice for primary staging of colorectal neoplasms.



Computed Tomography Overview

A unique imaging feature of CT is its ability to accurately visualise the bowel wall as well as adjacent structures. Therefore, abdominal CT provides a high sensitive method for the detection of intramural pathology as well as extraluminal extension of colonic diseases. CT is particularly valuable for characterisation of primary colonic malignancies and remains the study of choice for staging, as it can demonstrate regional extension of tumor as well as adenopathy and distant metastases. Accuracy rates for preoperative staging of colorectal cancer with CT range between 90-100% (Mays J-1980, van Waes PF-1983, Theoni RF-1981). Technical advancements and the introduction of spiral CT may continue to improve the accuracy of CT.



CT Technique

Abdominal CT is performed after the administration of oral and intravenous contrast. The patient routinely drinks approximately 1200-1250 cc of a 3% oral iodinated contrast solution 60-90 minutes before the scan. If specific colonic pathology is suspected, it is important to adequately opacify the entire colon. Oral contrast can be administered the night before the study as well as before the scan to ensure that contrast reaches the colon. Alternatively in urgent cases, or in patients in whom limited rectosigmoid disease is suspected, contrast can be gently administered through a rectal tube. The use of air and/or water to distend the colon has also been reported to be helpful (Amin Z-1996, Gazelle GS-1995). In situations where adequate opacification of the colon is not possible, air and faeces often provide natural contrast and sometimes allow detection of pathology.

The administration of intravenous contrast is essential for complete evaluation of colorectal disease, especially if extracolonic extension of disease is also suspected. At our institutions we routinely administer 100-120 cc of Iohexol 350 or Iopromide 300 intravenously, at a rate of 2-3 cc/sec.

Using a standard CT scanner, the abdomen should be routinely imaged from the level of the diaphragm through the symphysis pubis. Consecutive slices with 8mm collimation are obtained contiguously through the liver and at 10mm intervals thereafter. Thinner slices may be performed through a specific area of concern. With the arrival of spiral CT, faster scans are possible, allowing liver imaging during the arterial and/or venous phase of contrast enhancement. Using spiral CT 5-8mm collimation can be performed with a table speed of 8mm/sec., with reconstruction every 5-6mm. Our standard spiral CT protocol is to acquire data during the portal venous phases (45-60 seconds after initiation of contrast injection) of liver enhancement.



CT Imaging Findings

Primary Tumor

The sensitivity of CT for the detection of primary colon cancer is reported to range between 75-85% (Freeny PC-1986, Kerner BA-1993). CT is limited in its detection of small tumours (<2cm). Recent advancements in spiral CT and the use of 3D CT with interactive multiplanar views will likely improve the sensitivity of CT. Computed tomography in patients with colorectal cancer typically demonstrates a discrete soft tissue density mass with irregular borders. Larger masses may have a low density necrotic centre or occasionally may contain gas, resembling an abscess. However, a significant percentage of colorectal cancers will present, not as a discrete mass, but rather as a focal or circumferential area of wall thickening. In particular, rectal cancers may appear as asymmetric wall thickening which narrows the lumen. Thus, it is sometimes difficult to prospectively distinguish carcinoma from other processes which result in bowel wall thickening such as diverticulitis.

Complications of Primary Tumors

Complications of primary colonic malignancies, such as obstruction or perforation can be readily visualised on CT. The sensitivity of CT for detecting bowel obstruction is high, ranging between 90-94% (Megibow AJ-1991, Fukuya T-1992). With careful analysis of the images, the exact cause of the obstruction can be identified in greater than 70% of cases (Megibow AJ-1991).

Intussusception is a complication of colonic neoplasms which has a distinctive appearance. Intussusceptions can appear as a target mass with alternating rings of soft tissue and fat, representing the wall of the intussusceptum, mesenteric fat, and the wall of the intussuscipiens (Bar-Ziv J-1991). As oedema of the bowel wall increases, these distinct layers may be obscured. The presence of air within the bowel wall (pneumatosis) indicates significant ischaemia.

Perforation is another complication that can result from colorectal carcinoma. CT is extremely sensitive for detecting free air within the abdomen. Occasionally, extravasation of oral contrast material allows exact identification of the site of perforation. Often, however, perforations are more subtle and appear as small pockets of air outside the involved bowel loop associated with mesenteric thickening (Phatak MG-1984).



Local Disease and Adenopathy

Local extension of tumour on CT can appear as extracolonic tumour mass or infiltration of pericolonic fat. Invasion of nearby organs such as the bladder or pelvic muscles can be identified. CT is able to detect pericolonic extension of disease. CT is more accurate than MR in staging the local extent of tumour, particularly for rectal cancers and detection of penetration of the lamina propria (Zerhouni EA-1996). A recent study reported a 61% sensitivity and 81% specificity for the CT detection of local tumour extension (Freeny PC-1986). The lower sensitivity results from CT being unable to detect microscopic extramural tumour extension .

CT is able to detect pericolonic extension of disease. CT is more accurate than MR in staging the local extent of tumour, particularly for rectal cancers and detection of penetration of the lamina propria (Zerhouni EA-1996). A recent study reported a 61% sensitivity and 81% specificity for the CT detection of local tumour extension (Freeny PC-1986). The lower sensitivity results from CT being unable to detect microscopic extramural tumour extension .

CT can also reliably detect lymphadenopathy within the abdomen, with a high sensitivity but equal accuracy compared with MR (Zerhouni EA-1996). The presence of lymph nodes measuring greater than 1-1.5cm is considered to be pathologic. However enlarged nodes do not necessarily contain tumour, they may simply be reactive. Conversely, microscopic tumour may involve normal sized nodes. Therefore although CT has a high specificity (96%) for the detection of metastatic lymph nodes, the sensitivity is only twenty six percent (Freeny PC-1986). In most cases, however, the inability of CT to detect lymph nodes involved with tumour does not present a clinical problem, since regional lymph node dissection is routinely performed at surgery.



Metastases

Computed Tomography

The liver is the predominant organ to be involved with metastases from colorectal cancer. The accurate detection of liver metastases is essential, as patient management is altered if metastases are present. CT has an established role in the detection of liver metastases in patients with a variety of primary tumours, including colorectal cancer (Heiken JP-1989, Wernecke K-1991). The detection of liver masses with conventional CT varies from series to series but usually ranges between 60-75% (Wernecke K-1991, Miller WJ-1994, Matsui -1987, Soyer P-1992). However, the sensitivity of conventional CT for the detection of masses under 1cm in size is low and ranges from 0% to 47%.

Recent advancements in technology have made spiral CT scanners more widely available. Spiral CT allows more rapid scanning. In fact, the entire liver can be imaged in one breath-hold. Faster scanning eliminates respiratory misregistration, and allows imaging during the optimal window for lesion detection. Therefore spiral scanning with rapid IV contrast injection is considered the preferred technique for liver imaging. A recent study of the detection of hepatic masses using spiral CT demonstrated a better than 90% sensitivity for detecting liver lesions over 1cm in size, and a 56% sensitivity for detecting lesions less than 1cm (Kuszyk BS-1996) . This represents an improvement compared with traditional incremental CT scanning.

Another sensitive CT technique for the detection of hepatic metastases from colorectal cancer is CT with arterial portography (CTAP) (Soyer P-1994). This technique is invasive and expensive as it requires the placement of a catheter in the proximal superior mesenteric artery under fluoroscopic guidance for contrast injection. The objective is to scan the patient as the contrast enters the portal venous system. The results of several studies indicate that time delays must be chosen in conjunction with appropriate injection rate and iodine dose. The examination is therefore technically demanding (only one shot). Since hepatic metastases are usually supplied by the hepatic artery and not the portal vein, they will appear as low density lesions within an enhancing liver parenchyma.

Although highly sensitivity (81-93%) (Nelson RC-1989, Heiken JP-1989, Soyer P-1992 ), CTAP also has a high false positive rate (13-15%) (Soyer P-1994, Soyer P-1993). Non-tumourous perfusion defects are frequently seen and may mimic the appearance of tumours. Furthermore, CTAP is limited in the presence of liver cirrhosis and portal hypertension.

Magnetic Resonance Imaging

The accuracy of dynamic enhanced CT and unenhanced MRI in the detection of metastatic liver disease appears to be equal at 85% (Zerhouni EA-1996). High specificity for both CT (97%) and MRI (93%) is similar to most published series. When the sensitivity of the two techniques are compared, the sensitivity of CT ranges from 71-87% and of MRI from 78-88%. Thus MRI can detect small lesions but as with CT, these often lack morphological features and cannot be characterised as benign or malignant requiring follow-up with serial scanning. Respiratory motion artefacts are particularly problematic for MRI resulting in lesion blurring or unsharpness. Conventional MRI is also time consuming, requiring 3-10 minutes per sequence. Recent technological advances in hardware and software have resulted in the ability to acquire MRI images more rapidly in a breath-hold. Furthermore newer sequences are now available that have resulted in greater contrast between lesion and liver with improved spatial resolution. In addition, with ongoing developments in liver specific MR contrast agents a number of liver specific agents have been or are on the verge of regulatory approval which could help MRI surpass CT in liver imaging. Ferumoxides are liver specific (Kupffer cell uptake) MR contrast agents licensed for the detection of hepatic tumours (AMI-25 or Feridex; Advanced Magnetics, Cambridge, Mass: and Endorem; Laboratoire Guerbet, Aulnay-sous-Bois, France) and NHSU-555A (Schering AG, Berlin). Other hepatocyte uptake contrast agents are currently in clinical trials including mangafodipir trisodium Mn-DPDP (Nycomed Ltd.), Gd-BOPTA and Gd-EOB-DTPA.. Ferumoxides are superparamagnetic iron oxide particles that are sequested by Kupffer cells in the liver and spleen with approximately 80% of an injected dose being directly sequested in liver. Advantages of ferumoxide-enhanced imaging include the lack of invasiveness and the extension of the optimal time window for imaging of the liver which does not depend on the cardiovascular status of the patient. The latter has major clinical implications and yields greater flexibility in imaging. Local magnetic field inhomogeneities result in darkening of normal liver parenchyma which translates into improved contrast noise and therefore increased conspicuity allowing improved detection of lesions. On the basis of the results obtained to date, it appears that ferumoxide-enhanced MRI should be performed in patients with metastatic disease in whom less expensive routine screening tests have detected only a small number of metastases that make surgery a feasible option. It would appear that CTAP and ferumoxide-enhanced MR imaging are indicated in the same patients. The sensitivity of ferumoxide-enhanced MRI when compared to CTAP has not been adequately addressed particularly for metastases that are less than 1cm in diameter. Further assessment of CTAP with liver specific contrast agent MRI is required before the replacement of CTAP (Hagspiel KD-1995, Seneterre E-1996, Weisslender R-1994, Soyer P-1996, Rinck PA-1995).

Although the liver is the primary metastatic sight for colorectal cancer, metastases may also involve other abdominal organs such as the adrenals. In addition metastases may involve the chest, affecting the lung parenchyma, airways or mediastinum.



Tumor Recurrence

Introduction

After curative resection of colorectal cancer, recurrent disease occurs in 37-44% of patients (Cass AW-1976, Willett CG-1984, Tong D-1983, August DA-1984). Most recurrences (80%) occur within two years of surgical resection (Cass AW-1976). Recurrences both local and distant are more likely for rectal than colonic tumours (Pihl E-1981). Pihl et al. showed a 42% total recurrence rate for rectal cancer and 33% total recurrence rate for colon tumours in 1315 patients. Local recurrence for rectal tumours was 12% and 19% for patients with colonic cancer. The pattern of recurrence is largely dependent on the stage of the primary cancer with stages B3, C2 and C3 of the modified Duke’s classification system more likely to have local recurrence and distant metastasis. Local recurrence at the operative site accounts for 19-48% of recurrences, whereas distant metastases account for 25-44% (Cass AW-1976, Willett CG-1984, Tong D-1983). Multiple sites of recurrence are more common than single site disease (Charnsangavej C-19?). Extraluminal recurrence is more common than endoluminal recurrence (Thoeni RF-1981). Patient symptoms and physical examination can provide the first indication of recurrent tumour. However, it is generally believed that symptomatic or physically detected recurrent tumour is likely to be advanced and non-curable if detected this way (Cochrane JP-1980, Pihl E-1981, McCarthy SM-1985). An exception to this rule are lesions palpable at rectal examination and small anastomotic recurrences after low anterior resections (Kelly CJ-1992). CEA levels may be normal even with biopsy proven locally recurrent tumour (McCarthy SM-1985). As local recurrent disease is associated with substantial morbidity and mortality, surgical resection offers the only chance of cure (Baert RW-1983, Stulc JP-1986). Curative resection rates for locally recurrent disease are low (2-14%) but a higher curative rate (40%) has been reported in intensive follow-up including regular endoscopy is performed (Schiessel R-1986). The follow-up of patients who have undergone curative resection is based on anatomic and temporal patterns of tumour recurrence. Priority should be given to anatomic sites that response favourably to additional treatment (eg. liver, lung). Intensive early follow-up may improve curative or palliative treatments (Cathal JK-1992, Martin EW-1985, Attiyeh FF-1981). Colonoscopy and barium enema examinations are sensitive for detecting intraluminal anastomotic recurrence. Neither technique is sensitive in detecting the more common extraluminal local recurrence and distant metastases. Furthermore, neither technique is applicable for evaluating patients who have undergone abdominoperineal resections.

Computed Tomography

CT is a reliable method for the detection of tumour recurrence after resection (Baert RW-1983, Husband JE-1980, Moss AA-1982). Recurrent tumour after surgery usually appears as a soft tissue density mass with irregular borders appearing in or near the surgical bed. This can often be distinguished from post-operative fibrosis which usually appears more linear without a discrete mass. The presence of the uterus, bladder, seminal vesicles and small bowel in the resection bed can lead to interpretative errors as can myocutaneous flaps. Occasionally the distinction between postoperative fibrosis and recurrent tumour is not possible. CT findings clearly indicative of recurrent malignant disease include enlargement of soft tissue mass on serial scans, enlarging regional lymphadenopathy and invasion of contiguous structures. CT-guided biopsy for tissue confirmation is indicated when findings suggest recurrent colorectal carcinoma (Figure).

Magnetic Resonance Imaging

MRI has also been used to evaluate patients with suspected local recurrent colorectal cancer. Differences in signal intensity on T1 and T2 weighted scans have been used to distinguish between recurrent malignancy and benign processes such as radiation change and scarring (Thompson WM-1994, Sugimura K-1990) and so obviate the need for biopsy. Although early studies appeared encouraging, later investigations failed to confirm these findings (Moss A-1989, Thompson WM-1994, Thoeni RF-1991). Fibrotic tissue can demonstrate high signal on T2-weighted scans in the first year after surgery (Ebner F-1988, Rafto SE-1988) or after radiotherapy due to inflammation and oedema. Tumour can also include a desmoplastic response by a host tissue reducing signal intensity; furthermore, biopsy in this situation may only reveal fibrosis.

MRI techniques continue to evolve at a fast pace and the value of contrast enhancement has been evaluated recently. Dynamic contrast enhanced subtraction MR imaging appears to be better than using T2-weighted spin-echo signal characteristics (Kinkel K-1996, Muller-Schimpfle M-1993) for differentiating benign from malignant lesions in patients suspected of recurrent colorectal cancer. Early enhancement of an abnormal pelvic structure has high sensitivity and specificity (97% and 81%) compared with T2-weighted MR imaging 77% and 56%). Contrast enhancement by recurrent tumour can be explained as the basis neoangiogenesis. A minimum of six months is required between the end of treatment (surgery or radiation) to avoid false-positive results due to early benign inflammatory changes.

New Imaging Techniques

Recently new imaging methods, for example positron emission tomography (PET) has been evaluated in the context of recurrent colorectal cancer. Increased glucose metabolism and neoplastic and inflammatory tissue is shown by an increase in uptake of Fluorine-18-deoxyglucose (FDG). Increased uptake has been shown in primary colorectal tumours, nodal and liver metastases and in recurrence. In distinguishing recurrence from fibrotic tissue, FDG-PET is reported to be highly sensitive and specific comparing well with and complementary to MRI (Ito K-1992). However, increased FDG uptake does occur and persists after radiotherapy and this may limit the use of this technique in patients receiving pre-operative radiotherapy. Monoclonal antibodies against tumour associated antigens (eg. CEA or mucine-like glycoprotein (TAG-72) can be conjugated to radionuclides and used to localise tumours, regional and distant metastases in patients with colorectal cancer (Ryan JW-1993, Corbesiero RM-1991, Buraggi GC-1991). This technique may be useful in identifying recurrent disease that cannot be demonstrated conventionally (eg. by CT) (Abdel Nabi HH-1979). Factors that influence the success of antibody imaging include local expertise and the level of antigen expression.



Distant Recurrent Tumor

Postoperative development of liver metastases is reported to occur in up to 30% of patients within 2 years after curative surgery for colorectal cancer (Finlay IG-1992,1993). The development of hepatic metastases after surgery has a significant impact on patient survival. Once again, CT with intravenous contrast administration is the imaging modality of choice for detecting recurrent tumour within the liver. CT has been shown to be more helpful in the diagnosis of recurrent hepatic metastases than laboratory studies (CEA and LFT) (Freeny PC-1986). Other sites of recurrent tumour involvement after surgery for colorectal cancer, such as nodal recurrence or tumour recurrence involving the abdominal wall can also be successfully imaged with computed tomography.



Virtual Colonoscopy

One of the exciting potential changes in the evaluation of colon cancer is the coupling of spiral CT and advanced computer graphics for the generation of three dimensional images of the lumen of the colon. The technique termed "virtual colonoscopy" potentially could be a less invasive technique than fibreoptic colonoscopy and may possibly have a higher accuracy than classic barium enemas. However, significant advances in CT data acquisition and computer image processing techniques will be needed before the technology can become a reliable cost effective alternative (Vining DJ-1996).



Conclusion

In summary, colorectal cancer is a common malignancy resulting in significant morbidity and mortality. Although colonoscopy has gained widespread acceptance as a diagnostic tool, the air contrast barium enema continues to be safe, accurate and inexpensive. As the debate over instituting colon cancer screening continues, the barium enema may once again assume a more prominent role in the detection of colorectal cancer. However, although colonoscopy and barium enema are accurate at colon cancer detection, they are unable to evaluate extracolonic disease. CT has been proven to be valuable in the preoperative assessment and staging of patients with colorectal cancer as well in the postoperative surveillance for recurrence. Rapid advancements in CT technology will likely continue to improve its accuracy and utility. Similarly, ongoing developments in MRI (contrast agents, hardware and software developments) will enhance its utility in liver imaging and for the detection of locally recurrent tumour.



References

Abdel Nabi HH, Chan HW, Doerr RJ. Indium-labelled anti-colorectal carcinoma monoclonal antibody accumulation in non-tumoured tissue in patients with colorectal carcinoma. Journal of Nuclear Medicine 1990;31:197501979.

Amin Z, Boulos PB, Lees WR. Technical Report: spiral CT pneumoncolon for suspected colonic neoplasms. Clinical Radiology 1996;51:56-61.

Attiyeh FF, Sterans WW Jr. Second look laparotomy based on CEA evaluations in colorectal cancer. Cancer 1981;47:2119-2125.

August DA, Ottow RT, Sugarbaker PH. Clinical perspective of human colorectal cancer metastasis. Cancer Metastasis Rev. 1984;3:303-324.

Bar-Ziv J, Solomon A. Computed tomography in adult intussusception. Gastrointestinal Radiology 1991;16(3):264-6.

Beart RW Jr, O’Connell MJ. Postoperative follow-up of patients with carcinoma of the colon. Mayo Clin Proc 1983;58:361-363.

Beggs I, Thomas BM. Diagnosis of carcinoma of the colon by barium enema. Clin Radiology 1983;34:423-425.

Brady AP, Stevenson GW, Stevenson I. Colorectal cancer overlooked at barium enema examination and colonscopy: a continuing perceptual problem. Radiology 1994;192:373-378.

Buraggi GL, Gasparini M, Seregni E. Immunoscintigraphy of colorectal carcinoma with an anti-CEA monoclonal antibody: a critical review. Internation Journal of Radiology and Applied Instrumentation B 1991;18:45-50.

Cass AW, Million RR, Pfaff WW. Patterns of recurrence following surgery alone for adenocarcinoma of the colon and rectum. Cancer 1976;37:2861-2865.

Cathal JK, Daly JM. Colorectal cancer. Principles of postoperative follow-up. Cancer 1972;70:1397-1408.

Charnsangavej C. New imaging modlaities for follow-up of colorectal cancer. Cancer 1993;71:4236-4240.

Chen YM, Ott DJ, Wolfman NT, et al. Recurrent colorectal carcinoma: evaluation with barium enema examination and CT. Radiology 1987;163:307-310.

Cochrane JP, William JT, Faber RG, Slack WW. Value of out-patient follow-up after curative surgery for carcinoma of the large bowel. BMJ 1980;280:593-595.

Corbisiero RM, Yamauchi DM, Williams LE, et al. Comparison of immunoscintigraphy and computerized tomography in identifying colorectal cancer: individual lesion analysis. Cancer Research 1991; 51:5704-5711.

CRC Factsheet. Cancer of the large bowel - UK. Cancer Research Campaign Factsheet 18.1, 1993.

Ebner F, Kressel HY, Mintz MC, et al. Tumour recurrence versus fibrosis in the female pelvis: differentiation with MR imaging at 1.5 T. Radiology 1988; 166:333-340.

Fork FT, Lindstrom C, Ekelung GR. Reliability of routine double contrast examination of the large bowel in polyp detection: a prospective clinical study. Gastrointestinal Radiology 1983;8:163-172.

Fukuya T, Hawes DR, Lu CC, et al. CT diagnostic small bowel obstruction: efficacy in 60 patients. AJR 1992;158:765-769.

Finlay IG, Meek DR, Gray HW, Duncan JG, Mcardle CS. Incidence and detection of occult hepatic metastases in colorectal carcinoma. Br J Med 1982;284:803-805.

Freeny PC, Marks WM, Ryan JA, Bolen JW. Colorectal carcinoma evaluation with CT: preoperative staging and detection of postoperative recurrence. Radiology 1985;158:347-353.

Gazelle GS, Gaa J, Saini S, Shellito P. Staging of colon cancer using water enema CT. Journal of Computer Assisted Tomography 1995;19:87-91.

Glick SN, Teplick SK, Balfe SM, et al. Large colonic neoplasms missed by endoscopy. AJR Cancer 1989;152:513-517.

Heiken JP, Weyman PJ, Lee JK, et al. Detection of focal hepatic masses; prospective evaluation with CT, delayed CT, CT during arterial portography and MR imaging. Radiology 1989;171:47-51.

Husband JE, Hodson NJ, Parsons CA. The use of computed tomography in recurrent rectal tumours. Radiology 1980;134:677-682.

Ito K, Kato T, Tadokoro M, et al. Recurrent rectal cancer and scar: differentiation with PET and MR imaging. Radiology 1992;182:549-552.

Kelly CJ, Daly JM. Colorectal cancer. Principles of post-operative follow-up. Cancer 1992;70:1397-1408.

Kerner BA, Oliver GC, Eisenstat TE, Rubin RJ, Salvati EP. Is preoperative computerized tomography useful in assessing patients with colorectal carcinoma? Dis Colon and Rectum 1993;365(11):1050-1053.

Kinkel K, Tardivon AA, Soyer P, et al. Dynamic contrast-enhanced subtraction versus T2-weighted spin-echo MR imaging in the follow-up of colorectal neoplasm: A prospective study of 41 patients. Radiology 1996;200:453-458.

Kuszyk BS, Bluemke DA, Urban BA, et al. Portal-phase contrast enhanced helical CT for the detection of malignant hepatic tumours: sensitivity based on comparison with intraoperative and pathologic findings. AJR 1996;166:91-95.

Martin EW Jr, Minton JP, Carey LC. CEA-directed second look surgery in the symptomatic patient after primary resection of colorectal carcinoma. Ann. Surg. 1985;202:310-317.

Matsui O, Takashima T, Kadoya M, et al. Liver metastases from colorectal cancer: detection with CT during arterial portography. Radiology 1987;165:65-69.

Mays J, Zoronza J. Computed tomography of colon cancer. AJR 1980; 135:43-46.

McCarthy SM, Barnes D, Deveney K, et al. Detection of recurrent rectosigmoid carcinoma: prospective evaluation of CT and clinical features. AJR 1985;144:577-579.

Megibow AJ, Balthazar EJ, Cho KC, Medwid SW, et al. Bowel obstruction: evaluation with CT. Radiology 1991;180:307-308.

Miller WJ, Baron RL, Didd GD, Federle MP. Malignancies in patients with cirrhosis: CT sensitivity and specificity in 200 consecutive transplant patients. Radiology 1994;193:645-650.

Moss AA. Computed tomography in the staging of gastrointestinal carcinoma.. Radio Clin North Am 1982;20:761-780.

Moss A. Imaging of colorectal carcinoma. Radiology 1989;170:308-310.

Muller-Schimpfle M, Brix G, Layer G, et al. Recurrent rectal cancer: diagnosis with dynamic MR imaging. Radiology 1993;189:881-889.

Nelson RC, Chezmar JL, Sugarbaker PH, Bardardino ME. Hepatic tumours: comparison of CT during arterial protography, delayed CT, and MR imaging for pre-operative evaluation. Radiology 1989;171:47-51.

Parker SL, Tong T, Bolden S, Wingo PA. Cancer Statistics 1996. CA: A Cancer Journal for Clinicians 1996;46:5-27.

Phatak MG, Frank SJ, Ellis JJ. Computed tomography of bowel perforation. Gastrointestinal Radiology 1984;9:133-135.

Pihl E, Hughes ESR, McDermot FT, Milne BJ, Prince AB. Disease free survival and recurrence after resection of colorectal cancer. J. Surg. Oncol. 1981;16:333-341.

Rafto SE, Amendole MA, Gefter WB. MR imaging of recurrent colorectal carcinoma versus fibrosis. Journal of Computer Assisted Tomography 1988;12:521-523.

Rinck PA, Muller RN. Magnetic resonance imaging contrast agents. In: Rinck PA (Ed). The Rational Use of Magnetic Resonance Imaging. Blackwell Scientific, Oxford, 1995: 287-306.

Ryan JW. Immunoscintigraphy in primary colorectal cancer. Cancer 1993;71:4217-4224.

Schiessel R, Wanderlick M, Herbst F. Local recurrence of colorectal cancer: effect of early detection and aggressive surgery. Br. J. Surgery 1986;73:342-344.

Seneterre E, Taourel P, Bouvier G, et al. Detection of hepatic metastaes: Ferumoxides-enhanced MR imaging versus unenhanced MR imaging and CT during aterial portography. Radiology 1996; :785-792.

Soyer P, Levesque M, Elias D, et al. Detection of liver metastaes from colorectal cancer: comparison of intraoperative ultrasound and CT during arterial portography. Radiology 1992;183:541-544.

Soyer P, Bluemke DA, Hruban RH, Sitzmann JV, Fishman EK. Hepatic mtastases from colorectal cancer: detection and false positive findings with helical CT during arterial portography. Radiology 1994;192:389-392.

Soyer P, Lacheheb D, Levesque M. False positive CT portography: correlation with pathologic findings. AJR 1993;160:285-289.

Soyer P. Will ferumoxide enhanced MR imaging replace CT during arterial portography in the detection of hepatic metatases? Prologue to a promising future. Radiology 1996;20:610-611.

Stevenson G. Radiology in the detection and prevention of colorectal cancer. European Journal of Cancer 1995; ?:1121-1126.

Stulc JP, Petrelli NJ, Herrera L, et al. Anastomotic recurrence of adeno- carcinoma of the colon. Arch. Surg. 1986;121:1077-1080.

Sugimura K, Carrington BM, Quivey JM. et al. Post-irradiation changes in the pelvis. Assessment with MR imaging. Radiology 1990;175:805-813.

Theoni RF, Moss AA, Schnyder P, Margulis AR. Detection and staging of primary rectal and rectosigmoid ancer by computed tomography. Radiology 1981;141:135-138.

Theoni RF. Colorectal cancer: cross-sectional imaging for staging of primary tumour and detection of local recurrence. AJR 1991;156:909-915.

Thompson WM, Trekkner SW. Staging colorectal carcinoma. Radiol. Clin. North Am. 1994;32:25-37.

Tong D, Russell AH, Dawson LE, Wisbeck W. Second laparotomy for proximal colon cancer: sites of recurrence and implications for adjuvant therapy. Am. J. Surg. 1983;145:382-386.

van Waes PF, Koehler PA, Feldberg MA. Management of colorectal carcinoma: impact of comuted tomography. AJR 1983;140(6):1137-1142.

Vining DJ. Virtual endoscopy: is it reality? Radiolgy 1996;200:30-31.

Weissleder R. Liver imaging with iron oxide: Toward consensus and clinical practice. Radiology 1994;193:593-595.

Wernecke K, Rummeny E, Bongartz, et al. Detection of hepatic masses in patients with carcinoma: comparative sensitivities of sonography, CT, and MR imaging. AJR 1991;157:731-739.

Willett CG, Tepper JE, Cohen AM, et al. Failure patterns following curative resection of colonic carcinoma. Ann. Surgery 1984;200:685-690.

Zerhouni EA, Rutter C, Hamilton SR, et al. CT and MR imaging in the staging of colorectal carcinoma: report of Radiology Diagnostic Oncology Group II. Radiology 1996;200(2):443-451.