First published online October 29, 2007
British Journal of Radiology (2008) 81, 10-19
© 2008 British Institute of Radiology
doi: 10.1259/bjr/19492531
Accuracy of CT prediction of poor prognostic features in colonic cancer
S Burton, FRCS1,
G Brown, FRCR2,
N Bees, FRCR1,
A Norman, PhD2,
O Biedrzycki, FRCPath1,
A Arnaout, FRCPath1,
A M Abulafi, FRCS1 and
R I Swift, FRCS1
1 Departments of Colorectal Surgery, Radiology and Histopathology, Mayday University Hospital, London Road, Croydon, CR7 7YE, 2 Royal Marsden Hospital, Departments of Radiology and Statistics, Downs Road, Sutton, Surrey, SM2 5PT, UK
Correspondence: Dr Gina Brown, Department of Radiology, Downs Road, Sutton, Surrey, SM2 5PT, UK. E-mail: Gina.Brown{at}rmh.nhs.uk
 |
Abstract
|
|---|
Whilst imaging of poor prognostic features in rectal cancers has assisted pre-operative treatment stratification, such features have yet to be evaluated in colonic cancers. This study aims to develop criteria for identifying poor prognostic features in colonic tumours and assess the accuracy of CT prediction against histopathology. Criteria were developed for predicting T-stage and N-stage, the presence of extramural vascular invasion and involvement of the retroperitoneal surgical margin (RSM). These criteria were tested on 33 patients with colonic cancer who underwent pre-operative high-resolution CT of their tumour. Two radiologists (Obs 1 and Obs 2) identified independently these poor prognostic features and the results were compared with the final histopathological results. Histological agreement and interobserver variation were calculated using the kappa test. Accuracy of CT prediction of tumour extension beyond muscularis propria was 82% (Obs 1) and 70% (Obs 2). Correct prediction of RSM involvement was 76% (95% confidence interval (CI): 57.8–88.9%) and 79% (95%CI: 61.1–91%) for Obs1 and Obs 2, respectively, with significant agreement between observers (
= 0.455, p = 0.050). Prognosis was correctly predicted using CT in 82% (95%CI: 61.5–81.2%) (Obs1) and 85% (95%CI: 68.1–94.9%) (Obs2) with moderate agreement ( = 0.459, = 0.527, respectively) with histology. In conclusion, CT has potential as the imaging modality of choice in the pre-operative prediction of poor prognostic features in colonic cancers and could play a role in future treatment stratification.
|
Introduction
|
|---|
The potential advantages of neoadjuvant therapies in colorectal cancer include early treatment of micrometastatic disease in advanced tumours, greater patient compliance and improved completion rates, and significant downstaging of tumours to allow potential curative resection [1].
Prediction of poor prognostic features in rectal cancers has helped direct the application of neoadjuvant treatment strategies, resulting in reduced local recurrence rates and improved disease-free and overall survival. However, in colonic cancers, for which primary surgery remains the standard treatment, less importance has been placed on the pre-operative identification of poor prognostic features other than the presence of metastatic disease. More accurate identification of such features could allow future consideration of neoadjuvant treatments in locally advanced colonic tumours. Peritoneal involvement, N2 disease (presence of three or more lymph nodes) and the presence of extramural vascular invasion (EMV) are well-known poor prognostic features in both colonic and rectal tumours which, with appropriate imaging, can potentially be identified pre-operatively. In rectal cancer, circumferential margin involvement is one of the most powerful predictors of local recurrence and poor survival, but it is not yet known whether involvement of the retroperitoneal surgical margin in colonic tumours is of equivalent significance.
CT scanning has increasingly become a routine part of the pre-operative staging of colonic cancer. Its uses include assessing local extension of tumour and presence of nodal and distant metastases, providing a baseline for follow-up and planning surgery [2]. Several studies have questioned the routine use of pre-operative CT scanning for staging colonic cancers owing to its limitations in altering surgical management [3, 4].
Use of a neoadjuvant treatment strategy in colonic cancers directed by pre-operative imaging has yet to be explored. However, a better understanding of the pathology and anatomy of colonic tumours could improve the assessment of local spread and nodal involvement. CT also has the advantage of accurately assessing distant metastases simultaneously [2].
This pilot study evaluates the reliability of CT identification of known poor prognostic features against the gold standard of histopathology in our unit and other reported series.
|
Methods
|
|---|
Development of CT criteria
CT criteria were based on the following:
- (1) An understanding of the surgical anatomy and peritoneal coverings of the colon.
- (2) Knowledge of the morphological appearances of tumour spread, nodal disease and extramural vascular invasion based on MRI and histopathological criteria (Table 1
).
Anatomy
CT identification of the normal mesentery and its vasculature has previously been studied, showing higher rates of detection of the right-sided colonic vessels compared with left-sided and sigmoid vessels [5]. A basic knowledge of the expected routes of the ileocolic, right colic, middle colic and sigmoid vessels and potential variations in the normal colonic mesenteries is essential to the interpretation of whether the vessel is infiltrated with tumour (i.e. EMV) [6].
The caecum, ascending colon and descending colon are covered anteriorly by visceral peritoneum, which extends medially onto the rudimentary mesocolon and laterally onto the abdominal wall as parietal peritoneum. Although some anatomical variation exists, approximately 50% of the posterior circumference at these sites of the colon is free of peritoneum (Figure 1a,b). This has been described as the "bare area" or retroperitoneal surgical margin (RSM), and is the plane of dissection during surgical resection. This plane is similar to the mesorectal fascia encountered in rectal cancer surgery. Involvement of this area by tumour would indicate T3 disease with a positive margin rather than peritoneal involvement (T4 disease). In contrast, the transverse and sigmoid colon are completely invested in peritoneum and are suspended on their respective mesenteries (Figure 2a,b). Therefore, the RSM in these regions is minimal and can only potentially occur at the root of the mesentery where the colonic vessels will be ligated.
View larger version (121K):
[in this window]
[in a new window]
|
Figure 1. (a) Axial CT at the level of the ascending colon demonstrating the retroperitoneal surgical resection margin (solid white arrows) and the peritoneum (white dashed arrows). (b) Axial CT at the level of the descending colon demonstrating similar features.
|
|
View larger version (114K):
[in this window]
[in a new window]
|
Figure 2. (a) Axial CT demonstration of complete investment of the transverse colon by peritoneum (white dashed arrows). (b) Axial CT demonstration of complete investment of the sigmoid colon by peritoneum (white dashed arrows).
|
|
Study design
Ethical approval for the study was obtained from the local ethics committee. From December 2003 to March 2004, patients presenting with right-sided colonic cancer to a district general hospital were eligible for inclusion in the study. From April 2004 to September 2004, study participation was extended to all patients with colonic cancer 25 cm or further from the anal verge on withdrawal of the colonoscope. This extension of the inclusion criteria allowed evaluation of CT assessment at more anatomical variable sites (e.g. transverse and sigmoid colon). There were no age limits to patient recruitment. All patients underwent colonoscopic biopsy with histopathological confirmation of the diagnosis, and chest X-ray and abdominal CT scan to exclude distant metastatic disease. In the presence of distant metastatic disease, patients were excluded from the study only if their primary treatment was not surgical resection.
CT protocol
The night before the scan, patients were given 10 ml of gastrografin (Schering diagnostics, Berlin, Germany) diluted in 300 ml of tap water orally. 1 h prior to the scan, patients drank a further 10 ml of gastrografin diluted in 990 ml of tap water. The technique was performed using a CT scanner (Siemens AR Starr, Erlangen, Germany). Single slice spiral scans were taken with collimation of 10 mm and a pitch of 1 following injection of intravenous contrast medium (Iohexol 300; Omnipaque, Amersham Health, UK). The contrast medium was injected at a rate of 3 ml s–1 and the scan was performed after a 50 s delay.
Image interpretation
Two consultant radiologists with a specialist interest in gastrointestinal imaging independently reviewed the hard copy of each scan for T-stage, N-stage, presence of EMV and involvement of the RSM using the developed criteria (Table 1
). In addition, a verdict of good or bad prognosis was determined. Bad prognosis was predicted in the presence of at least two of the following poor prognostic factors — T4, N2 or EMV. Good prognosis was predicted in all other cases. Both observers were blinded to each other's predictions and the histopathological results. No consensus staging was attempted for the purposes of this pilot study. All patients underwent primary surgical resection of the tumour within 3 weeks of the initial CT scans.
Histopathological processing and staging
Specimens were processed according to UK Royal College of Pathologists guidelines [7]. 6 mm axial slices were cut throughout the tumour length and inserted into megablocks for further fixation and processing. All megablocks were embedded in paraffin, glass mounted and stained with haematoxylin & eosin for histological assessment. Separate microsections were also taken of any additional lymph nodes not included in the slices and distal margins if the tumour was within 3 cm. Each specimen's overall stage was reported and a minimum data set was recorded.
Statistical analysis
The sensitivity, specificity and accuracy of each of the poor prognostic features identified by CT were calculated by comparison with histology. The kappa test was used to infer agreement and standard error of the mean. The kappa test was also used to detect concordance between observers for each technique.
|
Results
|
|---|
33 patients were recruited to the study: 15 were female and 18 male. The average age was 72 years (range 46–93 years). The mean lymph node count was 20 (range 9–48). Table 2 shows the site and characteristics of the tumours as determined by histopathological assessment. 14 of the 33 tumours were T3 but approximately one-third were early tumours (T1 or T2). Only six tumours had nodal metastases, with only three of these having N2 disease. Most of the tumours had negative circumferential margins (30/33) but approximately half had evidence of EMV (15/33). 8 cases out of 33 (24%) had a bad prognosis on the basis of at least two poor prognostic features (T4, N2 or EMV) having been identified on histology.
T-stage
The rate of detection of primary tumour was 79% (26/33) for observer (Obs) 1 and 91% (30/33) for Obs 2. For Obs 1, three cases were small T1 tumours (two in the sigmoid colon), two were T2 tumours, one was T3, and one was T4. The same descending colon T4 tumour failed to be identified by Obs 2, and review of the histology confirmed it to be a small cicatrizing tumour. For Obs 1, an accuracy of 36% (95% confidence interval (CI): 20.4–54.9%) was achieved for T-staging with 36% understaged and 33% overstaged by CT with fair agreement ( = 0.204) (Table 3, Figure 3). Obs 2's accuracy was higher at 51.5% (95% CI: 33.6–69.1) with 30% understaging and 21% overstaging and fair agreement ( = 0.313; p = 0.016) with histopathology. When comparing CT T-staging between Obs 1 and Obs 2, there was 39% agreement (95% CI: 23–57.8%) ( = 0.214; p = 0.061).
View larger version (51K):
[in this window]
[in a new window]
|
Figure 3. Percentage agreement and 95% confidence intervals for CT staging in both observers versus histopathological staging.
|
|
Prediction of T4 disease (i.e. either serosal involvement alone or penetration of adjacent organs) on CT scanning appeared more promising. Obs 1 achieved 70% accuracy (95% CI: 51.3–84.4%), 71% sensitivity and 69% specificity when compared with histopathology; Obs 2 was correct in 85% of cases (95% CI: 68.1–94.9%), with 43% sensitivity and 96% specificity (Table 3, Figure 3). Agreement was fair ( = 0.31) for Obs 1 and moderate ( = 0.463) for Obs 2. Between observers, there was 67% agreement (95% CI: 48.2–82%) ( = 0.206; p = 0.3).
Figure 4 shows a T1 tumour without distortion of the bowel wall layers.
View larger version (110K):
[in this window]
[in a new window]
|
Figure 4. Axial CT image demonstrating a T1 tumour with intraluminal projection of the lesion(open arrow) but no distortion of the bowel wall layers (arrow).
|
|
Figure 5 demonstrates two examples of T3 tumours that were correctly predicted compared with histology. Peritumoral stranding or nodularity were the signs that helped to correctly stage these tumours.
View larger version (83K):
[in this window]
[in a new window]
|
Figure 5. (a) Axial CT image through the mid-ascending colon showing the primary tumour as a circumferential soft tissue density mass. The scan depicts peritumoral stranding posteriorly, suggestive of T3 disease (white arrows). (b) Axial CT image of tumour at the level of mid-sigmoid. The primary tumour is demonstrated as a segment of thickened sigmoid colon with enhancing soft tissue density and nodularity into pericolic fat, in keeping with a T3 tumour (white arrow).
|
|
Figure 6 shows tumour extending to the limit of the bowel wall, indicative of peritoneal involvement, but no evidence of tumour penetration of the anterior abdominal wall.
View larger version (178K):
[in this window]
[in a new window]
|
Figure 6. Axial CT image at the level of the caecum. The primary tumour is demonstrated as a bulky annular mass infiltrating through the bowel wall circumferentially. Anterior spread through peritonealized colon indicates T4 disease(white arrow).
|
|
N-stage
For Obs1, an accuracy of 64% (95% CI: 45.1–79.6%) was achieved for N-staging, with 12% understaged and 24% overstaged by CT with poor agreement ( = 0.185) (Figure 3). Obs 2 accuracy was lower at 54.5% (95% CI: 36.4–71.9%), with 15% understaging and 30.5% overstaging and poor agreement ( = –0.071). When comparing CT N-staging between Obs 1 and Obs 2, there was fair agreement ( = 0.341; p = 0.043).
Figure 7 gives two examples of CT-predicted N1 and N2 metastases in nodes >1 cm in diameter.
View larger version (127K):
[in this window]
[in a new window]
|
Figure 7. (a) Axial CT image of enlarged lymph nodes consistent with N1 disease. (b) Axial CT image of multiple enlarged lymph nodes which proved to be N2 disease.
|
|
Extramural vascular invasion
Obs 1 achieved an accuracy of 54.5% (95% CI: 36.4–71.9%) for CT prediction of EMV compared with histopathology, with poor agreement ( = 0.024) (Figure 3). Obs 2 achieved a higher accuracy of 61% (95% CI: 42.1–77.1%) for EMV prediction with fair agreement ( = 0.201), a sensitivity of 53% and a specificity of 67%. When comparing CT-predicted EMV between Obs 1 and Obs 2, there was poor agreement ( = 0.178). Sub-analysis of the EMV data according to the site of tumour (left or right colon) showed accuracy of 64% in the right colon for both observers, and 25% and 50% in the left colon for Obs 1 and Obs 2, respectively.
Two examples of CT-predicted EMV are shown in Figure 8 with the characteristic feature of nodularity of the vessel.
View larger version (56K):
[in this window]
[in a new window]
|
Figure 8. (a,b) Two examples of axial CT images at the level of the caecum. The ileocolic vessels are markedly thickened and irregular and show nodular enhancement (white arrows). This equates to tumour invasion within vessels.
|
|
Retroperitoneal surgical margin
Obs 1 achieved an accuracy of 76% (95% CI: 57.8–88.9%) for CT prediction of RSM status compared with histopathology with poor agreement ( = –0.128) (Figure 3). Obs 2 achieved a higher accuracy of 79% (95% CI: 54.5–86.7%) for RSM status prediction with 33% sensitivity, 83% specificity and moderate agreement ( = 0.436).
A negative RSM can be clearly visualized on the axial CT image of a descending colon tumour with histological confirmation (Figure 9). Figure 10 shows two right-sided colonic tumours with involved posterior resection margins (RSM).
View larger version (84K):
[in this window]
[in a new window]
|
Figure 9. (a) Axial CT image of a descending colon tumour, the posterior margin of which (dashed arrows) is clear of the retroperitoneal surgical resection margin (white arrows). (b) An axial pathological specimen of the same patient with the posterior margin of the tumour (dashed arrows) more than 1 cm from the RSM (solid arrows). (c) Histological demonstration of the negative RSM.
|
|
View larger version (93K):
[in this window]
[in a new window]
|
Figure 10. (a) Axial CT at the level of the ascending colon. A large ascending colon tumour with extension of tumour to within 1 mm of the retroperitoneal surgical margin — this can be seen as a blurring of the plane between the tumour and the posterior fascia (white arrows). (b) Axial CT at the level of the ascending colon. Extension of tumour to the retroperitoneal surgical margin can be seen as a tumour abutting the posterior fascia (white arrows).
|
|
Prognosis
Prediction of prognosis on CT scanning appeared more promising with Obs1, achieving 82% accuracy (95% CI: 61.5–81.2%), 92% sensitivity and 50% specificity when compared with histopathology, with 12% understaged and 6% overstaged. Obs 2 was correct in 85% of cases (95% CI: 68.1–94.9%), had 96% sensitivity and 50% specificity, with 12% being understaged and 3% overstaged (Figure 3). Agreement with histology was moderate: = 0.459 for Obs 1 and = 0.527 for Obs 2. Between observers, there was fair agreement in 79% of cases ( = 0.238).
|
Discussion
|
|---|
Colonoscopy or barium contrast studies remain the gold standards for detecting and diagnosing colonic cancers, although CT colonography is becoming increasingly established, especially as a screening tool [8–10]. However, traditional axial CT scanning offers advantages for obtaining information about the primary tumour beyond the lumen [11], although it has not been proven to consistently change management [4]. In the future, CT-detected prognostic information may have significant impact on pre-operative strategies in colonic cancer.
Many studies have looked at correct CT prediction of Dukes' classification as an indicator of prognosis and its potential for altering management [3, 4]. Freeny et al [12] demonstrated only 48% accuracy for predicting Dukes' classification in 80 patients compared with Acunas et al [13], who demonstrated accuracy rates of 71% in 28 patients. Our prognostic score was simplified into "good" or "bad" depending on absence of T4, N2 and EMV (good) or presence of any two of T4, N2 or EMV (bad). This resulted in a prognostic agreement with pathology in 82% (Obs 1) and 85% of cases (Obs 2); however, out of 33 patients, only 8 were deemed to have poor prognosis on histology using the criteria described, which limits the conclusions that can be drawn. Further work is needed to focus on more locally advanced colonic tumours to evaluate the potential for using CT scanning as a tool for pre-operatively identifying patients with poor prognosis who may benefit from neoadjuvant therapies.
CT staging often relies on size to help determine tumour stage, and one problem encountered is overstaging. This may be caused by the inability to recruit images of the tumours in the true axial plane. Spiral CT images are taken in the transverse plane only. With the increasing availability of thin collimation multislice CT and three-dimensional reconstruction of images, viewing of the true axial plane of colonic tumours should be possible, with an expected reduction in the overstaging rate [14]. Tumour understaging by CT can also occur as a result of the obliquity of images, although microscopic invasion of the serosa by tumour is unlikely to be detected by any imaging modality.
T4 disease or evidence of peritoneal involvement, with or without penetration of adjacent organs, has been shown to be a poor prognostic indicator in colonic cancer [15]. Peritoneal involvement without invasion of adjacent organs can be difficult to predict, as the serosa is a particularly thin layer of the bowel wall. Previous studies have described peritoneal involvement as an irregularity of the border of the tumour or clear invasion into adjacent organs shown by loss of fat planes [16]. Using similar criteria, our results demonstrated accuracies of 70% (69% sensitivity, 90% positive predictive value) and 85% (96% sensitivity, 86% positive predictive value) depending on the observer. Other CT characteristics previously described to define local extension include extracolonic tumour mass, strands of soft tissue extending beyond pericolic fat and invasion of adjacent organs; one series reported accuracy rates of 69% (55/80) for prediction of peritoneal involvement with 61% sensitivity. As a known poor prognostic factor for colonic cancer, such high rates of pre-operative identification of peritoneal involvement are encouraging and could be used in the future to help determine patients suitable for neoadjuvant therapies.
Size, rather than shape or type, of the border of lymph nodes appears to be the most reliable predictive criteria, with knowledge of the expected site of nodes assisting in their CT identification [17, 18]. In our study, accuracy rates for predicting N-stage of 64% and 54.5% were achieved, but only 6 of 33 patients had lymph node metastases on histopathological examination. Such rates were comparable to other published rates for N-stage prediction of 68–75% [12, 13, 19]. A weakness of any imaging technique remains the inability to identify micrometastatic disease within lymph nodes.
Knowledge of the normal and variant anatomy of the vasculature of the colonic mesocolon is essential before correct identification of tumour vascular invasion can be attempted [6]. One study evaluated colonic vessels within the mesocolon using CT and found that ileocolic and right colic vessels could be identified in 82% of cases, compared with 42% and 70% for left colic and sigmoid branches, respectively [5]. EMV is a reported poor prognostic factor in colonic cancer, resulting in reduced overall and disease-free survival [20]. It has previously been identified using MRI in rectal cancer and is described as "serpiginous extension of tumour within a vascular structure" [21]. EMV, however, has not been reported in the literature as a feature identifiable by CT. By using similar criteria as described above for MR imaging in rectal cancer, we were able to identify this feature on CT. Our results show accuracy rates of 61% (63% positive predictive value and 67% sensitivity) and 54.5% (55% positive predictive value and 89% sensitivity). However, correct identification of EMV status occurred less frequently in left-sided colonic tumours than in right-sided tumours for both observers, which may reflect the more variable and tortuous vasculature in the left mesocolon; this would also be consistent with findings from other studies of normal mesocolonic anatomy [5]. Nodularity of the vessel involved appeared to be the most consistent feature of EMV on CT.
In rectal cancer, positive circumferential margins have been shown to be a predictor for local recurrence and systemic failure [22–25]. The equivalent non-peritonealized "bare" area or RSM has recently been described for right-sided colonic tumours, with the incidence of positive margins at 7% [26] coinciding with other reported series of colonic local recurrences of 6–10% [27, 28]. Although it has previously been reported that the length and mobility of the mesocolon is variable at different sites [29], such variations of the fused RSM can be visualized using CT. Recognition of this posterior fascia and demonstration of its involvement by tumour has not previously been evaluated using CT. Our accuracy rates for correctly predicting the RSM status were 76% and 79% for Obs 1 and Obs 2, respectively, but only three margins were involved on histological assessment. These results may have been affected by the relative ease of achieving (for rectal cancers) a clear resection margin in a colonic resection compared with the rather limited clearance potential in the pelvis. Also, the frequency of RSM involvement in colonic cancers [26] is substantially less than the circumferential resection margin involvement in rectal cancers [22] and, therefore, may prove less useful as a predictor of poor prognosis.
This pilot study has demonstrated the potential of spiral CT not just in staging local tumour extension but also in identifying other poor prognostic features known to reduce disease-free and overall survival in colonic cancers. Whilst we acknowledge that our sample size was small, which accounts for some of the wide confidence limits presented, our results have shown the potential of CT scanning in identifying poor prognostic features in colonic cancers. Further work is needed on a larger scale before widespread use could be advocated; enhanced technology with multidetector CT scanning could also improve results [14]. In the future, reliable pre-operative identification of prognostic features and stratification of patients into prognostic groups may become important, especially if the concept of neoadjuvant therapies were to be applied to colonic cancers.
Received for publication December 5, 2006.
Revision received February 8, 2007.
Accepted for publication March 14, 2007.
|
References
|
|---|
- Chawla AK, Kachnic LA, Clark JW, Willett CG. Combined modality therapy for rectal and colon cancer. Semin Oncol 2003;30:101–12.[Medline]
- Horton KM, Abrams RA, Fishman EK. Spiral CT of colon cancer: imaging features and role in management. Radiographics 2000;20:419–30.[Abstract/Free Full Text]
- McAndrew MR, Saba AK. Efficacy of routine preoperative computed tomography scans in colon cancer. Am Surg 1999;65:205–8.[Medline]
- Barton JB, Langdale LA, Cummins JS, Stelzner M, Lynge DC, Mock CN, et al. The utility of routine preoperative computed tomography scanning in the management of veterans with colon cancer. Am J Surg 2002;183:499–503.[CrossRef][Medline]
- Silverman PM, Kelvin FM, Korobkin M, Dunnick NR. Computed tomography of the normal mesentery. AJR Am J Roentgenol 1984;143:953–7.[Abstract/Free Full Text]
- Charnsangavej C, DuBrow RA, Varma DG, Herron DH, Robinson TJ, Whitley NO. CT of the mesocolon. Part 1. Anatomic considerations. Radiographics 1993;13:1035–45.[Abstract/Free Full Text]
- Quirke P, Williams GT. Minimum dataset for colorectal cancer histopathology reports. London, UK: The Royal College of Pathologists; 1997: 1–11
- Munikrishnan V, Gillams AR, Lees WR, Vaizey CJ, Boulos PB. Prospective study comparing multislice CT colonography with colonoscopy in the detection of colorectal cancer and polyps. Dis Colon Rectum 2003;46:1384–90.[CrossRef][Medline]
- Mulhall BP, Veerappan GR, Jackson JL. Meta-analysis: computed tomographic colonography. Ann Intern Med 2005;142:635–50.[Abstract/Free Full Text]
- Taylor SA, Halligan S, Saunders BP, Morley S, Riesewyk C, Atkin W, et al. Use of multidetector-row CT colonography for detection of colorectal neoplasia in patients referred via the Department of Health "2-Week-wait" initiative. Clin Radiol 2003;58:855–61.[CrossRef][Medline]
- Husband JE, Hodson NJ, Parsons CA. The use of computed tomography in recurrent rectal tumors. Radiology 1980;134:677–82.[Abstract/Free Full Text]
- Freeny PC, Marks WM, Ryan JA, Bolen JW. Colorectal carcinoma evaluation with CT: preoperative staging and detection of postoperative recurrence. Radiology 1986;158:347–53.[Abstract/Free Full Text]
- Acunas B, Rozanes I, Acunas G, Celik L, Sayi I, Gokmen E. Preoperative CT staging of colon carcinoma (excluding the recto-sigmoid region). Eur J Radiol 1990;11:150–3.[CrossRef][Medline]
- Ji H, Rolnick JA, Haker S, Barish MA. Multislice CT colonography: current status and limitations. Eur J Radiol 2003;47:123–34.[CrossRef][Medline]
- Shepherd NA., Baxter KJ, Love SB. The prognostic importance of peritoneal involvement in colonic cancer: a prospective evaluation. Gastroenterology 1997;112:1096–102.[CrossRef][Medline]
- Thoeni RF. Colorectal cancer. Radiologic staging. Radiol Clin North Am 1997;35:457–85.[Medline]
- McDaniel KP, Charnsangavej C, DuBrow RA, Varma DG, Granfield CA, Curley SA. Pathways of nodal metastasis in carcinomas of the cecum, ascending colon, and transverse colon: CT demonstration. AJR Am J Roentgenol 1993;161:61–4.[Abstract/Free Full Text]
- Granfield CA, Charnsangavej C, Dubrow RA, Varma DG, Curley SA, Whitley NO, et al. Regional lymph node metastases in carcinoma of the left side of the colon and rectum: CT demonstration. AJR Am J Roentgenol 1992;159:757–61.[Abstract/Free Full Text]
- Balthazar EJ, Megibow AJ, Hulnick D, Naidich DP. Carcinoma of the colon: detection and preoperative staging by CT. AJR Am J Roentgenol 1988;150:301–6.[Abstract/Free Full Text]
- Chapuis PH, Dent OF, Bokey EL, Newland RC, Sinclair G. Adverse histopathological findings as a guide to patient management after curative resection of node-positive colonic cancer. Br J Surg 2004;91:349–54.[CrossRef][Medline]
- Brown G, Radcliffe AG, Newcombe RG, Dallimore NS, Bourne MW, Williams GT. Preoperative assessment of prognostic factors in rectal cancer using high-resolution magnetic resonance imaging. Br J Surg 2003;90:355–64.[CrossRef][Medline]
- Adam IJ, Mohamdee MO, Martin IG, Scott N, Finan PJ, Johnston D, et al. Role of circumferential margin involvement in the local recurrence of rectal cancer. Lancet 1994;344:707–11.[CrossRef][Medline]
- Birbeck KF, Macklin CP, Tiffin NJ, Parsons W, Dixon MF, Mapstone NP, et al. Rates of circumferential resection margin involvement vary between surgeons and predict outcomes in rectal cancer surgery. Ann Surg 2002;235:449–57.[CrossRef][Medline]
- Hall NR, Finan PJ, al-Jaberi T, Tsang CS, Brown SR, Dixon MF, et al. Circumferential margin involvement after mesorectal excision of rectal cancer with curative intent. Predictor of survival but not local recurrence? Dis Colon Rectum 1998;41:979–83.[CrossRef][Medline]
- Wibe A, Rendedal PR, Svensson E, Norstein J, Eide TJ, Myrvold HE, et al. Prognostic significance of the circumferential resection margin following total mesorectal excision for rectal cancer. Br J Surg 2002;89:327–34.[CrossRef][Medline]
- Bateman AC, Carr NJ, Warren BF. The retroperitoneal surface in distal caecal and proximal ascending colon carcinoma: the Cinderella surgical margin? J Clin Pathol 2005;58:426–8.[Abstract/Free Full Text]
- Willett C, Tepper JE, Cohen A, Orlow E, Welch C, Donaldson G. Local failure following curative resection of colonic adenocarcinoma. Int J Radiat Oncol Biol Phys 1984;10:645–51.[Medline]
- Carraro PG, Segala M, Cesana BM, Tiberio G. Obstructing colonic cancer: failure and survival patterns over a ten-year follow-up after one-stage curative surgery. Dis Colon Rectum 2001;44:243–50.[CrossRef][Medline]
- Saunders BP, Phillips RK, Williams CB. Intraoperative measurement of colonic anatomy and attachments with relevance to colonoscopy. Br J Surg 1995;82:1491–3.[Medline]
Top
Abstract
Introduction
