British Journal of Radiology 75 (2002),31-37 © 2002 The British Institute of Radiology
Histopathological correlates of abnormal pericolic fat on CT in the assessment of colorectal carcinoma
C S Ng, MRCP, FRCR
1
T C Doyle, FRACR
,1
A K Dixon, FRCP, FRCR
1
R Miller, MS, FRCS
2 and
M J Arends, MD, FRCPath
3
Departments of 1Radiology, 2Surgery and 3Pathology, Addenbrooke's NHS Trust and the University of Cambridge, Cambridge, UK
Correspondence: C S Ng, Department of Radiology, Division of Diagnostic Imaging, Box 57, MD Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas 77030-4095, USA
 |
Abstract
|
|---|
The aim of this study was to assess the pathological significance of abnormal pericolic fat shown by CT in the context of colorectal carcinoma. CT and histopathological findings of 63 resected colorectal carcinomas were retrospectively reviewed. CT examinations were assessed by two observers for the presence or absence of abnormal pericolic fat (typically linear or nodular opacities) at tumour sites. Specimens were reviewed histopathologically for depth of tumour invasion, extramuscular tissue reaction, and number and largest size of tumour-involved and tumour-free lymph nodes. The sensitivity, specificity, positive predictive value and negative predictive value of pericolic fat in identifying extension of tumour infiltration beyond the muscle coat were 79% (42/53), 33% (2/6), 91% (42/46) and 15% (2/13), respectively. Despite these indicators of efficacy, the association between the presence of pericolic fat abnormality on CT and extramuscular extension of tumour (infiltration and/or nodal disease) or tissue reaction alone or in combination did not reach statistical significance (p>0.3 in all cases). Abnormal ("misty" or "mucky") pericolic fat in the assessment of colorectal cancer on CT is not a precise indicator of extramuscular extension of tumour, as it cannot clearly distinguish between tumour infiltration and tissue reaction beyond the muscle coat, or pericolic nodal involvement. However, it is a very helpful CT sign that may draw attention to the presence and site of a potential colonic abnormality.
|
Introduction
|
|---|
Abnormal pericolic fat (stranding, nodularity, "mucky" or "misty" fat [1, 2]) may be observed on CT in association with colorectal cancers. It is generally taken to be suspicious of tumour extension beyond the muscle coat and a number of studies have used it as a criterion for tumour staging [3–10]. Abnormal pericolic fat was shown to be strongly associated with the presence of tumour in a previous study from our own institution examining the efficacy of CT in identifying colorectal carcinoma in elderly patients. The pathological correlate of this radiological sign has not been previously evaluated.
The objective of this study was to assess the pathological significance of pericolic fat abnormality on CT in the presence of colorectal carcinoma.
|
Methods and materials
|
|---|
Patients
Patients who had been referred with symptoms and signs suspicious for a colorectal malignancy but who had been considered too frail to undergo barium enema or colonoscopy, and who had been investigated by the CT technique described below between January 1995 and December 1998, were identified from records. The study group consisted of the subset of patients who went on to have colorectal resections for carcinomas.
CT technique
The CT technique consisted of helical acquisition from the domes of the diaphragm to the symphysis pubis with 10 mm collimation, 1.5 pitch and reconstruction at 10 mm or 15 mm intervals and a single breath-hold in the upper abdomen. Patient preparation was 1.2 l oral contrast medium containing 20 ml meglumine and sodium amidotrizoate (Gastrografin; Schering, Berlin, Germany) in four doses over 48 h, with no requirement for bowel cleansing. Studies were carried out on Somatom Plus (Siemens, Erlangen, Germany) and High Speed Advantage (General Electric, Milwaukee, WI) helical CT machines. Intravenous contrast medium was not administered routinely, but was available for evaluation of particular problems at the discretion of the supervising radiologist (100 ml iopamidol 300 mgI ml-1, pump injection 2 ml s-1), as were thinner CT sections.
Evaluation of CT images
The CT images of the subset of patients who had undergone colorectal resection (n=63) were reviewed by consensus by two observers (CSN, TCD), who were blinded to the histopathological findings, for the presence or absence of pericolic fat abnormality as assessed against fat in a contralateral location (Figures 1a,b
). This abnormality was characterized by increased density, linear or nodular opacities, or "mistiness". Images were reviewed on a workstation at soft tissue window settings (centre +20, width 400 Hounsfield units). Assessment of pericolic fat was not possible in those cases with little intraabdominal fat (n=4).
View larger version (59K):
[in this window]
[in a new window]
|
Figure 1. (a) 70-year-old man with change in bowel habit. CT shows bowel wall thickening in the proximal transverse colon with normal pericolic fat (arrow). Histopathology of the resected specimen showed a pT4 tumour invading the omentum, together with an extramuscular tissue reaction (grade 2 fibrosis and grade 2 inflammation). (b) 75-year-old woman with anaemia. CT shows bowel wall thickening in the ascending colon with abnormal pericolic fat (arrow). Histopathology of the resected specimen showed a pT2 tumour, together with an extramuscular tissue reaction (grade 2 fibrosis and grade 1 inflammation).
|
|
Histopathological evaluation
Histopathological slides of resection specimens were retrospectively reviewed by a histopathologist (MJA) with particular expertise in colorectal carcinoma and who was not aware of the CT findings. The following parameters were documented: (a) pT stage of tumour as a measure of extent of invasion; (b) presence or absence, character, and extent of reactive tissue beyond the muscle coat immediately adjacent to the invasive cancer; and (c) lymph node involvement by metastatic tumour, number of lymph nodes, and the size of the largest lymph node whether uninvolved or involved by tumour.
The pT staging, as used in TNM colorectal cancer staging, was evaluated as follows: pT1, tumour confined to the mucosa or submucosa; pT2, tumour invading the muscle coat (muscularis propria); pT3, tumour extending beyond the muscle coat into pericolic fat; and pT4, tumour infiltrating pericolic fat and reaching the peritoneal/serosal surface [11].
The character and extent of reactive tissue beyond the muscle coat immediately adjacent to the invasive cancer was classified into three categories: (a) fibrotic reaction, indicating significant reactive fibrosis in the pericolic fat; (b) inflammatory reaction, indicating significant chronic inflammation in the pericolic fat; or (c) follicular inflammatory reaction, indicating predominantly lymphoid follicle formation, with or without germinal centres, in the pericolic fat, sometimes in association with chronic inflammatory infiltration. When present, these features were graded to indicate the extent or severity of the changes: grade 1 (mild), grade 2 (moderate) and grade 3 (severe). An overall pericolic tissue reaction "score" was derived from the sum of the grades of the three individual categories, and a cut-off value of >3 was used in further evaluation.
Statistical analysis
Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were used as measures of diagnostic accuracy. All statistical tests were performed at the 5% significance level. The association between two categorical variables was tested using Fisher's Exact test if the expected count in one or more cells was less than 5, or 
2 tests otherwise, with a continuity correction in the case of 2 x 2 tables. Non-parametric tests were used for non-normal data.
|
Results
|
|---|
63 of 98 colorectal carcinomas identified in the total group of 1077 cases were resected and could be evaluated and thus formed the basis of the current analysis. One case was excluded because the primary colonic tumour and hence the site of the appropriate pericolic region could not be identified on CT, even in retrospect. The median age was 79 years (interquartile range 16 years), and 65% (41/63) of cases were women. The median interval between CT and resection was 42 days (interquartile range 79 days). One patient had two synchronous colonic tumours, which were treated as separate cases in this analysis.
Tumours were distributed as follows: caecum (n=22), ascending colon (n=8), hepatic flexure (n=2), transverse colon (n=6), splenic flexure (n=3), descending colon (n=2), sigmoid colon (n=16) and rectum (n=4).
Abnormal pericolic fat was identified on CT in 46 (73%) cases. It was normal in 13 cases, and not assessable in 4 cases because of insufficient intraabdominal fat.
Histopathological analysis showed tumour extending beyond the muscle coat in 57 (90%) of 63 cases (pT3=38, pT4=19; Dukes' stages B or C) (Figure 2a). Tumour was confined to the muscularis propria in the remaining six cases (pT2=6, pT1=0; Dukes' stage A), but pericolic fibrotic and chronic inflammatory reactions were seen even in these cases (Figure 2b). Fibrotic reaction in the pericolic fat was evident in all cases (grade 1=3, grade 2=40, grade 3=20). This was associated with inflammatory reaction in 51 cases (grade 1=25, grade 2=17, grade 3=9) and follicular inflammatory reaction in 11 cases (grade 2=10, grade 3=1).
View larger version (75K):
[in this window]
[in a new window]
|
Figure 2. (a) 76-year-old woman with anaemia. CT showed a caecal carcinoma with abnormal pericolic fat. The photomicrograph shows infiltration of colorectal cancer (C) beyond the muscularis propria (M), i.e. stage pT3, Dukes' B, together with a grade 2 host fibrotic tissue reaction (F). There was also a grade 1 inflammatory reaction (not shown), giving an overall pericolic tissue reaction score of 3. (P, uninvolved adipocytes). This case illustrates how invasion of malignant glands excites a fibrotic reaction directly in the adjacent pericolic tissue. Scale bar (upper left)=1 mm. (b) 80-year-old woman with rectal bleeding. CT showed a sigmoid carcinoma with normal pericolic fat. The photomicrograph shows colorectal cancer (C) superficially invading the muscularis propria (M), i.e. stage pT2, Dukes' A. There was an associated grade 2 host fibrotic reaction (F) and grade 2 chronic inflammatory reaction (I) in the underlying pericolic fat (P), giving an overall pericolic tissue reaction score of 4. This case illustrates that there may be a substantial fibrotic and inflammatory response within the pericolic tissue even though the invasive cancer lies 3–4 mm remote from the pericolic fatty tissue. Scale bar (left side)=1 mm.
|
|
Lymph nodes were identified on pathological examination in 59 cases. The largest tumour-involved nodes (median 11 mm, range 4–26 mm; n=36) were significantly larger than tumour-free nodes (median 7 mm, range 2–30 mm; n=51) (p<0.001, Mann-Witney U-test), although there was a small overlap of the interquartile ranges (Figure 3). 28 of these cases had both tumour-containing and tumour-free nodes, and within this group tumour-containing nodes were also significantly larger than uninvolved nodes (p<0.001, Wilcoxon signed rank test). However, in 18% (5/28) of these cases, the largest tumour-involved nodes were smaller than the largest tumour-free nodes (Figure 4).
View larger version (22K):
[in this window]
[in a new window]
|
Figure 3. Largest lymph node size measured histopathologically by lymph node status (tumour-involved vs tumour-free). Box plot showing median, interquartile range, and maximum and minimum values.
|
|
View larger version (69K):
[in this window]
[in a new window]
|
Figure 4. (a) 51-year-old woman with right iliac fossa pain. CT showed a caecal carcinoma with abnormal pericolic fat. Histopathology of the resected specimen found 2 out of 10 lymph nodes (LN) were involved by tumour, one of which (illustrated) was partially involved by colorectal cancer (C). Despite involvement by metastatic cancer, this small node is not enlarged. The largest involved node in this case measured 8 mm, smaller than the largest uninvolved node (13 mm). Scale bar (top right)=1 mm. (b) 45-year-old woman with abdominal pain. CT showed a caecal carcinoma (there was insufficient intraabdominal fat to allow satisfactory evaluation of the pericolic fat). Histopathology of the resected specimen found 3 out of 11 nodes were involved by tumour, the largest of which was 25 mm, and 8 out of 11 uninvolved nodes, the largest of which was 30 mm, as illustrated. This photomicrograph shows 25% of the latter (tumour-free) lymph node containing reactive follicular hyperplasia. Scale bar (top right)=1 mm (same magnification as Figure 4a for comparison).
|
|
The sensitivity, specificity, PPV and NPV of abnormal pericolic fat on CT in identifying various combinations of extramuscular disease, in the form of tumour infiltration, tumour-involved lymph nodes or tissue reaction are presented in Table 1, and in greater detail in Tables 2–6.
View this table:
[in this window]
[in a new window]
|
Table 1. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of abnormal pericolic fat in identifying various pathological processes beyond the muscle coat, and assessment of association (p-values) between abnormal pericolic fat and the various histopathological findings listed below
|
|
|
Discussion
|
|---|
Abnormal pericolic fat on CT appears as linear "stands", "wisps" or nodular opacities, or simply as a "mistiness" ("haziness" or "muckiness") in the surrounding fat. Although a wide range of causes were identified when originally described [1, 2], it has become generally inferred that in the context of colorectal cancer staging its presence is suspicious for extension of tumour beyond the muscle coat (muscularis propria) and suggestive of at least Dukes' stage B, or TNM stage pT3 or pT4 disease [3–10]. Our results lend only limited support to this notion. Although the sensitivity of abnormal pericolic fat on CT as an indicator of histopathologically confirmed extramuscular tumour infiltration was 79% and the PPV was 91%, the specificity and NPV were only 33% and 15%, respectively (Tables 1 and 2).
The presence of abnormal pericolic fat on CT therefore does not necessarily indicate the presence of extramuscular extension of tumour. In these circumstances, some authors have identified "reactive" tissue in the surrounding connective tissue on histopathology [6, 8, 10]. In our study, specificity and PPV improve (50% and 98%, respectively) when this possibility is added to the analysis (Table 5). However, caution should be exercised in using this interpretation, as a variable degree of extramuscular tissue reaction was identifiable on histopathology in all cases. Reactive fibrosis and chronic inflammation (Figure 2) adjacent to invasive cancer are long established host response phenomena, mostly resulting from a combination of tissue destruction by invasive islands and tongues of cancer cells as well as secretion of a mixture of cytokines both by tumour cells and host cells such as tissue macrophages. An inflammatory fibroblastic stromal response may also be excited following local perforation, as identified histopathologically in one of our cases. Lymphatic congestion following lymphatic metastasis and hyperaemia around tumours may also contribute to a complex tissue reaction [9]. Lymphoid follicular aggregates in response to colorectal cancer are associated more with mismatch repair defective colorectal carcinomas and so tend to be identified more often in right-sided tumours [12].
Conversely, normal pericolic fat on CT is not a reliable indicator of the absence of extramuscular extension of tumour. The NPV was only 15% (Table 2). This is perhaps not surprising, as CT cannot compete with histopathology in assessing microscopic degrees of extramuscular extension of tumour. However, high resolution MRI is showing promise in the area of rectal carcinoma [13].
The accuracy of pericolic fat in assessing tumour beyond the muscle coat, in the form of tumour-involved nodes alone or in combination with extramuscular extension of tumour, was not substantially different from the latter alone (Tables 3 and 4). Overall, pericolic fat abnormality on CT was not significantly associated with extension of disease in the form of tumour or reactive tissue.
Our data are in concordance with the generally held view that tumour-involved lymph nodes are typically larger than uninvolved nodes [6]. However, there was overlap between the two groups, with some small nodes containing tumour and conversely some large nodes being free of tumour (Figures 4a,b). Indeed, in those cases in which both tumour-involved and tumour-free nodes were identified, the former were smaller than the latter in as many as 18% of cases. Such difficulties are compounded by partial infiltration of lymph nodes, which may not necessarily enlarge the node, as was observed in some of our cases (Figure 4a). Conversely, reactive changes may enlarge otherwise tumour-free nodes. Indeed, some degree of reactive change, such as reactive follicular hyperplasia or sinus histiocytosis, was identified histopathologically in all the lymph nodes in our series that were free of tumour.
In common with other investigators evaluating colorectal and other malignancies, for example pelvic, gastric or lung cancers and lymphoma, our data suggest that nodal size is not a robust indicator of nodal status [3, 5–7, 9, 11, 14–16]. CT has further difficulties in visualizing locoregional nodes in the first place.
A limitation of this study was that it was retrospective and areas of pericolic fat assessed on CT may not necessarily have corresponded with those from which blocks had been taken for microscopy. There is also the possibility of sampling errors at histopathology. Summing the individual tissue reaction grades to derive an "overall" pericolic tissue reaction score and using a cut-off point for determining the presence or absence of tissue reaction beyond the muscle coat was to some extent arbitrary. In this study, CT reconstruction intervals were comparatively wide, limiting spatial resolution and increasing partial volume effects. More modern machines, particularly those with multidetector capability, would allow improvement in these aspects.
Another limitation was that the number of cases was not large and the case mix was somewhat skewed: there were a comparatively large proportion of cases with extramuscular tumour on histopathology (90%) and abnormal pericolic fat on CT (73%). The former reflects the selection criteria in our study, namely symptomatic elderly patients in whom tumours are inevitably comparatively advanced. Furthermore, there were variable delays between CT and surgical resection, reflecting pressures on investigational and surgical resources at our institution.
Evaluation of the presence of pericolic fat abnormality undoubtedly has a degree of subjectivity, but this can be reduced by using the "internal control" of intraabdominal fat in other locations, most usefully in a contralateral region. In this study we attempted to reduce this qualitative aspect further by utilizing two observers.
Abnormal pericolic fat is seen on CT in a number of pathological processes affecting the colon apart from tumours, for example inflammatory conditions, such as diverticulitis, appendicitis and colitis [9, 17–20]. It may also be manifest in other abdominal conditions including pancreatitis and renal obstruction [1, 2, 21]. The pathological significance of abnormal fat in this wide spectrum of conditions has not been evaluated in this study.
Our study suggests that abnormal pericolic fat abnormality as identified on CT is not a reliable sign of extramuscular extension of tumour. Nevertheless, pericolic fat abnormality is a very helpful CT sign in alerting the radiologist to the possibility and location of a potential colonic abnormality.
|
Footnotes
|
|---|
Current address for T C Doyle: Department of Radiology, Healthcare Otago, Dunedin Hospital, Dunedin, New Zealand. 
Received for publication June 15, 2001.
Revision received August 8, 2001.
Accepted for publication August 14, 2001.
|
References
|
|---|
-
Dixon AK, Nightingale RC. Abnormal fat: a useful marker of intra-abdominal disease at computed tomography. Clin Radiol 1984;35:469–73.[Medline]
-
Mindelzun RE, Jeffrey RB, Lane MJ, Silverman PM. The misty mesentery on CT: differential diagnosis. AJR 1996;167:61–5.[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]
-
Balthazar EJ, Megibow AJ, Hulnick D, Naidich DP. Carcinoma of the colon: detection and preoperative staging by CT. AJR 1988;150:301–6.[Abstract/Free Full Text]
-
Thoeni RF. CT evaluation of carcinomas of the colon and rectum. Radiol Clin North Am 1989;27:731–41.[Medline]
-
Angelelli G, Macarini L, Lupo L, Caputi-Jambrenghi O, Pannarale O, Memeo V. Rectal carcinoma: CT staging with water as contrast medium. Radiology 1990;177:511–4.[Abstract/Free Full Text]
-
Gazelle GS, Gaa J, Saini S, Shellito P. Staging of colon carcinoma using water enema CT. J Comput Assist Tomogr 1995;19:87–91.[Medline]
-
Okizuka H, Sugimura K, Shinozaki N, Watanabe K. Colorectal carcinoma: evaluation with ultrafast CT. Clin Imaging 1995;19:247–51.[Medline]
-
Harvey CJ, Amin Z, Hare CMB, Gillams AR, Novelli MR, Boulos PB, et al. Helical CT pneumocolon to assess colonic tumours: radiologic–pathologic correlation. AJR 1998;170:1439–43.[Abstract/Free Full Text]
-
Hundt W, Braunschweig R, Reiser M. Evaluation of spiral CT in staging of colon and rectum carcinoma. Eur Radiol 1999;9:78–84.[Medline]
-
Husband JES, Reznek RH. Imaging in oncology (1st edn). Oxford: Isis Medical Media, 1998.
-
Harrison JC, Dean PJ, El-Zeky F, van der Zwaag R. From Dukes through Jass: pathological prognostic indicators in rectal cancer. Hum Pathol 1994;25:498–505.[Medline]
-
Brown G, Richards CJ, Newcombe RG, Dallimore NS, Radcliffe AG, Carey DP, et al. Rectal carcinoma: thin-section MR imaging for staging in 28 patients. Radiology 1999;211:215–22.[Abstract/Free Full Text]
-
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]
-
Webb WR, Gatsonis C, Zerhouni EA. CT and MR imaging in staging non-small cell bronchogenic carcinoma: report of the Radiologic Diagnostic Oncology Group. Radiology 1991;178:705–13.[Abstract/Free Full Text]
-
Nielsen MB, Qvitzau S, Pedersen JF. Detection of pericolonic lymph nodes in patients with colorectal cancer: an in vitro and in vivo study of the efficacy of endosonography. AJR 1993;161:57–60.[Abstract/Free Full Text]
-
Johnson CD, Carlson HC, Taylor WF, Weiland LP. Barium enemas of carcinoma of the colon: sensitivity of double- and single-contrast studies. AJR 1983;140:1143–9.[Abstract/Free Full Text]
-
Birnbaum BA, Balthazar EJ. CT of appendicitis and diverticulitis. Radiol Clin North Am 1994;32:885–98.[Medline]
-
Rao PM, Rhea JT, Novelline RA, Dobbins JM, Lawrason JN, Sacknoff R, et al. Helical CT with only colonic contrast material for diagnosing diverticulitis: prospective evaluation of 150 patients. AJR 1998;170:1445–9.[Abstract/Free Full Text]
-
Chintapalli KN, Esola CC, Chopra S, Ghiatas AA, Dodd III GD. Pericolic mesenteric lymph nodes: an aid in distinguishing diverticulitis from cancer of the colon. AJR 1997;169:1253–5.[Abstract/Free Full Text]
-
Smith RC, Rosenfield AT, Choe KA, Essenmacher KR, Verga M, Glickman MG, et al. Acute flank pain: comparison of non-contrast-enhanced CT and intravenous urography. Radiology 1995;194:789–94.[Abstract/Free Full Text]
This article has been cited by other articles:
|
|
|
|
 
T H Saunders, H K Mendes Ribeiro, and F V Gleeson
New techniques for imaging colorectal cancer: the use of MRI, PET and radioimmunoscintigraphy for primary staging and follow-up
Br. Med. Bull.,
December 1, 2002;
64(1):
81 - 99.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
