| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Memorial Sloan-Kettering Cancer Center, 1275 York
Ave., New York, NY 10021.
2 Department of Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer
Center, New York, NY 10021.
Received August 13, 2001;
accepted after revision October 24, 2001.
Address correspondence to M. J. Gollub.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. After colonic cleansing, 61 patients (16 men and 45 women; mean age, 64 years; age range, 27-81 years) underwent 63 virtual colonoscopy examinations after using either single- or multidetector CT (slice thickness, 3.75-5.0 mm; table speed, 1.7-11.25 cm/sec; pitch, 1.5-3.0; and overlapped reconstructions, 1.95-2.5 cm) in supine and prone positions after IV administration of 1 mg of glucagon and rectal air insufflation. Conventional two-dimensional axial images were analyzed on a PACS (picture archiving and communication system) workstation. Two radiologists, who were unaware of patient history, independently evaluated the colonic distention on a 4-point scale (4 = optimal distention) and fluid retention on a 3-point scale (3 = no fluid) for all segments of the colon in patients who were imaged in both the supine and prone positions. Segmental and total average colon scores were calculated.
RESULTS. Forty-one patients (65%; 43 examinations, 67%) underwent prior surgery, radiation, or both (surgery, n = 29; radiation, n = 3; both, n = 11). The average overall colonic distention and fluid retention for this group was 3.13 and 2.38, respectively, versus 3.24 and 2.3 in the control group (p = not significant).
CONCLUSION. High-quality examinations were achieved in patients who had previously undergone radiation, surgery, or both with no clinically significant difference in distention or fluid retention compared with the controls.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Incomplete colonoscopy is one clinical setting in which the use of virtual colonoscopy is finding favor among gastroenterologists and radiologists [4,5,6]. Occurring in 4-29% of all colonoscopies [7], incomplete colonoscopy is usually due to tortuosity, colonic redundancy, poor colonic cleansing, or diverticulosis [8]. Previous surgery [9] and radiation [10] are additional explanations among the list of known causes.
For half a century, the barium enema has been the radiologic gold standard for the diagnosis of colonic abnormalities. Numerous publications describe the usefulness of the barium enema for achieving a total colon examination after incomplete colonoscopy. Most reports discuss completion rates and the capability for same-day performance after incomplete colonoscopy [11, 12]. Others discuss additional findings not seen on colonoscopy [13]. One study boldly suggests that the barium enema should not be used for surveillance after polypectomy as long as the more accurate colonoscopy is available [14]. The authors of this study, however, concede that in cases of incomplete colonoscopy, follow-up with the barium enema is recommended [15]. Unlike the occasional difficult colonoscopy, tortuosity, prior radiation, or surgery seem to pose little impediment to the completion of a good quality barium enema [9].
A sizable number of incomplete colonoscopies at our cancer center are due to prior radiation or surgery. The inability to safely complete colonoscopy is presumably caused by stiffening of the colon and its mesentery. At our institution, virtual colonoscopy has become the preferred method for total examination of the colon after incomplete colonoscopy; thus, we have had the opportunity to examine a number of patients who, for a variety of reasons, have undergone incomplete colonoscopy.
To our knowledge, no study has compared the quality and outcome of virtual colonoscopy after incomplete colonoscopy stratified according to the cause of incomplete colonoscopy. It is possible, for example, that not every failed colonoscopy should be followed by virtual colonoscopy and that some patients may, in fact, fare better with double-contrast barium enema, a concept that is undergoing further research in our department.
Our investigation was intended to answer the question of whether surgery or radiation presents a limitation to the performance of a high-quality virtual colonoscopy examination compared with other causes of incomplete colonoscopy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
All patients were scanned in the prone and then the supine position, and
all patients received 0.9 mg of IV glucagon hydrochloride immediately before
air insufflation. Fourteen patients received 150 mL of IV nonionic contrast
medium while in the supine position. A fellow or attending radiologist
inserted a 16-gauge flexible rubber catheter 6 inches (15 cm) into the rectum
followed by the insufflation of air using a blue manual compression bulb to
patient tolerance (
25-50 pumps, 1-2 L).
Initial and repeated scout images were obtained to assess the quality of air distention. Insufficient colonic distention led to additional attempts at insufflation.
Interpretation
All examinations had been previously officially interpreted and reports
dictated after analysis of two-dimensional (2D), 2D reformatted, and
three-dimensional perspective volume-rendered (virtual) images. Unaware of
patient medical and surgical history, two board-certified radiologists, with
experience interpreting virtual colonoscopies, retrospectively performed
independent analyses of 2D images on a picture archiving and communications
system (General Electric PACS; General Electric Medical Systems). Prone and
supine 2D data sets were evaluated for degree of colonic air distention and
degree of colonic fluid or stool retention.
Distention was subjectively rated 1-4 as follows: no distention, 1; poor distention, 2; good distention, 3; and optimal distention, 4 (Fig. 1A,1B,1C,1D). A high-quality examination was defined as having a distention rating of 3 or 4. Fluid was subjectively rated 1-3 as follows: more than one half of the luminal diameter fluid-filled, 1; less than one half of the luminal diameter fluid-filled, 2; no fluid or stool, 3 (Fig. 2A,2B,2C).
|
|
|
|
|
|
|
A high-quality examination was defined as having a fluid rating of 2 or 3. Ratings were applied to eight colonic segments, including the cecum, ascending colon, hepatic flexure, transverse colon, splenic flexure, descending colon, sigmoid, and rectum. The rating of the least distended area of each segment was recorded as the overall rating of that segment. The rating of the most fluid- or stool-filled area of a segment was recorded as the overall rating of that segment.
Adequate visualization of colonic segments required that, with the patient in at least one position (either prone or supine), distention was rated 3 or 4 and fluid was rated 2 or 3. If a given segment or focus within a given segment was given a distention rating of 1 or 2 or a fluid rating of 1, it was qualified with a plus (+) or a negative (-), depending on whether the same segment in the other body position was adequately (distention = 3 or 4 and fluid = 2 or 3) or inadequately (distention = 1 or 2 and fluid = 1) visualized, respectively (Figs. 3A,3B and 4A,4B).
|
|
|
|
Segments in which the entire segment or a portion of the segment received a negative qualifier in both positions (double negative [-/-]) were considered to have been inadequately evaluated.
Finally, identification of the ileocecal valve or the appendix outlined by air was recorded in each patient to evaluate completeness of colon visualization.
Data Calculations
The following four subgroups of patients were formed: all (all patients),
none (no treatment group), surgery (surgery only group), and both (surgery and
radiation treatment group). Three patients received radiation only, for which
a separate group was not formed. Surgery was defined as a procedure requiring
laparotomy. Laparoscopic procedures were excluded, as were herniorrhaphies.
Radiation included external beam therapy and brachytherapy.
In each group, mean values were calculated for distention and for fluid or stool retention for the whole colon by both observers and separately by each observer (Table 1). Mean and individual observer distention and fluid ratings were separately tabulated for the radiation-exposed rectum (Table 2). For all of the values mentioned previously, statistical significance of differences between patient groups was calculated using the analysis of variance F test.
|
|
The total number of double-negative qualifier segment examinations was tallied and compared across the four groups and summarized as to explainable or inexplicable cause (Table 3). Statistical significance was calculated using the exact tests for contingency tables.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Fifty (82%) of 61 patients had a history of malignancy, including 15, colorectal; nine, breast; six, ovarian; four, prostate; four, cervix; three, uterine; two, lymphoma; two, bladder; two, kidney; two, skin; and one each of esophageal, stomach, bone, bile duct, fallopian tube, myelodysplastic syndrome, and urethral cancer. Two of those patients had two or more cancers. Forty patients underwent surgery (surgery only, 29; combined with radiation, 11), including 21 hysterectomies, nine appendectomies, six colectomies, five cholecystectomies, two nephrectomies, two omentectomies, two splenectomies, two pelvic exenterations, two caesarian deliveries, two oophorectomies, and one each of the following: prostatectomy, gastrectomy, small-bowel resection, Whipple's procedure, and repair of colon perforation. Twelve patients had two or more surgical procedures. Fourteen patients underwent pelvic radiation. In three patients, only radiation was administered.
The mean values for distention ratings of the whole colon for the 61 patients were 3.29 and 3.51 for observers 1 and 2, respectively, and did not differ significantly from the means for patients who received no treatment (2.90 and 3.58), patients who had surgery only (2.78 and 3.49), or patients who had both surgery and radiation (2.76 and 3.5) (Table 1).
The mean value for distention ratings of the rectum for the 61 patients was 3.09 (Table 2). The prone rectum distention value for the surgery and radiation treatment groups was significantly lower (p < 0.001) statistically than the ratings for prone rectal distention in the surgery and radiation treatment group.
Mean prone distention values for each colonic segment in a given group (i.e., no treatment, surgery only, and surgery and radiation) are illustrated in Figures 5 and 6. For both observers, significant differences were noted in the distention values of the rectum among treatment groups (p < 0.001). No significant differences in distention were noted among groups on supine scans (not shown).
|
|
Note that high-quality ratings were achieved for fluid (2 or 3) and distention (3 or 4) in all groups when considering the mean ratings of both observers for the whole colon and rectum. Values less than high quality were recorded for the whole colon and rectum by observer 1 and for the rectum by both observers (mean). Values of observer 1 were lower for all categories, groups, and segments than those of observer 2. When using univariate p values to determine an effect between the treatment group and the observer, we found no correlation, and thus the variation in overall ratings applied by the observers was irrespective of treatment or no treatment.
Twenty-one segments (17 patients) were inadequately distended on scans obtained with the patients in both prone and supine positions (double negative). Carcinoma or stricture explained four segments (four patients). In 13 unexplained cases, 17 inadequately distended segments were distributed as follows: nontreatment group, two (9.5%) of 21 patients; surgery only group, seven (24%) of 29 patients; and surgery plus radiation group, four (36%) of 11 patients. The differences were not statistically significant for a two-sided p value using Spearman's rank correlation coefficient (Table 3).
In 57 (90%) of 63 examinations, the ileocecal valve (or ileocolic anastomosis) was identified on 2D images outlined by air. In 13 (21%) of 63 examinations, the appendiceal orifice was outlined by air. Cases in which neither the ileocecal valve nor the appendiceal orifice was visible were explained by excess fluid, luminal collapse, tumor, deep pelvic location of the cecum, or ventral hernia (Table 4).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Radiation therapy is commonly used to treat many abdominopelvic malignancies. Therapeutic levels are known to cause mesenteric and submucosal enteric fibrosis [16, 17]. Both conditions would be expected to decrease flexibility and pliability of the involved intestine compared with its natural state.
In addition to clinical factors such as pain tolerance and the ability to be sedated and manipulated with an invasive instrument, several anatomic factors related to the colon are required for successful performance of conventional colonoscopy, including adequate luminal diameter devoid of large or numerous false channels (diverticula); mobility of the colon to allow straightening over the colonoscope; and turns and curves that are not too angulated. Thus, it is easily understood that patients who have undergone radiation or surgery or those who have advanced diverticular disease could present an anatomic challenge to the passage of a colonoscope. In these patients, for the reasons stated, failure of completion of the examination might be anticipated.
Our hypothesis is that although a history of surgery and radiation may interfere with performance and completion of conventional colonoscopy, so that not all the colonic mucosa is examined from the rectum to the ileocecal valve, no such limitation would be encountered in virtual colonoscopy because no scope is inserted into the colon, and fewer anatomic prerequisites are necessary for a successful examination. Specifically, surgical adhesions that potentially cause angulation, radiation-induced mesenteric and submucosal fibrosis that potentially diminishes distensibility, and focal lack of distensibility would not reduce the quality of the examination of the colon by virtual colonoscopy.
In support of this hypothesis, mean values for whole colonic distention of all patients ranged from 3.29 to 3.51. Although the values are lower in the group who received both radiation and surgery and in the group who received surgery alone than in the nontreatment group, this was not statistically significant. Furthermore, all values met the criteria for high quality, and thus differences were also not clinically significant (Table 1).
For individual observers, a similar trend was noted, with lower distention ratings in the treatment groups and higher ratings in the nontreatment group. Again, differences were not significant. Observer 2 was found to be more exacting across all segments and groups than observer 1. Univariate p values testing for an association between the differences in observer ratings across different groups showed that this effect was independent of the group being rated.
Initially, values for the rectosigmoid were calculated from a combination of rectum and sigmoid values for the purpose of extracting that part of the colon that might be most exposed to radiation. Statistical analysis showed no differences among the groups. A separate analysis of rectal and sigmoid differences, however, showed that the distention ratings of the rectum with the patient in the prone position (most reliable position for rectal distention) in patients who had undergone both surgery and radiation were significantly lower than for the surgery alone group and the no treatment group (p < 0.001) (Table 2, Figs. 1A,1B,1C,1D and 2A,2B,2C). This effect could be explained by the effects of radiation and surgery on the rectum and might be absent for the overall rating of the sigmoid colon because much of this segment may lie outside the typical radiation portal. It would have been interesting to divide these segments of the colon into rectosigmoid colon (colon contained in the true pelvis or whole pelvis, where typical radiation portals are applied) and extrapelvic sigmoid colon. However, the values achieved for distention, nonetheless, were 3-4 and thus still of high quality.
Double-negative values for a given segment, indicating inadequate distention on both prone and supine scans, were caused by neoplasm in three patients, benign stricture in one patient, and an unknown cause in 13 patients. A modest trend of increasing numbers of patients in whom inadequately distended segments were noted occurred from the nontreatment group (9.5%) to the surgery alone group (24%) to the double treatment group (36%). Although the increase in patients was suggestive of an effect of treatment on examination adequacy, the number of cases is small, and there was no statistical significance (Table 3).
Similar to the standard used in conventional colonoscopy of the identification of cecal landmarks and the ileocecal valve to establish a complete examination, we were able to identify these landmarks in 57 (90%) of 63 cases outlined by air on the 2D images, and these segments were thus navigable using software for three-dimensional viewing. However, clearly, we are not reaching the ileocecal valve in each patient on the basis of a complete fly-through from the rectum with our virtual endoscope. Nonetheless, as a separate measure of quality control, it is of interest and may have implications for appropriateness of virtual colonoscopy to examine the cecum, ileocecal valve, and terminal ileum in patients with neoplasm.
Not surprisingly, the distention for various colonic segments varies with prone and supine postioning. For example, prone images are better for visualizing the rectum, ascending colon, descending colon, pelvic sigmoid, and, sometimes, cecum, depending on its location, whereas supine positioning is better for visualizing the transverse colon, upper sigmoid, and cecum. The necessity for both positions has been firmly established in the literature [18].
We rated fluid contents of the bowel and found no significant difference in patients across groups. The similar values noted indicate the independence of fluid retention from the history of treatment. Fluid retention depends on other factors such as compliance and successful colonic cleansing preparation, over which we have no control.
One limitation of this investigation is the subjective rating of distention. Rather than measuring the diameter of the loop under question, we used a 4-point Likert scale (none, poor, good, and excellent). The subjective rating of distention is clearly a matter of opinion that will vary with observer. Additionally, most of our patients represented a select population of individuals who have undergone incomplete colonoscopy. The quality of air distention noted in their colons may not necessarily reflect that achievable in a screening population.
No attempt was made to determine the accuracy of lesion detection in this study because infrequent pathologic correlation was available. Accuracy will be addressed in a similar surveillance population in a study that is ongoing at our institution.
Our results suggest several factors about the treatment of patients undergoing investigation of their colons and also opens other avenues for further research.
First, although most patients who undergo colonoscopy have a complete examination, those with prior radiation or prior surgery may not. Comparatively speaking, the literature does not mention the proportion of this subset of patients who will have unsuccessful examinations.
Second, although failure rates vary from institution to institution and among examiners, as well as among examinations for a given patient, in most cases, failure is due to colonic redundancy or tortuosity, diverticular disease, or a history of abdominopelvic surgery or radiation. Because our results suggest that no statistically significant qualitative difference exists in the virtual colonoscopy examinations of patients with or those without a history of prior radiation or surgery, and completion of a good-quality virtual colonoscopy is possible in most cases, these patients can be referred to virtual colonoscopy with confidence that completion of their total colon examination will be achieved on the same day.
Third, in patients who need repeated colonic evaluation, who have a history of radiation or surgery, and who are undergoing surveillance after removal of known polyps or colon cancer, virtual colonoscopy will succeed, whereas endoscopic colonoscopy could fail (the delayed repeated colonoscopy failure rate after one failed attempt is 50%) [19].
Finally, if virtual colonoscopy is shown to be a useful screening test in the general population, those patients with a history of radiation or surgery for noncolorectal cancer, in choosing among available screening tests, may opt to undergo virtual colonoscopy because their treatment history may put them at risk for incomplete conventional colonoscopy [8]. We have noted that the quality of examinations subjectively differs in patients in whom the cause of incomplete colonoscopy was redundancy or diverticular disease. In some of these cases barium enema follow-up was needed. No formal evaluation or comparison of these subsets was performed. However, this observation and question are leading to further investigation in our department. We will attempt to better define the proper triage of patients to the examination most likely to succeed after a failed colonoscopy and will not assume that virtual colonoscopy is the choice in all patients.
In conclusion, high-quality virtual colonoscopy examinations were achieved in all patients who previously underwent radiation or surgical treatment. Conventional colonoscopy failed in some of these patients because of this history. The completion rates and overall quality of colonic distention did not differ from patients who had no prior treatment. Thus, we recommend performance of virtual colonoscopy in these patients. We also recommend virtual colonoscopy in patients after previous failed surveillance colonoscopy and, potentially in the future, in those patients contemplating a screening examination of the colon in which a personal history of surgery or radiation treatment may be considered a risk factor for incomplete colonoscopy.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M. Macari and E. J. Bini CT Colonography: Where Have We Been and Where Are We Going? Radiology, December 1, 2005; 237(3): 819 - 833. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
