| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Imaging, Royal Marsden Hospital, Downs Rd., Sutton, Surrey,
United Kingdom SM2 5PT.
2 Department of Imaging, Mount Vernon Cancer Centre, Northwood, Middlesex,
United Kingdom HA6 2RN.
3 Imaging Research, EPIX Pharmaceuticals, Inc., Cambridge, MA.
4 Department of Clinical Oncology, Mount Vernon Cancer Centre, Northwood,
Middlesex, United Kingdom HA6 2RN.
Received November 10, 2005;
accepted after revision February 28, 2006.
Address correspondence to S. D. Allen
(sdallen{at}doctors.org.uk).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. High-resolution T2-weighted MRI was performed before and immediately after a standardized 5-week course of chemoradiation therapy in the care of 30 patients with locally advanced adenocarcinoma of the rectum. Changes in morphologic features were evaluated with respect to primary tumor and nodal downstaging. The MRI findings after chemoradiation therapy were compared with the histologic findings in the resected specimens with respect to prediction of tumor stage and showing the relation between the tumor and the circumferential margin of resection.
RESULTS. Tumor shrinkage > 30% was found in 19 (63%; 95% CI, 46-81%) of 30 patients, but changes in MRI T stage occurred in only five (17%; 95% CI, 3-30%) of 30 patients. Tumor regression from the circumferential resection margin was found in five patients, all findings confirmed at histologic examination. Nodal downstaging was observed in 13 (68%; 95% CI, 48-89%) of 19 patients; 11 patients were node free on the basis of both MRI findings and subsequent histologic results. Overall prediction of distance between tumor and circumferential resection margin was good, with a mean difference of -0.2 mm and an interclass correlation coefficient of 0.74. MRI was not useful for gauging disease activity of persistent abnormalities in mucinous tumors that often represented inactive mucin lakes.
CONCLUSION. Decreases in tumor size and nodal downstaging can be seen on MRI after chemoradiation therapy in approximately two thirds of patients. The surgically more relevant parameter—distance between tumor and circumferential resection margin—can be accurately predicted. Errors were caused by the presence of considerable tumor, rectal wall fibrosis, and mucinous tumors.
Keywords: chemoradiation • MRI • rectal carcinoma
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Neoadjuvant chemoradiation therapy is often used in the care of patients with extramural spread of rectal cancer in whom the circumferential resection margin may be at risk of involvement at surgery. Preoperative chemoradiation therapy has been the subject of many clinical trials and in Europe has become standard adjunctive preoperative treatment of patients in whom there is a high likelihood of not achieving curative resection [6, 7]. Such patients have locally advanced disease, have extramural tumor abutting the levator muscles, have a tumor that lies too close to the external anal sphincter for which sphincter-sparing surgery is being considered, or have a tumor potentially infiltrating the intersphincteric space. Different radiation treatment regimens are used and generally are classified as short- or long-course therapies.
The role of imaging appears to be changing, with a shift in emphasis toward preoperative assessment with MRI [8, 9]. This approach has been widely adopted and used for some time in Europe and Japan and is being increasingly adopted in the United States, notably in a number of major cancer centers. Patients previously were considered for surgery without undergoing preoperative crosssectional pelvic imaging. With advances in MRI techniques, however, such as the use of endorectal and phased-array coils, spatial resolution has dramatically improved [10, 11]. Current evidence suggests that MRI is the most accurate technique for predicting tumor stage [12]. Moreover, it has been repeatedly suggested that preoperative MRI prediction of the distance between the tumor and the circumferential resection margin, and hence of tumor-free total mesorectal excision, reduces the frequency of finding involved surgical margins at pathologic examination [13]. This benefit should lead to reduced rates of local tumor recurrence and lower patient morbidity.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We identified 30 subjects (20 men, 10 women; mean age, 59 years; age range, 21-76 years) with high-quality MR images before and after chemoradiation therapy (48 patients had no MRI assessments before or after chemoradiation therapy and so were excluded, one patient had inadequate pretreatment MRI, one patient an inadequate posttreatment MRI) and adequate surgical and histologic records confirming complete surgical resection (16 patients had inadequate surgical or histologic records or did not undergo surgery). The standardized chemoradiation therapy regimen was 45 Gy in 25 fractions over 33 days with IV infusions of 5-fluorouracil (350 mg/m2) on days 1-5 and days 29-33. The median time between the paired MRI examinations was 112 days (range, 85-183 days), and the median time between completion of treatment and second MRI examination was 38 days (range, 20-81 days). The median time between preoperative MRI and surgery was 43 days (range, 11-89 days).
MRI Technique
MRI was performed with a 1.5-T whole-body MR imager (Symphony, Siemens
Medical Solutions) with a pelvic phased-array coil. An endorectal coil was not
used, and all patients underwent the same imaging protocol. The examination
protocol did not change over the 4-year period in terms of imaging coils and
sequences. The techniques used are still considered applicable for 1.5-T
systems.
Patients received a bowel relaxant before imaging, either hyoscine-N-butyl bromide 20 mg or glucagon 1 mg intramuscularly. The former was administered to most patients except those with known cardiac disease, prostatic enlargement, or glaucoma. After initial localization imaging, images of the pelvis and rectum were acquired as follows. Imaging began with axial whole-pelvis spinecho T1-weighted and T2-weighted sequences below the L4-L5 disk and covered the entire pelvis. The imaging parameters for the T1-weighted sequences were TR/TE, 500/20; flip angle, 90°; echotrain length, 1. The parameters for the T2-weighted sequences were 6,000-6,500/130; flip angle, 175°; echo-train length, 23. Slice thickness was 6 mm; field of view, 350 cm; and matrix size, 256 x 512.
T2-weighted sequences with a smaller field of view were then performed in the coronal and sagittal planes (Fig. 1). The imaging parameters for the coronal sequences were 4,500-5,000/135; flip angle, 180°; echo-train length, 23. The parameters for the sagittal sequences were 6,500-7,000/135; flip angle, 180°; echo-train length, 23. Slice thickness was 3 mm; field of view, 20 cm; and matrix size, 292 x 512. The sagittal T2-weighted images were used for planning T2-weighted thin-section axial images through the rectal cancer and adjacent perirectal tissue. The imaging plane of the axial image was perpendicular to the long axis of the rectum as described by Brown et al. [14]: 7,000-7,500/130; flip angle, 180°; echo-train length, 23; slice thickness, 3 mm; field of view, 20 cm; matrix size, 318 x 512.
MRI Interpretation
MRI staging of tumors was undertaken by two oncology-trained radiologists
with 4 and 7 years of pelvic MRI experience. The images were assessed on
hard-copy film independently and in consensus. Analysis was performed on
hard-copy film for organizational reasons due to unavailability of a PACS. All
measurements were made to the nearest millimeter using the scale on the
images. All image analysis was completed without knowledge of the pathology
results.
Assessments consisted of tumor localization, measurement of longest dimension, and evaluation of signal intensity and morphologic type. A tumor was defined as low rectal if its lowest aspect was less than 5 cm above the anal verge and as midrectal if its lowest aspect was 5-10 cm above the anal verge (Fig. 2). A tumor was defined as high rectal if its lowest aspect was more than 10 cm above the anal verge or if it was above the peritoneal reflection. A tumor was classified as being largely mucinous if more than 50% of the tumor was of greater signal intensity than the obturator internus muscle [15] (Fig. 3).
|
|
|
When an enlarged lymph node (defined later) or extramural tumor deposit was located nearer to the mesorectal fascia than the primary tumor, this shorter distance was measured to the circumferential resection margin (Fig. 2). When the tumor reached or penetrated the mesorectal fascia or was less than 2 mm from the circumferential resection margin or invaded adjacent organs or peritoneum (stage T4), 0 was recorded for distance to the circumferential resection margin, because such a margin is considered involved or surgically threatened. When a tumor was MRI stage T1 or T2, the distance to the circumferential resection margin was not recorded because it was not considered relevant. Particular care was taken to differentiate as accurately as possible posttreatment fibrosis from active extramural tumor. Fibrosis has lower signal intensity and is spiculated. Active tumor was predicted if there was any degree of broad-based pushing or nodular configuration of the lateral tumor margin [8].
Regression of tumor after treatment was defined as a greater-than-2-cm increase in distance between the tumor edge and the circumferential resection margin. The tumor was considered to have decreased in size when the longest dimension of the MRI-measurable tumor decreased more than 30%. Tumor downstaging was defined as reduction in T stage of at least one category (e.g., T4 to T3/T2; T3 to T2). Changes within a T-stage category (e.g., T3c to T3a) were not considered downstaging. Similarly, nodal downstaging was a reduction in N-stage category (Appendix 1). A prediction was made about whether R0 resection would be possible. R0 resection is defined as complete resection with histologically negative margins and no residual tumor after resection [16].
The number and characteristics of pelvic lymph nodes were reviewed. Maximum short-axis nodal length was recorded. A node was considered enlarged if the length was more than 5 mm (mesorectal), 7 mm (internal iliac), 10 mm (external iliac), or 9 mm (common iliac) [17]. Locoregional nodal involvement was recorded. For rectal carcinoma, involvement was designated perirectal, sigmoid mesenteric, inferior mesenteric, lateral sacral, presacral, internal iliac, sacral promontory, superior rectal, middle rectal, or inferior rectal [16]. Lymph node signal intensity was recorded, as was border, which was classified as smooth and well defined or as irregular and ill defined [18].
Histologic Review
Type of surgery (anterior or abdominoperineal resection) and whether total
mesorectal excision was performed were recorded. Nineteen patients underwent
abdominoperineal resection, and 11 patients underwent sphincter-sparing
anterior resection. Surgical complications were documented, as was the need
for more extensive nonstandard resection at operation. Total mesorectal
excision was documented in 20 patients. In the cases of the other 10 patients,
we were unable to confirm or refute whether total mesorectal excision was
used. All surgeons involved were trained in total mesorectal excision, and it
seems likely that total mesorectal excision was performed in all cases.
The histologic reports were reviewed, although no special central specimen review was performed by a pathologist. The histologic examination was standardized: The circumferential resection planes of the specimen were inked and then opened anteriorly proximal to the tumor before being fixed in buffered formalin for a minimum of 24 hours. The tumor was sectioned transversely at 3-mm intervals according to the method described by Quirke et al. [2]. Pathologic tumor and nodal staging was performed according to the TNM system [16] along with assessment of tumor size and maximum depth of extramural tumor spread. The shortest relation between tumor and circumferential resection margin and whether R0 resection was performed were recorded. Pathologists classified a tumor as histologically mucinous if more than 50% of the volume of the tumor was extracellular mucinous pools [19].
Statistical Methods
Histologic tumor stage and distance to mesorectal fascia were taken as the
standards with which posttreatment MRI findings were compared. The linear
weighted kappa statistic was used to assess agreement between MRI findings and
histologic tumor stage for all tumors and for nonmucinous tumors. Agreement of
tumor length and distance between tumor and circumferential resection margin
in measurements on MRI and histologic sections were assessed with the method
comparison analysis described by Bland and Altman
[20,
21] and with the interclass
correlation coefficient.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
Decreases in tumors size of more than 30% (disease response according to the Response Evaluation Criteria in Solid Tumors group [22]) occurred in 19 (63%; 95% CI, 46-81%) of 30 tumors (Figs. 5A, 5B and 6A, 6B). Eight of 30 tumors were downstaged on the basis of imaging findings (Table 2); no patient had an increase in T stage. Overall, posttreatment histologic staging was as follows: four (13.3%) of the patients had no viable tumor; four (13.3%) of the patients had T1 or T2 N0 disease; two (6.7%) of the patients had T1 or T2 N1 disease; 11 (36.7%) of the patients had T3 N0 disease; and nine (30%) of the patients had T3 N1+ or T4 disease. The concordance between postchemoradiation therapy MRI and histologic findings is given in Table 1.
|
|
|
|
|
Mucinous tumors were a difficult tumor subtype to interpret on MRI. On
T2-weighted imaging, mucinous tumors had mixed intermediate and high signal
intensity in the tumor stroma, which often achieved closer to uniform
hyperintensity after chemoradiation therapy (Fig.
7A,
7B). Consequently, for
statistical analyses we assessed all cases and nonmucinous tumors only (Tables
1 and
3). For the alltumors group, 18
(60%) of 30 tumors were correctly T staged after treatment. Concordance
between postchemoradiation therapy MRI and histologic stage was fair to
moderate (
= 0.40; 95% CI, 0.10-0.66). A similar proportion (61%
[14/23]) of nonmucinous tumors were correctly staged after treatment (
= 0.44; 95% CI, 0.14-0.75). However, nodal staging improved from 70% to 87%
when nonmucinous tumors alone were considered. A subset of nonmucinous tumors
separately assessed were tumors that had only mucinous differentiation after
chemoradiation therapy. All four of these tumors were correctly staged for
both tumors and nodes. Three of these tumors shrank, and two were downstaged,
one showing nodal downstaging (the other three were node free throughout),
after chemoradiation therapy.
|
|
|
Staging errors were mainly in the small T3 tumor group (T3a), in which less than 5 mm of extramural tumor spread was present (Table 3). Seven of the 12 incorrectly staged tumors were either T3a tumors staged as being confined to the bowel wall or vice versa. Three of the 12 incorrectly staged tumors were staged as larger T3 tumors (> 5 mm of extramural spread) or T4 tumors that had mucinous histologic features. Histologic examination showed that the latter had no viable tumor cells present, only inactive mucin lakes. The other two errors were in tumors staged T4 because of circumferential resection margin involvement, whereas histologic examination revealed a clear circumferential resection margin but by only 2 mm, making the tumors large T3 with more than 10 mm of extramural spread.
More than 2 mm of circumferential resection margin regression was observed on MRI in 10 (45%; 95% CI, 25-66%) of 22 patients with radiologic T3 or T4 tumors. In the nonmucinous tumor group, mean regression was 2 mm (95% CI, -12 to 17 mm). In the tumor group in which mucinous differentiation developed after chemoradiation therapy, mean regression was 5 mm (95% CI, -10 to 19 mm). In the mucinous tumor group, mean regression was -1 mm (95% CI, -4 to 3 mm). Tumors therefore tended to progress slightly rather than regress in response to chemoradiation therapy.
In all 13 patients with tumors radiologically staged T3 and radiologically predicted to have a clear circumferential resection margin, these findings were confirmed at histologic examination. No tumors staged T3 were found at histologic examination to have circumferential resection margin involvement. Of the eight tumors staged radiologically as having circumferential resection margin involvement (T4 tumors), only four had the histologic finding of circumferential resection margin involvement. Two of the incorrectly staged tumors were mucinous in type and had mucinous lakes at the circumferential resection margin but no active tumor. One tumor was radiologically considered to involve the peritoneum, but this finding was not histologically proven. In the other case, posttreatment fibrosis was present at the circumferential resection margin, but no active tumor was found.
For tumors that were stage T3 or greater radiologically (n = 22), we assessed the utility of MRI in prediction of the distance between the tumor and the circumferential resection margin using histologic findings as the reference standard (Fig. 8A, 8B, 8C, 8D). For all tumors, the 95% limits of agreement between histologic and MRI findings were -8 to 5 mm (mean difference, 0.2 mm) (Fig. 8A, 8B, 8C, 8D) with an interclass correlation coefficient of 0.74. Lesion length was less well predicted; for all tumors, 95% limits of agreement were 35 to 46 mm (mean difference, 5.4 mm).
|
|
|
|
Of the eight tumors with anal sphincter involvement found at histologic analysis, six were correctly identified on MRI. Overall, the prediction based on MRI findings was that eight patients would have anal sphincter involvement, and in six of the patients the involvement was found at histologic examination. Overall, achievement of R0 resection was correctly predicted on the basis of MRI findings for 18 (82%; 95% CI, 66-98%) of 22 tumors staged radiologically as T3 or T4. Excluding mucinous tumors, a similar proportion (17/18; 94%; 95% CI, 84-100%) was correctly predicted (Table 1).
Other MRI Changes After Chemoradiation Therapy
The most common postchemoradiation therapy MRI change other than staging
characteristics was bone marrow fatty atrophy in 93% (n = 28) of the
patients. Other changes included rectal muscular wall fibrosis in 63%
(n = 19) of the patients, new presacral edema either behind or in
front of the presacral fascia in 60% (n = 18), rectal mucosal edema
in 27% (n =8), peritoneal fibrosis in 10% (n = 3), and
presacral fibrotic thickening in 7% (n = 2) of the patients.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods of selection of patients for longcourse chemoradiation therapy vary widely throughout the world. At our institution, we do not manage all bulky T3 tumors with neoadjuvant chemoradiation therapy but instead treat patients who have extramural tumor spread and a circumferential resection margin at risk of involvement at surgery. This criterion includes patients with bulky T3 disease at the circumferential resection margin, patients with T3 disease of small volume with extramural tumor close to the levator muscles, patients with tumors that lie too close to the external anal sphincter, and patients with tumors potentially infiltrating the intersphincteric space. The decision to treat with preoperative chemoradiation therapy is made by multidisciplinary teams on the basis of staging MRI findings and on clinical grounds (tumor fixity at rectal examination or evaluation under anesthesia).
Most of our patients had intermediate- to high-grade tumors that were low rectal in position and annular in structure. A considerable proportion were mucinous according to signal intensity criteria. After chemoradiation therapy, 63% of tumors shrank sufficiently (> 30% decrease in longest dimension) to constitute a partial response to treatment. Tumor downstaging occurred in only 17% of cases.
Most tumors were stage T3, and although the tumors often became smaller, rectal wall penetration usually persisted to some extent. Our accuracy in prediction of T3 disease was considerably poorer when extramural tumor extension of < 5 mm was present. This finding clearly was not always differentiated from posttreatment tumor and muscle wall fibrosis. In the case of larger T3 tumors, our prediction was highly accurate. Except for a solitary mucinous tumor, T3 disease was predicted exactly. Because of this degree of accuracy, a future study should assess how MRI prediction of tumor stage and circumferential resection margin involvement influences treatment of patients. These findings may be used for accurate definition of optimal timing of surgery. The current recommendation of surgery 6-8 weeks after chemoradiation therapy is derived from results of old studies of radiation therapy to the bladder followed by cystectomy. MRI findings may also be used to provide a road map for showing surgeons areas where surgical dissection is likely to be difficult. Finally, if these MRI findings are confirmed, it is possible to give further therapy if the imaging findings suggest that R0 resection is not going be achieved.
Nodal downstaging was more prevalent than tumor downstaging, occurring in 68% of the patients, and it occurred regardless of decreases in tumor size or stage. MRI findings also concurred closely with histologic findings on prediction of exact nodal stage. In this study, nodal involvement was assessed with size criteria [16] and morphologic features such as border, shape, and internal structure, as described by Brown et al. [23].
Imaging identification of nodal disease in patients with rectal cancer is diagnostically difficult. Reliable differentiation between reactive and metastatic nodes is not always possible because there is a high incidence of microscopic metastasis in normal-appearing nodes [24], and enlarged but reactive mesorectal nodes can be present in patients with rectal cancer. In our study, fewer nodes were found after chemoradiation therapy, and according to our criteria for nodal involvement, nodal downstaging occurred in 68% of cases. This finding concurs with the finding of DeVries et al. [25], who also found fewer nodes after chemoradiation therapy. MR lymphography with ultrasmall paramagnetic iron oxide must be assessed in this setting. We believe this technique is unlikely to be helpful because of the exquisite sensitivity of macrophages to radiation, as has been well documented in studies of the liver in which depletion of Kupffer cells occurred after radiation therapy [18, 26-28].
Mucinous tumors tended to retain high signal intensity after chemoradiation therapy both in primary tumors and in nodes and did not seem to shrink or change much morphologically. These tumors did not seem to respond well to chemoradiation therapy, with persistent areas of high signal intensity, and apparent circumferential resection margin involvement was found more frequently. This finding proved to be a source of error; in many of these cases, histologic evaluation showed no active tumor but only inactive mucin lakes. To our knowledge, this finding has not been specifically reported, although this tumor subtype is well recognized as a radiologic entity [15, 29] and is accepted as having an overall poorer outcome than other histologic subtypes [30, 31].
We found four additional mucinous tumors after chemoradiation therapy (compared with pretreatment MRI findings). This morphologic change was found in a similar proportion in a previous pathologic study [32]. Furthermore, demographic studies have shown mucinous tumors in patients receiving both radiation therapy and chemoradiation therapy [33], suggesting that chemoradiation therapy may increase the degree of mucinous differentiation of rectal carcinoma. Mucinous differentiation due to chemoradiation therapy does not appear to be a poor prognostic factor in itself [33]. This finding also occurred in our study. Tumors showing mucinous differentiation only after chemoradiation therapy were responsive to treatment, showing a larger amount of regression from the circumferential resection margin. Imaging findings suggested that tumors that were mucinous throughout did not appear to respond to chemoradiation therapy, but this finding did not have much influence on the overall accuracy of radiologic prediction of tumor stage. However, the rate of R0 resection was slightly more difficult to predict for mucinous tumors (82%), although with the exclusion of mucinous tumors the rate remained high (94%).
Another pitfall included differentiating active tumor from posttreatment fibrosis. This problem was most relevant in differentiating T2 and T3 disease, in which most of the errors occurred. This phenomenon occurred even though the observers based the MRI diagnosis of T3 lesions on the method described by Brown and colleagues [8]. Diagnosis was based on the presence of tumor extending into the perirectal fat with a broad-based bulging configuration and in continuity with the intramural portion of the tumor. This method of classification and other criteria are based on observations of patients receiving short-course chemoradiation therapy [14].
In our study, however, a full 6 weeks (long course) of chemoradiation therapy was used, and the observed 63% rate of tumor or muscle wall fibrosis on MRI was replicated by similar histologic changes. In our study, most of the subjects had advanced local disease (clinically and on MRI), so we used long-course chemoradiation therapy before surgery. We were unable to reliably differentiate active tumor from posttreatment fibrosis, because the latter was more florid and prevalent in our study than in others. Brown and colleagues [8] stated that their short-course (1 week) treatment resulted in no discernible histopathologic evidence of a radiation therapy effect on the tumors. We therefore do not believe that the appearance of fibrosis in studies in which short-course chemoradiation therapy is used can be extrapolated to studies such as ours in which long-course chemoradiation therapy was used because differentiation from active tumor often is not possible.
This assertion is supported by Shia et al. [34], who in a pathologic study showed that the morphologic patterns of the rectum after long-course chemoradiation therapy are quite distinct from the patterns after short-course chemoradiation therapy. Chen et al. [35] also used long-course chemoradiation therapy and had considerable difficulty reliably differentiating active tumor from posttreatment fibrosis; tumor staging accuracy was only 52%. Kuo et al. [36] also used long-course chemoradiation therapy and had similar difficulties; overall tumor staging accuracy was 47%. In none of those studies did the investigators assess the clinically important factor of distance between tumor and circumferential resection margin, which is what we investigated. Differentiating fibrosis and mucinous degeneration from active disease is difficult with morphologic imaging. It one day may be possible to use functional techniques such as dynamic MRI and diffusion-weighted imaging to differentiate fibrosis from active tumor during imaging at high spatial resolution.
Circumferential resection margin tumor regression was frequently found, and the distance between tumor and circumferential resection margin was accurately predicted to within a few millimeters. However, the 95% limits of -8 to 5 mm (mean difference, 0.2 mm; interclass correlation coefficient, 0.74) for tumors managed with total mesorectal excision was greater than that found in other studies [8, 9]. Those studies were conducted with patients receiving no preoperative therapy or only short-course radiation therapy and showed that circumferential resection margins could be predicted to ± 5 mm [9]. The explanation for the difference is the amount of posttreatment fibrosis, which is greater when patients receive long-course chemoradiation therapy. The difference caused significant errors in interpretation of the amount of residual active tumor present and hence reduced the accuracy of prediction of measurements of circumferential resection margin on MRI. We believe that our accuracy of prediction of circumferential resection margin is a more realistic estimate of what would be practically achievable in daily practice after chemoradiation therapy.
Because of this degree of accuracy, a future study should assess how MRI prediction of distance between tumor and circumferential resection margin influences surgical management. Peschaud et al. [37] examined the accuracy of MRI in prediction of involvement of the circumferential resection margin in patients receiving long- or short-course chemoradiation therapy. Those investigators considered the circumferential resection margin as involved or uninvolved by tumor rather than as a distance. Involvement was scored if the measurable distance on MRI was less than 2 mm from the circumferential resection margin. Circumferential resection margin involvement in the patients who had received long-course chemoradiation therapy was predicted in 70% of cases, but most of the tumors were midrectal. The rate of prediction of the circumferential resection margin of low rectal tumors was a poor 22%.
In conclusion, high-resolution MRI is only moderately accurate in prediction of tumor stage in patients with rectal cancer who have undergone long-course preoperative chemoradiation therapy. The surgically more relevant parameter, distance between tumor and circumferential resection margin, can be predicted to a few millimeters. Errors in predicting stage and circumferential resection margin arise in the case of mucinous tumors. The errors are related to marked tumor and rectal wall fibrosis, which probably is caused by effective long-course preoperative chemoradiation therapy and which causes difficulty in differentiating fibrosis from active tumor. These effects must be considered when patients are being assessed for preoperative identification of high risk of the presence of an involved surgical margin. With this knowledge of the accuracy of MRI in prediction of the histologic findings after long-course chemoradiation therapy, investigators conducting future studies can assess how MRI findings influence surgical management.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||
