| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Edward Mallinckrodt Institute of Radiology, Washington University School of
Medicine, 510 S. Kingshighway Blvd., St. Louis, MO 63110.
2
Siteman Cancer Center, Washington University School of Medicine, St. Louis, MO
63110.
3
Department of Surgery, Washington University School of Medicine, St. Louis, MO
63110.
Received April 13, 2001;
accepted after revision August 28, 2001.
Address correspondence to F. Dehdashti.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. We retrospectively identified 47 patients with recurrent colorectal cancer who underwent surgical exploration for possible curative resection of hepatic metastases. All patients underwent CT or MR imaging and FDG PET preoperatively and intraoperative sonography. The performance of the imaging techniques was evaluated through review of the radiologic reports and correlation with surgical and histopathologic findings.
RESULTS. Eighty-seven malignant hepatic lesions were identified by histopathologic analysis of liver specimens, and 23 benign hepatic abnormalities were documented histopathologically or by uroradiologic imaging. For hepatic sections characterized as containing metastases by radiologic imaging, the positive predictive value for FDG PET was 93% (54/58); for intraoperative sonography, 87% (52/60); and for conventional imaging, 83% (43/52). For individual lesions characterized as probably malignant, the positive predictive value for FDG PET was 93% (62/68); for intraoperative sonography, 89% (63/71); and for conventional imaging, 78% (46/59). The findings at intraoperative sonography led to a change in the clinical treatment of only one patient (2%).
CONCLUSION. The results indicate that FDG PET effectively screens potential candidates for curative liver resection. Although intraoperative sonography helps to determine the anatomic location of metastases thus facilitating surgical resection, its adjunctive use in patients screened preoperatively by FDG PET has limited impact on treatment selection.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Unfortunately, only 10-20% of patients with hepatic metastases are considered surgical candidates [6]. Accurate preoperative identification of the number and location of hepatic lesions is essential for planning the surgical procedure [7]. Thus, considerable attention has been directed toward defining the most efficient preoperative method for selecting patients who may benefit from surgical resection of their metastases. Such a method would decrease the morbidity and mortality rates associated with unnecessary surgery in patients with advanced disease.
Both CT during arterial portography (CT portography) and intraoperative sonography have been studied as potential methods for accurate assessment of hepatic metastatic disease. CT portography has been shown to be more sensitive (85%) than CT and MR imaging (68%) in the detection of hepatic metastatic lesions [8]. Other studies have shown that intraoperative sonography also is sensitive in identifying hepatic metastases [9,10,11]. However, an inherent disadvantage of intraoperative sonography is that it is a part of the surgical procedure, and its results can only change the surgical treatment after the patient has already undergone a laparotomy. Although the sensitivity of intraoperative sonography for the detection of hepatic lesions is high, the literature addressing the specificity of this technique is limited.
Over the last decade, PET with 2-[18F]-fluoro-2-deoxy-D-glucose FDG has been increasingly used for the evaluation of patients with cancer. FDG PET has been repeatedly shown to have high sensitivity (93-100%) and specificity (78-100%) for evaluating patients with colorectal cancer [12]. The literature has focused mainly on the comparison of FDG PET with conventional imaging techniques, such as CT, in identifying the primary colorectal cancer, evaluating metastatic or recurrent disease [12], and monitoring treatment. The aim of this study was to compare the diagnostic performance of preoperative FDG PET and intraoperative sonography with the standard of histologic examination of resected liver specimens to determine their relative merits for identifying suitable candidates for curative resection of liver metastases from colorectal cancer.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
PET
PET imaging was performed with an ECAT EXACT scanner (Siemens-CTI,
Knoxville, TN) beginning approximately 40 min after IV administration of 10-15
mCi (370-555 MBq) of FDG. A series of four to six overlapping 47-slice
emission and transmission PET images were obtained to include the region from
the middle of the neck to the upper thighs. Renal, ureteral, and bladder
activity were minimized by IV hydration, diuretic administration, and bladder
catheterization during the study, as previously described
[14]. The PET images in 14
patients were reconstructed by filtered back projection with the use of a
Hanning filter (frequency cutoff, 0.6 x Nyquist value). In the remaining
32 patients, the PET images were reconstructed by use of an ordered-subset
estimation-maximization iterative algorithm and Butterworth filter (frequency
cutoff, in the range of 0.2-0.4 cycles per pixel). The emission images were
corrected for attenuation with the segmentation method
[15]. Images were viewed on
the computer display monitor in three orthogonal projections and as whole-body
maximum-pixel-intensity reprojection images for visual interpretation.
All PET images were evaluated by an experienced nuclear radiologist in routine clinical fashion, including correlation with other imaging studies and with pathologic results. Foci of FDG uptake greater than that of normal liver were identified as tumor. Areas with FDG accumulation similar to that of normal liver (in the case of an abnormal finding identified on CT or MR imaging with no corresponding abnormal finding on PET) were considered to represent normal tissue or benign disease. All patients had only one FDG PET imaging study.
Cross-Sectional Imaging
Cross-sectional imaging of the abdomen was performed in all patients. The
cross-sectional studies available at the time of FDG PET interpretation (a
total of 39 CT and seven MR imaging examinations for 45 patients) were
included in the data analysis. The cross-sectional studies were performed an
average of 37 days before FDG PET and 51 days before surgery. Hardcopy images
were available at the time of interpretation of the PET studies. Because
studies in 24 patients were obtained from outside institutions, various
imaging protocols were used. However, all cross-sectional imaging studies were
judged to be of quality comparable with the images obtained at our
institution. In general, the CT examinations were all performed with use of
oral and IV contrast agents with image acquisition during the portal venous
phase of contrast enhancement. The CT examinations were interpreted in a
routine clinical fashion by the attending radiologist.
Intraoperative Sonography
Intraoperative sonography was performed an average of 23 days after FDG PET
and 51 days after cross-sectional imaging studies. All sonographic
examinations were obtained using 2200 (Tetrad, Englewood, CO), SSD-2000 (Aloka
Ultrasound, Wallingford, CT), HDI 3000 (ATL Ultrasound, Bothell, WA), or 128XP
(Acuson, Mountain View, CA) scanners. All studies were performed by a
radiologist experienced in intraoperative sonography with a high-resolution
linear array transducer ranging in frequency from 5 to 10 MHz. In five of the
47 patients, the intraoperative sonography was targeted to the hemiliver that
was to be preserved (and was presumed to be free of metastases on the basis of
the results of prior imaging studies). In one patient, the examination was
primarily targeted to the hemiliver with known metastatic involvement to
determine the extent of disease. In the other 41 patients, the reports
indicated that the entire liver was examined.
Hepatic Resection
All surgeries were performed by experienced hepatic surgeons. In each
patient, the liver was fully inspected and palpated by the surgeon, and
suspicious areas detected by direct examination or intraoperative sonography
were biopsied or resected. Occasionally, random biopsies were obtained. The
remainder of the abdomen was inspected for extrahepatic metastasis. The
surgical procedure in all patients was resection of a single or multiple
hepatic sections, with occasional wedge resections of additional lesions.
Data Analysis
For a lesion to be included in the analysis, it had to be detected by
cross-sectional imaging, FDG PET, intraoperative sonography, or intraoperative
liver inspection and then analyzed histologically.
The effectiveness of PET, intraoperative sonography, and cross-sectional imaging was determined on the basis of the correct identification of metastatic involvement of the hepatic sections by a given technique. In the analysis, we used a widely accepted anatomic division of the liver into the right posterior, right anterior, left medial, and left lateral sections. We considered an imaging study or histopathologic analysis to be positive for a given hepatic section if any lesion was detected within it.
Because a large number of metastatic foci, even when limited to one or two hepatic sections, have been shown to predict a poor prognosis [3, 16], the data also were analyzed on the basis of the number of correctly identified malignant lesions, regardless of the sectional involvement. An imaging study was considered to have correctly identified a malignant lesion only if the location of the lesion correlated precisely with its anatomic location described in the surgical pathology and operative reports. In five patients (11%), six or more hepatic lesions were seen on imaging studies. Radiologic reports stated that "multiple" or "numerous" hepatic metastatic foci were seen. In these cases, an arbitrary number of six foci were used in the calculation of sensitivity and positive predictive value of the imaging test.
On the basis of the available data, we could calculate the sensitivity and positive predictive value for FDG PET and cross-sectional imaging. Because intraoperative sonography in some patients was targeted to only a portion of the liver parenchyma, as noted previously, the calculation of true sensitivity was not possible. On the basis of the available data, a positive predictive value was thus calculated for intraoperative sonography and compared with those for FDG PET and cross-sectional imaging studies.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The number of metastatic foci also is important in determining resectability of hepatic lesions. The presence of a large number of metastatic foci, even when limited to one or two hepatic sections, has been shown to predict a poor prognosis [3, 16]. Patients with four or more hepatic lesions are considered poor surgical candidates even if the lesions are confined to one hepatic section. Therefore, the data also were analyzed on the basis of the number of correctly identified malignant lesions, regardless of the sectional involvement. A total of 87 metastatic lesions was identified by histopathologic analysis of biopsied or resected lesions. FDG PET correctly revealed 62 of 87 malignant lesions (71% sensitivity). Although, the true sensitivity of intraoperative sonography could not be calculated because of the targeted nature of this examination in six patients, it detected 63 of 87 lesions, a result comparable to that for FDG PET.
An additional 23 histopathologically benign focal hepatic abnormalities were detected by imaging studies, by liver palpation at the time of surgery, or by liver sectioning after surgical resection. The benign "lesions" were granulomas, benign mesenchymal tumors, bile duct hamartomas, cavernous hemangiomas, and normal hepatic parenchyma. In five instances, FDG PET was falsely positive. These lesions were found to represent granuloma, hamartoma, or normal hepatic parenchyma. Detection of histopathologically confirmed normal hepatic parenchyma as an abnormality by any of the imaging techniques may be due to multiple factors, such as focal fat infiltration, regionally altered blood flow resulting in heterogeneous contrast enhancement, intrinsic heterogeneity of hepatic parenchyma, or image artifact. On a lesion basis, the calculated positive predictive value of FDG PET is 93% (62/68). This compares with a positive predictive value of 89% (63/71) for intraoperative sonography, which incorrectly characterized eight of 23 benign lesions (Fig. 2A,2B,2C). By comparison, conventional imaging (CT and MR imaging) showed a sensitivity of 55% (46/83) and a positive predictive value of 78% (46/59).
|
|
|
Finally, we determined the number of cases in which intraoperative sonography changed the surgical approach either because of the detection of hepatic lesions that were unresectable or because of the detection of additional lesions. In our series of 47 patients, surgical resection was not performed in two patients (Fig. 1). In one of these patients, the change in treatment was not dependent on intraoperative sonographic results but rather was based on the finding of extrahepatic disease by inspection at the time of surgery. In the other patient, intraoperative sonography revealed extensive hepatic metastatic disease that was not detected on preoperative imaging studies including FDG PET and that was not apparent by direct inspection and palpation of the liver.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In patients with colorectal cancer, radiologic imaging is used to select patients suitable for curative resection of hepatic metastases and in the planning of curative surgery. Intraoperative sonography provides a valuable adjunct to surgical palpation and inspection of the liver and is used by many surgeons for lesion detection and assessment of vascular anatomy, exclusion of vascular invasion, and the mapping of sectional anatomy. The sensitivity of intraoperative sonography for the detection of focal metastatic disease of the liver has been reported to be as high as 96%, even with detection of lesions smaller than 10 mm in diameter, thus exceeding the sensitivity of other imaging methods [18]. Kruskal and Kane [19] have reported that intraoperative sonography consistently reveals 25-35% more lesions than noninvasive preoperative routine cross-sectional imaging. Although intraoperative sonography has a high sensitivity, its relatively low specificity often results in the need for biopsy to characterize some detected hepatic abnormalities, as we found in this study and as reported by Stone et al. [20]. CT portography is also considered a highly sensitive preoperative imaging technique for the detection of hepatic lesions; however, it also has a considerable false-positive rate [8]. It has been reported that CT portography detects 76-86% of lesions overall [8, 21] and 67% of lesions that are smaller than 10 mm in diameter [8].
FDG PET is an established imaging technique for in vivo measurement of tissue glucose metabolism. The increased use of glucose by tumor cells has been shown in many tumor types, including colorectal cancer [12]. It has been shown that FDG PET has greater sensitivity than CT in staging of certain tumors, including esophageal cancer and lung cancer [14, 22], and in detecting recurrent or residual tumor, including brain tumors and colorectal cancer [12, 23]. We have recently found that FDG PET is more sensitive for the detection of locoregional recurrence of colorectal cancer than CT and colonoscopy combined [24]. We have also found that FDG PET has a sensitivity of 95% and specificity of 100% for detection of hepatic metastases in patients treated for colorectal cancer, whereas CT has a sensitivity of 74% and specificity of 85% in the same group of patients [25]. Schiepers et al. [26] have reported a sensitivity of 94% and accuracy of 94% for FDG PET and 85% and 93%, respectively, for conventional imaging (CT and sonography) for detection of hepatic metastases. In another study, FDG PET was found to be more accurate than CT portography (93% vs 76%) mainly because of the much higher specificity (100% vs 9%) of PET [27].
In patients with hepatic metastases from colorectal cancer, FDG PET is being used increasingly to distinguish patients eligible for curative resection from those with occult advanced disease who are not candidates for surgery. Preoperative screening by FDG PET decreases the failure rate after presumed curative resection. Once a patient is deemed a suitable candidate for curative resection of hepatic metastases, resection will be performed unless intraoperative sonography or the surgeon's direct examination discloses extrahepatic or extensive hepatic disease at surgery. In our study, we compared the performance of FDG PET and intraoperative sonography for the detection of hepatic metastatic disease in patients with colorectal cancer. For identification of hepatic sections involved with metastatic disease and detection of individual hepatic metastases, our results show similar positive predictive values for FDG PET and intraoperative sonography. Although not a primary focus of the current study, both the sensitivity and positive predictive value of cross-sectional imaging were inferior to those of FDG PET and intraoperative sonography.
In our overall study population, 70 of 120 patients were considered unsuitable candidates for surgery on the basis of clinical and imaging data that included the results of FDG PET. As is true of all imaging methods, the specificity of FDG PET in the detection of liver metastases from colorectal cancer is not perfect. Therefore, in those instances in which the findings of FDG PET would result in the patient being denied potentially curative surgery, we advocate biopsy confirmation of lesions (hepatic or extrahepatic) detected by FDG PET, unless they can be clearly shown to be benign by other imaging methods. If the lesion cannot be localized (by imaging or by inspection at surgery) and the patient is otherwise considered a surgical candidate, the surgeon will proceed with hepatic resection. However, when PET reveals widespread metastatic disease, hepatic surgery will not be performed, and the patient will be followed closely. Typically within a few months, the disease will become apparent on cross-sectional imaging studies.
The redundancy of the current preoperative staging algorithm, which includes multiple imaging techniques, is evident [28]. Unlike intraoperative sonography, FDG PET has an inherent advantage of preventing unnecessary laparotomies by preoperatively depicting extra-hepatic metastatic spread of colorectal cancer [27]. FDG PET also has the advantage of greater accuracy compared with conventional cross-sectional imaging techniques. In our study, the sensitivity and positive predictive value of CT and MR imaging were lower than those of FDG PET. This finding is in agreement with other reports in the literature indicating that FDG PET is more effective than conventional imaging for screening of potential candidates for liver resection [26, 27]. Cross-sectional imaging, nonetheless, provides anatomic details that are often necessary for correct interpretation of FDG PET studies.
One of the limitations of this study was that cross-sectional imaging was performed an average of 37 days before FDG PET; thus, some hepatic lesions might have become more apparent during this time. This fact alone, however, cannot fully account for the lower sensitivity of cross-sectional imaging compared with FDG PET.
Another limitation of our study is that intraoperative sonography was performed with prior knowledge of cross-sectional imaging results, thus biasing the radiologist's interpretation of some benign lesions, such as cavernous hemangiomas or granulomas. Thus, the true positive predictive value of intraoperative sonography for the detection of malignant hepatic foci may be somewhat lower than that reported in our study. On the other hand, the intraoperative sonography of five of our patients was targeted toward the portion (or portions) of the liver parenchyma thought to be free of documented metastatic disease. In these patients, the portions of the liver that were not evaluated by intraoperative sonography were known to contain metastatic disease on the basis of the results of other imaging studies (FDG PET and CT or MR imaging). Although it is possible that intraoperative sonography of the involved hepatic segments in these patients would have revealed additional lesions, the surgeons at our institution believed that this would not impact the surgical treatment of these patients and thus did not warrant the additional scanning time that would be required. Even though extending the time and scope of the intraoperative sonography would probably have improved its sensitivity, such changes would also likely decease the positive predictive value of the technique.
Our analysis is based on a retrospective review of radiologic and pathologic reports. Thus, correlation between the hepatic lesions revealed by imaging and those seen by gross examination of pathology specimens relied on and was limited by the description of their location found in the reports. In fact, our analysis excluded three patients with ambiguous pathology reports.
Another potential limitation of our study is directly related to its retrospective nature. There is the potential for bias in the patient population because only patients who had preoperative cross-sectional imaging, FDG PET, and intraoperative sonography were included. However, this population reflects the patients with suspected hepatic metastatic disease currently being evaluated at our institution and at most institutions with the capability for PET imaging.
One of the strengths of intraoperative sonography is its high resolution. In one patient in our series, FDG PET revealed a single hepatic lesion that was later shown by intraoperative sonography and histopathologic analysis to represent three adjacent, but separate, metastatic foci. This apparent error, which lowered the calculated sensitivity of FDG PET, was likely related to the relatively low resolution of the technique but had no impact on the patient's treatment. The relatively low resolution of current PET scanners could potentially result in the erroneous localization of a metastatic focus to an incorrect hepatic section, thus misdirecting the surgical approach. This potential limitation of FDG PET provides justification for its use in conjunction with intraoperative sonography.
The ultimate test of the usefulness of intraoperative sonography in surgical planning is its impact on clinical treatment. In this series of 47 patients screened preoperatively with FDG PET, the findings of intraoperative sonography altered the surgical approach in only one patient (2%). In one additional patient, hepatic resection was not performed, but the decision to change the treatment plan was based on the detection of metastatic disease by direct inspection at laparotomy rather than by the intraoperative sonography findings. It could, therefore, be argued that the role of intraoperative sonography during surgery should be limited, and possibly its use restricted to patients with questionable FDG PET findings or when further anatomic delineation of the hepatic lesions is necessary.
Despite the limitations of this study, the results strongly suggest that FDG PET can effectively identify patients suitable for hepatic resection with the best chance for cure. Our recent finding of a 77% overall survival and 40% disease-free survival at 3 years in patients evaluated preoperatively by FDG PET supports this premise [13]. These survival rates are substantially better than those reported for patients evaluated preoperatively with conventional imaging only, such as CT and sonography [29, 30]. Additional studies to assess clinical outcomes in patients undergoing different presurgical screening algorithms are needed.
Acknowledgments
We thank Andrea Sykes for her assistance in preparation of this
manuscript
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  O J Garden, M Rees, G J Poston, D Mirza, M Saunders, J Ledermann, J N Primrose, and R W Parks Guidelines for resection of colorectal cancer liver metastases Gut, August 1, 2006; 55(suppl_3): iii1 - iii8. [Full Text] [PDF]
|
|
|
|
 |
|
 J. B. Kruskal and R. A. Kane Intraoperative US of the Liver: Techniques and Clinical Applications. RadioGraphics, July 1, 2006; 26(4): 1067 - 1084. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. P. Khatri, N. J. Petrelli, and J. Belghiti Extending the Frontiers of Surgical Therapy for Hepatic Colorectal Metastases: Is There a Limit? J. Clin. Oncol., November 20, 2005; 23(33): 8490 - 8499. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Bipat, M. S. van Leeuwen, E. F. I. Comans, M. E. J. Pijl, P. M. M. Bossuyt, A. H. Zwinderman, and J. Stoker Colorectal Liver Metastases: CT, MR Imaging, and PET for Diagnosis--Meta-analysis Radiology, October 1, 2005; 237(1): 123 - 131. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. J. Kelloff, J. M. Hoffman, B. Johnson, H. I. Scher, B. A. Siegel, E. Y. Cheng, B. D. Cheson, J. O'Shaughnessy, K. Z. Guyton, D. A. Mankoff, et al. Progress and Promise of FDG-PET Imaging for Cancer Patient Management and Oncologic Drug Development Clin. Cancer Res., April 15, 2005; 11(8): 2785 - 2808. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. Prati, F. Trotta, L. Roveda, C. Filice, F. Dehdashti, S. M. Strasberg, and B. A. Siegel Intraoperative Sonography Still the Standard Technique in Hepatic Surgery Am. J. Roentgenol., November 1, 2002; 179 (5): 1347 - 1348. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
