|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 1999 American Society for Clinical Oncology Additional Value of Whole-Body Positron Emission Tomography With Fluorine-18-2-Fluoro-2-deoxy-D-glucose in Recurrent Colorectal CancerFrom the Departments of Nuclear Medicine, Internal Medicine, Abdominal Surgery, and Radiology, University Hospital Gasthuisberg and Catholic University of Leuven, Leuven, Belgium. Address reprint requests to Patrick Flamen, MD, Department of Nuclear Medicine, University Hospital Gasthuisberg, Catholic University of Leuven, Herestraat 49, B-3000 Leuven, Belgium; email patrick.flamen_uz.kuleuven.ac.be
PURPOSE: To assess the additional value of the whole-body [18F]-2-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) scan as a staging modality complementing conventional diagnostic methods (CDM) in patients suspected of having recurrent colorectal adenocarcinoma. PATIENTS AND METHODS: In 103 patients, the discordances between FDG-PET and CDM results were identified and related to the final diagnosis obtained by histopathology or clinical follow-up (> 1 year). All FDG-PET studies were reviewed with full knowledge of the CDM findings. RESULTS: In a region-based analysis, discordances between CDM and FDG-PET findings were found in 40 of 412 regions (10%). In these, FDG-PET had additional diagnostic value in 14 of 16 locoregional, six of seven hepatic, seven of eight abdominal, and eight of nine extra-abdominal regions. In a patient-based analysis, CDM categorized a subgroup of 60 patients as having resectable recurrent disease limited to the liver (n = 37) or locoregional region (n = 23). In 13 of these patients, there were discordant FDG-PET findings, detecting additional tumor sites in nine patients and excluding disease in three patients and yielding an additional diagnostic value in 20% of the patients. A second subgroup consisted of 13 patients with inconclusive CDM findings (n = 5) or with elevated plasma carcinoembryonic antigen levels and an otherwise negative conventional work-up (n = 8). In these patients, FDG-PET results were correct in eight of nine discordances, yielding a positive additional diagnostic value in 62% of the patients. CONCLUSION: Whole-body FDG-PET can have a clear impact on the therapeutic management in the follow-up of patients with colorectal cancer.
IN PATIENTS WITH colorectal cancer, the recurrence rate after apparently curative resection of the primary tumor is 30% to 40%, and recurrent disease develops predominantly within 3 years after surgery.1 Approximately 25% of first colorectal cancer recurrences are isolated locoregional failures, and an additional 15% to 20% are metastatic deposits that are potentially resectable for cure.2 In those patients, for resection of liver metastases or pelvic recurrence to be curative, it is imperative that there be no unrecognized foci of tumoral disease outside the operation field. Presently, using standard preoperative diagnostic procedures, surgical treatment of recurrences leads to cure in 25% to 40% of the patients.3,4 Although the value of the conventional diagnostic methods (CDM) has improved in this setting, there still is a need for a highly accurate and noninvasive imaging modality to detect inoperable disease and select patients for curative surgery. Recently, whole-body positron emission tomography (PET) has been advocated as a promising tool for staging recurrent colorectal adenocarcinoma.5-8 With the use of a labeled glucose analog, [18F]2-fluoro-2-deoxy-D-glucose (FDG), this technique detects malignancy by depicting increased glucose uptake and metabolism in cancer cells. The agent is transported into cells by means of epithelial glucose transporter proteins. Because FDG lacks a hydroxyl group in the two-position, its first metabolite, FDG-6-phosphate, is not a substrate for glucosephosphate isomerase and cannot be converted to the fructose analog. Owing to the negative charge on FDG-6-phosphate, it has a low membrane permeability and will accumulate intracellularly.9 Recent advances in technology allow the screening of the whole body in one examination session.10 The major impediment to the widespread application of this unique metabolic imaging tool is its limited availability owing to high cost and complexity. Therefore, definition of patient subsets most likely to benefit from this technique in a realistic environment of competing diagnostic modalities is mandatory. The purpose of this retrospective study was to evaluate the additional value of the whole-body FDG-PET scan as a staging modality complementary to CDM in patients with known or suspected recurrence of colorectal carcinoma.
Patient Population From June 1991 to June 1996, the nuclear medicine department at the University Hospital of Leuven performed a whole-body FDG-PET scan in 172 consecutive patients with a clinical suspicion of recurrent colorectal cancer or for preoperative staging in the case of previously proven recurrence. Only patients with a clinical and radiologic follow-up by the clinician at the university hospital were considered (n = 124). Patients who had undergone previous chemotherapeutic treatment for advanced disease (n = 8) and those with known inflammatory bowel disease (n = 6) were excluded. Patients with a time interval of more than 2 months between the abdominal computed tomographic (CT) scan and the PET scan were also excluded (n = 7). The selected group consisted of 103 patients (62 men and 41 women), with a mean age of 61 years (range, 35 to 81 years). All patients had undergone surgical resection of a primary adenocarcinoma located in the colon (n = 24) or in the rectum (n = 79). The mean time interval between primary surgery of the colorectal cancer and the FDG-PET scan was 580 days (range, 2 months to 11 years).
CDM
PET
Data Analysis Suspected lesions identified by one of the CDM studies or FDG-PET were grouped into four regions: locoregional (ie, the operative site), abdominal cavity (including mesenteric, peritoneal, and retroperitoneal metastases), liver (divided into right and left lobe), and extra-abdominal organs.
Region-based analysis.
Patient-based analysis.
Statistical Analysis
Region-Based Analysis Table 1 shows the distribution of the discordances between CDM and FDG-PET per anatomic region and the methodology used to define the true lesion status. A total of 40 regional discordances were found in 103 patients. The true lesion status of these discordances was determined by histologic examination in 22 cases (55%) and by follow-up in 18 cases (45%). The results of the analysis of the discordances between CDM and FDG-PET findings in the different anatomic regions are summarized in Table 2.
Locoregional recurrence.
Liver metastasis.
Abdominal cavity.
Extra-abdominal metastasis.
Patient-Based Analysis
In the patient subgroup with presumed resectable locoregional recurrence (n = 23), a discordant FDG-PET status was found in eight patients. Additional diagnostic value of FDG-PET was found in eight of 23 patients (35%), based on the detection of additional unsuspected tumor sites in five patients and the exclusion of recurrent disease activity in three patients. The additional true-positive tumor sites detected by FDG-PET were located in the axilla (n = 1), abdominal cavity (n = 2), and lungs (n = 2). An example of the additional value of FDG-PET in a patient with presumed limited pelvic recurrence is shown in Fig 1.
A third patient subgroup consisted of 22 patients with extended disease, documented by conventional imaging techniques, who were not eligible for curative surgery. In these patients, a discordant FDG-PET result was found in only two patients; FDG-PET result was true in one patient and false in the other. The latter patient's classification was incorrectly downstaged by FDG-PET owing to a normal FDG distribution in the liver despite the presence of bilobar metastases. The other patient's classification was correctly downstaged by FDG-PET because of a true-negative FDG-PET finding in the pelvis; the patient was thus classified as having an operable solitary liver metastasis. Thus, in this patient subgroup, FDG-PET imaging did not yield significant additional diagnostic information. A fourth subset of patients consisted of eight patients with completely normal CDM findings. In these patients, FDG-PET did not detect any additional tumor site. They all remained tumor-free during follow-up (range, 12 to 39 months). In another patient subgroup (n = 8), the only indicator for recurrent disease activity was an elevated plasma CEA level (mean, 38 ng/mL; range, 9 to 56 ng/mL; normal range, 0 to 2.5 ng/mL). All conventional imaging studies were negative. FDG-PET correctly detected tumor recurrence in three of these patients. The lesions were located in the liver, at the postsurgical site in the pelvis, and in a neck lymph node. In this subgroup, one false-positive FDG-PET lesion was reported in the lung on the basis of FDG uptake in a tuberculotic lesion. During follow-up of the patients with normal FDG-PET findings, one patient remained tumor-free (follow-up, 48 months), two patients had a local recurrence, and one patient developed diffuse peritoneal metastasis 8 months later. A last patient subgroup consisted of five patients whose only abnormal CDM sign was an equivocal CT scan reading at the postsurgical site in the pelvis (n = 4) or in the liver (n = 1). FDG-PET correctly classified these findings as true-negative in two patients and true-positive in three patients. In terms of patient management, these subgroups can be regrouped into three patient categories. The first category consists of 60 patients in whom, on the basis of CDM findings, surgical resection of the locoregional or liver recurrences could be considered. In this category, additional FDG-PET value was found in 12 of 60 patients (20%). The additional value of FDG-PET in this patient subset was statistically significant (P = .006). A second category consists of 13 patients with inconclusive CDM findings or with elevated CEA plasma levels with normal conventional imaging results. Additional FDG-PET value was found in eight of 13 patients (62%). The additional value of FDG-PET in this patient subset was statistically significant (P = .045). A third category consists of eight patients with a clinical suspicion of recurrent disease but with completely normal CDM findings and 22 patients in whom CDM detected an extended-disease stage. FDG-PET scanning did not offer additional diagnostic information in these latter categories.
This study confirms the significant role for whole-body FDG-PET imaging as a complementary diagnostic modality to conventional staging techniques in the follow-up of patients with resected colorectal adenocarcinoma. Major areas in which FDG-PET had a particular positive impact on diagnosis were the biochemical characterization of lesions that are indeterminate on anatomic imaging, the selection of patients most likely to benefit from surgical resection of recurrences, and the localization of occult metastatic disease. One of the main applications of FDG-PET in recurrent colorectal cancer patients has been the biochemical characterization of pelvic lesions that are indeterminate on structural imaging procedures, such as CT or magnetic resonance imaging.13,14 Most of these studies have relatively small numbers of patients, but in all studies, the sensitivity and specificity of FDG-PET imaging were very high and better than those of CT, magnetic resonance imaging, or radioimmunoscintigraphy. Our study specifically addressed the concordance of FDG-PET with the referring clinician's diagnosis inferred from the integrated data of the physical examination, plasma CEA monitoring, CT scan, endoscopy, and TREUS. From this point of view, a true-discordant FDG-PET diagnosis was found in 14% of the patients, which is significantly lower than expected from the literature. This could be explained by more extensive and accurate CDM studies, including high-quality spiral CT scan and TREUS, although a formal comparison with CDM performed in several centers is not feasible. The major part of the additional value of FDG-PET in this situation resided in the subgroup of patients with inconclusive CDM findings and consisted in a significant decrease of equivocal FDG-PET readings compared with the CDM. This is illustrated by the fact that if TREUS is part of the CDM, it reduces the additional diagnostic value of FDG-PET in parallel to a dramatic decrease of equivocal reports. In the subset of patients (n = 25) who underwent both TREUS and CT scanning, the additional diagnostic value of FDG-PET for the detection of pelvic tumor recurrence fell from 56% if only the CT scan was considered, to 20% if the CT scan and TREUS were considered together. Several reports have indeed advocated TREUS as a more effective noninvasive technique than CT for the detection of local recurrences of rectal carcinoma in the pelvis.15-17 However, a direct comparison of the efficacy of TREUS versus FDG-PET has not been reported. Locoregional pelvic recurrence and liver metastases are the major sites of relapse after resection of colorectal cancer. Surgical resection is the only curative therapy in these patients, but recurrence is frequent, and considerable mortality and morbidity have been reported. Therefore, before these major surgical interventions are embarked on, it is imperative that there be no disease outside the operation field. The conventional imaging modality of choice for evaluating these patients is the CT scan, which is relatively sensitive for detecting liver metastases but less so for lymph node metastases and mesenteric involvement. Moreover, the limited field of view can lead to the overlooking of unsuspected metastatic lesions outside the abdomen and thorax. In this study, 60 patients were considered for surgical resection of presumed solitary recurrent local or hepatic disease, on the basis of the findings of a complete conventional work-up. Adding FDG-PET imaging in this patient group potentially changed the therapeutic management in 12 patients (20%) by correctly upstaging the classification of nine patients to an inoperable, extended-disease stage and by correctly downstaging the classification of three patients to a disease-free status. These findings are in good agreement with recently published data of Delbeke et al,5 who reported a change in surgical management in 28% of the 52 studied patients. More successful results of the FDG-PET technique in this particular setting have been reported by Lai et al,6 who detected unsuspected extrahepatic malignant disease that was missed by CDM in 32% of 34 patients. The region-based analysis indicated that the beneficial impact of FDG-PET on disease staging mainly results from a significant additional diagnostic accuracy for the diagnosis of locoregional recurrence, from a higher sensitivity for abdominal lymph node metastasis, and from the detection of unsuspected extra-abdominal metastatic sites. Our study therefore shows a definite potential of FDG-PET for improving patient selection and reducing inappropriate surgical interventions. The potential benefit of FDG-PET is particularly appealing in the recurrent colorectal cancer patient with a rising CEA who has a negative conventional work-up. Several studies have independently demonstrated the high accuracy of FDG-PET in localizing the source of recurrence in patients with occult disease based on elevated CEA serum levels. Valk et al18 reported true-positive FDG-PET findings in 70% of 27 patients with CEA elevation and a normal CT scan. In our study, 57 patients had elevated CEA levels; only eight (14%) of these patients had normal CDM findings, suggesting the relative rarity of this clinical situation. FDG-PET correctly localized metastatic sites in three patients but was false-positive in one of them. These findings indeed suggest a significant additional value of FDG-PET imaging in such patients. However, larger prospective and multicenter studies are needed to confirm the role of FDG-PET imaging in this particular setting. FDG is not a very tumor-specific substance, inasmuch as the leukocytes and macrophages of inflammatory processes also accumulate the tracer. This is a major source of false-positive diagnoses in the application of FDG-PET in oncology.19 Moreover, physiologic tracer activity in the normal gastrointestinal or urinary tract can sometimes mimic pathologic lesions. Nevertheless, we reported only two false-positive FDG-PET readings due to increased FDG uptake in inflammatory lesions in lungs and colon. Importantly, these two false-positive FDG-PET readings resulted in an overstaging of the classification of two patients with presumed limited and resectable recurrent disease to an extended-disease stage, possibly leading to the cancellation of potentially curative surgery. Therefore, to reduce the potentially negative impact of occasional false-positive FDG-PET results on patient management, it is absolutely mandatory to carefully select the candidates for an FDG-PET scan, excluding those with known inflammatory lung or bowel disease, to correlate closely the FDG-PET images with data from conventional imaging methods, and to carefully confirm by other means the additional FDG-PET lesions that significantly alter patient management. A second limitation of PET is the limited spatial resolution of the imaging tool, leading to false-negative reports in small lesions. Importantly, three of the four false-negative discordant PET findings in our study were indeed small-volume tumors (< 1 cm in diameter). This finding indicates that the limited spatial resolution of the imaging method, rather than an insufficient tracer concentration in the tumor, is the major reason for the nonvisualization of these lesions. It is clear that future technologic innovations that optimize the spatial resolution of the PET systems will significantly increase the additional diagnostic value of PET imaging in staging of recurrent colorectal cancer. The observation that false-negative FDG-PET findings are due to resolution rather than metabolic constraints again emphasizes the importance of having a detailed knowledge of the morphologic and structural characteristics of the suspected tumor before excluding tumor recurrence on the basis of FDG-PET scan findings. A major criticism of the current study could be the retrospective character and the interpretation of the PET images with a priori knowledge of results of the conventional diagnostic procedures. However, the purpose of this study was to assess the diagnostic yield of FDG-PET in addition to the CDM in different patient subsets. For this, we specifically focused on the discordances between both modalities to characterize the type of patients that would most likely benefit from performing an additional PET scan after a set of conventional investigations. Differences in sensitivity or specificity between CDM and FDG-PET or the value of FDG-PET as a first-line screening technique can only be inferred from a prospective study with blinded readings. A second limitation of this study could be the methodology to define the true lesion status of the discordances. Whereas it would have been ideal to have a histologic diagnosis of all discordant lesions, 22 (55%) of 40 lesions were determined only by clinical follow-up. In our opinion, the possibility that this had a significant impact on the results and conclusions is very small, mainly because a minimal follow-up of 12 months was required. A relapse of a colorectal carcinoma, once present, is usually a relatively fast-evolving tumor. It is known that the median survival of patients with advanced or metastatic colorectal cancer is 6 months. Therefore, it would be extremely unusual that metastases did not progress over the period of 1 year, especially when the masses were evaluated by means of a CT scan. However, a possible error in the true lesion status could occur when a nonsurgical antineoplastic therapy was initiated in a patient with a positive or equivocal CDM and a negative PET result. Under these circumstances, a slower tumor progression could indeed have led to a false-negative instead of a true-negative FDG-PET discordance. In our study, this potential bias was of marginal importance, as this particular condition was found only once, in a patient with extended recurrent cancer who had a positive CDM and a negative PET finding in the liver. Palliative chemotherapy was initiated, but a follow-up CT scan showed a fulminatory progression of the extrahepatic recurrences, whereas the hepatic lesions remained unchanged and were at that moment interpreted as benign biliary cysts. In conclusion, significant additional diagnostic value of FDG-PET performed in conjunction with CDM was found in 20% of patients with presumed resectable hepatic or pelvic recurrence, as well as in 62% of patients with inconclusive CDM findings or with elevated plasma CEA levels with a normal work-up. The region-based analysis indicates that the beneficial impact of FDG-PET on disease staging mainly results from a significant positive impact of FDG-PET in the case of inconclusive CDM findings in the pelvis, from a higher sensitivity for abdominal lymph node metastasis, and from the detection of unsuspected extra-abdominal metastatic sites.
Presented in part at the Annual Meeting of the American Society of Clinical Oncology, Los Angeles, CA, May 1998.
1. Cancer Facts and Figures: 1996. New York, NY, American Cancer Society, 1996 2. August DA, Ottow RT, Sugarbaker PH: Clinical perspective of human colorectal cancer metastasis. Cancer Metastasis Rev 3:303-324, 1984[Medline] 3. Beart RW Jr: Colon and Rectum, in Abeloff MD, Armitage JQ, Lichter AS, et al (eds): Clinical Oncology. New York, NY, Churchill Livingstone, 1995, pp 1267-1286 4. Wanebo HJ, Chu QD, Vezeridis MP, et al: Patient selection for hepatic resection of colorectal metastases. Arch Surg 131:322-329, 1996[Abstract]
5.
Delbeke D, Vitola JV, Sandler MP, et al: Staging recurrent metastatic colorectal carcinoma with PET. J Nucl Med 38:1196-1201, 1997 6. Lai DT, Fulham M, Stephen MS, et al: The role of whole-body positron emission tomography with [18F]fluorodeoxyglucose in identifying operable colorectal cancer metastases to the liver. Arch Surg 131:703-707, 1996[Abstract] 7. Schiepers C, Penninckx F, De Vadder N, et al: Contribution of PET in the diagnosis of recurrent colorectal cancer: Comparison with conventional imaging. Eur J Surg Oncol 21:517-522, 1995[Medline] 8. Beets G, Penninckx F, Schiepers C, et al: Clinical value of whole-body positron emission tomography with [18F]fluorodeoxyglucose in recurrent colorectal cancer. Br J Surg 81:1666-1670, 1994[Medline] 9. Pauwels E, McCready VR, Stoot JH, et al: The mechanism of accumulation of tumour-localising radiopharmaceuticals. Eur J Nucl Med 25:277-305, 1998[Medline]
10.
Dahlbom M, Hoffman EJ, Hoh CK: Whole-body positron emission tomography: I. Methods and performance characteristics. J Nucl Med 33:1191-1199, 1992 11. Shepp LA, Vardi Y: Maximum likelihood reconstruction for emission tomography. IEEE Trans Med Imaging 1:113-122, 1982 12. Lange K, Carson R: EM reconstruction algorithm for emission and transmission tomography. J Comput Assist Tomogr 8:306-316, 1984[Medline]
13.
Strauss LG, Clorius JH, Schlag P, et al: Recurrence of colorectal tumors: PET evaluation. Radiology 170:329-333, 1989
14.
Ito M, Takashi K, Tadokoro M, et al: Recurrent rectal cancer and scar: Differentiation with PET and MR imaging. Radiology 182:549-552, 1992 15. Hiele M, Gevers AM, Mertens L, et al: Comparison of endorectal ultrasound and CT scan in the preoperative staging of colorectal cancer. Gastroenterology 114:A610, 1998 (abstr) 16. Mascagni D, Corbellini L, Urciuoli P, et al: Endoluminal ultrasound for early detection of local recurrence of rectal cancer. Br J Surg 76:1176-1180, 1989[Medline] 17. Thoeni RF: Gastrointestinal cancer: Radiologic diagnosis and staging. Radiol Clin North Am 35:457-485, 1997[Medline] 18. Valk P, Abella-Columna E, Tesar D, et al: Detection of recurrent colorectal cancer by FDG-PET in patients with serum CEA elevation. J Nucl Med 39:135P, 1998 (abstr) 19. Strauss LG: Fluorine-18 deoxyglucose and false-positive results: A major problem in the diagnostics of oncological patients. Eur J Nucl Med 23:1409-1415, 1996[Medline] Submitted July 22, 1998; accepted November 20, 1998. This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
