- Email: [email protected]
The role of fluorodeoxyglucose and positron emission tomography in the evaluation of pancreatic disease
The role of fluorodeoxyglucose and positron emission tomography in the evaluation of pancreatic disease Ashwani Rajput, MD, Thomas A. Stellato, MD, Peter F. Faulhaber, MD, Hubert J. Vesselle, MD, and Floro Miraldi, MD, Cleveland, Ohio
Background. The difficulties involved in the timely and accurate diagnosis of pancreatic disease are well known. The usual imaging modalities usually identify abnormalities but may not always differentiate malignancy from other condition such as scar tissue or chronic inflammation. The purpose of our study was to determine if fluorodeoxyglucose positron emission tomography (FDG PET) can accurately diagnose pancreatic disease. Methods. The records of 15 patients presenting with pancreatic disease were retrospectively reviewed. The diagnosis suspected by imaging modalities was compared with the final tissue diagnosis. Two patients were excluded because no tissue was obtained. Results. Adenocarcinoma was diagnosed in 9 patients. A mass consistent with this diagnosis was seen in 8 of 9, 6 of 9, 6 of 8, and 5 of 5 patients by PET, computed tomography (CT), endoscopic retrograde cholangiopancreatography (ERCP), and endoscopic ultrasound (EUS), respectively. Chronic pancreatitis (CP) was diagnosed in 2 patients. The unique appearance on FDG PET made the diagnosis in both these patients. Both patients with CP were thought to have a malignancy by CT and EUS and 1 of 2 by ERCP. Neuroendocrine tumors were diagnosed in 2 other patients. One of 2 was seen by FDG PET and both by CT. Conclusions. FDG PET can accurately differentiate a pancreatic adenocarcinoma from chronic pancreatitis in a patient with a suspicious pancreatic mass. Thus, FDG PET may help in establishing a diagnosis and subsequently managing a patient with pancreatic disease. (Surgery 1998;124:793-8.) From the Departments of Surgery and Radiology, Case Western Reserve University, Cleveland, Ohio
IN 1998 IT IS estimated that in the United States 29,000 new cases of pancreatic cancer will be diagnosed and an estimated 28,900 people will die of this disease.1 Early diagnosis is essential for the treatment of this aggressive malignancy. The timely and acurate diagnosis of pancreatic disease, however, is difficult. The usual imaging modalities such as computed tomography (CT), ultrasound, endosocpic retrograde cholangiopancreatography (ERCP), and endoscopic ultrasound (EUS) are able to identify abnormalities but are not always able to differentiate a malignancy from other disease processes such as scar tissue, necrosis, or chronic inflammation. The unique biology of cancer cells allows the functional imaging performed by positron emisPresented at the Fifty-fifth Annual Meeting of the Central Surgical Association, Ann Arbor, Mich, March 5-7, 1998. Reprint requests: Thomas A. Stellato, MD, Department of Surgery, Case Western Reserve University, 10900 Euclid Ave, Cleveland, OH 44106. Copyright © 1998 by Mosby, Inc. 0039-6060/98/$5.00 + 0
11/6/91226
sion tomography (PET) to differentiate malignancy from other disease processes. PET scanning with fluorine-18-labeled fluorodeoxyglucose (FDG) has been used successfully to diagnose a malignancy in various organ systems.2-4 The purpose of our study was to determine whether FDG PET could accurately diagnose pancreatic disease. MATERIAL AND METHODS The records of 15 patients presenting between March 1995 and August 1996 with possible pancreatic disease were retrospectively reviewed. There were 8 men and 7 women between the ages of 22 and 83 years in the series. The diagnosis suspected by the modalities of CT, ERCP, EUS , and FDG PET were compared with the final histologic diagnosis. Two patients were excluded because no tissue was obtained. PET scans were performed using an ECAT EXACT (CTI, Knoxville, Tenn) scanner. The scanner was operated in two-dimensional mode. The intrinsic resolution is 6 mm in both the transverse and axial planes. The axial field of view is 16 cm. Patients were fasted overnight before scanning. An SURGERY 793
794 Rajput et al
Surgery Octoaber 1998
Table I. Mass lesion seen by Patient 1 2 3 4 5 6 7 8 9 10 11 12 13
PET
CT
ERCP
EUS
Size (cm)
+ – + + + + + + + + – – –
+ + + – + – – + + + + + +
nd + + + + + + – – – – – +
nd + nd + + + nd + nd nd nd + +
1.5 5.0 2.0 6.0 3.5 3.0 4.0 4.5 2.0 2.7 3.0 2.5 5.3
Histology Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Neuroendocrine tumor Neuroendocrine tumor Chronic pancreatitis Chronic pancreatitis
Size of tumor is given as maximum dimension in cm. +, Positive for mass lesion; –, negative for mass lesion; nd, not done.
imaging protocol was established at our institution after the first 5 patients were scanned. Patients 6 through 13 underwent scanning with the following protocol. These patients had a bowel preparation with an isosmotic solution the evening before examination. Between 407 MBq and 802 MBq of F-18 FDG was administered intravenously along with 500 mL crystalloid. Thirty minutes after administration of the radioisotope, 20 mg of furosemide was administered for diuresis and subsequent prevention of F-18 FDG accumulation in the urinary system. Forty-five minutes after the administration of the F-18 FDG, tomographic images were obtained. Initially, images of one bed to include the pancreas were obtained. Four additional beds were imaged to cover the torso. FDG PET images were qualitatively reviewed by 3 radiologists who were blinded to the patient’s diagnosis. Quantitative analysis using standardized uptake values was not done. The following criteria were used for reading the scans. Focal areas of increased activity were classified as malignant. Diffuse uptake within the pancreas was defined as pancreatitis. RESULTS As shown in Table I,the diagnosis of adenocarcinoma was made in 9 of the 13 patients. Focally increased FDG on FDG PET scan consistent with adenocarcinoma was seen in 8 of the 9 patients. CT, ERCP, and EUS showed a distinct mass in 6 of 9, 6 of 8, and 5 of 5 patients, respectively (Figs 1 and 2). Two patients were found to have chronic pancreatitis. The pattern of uptake on FDG PET scans in these two patients was consistent with the diagnosis of chronic pancreatitis; however, CT scans showed
masses in both of these patients suspicious for an andenocarcinoma (Figs 3 and 4). ERCP and EUS were also suspicious for adenocarcinoma in 1 of 2 and 2 of 2 patients, respectively. Two other patients were diagnosed with neuroendocrine tumors. CT successfully identified a mass lesion in both patients with neuroendocrine tumors. FDG PET showed an increased focal uptake in a patient with a vasoactive peptide-secreting tumor but failed to identify a multiple islet cell tumor. Thus there were no false-positive results by FDG PET, but CT had 2 false-positive results for malignancy. Both of these false-positive results were in patients who had a tissue diagnosis of pancreatitis that was successfully diagnosed by the unique appearance on FDG PET. CT had 3 false-negative results, whereas PET had one false-negative result for adenocarcinoma (Table II). DISCUSSION The physiologic basis for the use of FDG PET technology rests on the discovery of Warburg5 that tumor cells as compared with normal cells have an increased rate of glycolysis and thus an increased uptake of glucose. Advances in molecular and cellular biology have revealed that this increased uptake is accomplished by the activation of glucose transporters in malignant cells.6-8 FDG is a glucose analog that is taken up by cellular glucose transporters. Once in the cell, FDG is phosphorylated by hexokinase. FDG-6-phosphate cannot be further metabolized by the cell, and it cannot diffuse out of a cell. Thus it remains within the cell, and as it slowly degrades, it can serve as a marker for cells involved in increased glucose uptake such as cells comprising a pancreatic adenocarcinoma.
Rajput et al 795
Surgery Volume 124, Number 4
Our results show that when a patient presents with a question of pancreatic disease and with CT, ERCP, and EUS results that are indiscriminate, FDG PET can successfully differentiate an adenocarcinoma from chronic pancreatitis in a suspicious pancreatic mass. Both patients in our study with chronic pancreatitis were falsely thought to have malignant lesions based on the results of CT and EUS. FDG PET results, however, were typical of chronic pancreatitis. In the 9 patients with adenocarcinoma, CT scan resulted in 3 falsely negative diagnoses. The FDG PET scan correctly diagnosed 8 of the 9 patients. The one false-negative result (patient 2) with the FDG PET, despite a 5 cm mass, was in a patient with insulin-dependent diabetes mellitus. Serum glucose levels at the time of PET scan were not determined in our study. Uptake values of FDG have been shown to be higher in the fasted as opposed to the glucose loaded state. In diabetics, FDG uptake has also been shown to be decreased and therefore results in a false-negative finding.9 Furthermore, this patient underwent FDG PET scanning before the institution of our protocol of bowel cleansing, hydration, and diuresis. Reexamining this patient’s PET study shows the increased uptake in her right kidney that masks a less intense area of uptake that correlates with mass seen on CT scan. Perhaps if the protocol had been in place, this patient would not have been missed. The decreased uptake as compared with the accumulation in the kidney is explained by the necrotic nature of this patient’s tumor as seen on pathologic examination. There were no false-positive results in our series. This may be due to the small number of patients. Nonetheless, clinicians using the data generated by FDG PET scanning should be aware of the possible reasons for false-positive results. The work by Kubota et al10 has demonstrated that FDG has high levels of accumulation in macrophages and granulation tissue. Thus acute inflammatory processes such as an abdominal abscess will accumulate FDG.11 In one of 2 patients with a neuroendocrine tumor, a mass was identified with FDG PET. These tumors that arise from islet cells within the pancreas have unique receptors on the cell surface and are best studied by peptide receptor scintigraphy.12 This method has been shown to be effective in diagnosing islet cell tumors and differentiating islet cell tumors from pancreatic ductal cell cancers.13,14 The goal of this paper was not to evaluate FDG PET for its ability to detect metastatic disease at the time of diagnosis or in follow-up. However, while reviewing the patients’ records, the possible role of FDG PET in staging and follow-up was highlighted.
Fig 1. CT scan of patient with a mass in the head of the pancreas. The tissue diagnosis was of adenocarcinoma.
Table II. Accuracy of diagnostic modalities Accurately diagnosed by n
Histology
9 Adenocarcinoma 2 Chronic pancreatitis
PET
CT
ERCP
EUS
8/9 2/2
6/9 0/2
6/8 0/2
5/5 0/2
Patient 1, for example, had a CT scan done 2 months before her FDG PET scan. The CT scan showed the pancreatic mass but did not reveal any metastatic disease. The PET scan, however, showed abnormal accumulations in the liver consistent with metastatic disease. Patient 3 had a similar course of events with a CT scan negative for metastasis, whereas a PET scan showed abnormal foci in the liver. This patient eventually succumbed to widespread metastasis. A final example is patient 4. This patient underwent a pyloric-sparing pancreatoduodenectomy for an adenocarcinoma in the head of the pancreas. Eight months after the operation, the patient presented with complaints of vague pain, poor appetite, and weight loss. A work-up including physical examination, blood work including liver function tests and CA 19-9, plain x-ray films, CT scans of the chest and abdomen, and a magnetic resonance imaging (MRI) scan failed to reveal any metastatic disease. An FDG PET scan, however, revealed tumor recurrence in the pancreatic bed and metastases to the lungs (Fig 5). There are some preliminary data suggesting that FDG PET may be used to define the prognosis of patients with pancreatic adenocarcinoma.15 This is based on quantitative analysis of FDG uptake and correlation with the mean time of survival. Further study with a larger series of patients is needed to help define the role of FDG PET in staging, follow-up, and prognosis.
796 Rajput et al
Surgery Octoaber 1998
Fig 2. Shown are representative whole body FDG PET images in coronal, sagittal, and axial planes of the patient in Fig 1. The intense FDG uptake in the pancreatic head is typical for an adenocarcinoma.
Fig 3. CT scan shows a mass in the head of the pancreas. The tissue diagnosis was of chronic pancreatitis.
Fig 4. Shown are representative whole body FDG PET images in the coronal, sagittal, and axial planes of the patient in Fig 3. The less intense, diffuse uptake in the pancreas is typical for chronic pancreatitis.
Surgery Volume 124, Number 4
Rajput et al 797
Fig 5. Shown are the coronal and sagittal views of a FDG PET scan done for a metastatic evaluation in a patient who underwent a pancreatoduodenectomy 8 months previously for an pancreatic adenocarcinoma. Increased foci of uptake in the pancreatic bed, liver, and lung fields are diagnostic of recurrence and metastasis.
CONCLUSION The increasing knowledge in the basic science of tumor biology and the advances in technology arm the clinician with new modalities for diagnosing disease. One such modality that is proving to be highly effective in the diagnosis of pancreatic disease is FDG PET. Our results show that FDG PET can accurately differentiate a pancreatic adenocarcinoma from chronic pancreatitis in a patient with a suspicious pancreatic mass. Further study may delineate the role in staging and follow-up. Thus, FDG PET can aid the surgeon in establishing a diagnosis and subsequently managing a patient with pancreatic disease. REFERENCES 1. Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1998. CA Cancer J Clin 1998;48:6-29. 2. Bergstrom M. Positron emission tomography in tumor diagnosis and treatment follow-up. Acta Oncologica 1993;32:183-8. 3. Blahd WH, Brown CV, Khonsary SA, Farahi JB, Quinones N, Ribe JY, et al. PET scans of abdominal malignancy. World J Surg 1996;20:245-7. 4. Gupta N, Bradfield H. Role of positron emission tomography scanning in evaluating gastrointestinal neoplasms. Semin Nucl Med 1996;26:65-73. 5. Warburg O. On the origin of cancer cells. Science 1956;123:309-14. 6. Schonberger JA, Hartung T, Stollfuss JC, Beger HG, Reske SN. Increased FDG-uptake in human pancreatic carcinoma is associated to over-expression of glucose-transporter gene (GLUT-1)[abstract]. J Nucl Med 1994;35:117P. 7. Kornrumpf D, Buchler M, Langhans J, Beger HG, Reske SN. Concordantly increased glucose transporter 1 (GLUT 1) gene expression and FDG uptake in human pancreatic carcinoma [abstract]. J Nucl Med 1993;34:223P.
8. Flier JS, Mueckler MM, Usher P, Lodish HF. Elevated levels of glucose transport and transporter messsenger RNA are induced by ras or src oncogenes. Science 1987;235:1492-5. 9. Bares R, Klever P, Hellwig D, Hauptmann S, Fass J, Hambuechen U, et al. Pancreatic cancer detected with 18Flabelled deoxyglucose: method and first results. Nucl Med Commun 1993;14:596-601. 10. Kubota R, Yamada S, Kubota K, Ishiwata K, Tamahashi N, Ido T. Intratumoral distribution of fluorine-18-fluorodeoxyglucose in vivo: high accumulaiton in macrophages and granulation tissues studied by microautoradiography. J Nucl Med 1992;33:1972-80. 11. Tahara T, Ichiya Y, Kuwabara Y, Otsuka M, Miyake Y, et al. High 18 F fluorodeoxyglucose uptake in abdominal abscesses: a PET study. J Comput Assist Tomogr 1989; 13:829-31. 12. Lamberts SWJ, Reubi JC, Krenning EP. Somatostatin and the concept of peptide receptor scintigraphy in oncology. Semin Oncol 1994;21:1-5. 13. van Eyck CHJ, Bruining HA, Reubi JC, Bakker WH, Oei HY, Krenning EP, et al. Use of isotope-labeled somatostatin analogs for visualization of islet cell tumors. World J Surg 1993;17:444-7. 14. van Eijck CHJ, Lamberts SWJ, Lemaire LCJM, Jeekel H, Bosman FT, Reubi JC, et al. The use of somatostatin receptor scintigrapy in the differential diagnosis of pancreatic duct cancers and islet cell tumors. Ann Surg 1996;224:119-24. 15. Nakata B, Chung YS, Nishimura S, Nishihara T, Sakurai Y, Sawada T, et al. 18 F-Fluordeoxyglucose positron emission tomography and the prognosis of patients with pancreatic adenocarcinoma. Cancer 1997;79:695-9.
DISCUSSION Dr James A. Madura (Indianapolis, Ind). When I first received the manuscript, I thought it represented a fairly small and insignificant series of nonhomogeneous patients. However, a MEDLINE search failed to produce any other series, large or small, in the surgical literature. Most of the reports are in the radiologic and nuclear medicine literature, and very few are from the United
798 Rajput et al
States. I believe there is a single large series from St Louis, and the rest are from Japan and Europe. In the literature I was able to gather about 300 patients with pancreatic diseases, and the results of PET scanning were about the same as yours. There was a 90% to 93% accuracy rate, but a 5% false-positive rate for pancreatic malignancies. Many of the false-positive results were in less lethal neoplasms such as cystic neoplasms, intraductal malignancies, and neuroendocrine tumors. I know you had several of these lesions in your series. In addition, in a series from Osaka, Japan, in a group of patients about equal in number to your series, they report that a higher uptake of the FDG resulted in a poorer prognosis. Would you comment on that issue as it relates to your patients? Does the PET scan really add anything to your workup of these patients? From the literature, the PET scan seems to have the same tumor size limitations as the CT and MRI scans. The maximal size detected in the few series reporting it was about 10 mm. Many times smaller undetected lesions make these cases unresectable. If you find a tumor on PET scan, does it change your therapeutic approach? Does it eliminate the need to operate? We believe that a lesion occluding the common or pancreatic duct probably needs an operation anyway. Although your PET scan photographs are nice, they are not as anatomically elegant as a CT or MRI scan, which show surgeons critical structures like blood vessels. My last comment is about cost. At our hospital a PET scan costs the patient about $2100. To do all of the studies you did in these patients would cost about $8500. If you add fine-needle aspiration and pathology charges, this work-up is in excess of $10,000. Therefore tell us what you are doing now clinically to evaluate these patients. Is this simply a study you wish to add to the surgical literature, or are you really using it clinically in the patient with suspected pancreatic disease? Dr Rajput. You have a valid criticism that we have a small series. We hope to increase the number of patients in ongoing investigations. One of our goals, however, is to bring the knowledge of this technology to the surgical literature. In regards to your concern about false-positive studies, one must be aware that any acute inflammatory process may have an increased uptake of FDG. The use of FDG PET in cystic malignancies may result in false-negative studies. PET is based on the metabolic activity of cells. Thus, if there is not a significant cellular component to a cystic malignancy, the lesion may be missed. Neuroendocrine tumors are unique in that they have certain peptide receptors on their surfaces. Thus peptide receptor scintigraphy rather that FDG PET is being pursued for their diagnosis. The study you mention from Osaka, Japan, quantitatively analyzed the data in terms of standard uptake value
Surgery Octoaber 1998
of FDG and outcome. This study was based on the fact that a more metabolically active lesion will have a higher uptake value of FDG. They hypothesized that a more metabolically active lesion would be a more aggressive malignancy. The study found that patients with higher uptake values of FDG had a mean survival of 5 months and those with lower uptake values had a mean survival of 14 months. However, there were only 14 patients in the study. PET scanning is not in our standard armamentarium. Our usual work-up for pancreatic masses includes CT scan, ERCP, EUS, and laparoscopy. We are attempting to define the clinical role of PET scan. It may prove to be helpful in patients in whom the usual modalities do not provide a diagnosis or serious comorbid disease may make surgery a relative contraindication. PET may be most useful in the work-up of metastatic disease as evidenced by the last case I presented. If metastatic disease is identified, a palliative bypass, as opposed to a curative resection, would be offered and laparoscopy would be obviated. PET will not help assess for resectability; it is functional imaging as opposed to anatomic imaging. In regards to the size of the lesion that can be assessed, our PET scanner produces 6 mm sections. Thus a lesion would have to be 6 mm. There are preliminary data, however, that show very metabolically active lesions smaller than 6 mm can be detected. Cost is of concern to all of us in today’s health care climate. At the time our study was done, this was considered an experimental protocol and none of the patients were charged. Now, however, our institution charges about $2000 per study. We are in the process of negotiating with our “administocracy” to reduce the costs. Dr Henry Pitt (Milwaukee, Wis). I would like to compliment you for introducing this new, accurate modality. However, I share Dr Madura’s concerns about cost and the effort required to get good studies. In addition, I was surprised at the low accuracy of CT scanning in your patients. With newer spiral CT techniques the quality of the study is operator dependent. Were these CT scans done at your institution or elsewhere and what protocol was used? My second comment is that serum CA19-9 should be used when trying to differentiate pancreatic cancer from chronic pancreatitis. CA19-9 got a bad name when it was introduced with normal values below 37 to 40 U/mL. However, if 120 or 200 is used as the cutoff value, this test becomes very accurate especially in nonjaundiced patients. This test is relatively inexpensive and should probably be used more often in evaluating these patients. Dr Rajput. As we are a tertiary referral center, all the CT scans were not done at our institution with a rigid protocol. We appreciate your comments in regards to CA19-9. Because our study was retrospective, not all the patients had a CA19-9 level to help make the differentiation in suspected pancreatic disease.
Background. The difficulties involved in the timely and accurate diagnosis of pancreatic disease are well known. The usual imaging modalities usually identify abnormalities but may not always differentiate malignancy from other condition such as scar tissue or chronic inflammation. The purpose of our study was to determine if fluorodeoxyglucose positron emission tomography (FDG PET) can accurately diagnose pancreatic disease. Methods. The records of 15 patients presenting with pancreatic disease were retrospectively reviewed. The diagnosis suspected by imaging modalities was compared with the final tissue diagnosis. Two patients were excluded because no tissue was obtained. Results. Adenocarcinoma was diagnosed in 9 patients. A mass consistent with this diagnosis was seen in 8 of 9, 6 of 9, 6 of 8, and 5 of 5 patients by PET, computed tomography (CT), endoscopic retrograde cholangiopancreatography (ERCP), and endoscopic ultrasound (EUS), respectively. Chronic pancreatitis (CP) was diagnosed in 2 patients. The unique appearance on FDG PET made the diagnosis in both these patients. Both patients with CP were thought to have a malignancy by CT and EUS and 1 of 2 by ERCP. Neuroendocrine tumors were diagnosed in 2 other patients. One of 2 was seen by FDG PET and both by CT. Conclusions. FDG PET can accurately differentiate a pancreatic adenocarcinoma from chronic pancreatitis in a patient with a suspicious pancreatic mass. Thus, FDG PET may help in establishing a diagnosis and subsequently managing a patient with pancreatic disease. (Surgery 1998;124:793-8.) From the Departments of Surgery and Radiology, Case Western Reserve University, Cleveland, Ohio
IN 1998 IT IS estimated that in the United States 29,000 new cases of pancreatic cancer will be diagnosed and an estimated 28,900 people will die of this disease.1 Early diagnosis is essential for the treatment of this aggressive malignancy. The timely and acurate diagnosis of pancreatic disease, however, is difficult. The usual imaging modalities such as computed tomography (CT), ultrasound, endosocpic retrograde cholangiopancreatography (ERCP), and endoscopic ultrasound (EUS) are able to identify abnormalities but are not always able to differentiate a malignancy from other disease processes such as scar tissue, necrosis, or chronic inflammation. The unique biology of cancer cells allows the functional imaging performed by positron emisPresented at the Fifty-fifth Annual Meeting of the Central Surgical Association, Ann Arbor, Mich, March 5-7, 1998. Reprint requests: Thomas A. Stellato, MD, Department of Surgery, Case Western Reserve University, 10900 Euclid Ave, Cleveland, OH 44106. Copyright © 1998 by Mosby, Inc. 0039-6060/98/$5.00 + 0
11/6/91226
sion tomography (PET) to differentiate malignancy from other disease processes. PET scanning with fluorine-18-labeled fluorodeoxyglucose (FDG) has been used successfully to diagnose a malignancy in various organ systems.2-4 The purpose of our study was to determine whether FDG PET could accurately diagnose pancreatic disease. MATERIAL AND METHODS The records of 15 patients presenting between March 1995 and August 1996 with possible pancreatic disease were retrospectively reviewed. There were 8 men and 7 women between the ages of 22 and 83 years in the series. The diagnosis suspected by the modalities of CT, ERCP, EUS , and FDG PET were compared with the final histologic diagnosis. Two patients were excluded because no tissue was obtained. PET scans were performed using an ECAT EXACT (CTI, Knoxville, Tenn) scanner. The scanner was operated in two-dimensional mode. The intrinsic resolution is 6 mm in both the transverse and axial planes. The axial field of view is 16 cm. Patients were fasted overnight before scanning. An SURGERY 793
794 Rajput et al
Surgery Octoaber 1998
Table I. Mass lesion seen by Patient 1 2 3 4 5 6 7 8 9 10 11 12 13
PET
CT
ERCP
EUS
Size (cm)
+ – + + + + + + + + – – –
+ + + – + – – + + + + + +
nd + + + + + + – – – – – +
nd + nd + + + nd + nd nd nd + +
1.5 5.0 2.0 6.0 3.5 3.0 4.0 4.5 2.0 2.7 3.0 2.5 5.3
Histology Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Adenocarcinoma Neuroendocrine tumor Neuroendocrine tumor Chronic pancreatitis Chronic pancreatitis
Size of tumor is given as maximum dimension in cm. +, Positive for mass lesion; –, negative for mass lesion; nd, not done.
imaging protocol was established at our institution after the first 5 patients were scanned. Patients 6 through 13 underwent scanning with the following protocol. These patients had a bowel preparation with an isosmotic solution the evening before examination. Between 407 MBq and 802 MBq of F-18 FDG was administered intravenously along with 500 mL crystalloid. Thirty minutes after administration of the radioisotope, 20 mg of furosemide was administered for diuresis and subsequent prevention of F-18 FDG accumulation in the urinary system. Forty-five minutes after the administration of the F-18 FDG, tomographic images were obtained. Initially, images of one bed to include the pancreas were obtained. Four additional beds were imaged to cover the torso. FDG PET images were qualitatively reviewed by 3 radiologists who were blinded to the patient’s diagnosis. Quantitative analysis using standardized uptake values was not done. The following criteria were used for reading the scans. Focal areas of increased activity were classified as malignant. Diffuse uptake within the pancreas was defined as pancreatitis. RESULTS As shown in Table I,the diagnosis of adenocarcinoma was made in 9 of the 13 patients. Focally increased FDG on FDG PET scan consistent with adenocarcinoma was seen in 8 of the 9 patients. CT, ERCP, and EUS showed a distinct mass in 6 of 9, 6 of 8, and 5 of 5 patients, respectively (Figs 1 and 2). Two patients were found to have chronic pancreatitis. The pattern of uptake on FDG PET scans in these two patients was consistent with the diagnosis of chronic pancreatitis; however, CT scans showed
masses in both of these patients suspicious for an andenocarcinoma (Figs 3 and 4). ERCP and EUS were also suspicious for adenocarcinoma in 1 of 2 and 2 of 2 patients, respectively. Two other patients were diagnosed with neuroendocrine tumors. CT successfully identified a mass lesion in both patients with neuroendocrine tumors. FDG PET showed an increased focal uptake in a patient with a vasoactive peptide-secreting tumor but failed to identify a multiple islet cell tumor. Thus there were no false-positive results by FDG PET, but CT had 2 false-positive results for malignancy. Both of these false-positive results were in patients who had a tissue diagnosis of pancreatitis that was successfully diagnosed by the unique appearance on FDG PET. CT had 3 false-negative results, whereas PET had one false-negative result for adenocarcinoma (Table II). DISCUSSION The physiologic basis for the use of FDG PET technology rests on the discovery of Warburg5 that tumor cells as compared with normal cells have an increased rate of glycolysis and thus an increased uptake of glucose. Advances in molecular and cellular biology have revealed that this increased uptake is accomplished by the activation of glucose transporters in malignant cells.6-8 FDG is a glucose analog that is taken up by cellular glucose transporters. Once in the cell, FDG is phosphorylated by hexokinase. FDG-6-phosphate cannot be further metabolized by the cell, and it cannot diffuse out of a cell. Thus it remains within the cell, and as it slowly degrades, it can serve as a marker for cells involved in increased glucose uptake such as cells comprising a pancreatic adenocarcinoma.
Rajput et al 795
Surgery Volume 124, Number 4
Our results show that when a patient presents with a question of pancreatic disease and with CT, ERCP, and EUS results that are indiscriminate, FDG PET can successfully differentiate an adenocarcinoma from chronic pancreatitis in a suspicious pancreatic mass. Both patients in our study with chronic pancreatitis were falsely thought to have malignant lesions based on the results of CT and EUS. FDG PET results, however, were typical of chronic pancreatitis. In the 9 patients with adenocarcinoma, CT scan resulted in 3 falsely negative diagnoses. The FDG PET scan correctly diagnosed 8 of the 9 patients. The one false-negative result (patient 2) with the FDG PET, despite a 5 cm mass, was in a patient with insulin-dependent diabetes mellitus. Serum glucose levels at the time of PET scan were not determined in our study. Uptake values of FDG have been shown to be higher in the fasted as opposed to the glucose loaded state. In diabetics, FDG uptake has also been shown to be decreased and therefore results in a false-negative finding.9 Furthermore, this patient underwent FDG PET scanning before the institution of our protocol of bowel cleansing, hydration, and diuresis. Reexamining this patient’s PET study shows the increased uptake in her right kidney that masks a less intense area of uptake that correlates with mass seen on CT scan. Perhaps if the protocol had been in place, this patient would not have been missed. The decreased uptake as compared with the accumulation in the kidney is explained by the necrotic nature of this patient’s tumor as seen on pathologic examination. There were no false-positive results in our series. This may be due to the small number of patients. Nonetheless, clinicians using the data generated by FDG PET scanning should be aware of the possible reasons for false-positive results. The work by Kubota et al10 has demonstrated that FDG has high levels of accumulation in macrophages and granulation tissue. Thus acute inflammatory processes such as an abdominal abscess will accumulate FDG.11 In one of 2 patients with a neuroendocrine tumor, a mass was identified with FDG PET. These tumors that arise from islet cells within the pancreas have unique receptors on the cell surface and are best studied by peptide receptor scintigraphy.12 This method has been shown to be effective in diagnosing islet cell tumors and differentiating islet cell tumors from pancreatic ductal cell cancers.13,14 The goal of this paper was not to evaluate FDG PET for its ability to detect metastatic disease at the time of diagnosis or in follow-up. However, while reviewing the patients’ records, the possible role of FDG PET in staging and follow-up was highlighted.
Fig 1. CT scan of patient with a mass in the head of the pancreas. The tissue diagnosis was of adenocarcinoma.
Table II. Accuracy of diagnostic modalities Accurately diagnosed by n
Histology
9 Adenocarcinoma 2 Chronic pancreatitis
PET
CT
ERCP
EUS
8/9 2/2
6/9 0/2
6/8 0/2
5/5 0/2
Patient 1, for example, had a CT scan done 2 months before her FDG PET scan. The CT scan showed the pancreatic mass but did not reveal any metastatic disease. The PET scan, however, showed abnormal accumulations in the liver consistent with metastatic disease. Patient 3 had a similar course of events with a CT scan negative for metastasis, whereas a PET scan showed abnormal foci in the liver. This patient eventually succumbed to widespread metastasis. A final example is patient 4. This patient underwent a pyloric-sparing pancreatoduodenectomy for an adenocarcinoma in the head of the pancreas. Eight months after the operation, the patient presented with complaints of vague pain, poor appetite, and weight loss. A work-up including physical examination, blood work including liver function tests and CA 19-9, plain x-ray films, CT scans of the chest and abdomen, and a magnetic resonance imaging (MRI) scan failed to reveal any metastatic disease. An FDG PET scan, however, revealed tumor recurrence in the pancreatic bed and metastases to the lungs (Fig 5). There are some preliminary data suggesting that FDG PET may be used to define the prognosis of patients with pancreatic adenocarcinoma.15 This is based on quantitative analysis of FDG uptake and correlation with the mean time of survival. Further study with a larger series of patients is needed to help define the role of FDG PET in staging, follow-up, and prognosis.
796 Rajput et al
Surgery Octoaber 1998
Fig 2. Shown are representative whole body FDG PET images in coronal, sagittal, and axial planes of the patient in Fig 1. The intense FDG uptake in the pancreatic head is typical for an adenocarcinoma.
Fig 3. CT scan shows a mass in the head of the pancreas. The tissue diagnosis was of chronic pancreatitis.
Fig 4. Shown are representative whole body FDG PET images in the coronal, sagittal, and axial planes of the patient in Fig 3. The less intense, diffuse uptake in the pancreas is typical for chronic pancreatitis.
Surgery Volume 124, Number 4
Rajput et al 797
Fig 5. Shown are the coronal and sagittal views of a FDG PET scan done for a metastatic evaluation in a patient who underwent a pancreatoduodenectomy 8 months previously for an pancreatic adenocarcinoma. Increased foci of uptake in the pancreatic bed, liver, and lung fields are diagnostic of recurrence and metastasis.
CONCLUSION The increasing knowledge in the basic science of tumor biology and the advances in technology arm the clinician with new modalities for diagnosing disease. One such modality that is proving to be highly effective in the diagnosis of pancreatic disease is FDG PET. Our results show that FDG PET can accurately differentiate a pancreatic adenocarcinoma from chronic pancreatitis in a patient with a suspicious pancreatic mass. Further study may delineate the role in staging and follow-up. Thus, FDG PET can aid the surgeon in establishing a diagnosis and subsequently managing a patient with pancreatic disease. REFERENCES 1. Landis SH, Murray T, Bolden S, Wingo PA. Cancer statistics, 1998. CA Cancer J Clin 1998;48:6-29. 2. Bergstrom M. Positron emission tomography in tumor diagnosis and treatment follow-up. Acta Oncologica 1993;32:183-8. 3. Blahd WH, Brown CV, Khonsary SA, Farahi JB, Quinones N, Ribe JY, et al. PET scans of abdominal malignancy. World J Surg 1996;20:245-7. 4. Gupta N, Bradfield H. Role of positron emission tomography scanning in evaluating gastrointestinal neoplasms. Semin Nucl Med 1996;26:65-73. 5. Warburg O. On the origin of cancer cells. Science 1956;123:309-14. 6. Schonberger JA, Hartung T, Stollfuss JC, Beger HG, Reske SN. Increased FDG-uptake in human pancreatic carcinoma is associated to over-expression of glucose-transporter gene (GLUT-1)[abstract]. J Nucl Med 1994;35:117P. 7. Kornrumpf D, Buchler M, Langhans J, Beger HG, Reske SN. Concordantly increased glucose transporter 1 (GLUT 1) gene expression and FDG uptake in human pancreatic carcinoma [abstract]. J Nucl Med 1993;34:223P.
8. Flier JS, Mueckler MM, Usher P, Lodish HF. Elevated levels of glucose transport and transporter messsenger RNA are induced by ras or src oncogenes. Science 1987;235:1492-5. 9. Bares R, Klever P, Hellwig D, Hauptmann S, Fass J, Hambuechen U, et al. Pancreatic cancer detected with 18Flabelled deoxyglucose: method and first results. Nucl Med Commun 1993;14:596-601. 10. Kubota R, Yamada S, Kubota K, Ishiwata K, Tamahashi N, Ido T. Intratumoral distribution of fluorine-18-fluorodeoxyglucose in vivo: high accumulaiton in macrophages and granulation tissues studied by microautoradiography. J Nucl Med 1992;33:1972-80. 11. Tahara T, Ichiya Y, Kuwabara Y, Otsuka M, Miyake Y, et al. High 18 F fluorodeoxyglucose uptake in abdominal abscesses: a PET study. J Comput Assist Tomogr 1989; 13:829-31. 12. Lamberts SWJ, Reubi JC, Krenning EP. Somatostatin and the concept of peptide receptor scintigraphy in oncology. Semin Oncol 1994;21:1-5. 13. van Eyck CHJ, Bruining HA, Reubi JC, Bakker WH, Oei HY, Krenning EP, et al. Use of isotope-labeled somatostatin analogs for visualization of islet cell tumors. World J Surg 1993;17:444-7. 14. van Eijck CHJ, Lamberts SWJ, Lemaire LCJM, Jeekel H, Bosman FT, Reubi JC, et al. The use of somatostatin receptor scintigrapy in the differential diagnosis of pancreatic duct cancers and islet cell tumors. Ann Surg 1996;224:119-24. 15. Nakata B, Chung YS, Nishimura S, Nishihara T, Sakurai Y, Sawada T, et al. 18 F-Fluordeoxyglucose positron emission tomography and the prognosis of patients with pancreatic adenocarcinoma. Cancer 1997;79:695-9.
DISCUSSION Dr James A. Madura (Indianapolis, Ind). When I first received the manuscript, I thought it represented a fairly small and insignificant series of nonhomogeneous patients. However, a MEDLINE search failed to produce any other series, large or small, in the surgical literature. Most of the reports are in the radiologic and nuclear medicine literature, and very few are from the United
798 Rajput et al
States. I believe there is a single large series from St Louis, and the rest are from Japan and Europe. In the literature I was able to gather about 300 patients with pancreatic diseases, and the results of PET scanning were about the same as yours. There was a 90% to 93% accuracy rate, but a 5% false-positive rate for pancreatic malignancies. Many of the false-positive results were in less lethal neoplasms such as cystic neoplasms, intraductal malignancies, and neuroendocrine tumors. I know you had several of these lesions in your series. In addition, in a series from Osaka, Japan, in a group of patients about equal in number to your series, they report that a higher uptake of the FDG resulted in a poorer prognosis. Would you comment on that issue as it relates to your patients? Does the PET scan really add anything to your workup of these patients? From the literature, the PET scan seems to have the same tumor size limitations as the CT and MRI scans. The maximal size detected in the few series reporting it was about 10 mm. Many times smaller undetected lesions make these cases unresectable. If you find a tumor on PET scan, does it change your therapeutic approach? Does it eliminate the need to operate? We believe that a lesion occluding the common or pancreatic duct probably needs an operation anyway. Although your PET scan photographs are nice, they are not as anatomically elegant as a CT or MRI scan, which show surgeons critical structures like blood vessels. My last comment is about cost. At our hospital a PET scan costs the patient about $2100. To do all of the studies you did in these patients would cost about $8500. If you add fine-needle aspiration and pathology charges, this work-up is in excess of $10,000. Therefore tell us what you are doing now clinically to evaluate these patients. Is this simply a study you wish to add to the surgical literature, or are you really using it clinically in the patient with suspected pancreatic disease? Dr Rajput. You have a valid criticism that we have a small series. We hope to increase the number of patients in ongoing investigations. One of our goals, however, is to bring the knowledge of this technology to the surgical literature. In regards to your concern about false-positive studies, one must be aware that any acute inflammatory process may have an increased uptake of FDG. The use of FDG PET in cystic malignancies may result in false-negative studies. PET is based on the metabolic activity of cells. Thus, if there is not a significant cellular component to a cystic malignancy, the lesion may be missed. Neuroendocrine tumors are unique in that they have certain peptide receptors on their surfaces. Thus peptide receptor scintigraphy rather that FDG PET is being pursued for their diagnosis. The study you mention from Osaka, Japan, quantitatively analyzed the data in terms of standard uptake value
Surgery Octoaber 1998
of FDG and outcome. This study was based on the fact that a more metabolically active lesion will have a higher uptake value of FDG. They hypothesized that a more metabolically active lesion would be a more aggressive malignancy. The study found that patients with higher uptake values of FDG had a mean survival of 5 months and those with lower uptake values had a mean survival of 14 months. However, there were only 14 patients in the study. PET scanning is not in our standard armamentarium. Our usual work-up for pancreatic masses includes CT scan, ERCP, EUS, and laparoscopy. We are attempting to define the clinical role of PET scan. It may prove to be helpful in patients in whom the usual modalities do not provide a diagnosis or serious comorbid disease may make surgery a relative contraindication. PET may be most useful in the work-up of metastatic disease as evidenced by the last case I presented. If metastatic disease is identified, a palliative bypass, as opposed to a curative resection, would be offered and laparoscopy would be obviated. PET will not help assess for resectability; it is functional imaging as opposed to anatomic imaging. In regards to the size of the lesion that can be assessed, our PET scanner produces 6 mm sections. Thus a lesion would have to be 6 mm. There are preliminary data, however, that show very metabolically active lesions smaller than 6 mm can be detected. Cost is of concern to all of us in today’s health care climate. At the time our study was done, this was considered an experimental protocol and none of the patients were charged. Now, however, our institution charges about $2000 per study. We are in the process of negotiating with our “administocracy” to reduce the costs. Dr Henry Pitt (Milwaukee, Wis). I would like to compliment you for introducing this new, accurate modality. However, I share Dr Madura’s concerns about cost and the effort required to get good studies. In addition, I was surprised at the low accuracy of CT scanning in your patients. With newer spiral CT techniques the quality of the study is operator dependent. Were these CT scans done at your institution or elsewhere and what protocol was used? My second comment is that serum CA19-9 should be used when trying to differentiate pancreatic cancer from chronic pancreatitis. CA19-9 got a bad name when it was introduced with normal values below 37 to 40 U/mL. However, if 120 or 200 is used as the cutoff value, this test becomes very accurate especially in nonjaundiced patients. This test is relatively inexpensive and should probably be used more often in evaluating these patients. Dr Rajput. As we are a tertiary referral center, all the CT scans were not done at our institution with a rigid protocol. We appreciate your comments in regards to CA19-9. Because our study was retrospective, not all the patients had a CA19-9 level to help make the differentiation in suspected pancreatic disease.