Update of PET and PET/CT for hepatobiliary and pancreatic malignancies

Tumors of the hepatobiliary system and the pancreas

Metastases are the most common malignant tumors affecting the liver, occurring 20 times more often than primary carcinoma. The most common primaries producing liver metastases are colorectal, gastric, pancreatic, lung and breast carcinoma. Ninety percent of malignant primary liver tumors are tumors from epithelial origin: hepatocellular carcinoma (HCC) and cholangiocarcinoma. Gallbladder carcinoma is uncommon and is associated with cholelithiasis in 75% of the cases. These tumors are insidious, not suspected clinically, and often discovered at surgery or incidentally in the surgical specimen.

Pancreatic carcinomas usually arise from the pancreatic ducts and are the third most common malignant tumor of the gastrointestinal tract and the fifth leading cause of cancer-related mortality. Most tumors arise in the head of the pancreas, and patients present with bile duct obstruction, pain and jaundice. Carcinoma of the ampulla of Vater may be difficult to differentiate from those arising from the head of the pancreas. Acinar cell carcinomas comprise no more than 1%–2% of all pancreatic cancer, and the prognosis is as poor as for ductal cell carcinoma. Cystic neoplasms can arise in the pancreas and differentiation of benign from malignant is critical. Islet cell tumors and other endocrine tumors make up a small fraction of all pancreatic neoplasms and are most often located in the body and tail of the pancreas. They are usually slow-growing tumors and are associated with endocrine abnormalities.

Methods of diagnosis

The diagnostic issues include early detection of these tumors, differentiation of malignant tumors from benign tumors (lesion characterization), staging for resection that include lesion localization, evaluation of proximity to vessels, invasion of adjacent structures, metastasis to regional lymph nodes and distant sites, and assessment of therapeutic response.

Various imaging modalities are available to achieve these goals including ultrasound (US), computed tomography (CT), magnetic resonance imaging (MRI), and functional imaging using radiophar-maceuticals (Nuclear Medicine). Tomographic imaging for functional radioisotopic studies can be performed using single photon emission tomography technique (SPECT) if the radiopharmaceutical is a single photon emitter, and positron emission tomography technique (PET) if the radiopharmaceutical is a positron emitter. Some of these goals are better achieved with the high resolution of anatomical imaging techniques and others with molecular imaging using PET.

The rapid advances in imaging technologies are a challenge for both radiologists and clinicians who must integrate these technologies for optimal patient care and outcomes at minimal cost. Since the early 1990s, numerous technological improvements have occurred in the field of radiological imaging. These include: (1) multislice spiral computed tomography (CT) which permits fast acquisition of CT angiographic images and multiphase enhancement techniques; and (2) PET using ¹⁸F-fluorodeoxyglucose (FDG) as a radiopharmaceutical that provides the capability for imaging tumor metabolism.

Some of these tumors are associated with elevated serum levels of tumor markers that can be helpful for the diagnosis and surveillance of these patients, such as serum levels of carcinoembryonic antigen (CEA) for patients with colorectal carcinoma, alpha-fetoprotein (AFP) for screening patients at risk for hepatocellular carcinoma (HCC), and Ca 19-9 for surveillance of patients with pancreatic carcinoma, as well as various peptides for islet cell neoplasms.

Conventional imaging modalities

Transabdominal US is well established as a valuable screening technique that is inexpensive, portable, sensitive for evaluating bile duct dilatation, the pancreas, and can detect hepatic lesions as small as 1 cm. It can also provide guidance for biopsy and drainage procedures. Its limitations include poor sensitivity (50%) for detection of small hepatic lesions and regional lymphadenopathy compared to CT and MRI.

Endoscopic ultrasound is a promising new technique for the evaluation of the extrahepatic bile ducts and pancreatic ducts. It is sensitive in the detection cho-ledocholithiasis and pancreatic masses. However, it is highly operator-dependent, and requires sedation.

CT remains the first choice for a screening abdominal examination at many institutions. CT and MRI for hepatic imaging are based on the dual perfusion of the liver: 80% of the blood flow to normal liver parenchyma is derived from the portal vein, whereas nearly all of the blood flow to hepatic neoplasms is derived from the hepatic artery. Therefore some lesions are better seen at different times after intravenous contrast injection. Typically, hypervascular tumors and metastases (HCC, metastases of carcinoid carcinoma, islet cell tumor, malignant pheochromocytoma, renal cell carcinoma, sarcoma, melanoma, and breast carcinoma) may be best seen during the arterial phase of enhancement, or before contrast is administered; whereas hypovascular metastases (colorectal carcinoma, and most metastases of other primaries) are best seen during the portal venous phase of enhancement ¹.

After invasive arterial catheterization, contrast material can also be injected into the superior mesenteric artery, which increases the sensitivity for detection of small hepatic lesions, but decreases the specificity because of non-specific perfusion defects. This technique of CT portography is invasive and more expensive (approximately eight times that of CT), but has the potential to detect HCC less than 1 cm in diameter.

MRI imaging is certainly as sensitive as CT for detection of focal hepatic lesions, but it too is inferior to CT portography. A multitude of pulse sequences have been developed to characterize lesions. Gadolinium chelate contrast agents are used like the intravenous CT contrast agents, rapidly leaving the vascular space and reaching equilibrium throughout the extracellular fluid compartment after about 3 minutes ². MR cholangiopancreatography (MRCP) permits visualization of the biliary tree non-invasively without the administration of contrast agents ³. Using a heavily T2-weighted pulse sequence, solid organs and moving fluid have a low signal, whereas relatively stagnant fluid (such as bile) has a high signal intensity, resulting in the biliary tract appearing as a bright well-defined structure. Although MRCP does not provide the resolution of percutaneous transhepatic cholangiography (PTC) or endoscopic retrograde cholangiopancreatography (ERCP), it is able to clearly demonstrate intraluminal filling defects and luminal narrowing. MRCP provides invaluable information in both benign and malignant biliary tract disease.

Cholangiopancreatography via PTC or ERCP is an invasive technique but remains the procedure of choice for high-resolution assessment of the biliary tree anatomy. ERCP is performed by endoscopic cannulation of anatomic tracts and is therefore less invasive than PTC, which requires passage of a needle through the liver parenchyma. Contrast material is then injected directly into the biliary tree. Both techniques offer the advantage of allowing interventional procedures such as stent placement in the same setting as the imaging procedure. PTC demonstrates the intrahepatic ducts better than ERCP, which better depicts the extra-hepatic ducts.

Functional imaging with conventional radiophar-maceuticals can help to characterize lesions. ^99mTc-labeled red blood cells are a tracer of the blood pool and are highly accurate in differentiating cavernous hemangiomas from other lesions. ^99mTc-sulfur colloid accumulation in hepatic Kupffer cells allows characterization of focal nodular hyperplasia. ¹³¹I- or ¹²³I-metaiodobenzylguanidine (MIBG), which localizes through the norepinephrine reuptake mechanism into the catecholamine storage vesicles, can be used to image neuroendocrine tumors and their metastases. ¹¹¹In-octreotide is a somatostatin analog that accumulates in a variety of neuroendocrine tumors expressing somatostatin receptors, but may also help characterize other pathologic processes such as lymphoma, sarcoidosis, and autoimmune diseases. ⁶⁷Gallium, ²⁰¹thallium, ^99mTc-isonitriles (MIBI), and radiolabeled monoclonal antibodies are poor imaging agents for hepatic lesions due to physiological high liver background activity.

In summary, triple phase CT functions as the standard imaging modality for the detection and characterization of hepatic lesions, whereas US, MRI, MRCP, and ERCP/PTC provide complementary techniques for further characterization of lesions in specific circumstances. Conventional radiopharmaceuticals with or without SPECT may contribute as well but the development and proliferation of PET may yet provide unprecedented utility in the evaluation of these patients.

Positron emission tomography

Molecular imaging using positron imaging is unique in that positron emitters allow labeling of radiopharmaceuticals that closely mimic endogenous molecules, and there are continuing efforts to develop new biological tracers. ¹⁸Fluorodeoxyglucose (FDG) PET, allowing evaluation of glucose metabolism, is the radiopharmaceutical most widely used with the PET technology and that has been approved by the Health Care Financing Administration (HCFA) for reimbursement by Medicare in the evaluation of patients with various body tumors.

Radiopharmaceuticals for molecular imaging with PET

¹⁸F-fluorodeoxyglucose (FDG)

Because of its ability to demonstrate glucose metabolism and its longer, more practical half-life (110 min) as compared to other short-lived positron emitters, FDG is the most frequently used tracer for clinical applications. Multiple indications for molecular imaging using FDG are now well accepted in the fields of neurology, cardiology, and oncology ⁴.

Although variations in uptake are known to exist among tumor types, elevated uptake of FDG has been demonstrated in most primary malignant tumors ⁵. This is due to the expression of increased numbers of glucose transporter proteins and increased intracellular enzyme levels of hexokinase and phosphofructokinase, among others, which promote glycolysis ⁶,⁷,⁸,⁹. FDG PET imaging can be used to exploit the metabolic differences between benign and malignant cells for imaging purposes ¹⁰,¹¹.

Improvements in the distribution of FDG by commercial companies and the widespread oncologic applications, including differentiation of benign from malignant lesions, staging malignant lesions, detection of malignant recurrence, and monitoring therapy, have contributed to the establishment of the PET technology in many medical centers in the United States, Europe and progressively throughout the world. At present, the major indications for FDG PET imaging is staging and restaging of malignant tumors leading to detection of unsuspected metastases in 25%–30% of the patients and major changes in therapy. Because of the limitations of FDG related to variations of physiological uptake and overlap of uptake between inflammatory and malignant lesions, other PET radiopharmaceuticals have been investigated for clinical use.

Potential new PET tracers for clinical use

Besides evaluation of glucose metabolism with FDG, PET can assess various other biologic parameters such as perfusion, metabolism of other compounds, hypoxia and receptor expression. Some of these radiopharmaceuticals are labeled with positron emitters that have a short half-life, such as ¹⁵O (T1/2 = 2 min), ¹³N (T1/2 = 10 min), and ⁿC (T1/2 = 20 min). The short half-life of these radioisotopes prevents any timely distribution of the radiopharmaceuticals labeled with them and therefore their use is restricted to institutions that have a cyclotron and associated laboratories and personnel on-site.

Tracer of bone metabolism

¹⁸F-fluoride was first described as a skeletal imaging agent in the 1960s but then was replaced by the ^99mTc-labelled diphosphonate radiopharmaceuticals ¹². With the widespread applications of FDG PET in oncology, PET imaging systems are becoming more widely available, and there is a renewed interest in ¹⁸F-fluoride. Although the mechanism of uptake for ¹⁸F-fluoride is similar to that for other bone-imaging radiopharmaceuticals ¹³, the spatial resolution of the PET technology is superior to that of both planar and SPECT imaging using the ^99mTc-radiopharma-ceuticals. Because of the better spatial resolution and routine acquisition of tomographic images, ¹⁸F-fluoride PET imaging offers potential advantages over conventional bone scintigraphy in detecting metastases.

Tracers of DNA synthesis

The rate of DNA synthesis can be assessed using ¹¹C-thymidine or ¹⁸F-fluorothymidine (FLT) ¹⁴. Thymidine is a DNA precursor and allows direct assessment of tumor proliferation. In the early nineties, Higashi et al.¹⁵ demonstrated that uptake in vitro correlates with the tumor proliferative rate. Compared to FDG, FLT uptake is lower but with a significant linear correlation ¹⁶. The high physiologic uptake in the liver and bone marrow is a limitation for detection of metastases in these organs. A report of 17 patients with colorectal cancer concludes that the high physiologic liver background limits detection of hepatic metastases, so it is unlikely that FLT will play a role in the evaluation of patients with colorectal carcinoma ¹⁷.

Tracers of amino-acid transport and protein metabolism

Assessment of amino acid transport and protein metabolism is possible with ¹¹C-methionine and ¹⁸F-tyrosine. ¹¹ C-methionine is used more commonly in Europe than in the United States and has some advantages over FDG for evaluation of brain tumors, for example, because of the low physiologic background uptake in the cortex. Little data is available for evaluation of gastrointestinal tumors with tracers of amino-acid transport and protein metabolism.

Tracers of membrane lipid synthesis

¹¹C-acetate, ¹¹C-choline and ¹⁸F-fluorocholine ¹⁸,¹⁹ are markers of membrane lipid synthesis and are promising PET radiopharmaceuticals for evaluation of some malignancies for which FDG has limitations such as prostate carcinoma. ¹¹C-acetate is also a promising tracer for imaging HCC as described in the section below. Tumor cells incorporate acetate preferentially into lipids rather than into amino acids or CO₂ as a necessary condition for cell proliferation ²⁰.

Tracers for receptor imaging

In the field of neurology, neuroreceptor imaging has been investigated extensively. Various PET radio-pharmaceuticals have been developed for dopamine-, serotonin- and benzodiazepine-receptor imaging. Some of these tracers for receptor imaging may have a role for detection of neuroendocrine tumors.

Instrumentation for molecular imaging with PET

The clinical utility of FDG imaging was first established using dedicated PET tomographs equipped with multiple rings of bismuth germanate oxide (BGO) detectors, but a spectrum of equipment is now available for positron imaging including gamma camera-based PET at the low end of the spectrum and dedicated PET tomographs equipped with newer detector materials. The advantages and limitations of each of these systems are beyond the scope of this review ²¹.

Anatomical and functional imaging are complementary

Although numerous studies have shown that the sensitivity and specificity of FDG imaging is superior to that of CT in many clinical settings, the inability of FDG imaging to provide accurate anatomical localization remains a significant impairment in maximizing its clinical utility. Because FDG is a tracer of glucose metabolism, its distribution is not limited to malignant tissue. To avoid misinterpretations, the interpreter must be familiar with the normal pattern and physiologic variations of FDG distribution and with clinical data relevant to the patient ²²,²³. It is also important to standardize the environment of the patient during the uptake period so as to limit physiologic variations of FDG uptake, (e.g., in activated muscular tissue). The problem of precise anatomical localization of the foci of abnormal uptake and differentiation of physiologic from pathologic foci of uptake is compounded by the lower resolution and increased noise in the images of many of the systems at the low end of the spectrum and the hybrid gamma camera-based systems.

The limitations of anatomical imaging with CT and MRI are related to size criteria for differentiation of benign from malignant lymph nodes, difficulty differentiating post-therapy changes from tumor recurrence, and difficulty differentiating non-opacified loops of bowel from metastases in the abdomen and pelvis.

Close correlation of FDG studies with conventional CT scans helps to minimize these difficulties. In practice for the past ten years, interpretation has been accomplished by visually comparing corresponding FDG and CT images. The interpreting physician visually integrates the two image sets in order to precisely locate a region of increased FDG uptake on the CT scan. To aid in image interpretation, computer software has been developed to co-register the FDG PET emission scans with the high-resolution anatomical maps provided by CT ²⁴. These methods offer acceptable fusion images for the brain that is surrounded by a rigid structure, the skull. For the body, co-registration of two image sets often obtained at different points in time is technically more difficult. Identical positioning of the patient on the imaging table is critical. Shifting internal organ movement and peristalsis compound the problem.

Another approach that has gained wider acceptance recently is the hardware approach to image fusion using multimodality imaging with an integrated PET/CT imaging system ²⁵.

Integrated PET/CT systems

The recent development of integrated PET/CT systems provides CT and FDG PET images obtained in a single imaging setting, allowing optimal co-registration of images. With these integrated systems, a diagnostic CT scan and a PET scan can be acquired sequentially with the patient lying on the imaging table and simply being translated between the two systems. Accurate calibration of the position of the imaging table and the use of common parameters in data acquisition and image reconstruction permit the fusion of images of anatomy and metabolism from the same region of the body that are registered in space and only slightly offset in time. The fusion images provided by these systems allow the most accurate interpretation of both CT and FDG PET studies. Because of the high photon flux of X-rays, CT attenuation maps from these integrated PET/CT systems also allow for optimal attenuation correction of the PET images.

Clinical impact of integrated PET/CT images

From the diagnostic point of view, the CT obtained for attenuation maps can also be used for precise localization of the foci of uptake with the help of the fusion of anatomical and molecular images. Published data regarding the incremental value of integrated PET/CT images compared to PET alone, or PET correlated with a CT obtained at a different time, are limited but conclude the following: (1) Improvement of lesion detection on both CT and FDG PET images; (2) Improvement of the localization of foci of FDG uptake resulting in better differentiation of physiologic from pathologic uptake; and (3) Precise localization of the malignant foci, for example in the skeleton versus soft tissue, or liver versus adjacent bowel or node. PET/CT fusion images affect the clinical management by guiding further procedures, excluding the need of further procedures, and changing both inter- and intra-modality therapy ²⁶. For example, precise localization of metastatic lymph nodes could result in a less invasive and more efficient surgical procedure. PET/CT fusion images have the potential to provide important information to guide the biopsy of a mass to more metabolically active regions of the tumor and to provide better maps, than CT alone, to modulate field and dose of radiation therapy ²⁷.

Concurrent PET/CT imaging with an integrated system may be especially important in the abdomen and pelvis ²⁸. PET images alone may be difficult to interpret owing to the absence of anatomical landmarks (other than the kidneys and bladder), and to the presence of non-specific uptake in the stomach, small bowel and colon, and urinary excretion of FDG. If possible, images of the abdomen and pelvis should be obtained with the arms elevated to avoid artifacts due to motion and to beam hardening artifacts on the CT transmission images. Concurrent PET/CT imaging is helpful for differentiating focal retention of urine in the ureter, for example, versus an FDG-avid lymph node. A review of PET/CT for gastrointestinal tumors has been published recently ²⁹. Some data comparing PET/CT versus PET alone are available for patients with colorectal carcinoma and are described in the section below.

FDG PET for evaluation of colorectal carcinoma

About 14,000 patients per year with colorectal carcinoma present with isolated liver metastases as their first recurrence, and about 20% of these patients die with metastases exclusively to the liver ³⁰. Hepatic resection is the only curative therapy in these patients, but it is associated with a mortality of 2%–7% and significant morbidity ³¹. Early detection and prompt treatment of recurrences may lead to a cure in up to 25% of patients. However, the size and number of hepatic metastases and the presence of extra-hepatic disease adversely affect the prognosis. The poor prognosis of extra-hepatic metastases is believed to be a contraindication to hepatic resection ³². Therefore, accurate non-invasive detection of inoperable disease with imaging modalities plays a pivotal role in selecting patients who would benefit from hepatic surgery.

A meta-analysis of 11 clinical reports encompassing 577 patients determined that the sensitivity and specificity of FDG PET for detecting recurrent colorectal cancer were 97% and 76% respectively ³³. A comprehensive review of the PET literature (2,244 patients studies) has reported a weighted average for FDG PET sensitivity and specificity of 94% and 87% respectively, compared to 79% and 73% for CT ³⁴. A meta-analysis performed to compare non-invasive imaging methods (US, CT, MRI and FDG PET) for the detection of hepatic metastases from colorectal, gastric and esophageal cancers demonstrated that FDG PET had a higher sensitivity (90%) than MRI (76%), CT (72%) and US (55%), with an equivalent specificity of 85% ³⁵.

In the meta-analysis of the literature, FDG PET imaging changed the management in 29% (102/349) of patients ³³. The comprehensive review of the PET literature has reported a weighted average change of management related to FDG PET findings in 32% of 915 patients ³⁴. Although survival is not an end point for a diagnostic test, Strasberg et al.³⁶ have estimated the survival of patients who underwent FDG PET imaging in their preoperative evaluation for resection of hepatic metastases. The Kaplan-Meier test estimate of the overall survival at three years was 77% and the lower confidence limit was 60%. These percentages are higher than those in previously published series (without PET) that ranged from 30% to 64%. In the patients undergoing FDG PET imaging prior to hepatic resection, the three-year disease-free survival rate was 40%, again higher than that usually reported. Figure 1 illustrates detection of extrahepatic disease precisely localized on PET/CT in a patient with a history of colorectal carcinoma and recently diagnosed with hepatic metastases.

Figure 1. .

A 57-year-old male with a history of colon carcinoma and recently diagnosed hepatic metastases underwent a PET/CT study for restaging. FDG PET maximum intensity projection (MIP) image demonstrated: (1) Two foci of FDG uptake in the liver consistent with the known hepatic metastases; (2) Two foci of FDG uptake in the lower chest to the left of the midline (arrows); (3) A focus of FDG uptake in the pelvis (arrow). In addition, there is an artifact over the arms of the patient because they were touching the gantry and there is physiological FDG uptake in the ascending colon, kidneys and bladder. Transaxial views through the focus of FDG uptake in the pelvis demonstrated that the focus of uptake corresponded to a presacral lymph node indicating a metastasis (arrows). Similar images demonstrated that the foci of FDG uptake in the chest corresponded to pleural-based soft tissue densities indicating metastases in the chest (not shown).

FDG PET for monitoring therapy of hepatic metastases

Hepatic metastases can be treated with systemic chemotherapy or regional therapy to the liver. A variety of procedures to administer regional therapy to hepatic metastases has been investigated including chemotherapy administered through the hepatic artery using infusion pumps, selective chemoembolization, radio-frequency ablation, cryoablation, alcohol ablation and radiolabeled ⁹⁰Y-microspheres administered via the hepatic artery ³⁷,³⁸,³⁹,⁴⁰. There are preliminary reports suggesting that the response to chemotherapy in patients with hepatic metastases can be predicted with PET. Responders may be discriminated from non-responders after four to five weeks of chemotherapy with fluorouracil by measuring FDG uptake before and during therapy ⁴¹. Regional therapy to the liver by chemoembolization can also be monitored with FDG PET imaging ⁴²,⁴³. FDG uptake decreases in responding lesions, and the presence of residual uptake in some lesions can help in guiding further regional therapy. Langenhoff et al.⁴⁴ have prospectively monitored 23 patients with liver metastases following radiofrequency ablation and cryoablation. Three weeks after therapy, 51/56 metastases became FDG negative, and there was no recurrence during 16 months of follow-up; whereas among the 5/56 lesions with persistent FDG uptake, 4/5 recurred. Data in smaller series of patients support their findings ⁴⁵,⁴⁶. Wong et al.⁴⁷ have compared FDG PET imaging, CT or MRI and serum levels of CEA to monitor the therapeutic response of hepatic metastases to ⁹⁰Y-glass microspheres. They found a significant difference between the FDG PET changes and the changes on CT or MRI; the changes in FDG uptake correlated better with the changes in serum levels of CEA.

In summary, preliminary data suggest that FDG PET imaging may be able to effectively monitor the efficacy of regional therapy to hepatic metastases, but these data need to be confirmed in larger series of patients.

PET/CT versus PET for evaluation of colorectal carcinoma

A study of 204 patients (34 with gastrointestinal tumors), performed at Rambam Medical Center ⁴⁸ using an integrated PET/CT system, concluded that the diagnostic accuracy of PET is improved in approximately 50% of patients. In that study, PET/CT fusion images improved characterization of equivocal lesions as definitely benign in 10% of sites and definitely malignant in 5% of sites. Fusion images precisely defined the anatomic location of malignant FDG uptake in 6% and led to retrospective lesion detection on PET or CT in 8%. The results of fusion PET/CT images had an impact on the management of 14% (28/204) of patients, 7/28 patients with a change of management had colorectal cancer, representing 20% (7/34) of patients with gastrointestinal tumors. The changes in management in the 7 patients with colorectal cancer included guiding colonoscopy and biopsy for a local recurrence (N = 2), guiding biopsy to a metastatic supraclavicular lymph node (N = 1), guiding surgery to localized metastatic lymph nodes (N = 3) and referral to chemotherapy (N = 2). Similar conclusions were found in a study of 173 patients performed at Vanderbilt University, 24 of whom had colorectal carcinoma ⁴⁹. A study of 45 patients with colorectal cancer referred for FDG PET imaging using an integrated PET/CT system concluded that PET/CT imaging increases the accuracy and certainty of locating lesions. In their study, the frequency of equivocal and probable lesion characterization was reduced by 50% with PET/CT compared to PET alone, the number of definite locations was increased by 25%, and the overall correct staging increased from 78% to 89% ⁵⁰. At the time of this writing, most institutions acquire CT transmission images without intravenous contrast to permit optimal attenuation correction, but CT images without intravenous contrast do not allow visualization of many hepatic metastases. Therefore, although hepatic metastases are commonly seen as FDG-avid on the PET images, no corresponding lesions are seen on the non-contrasted CT transmission images. A standard of care CT with intravenous and oral contrast needs to be performed if surgery is contemplated. Concurrent PET/CT fusion images have the potential to provide better maps than CT alone to modulate field and dose of radiation therapy including in patients with colorectal carcinoma ²⁷.

FDG PET for the evaluation of hepatocellular carcinoma

Differentiated hepatocytes normally have relatively high glucose-6-phosphatase activity. Although experimental studies have shown that glycogenesis decreases and glycolysis increases during carcinogenesis, the accumulation of FDG in HCC is variable due to varying degrees of activity of the enzyme glucose-6-phosphatase in these tumors preventing intracellular accumulation of FDG ⁵¹,⁵²,⁵³,⁵⁴,⁵⁵. FDG PET detects only 50%–70% of HCC, but has a sensitivity greater than 90% for all other primary (cholangiocarcinoma and sarcoma) and metastatic tumors to the liver ⁵⁶,⁵⁷. All benign tumors, including focal nodular hyperplasia, adenoma, and regenerating nodules, demonstrate FDG uptake at the same level as normal liver, except for rare granulomatous abscesses. In a report involving one of the largest series of FDG PET imaging for HCC (N=91), the sensitivity of FDG PET for detection of HCC was 64% with a clinically significant impact in 28% of patients overall ⁵⁸. A correlation was found between the degree of FDG uptake and the grade of malignancy ⁵⁵,⁵⁶. Therefore, FDG imaging may have a prognostic significance in the evaluation of patients with HCC. HCC that accumulate FDG are associated with markedly elevated alpha-fetoprotein levels ⁵⁹,⁶⁰. However, FDG PET has limited value for the differential diagnosis of focal liver lesions in patients with chronic hepatitis C virus infection because of the low sensitivity for detection of HCC and the high prevalence of this tumor in that population of patients ⁶¹.

In patients with HCC that accumulate FDG, PET imaging is able to detect unsuspected regional and distant metastases, as with other tumors. In a series of 23 patients with HCC who underwent FDG PET scanning in an attempt to identify extrahepatic metastases, 13 of the 23 patients (57%) had increased uptake in the primary tumor and 4 of the 13 had extrahepatic metastases demonstrated by FDG PET images ⁶². Figure 2 illustrates detection of extrahepatic disease precisely localized on PET/CT in a patient just diagnosed with HCC.

Figure 2. .

A 77-year-old male with a history of cirrhosis and elevated AFP presented with a large hepatic mass. FDG PET maximum intensity projection (MIP) image demonstrated: (1) Heterogenous FDG uptake in the large hepatic mass consistent with hepatocellular carcinoma; (2) A focus of FDG uptake at the level of the renal hila on the midline (arrow). In addition there is physiological FDG uptake in the ascending colon and kidneys, ureters and bladder. Transaxial views through the focus of FDG uptake in the abdomen demonstrated that the focus of uptake corresponded to a retroperitoneal lymph node indicating a metastasis (arrows). Systematic review of the transaxial images demonstrated another abnormal focus of FDG uptake in the chest at the right lung base corresponding to a pulmonary nodule indicating a lung metastasis (arrow).

Because the majority of patients with HCC have advanced-stage tumors and/or underlying cirrhosis with impaired hepatic reserve, surgical resection is often not possible. Therefore, other treatment strategies have been developed, including hepatic arterial chemoembolization, systemic chemotherapy, surgical cryoablation, ethanol ablation, radiofrequency ablation, and, in selected cases, liver transplantation. In patients treated with hepatic arterial chemoembolization, FDG PET is more accurate than lipiodol retention on CT in predicting the presence of residual viable tumor. The presence of residual uptake in some lesions can help in guiding further regional therapy ⁶³,⁶⁴,⁶⁵.

¹¹C-acetate for hepatocellular carcinoma

PET with ¹¹C-acetate has been shown to be useful in detection of HCC. Possible biochemical pathways that lead to accumulation of ¹¹C-acetate in tumors include: (1) entry into the Krebs cycle from acetyl coenzyme A (acetyl CoA) or as an intermediate metabolite; (2) esterification to form acetyl CoA as a major precursor in β-oxidation for fatty acid synthesis; (3) combining with glycine in heme synthesis; and (4) through citrate for cholesterol synthesis. Among these possible metabolic pathways, participation in free fatty acid synthesis is believed to be the dominant method of incorporation into tumors. Ho et al.⁶⁶ reported a study of 57 patients with various hepato-biliary tumors who underwent both FDG and ¹¹C-acetate PET imaging. For the 23 patients with HCC, the sensitivities of FDG and ¹¹C-acetate imaging were 47% and 87% respectively, with a combined sensitivity of 100%. Well-differentiated tumors tended to be ¹¹C-acetate-avid whereas poorly differentiated tumors tended to be FDG-avid. Other malignant tumors were FDG-avid but not ¹¹C-acetateavid. Benign tumors, such as adenoma and hemangioma were not ¹¹C-acetate-avid except for mild uptake in FNH.

FDG PET for the evaluation of cholangiocarcinoma and gallbladder carcinoma

There is preliminary evidence that FDG PET imaging may be useful in the diagnosis and management of small cholangiocarcinomas in patients with sclerosing cholangitis ⁶⁷. Anderson et al.⁶⁸ reviewed 36 consecutive patients who underwent FDG PET for suspected cholangiocarcinoma. Patients were divided into group 1: nodular type (mass >1 cm) (N = 22) and group 2: infiltrating type (N = 14). The sensitivity for nodular morphology was 85%, but only 18% for infiltrating morphology. Sensitivity for metastases was 65% with three false negative for carcinomatosis and one false positive result in a patient with primary sclerosing cholangitis who had acute cholangitis. Seven of 12 patients (58%) had FDG uptake along the tract of a biliary stent. FDG PET led to a change in surgical management in 30% of patients (11/36) because of detection of unsuspected metastases.

Unsuspected gallbladder carcinoma is discovered incidentally in 1% of routine cholecystectomies. At present, the majority of cholecystectomies are performed laparoscopically, and occult gallbladder carcinoma found after laparoscopic cholecystectomy has been associated with reports of gallbladder carcinoma seeding of laparoscopic trocar sites ⁶⁹,⁷⁰. Increased FDG uptake has been demonstrated in gallbladder carcinoma ⁷¹ and has been helpful in identifying recurrence in the area of the incision when CT could not differentiate scar tissue from malignant recurrence ⁷². In a study reviewing 14 patients with gallbladder carcinoma, the sensitivity for detection of residual gallbladder carcinoma was 78%. Sensitivity for extrahepatic metastases was 50% in 8 patients; 6 of these had carcinomatosis ⁶⁸.

FDG PET for the evaluation of pancreatic carcinoma

The difficulty in correctly determining a preoperative diagnosis of pancreatic carcinoma is associated with two types of adverse outcomes. First, less aggressive surgeons may abort attempted resection due to a lack of tissue diagnosis. This is borne out by the significant rate of'reoperative' pancreaticoduodenectomy performed at major referral centers ⁷³,⁷⁴,⁷⁵. In a recent review of the MD Anderson Cancer Center involving 29 patients undergoing successful pancreaticoduodenectomy after failure to resect at the time of initial laparotomy, 31% did not undergo resection at the time of the initial procedure because of the lack of tissue confirmation of malignancy ⁷⁵. A second type of adverse outcome generated by failure to obtain a preoperative diagnosis occurs when more aggressive surgeons inadvertently resect benign disease. This is particularly notable in those patients who present with suspected malignancy without an associated mass on CT scan, occurring in up to 55% of patients ⁷⁶.

In order to avoid these adverse outcomes, metabolic imaging with FDG PET may be used to improve the accuracy of the preoperative diagnosis of pancreatic carcinoma. The summary of the literature published in 2001 reported an average sensitivity and specificity of 94% and 90% respectively ²⁶. Although the sensitivity of CT imaging rises with increasing lesion size, the sensitivity of FDG PET does not appear to be affected by size, ⁷⁷. Figure 3 illustrates detection of a pancreatic carcinoma with PET/CT in a patient with no definite mass on CT and a non-diagnostic fine needle biopsy. The degree of FDG uptake is reported to have a prognostic value. Nakata et al.⁷⁸ noted a correlation between SUV and survival in 14 patients with pancreatic adenocarcinoma. Patients with an SUV 4 3.0 had a mean survival of 5 months compared to 14 months in those with an SUV <3.0. Zimny et al.⁷⁹ performed a multivariate analysis on 52 patients with pancreatic cancer to determine the prognostic value of FDG PET. The median survival of 26 patients with an SUV >6.1 was 5 months compared to 9 months for 26 patients with an SUV >6.1. The multivariate analysis revealed that SUV and Ca 19-9 were independent predictors of prognosis.

Figure 3. .

A 7 8-year-old female presented with abdominal pain and jaundice. An ERCP was performed and stent placed in the pancreatic duct. No definite mass was seen on CT and a fine needle biopsy was non-diagnostic. FDG PET maximum intensity projection (MIP) image demonstrated a focus of FDG uptake in the mid-abdomen on the midline (arrow). In addition there is physiological FDG uptake in the myocardium, colon, kidneys and bladder. Transaxial views through the focus of FDG uptake in the mid-abdomen demonstrated that the focus of uptake corresponded to the head of the pancreas indicating pancreatic cancer (arrows), which was proven at surgery.

As for other tumors, FDG imaging has not been superior to helical CT for regional N staging, but is more accurate than CT for M staging. In the study by Delbeke et al.⁸⁰, metastases were diagnosed both on CT and PET in 10/21 patients with stage IV disease, but PET demonstrated hepatic metastases not identified or equivocal on CT and/or distant metastases unsuspected clinically in 7 additional patients. In four patients, neither CT nor PET imaging showed evidence of metastases, but surgical exploration revealed carcinomatosis in three and a small liver metastasis in one patient. The addition of FDG PET imaging to CT altered the surgical management in 41% of the patients: 27% by detecting CT-occult pancreatic carcinoma and 14% by identifying unsuspected distant metastases, or by clarifying the benign nature of lesions equivocal on CT. Another report of 19 patients concludes that FDG PET added important additional information to clinicians in 50% of the patients, resulting in a change of therapeutic procedure ⁸¹. This includes patients with elevated tumor markers levels, and no findings on anatomical imaging.

Preliminary data suggest that FDG PET imaging is useful for the assessment of tumor response to neo-adjuvant therapy and the evaluation of possible recurrent disease following resection ⁷⁷.

Therefore, FDG PET may be particularly useful when CT identifies an indistinct abnormality in the bed of the resected pancreas that is difficult to differentiate from postoperative or postradiation fibrosis, for the evaluation of new hepatic lesions that may be too small to biopsy, and in patients with rising tumor markers levels and a negative conventional work-up.

FDG PET for evaluation of islet cells and other endocrine tumors of the pancreas

Most neuroendocrine tumors, including carcinoid, paraganglioma, and islet cell tumors, express somato-statin receptors (SSR) and can, therefore, be imaged effectively with somatostatin analogs such as ¹¹¹In-octreotide. This modality has been reported to be more sensitive than CT for defining the extent of metastatic disease, especially in extrahepatic and extra-abdominal sites. However, there may be significant heterogeneity in regard to SSR expression, even in the same patient in adjacent sites, probably related to dedifferentiation of the tumor. Absence of SSRpositivity is reported to be a poor prognostic sign, but virtually all of these SSR-negative neuroendocrine tumors will accumulate FDG and can therefore be imaged with PET ⁸². More differentiated SSR-positive tumors do not reliably accumulate significant FDG and may, therefore, be false negative with FDG PET imaging ⁸³. There is controversy in the literature about the sensitivity of FDG imaging for detection of carcinoid tumors ⁸⁴,⁸⁵, but at least in some reports, ¹¹¹In-octreotide scintigraphy is more sensitive than FDG PET imaging; FDG positive/octreotide-negative tumors tend to be less differentiated and may have a less favorable prognosis.

Other positron emitting tracers seem to be more promising. A serotonin precursor 5-hydroxytryptophan (5-HTP) labeled with ¹¹C has shown an increased uptake in carcinoid tumors. This uptake seems to be selective and some clinical evidence has demonstrated that it allows the detection of more lesions with PET than with CT or octreotide scintigraphy. Another radiopharmaceutical in development for PET is ¹¹C L-DOPA and ¹⁸F-DOPA, which seems to be useful in visualizing gastrointestinal neuroendocrine tumors ⁸⁶,⁸⁷. A study of 17 patients with 92 carcinoid tumors comparing FDG PET, ¹⁸F-DOPA PET and somatostatin-receptor scintigraphy demonstrated the following sensitivities: 29% for FDG PET, 60% for ¹⁸F-DOPA, 57% for somatostatin-receptor scintigraphy and 73% for morphologic procedures ⁸⁸.

An octreotide derivative can be labeled with ⁶⁴Cu (half-life, 12.7 h; beta+, 0.653 MeV (17.4%); beta-, 0.579 MeV (39%)) and has shown potential as a radiopharmaceutical for PET imaging and radiotherapy. A pilot study in humans has demonstrated that ⁶⁴Cu-TETA-octreotide (where TETA is 1,4,8,11 -tetraazacyclotetradecane-N,N^′,N^′′,N^′′′-tetra-acetic acid) and PET can be used to detect somatos-tatin-receptor-positive tumors in humans. The high rate of lesion detection, sensitivity, and favorable dosimetry and pharmacokinetics of ⁶⁴Cu-TETA-OC indicate that it may be a promising radiopharmaceutical for PET imaging of patients with neuroendocrine tumors ⁸⁹.

Limitations of FDG imaging

Scintigraphic tumor detectability depends on both the size of the lesion and the degree of uptake, as well as surrounding background uptake and intrinsic resolution of the imaging system. Small lesions may yield false negative results which can be due to partial volume averaging, leading to underestimation of the uptake in small lesions (of less than twice the resolution of the imaging system, for example small ampullary carcinoma, cholangiocarcinoma of the infiltrating type and military carcinomatosis) or in necrotic lesions with a thin viable rim, falsely classifying these lesions as benign instead of malignant. The sensitivity of FDG PET for detection of mucinous adenocarcinoma is lower than for non-mucinous adenocarcinoma (41%–58% versus 92%), probably because of the relative hypocellularity of these tumors ⁹⁰. Other false negatives include differentiated neuroendocrine tumors and HCC. The high incidence of glucose intolerance and diabetes exhibited by patients with pancreatic pathology represents a potential limitation of this modality in the diagnosis of pancreatic cancer, because elevated serum glucose levels result in decreased FDG uptake in tumors due to competitive inhibition. Low SUV values and false negative FDG PET scans have been noted in hyperglycemic patients.

In view of the known high uptake of FDG by activated macrophages, neutrophils, fibroblasts and granulation tissue, it is not surprising that inflammatory tissue demonstrates FDG activity. Mild to moderate FDG activity seen early after radiation therapy, along recent incisions, infected incisions, biopsy sites, drainage tubing and catheters, as well as colostomy sites, can lead to errors in interpretation if the history is not known. Some inflammatory lesions, especially granu-lomatous ones, may be markedly FDG-avid and can be mistaken for malignancies; this includes inflammatory bowel disease, abscesses, acute cholangitis, acute cholecystitis, acute pancreatitis, and chronic active pancreatitis with or without abscess formation. False positive studies are frequent in patients with elevated C-reactive protein and/or acute pancreatitis with a specificity as low as 50% ⁹¹,⁹². Therefore, screening for C-reactive protein has been recommended. Other lesions with false positive images that have been reported are serous cystadenoma and retroperitoneal fibrosis.

FDG uptake normally present in the gastrointestinal tract can occasionally be difficult to differentiate from a malignant lesion. Incidental colonic FDG uptake in 27 patients without colorectal carcinoma has been correlated with colonoscopic and/or histolopathologic findings ⁹³. Diffuse uptake in eight patients was normal and associated with a normal colonoscopy. Segmental uptake was due to colitis in 5/6 patients. Focal uptake in seven patients was associated with benign adenomas. The clinical history, physical examination, pattern of uptake and correlation with anatomy as seen on the CT images are more helpful in avoiding false positive interpretations than semiquantitative evaluation by SUV.

Summary

Evaluation of patients with known or suspected recurrent colorectal carcinoma is now an accepted indication for FDG PET imaging. FDG PET does not replace imaging modalities such as CT for preoperative anatomic evaluation, but is indicated as the initial test for diagnosis and staging of recurrence, and for preoperative staging (N and M) of known recurrence that is considered to be resectable. FDG PET imaging is valuable for differentiation of post-treatment changes from recurrent tumor, differentiation of benign from malignant lesions (indeterminate lymph nodes, hepatic and pulmonary lesions) and evaluation of patients with rising tumor markers in the absence of a known source. FDG PET has an impact on the treatment of 25%–30% of patients. Addition of FDG PET to the evaluation of these patients reduces overall treatment costs by accurately identifying patients who will and will not benefit from surgical procedures.

FDG PET imaging seems promising for monitoring patient response to therapy, including regional therapy to the liver, but larger studies are necessary.

FDG PET imaging appears helpful to differentiate malignant from benign hepatic lesions, with the exception of false negative HCC, false negative infiltrating cholangiocarcinoma, and false positive inflammatory lesions. It is not helpful to identify HCC in patients with cirrhosis and regenerating nodules. In patients with hepatic primary malignant tumors trapping FDG, FDG PET imaging does identify unexpected distant metastases (although military carcinomatosis is often false negative) and can help in monitoring therapy.

FDG PET imaging is especially helpful for the pre-operative diagnosis of pancreatic carcinoma in patients with suspected pancreatic cancer in whom CT fails to identify a discrete tumor mass or in whom FNAs are non-diagnostic. By providing preoperative documentation of pancreatic malignancy in these patients, laparotomy may be undertaken with a curative intent, and the risk of aborting resection due to diagnostic uncertainty is minimized. FDG PET imaging is also useful for M staging and restaging by detecting CT-occult metastatic disease, and allowing non-therapeutic resection to be avoided altogether in this group of patients. As is true with other neoplasms, FDG PET can differentiate post-therapy changes from recurrence and holds promise for monitoring neo-adjuvant chemoradiation therapy.

FDG PET imaging is complementary to morphological imaging with CT; therefore, integrated PET/CT imaging provides optimal images for interpretation. The diagnostic implications of integrated PET/CT imaging include improved detection of lesions on both the CT and FDG PET images, better differentiation of physiologic from pathologic foci of metabolism, and better localization of the pathologic foci. This new powerful technology provides more accurate interpretation of both CT and FDG PET images and therefore more optimal patient care. PET/CT fusion images affect the clinical management by guiding further procedures (biopsy, surgery, radiation therapy), excluding the need for additional procedures, and changing both inter- and intra-modality therapy.