The role of nuclear medicine in diagnosis and management of hepatopancreatobiliary diseases

Nuclear medicine uses radioactive pharmaceuticals, or radiopharmaceuticals , for diagnostic imaging and internal radiotherapy of a variety of diseases. This chapter discusses clinical applications of diagnostic nuclear medicine imaging for the care of patients with hepatic, pancreatic, and biliary (hepatopancreatobiliary [HPB]) diseases. Radioembolization of liver tumors with radiolabeled microspheres is discussed more extensively in Chapter 94B .

In general, the role of diagnostic nuclear medicine imaging (NMI), or scintigraphy, including positron emission tomography (PET), is to provide HPB clinicians with a noninvasive method to aid in detecting and localizing certain types of HPB disease and to evaluate HPB organ function and the effects of treatment. In general, NMI can be considered a clinical assay of cellular biology in the tissues of patients; the in vivo tissue accumulation, or uptake, of most radiopharmaceuticals depends on the biomolecular composition of living cells in body tissues, as well as tissue perfusion.

The diagnostic accuracy of scintigraphy varies according to the specific scintigraphic study (including the specific radiopharmaceutical used and how it is assayed) and the specific disease or condition being studied. The HPB specialist must integrate diagnostic data from any scintigraphic study of a particular patient with signs, symptoms, and data from other relevant assays for optimal diagnostic accuracy and therapeutic decision making. NMI has a major positive impact on patient care, improving therapeutic strategy.

This chapter discusses the published clinical evidence regarding the impact of nuclear medicine in HPB diseases and focuses on state-of-the-art nuclear medicine. As such, it concentrates predominantly on published medical literature from the past 15 years. In our experience, most clinical nuclear medicine research publications before then often employ methodology and technology that is no longer reflective of current state-of-the-art clinical practice in nuclear medicine. The state-of-the-art in nuclear medicine, in its diagnostic and therapeutic procedures, has improved rapidly in the past 15 years and continues to evolve and innovate, including major improvements in commercially available nuclear imaging camera systems (particularly the advent of hybrid “fusion imaging” camera systems), image data processing, new types of instrumentation, and clinical introduction of new radiopharmaceuticals, both for diagnostic imaging and nuclear therapy.

Therefore we strongly advise the reader to note the dates of nuclear medicine references cited in HPB bibliographies and other guidelines, especially when these make judgments on the diagnostic accuracy or clinical impact of nuclear medicine; sometimes guidelines cite outdated nuclear medicine research from decades past. Such guidelines might be designed in recognition that nuclear medicine clinical practice varies worldwide, as reflected in the often widely varying diagnostic sensitivities and specificities reported by different medical centers performing a particular NMI procedure and the variation of hardware (e.g., scanners) and techniques (e.g., administered tracer doses, software-based data-processing algorithms) employed in different centers.

After an introduction to the pharmacology and technology of diagnostic imaging and therapy in radiopharmaceuticals and the general role of nuclear medicine in HPB diseases, we discuss current nuclear medicine procedures for specific HPB clinical indications. For diagnostic imaging procedures, discussion focuses on how well a particular clinical NMI study performs for a specific HPB indication, in terms of its diagnostic accuracy (sensitivity, specificity) and potential pitfalls, including necessary patient preparation, when applicable. The chapter also includes a concise look at select new, currently investigational radiopharmaceuticals relevant to HPB disease.

Radiopharmaceuticals

Nuclear medicine specialists prescribe radiopharmaceuticals for diagnostic imaging and internal radiotherapy of a variety of diseases. Radiopharmaceuticals can be placed into three major categories of applications in HPB disease: detection and evaluation of cancerous HPB tumors, treatment of HPB cancers, and evaluation of HPB organ function (and indirectly for detection of disease entities causing HPB organ dysfunction).

A radiopharmaceutical is a radioactive compound containing a radionuclide, also referred to as a “radioisotope” (radioactive isotope). A radioisotope is an energetically unstable atom that will achieve a stable or more stable, lower-energy state (transitioning from a parent to a daughter state) by releasing (radiating) energy (radiation) in some form (e.g., emitting a gamma ray, positron particle, or beta particle, as discussed later). The release of energy by the (parent) radioisotope atom may be called a physical decay, disintegration, or transition. The energy decay makes the elemental atom either become a different isotope of the same element (e.g., the radioisotope technetium 99m [ ^99m Tc] decays to the stable isotope technetium 99 [ ⁹⁹ Tc]) or become a different element by transmutation (e.g., the radioisotope ¹⁸ F decays to become a stable form of the element oxygen, ¹⁸ O). Other forms of nuclear decay are possible (e.g., transitions from a higher-energy unstable radioisotope to a lower-energy, but still unstable, daughter radioisotope). A radiopharmaceutical is administered in a trace amount (with no detectable radiobiologic effects) or therapeutic amount for use as a diagnostic imaging agent or therapeutic agent. A radiopharmaceutical also contains other active and inactive ingredients in the compound formulation. In the radiopharmaceutical, the elemental radioisotope atom typically is incorporated within a molecule by chemical bonding. The molecule is said to be radiolabeled.

As with any pharmaceutical, each type of radiopharmaceutical has in vivo pharmacokinetic (PK) properties specific to and determined by its molecular structure and associated physicochemical properties. PK properties include the radiopharmaceutical’s distribution in tissues throughout the body (biodistribution), metabolism, and bodily elimination (by hepatobiliary and urinary excretion for all relevant radiopharmaceuticals). The in vivo PK properties are also determined, to some degree, by the physicochemical properties of excipients (vehicles) in the radiopharmaceutical formulation (e.g., formulation of an orally administered radiopharmaceutical compound may affect its bioavailability and biodistribution), as well as by the route of administration (e.g., peripheral intravenous [IV] injection, hepatic arterial catheter infusion).

The mass-amount of radioactive molecules in any prescribed radiopharmaceutical formulation is only a trace amount, typically in the picogram (pg) range. This tiny mass-dose of radioactive molecule is incapable of exerting detectable pharmacologic effects on body tissues in vivo, but the typical pg amounts of radioactive molecules emit radioactivity sufficient for diagnostic imaging and therapeutic applications. With exceptions, the nonradioactive constituents of radiopharmaceutical compounds typically used only for clinical diagnostic imaging are present in somewhat higher mass-amounts but are still scant, typically less than 100 micrograms (µg), and allergic reactions, other side effects, or pharmacodynamic effects are rarely reported. Nuclear medicine specialists may prescribe the radiopharmaceutical compound to be administered in conjunction with a relatively high and biologically effective mass-amount of a nonradioactive, or unlabeled, version of the same compound or a related compound, with therapeutic intent (relevant compounds are discussed later).

The terms “radiotracer,” “tracer dose,” and “radiotracer dose” commonly refer to the use of trace amounts of a radiolabeled molecule to study molecular biology. The trace amount of radioactivity and the trace mass of the administered radiotracer are unable to affect (and therefore unable to interfere with measurements of) the biomolecular system or target being assayed. Following this common convention, in this chapter we use radiotracer to refer to radiopharmaceutical administered for diagnostic imaging. We use therapeutic radionuclide to refer to administration of a relatively high amount of radioactivity with the intent of inducing therapeutic radiobiologic effects in vivo, as discussed later. The radioactivity emitted by a therapeutic radiopharmaceutical may be useful for diagnostic imaging, as well as radiotherapy. The approach of combining diagnostic imaging and therapy using a same molecule or at least very similar molecules, which are either radiolabeled differently or given in different dosages, is known as theranostics.

Fluorodeoxyglucose positron emission tomography

In the past 25 years, fluorodeoxyglucose (FDG) PET has rapidly emerged as a revolutionary imaging modality in clinical oncology, demonstrating diagnostic efficacy in tumor staging and tumor-response evaluation for histologies across a variety of cancers. FDG, or 2-deoxy-2-( ¹⁸ F)fluoro-d-glucose, is an analogue of glucose; fluorine-18 occupies the molecular 2′ position in which a hydroxyl group is found in glucose. The substitution affects the metabolism of FDG compared with glucose. In vivo, IV FDG extravasates into tissues, followed by its uptake into tissue cells by glucose transporter proteins. Once FDG enters the cell, hexokinase converts FDG to FDG-6-phosphate, which cannot be metabolized further, thus trapping the tracer intracellularly. Blood FDG concentrations decrease to relatively low levels 45 to 90 minutes after injection; at this point, further FDG uptake in most tissues is relatively minor, and after that time, FDG concentrations in most tissues and tumors remain relatively stable. Acquiring a single FDG PET scan, beginning 45 to 90 minutes after injection, has become the standard clinical approach. Usually, as for HPB cancer imaging, the scan spans from “eyes to thighs,” including the entire head or extremities only if there is a patient-specific clinical reason.

The basic rationale for using FDG PET for tumor detection is the observation that neoplastic cells typically accumulate FDG more than the non-neoplastic cells of origin and that the difference in FDG concentration between the tumor and surrounding normal tissues in an organ is detectable by PET. This avidity of tumors for FDG manifests on PET images as a “hot spot,” or a focus, of FDG accumulation that is of abnormally high concentration relative to other, healthy tissues.

Why are some tumors FDG avid and other are not very avid? The physician-scientist and Nobel laureate Otto Warburg long ago observed in multiple tumor cell lines that he studied an abnormally high rate of glycolysis in cancer cells compared with their normal cellular counterparts, even in the presence of normal levels of oxygen. In normal cells with adequate environmental oxygen, glucose metabolism is typically directed into the mitochondrial tricarboxylic acid (TCA) cycle; glucose–TCA cycle metabolism yields the maximal amount of energy substrate (adenosine triphosphate [ATP]) from each glucose molecule metabolized for meeting the bioenergetic needs of the cell. The TCA cycle depends on oxygen to function; in normal cells, if environmental oxygen is low, glucose metabolism instead occurs in the cytosol by an oxygen-independent glycolytic process that yields much less ATP per glucose molecule. In cancer cells, however, Warburg observed that glucose metabolism occurred predominantly by glycolysis in the cytosol, regardless of whether or not the tumor cells were well oxygenated. This preference of tumor cells is the Warburg effect. According to the Warburg hypothesis, cancer cell metabolism of glucose was inefficient because it yielded fewer ATP molecules per glucose cell, and this inefficiency was caused by a defect in the mitochondrial metabolism of cancer cells. The Warburg effect remains a valid observation, although not a universal phenomenon among all cancer cell lines and types (i.e., glucose metabolism of some cancer cell lines is essentially the same as the glucose metabolism of normal cell counterparts). The Warburg hypothesis, however, is outdated; the shift of glucose metabolism from the mitochondrial TCA cycle to cytosolic glycolysis is not an inefficient use of glucose. Rather, it is a “repurposing” of glucose. In multiple cell lines, abnormal cytosolic glycolysis has become understood as advantageous to cancer cell proliferation. Cytosolic glycolysis yields fewer ATP molecules, but it yields glucose-derived metabolites during the multiple intermediate steps of glycolysis that the cell can use in other anabolic pathways as components for synthesizing macromolecules necessary for building cellular biomass before cell division and tumor growth. To meet the bioenergetic needs of these cells, instead of predominantly relying on glucose, these cells depend on other nutrient molecules, notably glutamine, to fuel the TCA cycle.

The Warburg effect explains the avidity of tumors for FDG, the glucose analogue, when visualized by PET, but only in part. PET visualizes the FDG avidity of tumors at the macroscopic tissue level (again, with spatial resolution of ~2 mm). The “FDG-avid tumor” visualized by PET and described in PET/computed tomography (CT) reports represents a complex composite of FDG avidities of tumor cells and nonneoplastic cell constituents within the tumor internal microenvironment under the influence of complex biomolecular and other processes. Detected FDG avidity in a particular tumor often does primarily represent tumor cell FDG avidity (i.e., the sum of FDG uptake from all tumor cells within the tumor) more than the FDG avidity of other cells in the same tumor. However, the FDG avidity of other constituents of the tumor microenvironment sometimes contributes to a clinically significant degree, particularly in the posttreatment setting, potentially causing diagnostic confusion, as discussed. In certain cases, overall tumor FDG avidity can be caused, in relatively large part or even primarily, by tumor cellular constituents other than the neoplastic cells, such as infiltrating inflammatory cells, especially when the tumor cells do not have intrinsically high FDG avidity and when inflammatory cells are present in relatively high tissue concentrations. For example, inflammatory cells can accumulate around the necrotic cores of tumors before treatment and can infiltrate heavily throughout tumors after treatment. Standard FDG PET guidelines often advise that posttreatment PET be deferred for 6 to 8 weeks after chemotherapy and 2 to 3 months after radiotherapy. It was empirically observed in clinical PET trials that successfully treated tumors often demonstrate apparently suspicious residual FDG avidity in the first few weeks after treatment because of inflammatory cells infiltrating the treated tumors, presumably to clear the necrotic/apoptotic debris associated with successful treatment.

Whenever using PET to characterize or localize tumors, the oncologist (and imaging specialist) should be aware of key factors that affect the apparent FDG avidity of a tumor and the diagnostic sensitivity of FDG PET: tumor size; cancer treatment(s); and background organ FDG avidity. As the Nyquist principle indicates, the apparent FDG avidity of subcentimeter tumors will be underestimated because such small lesions fall below the spatial resolution of PET technology. It is still possible for PET to detect a subcentimeter tumor if the tumor is so FDG avid that the FDG accumulation in tumor is detectably higher than that in background tissues of the organ involved, but many subcentimeter tumors lack apparent FDG avidity and may be reported as “too small to characterize by FDG PET.” Additionally, the specificity of FDG can vary in the setting of coexisting benign pathologies, leading to false-positive results. For example, specificity of FDG PET is lower for patients who live in areas where tuberculosis (TB) is endemic. Furthermore, TB lesions absorb FDG and can mimic tumors on FDG PET.

Cancer treatments, depending on action and efficacy, also affect apparent tumor FDG avidity. Tumor FDG avidity represents the sum of the FDG avidity of constituent tumor cells. Various studies indicate FDG PET is unable to detect microscopic residual disease; for example, a partially treated tumor containing FDG-avid cells may be of macroscopic size on CT or magnetic resonance imaging (MRI) but may contain a depleted cell population with a sum FDG avidity that appears minimal to nil on PET imagery. For staging, FDG PET is expected to be less accurate after therapy than before therapy. FDG PET can also be false positive, detecting FDG uptake at a former tumor site in the absence of residual disease. With systemic therapy, residual FDG uptake may indicate inflammatory cells (extremely FDG avid, when active) infiltrating tissues to remove the debris of treated disease. Radiotherapy and surgery, for tumor treatment or resection, both evoke local tissue inflammation that can be greatly FDG avid, mimicking local residual or recurrent neoplastic disease on PET imagery. FDG PET is usually deferred for several weeks after surgery or radiotherapy, when evaluation for local disease is desired.

Certain organs and organ systems have marked FDG avidity consistently or variably that may exceed that of primary tumors and metastases, obscuring tumor detection. For example, the brain is consistently FDG avid because it normally depends on glucose metabolism; FDG PET has limited sensitivity for detection of brain metastases. FDG is excreted through the urinary tract; the radioactive signal from excreted FDG in the collecting systems typically obscures PET evaluation of the kidneys and urinary bladder. The liver, lungs, and other tissues have lesser degrees of background FDG avidity that usually do not obscure tumor detection significantly.

Besides tumor detection, for disease (re)staging, FDG PET may be used to evaluate tumor response to cancer therapy. Frequently, FDG PET is performed twice, before and after therapy, for comparison, using changes in tumor FDG avidity as an index of changes in tumor cell population size (i.e., tumor response). Marked decreases in tumor FDG avidity during therapy have frequently predicted favorable clinical outcomes, whereas stable or increasing FDG avidity portend worse outcomes across a variety of cancers. Evaluation of tumor FDG avidity after cytotoxic therapy without a pretreatment PET study for comparison can be performed but can yield potentially confusing findings; for example, reactive lymph nodes and partially treated metastatic adenopathy can have similar appearances on PET. Tumors may have marked residual FDG avidity after treatment, which may provoke concerns about tumor resistance. If a pretreatment FDG PET had been obtained, however, the residual FDG avidity might have been observed as a marked decrease from baseline tumor FDG avidity, suggesting a favorable tumor response. In other words, the change in tumor FDG avidity before versus after treatment can be more predictive of tumor response than merely the posttreatment FDG avidity alone. As mentioned, certain tumor histologies seem frequently to lack FDG avidity, despite the presence of viable, macroscopic neoplastic disease. A lack of tumor FDG avidity on a posttreatment PET scan can be potentially misleading as an indicator of tumor response, unless a pretreatment PET scan has demonstrated the tumor being treated was originally FDG avid.

Diagnostic imaging in nuclear medicine

In general, diagnostic NMI is a noninvasive procedure that uses scanning hardware to examine the distribution of a radiopharmaceutical within the internal environment of the body. As discussed, imaging the in vivo distribution of a radiopharmaceutical can be considered as an in vivo assay of radiopharmaceutical pharmacokinetics, not just in blood but also in tissues/organs throughout the body. No radiopharmaceutical compound yet designed has been found to bind exclusively to one particular biologic molecule. Some compounds, however, such as radiolabeled antibodies, radiolabeled “small molecules,” and other types of agents, do bind with very high selectivity and affinity to relatively few biologic molecules and not at all to other types of molecules and are called “targeted agents.” Still, the biophysiologic processes and biologic molecules targeted by such agents for diagnostic imaging (or “targeted therapy”) of a particular condition of interest can almost invariably be found in other physiologic or pathologic conditions, again precluding 100% specificity. For example, the biologic molecule prostate-specific membrane antigen (PSMA, now more properly referred to as glutamate carboxypeptidase II), once thought to be uniquely expressed by prostate tissues and thus a good biomarker for prostate cancer (e.g., for imaging by PSMA-targeted radiolabeled antibody), was later found to be expressed by certain other tissues in the body and in the neovasculature of most tumors. However, PSMA is highly expressed in only a few types of nonprostatic tissues and therefore still possesses high selectivity for prostatic tissues. Therefore “perfect” specificity should not be expected for diagnostic imaging agents, even radiolabeled antibodies, considering the underlying imperfect pharmacologic and biologic specificity, as well as potentially misleading imaging artifacts.

In diagnostic NMI, the image is produced by the radiopharmaceutical administered to the patient. Once administered, the radioisotope physically decays with a characteristic radioactive emission pattern, producing energy or photons . These photons are detected by nuclear scanner (e.g., PET scanner or gamma camera) and an image is created.

Does the biodistribution of the radioisotope atoms visualized by the nuclear scan represent the biodistribution of the administered radiopharmaceutical (molecules)? If the radiopharmaceutical does not undergo in vivo chemical transformation to another form (e.g., catabolite or metabolite) before imaging of the patient, the answer is yes. Otherwise, the radioisotope biodistribution imagery may represent a composite of biodistributions, including those of the (unmodified) administered radiopharmaceutical and the radioactive products of in vivo chemical reactions (i.e., reaction products that still incorporate the radioisotopic atom). Usually, in vivo metabolism of the radiopharmaceutical causes in vivo production of metabolites, one or more of which include the radioisotope; these are radiometabolites. Such metabolism may be the diagnostic imaging target of the nuclear scan (e.g., PET imaging with F-18 FDG to detect tumor concentrations of the FDG metabolite). Some radiometabolites may be radiolabeled molecules, or in vivo metabolism may yield radioisotope in free, unattached elemental form. These radiometabolites often have different in vivo PK properties from the intact parent radiopharmaceutical. Thus the radiotracer biodistribution visualized by nuclear imagery will represent a combination of biodistributions: that of the intact radiopharmaceutical and that of one or more radiometabolites. In vivo metabolism occurs but typically does not interfere with diagnostic interpretation. On the contrary, metabolism may yield a radiometabolite “trapped” in a tissue of interest, such as enzymatic trapping of the PET imaging FDG in tumor cells; the cytoplasmic enzyme hexokinase yields the radiometabolite FDG-6-phosphate, which is trapped intracellularly. This chapter discusses the meaning of each radiopharmaceutical scintigraphic biomarker scan relevant in HPB diseases. Once administered to a patient, the radioisotope used for diagnostic imaging emits radiation that can be detected by a nuclear scanner.

Diagnostic imaging with radiopharmaceuticals, in standard clinical practice, may be referred to in various ways, including (1) using general terms such as nuclear imaging or scintigraphy, (2) referring to one of two general types of scintigraphic camera technology (PET, single-photon emission computed tomography [SPECT]/single-photon emission tomography [SPET]), and (3) using the procedure involved (e.g., theranostic imaging). The term scintigraphy (Latin scintilla , “spark”) in medicine refers to the light produced by crystalline detectors in clinical scintigraphic cameras when those crystals are struck by gamma rays emitted from radiopharmaceuticals (e.g., as emitted from within a patient scanned after receiving a radiopharmaceutical injection). These scintillations produced in the crystalline detectors are recognized and processed by the camera system to yield nuclear imagery.

Of the basic types of scans found in a radiology department (e.g., plain radiography, CT, MRI, ultrasound [US]), diagnostic NMI scans are typically of the longest duration, in terms of both the time the patient must physically spend with the scanner and the time required for the entire study (start to finish), often with necessary delays before scanning or between scanning (i.e., if the patient is scanned more than once after a single radiopharmaceutical administration) to allow the radiopharmaceutical time to undergo desired in vivo physiologic processes. The total duration of a diagnostic NMI study thus depends on a variety of technical, biologic, and typical clinical logistical variables. Most frequently, a radiopharmaceutical is administered intravenously by bolus injection. After the injection, a standard time-delay may be necessary before the patient undergoes scanning to allow the radiopharmaceutical to spread throughout the body and achieve a biodistribution considered optimal for imaging. To acquire data for a single image, the time that a patient spends “in front of the camera” must be of sufficient duration for the scanner to collect a statistically robust number of radioactive signals, or counts, to ensure that the derived imagery will be satisfactory for visual analysis. Low-count images are visually “noisy.” How long it takes for the camera to collect enough photons for a sufficient-quality diagnostic image depends primarily on the intrinsic properties of the radioisotope involved, how much radiopharmaceutical is administered, how well the radiopharmaceutical concentrates in tissues of interest (e.g., tumors) compared with surrounding tissues in vivo, how well the camera system detects photons, and how the photon data are constructed into imagery. Depending on the type of nuclear imaging study, before imaging even starts, there may be a standardized delay after the radiopharmaceutical administration to allow the radiopharmaceutical time to achieve an in vivo biodistribution considered optimal (i.e., one hour after FDG injection before PET scan acquisition). Lastly, the imaging specialist decides whether to have the patient undergo scanning at additional time points or using special techniques, if it is thought necessary to increase the diagnostic accuracy of the study. The referring clinician’s staff can help prepare patients mentally by advising them of the prolonged duration typical of diagnostic NMI.

Nuclear scanners may be categorized into two general types: PET scanners and standard gamma cameras . Their designs are tailored to image two fundamentally different types of radiopharmaceuticals (radioisotopes): those that emit positrons (for PET cameras) and those that emit gamma rays (for standard gamma cameras). As mentioned previously, images are created from photons produced by decaying radioisotopes administered to the patient, which are detected by nuclear scanners. The scanner system processes the photon data and reconstructs it into an image that can be presented as a two-dimensional (2D), or planar , image or as a (virtual) three-dimensional (3D) image (e.g., allowing display of sections of data in conventional axial, coronal, and sagittal views, similar to CT). Images may represent biodistribution at one or a few time points or can display time-dependent changes in biodistribution in cinematic fashion.

Planar images of radiotracer biodistribution in the anterior-posterior plane will result in an image in which in vivo tracer accumulations in two or more organs or other tissues may overlap (in the 2D plane) and thus potentially obscure detection or evaluation of the radiotracer uptake of interest (e.g., tumor detection). Tomographic (SPECT and PET) nuclear imaging can help avoid this potential issue by permitting tracer biodistribution to be evaluated in three dimensions. However, the limited spatial resolution of scintigraphic imaging may make it difficult to localize a particular tracer accumulation in a small tissue structure (e.g., tracer uptake in a small tumor may be hidden if the tumor is located within or immediately adjacent to a normal organ that also accumulates tracer). Additionally, for single-photon imaging agents, SPECT often requires a significantly longer duration scan than a 2D planar scan using standard SPECT camera systems, and often 2D imaging may be sufficient for the clinical data desired. The necessity for SPECT imaging is guided by the reason for a particular examination, available clinical research, and the particular patient context. For PET, by definition, tomography (3D imaging) is always used, involving ring-type dedicated PET camera systems with sophisticated signal analysis algorithms.

The advantage of PET imaging versus single-photon gamma imaging is that the PET permits a more precise determination of where the radiation originated, when using the coincidence-detection method. Thus the scan imagery reconstructed from PET data has a much better spatial resolution (typically 4–5 mm, vs. <1 mm on CT) than that reconstructed from single-photon gamma data (typically 12–15 mm). According to the Nyquist principle, this superior resolution results in PET-acquired data providing superior quantification of radioactivity concentrations in imaging data analyses compared with single-photon imaging. As one potential advantage versus PET imaging, single-photon imaging can simultaneously detect and distinguish two or more different gamma-emitting radiopharmaceuticals in a single patient in vivo, whereas PET imaging cannot distinguish between different PET isotopes. Gamma rays emitted by non-PET isotopes for single-photon imaging can have a variety of signature energy levels, which can distinguish it and be separated by signal processing.

Fusion imaging

Scintigraphic imaging of PET and single-photon emission radiopharmaceuticals are often combined with CT imaging for fusion imaging: PET/CT and SPECT/CT, respectively. In fusion imaging, the 3D imagery of PET or SPECT is combined with CT data so that tracer biodistribution/localization is visualized within the internal anatomy. Clinical studies have, overall, demonstrated that fusion imaging can have a synergistic effect on the accuracy of scintigraphy and CT image analyses for various clinical applications. A notable general example is improved accuracy for detection of radiotracer-avid tumors; often, on fusion imagery, scintigraphy (PET or SPECT) highlights findings poorly detected or easily overlooked on CT, or vice versa. The CT information also serves as data for an important technical function, called attenuation correction , which improves the quantitative accuracy of measurements derived from PET or SPECT analyses. Fusion imaging has done much to rescue diagnostic nuclear medicine from its former moniker (deserved or not) of “unclear medicine.”

One important caveat remains regarding the CT scans involved in fusion PET/CT and SPECT/CT imaging: the companion CT scan. The quality of the companion CT image can vary considerably; it can be of standard diagnostic quality (i.e., exactly the same technical-quality CT scan as obtained from a separate, stand-alone, state-of-the-art CT scanner), or it can be of inferior diagnostic quality (e.g., if acquired with a lower current [mA], yielding relatively noisier images, with less detail and greater susceptibility to certain artifacts, such as beam hardening). Chest imaging may be acquired at pulmonary end expiration rather than the standard maximal pulmonary inspiration, limiting evaluation. The CT might be acquired with the patient’s upper extremities positioned along the torso if the patient cannot tolerate having the arms raised (the standard chest CT position) for the 15 to 25 minutes of a standard torso FDG PET/CT, often creating a beam hardening artifact. Additionally, oral and IV iodinated contrast material more routinely used in standard CT protocol are not always used. Nevertheless, a noncontrast low-dose companion CT performed typically is sufficient for the basic needs of the PET or SPECT scan, providing sufficient anatomic detail to identify what tissues are involved in radiotracer uptake within an organ and providing attenuation correction, a modification of the scintigraphic data based on tissue densities and depth measured by CT that improves scintigraphic image quality, especially for quantification (e.g., for PET quantitative measurements known as the standardized uptake value [SUV]). Decay photons emitted from within deeper or denser body tissues will lose more energy from tissue interactions than photons from superficial or less dense tissues. The attenuation correction attempts to account for that artifact to provide more accurate measurement of tissue tracer concentrations.

Beyond CT, clinical PET/MRI systems are already available at a few major medical centers with increasing use in the setting of certain pathologies (e.g., prostate cancer and neuroimaging). The advantages of PET/MRI include simultaneous acquisition to improve registration of fusion images, lower radiation dose, superior soft-tissue contrast, and availability of multiparametric imaging. To date, the clinical use of PET/MRI has been limited by availability of accurate attenuation correction algorithms and problem of motion artifacts, including respiratory motion. With the advent of machine learning programs, however, these problems are being overcome and its use in clinical practice is expected to grow accordingly.

“Can I order two nuclear medicine scans on the same day?”

The distinction between positron-emitting (PET) and solely gamma photon–emitting (non-PET) isotopes is relevant for referring specialists primarily because of this question. Unlike positron-emitting radioisotopes, the gamma ray–emitting isotopes of radiopharmaceuticals used for single-photon imaging come in a variety of energy levels. Gamma cameras can distinguish and potentially use these energy levels to separate the biodistributions of one radioisotope from another if a patient were to receive two radiopharmaceuticals with different radioisotopes simultaneously (i.e., by having the camera only accept detected photons of the energy level characteristic of the particular radioisotope of interest, then doing the same for the other radioisotopes). However, the energy emissions of different single-photon imaging radioisotopes can overlap, particularly if one of the radioisotopes emits relatively high-energy emissions, because some emitted rays will lose energy and fall into the energy (keV) range of the other radioisotope. The 511 keV coincidence gamma rays of PET radiopharmaceuticals are relatively high energy and will interfere with imaging of single-photon radiopharmaceuticals (whether to a significant degree depends on other technical factors). Thus, after a PET scan, some period of delay is necessary, to allow the PET radiopharmaceutical to undergo physical decay and biologic clearance from the patient, before performing another imaging study with a single-photon radiopharmaceutical or another PET radiopharmaceutical.

Practically, if a patient receives ¹⁸ F-FDG for FDG PET, the patient should wait more than 20 hours before a subsequent nuclear study is performed, based on the known half-life of ¹⁸ F (~1.9 hours), the amount of tracer we administer to a patient (maximal tracer dose of 12 mCi), and a common rule of thumb of waiting 10 half-lives for a radioisotope to decay sufficiently before allowing imaging with another radiopharmaceutical. In general single-photon radiopharmaceutical imaging, however, studies can be performed immediately before PET scan as the photons of non-PET radioisotopes typically are relatively low energy and thus cannot interfere with detection of the relatively high-energy 511 keV photons of PET radiopharmaceuticals. A few non-PET radioisotopes do emit high-energy photons at or above the 511 keV range, but even the presence of these high-energy photons from a prior non-PET radiopharmaceutical injection do not necessarily preclude immediately performing a PET study. Because the high-energy photons of the non-PET radioisotope are not produced as pairs, these photons will strike PET detectors in random directions. The PET scanner, however, will ignore photons that are not detected in the coincidence pattern typical of annihilation photon pairs. As such, the PET scanner often can detect the desired PET radiopharmaceutical signal without significant interference from any residual non-PET radiopharmaceutical in the patient’s system.

Radiation dose in nuclear medicine

When administered solely for diagnostic imaging, conventional radiopharmaceuticals expose the receiving patient to a low radiation dose, typically one or more orders of magnitude below the level of radiation exposure conventionally accepted as being associated with an increased risk for harmful radiation effects, based on decades of dosimetric research. Certain radiopharmaceuticals are administered at relatively high doses of radioactivity with therapeutic intent; the administered radioactivity is sufficiently high (and concentrates in body tissues at levels sufficient) to induce acute radiobiologic effects on diseased tissues and other organs within a patient, with potential therapeutic benefit as well as adverse side effects. The radiation dose absorbed by the patient from the radiopharmaceutical is determined by the specific type of radioactive isotope involved, the administered activity (amount) of radiopharmaceutical (quantified in becquerels, typically megabecquerels [MBq], or curies, typically millicuries [mCi]), and the pharmacokinetics and distribution of radiopharmaceutical throughout the body (biodistribution), which is again determined by the physicochemical characteristics of the radiopharmaceutical molecule (or element).

Nuclear medicine and liver cancer

Hepatocellular carcinoma

For more information on hepatocellular carcinoma (HCC), see Chapter 89 .

Detection and staging

HCC is detected and characterized by CT and MRI (see Chapter 14 ). Both 2020 National Comprehensive Cancer Network (NCCN) guidelines and 2018 European Association for the Study of the Liver (EASL) guidelines do not recommend PET/CT for detection of HCC because of its limited sensitivity.

HCC is FDG avid in less than 40% of HCC cases, with the majority of well-differentiated HCC tumors not avid or demonstrating avidity close to background liver on 18F-FDG PET, thus limiting sensitivity of detection ( Fig. 18.1 ). Inherently low FDG avidity is because of the (1) HCC lower levels of glucose transporter-1 expression and (2) overexpression of the enzyme glucose-6-phosphatase, which hydrolyzes FDG-6-phosphate to FDG, which can be transported out of the cell. However, increased glucose transporter expression has been demonstrated in poorly differentiated HCC, with increased FDG avidity on PET ( Fig. 18.2 ); thus non-avid HCC appears to be associated with a less aggressive tumor and a more favorable patient prognosis.

FIGURE 18.1, Well-differentiated hepatocellular carcinoma (HCC) on fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT).

FIGURE 18.2, Poorly differentiated hepatocellular carcinoma (HCC) on fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT).

Tumor-to-liver uptake ratio (TLR) has been shown to correlate more closely with HCC doubling time and represents metabolic activity of HCC more precisely than SUV, with one study performed in 116 patients showing that a higher TLR (>1.62) was associated with poorer prognosis and presence of extrahepatic metastases. ^, Additionally, FDG-avid portal vein thrombosis has been shown to be an independent predictor for progression-free survival (PFS) and overall survival (OS), irrespective of avidity of primary HCC. Thus PET/CT may be of potential prognostic value before surgical resection, liver transplantation, or locoregional therapy.

Extrahepatic metastasis (most commonly lung, abdominal nodes, and bone) have been found in 37% of patients during staging with poorly differentiated HCC more frequently likely to metastasize. A positive statistical correlation between FDG avidity of primary HCCs and tendency of extrahepatic metastasis has been shown, suggesting that metastatic HCC lesions would also have increased FDG uptake. FDG PET has demonstrated high sensitivities of 77% to 100% for detecting extrahepatic metastasis, notably bone metastasis, with one study demonstrating superior diagnostic detection compared with bone scintigraphy. Thus PET may have a role in detecting extrahepatic disease, but current NCCN and EASL guidelines do not support routine use. NCCN recommends continued use of bone scintigraphy with ^99m Tc–radiolabeled bisphonates, such as methylenediphosphate (MDP), for staging of HCC patients with bone lesions.

Additional PET tracers have been studied in HCC with Carbon 11 ( ¹¹ C) acetate and ¹¹ C choline and perform better than FDG PET for detection of well-differentiated HCC but are inferior compared with CT/MRI. Also, ¹¹ C has a physical half-life of 20 minutes, requiring an onsite cyclotron or nearby producer for manufacturing of tracer, thus further limiting its widespread routine clinical use. ⁶⁸ Ga-PSMA has recently been proposed as a potential tracer in HCC. PSMA plays a major role in regulating angiogenesis and endothelial cell recruitment, which occurs early in HCC and throughout hepatic tumorigenesis. Nearly 95% of HCCs stain positive for PSMA in the tumor vasculature and early prospective trials have shown that ⁶⁸ Ga-PSMA PET has outperformed ¹⁸ F-FDG PET in the detection of HCC and extrahepatic disease.

Tumor-response evaluation

FDG PET/CT has become a tumor-response radiologic biomarker with a major clinical impact on the management of a growing number of other cancers (e.g., it is standard of care in breast cancer, lymphoma, esophageal cancer, and gastric cancer). In these types of cancer, overall robust clinical literature demonstrates that a therapy-induced decrease in tumor FDG avidity, as a biomarker of a decrease in tumor metabolism and tumor cell mass, correlates with histopathologic response and is often prognostically powerful, especially when conventional changes in tumor volume (e.g., anatomic criteria such as RECIST) measured by CT or MRI failed to predict either pathologic response or prognosis. Nevertheless, unlike cancers such as invasive ductal breast cancer or high-grade lymphoma, which are frequently FDG avid, HCC frequently demonstrates minimal to no FDG avidity, therefore limiting the role of PET-response evaluation. As such, PET is not currently routinely recommended for HCC tumor response to therapy.

Still, the role of PET/CT has been shown to be potentially useful in certain post-treatment scenarios. In previous studies of HCC patients who underwent curative surgical resection or liver transplantation, high FDG uptake in an HCC showed a significant association with tumor recurrence, especially early recurrence. Furthermore, retrospective studies have shown that when PET/CT is used in combination with Milan criteria (a solitary tumor ≤5 cm in diameter or 2 to 3 tumors ≤ 3 cm in diameter) with or without serum alpha fetal protein (AFP) for selecting candidates for liver transplantation, patients who are beyond the Milan criteria but have a negative PET/CT had clinical outcomes comparable with those within Milan criteria. Furthermore, in candidates who met Milan criteria but had an FDG-avid HCC, higher rates of recurrence were seen than in those who had low FDG-avid HCC. ^, Thus a FDG PET finding has been found to be an independent predictive factor for tumor recurrence. Along with Milan criteria and serum AFP and FDG PET, it could provide additional information for making decisions regarding liver transplantation for HCC patients.

Locoregional therapy (e.g., ablation or transarterial chemoembolization [TACE]) is a preferred treatment option for patients with unresectable or nonoperable liver-confined disease (see Chapters 94 and 96 ). The prognostic value of FDG PET has been assessed in HCC patients treated with locoregional therapy, which suggests longer PFS and OS in patients with low FDG uptake of HCCs, again suggesting significant associations between FDG avidity of HCCs and clinical outcomes.

Recurrence

HCC recurrence typically manifests as tumor regrowth at a prior site of treatment (e.g., ablation) or as tumor appearing at a new site in the liver. PET/CT is not routinely recommended for surveillance in the post-treatment HCC patient. However, Hayakawa et al. evaluated FDG PET/CT for detecting recurrent HCC postoperatively in patients with either suspected recurrence on CT or MRI (group 1) or suspected recurrence because of abnormal serum tumor markers but with no disease evident on CT or MRI (group 2). FDG PET/CT had a 53% and 41% sensitivity and 100% specificity for recurrent tumor in both groups 1 and 2, respectively. The data from group 1 support the idea that FDG PET/CT cannot replace CT or MRI as the first-line imaging modality for detection of recurrence; the data from group 2 support the hypothesis that FDG PET/CT may be of value as second-line imaging if CT or MRI fails to detect recurrence. Wang et al. reported a remarkably high sensitivity of FDG PET/CT for HCC detection of 97%, with a specificity of 83% ( n = 32). FDG PET/CT has reported sensitivity no greater than 90% (usually much lower) for HCC in the pretreatment setting.

Colorectal cancer metastasis to liver

For more information on colorectal cancer (CRC) metastasis to liver, see Chapter 90 .

The most recent guidelines from the NCCN (2020), European Society of Medical Oncology (ESMO), and European Registration of Cancer Care (EURECCA) agree on the potential roles and limitations of FDG PET/CT in clinical management of CRC and only recommend PET/CT in certain clinical circumstances (e.g., potentially surgically curable metastatic disease). As such, this discussion focuses on the CRC patient with potentially resectable liver metastases.

In CRC patients with liver metastases, but no extrahepatic metastases, complete resection of liver metastases improves long-term survival better than other treatments currently available. (see Chapters 90 , 97 , and 98 ). Before surgery, it is essential to confirm that liver metastases are likely resectable (based on CT and/or MRI imaging) and that no extrahepatic metastatic disease is present. CRC metastases occur most commonly in the liver, followed by the lungs, with metastases in the central nervous system, bones, adrenal glands, and spleen occurring in less than 10% of CRC patients.

FDG PET/CT is extremely sensitive in detection of liver and lung metastases, the two most common viscera to be involved by metastatic disease at initial presentation. FDG PET/CT has long been recognized as superior to conventional CT for detecting hepatic metastases, with decreasing sensitivity for both modalities in characterizing subcentimeter hepatic lesions. ^, FDG PET/CT and MRI are equivalent in diagnostic sensitivity for liver metastases, but MRI is more sensitive in identifying subcentimeter liver lesions than FDG PET/CT (i.e., with greater sensitivity). As such, PET/CT is not routinely indicated in primary staging of CRC in current NCCN guidelines. ESMO 2014 colorectal cancer guidelines agree with current NCCN guidelines regarding the role of FDG PET-CT.

However, FDG PET/CT is considered a diagnostic adjunct to staging in patients with an equivocal finding on CT/MRI or in patients with potentially surgically curable metastatic disease (M1), as demonstrated on CT/MRI, but require evaluation for unrecognized metastatic disease that would preclude the possibility of surgical management. Patients planning to undergo hepatic resection based on conventional imaging will be found to have extrahepatic disease by FDG PET in 18% to 32% of cases ( Fig. 18.3 ), changing management in 20% to 40% of cases in early clinical PET trials ^, and changing management in curative-intent surgery in as many as 25% of patients in a later trial. Fernandez et al. found that patients with hepatic metastases who underwent FDG PET for preoperative staging had a much higher 5-year survival rate than historic controls. A randomized, controlled trial (RCT) of patients with resectable metachronous metastases assessed the role of PET/CT in the workup of potential curable disease. This study showed that while PET/CT did not impact survival, surgical management was changed in 8% of patients after PET/CT, with additional sites of metastatic disease detected in 2.7% of patients (bone, peritoneum/omentum, abdominal nodes), thus precluding surgical resection. In addition, 1.5% of patients had more extensive hepatic resections and 3.4% had additional organ surgery. However, 8.4% of patients in the PET/CT arm had false-positive results, many of which required additional imaging with biopsies or other imaging modalities.

FIGURE 18.3, Role of fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT) in colon cancer.

Additionally, current NCCN guidelines for CRC do not advocate routine use of FDG PET/CT for post-treatment follow-up imaging of patients with no evidence of distant metastatic disease by contrast-enhanced CT/MRI because of the potential risk for false-negative (e.g., non-avid necrotic liver lesions after chemotherapy) or false-positive findings (e.g., post treatment or surgery tissue inflammation). NCCN guidelines also do not recommend PET/CT in long-term monitoring: a small RCT reported earlier detection of recurrences with PET and suggested improved clinical outcomes compared with conventional CE imaging, but larger trials are required to further investigate its utility.

The utility of FDG PET/MRI in the follow-up of treated CRC patients has been investigated, with initial studies showing promising results. When compared with current standard-of-care imaging, PET/MRI changed clinical management in 35.7% of cases: 21.5% upstaging cases and 14.2% downstaging cases ( P < .001). However, larger multicenter prospective studies with larger patient numbers are required to confirm these preliminary results.

NCCN guidelines also suggest clinicians consider FDG PET/CT for evaluation of patients with serial elevations of serum carcinoembryonic antigen (CEA) levels without an identifiable source after standard IV contrast CT or colonoscopy. A systematic review and meta-analysis of 11 trials (510 patients) reported pooled estimates of sensitivity and specificity for FDG PET/CT for detection of occult tumor recurrence (after conventional CT workup) of 94% (95% confidence interval [CI], 89%–97%) and 77% (95% CI, 66%–86%), respectively.

The role of nuclear medicine in locoregional liver therapy

Although surgical resection often provides the best patient outcomes for patients with HCC or metastatic liver lesion, some patients are not surgical candidates (e.g., because of the extent of tumor, underlying liver disease, or comorbid conditions). In these patients, locoregional therapies (e.g., ablation, transarterial chemoembolization, or radioembolization) can be offered typically as a palliative therapy, although it can offer curative extent in certain cases. Locoregional therapies will be discussed in more detail in Chapters 94 and 96 , but we will briefly touch on the role of nuclear medicine in workup, therapy, and surveillance in this setting.

The role of nuclear medicine in hepatic arterial infusion therapy

Hepatic arterial infusion therapy (HAIT) is direct infusion of a therapeutic compound (e.g., chemotherapeutic, embolic, or radioembolic agents) to treat malignant liver tumors (primary or secondary; see Chapters 94 , 97 , and 100 ). Clinical studies demonstrate that arterial infusion improves tumor uptake of certain chemotherapeutic agents compared with portal venous or systemic venous infusion. Cancerous liver tumors derive/stimulate a nutrient blood supply from the arterial system by tumor neoangiogenesis as opposed to normal hepatic parenchyma, which receives blood supply mostly from the portal venous system.

Hepatic arterial infusion scintigraphy (HAIS), also called hepatic arterial perfusion scintigraphy, is the imaging of the biodistribution of a radionuclide delivered directly into liver through an arterial catheter, typically in the (proper) hepatic artery, for delivery of chemotherapy, or potentially into hepatic lobar or segmental arterial branches, when HAIS is conducted associated with HAIT radioembolization (e.g., treatment with ⁹⁰ Y-radiolabeled resin or glass microspheres or ¹³¹ I-radiolabeled lipiodol).

The role of HAIS is to:

(1)
predict that infused macroaggregated albumin (MAA) is properly distributed throughout the downstream liver or targeted hepatic subregion before HAIT,
(2)
ensure that no extrahepatic infusion caused by variant anatomy is present before HAIT,
(3)
calculate lung shunt volumes before radioembolic HAIT is performed,
(4)
predict or measure hepatic radiation dosimetry, before or after HAIT, and
(5)
ensure the integrity and function of an infusion system before administration of medication (e.g., post placement of a hepatic intraarterial pump reservoir-catheter system, which is used to deliver local chemotherapy).

Placement of the arterial catheter used for HAIT and HAIS involves different possible techniques, most commonly placed intermittently into the hepatic artery during an interventional radiology (IR) procedure. In certain institutions, catheters may connect to a subcutaneously implanted port or an infusion pump system to allow slow, continuous drug infusion of hepatic arterial chemotherapy (more commonly performed in the United States). The tracer used for HAIS for treatment planning is ^99m Tc MAA, a radiolabeled particulate, 25 to 50 μm in mean diameter, which physically occludes the first microvascular lumen it encounters. The infusate tracer, several hundred thousand MAA particles, labeled with a trace amount of radioactivity, are suspended in a small volume of saline (e.g., 2 cc) and injected into a transiently placed hepatic artery catheter or hepatic pump (see Chapter 97 ) by a trained operator (e.g., physician or nurse) and remain intact in vivo for several hours, allowing imaging of the distribution of tracer. Injection of an infusate can be slow (e.g., <1 cc/min), which mimics somewhat the slow rate typical of pump-infusions in HAI chemotherapy. Fast injection of infusate (e.g., ≥1 cc/sec; bolus) allows rapid introduction of infusate into the hepatic artery, creating turbulence in the bloodstream that mixes infusate and blood more homogeneously and more rapidly and, as a result, bolus infusions of tracer infusate are typically not advised for HAIS because the bolus pressure may cause retrograde flow of tracer into other celiac branches and artifactual extrahepatic tracer accumulations. Extrahepatic tracer accumulations, when tracer is infused slowly, are usually because of variant hepatic arterial anatomy, such as branches supplying extrahepatic viscera (e.g., stomach, pancreas, gastrointestinal [GI] tract) and could lead to complications (e.g., GI ulceration) if not identified before HAIT. True aberrant arterial branches typically must be embolized before HAIT to avoid extrahepatic organ toxicity.

SPECT/CT provides excellent anatomic localization of infused ^99m Tc MAA. Alternatively, ^99m Tc-labeled SC scintigraphy can immediately precede HAIS, to provide a gross outline of hepatic (and splenic) anatomic contours ( Fig. 18.4 ). Because ^99m Tc-MAA HAIS and ^99m Tc-SC scintigraphy use the same ^99m Tc isotope, a larger amount (activity, MBq) of ^99m Tc MAA is administered than ^99m Tc sulphur colloid (SC), so that the signal on the subsequent HAIS scan predominantly represents ^99m Tc MAA. However, confusion can arise regarding overlap between ^99m Tc-SC and ^99m Tc-MAA signals, as in evaluation of the homogeneity of ^99m Tc-MAA hepatic distribution. Additionally, if any delay occurs after ^99m Tc SC is injected, before ^99m Tc-MAA scintigraphy is performed, ^99m Tc-SC catabolic breakdown and release of free pertechnetate can yield gastric tracer-uptake, potentially mimicking extrahepatic infusion of the stomach on ^99m Tc-MAA scintigraphy. Obtaining dynamic images of ^99m Tc-MAA uptake as it is infused slowly for two minutes or more permits detection of increases in hepatic or extrahepatic tracer uptake indicative of true ^99m Tc-MAA signal, whereas any residual ^99m Tc-SC or free-pertechnetate signal will remain unchanged during dynamic imaging of the ^99m Tc-MAA infusion ( Fig. 18.5 ). ^99m Tc-MAA SPECT/CT without ^99m Tc SC has the advantage of having SPECT imagery that solely represents ^99m Tc-MAA biodistribution, although CT typically is associated with a higher radiation dose to the patient than ^99m Tc-SC scintigraphy.

FIGURE 18.4, Normal appearance of technetium 99m ( 99m Tc)–labeled sulfur colloid scintigraphy.

FIGURE 18.5, Normal appearance of hepatic arterial infusion scintigraphy.

Abnormal HAIS findings include subtotal hepatic infusion (in nonselective angiography cases), extrahepatic organ infusion, catheter obstruction, and catheter leakage ( Fig. 18.6 ) and typically require additional investigation (e.g., fluoroscopy or CT hepatic angiography) to guide subsequent management. In a clinical study of patients with implanted hepatic arterial pump-catheter systems, HAIS studies revealed abnormalities in 9% of patients after pump implantation; the abnormalities were predominantly extrahepatic infusion (63% of abnormal studies), but 12% demonstrated abnormal subtotal intrahepatic infusion (i.e., infusion distributed to only a portion of the liver, when infusion of the entire liver was expected). Abnormal subtotal intrahepatic infusion (i.e., regions of devoid of infusate uptake) should be distinguished from heterogenous intrahepatic distribution of MAA infusate, which occurs probably in part because of laminar flow phenomenon. The laminar flow phenomenon can occur in the presence of a large hepatic tumor mass, which appears to draw hepatic arterial flow away from other hepatic regions. Heterogenous MAA infusate distribution throughout the liver has no clear clinical significance and in our experience is relatively common and not clearly associated with adverse outcomes. If MAA infusate is clearly, unexpectedly absent (not merely relatively low) in one or more hepatic subregions, however, the finding is potentially clinically significant because it suggests the possibly of aberrant intrahepatic arterial anatomy, a misplaced catheter, or an occluded arterial branch (e.g., stenotic or thrombosed). Anecdotal cases, including our own experiences, have found that abnormal subtotal hepatic infusion on HAIS can predict poor tumor response in those subregions without detectable receipt of infusate, whereas tumors in the same patient that appear well perfused on HAIS respond favorably. As such, abnormal subtotal hepatic infusion always warrants further investigation.

FIGURE 18.6, Abnormal hepatic arterial infusion scintigraphy (HAIS).

In the setting of hepatic pump HAIS abnormalities, pump fluoroscopy evaluation fails to find corresponding abnormalities in approximately 25% of cases. In those cases, HAIS was repeated, and the previously identified HAIS abnormality was no longer evident scintigraphically in almost all cases. In the few cases in which the HAIS abnormality persisted, fluoroscopic or CT studies were again repeated, and a corresponding aberrant vessel was successfully identified. This study recommended evaluating abnormal HAIS findings by fluoroscopic correlative imaging initially; if radiographic correlation is found, HAIS is repeated in two to three weeks, seeking spontaneous normalization of HAIS.

Radioembolization therapy with yttrium 90 ( ⁹⁰ Y)–labeled glass or resin microspheres delivered by hepatic intraarterial catheter infusion is used for treatment of colorectal metastases or HCC (see Chapter 94B ). Patients before therapy undergo HAIS to delineate hepatic perfusion and extrahepatic supply, as previously mentioned, but also dosimetry to calculate dose to tumor, and lung shunt fraction (LSF) needs to be calculated on ^99m Tc-MAA imaging before treatment because radiation-induced pneumonitis and sclerosis can occur because of hepatopulmonary shunting of radioembolic microspheres and is a potential major toxicity concern. Significant shunting is estimated at LSF greater than 15%, which requires treatment modification before HAIT to reduce complications or not treating if LSF is greater than 20%. Immediately after therapy, typically SPECT/CT is performed to evaluate technical success, although in certain centers, ⁹⁰ Y-PET/CT is used instead.

After radioembolization dosimetry, SPECT or PET/CT, performed within 24 hours of treatment, has been employed in certain institutions because it enables rapid and precise prediction of efficacy on a per-lesion basis and allows for early treatment adaptation in case of undertreatment of the lesions. One study showed that in chemo-refractory mCRC, patients treated with radioembolization that absorbed dose determined on post radioembolization ⁹⁰ Y-PET/CT correlated with metabolic response, and higher lesion mean absorbed doses were associated with prolonged OS.

There is no standard protocol for pre- and postradioembolization imaging, with CT being the most commonly used modality. The role of FDG PET/CT has been suggested as a promising radiologic assay for detecting a favorable liver tumor response to ⁹⁰ Y microsphere radioembolization. One study showed that increased metabolic activity of the lesion pretherapy was associated with decreased liver PFS postradioembolization with authors recommending use of FDG PET/CT as part of work-up before therapy. Furthermore, the current data are heterogenous with regards to FDG PET/CT in the setting of evaluating postembolization treatment response of liver metastasis versus HCC because well-differentiated HCC does not typically accumulate FDG as previously discussed. As such, current guidelines do not recommend use of PET/CT routinely in pre- and postradioembolization imaging.

The role of positron emission tomography in percutaneous liver ablation

Percutaneous ablation of hepatic malignancies, HCC or liver metastases, can be considered in patients who are not surgical candidates or as a bridging strategy before curative therapy (e.g., resection or transplantation ; see Chapter 96 ). As discussed previously, FDG PET/CT has been shown to be superior to CT or MRI alone in staging patients with hepatic metastases and can often lead to a change in the treatment plan in patients being considered for percutaneous ablation. In one study, pre-ablation PET/CT led to a change in clinical management in 26% of patients in whom extrahepatic disease was identified and ablation was not performed. Another study found that PET/CT altered the clinical management in 25% of potential radiofrequency ablation (RFA) candidates with colorectal hepatic metastases in whom extrahepatic disease was missed by conventional imaging, and systemic chemotherapy was offered instead of performing ablation. PET/CT imaging can be used within the IR suite to guide ablation in PET-avid CT occult lesions. Additionally, periprocedural PET/CT can be used to determine technical success and permits immediate repeat tumor ablation as needed.

Recurrence after percutaneous ablation of hepatic malignancy is, unfortunately, not uncommon. After ablation of HCC, local recurrence rates of 11% to 36% have been reported, with more frequent recurrence occurring after treatment of larger lesions and lesions close to large vessels. Metastatic colorectal cancer recurs even more frequently after ablation than in HCC, with reported rates over 50%. High quality post-ablation surveillance imaging is required to identify residual or recurrent tumor to allow early identification of recurrent or residual disease and retreatment. However, it can be challenging to distinguish recurrent or residual disease from normal treatment response. Immediately after thermal ablation of the liver, the central area of post-ablation necrosis is surrounded by a zone of hyperemia and a peripheral rim of mild reactive change, which appears as a central nonenhancing area of necrosis, with a rim of increased enhancement compared with normal liver tissue in the hyperemic zone on contrast enhanced CT or MRI (see Chapters 14 and 15 ). The rim of increased enhancement may mask residual viable tumor in the early post-ablation period. Cells destroyed by thermal ablation lose their ability to concentrate glucose within the cell; thus, on FDG PET/CT, there is a corresponding photopenic area. Glucose metabolism within the zone of hyperemia, however, is unaltered and normal FDG uptake or a rim of uniform, low-grade FDG uptake surrounding the ablation site may be present. If focal areas of increased FDG uptake are seen adjacent to the photopenic area of necrosis within 48 hours of ablation, residual macroscopic tumor is suspected. From 72 hours to 6 months after ablation, a band of regenerative tissue containing neutrophils and fibroblasts develops surrounding the ablation zone that demonstrates variable degrees of both peripheral enhancement and also increased FDG uptake surrounding the ablation site because of inflammatory changes; however, FDG PET/CT seems more sensitive for determining the treatment effect and detecting local recurrence because of combined functional and anatomic information. Chen et al. retrospectively reviewed 33 lesions treated with RFA in 28 patients and reported that PET/CT demonstrated superior sensitivity and accuracy (94.1% and 87.9%, respectively) compared with MRI (66.7% and 75%, respectively) and multiphase contrast-enhanced CT (66.7% and 64.3%, respectively). To date, however, there are no established guidelines for when to perform post-ablation PET/CT, with some centers proposing intervals of 3 to 6 months for the first year after ablation, and over 95% of local recurrences identified within 1 year of treatment.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here