Dr. Kipper is the Physician Coordinator for Nuclear Imaging and PET in the Department of PET and Nuclear Medicine Services, Radiology Medical Group, San Diego, CA.

Positron emission tomography (PET) has become an important diagnostic modality for radiologists and nuclear medicine physicians. Developed nearly 3 decades ago for research purposes, it is now recognized as a powerful clinical tool in the investigation of oncologic patients. Efficacy has also been shown for selected neurologic and cardiac disorders (Figure 1). The purpose of this brief overview is to provide the diagnostic imager with practical information about the indications, performance, interpretation, and potential pitfalls of PET imaging. A limited number of references are provided for the reader interested in more-in-depth reviews of specific topics.

Indications

More than 90% of PET studies are performed for cancer in the clinical arena. The basis for in vivo imaging of F-18 fluorodeoxyglucose (FDG) distribution is the enhanced glycolysis of tumor cells, recognized more than 60 years ago. Uptake and concentration of glucose (and FDG) are facilitated by two factors: increased numbers of glucose transporters on malignant cell surfaces and a block in the intracellular glycolytic pathway due to relative glucose-6-phosphatase deficiency. Ultimately, images reflect the relative state of glycolysis that is increased in tumor cells. Tumors with the highest metabolic rates demonstrate the greatest FDG accumulation. Allowing for variability between tumor types, as well as potential differences in metabolic activity for a specific tumor, generalizations are possible as a guide in selecting appropriate patients for study.

Lung cancer

Positron emission tomography is extremely valuable for the diagnosing, staging, evaluating for recurrence, and, possibly, monitoring of response to treatment in non-small-cell lung cancer (Figure 2). It has demonstrated >90% sensitivity and specificity in characterizing solitary pulmonary nodules as benign or malignant based upon FDG concentration (Figure 3). This activity can be characterized by visual/qualitative assessment (eg, compared with normal mediastinal activity) or semi-quantitatively by calculating a standardized uptake value (SUV). The SUV is a numerical representation of the concentration of FDG in a focus/lesion compared with body activity. Most PET imagers use a cutoff SUV of 2 to 2.5 to separate a benign from malignant focus. Small-cell lung cancer is nearly always disseminated by the time of diagnosis; and therefore the use of PET in this disease is limited to select patients to answer specific questions. Major advantages of PET over conventional, anatomic-based modalities are its ability to characterize mediastinal and hilar lymph nodes as benign or malignant (not relying on size or shape of the node with the well-recognized inaccuracy of these criteria), whole-body survey capability (including accurate adrenal evaluation), and differentiation of post-therapy changes from those of recurrent disease. Caution must be exercised with bronchoalveolar carcinoma, and most authors report reduced sensitivity (likely due to lower metabolic rate). ¹

Colorectal cancer

Positron emission tomography imaging is utilized for staging and for assessing recurrence of colorectal cancer (Figure 4). There is preliminary data to suggest a potential role for monitoring response to therapy. ² Longitudinal surveillance is often performed with measurements of carcinoembyonic antigen (CEA) levels; however, the sensitivity of this test is only 60%, and localization of lesion(s) is not possible. The initial imaging study is appropriately a computed tomography (CT) scan of the abdomen/pelvis, but again, the sensitivity of CT is less than whole-body PET. Many oncologists now utilize PET at the first sign of a rising CEA and certainly in cases in which the CT is negative. Based upon PET's ability to characterize the metabolic status of a mass, it is recommended to differentiate post-therapy change (eg, scar, radiation fibrosis) from active tumor, which is difficult at best with CT and MR. The importance of early identification of tumor recurrence is twofold: first, solitary metastases may be amenable to curative surgery, and second, multiple sites of recurrence will spare the patient an unsuccessful curative-intent operation and will redirect therapy (Figure 5).

Lymphoma

Both Hodgkin's disease (HD) and non-Hodgkin's lymphoma (NHL) are FDG-avid. Many reports have documented the high accuracy of PET in nodal, extranodal, and bone marrow staging of lymphomas (Figure 6). ³ It also appears that the degree of FDG uptake may have prognostic significance. Presently, conventional staging is most often performed with contrast-enhanced CT of the neck, chest, abdomen, and pelvis; gallium whole body imaging; and bone marrow aspiration. Although there are studies to suggest that a PET scan is more accurate than the combination of all of the above, ⁴ a reasonable recommendation is to use PET as a complementary examination in conjunction with CT (and possibly bone marrow biopsy), while eliminating the gallium scan. After PET imaging, up to 15% to 20% of patients will be restaged. ⁵ Note that low-grade lymphomas may show diminished FDG accumulation due to reduced metabolic rate/low proliferative activity. Positron emission tomography is also extremely accurate for identifying recurrence and monitoring response to therapy. A pooled series in 262 patients for monitoring response showed a positive predictive value of 88% and a negative predictive value of 94%. ⁶

Melanoma

Positron emission tomography can contribute to the evaluation and management of melanoma patients in several areas: staging, identifying recurrence, monitoring therapy, and assessing a patient for curative-intent surgery after the discovery of a metastasis. Most PET whole-body surveys extend from mid-thigh to mid-nose; however, because of the potential involvement of any bodily site with melanoma, images should include the top of the skull to the bottom of the feet. Performing the study with attenuation correction may take up to 2 hours. Alternatives include elimination of attenuation correction or "splitting the body," acquiring attenuation corrected images above the inguinal regions with noncorrected images below. The usual, high metabolic rate of melanoma contributes to the excellent sensitivity (83%) and specificity (91%) for PET in this tumor (Figure 7).

Head and neck

Cancers of the head and neck, as well as therapy for these tumors, may result in high levels of morbidity. Also, it is not uncommon to identify cancer in a cervical node and be unable to find the primary site with conventional imaging studies or endoscopic examination. Positron emission tomography may be helpful in locating the primary site, staging, detecting recurrence, and monitoring therapeutic response. The accuracy (including sensitivity and specificity) of PET is significantly better than CT for diagnosis, staging, and monitoring response. However, caution must be used in interpreting PET studies in the patient following radiation therapy. ⁷ Many authors feel that increased FDG accumulation (ie, hypermetabolic activity) may persist for months following radiation. ⁸ Until more data are available, it is advisable to delay PET scanning for 6 months, if possible, in radiated patients with head and neck cancer. Most other tumors will likely require less time, but, again, these data are still preliminary. False-positive results may also occur in reactive nodes, reducing test specificity.

Breast cancer

As of October 1, 2002, the Centers for Medicare and Medicaid Services has initiated reimbursement for PET in breast cancer, including staging, restaging, and monitoring therapy. Screening will remain the purview of mammography, with PET used in patients with dense breasts or implants. Axillary lymph node staging will continue to be done by sentinel node sampling after lymphoscintigraphy. Lymphoscintigraphy is not, however, helpful for internal mammary nodes that are suspect in up to 20% of breast cancer patients, especially those with medial lesions. Staging and restaging may result in equivocal findings, necessitating a whole-body PET survey. There are data to support a PET study to replace the standard battery of conventional examinations for staging (eg, bone scan, radiographs, CT, MRI, ultrasound, biopsy). Monitoring therapy may be especially important in breast cancer patients because of the toxicity of chemotherapeutic agents and the availability of alternate drugs. Positron emission tomography is uniquely positioned to provide this information based upon the ability to characterize the metabolic response of tumor sites. These responses antedate anatomic changes by weeks to months.

Other tumors

This broad category includes tumors that can be subdivided into three groups: category 1--PET efficacy well documented; category 2--PET recommended for problem solving and in specific patients; and category 3--PET rarely indicated.

Category 1

Conditions in which PET has been shown to be effective include esophageal carcinoma and brain tumors. For patients with esophageal cancer, PET is the best way to determine metastatic spread, while staging of the primary site is performed with anatomic imaging (Figure 8). A positive PET study at distant sites may obviate unnecessary surgery. Applications of PET in brain tumor patients include differential diagnosis between recurrence (increased FDG uptake) and radiation necrosis (reduced FDG uptake), identification of potential biopsy site, monitoring response to treatment, and grading the malignancy (ie, the greater the FDG uptake, the higher the tumor grade) (Figure 9). Note that FDG does not typically accumulate in low-grade gliomas.

Category 2

Positron emission tomography is recommended for problem solving and in specific patients for almost all other tumor types. An exhaustive review of these tumors is beyond the scope of this brief report; however, based on literature series, anecdotal data, and personal experience/communication, many patients with cervical, ⁹ pancreatic, ¹⁰ testicular, ¹¹ renal, ¹² and ovarian cancer ¹³ are excellent candidates for whole-body PET. Although there are few reported cases, PET appears to be helpful in certain musculoskeletal tumors and in thyroid cancer patients with a negative I-131 whole-body survey and elevated serum thyroglobulin level. In problem solving applications, some patients presenting with an unknown primary cancer, and some with rising tumor markers and a negative standard work-up, may have positive PET findings.

Category 3

Positron emission tomography is rarely indicated in most patients with prostate cancer, neuroendocrine tumors, hepatocellular carcinoma, and, most likely, bronchoalveolar lung cancers. While research is ongoing to identify more efficacious radiopharmaceuticals for these malignancies and some authors have reported positive results with FDG, PET is recommended to answer very specific questions.

In most departments, >90% of all PET work is performed to evaluate cancer. However, there is clear evidence of the value of FDG PET in determining myocardial viability ^14,15 and in the evaluation of dementia and seizures. ¹⁶ Again, the reader is directed to the references for additional information.

Performance

A helpful, practical way to understand the performance of a whole-body PET survey is to follow the course of a typical patient from request to report, highlighting areas of importance as well as ongoing debate. Since the overwhelming majority of PET studies are ordered on outpatients, we will assume that our patient is in this group.

I. Request received

There are three major reasons to review the request prior to scheduling. First, determine whether PET scanning is appropriate. Oftentimes the tumor type is not FDG-avid (ie, prostate) or an anatomical study is preferable (eg, primary site evaluation for esophageal cancer). Second, assess the urgency of the request. Many times, treatment decisions are made based upon the results of the PET scan. This is especially true for patients initiating or undergoing potentially toxic chemotherapy. Third, if there are authorization or other insurance issues, the staff should begin to address these immediately. Support personnel must document whom they speak with, date and time of contact, and any potential problems with getting approval for the study. Additionally, the request form should indicate where and when correlative studies were performed so that these may be requested, identify diabetic or claustrophobic patients, and list all physicians who should receive a report.

II. Patient contact

Prior to the study, the patient should be contacted (and the contact should be documented) for an explanation of the basics of the test (ie, safety, duration, restrictions, etc.) and to enlist his or her help in securing all prior correlative studies conducted at other medical facilities. Most hospitals and imaging facilities are willing to loan the patient his or her films for comparison. A patient contact prior to the study can be invaluable in allaying patient anxiety and in dramatically reducing the number of "no-shows." Additional items to cover with the patient include fasting for 6 or more hours prior to the study (fluid hydration is fine), limitation of physical activity for the day before the examination (to reduce muscular uptake), documentation of recent surgery, chemotherapy, or radiation therapy, and directions to the center. A patient who is lost or is significantly late can have a major impact on the schedule.

III. Day of the study

The patient should have fasted for 6 or more hours; should wear warm, comfortable clothes; and should arrive at least 30 minutes prior to the scheduled time of the FDG injection. In these 30 minutes, the patient should complete forms, submit his or her outside studies, be taken to the uptake room, and have a blood glucose level performed. Most centers use an upper limit of 150 to 200 mg/dL. Elevated glucose levels compete with injected FDG, potentially affecting the sensitivity and quality of the scan.

A patient with a significantly elevated glucose level is best rescheduled. It is not recommended to administer insulin, since this drives FDG into the muscles. The management of a patient with known diabetes is left to the discretion of the interpreting physician, often after consultation with the patient's doctor and/or the patient. General guidelines: diet-controlled diabetics can be treated as nondiabetics (ie, fast for 6 hours); and oral agent-controlled diabetics may be placed into an early-morning time slot and instructed to withhold their pills and food until after their scan. Insulin-dependent diabetics can check their own glucose level prior to the scan (morning study preferred) and either withhold insulin until after their examination or modify/reduce their insulin dose (eg, a half dose) based upon anticipated study completion time and prior history of hypoglycemia or hyperglycemia. The patient should bring (and the center should supply) foods to counter hypoglycemia. An insulin-dependent diabetic with a glucose level <100 mg/dL prior to the study should be watched carefully and questioned for any signs or symptoms of hypoglycemia (eg, sweating, nausea, tachycardia, disorientation, etc.).

The dose of FDG is typically 10 to 15 mCi in adults and 0.14 mCi/Kg in children. This may vary depending on the type of study (eg, may be less for myocardial viability or dedicated brain examinations), patient size, and even which camera is used. Based upon system sensitivity, the dose could be less--it is recommended that the manufacturer be consulted prior to initiation of studies. The uptake phase (time from injection until time of imaging) should take place in a quiet room with minimal conversation or activity. Typically, uptake is approximately 1 hour; however, some data are accumulating to indicate that in selected tumors a longer time frame may be better (eg, breast = 1.5 to 2 hours).

Imaging

Most whole-body studies take approximately 45 to 60 minutes; less time for shorter patients, more time for taller patients and for melanoma studies that extend from the top of the skull to the bottom of the feet. A typical whole-body survey for cancer begins at mid-thigh level and goes to mid-face. Studies may be performed with or without attenuation correction (ie, transmission scan) and at this time the majority of centers perform both emission and transmission scans. ¹⁷

Debatable issues

Bladder catheterization

F-18 fluorodeoxyglucose is excreted into the bladder, and may obscure pelvic findings. Voiding prior to initiation of scanning, starting at the lower body first, and reconstruction in three planes (especially the sagittal) all can be helpful in assessing the perivesicular region. Also, a dedicated pelvic stop immediately after (or before) the whole-body study can be helpful. In some patients, none of these practical steps are successful. Some authors recommend bladder catheterization in all patients, including administration of furosemide before imaging. Others instill sterile saline to "dilute out" the radioactive urine. While we have not found this necessary as a routine, consideration should be given to catheterization in patients with known or suspected pelvic malignancies (ie, gynecologic tumors or colorectal cancers).

Laxatives

Most centers do not administer laxatives for bowel cleansing, accepting some bowel activity as a normal pattern. A word of caution--variability in distribution, intensity, and shape of bowel activity is the rule rather than the exception, including differences in the same patient from scan to scan. A tubular distribution, often confirmed by multiplane analysis, is often reassuring, while a more focal site mandates careful correlation with CT.

Sedation/muscle relaxants

We have not found muscle relaxants necessary for routine studies. However, for extremely claustrophobic patients, we have found 0.5 to 1.0 mg of Xanax to be very helpful. Claustrophobic patients can often be reassured by a visit to the scanner on a day prior to their study for viewing or even a "trial run." Muscle relaxants are rarely necessary but if the patient truly feels he or she needs one, diazepam (5 to 10 mg) may be effective. Transportation home is mandatory for any patient receiving sedatives or relaxants.

Interpretation

The basics of interpretation include knowledge of the normal distribution of FDG (eg, brain, thyroid, heart, renal collecting systems, stomach, bowel, bladder, etc.), benign variants (eg, muscular uptake, injection sites, recent wounds, inflammatory sites, etc.), and the differences between attenuation-corrected and noncorrected images. Next, comparison with anatomic imaging studies is critical. The corresponding findings on CT or MR will often confirm the PET impression, or if incongruent, suggest an appropriate clinical strategy (ie, further testing versus observation). There is no substitute for experience in reading PET studies, and many fine training opportunities are available. Choose a center that offers didactic "read-out" sessions as well as an opportunity to access their case files for maximimal exposure. Actual interpretation may be based upon visual assessment (qualitative) or may include semi-quantitative measurements, of the SUV, which is the ratio of tumor tracer concentration to the distributed body concentration. However, benign, generally very active inflammatory processes may show SUVs >2, and low-metabolic-rate tumors may demonstrate SUVs <2. There are requirements for reliable SUV determination; two important variables are proper attenuation correction (therefore, one should not perform SUV measurements on noncorrected data) and calculation based upon a standardized time (ie, 60 minutes). For visual interpretation of the chest, comparison is made to normal mediastinal uptake.

Potential pitfalls

The performance and interpretation of a PET study, as in all imaging tests, are subject to error. Most errors can be explained by careful review of all aspects of the examination. False-positive findings may be due to nontumorous processes that actively concentrate FDG, such as active inflammation, granulomatous disease, muscular uptake, normal patient variation (eg, bowel, thyroid, GU tract), radiation therapy sites, marrow activation following chemotherapy or with anemia, bone/joint uptake secondary to fracture or the arthritides, and reactive lymph nodes (Figure 10). Uptake secondary to postradiation changes may persist for months, while uptake due to che-motherapy generally lasts for weeks. An extensive listing by body regions is referenced. ¹⁸ False-negative studies may occur in patients with hyperglycemia, lesions below the level of resolution of current PET cameras (1 cm), microscopic nodal disease, low-metabolic rate tumors (eg, low-grade glioma), juxtaposition to normal FDG-avid regions, or sole reliance on an arbitrary SUV cut-off value.

Conclusion

The unique information provided by PET is a remarkable addition to our diagnostic imaging armamentarium. However, the multiple factors that must come together to produce a successful study mandate careful attention to detail and a clear understanding of each aspect of the examination, from receipt of the request to completion of scan review. AR