PET-CT Fusion. imaging in differentiating

Transcription

1 EDUCATION EXHIBIT 1411 PET-CT Fusion Imaging in Differentiating Physiologic from Pathologic FDG Uptake 1 CME FEATURE See accompanying test at /education /rg_cme.html LEARNING OBJECTIVES FOR TEST 5 After reading this article and taking the test, the reader will be able to: Describe the physiologic uptake patterns that can mimic disease at FDG PET. Discuss the advantages and limitations of PET-CT fusion imaging in differentiating normal from pathologic findings. Lale Kostakoglu, MD Ruth Hardoff, MD Rosna Mirtcheva, MD Stanley J. Goldsmith, MD Interpretation of positron emission tomographic (PET) scans in the absence of correlative anatomic information can be challenging. PET computed tomography (CT) fusion imaging is a novel multimodality technology that allows the correlation of findings from two concurrent imaging modalities in a comprehensive examination. CT demonstrates exquisite anatomic detail but does not provide functional information, whereas 2-[fluorine 18]fluoro-2-deoxy-d-glucose (FDG) PET reveals aspects of tumor function and allows metabolic measurements. Subtle findings at FDG PET that might otherwise be disregarded or interpreted as physiologic variants may lead to detection of a malignant process after being correlated with simultaneously acquired CT findings. Alternatively, equivocal CT findings, which could represent malignant tumor, reactive changes, or fibrosis, can be clarified with the help of the additional metabolic information provided by concurrent FDG PET. Accurate interpretation of FDG PET scans requires a thorough knowledge of the normal physiologic distribution of FDG and of normal variants that may reduce the accuracy of PET studies, thereby significantly affecting patient treatment. Although in rare instances PET-CT cannot help resolve the diagnostic dilemma, it is enjoying widespread acceptance in the medical imaging community, usually allowing differentiation of physiologic variants from juxtaposed or mimetic neoplastic lesions and more accurate tumor localization. RSNA, 2004 Abbreviation: FDG 2-[fluorine 18]fluoro-2-deoxy-d-glucose Index terms: Computed tomography (CT), ** Computed tomography (CT), utilization Fluorine, radioactive Positron emission tomography (PET), ** Radionuclide imaging, in diagnosis of neoplasms, ** ; 24: Published online /rg Content Codes: 1 From the Division of Nuclear Medicine, Department of Radiology, New York Presbyterian Hospital, Weill Medical College of Cornell University, 525 E 68th St, Starr No 221, New York, NY (L.K., R.M., S.J.G.); and the Department of Nuclear Medicine, Rabin Medical Center, Sackler School of Medicine, Tel-Aviv University, Tel-Aviv, Israel (R.H.). Received October 6, 2003; revision requested November 25 and received January 24, 2004; accepted January 30. All authors have no financial relationships to disclose. Address correspondence to L.K. ( lak2005@med.cornell.edu). 2 ** multiple body systems RSNA, 2004

2 1412 September-October 2004 RG f Volume 24 Number 5 Introduction Accurate evaluation of disease extent prior to therapy and of response to therapy have a significant impact on the clinical management of oncologic disorders. Positron emission tomography (PET) with 2-[fluorine 18]fluoro-2-deoxy-d-glucose (FDG) provides valuable metabolic information and has recently become an essential diagnostic modality in the staging and restaging of various cancers. Increased FDG accumulation in neoplastic tissues is a function of increased expression and activity of glucose transporter proteins and of the glucose phosphorylating enzyme hexokinase, which result from increased anaerobic metabolism in cancer cells as well as metabolic trapping of FDG within the tumor cells due to the lack of further metabolic pathways for FDG (1). Nevertheless, FDG is not specific for neoplastic processes; it accumulates physiologically in various normal organs, including the brain, muscles, salivary glands, myocardium, gastrointestinal tract, urinary tract, brown adipose tissue, thyroid gland, and gonadal tissues (2 4). FDG PET is a strictly functional modality and lacks anatomic landmarks for precise morphologic orientation. Unless anatomic correlation is available to delineate normal structures, pathologic sites of FDG accumulation can easily be confused with normal physiologic uptake, leading to false-positive or false-negative findings. This is an important shortcoming in the determination of active disease sites, particularly for small lesions and lesions located near sites of physiologic uptake. Coregistration of PET scans (functional and morphologic information) with computed tomographic (CT) scans (anatomic information) using a combined PET-CT scanner improves the overall sensitivity and specificity of information provided by PET or CT alone (4,5). The unique advantage of PET-CT fusion imaging is the ability to correlate findings at two complementary imaging modalities in a comprehensive examination. Hence, PET-CT provides more precise anatomic definition for both the physiologic and pathologic uptake seen at FDG PET. In particular, in the posttherapy period, subtle metabolic findings at FDG PET that might otherwise be overlooked may allow detection of residual disease after correlation with the simultaneously acquired CT data. Conversely, equivocal CT findings can be better evaluated with the help of the additional functional information provided by FDG PET. In this article, we discuss and illustrate normal physiologic patterns of FDG uptake and pathologic patterns of uptake at PET in various anatomic locations (head and neck, chest, abdomen and pelvis). We also discuss the utility of PET-CT in differentiating malignant processes from physiologic uptake, since in some cases such differentiation cannot be made with certainty with FDG PET alone. Sites of Physiologic FDG Uptake There are several sites of normal physiologic accumulation of FDG (Fig 1). FDG accumulation is most intense in the cerebral cortex, basal ganglia, thalamus, and cerebellum, since the brain is exclusively dependent on glucose metabolism. The myocardium expresses insulin-sensitive glucose transporters, which facilitate the transport of glucose into muscle. Although the myocardium uses free fatty acids as its primary substrate, in the nonfasting state the antilipolytic effect of insulin reduces the delivery of free fatty acids, and the heart relies more on glycolytic metabolism. A recent meal often causes intense myocardial FDG uptake because of the associated elevated serum insulin levels. Fasting for 4 6 hours before FDG administration decreases the availability of both glucose and insulin in the circulation, which usually leads to decreased accumulation of FDG within the myocardium.

3 RG f Volume 24 Number 5 Kostakoglu et al 1413 Figure 1. Normal distribution of FDG. Coronal CT (a), PET (b), and PET-CT fusion (c) images demonstrate the physiologic accumulation of FDG in the cerebral-cerebellar cortex at the base of the skull and in the myocardium, liver, kidneys, renal pelvis, bone marrow, and urinary bladder. Note also the minimal uptake in the mediastinum and bilaterally in the lower cervical and psoas muscles. Because FDG appears in the glomerular filtrate and, unlike glucose, is not reabsorbed in the tubules, intense FDG activity is seen in the intrarenal collecting systems, ureters, and bladder. Less intense radiotracer activity is present in the liver, spleen, bone marrow, and renal cortex. At 1 hour after radiotracer injection, blood pool activity results in moderate background activity in the mediastinum, whereas lung activity is low. Significant muscle uptake is observed in the skeletal muscles with exercise, in the breathing muscles with hyperventilation, in the cervical muscles with tension, and in the laryngeal muscles with vocalization. Uptake in lymphatic tissues and salivary glands may also be seen as a normal variant. Uptake in the gastrointestinal tract is variable. Normal stomach, small intestine, and colon may demonstrate increased FDG uptake due to a combination of factors, including smooth muscle contraction and metabolically active mucosa (2). FDG uptake in bone marrow is normally modest. Patients undergoing treatment with granulocyte-stimulating factors have diffuse intense FDG uptake in the bone marrow (6).

4 1414 September-October 2004 RG f Volume 24 Number 5 Figure 2. Contraction-induced FDG uptake in a 50-year-old woman with advanced non small cell lung cancer who was referred for presurgical evaluation. (a) Axial FDG PET scans demonstrate multiple hypermetabolic foci in the anterior cervical region. Intense symmetric and superficial uptake in the anterior neck (arrowheads) suggests tense sternocleidomastoid muscle uptake. An additional asymmetric focus of uptake posterior to the right sternocleidomastoid muscle (arrows) suggests cervical lymph node metastasis, which represents a more advanced disease stage and renders the patient ineligible for surgery until after neoadjuvant chemotherapy. (b, c) Axial CT (b) and PET-CT fusion (c) images demonstrate that the unilateral uptake in the right cervical region corresponds to the anterior scalene muscle (long arrow) and that the superficially located symmetric FDG uptake corresponds to the sternocleidomastoid muscles (short arrows). Evaluation with PET-CT excluded the possibility of a false-positive finding (ie, metastatic disease in the neck), which would subject the patient to unnecessary chemotherapy. Physiologic versus Pathologic FDG Uptake Head and Neck Although muscle uptake anywhere in the body may make the interpretation of FDG PET scans difficult, the abundance of small muscle groups in the neck constitutes a diagnostic dilemma, particularly in patients with head and neck neoplasms. Differentiation of physiologic muscle uptake from pathologic uptake is even more critical in the posttherapy follow-up period. Moderate to high FDG uptake is noticeable in the muscles, including the ocular muscles, and may be a potential source of false-positive findings in patients with malignant head and neck tumors and central nervous system tumors. Even without the help of CT, the origin of such FDG uptake is usually obvious due to the symmetric uptake pattern and the typical anatomic location. Contraction-induced FDG uptake in cervical muscles in tense patients can be confused with lymph node metastasis or, alternatively, may lead to false-negative findings of disease in the underlying lymph nodes (Fig 2), which constitutes a serious problem in patients with asymmetric muscle uptake due to prior neck dissection (Fig 3). Physiologic FDG uptake in the normal thyroid gland is usually absent or minimal, whereas in adenomas, uptake can be as high as that which is observed in malignant processes (7). In the absence of correlative imaging with an anatomic modality, focal uptake in the thyroid gland can be falsely interpreted as metastatic disease in the lower cervical lymph node stations (Figs 4, 5).

5 RG f Volume 24 Number 5 Kostakoglu et al 1415 Figure 3. Asymmetric muscle uptake in a 62-year-old woman with squamous cell carcinoma of the tongue who had undergone neck dissection. (a) Axial FDG PET scans demonstrate multiple hypermetabolic foci in the right anterolateral cervical region. Asymmetric and superficial uptake in the right anterior portion of the neck (arrowheads) suggests tense sternocleidomastoid muscle uptake. An additional asymmetric focus of uptake is noted posterior to the right sternocleidomastoid muscle (long arrow) and suggests cervical lymph node metastasis, although unilateral cervical muscle uptake cannot be excluded (cf Fig 4). Symmetric foci seen within the larynx (short arrows) probably represent the intrinsic laryngeal muscles. (b, c) Axial CT (b) and PET-CT fusion (c) images help confirm that the superficial asymmetric FDG uptake in the anterior cervical region corresponds to the right sternocleidomastoid muscle (arrowhead). The absence of contralateral muscle uptake is due to prior neck dissection. The unilateral focus of uptake posterior to the right sternocleidomastoid muscle (long arrow) corresponds to a cervical lymph node, and symmetric foci within the larynx (short arrows) correspond to the intrinsic laryngeal muscles and cricoarytenoid muscles posteriorly. The results of simultaneous evaluation with PET and CT confirmed the diagnosis of (advanced) metastatic disease in the neck and had a significant impact on clinical management. The patient was started on therapy on the basis of the PET-CT findings. Figure 4. Thyroid carcinoma in a 45-year-old woman with a history of breast cancer who was referred for posttherapy evaluation. Axial PET scan (b) reveals a hypermetabolic focus in a cervical lymph node in the left side of the neck (arrowhead) that is highly suspicious for metastatic disease. (a, c) Axial CT (a) and PET-CT fusion (c) images demonstrate that the focus (arrowhead) corresponds to the left thyroid lobe and is consistent with a thyroid nodule. Subsequent ultrasonography demonstrated a multinodular gland with a dominant nodule in the left lower pole. Further investigation with needle biopsy revealed follicular carcinoma of the thyroid gland. PET-CT findings confirmed that the patient had thyroid disease (adenoma or carcinoma) rather than lymph node metastasis; however, the nodule could not be characterized on the basis of PET-CT findings alone.

6 1416 September-October 2004 RG f Volume 24 Number 5 Figure 5. Thyroid nodule in a 51-year-old man with a history of squamous cell carcinoma of the tonsil who was referred for postsurgical evaluation. Axial PET scan (b) reveals a hypermetabolic focus in the right side of the neck (arrow) that is highly suspicious for metastatic disease. (a, c) Axial CT (a) and PET-CT fusion (c) images demonstrate that the focus (arrow) corresponds to the right thyroid lobe and is consistent with a thyroid nodule, a finding that was confirmed with subsequent ultrasonography. PET-CT findings confirmed that the patient did not have lymph node metastasis; however, a thyroid nodule requires further investigation with biopsy to rule out a malignant cause within the thyroid gland. Figure 6. Primary tumor of the larynx in a 45-year-old man with epiglottic carcinoma who was referred for presurgical evaluation. Axial CT (a), PET (b), and PET-CT fusion (c) images show a focus of uptake (arrow) that corresponds to a mass that originates from the larynx and nearly obliterates the lumen on the CT scan. Correlation with CT and PET-CT helped greatly in establishing the diagnosis of primary laryngeal malignancy rather than activated laryngeal muscles. In cases of occult head and neck tumor, differentiation may be difficult without CT correlation. It is essential that the patient remain silent during the period of FDG uptake. Figure 7. Physiologic laryngeal uptake in a 52-year-old woman with squamous cell carcinoma of the floor of the mouth who had undergone surgery and radiation therapy. Axial CT (a), PET (b), and PET-CT fusion (c) images clearly demonstrate a midline focus of uptake (arrow) that corresponds to the cricoarytenoid muscles located posterior to the thyrocricoid cartilage. Although this site and pattern of uptake are typical for activated vocal cords, CT helps confirm the physiologic nature of the FDG accumulation. Without CT correlation, this focus of uptake may lead to a false-positive finding in patients with malignant processes in this location, particularly in the posttherapy setting.

7 RG f Volume 24 Number 5 Kostakoglu et al 1417 Figures 8, 9. (8) Extranodal disease in a 33-year-old man with a history of aggressive non-hodgkin lymphoma who had undergone chemotherapy 1 year earlier. Axial PET scan (b) demonstrates intense FDG uptake in the posterosuperior oral cavity (arrows). (a, c) Axial CT (a) and PET-CT fusion (c) images demonstrate that the uptake (arrows) corresponds to the palatine tonsils, with minimal asymmetry seen at CT. In the tonsillar lymphatic tissues, physiologic uptake cannot be differentiated from extranodal lymphoma, especially in cases of subtle CT findings. (9) Lymph node metastasis in a 48-year-old man with newly diagnosed squamous cell cancer of the tonsil. Axial PET scan (b) demonstrates intense FDG uptake in the region of the left tonsil (arrow) and in a left cervical lymph node (arrowhead), findings that are consistent with primary and metastatic disease, respectively. (a, c) Axial CT (a) and PET-CT fusion (c) images help confirm that the uptake corresponds to the left palatine tonsil (arrow) and left jugular lymph node (arrowhead). The asymmetric nature of the tonsillar uptake and the presence of lymph node metastasis allow definitive diagnosis. The intrinsic laryngeal muscles serve both as a sphincter and as an organ of phonation. FDG accumulates in the striated laryngeal muscles in proportion to contractile activity during speech. This phenomenon is a major concern and may lead to false readings in patients with head and neck cancers (Figs 6, 7) (8,9). A rigorous approach to preventing physiologic FDG uptake in the laryngeal muscles should be adopted to avoid false-positive findings. In practice, patients should be encouraged to remain silent beginning 15 minutes prior to injection and continuing until the study is completed. Low to moderate FDG uptake occurs in the lymphatic tissues in the pharynx, consisting of the nasopharyngeal, palatine, and lingual tonsils (Waldeyer ring) (Figs 8 10 ) (4). The Waldeyer ring is a common site of head and neck manifestations of extranodal non-hodgkin lymphoma (Fig 8). Furthermore, primary squamous cell carcinoma of the head and neck may occur within the crypts of the Waldeyer ring (Fig 9). In cases of tonsillar lymphoma, the asymmetric nature of FDG uptake suggests a pathologic process (Fig 10). However, in the absence of CT guidance, malignant processes arising from the lymphatic tissues may be difficult to identify at FDG PET.

8 1418 September-October 2004 RG f Volume 24 Number 5 Figure 10. Tonsillar lymphoma in a 20-year-old man with Burkitt lymphoma of the abdomen who was referred for posttherapy evaluation. Axial PET scan (b) demonstrates intense FDG uptake in the region of the tonsils with minimal asymmetry (arrows), a finding that is suspicious for extranodal lymphoma in the regional lymphatic tissues. (a, c) Axial CT (a) and PET-CT fusion (c) images help confirm that the uptake (arrows) corresponds to the palatine tonsils. Furthermore, asymmetric uptake in the left tonsil is seen as a mass that bulges into the lumen on the CT scan. The minimal asymmetry of the tonsillar uptake suggests disease; however, definitive diagnosis cannot be made on the basis of PET-CT findings alone. Subsequent biopsy revealed Burkitt lymphoma of the tonsils. Low to moderate FDG uptake is noted in the salivary glands, most prominently in the floor of the mouth. In the saliva, FDG concentration has been observed to be higher than physiologic glucose content (10). The difference between physiologic expected values and observed values may reflect a difference between the reabsorption process for glucose and that for FDG, similar to the processing of FDG by the renal tubules. Symmetric diffuse uptake in the parotid glands is usually physiologic, whereas focal and heterogeneous uptake is suggestive of a malignant process. Chest The thymus lies in the upper part of the mediastinum anterior to the great vessels and extends upward into the neck. Involution of the gland begins in adolescence. The proposed mechanism for therapy-related thymic rebound includes initial thymic regression secondary to chemotherapy or corticosteroids that trigger apoptosis of T cells and thymocyte death, with subsequent rebound on reversal of the predisposing condition (11,12). Moderate to high FDG uptake is noted in patients with thymic rebound (Fig 11) and should not be confused with asymmetric uptake due to lymphoma in this location (Fig 12). In pediatric patients, anatomic correlation is necessary following chemotherapy to differentiate the enlarged thymus from residual or recurrent disease at this location, especially with focal thymic uptake (Fig 13). High FDG uptake is seen in the brown adipose tissue in the supraclavicular regions, midaxillary line, and paraspinal regions in the posterior mediastinum (13,14). In contrast to other tissue, brown adipose tissue expresses the mitochondrial uncoupling protein, which allows the cell mitochondria to uncouple oxidative phosphorylation and generate heat rather than adenosine triphosphate (15). Metabolism within the brown adipose tissue is increased by means of anaerobic metabolism to prevent this highly metabolic tissue from becoming adenosine triphosphate deficient. In patients with increased activity in the brown adipose tissue, usually pediatric patients and females, symmetric FDG uptake can mimic pathologic uptake (13,14). Simultaneously acquired PET and CT scans allow delineation of normal anatomic structures and precise localization of FDG uptake. Discordant findings in which in-

9 RG f Volume 24 Number 5 Kostakoglu et al 1419 Figures (11) Thymic rebound in a 25-year-old woman with a history of Hodgkin disease who was referred for posttherapy evaluation. Axial PET scan (b) demonstrates findings that are consistent with posttherapy thymic rebound (arrows). (a, c) Axial CT (a) and PET-CT fusion (c) images reveal a symmetric hypermetabolic focus (arrows) that corresponds to the thymus and appears as a bilobed structure with convex lateral borders in the anterosuperior mediastinum at CT. (12) Lymphoma in a 36-year-old man with a history of Hodgkin disease who was referred for posttherapy evaluation for possible residual disease. Axial CT (a), PET (b), and PET-CT fusion (c) images demonstrate a convex right mediastinal hypermetabolic focus (arrow) that is suspicious for residual disease. Although the unilateral manifestation and midmediastinal location of this focus and the presence of associated other abnormal foci of mediastinal uptake do not support thymic rebound, the pattern of uptake resembles thymic uptake. This pattern may be confusing in certain clinical settings, particularly when the pathologic uptake is located more anteriorly in the mediastinum. (13) Physiologic thymic uptake in a 23-year-old woman with a history of Hodgkin disease of the chest who was referred for posttherapy evaluation. Axial PET scan (b) demonstrates a unilateral focus of FDG uptake in the superoanterior mediastinum (arrow), a finding that is suspicious for residual disease. Although the uptake is located in the thymic region, its midmediastinal location may suggest viable lymphoma (cf Fig 12). (a, c) Axial CT (a) and PET-CT fusion (c) images help confirm that the anterosuperior mediastinal uptake (arrow) corresponds to thymic tissue, thereby excluding the possibility of residual disease. Although PET-CT is convenient for immediate correlation of images obtained in the same anatomic planes, a recent, separately acquired CT scan would also be helpful in confirming thymic uptake. tense FDG uptake is seen on high-quality PET scans but no corresponding abnormality is seen at CT suggest brown adipose tissue (Fig 14). Because brown adipose tissue has rich sympathetic innervation, administration of diazepam can prevent observed uptake in the brown adipose tissue by means of a sympathetic nervous system blockade (16). However, -blockers of the sympathetic nervous system may be more useful in inhibiting FDG uptake in the brown adipose tissue.

10 1420 September-October 2004 RG f Volume 24 Number 5 Figure 14. Residual disease in a 36-year-old man with a history of Hodgkin disease who was referred for postchemotherapy evaluation. (a c) Coronal PET scan (b) shows symmetric hypermetabolic foci in the cervical and supraclavicular regions (small arrows). In addition, there is asymmetric uptake in the right axillary region (large arrow). Coronal CT (a) and PET-CT fusion (c) images help confirm persistent lymphoma in the right axillary lymph nodes (large arrow). No abnormality is seen in the cervical supraclavicular regions that corresponds to the FDG uptake in these regions (small arrows), a finding that is consistent with brown adipose tissue. (d f) Axial PET scan (e) demonstrates multiple foci of increased FDG uptake in the cervical region. There is also a unilateral focus of uptake in the right jugular region (arrow) that is suspicious for malignancy. CT (d) and PET-CT fusion (f) images help confirm that the focal FDG uptake on the right side (arrowhead) corresponds to a jugular lymph node, a finding that is consistent with residual Hodgkin disease. The uptake in the brown adipose tissue renders interpretation difficult by obscuring the underlying lymph nodes that harbor viable residual disease. It is essential that this pattern of uptake be evaluated simultaneously with CT. Low to moderate FDG uptake is noted in the distal esophagus, particularly in patients with gastroesophageal reflux secondary to inflammatory changes (Fig 15). In some patients, activity in the distal esophagus without supporting evidence of disease is observed due to muscle contraction at the gastroesophageal junction or inflammatory changes caused by gastroesophageal reflux. However, high-grade uptake in the same location may be caused by malignant processes (ie, carcinoma of the distal esophagus, usually associated with morphologic esophageal changes) (Fig 16). In some malignant lesions that demonstrate subtle FDG uptake and equivocal morphologic changes such as lack of significant mucosal and wall thickening, correlative imaging may not be useful in characterizing the function and morphologic features of the lesions. Abdomen and Pelvis Coregistration of simultaneously acquired PET and CT scans is particularly helpful in the abdomen and pelvis. PET-CT provides significantly improved localization of abnormal or unexpected FDG findings compared with PET or CT alone. Moderate to high FDG uptake is visible in the muscles that contribute to breathing in patients with chronic obstructive pulmonary disease due to difficulty in breathing and use of accessory muscles to facilitate breathing. In addition, due to an imbalance between oxygen supply and increased demand, the decrease in oxygen delivery causes a switch to anaerobic metabolism (17). Hence, the increased uptake seen in the diaphragmatic cruces may be the result of accentuated abdominal breathing effort and the anaerobic metabolism that leads to increased FDG uptake similar to the physiologic alterations in cancer

11 RG f Volume 24 Number 5 Kostakoglu et al 1421 Figures 15, 16. (15) Nonneoplastic esophageal uptake in a 21-year-old woman with a history of non- Hodgkin lymphoma who was referred for restaging following therapy. Axial PET scan (b) shows a focus of metabolic activity in the midline in the lower chest, inferior to the heart and adjacent to the abdominal aorta (arrowhead). A malignant process in a lymph node in this region cannot be excluded on the basis of PET findings alone. (a, c) Axial CT (a) and PET-CT fusion (c) images reveal that the focal FDG uptake in the inferoposterior mediastinum (arrowhead) corresponds to the distal esophagus. This focus of uptake is probably due to physiologic muscle uptake at the gastroesophageal junction or inflammatory changes secondary to gastroesophageal reflux. No tumor, wall thickening, or lymphadenopathy is appreciated in this region. (16) Esophageal adenocarcinoma in a 52-year-old man who was referred for presurgical evaluation. Axial PET scan (b) demonstrates an intense focus of metabolic activity (arrowhead) (cf Fig 15). If no clinical information were available, it would be unclear whether this focus represented physiologic FDG uptake in the distal esophagus or a lymphatic or esophageal malignancy. (a, c) Axial CT (a) and PET-CT fusion (c) images show significant thickening of the distal esophagus that almost obliterates the lumen (arrowhead). This finding corresponds to the intense uptake seen at PET and is consistent with esophageal adenocarcinoma. Figure 17. Physiologic diaphragmatic uptake in a 49-year-old woman with a history of abdominal lymphoma and severe chronic obstructive pulmonary disease who was referred for posttherapy follow-up. Axial PET scan (b) demonstrates bilateral hypermetabolic foci in the upper midabdomen that are slightly more prominent on the right side than on the left (arrowheads). These foci of uptake may be attributed to recurrent lymphoma in the regional abdominal lymph nodes. (a, c) Axial CT (a) and PET-CT fusion (c) images demonstrate that these foci (arrowheads) correspond to the diaphragmatic cruces. The combined use of PET and CT in this case allowed definitive exclusion of relapse of lymphoma. cells. Any disease process involving the celiac or perigastric lymph nodes (eg, lymphoma, nodal metastatic disease) can be difficult to interpret in patients with diaphragmatic uptake, especially in the posttherapy setting. Nevertheless, PET-CT is particularly helpful in this group of patients in distinguishing physiologic uptake in the diaphragmatic cruces from neoplastic processes (Figs 17, 18).

12 1422 September-October 2004 RG f Volume 24 Number 5 Figure 18. Recurrent disease in a 56-year-old man with esophageal carcinoma. The patient had undergone esophagectomy and was referred for follow-up evaluation. Axial PET scan (b) demonstrates a hypermetabolic focus left of the midline in the upper abdomen (arrow) (cf Fig 17). This focus is consistent with metastatic esophageal carcinoma in the regional lymph nodes or unilateral uptake in the left diaphragmatic crux. (a, c) Axial CT (a) and PET-CT fusion (c) images demonstrate that the hypermetabolic focus seen at PET corresponds to an enlarged lymph node (arrow), a finding that is consistent with recurrent disease. The patient subsequently underwent chemotherapy on the basis of the PET-CT findings. Figures 19, 20. (19) Physiologic gastric uptake in a 52-year-old man with colorectal cancer who had undergone surgical tumor resection. (a) Coronal PET scans demonstrate a diffuse hypermetabolic focus in the left upper quadrant (arrow) that corresponds to the stomach or the transverse colon. (b d) Coronal CT (b), PET (c), and PET- CT fusion (d) images help confirm that the focus of activity (arrow) corresponds to physiologic uptake in the stomach. (20) Gastritis in a 47-year-old woman with a history of breast cancer. (a) Coronal PET scans demonstrate a hypermetabolic area in the left upper quadrant of the abdomen in the stomach region (arrow). (b d) Coronal PET scan (c) shows linear activity (black arrow). Coronal CT (b) and PET-CT fusion (d) images demonstrate that this area of uptake (white arrow) corresponds to the contour of the stomach. It was later confirmed that the patient suffered from gastritis at the time of the PET-CT study. Note the asymmetric uptake in the left breast, a finding that is consistent with breast cancer. Low to moderate uptake is usually observed in the stomach. The diffuse uptake pattern, typically located in the fundus, is rarely confused with pathologic FDG uptake. However, focal FDG accumulation can give rise to misinterpretations in the absence of correlative imaging, which provides exquisite delineation of the local anatomy,

13 RG f Volume 24 Number 5 Kostakoglu et al 1423 Figure 21. Newly diagnosed gastric cancer in a 59-year-old woman who was referred for presurgical evaluation. (a) Coronal PET scans demonstrate a hypermetabolic focus in the left upper quadrant (arrow) in the region of the stomach, transverse colon, or lymph nodes. (b d) Coronal PET scan (c) demonstrates irregular FDG uptake (black arrow). Coronal CT (b) and PET-CT fusion (d) images show that this uptake (white arrow) corresponds to a mass that originates from the antral portion of the stomach and nearly obliterates the lumen, a finding that is consistent with gastric cancer. PET-CT correlation helped establish the diagnosis of gastric cancer with no metastasis in the locoregional lymph nodes. Figure 22. Physiologic bowel uptake in a 36-year-old man with malignant thymoma who had undergone surgical tumor resection. Coronal PET scan (b) demonstrates two hypermetabolic foci in the right lower quadrant (arrows). These foci may represent physiologic bowel activity; however, mesenteric lymph node metastasis or synchronous colon cancer cannot be definitively excluded. (a, c) Coronal CT (a) and PET-CT fusion (c) images reveal that the hypermetabolic foci seen at PET correspond to small bowel loops (arrows). Simultaneous evaluation with CT helps exclude the possibility of lymph node metastasis or synchronous colonic malignancy. Although it is rare, small bowel malignancy is still a possibility, in which case correlation with CT would not be helpful. including the lymph nodes, liver, and pancreas (Figs 19, 20). Focal and irregular uptake in the stomach is usually due to a malignant process; nevertheless, local gastritis cannot be excluded with certainty without the help of CT (Fig 21) (18). The importance of FDG PET in the evaluation of colorectal cancer is well established. Both small and large bowel may demonstrate varying degrees of FDG uptake, usually with a diffuse and linear pattern. However, focal physiologic uptake is not an uncommon finding in short segments of the bowel (Figs 22, 23). Unless CT correlation is

14 1424 September-October 2004 RG f Volume 24 Number 5 Figure 23. Physiologic bowel uptake in a 44-year-old man with squamous cancer of the oropharynx who was referred for posttherapy evaluation. Coronal PET scan (b) reveals hypermetabolic foci in the right lateral midabdomen near the hepatic flexure (short arrow) and in the region of the renal pelvis (long arrow). These foci may represent mesenteric lymph node metastasis, synchronous colon cancer, or physiologic bowel uptake. (a, c) Coronal CT (a) and PET-CT fusion (c) images demonstrate that the uptake corresponds to a bowel loop (white arrow in a, short arrow in c). In addition, the superior focus of uptake in the region of the renal pelvis (black arrow in a, long arrow in c) is most consistent with the ureter. PET-CT correlation helped exclude the possibility of mesenteric lymph node metastasis; however, a second primary tumor in the colon could not be excluded with this study. Figure 24. Primary carcinoid tumor of the bowel in a 47-year-old woman with a history of breast cancer and a recent diagnosis of metastatic carcinoid tumor in the lung. Coronal PET scan (b) demonstrates an intense hypermetabolic focus in the right lower quadrant (arrow), a finding that may represent physiologic bowel uptake, carcinoid tumor in the bowel, or mesenteric lymph node. The faint focus in the right lower lung (arrowhead) is consistent with the known carcinoid metastasis. (a, c) Coronal CT (a) and PET-CT fusion (c) images reveal that the hypermetabolic focus in the right lower quadrant (arrow) corresponds to the bowel. This finding is highly suspicious for a primary carcinoid tumor, which was later confirmed at biopsy. Evaluation with PET-CT helped characterize the FDG uptake in this location, which would otherwise be less specific. available, the configuration of uptake in these cases may be indistinguishable from malignant processes (Figs 24, 25) (19). Nevertheless, if the malignant lesion is not well defined at CT, the guidance obtained from CT may not be sufficient. Gallbladder uptake of FDG is not a common finding. In our limited experience, patients in whom the gallbladder is visualized at FDG PET have described postprandial discomfort, a symptom that suggests a chronic disorder such as chronic cholecystitis (Fig 26). When activity is

15 RG f Volume 24 Number 5 Kostakoglu et al 1425 Figure 25. Adenocarcinoma of the cecum in a 77-year-old man with a cecal polyp that had recently been detected at colonoscopy. The histologic features of the lesion were consistent with tubular adenoma with highgrade dysplasia. The patient was referred for assessment of the metabolic status of the lesion. Coronal PET scan (b) demonstrates a hypermetabolic focus in the right lower quadrant (arrow) that corresponds to the cecal area and is consistent with either physiologic bowel uptake or a malignant process. (a, c) Coronal CT (a) and PET-CT fusion (c) images help confirm that this hypermetabolic focus (arrow) corresponds to the cecum. This focus was confirmed to be adenocarcinoma at histologic analysis. Although PET-CT helped localize the focus, physiologic uptake in the colon could not be excluded with certainty. In the absence of clinical information and biopsy, PET-CT may still fall short in differentiating physiologic bowel uptake from a malignant process. Figure 26. Chronic cholecystitis in a patient with papillary thyroid cancer who underwent thyroidectomy and radioiodine ablation. Findings at iodine-131 whole-body scintigraphy were negative, but the patient was referred for evaluation of elevated thyroglobulin levels. (a) Coronal PET scans demonstrate a slightly hypermetabolic focus in the lower medial aspect of the liver (arrow) that is suggestive of liver metastasis. (b d) Axial PET scan (c) demonstrates a focus of uptake (black arrow). CT (b) and PET-CT fusion (d) images reveal that the focus (white arrow) localizes to the gallbladder, thereby excluding the possibility of liver metastasis. Note the intense bilateral uptake in the renal collecting system. The patient subsequently experienced postprandial discomfort, raising the possibility of chronic cholecystitis, although further tests for a definitive diagnosis were not performed.

16 1426 September-October 2004 RG f Volume 24 Number 5 Figure 27. Liver metastasis in a 55-year-old man with rectal adenocarcinoma who was referred for posttherapy evaluation. (a) Coronal PET scans demonstrate a hypermetabolic focus in the lower medial aspect of the liver (arrow) (cf Fig 32). This lesion is suggestive of liver metastasis but can also be attributed to gallbladder activity. (b d) Axial PET scan (c) shows a focus of uptake in the region of the gallbladder fossa (black arrow). Axial CT (b) and PET-CT fusion (d) images show that this focus (white arrow) is in fact within the hepatic parenchyma. All three images also demonstrate a hypermetabolic lesion (arrowhead) that is consistent with liver metastasis. observed in this anatomic location, choleductal cancer, adenocarcinoma of the gallbladder, and primary or metastatic disease of the liver should be considered in the differential diagnosis (Fig 27). CT correlation is most helpful in delineating anatomic landmarks and distinguishing a benign gallbladder variant from juxtaposed malignant lesions. Osteophytes are an outgrowth or excrescence of bone and usually develop in areas of joints that are subject to low stress. Osteophytes can demonstrate FDG uptake, which depends on the degree of metabolic activity. Although osteophytes can occur at any level of the vertebral column, those that are located anteriorly may be confused with the paravertebral lymph nodes in the absence of correlative imaging (Fig 28). Unlike glucose, FDG is not reabsorbed by the renal tubules after filtration. Thus, significant FDG accumulation is seen in the intrarenal collecting system and renal pelvis. This accumulation may interfere with the identification of renal parenchymal or pelvic urothelial tumors. However, contemporaneous anatomic information provided by CT allows proper assessment and characterization of renal masses (Fig 29). Focal FDG accumulation in the ureters is a common finding due to the pooling of radiotracer in the recumbent patient, although the intensity and location of uptake usually allow accurate identification of the ureters in patients with abdominal malignancies. This finding can be misdiagnosed as pelvic lymph node metastasis or nodal lymphoma (Figs 30, 31). Unrecognized renal transplants may also lead to false-positive findings (20). Although the absence of the native kidneys should be alarming, simultaneously acquired CT scans delineate the anatomy and help avoid falsepositive findings (Fig 32).

17 RG f Volume 24 Number 5 Kostakoglu et al 1427 Figure 28. Osteophyte in a 65-year-old man with a history of colorectal cancer who was referred for posttherapy evaluation. (a, b) Axial PET scans demonstrate a small focus of FDG uptake in the right paravertebral region (arrow), a finding that is consistent with a mesenteric or paravertebral lymph node. Note also the uptake in the midabdomen (white arrowhead), a finding that is suggestive of a malignant process in the mesenteric lymph nodes. The faint uptake in the anterior abdominal wall (black arrowhead) is consistent with postsurgical changes. (c e) Axial PET scan (d) demonstrates a focus of uptake in the right paravertebral region (black arrow). Axial CT (c) and PET-CT fusion (e) images clearly show that this focus (white arrow) corresponds to an osteophyte in the lateral portion of the vertebral body. The midline abdominal uptake seen on all three images (arrowhead) is consistent with physiologic bowel uptake. PET-CT provided valuable anatomic information and helped exclude the possibility of metastatic disease in the paravertebral and mesenteric lymph nodes in the midabdomen. Figure 29. Renal cell carcinoma in a 60-year-old woman with a newly diagnosed renal mass who was referred for presurgical evaluation. (a, b) Axial PET scans demonstrate intense FDG uptake in the midportion of the right kidney (arrow), a finding that is indistinguishable from physiologic renal uptake. (c e) Axial PET scan (d) shows uptake in the right kidney (black arrow). Axial CT (c) and PET-CT fusion (e) images help confirm that this focus (white arrow) corresponds to the known renal mass, which is most consistent with renal carcinoma. In the absence of correlative CT, this pattern of uptake can be misinterpreted as physiologic uptake in the intrarenal collecting system. Although evaluation of renal masses with FDG PET is difficult, CT provides valuable information for accurate interpretation of PET findings.

18 1428 September-October 2004 RG f Volume 24 Number 5 Figures 30, 31. (30) Physiologic uptake in the renal pelvis in a 66-year-old man with a history of colorectal cancer who had undergone surgery and chemotherapy. Axial PET scan (b) demonstrates a focus of intense FDG uptake in the left upper quadrant (arrow). This focus of uptake probably represents the renal pelvis; however, mesenteric lymph node involvement cannot be definitively excluded. A focus of increased FDG uptake in the lateral aspect of the liver (arrowhead) is consistent with liver metastasis. (a, c) Axial CT (a) and PET-CT fusion (c) images help confirm that the focal uptake in the left upper quadrant (arrow) corresponds to the renal pelvis, thereby excluding the possibility of metastatic disease. (31) Recurrent nodal disease in a 49-year-old man with a history of abdominal lymphoma who had undergone chemotherapy. Axial PET scan (b) demonstrates a focus of intense FDG uptake in the left upper quadrant near the spleen (arrow) (cf Fig 30), a finding that is consistent with the renal pelvis or recurrent lymphoma in the splenic hilum. (a, c) Axial CT (a) and PET-CT fusion (c) images reveal that the focus of uptake (arrow) corresponds to the splenic hilum, a finding that is consistent with recurrent lymphoma in the regional lymph nodes. There is usually no FDG accumulation in the uterus, although focal FDG uptake in the uterus during menstruation has been described (21). FDG accumulation during menstruation can be attributed to heavy bleeding or to necrotic endometrial epithelium due to sudden reduction of estrogen and progesterone levels at the end of the secretory phase of the menstrual cycle. It may not be possible to differentiate this uptake pattern from uterine carcinoma, even with the help of PET-CT (Figs 33, 34). However, FDG uptake is usually more irregular, diffuse, and extensive in uterine cancer (Fig 35). Conclusions The primary advantage of PET-CT fusion technology is the ability to correlate findings from two concurrent imaging modalities in a comprehensive examination that combines anatomic data with functional and metabolic information. CT demonstrates exquisite anatomic detail but does not provide functional information, whereas FDG PET lacks anatomic landmarks but reveals aspects of tumor function and allows metabolic measurements. Physiologic FDG uptake in nonmalignant conditions limits the specificity of PET, particularly in the posttherapy setting. Hybrid PET-CT scanners allow PET and CT image fusion for differentiation of physiologic variants from juxtaposed or mimetic neoplastic lesions and more accurate tumor localization. Softwarebased fusion of separately acquired PET and CT scans is more likely to lead to misregistration due to independent parameters and differences in patient positioning (22,23). In addition, CT permits rapid acquisition of attenuation correction data for the PET scan. Charron et al (24) compared combined PET-CT scans of different cancers with PET scans alone. In 31% of cases, variable amounts of normal physiologic FDG uptake were distinguished from pathologic FDG uptake. In our experience, the majority of cases in which PET-CT

19 RG f Volume 24 Number 5 Kostakoglu et al 1429 Figures (32) Bone and bone marrow involvement in a 42-year-old woman with a recent diagnosis of lymphoma. (a c) Coronal PET scan (b) demonstrates an area of intense FDG uptake in the left side of the pelvis (black arrow), a finding that is consistent with lymphomatous involvement in the soft tissues of the pelvis or a transplanted kidney. Coronal CT (a) and PET-CT fusion (c) images demonstrate uptake in the left lower quadrant (white arrow), thereby helping confirm that the uptake is in the renal collecting system of a transplanted kidney. (d f) Axial PET scan (e) shows uptake in the anterior left side of the pelvis (black arrow). Axial CT (d) and PET-CT fusion (f) images help further confirm that this uptake (white arrow) corresponds to a transplanted kidney. In addition, the PET and PET-CT fusion images show multiple hypermetabolic foci in the osseous structures of the pelvis (arrowhead), a finding that is consistent with lymphomatous involvement. PET-CT helped differentiate the renal transplant from lymphoma and helped identify other sites of lymphomatous involvement in the osseous structures. (33) Physiologic uterine uptake in a 40-year-old woman with a history of lymphoma who was referred for posttherapy evaluation. Axial PET scan (b) demonstrates a hypermetabolic focus in the posterior midline pelvic region (arrow), a finding that is consistent with physiologic FDG uptake in the bowel or colorectal malignancy. (a, c) Axial CT (a) and PET-CT fusion (c) images demonstrate that this focus (arrow) corresponds to a retroverted uterus. Further inquiry revealed that the patient was menstruating at the time of imaging. (34) Uterine carcinoma in a 45-year-old woman with a history of colorectal cancer. Axial PET scan (b) demonstrates a hypermetabolic focus in the anterior right pelvic region (arrow), a finding that is consistent with physiologic FDG uptake in the bowel, uterine malignancy, or physiologic uptake in the uterus during menstruation. (a, c) Axial CT (a) and PET-CT fusion (c) images demonstrate that this focus (arrow) corresponds to the anterior portion of the uterus. Results of laparoscopy confirmed malignant invasion of the uterus by colorectal cancer. Although PET-CT helped localize the uptake to the uterus, differentiation of pathologic from physiologic uptake cannot be achieved on the basis of PET-CT findings alone without the patient s medical history and inquiry regarding menstruation.

20 1430 September-October 2004 RG f Volume 24 Number 5 Figure 35. Recently diagnosed endometrial cancer in a 66-year-old woman. (a c) Axial PET scan (b) demonstrates a hypermetabolic focus of uptake in the midpelvic region (black arrow), a finding that is consistent with physiologic FDG uptake in the bowel, uterine malignancy, or physiologic uptake in the uterus during menstruation. Axial CT (a) and PET-CT fusion (c) images demonstrate that this focus (white arrow) corresponds to the uterine cavity. (d f) Sagittal PET scan (e) shows a focus of FDG uptake (black arrow). Sagittal CT (d) and PET-CT fusion (f) images help further confirm that the focus of uptake (white arrow) corresponds to the uterine cavity and involves the entire endometrium. PET-CT helped localize the uptake to the uterine cavity, and irregular and extensive FDG uptake strongly suggests endometrial cancer. Nevertheless, PET-CT cannot supplant biopsy for confirmation. findings either provided more diagnostic confidence or changed the interpretation of the study involved the abdomen and pelvis (unpublished data). PET-CT is not helpful in differentiating (a) inflammatory changes from neoplastic processes in lymph node stations or lymphatic tissues (Waldeyer ring or appendix), (b) residual tumor from posttherapy changes immediately after surgery or radiation therapy, (c) benign thyroid adenoma from thyroid cancer, (d) focal physiologic bowel uptake from large or small bowel malignancies, or (e) focal physiologic uptake in the uterus during menstruation from uterine cancer. Awareness of the pitfalls associated with PET-CT allows accurate image interpretation. The use of CT attenuation correction may cause diaphragmatic artifacts on reconstructed emission images if significant differences in respiratory phases exist between the two methods. These artifacts may give the spurious impression of subdiaphragmatic lesions in the lungs and vice versa (25,26). Because metallic devices or implants can cause artifacts on CT scans, CT-based attenuation correction may induce artifacts on PET scans (4). The unenhanced portion of the PET-CT fusion image is appropriate for localization but could potentially fail to depict lesions seen at routine contrast material enhanced CT (27). In summary, combined PET-CT scans are more effective than PET scans alone for precise localization of neoplastic lesions and differentiation of normal variants from juxtaposed neoplastic lesions. Hence, PET-CT may significantly affect patient treatment by improving diagnostic specificity more than sensitivity.