Oncologic PET by Anatomical Region

III
Oncologic Applications



10
Oncologic PET by
Anatomical Region

Eugene C. Lin and Abass Alavi


Image General Principles


The primary attribute of positron emission tomography (PET) imaging is its great contrast resolution compared with anatomical imaging techniques. Because of this contrast resolution, in oncologic imaging PET has substantial advantages over anatomical techniques in the early detection of disease in staging and recurrence and the accurate evaluation of therapy response. In general, the disadvantages of PET are related to its relatively poor resolution relative to anatomical imaging techniques and the numerous sources of nonneoplastic uptake of fluorodeoxyglucose (FDG). However, many of the disadvantages of PET are greatly mitigated with the use of combined PET/computed tomography (CT) imaging (see Chapter 8).



  1. Advantages. The advantages of PET over conventional imaging techniques in assessing cancer are found in two areas:

    1. PET can detect early disease before gross anatomical changes (Fig. 10.1). PET can detect disease in areas that would be interpreted as normal on anatomical imaging techniques due to the lack of structural changes.
    2. There is a greater contrast-to-noise ratio of abnormal to normal structures on PET compared with that of the anatomical modalities (Fig. 10.2). Because of this, abnormalities clearly seen on PET are often not prospectively detected on anatomical imaging techniques. In many cases, these abnormalities can be retrospectively detected after attention is specifically directed to the abnormal area by PET.

  2. Disadvantages

    1. Limited sensitivity for small lesions. In general, the sensitivity of PET is low for lesions < 1 cm. PET (or any other gross imaging modality) cannot detect micro-metastases.
    2. False-positives. Potential false-positive results can be seen from a wide range of inflammatory/infectious processes or other benign etiologies.
    3. Anatomical localization. Occasionally, it is difficult to localize lesions to their exact anatomical sites based on PET, particularly in the head, neck, and pelvic sites. This is largely remedied by combined PET/CT, although misregistration is sometimes encountered with this approach.
    4. Low sensitivity in specific areas. PET has poor sensitivity for brain and lung metastases and sclerotic bone lesions.
    5. Decreased sensitivity for specific tumors. PET has poor sensitivity for specific tumors, such as prostate cancer, bronchoalveolar carcinoma, and mucinous adenocarcinoma.


  3. Metastatic disease demonstrated by PET will often alter patient management. Given the possibility of false-positive findings, confirmation with anatomical imaging techniques or biopsy should be strongly considered for PET findings that will alter patient management. This is particularly true if there is a solitary lesion demonstrated on PET that will potentially alter patient management.


image

Fig. 10.2 Peritoneal metastasis from colon cancer. (A) Axial positron emission tomography (PET) scan shows focal uptake (arrow) adjacent to small bowel (arrowhead). Although mild diffuse small bowel uptake is usually normal, focal areas of uptake greater than the surrounding small bowel are usually not normal. It is normal to have substantially greater uptake in the right colon (open arrow) than the small bowel. (B) Computed tomography (CT) scan at the same level demonstrates a small focus of abnormal soft tissue (arrow) adjacent to small bowel. This was not seen prospectively on the CT and was present only on one slice. The contrast-to-noise ratio of the abnormality on PET is much greater than on CT.


Image Liver



  1. Liver metastases. PET is very sensitive for detecting liver metastases that are > 1 cm in size. In general, PET is more specific than CT or magnetic resonance imaging (MRI).1 Some studies have suggested that PET is more sensitive than CT or MRI,2 but other studies suggest that PET is less sensitive.1 Note that liver MRI performed with specific liver imaging agents such as superparamagnetic iron oxide1 or mangafodipir trisodium3 is more sensitive than PET.
  2. Differentiation of benign from malignant lesions. Focal uptake in a liver lesion is very specific for malignancy and may represent metastases, hepatocellular carcinoma (HCC), or cholangiocarcinoma.4

    1. Benign lesions such as hemangiomas, focal nodular hyperplasia, and hepatic adenomas typically do not have increased FDG uptake.

      • There is a single report of a focal nodular hyperplasia with increased uptake.5

    2. False-positives. Hepatic abscess, nodular lymphoid hyperplasia (pseudolymphoma), inflammatory pseudotumor, sarcoidosis6,7
    3. False-negative. Low-grade HCC (see Chapter 20)

  3. Evaluation of small liver lesions. Although FDG PET has a limited sensitivity for liver lesions < 1 cm in size, it still has a high specificity. This is particularly useful when CT detects liver lesions < 1 cm but cannot further characterize the nature of these lesions. Statistically, these lesions most likely represent small cysts or hemangiomas, but in the setting of primary malignancy elsewhere, it is not possible to determine non-invasively the true nature of these lesions. Given the low sensitivity, PET should not be used primarily for evaluating these small liver lesions, but PET is indicated for assessing for disease activity elsewhere. If focal FDG uptake is noted in these small lesions, they are almost certainly metastatic (Fig. 10.3). However, even when there is no increased uptake seen on PET, these lesions cannot be diagnosed as benign, as they could be malignant and not detected on PET due to their small size.

    image

    Fig. 10.3 Subtle liver metastasis. A metastasis in the dome of the liver (arrow) is much better seen on (A) an arterial-phase axial computed tomography (CT) scan than on (B) a portal phase scan. (C) This lesion has mild FDG uptake on an axial PET scan. This lesion might be interpreted as artifactual secondary to image noise on the positron emission tomography scan alone, or if correlated with the portal phase CT. Correlation with arterial phase CT was necessary. The lesion itself is nonspecific in appearance on CT, but the presence of visible fluorodeoxyglucose uptake in this small lesion is consistent with metastatic disease.


  4. Artifacts. The liver is often heterogeneous due to image noise on attenuation-corrected images. Image noise is more prominent in the liver than in any other organ. Due to the normal heterogeneity of the liver, caution must be exercised before interpreting subtle findings as liver metastases (Fig. 10.4).
  5. How should one interpret subtle foci of uptake in the liver? Most metastases in the liver will demonstrate greatly increased activity relative to normal liver. A subtle focus of activity in the liver on attenuation-corrected images may be due to image noise rather than a lesion. Before interpreting a subtle focus as a lesion:


    image

    Fig. 10.5 Liver metastases seen by positron emission tomography (PET) only. (A) Axial computed tomography (CT) scan does not demonstrate any liver lesions. (B) Corresponding PET/CT scan demonstrates multiple hepatic metastases. This is a rare finding, as the majority of liver metastases seen on PET can be visualized by contrast-enhanced CT. However, liver metastases are occasionally not visualized on portal-phase CT scans. The fatty infiltration of the liver in this case may contribute to the lack of visualization of the metastases. An arterial-phase CT was not done in this case, so it is unclear whether the metastases would have been visualized in a different phase of contrast enhancement.



    1. Review the nonattenuation-corrected (NAC) images. If the focus is seen on the less noisy NAC images, the level of confidence for a lesion is greatly increased. If the focus is not detected on NAC images, it may be artifactual secondary to image noise. Note that this method may be less helpful for central lesions, which may not be seen on NAC images due to attenuation.
    2. Correlation with CT must be performed. CT is very sensitive for detecting small liver lesions, but it often cannot characterize these lesions. It is likely that if a focus seen on PET is a true lesion, it will be recognized on a contrast-enhanced CT, often in retrospect. Rarely, true-positive lesions may not be seen on CT, particularly if the liver is fatty infiltrated and/or imaged in only one phase of enhancement (Figs. 10.3 and 10.5). However, if a non-contrast CT is obtained as part of a PET/CT exam, many true-positive liver lesions identified on PET will not be detected on the noncontrast CT.

Image Spleen


There are limited data on PET/CT imaging for the evaluation of solid splenic masses.8




  1. Known FDG-avid malignancy

    1. PET/CT. Sensitivity 100%, specificity 100%
    2. Standardized uptake value (SUV). An SUV of 2.3 is useful for differentiating benign from malignant lesions.

  2. No known malignancy

    1. PET/CT. Sensitivity 100%, specificity 83%
    2. PET has a high negative predictive value in this setting.

      • However, a non-FDG-avid primary tumor should be excluded before diagnosing the splenic mass as a benign process.

    3. An FDG-avid splenic mass is likely (80%) to be malignant even without a known primary.

Image Peritoneum



  1. Peritoneal metastases. Peritoneal metastases are most common in ovarian and gastrointestinal cancers. PET is more accurate than CT in detecting peritoneal metastases, but it will not detect very small implants (thus, it will not replace second-look laparotomy). PET is particularly helpful when intraperitoneal fluid is present—it can detect malignant ascites or small solid implants (Fig. 10.6).
  2. Patterns of spread. It is important to understand the classic patterns of peritoneal spread. Specific attention should be directed to these areas (Fig. 10.7):

    1. Serosal surfaces of the liver and spleen (note that when correlating with CT, splenic serosal metastases can appear cystic)
    2. Omentum
    3. Paracolic gutters, particularly on the right
    4. Medial to the cecum (this must be differentiated from normal physiologic cecal uptake) (Fig. 10.8)
    5. Sigmoid mesocolon
    6. Pelvis—particularly between the bladder and uterus and between the uterus and rectum


    image

    Fig. 10.7 Peritoneal seeding pattern. Coronal positron emission tomography scan demonstrates multiple foci of peritoneal seeding medial to the cecum (open arrow) and on the sigmoid mesocolon (arrowheads) from a gastroesophageal carcinoma (arrow). This is a classic pattern of peritoneal spread. (From Lin EC, Lear J, Quaife RA. Metastatic peritoneal seeding patterns demonstrated by FDG positron emission tomographic imaging. Clin Nucl Med 2001;26(3):249–250. Reprinted with permission.)



  3. Distinguishing peritoneal metastasis from bowel activity. It is sometimes difficult to differentiate peritoneal metastasis from bowel activity without PET/CT.

    1. The focus should be reviewed in all three planes, with an attempt to link the activity to a segment of bowel in at least one plane. Often foci that appear outside the bowel in one plane are clearly visualized in the bowel on another plane.
    2. Outside the right colon, cecum, and rectosigmoid where normal intense areas of uptake are often noted, it is unusual to see foci of bowel activity, which are substantially more intense than the surrounding structures; this is particularly true in the small bowel (Fig. 10.2).
    3. CT correlation is very helpful, as peritoneal metastases are often identified retrospectively with this modality after attention is directed to the area of abnormality by PET.
    4. SUV. An SUV cutoff of 5.110 may be helpful for the diagnosis of peritoneal carcinomatosis.
    5. PET/CT. Although PET/CT can often allow for differentiation of bowel activity from peritoneal disease, PET/CT has the potential disadvantage of causing areas of artifactually increased and decreased uptake in the bowel (see Chapter 8). Potential peritoneal disease should be interpreted with caution on PET/CT, as bowel motion between the CT and PET scans can result in both false-positive and false-negative findings. NAC images should always be reviewed in conjunction with the corrected scans.

  4. Diffuse peritoneal carcinomatosis. Diffuse peritoneal carcinomatosis can result in diffuse peritoneal uptake, which may be difficult to clearly define, as no focal lesions are seen in this setting.11 Clues to identifying diffuse disease are (Fig. 10.9)12

    1. Liver border. The liver border is poorly visualized, as peritoneal activity is close in intensity to that of liver.
    2. Straight-line sign. On sagittal and axial images, the retroperitoneum is less intense than the peritoneum in peritoneal carcinomatosis (normally, the peritoneum and retroperitoneum are of comparable intensity). This results in a straight line demarcating the peritoneum and retroperitoneum on sagittal images.

Image Lymph Nodes



  1. Lymph node metastases. PET is more sensitive than CT in detecting lymph node metastases. Malignant nodes can be detected by PET before nodal enlargement (> 1 cm) on CT.
  2. Size. The sensitivity of PET for metastatic deposits in nodes measuring 6 to 10 mm is 83%; it drops to 23% for those ≤ 5 mm.13 Because PET will not detect micrometastases, it is not a substitute for sentinel node imaging procedures. Thus, PET is not a substitute for sentinel node imaging for axillary nodal staging in breast cancer and local nodal staging in melanoma.
  3. Inflammatory versus malignant lymph nodes. A common problem is differentiating benign from malignant nodal uptake.

    1. SUV. As a general rule, malignant nodes have higher SUV values (usually > 2.5) compared with inflammatory nodes. However, small nodes with metastases, due to partial volume effects, may have a low SUV and low uptake by visual interpretation.
Tags:
Sep 3, 2016 | Posted by in GENERAL RADIOLOGY | Comments Off on Oncologic PET by Anatomical Region

Full access? Get Clinical Tree
Get Clinical Tree app for offline access