Imaging studies in melanoma

INTRODUCTION — Melanoma is an aggressive cutaneous neoplasm that may spread to involve virtually any organ of the body. Dissemination may occur by direct extension from the primary site, by lymphatic spread to regional lymph nodes, or hematogenously. Almost all patients who die from melanoma do so with disseminated disease involving multiple organs; lung and intracranial metastases are the most frequent cause of death.

Imaging studies play an essential role in the initial staging and subsequent management of patients with melanoma by evaluating the extent of disease, detecting clinically occult disease, and/or determining the response to treatment [1-3]. The imaging modalities useful in patients with melanoma are reviewed here.

Topics summarizing the application of information from imaging studies include:

●(See "Staging work-up and surveillance of cutaneous melanoma".)

●(See "Evaluation and management of regional nodes in primary cutaneous melanoma".)

●(See "Overview of the management of advanced cutaneous melanoma".)

IMAGING MODALITIES — Various imaging modalities can be used to assess the extent of disease at presentation, detect recurrence after initial treatment, or monitor disease progression and response to treatment.

Most radiology departments have established protocols for imaging neoplastic disease. Although the referring clinician is not expected to know the details of specific protocols when ordering imaging studies, consultation with the radiologist may assist in optimizing the use of imaging studies for patient management decisions.

Our approach to imaging recommendations for patients with melanoma is generally consistent with guidelines from the National Comprehensive Cancer Network (NCCN) [4] and the American Academy of Dermatology [5].

The judicious use of these imaging studies for staging and restaging of disease is important, as false-positive findings may result in unnecessary expensive and invasive procedures. As an example, in one study of 154 patients with Stage III melanoma who underwent surveillance using computed tomography (CT) or positron emission tomography (PET)/CT, the false positive rate was 53 percent, with a majority of lesions (88 percent) found to be benign [6].

This section will review the primary imaging modalities used in patients with melanoma.

Chest radiographs — Plain chest radiographs ("x-rays") are inexpensive and use a relatively low radiation dose equivalent to the background radiation a person would receive from the environment in less than two weeks. (See 'Radiation exposure' below.)

Chest radiographs have a relatively poor sensitivity for the detection of early (small) lung metastases, adenopathy, or soft tissue disease and are not indicated for asymptomatic patients with stage 0 to II disease [7]. However, in patients with pulmonary symptoms, chest radiographs should be obtained as an initial imaging study to observe for gross anatomical changes that may be seen in both benign and malignant disease. Most patients with suspected metastatic disease will be evaluated with more sensitive and specific imaging procedures, such as CT, PET/CT, or magnetic resonance imaging (MRI).

Computed tomography — CT imaging is used to image the chest, abdomen, and pelvis and may also have a role in brain imaging when MRI is either not available or contraindicated. CT is frequently combined with PET to increase sensitivity and specificity. (See 'FDG PET and PET/CT' below.)

Rapid CT scanning techniques are less prone to bowel and respiratory motion artifacts during imaging, enhancing the sensitivity and specificity of CT scanning. Reported results vary widely, and results with older equipment or techniques do not reflect results that can be obtained with current generation scanners.

CT scans can be done with or without contrast, and contrast can be administered intravenously, orally, or via both routes.

●Oral contrast administration is very helpful to distinguish bowel from nodes, vessels, female reproductive organs, and other structures in the abdomen and pelvis. However, high-density oral contrast agents can obscure the bowel wall, decreasing CT sensitivity for bowel metastases. Low-density, water-based contrast agents or even plain water can be used, which allow excellent visualization of the bowel wall [8,9].

●Intravenous contrast agents improve the sensitivity of CT to detect metastatic disease compared with noncontrast CT. Metastatic melanoma is typically hypervascular, with lesions appearing hyperdense compared with surrounding tissue on early-phase contrast-enhanced CT. Disadvantages of intravenous contrast agents include potential allergic reactions and renal damage. The risk of allergic reactions can often be decreased by using nonionic contrast agents and by avoiding contrast injection in patients with a history of allergies. While renal toxicity secondary to iodinated contrast is rare in patients with normally functioning kidneys, it is a significant risk in patients with preexisting diminished renal function or congestive heart failure.

Multiple-phase imaging consists of several sets of CT images obtained at different times following contrast injection to increase the detection of metastatic melanoma. For melanoma metastases, CT can be optimized with two image sets, one prior to contrast injection and the other during the portal venous phase (60 seconds following the start of contrast injection) [10].

MRI and PET/MRI — Magnetic resonance imaging (MRI) is a sensitive, specific test for imaging melanoma, particularly in the brain, with T1 shortening and hypointensity on short tau inversion recovery (STIR) sequences correlating with the melanin content of the tumor [11]. However, whole-body imaging with MRI is often not practical and is less sensitive than fluorodeoxyglucose (FDG) positron emission tomography (PET)/CT [12].

MRI is used more frequently to better define the anatomic extent of metastases that may not be delineated by other imaging tests, such as CT and PET/CT, particularly in the brain. MRI with contrast enhancement is the most sensitive test for imaging brain metastases; therefore, it is the procedure of choice for patients with symptoms suggesting brain metastases, as well as for those with aggressive histology and clinical findings that may make brain metastasis more likely. Contrast-enhanced T1 acquisition delineates blood-brain barrier breakdown, while T2-weighted fluid-attenuated inversion recovery (FLAIR) imaging delineates areas of vasogenic edema caused by the neoplasm. The Response Assessment in Neuro-Oncology Brain Metastases (RANO-BM) group suggests standard MRI and/or CT response and progression criteria for monitoring response to therapy [13].

Because it may take weeks for metastases to show anatomic changes following effective therapy, new MRI techniques evaluating tumor and cellular metabolism, such as magnetic resonance spectroscopy (MRS), susceptibility weighted imaging (SWI), and chemical exchange saturation transfer (CEST), are being evaluated [14]. (See "Principles of magnetic resonance imaging".)

Patients presenting emergently with symptoms suggesting central nervous system (CNS) disease will often have a CT scan as the initial exam. CT is usually less expensive, easier to perform, and often adequate to diagnose the cause of presenting symptoms. As mentioned above, MRI is more sensitive for brain metastases than contrast-enhanced CT; therefore, brain MRI may be obtained in patients with CT-documented CNS disease if additional information about the number or location of lesions in the brain may affect management. (See "Management of brain metastases in melanoma".)

FDG PET/MRI or whole-body MRI has not replaced FDG PET/CT as the procedure of choice for whole-body staging of melanoma. The advantage of FDG PET/MRI is reduction in radiation dose, an important consideration as cancer survivorship increases. While FDG PET/MRI may also be more sensitive for melanoma metastasis in some tissues such as brain, bone, and liver, it does not do as well as FDG PET/CT for evaluation of lung and lymph node metastases [15]. The improved accuracy of MRI over FDG PET/CT for imaging melanoma metastases in some tissues must be weighed against the cost and availability of FDG PET/MRI as well as the small incremental benefit. Because FDG PET/CT is very accurate, a small improvement in sensitivity outside the brain may often not be of clinical consequence [16,17]. Because whole-body MRI and FDG PET/MRI are not yet as widely available as FDG PET/CT, their role in the management of patients with melanoma remains to be defined

Nuclear medicine techniques — The anatomic detail provided by CT and MRI is complemented by the physiologic information provided by nuclear medicine studies, including PET/CT for staging and following response to therapy, as well as lymphoscintigraphy for sentinel lymph node localization.

Nuclear medicine modalities such as gallium scans and bone scans were used in the past for detecting occult disease; however, FDG PET/CT has almost entirely replaced these techniques

FDG PET and PET/CT — Fluorodeoxyglucose (FDG) positron emission therapy (PET)/computed tomography (CT) combines the anatomic imaging information from CT with the physiologic information that is derived from metabolic differences between tumor and normal tissue. FDG PET/CT is a preferred modality to define the total body burden of metastases, as well as to provide more information about specific sites of metastasis, such as the lung, mediastinum, and abdomen.

●Sensitivity and specificity — In patients with advanced or metastatic disease, PET/CT is at least as sensitive and is more specific than anatomic imaging modalities such as CT and MRI for the detection of metastatic disease [18-23]. Sensitivity is highest (90 percent or better) for metastases that are greater than 1 cm in diameter, but tumor deposits as small as 0.6 cm can be reliably seen in areas of low background activity, such as the lungs.

The reported specificity for PET ranges from 43 to 97 percent [18,23,24], with one study suggesting a false positive rate of up to 53 percent [6]. False positive results may be due to inflammation (recent surgery, infection, granulomatous disease), uptake in skeletal muscle, brown fat or myocardium, immune response, normal activity in the bowel and urinary collecting system, or attenuation artifacts [16,25-27]. False negatives are seen with bone marrow stimulation and in nodes adjacent to brown fat uptake. Normal bowel concentration of tracer may obscure bowel metastases. The specificity of PET (lack of false positives) has improved substantially with the addition of CT to PET (PET/CT). Since the development of PET/CT scanners, PET-only scanners have virtually disappeared, although much of the early literature on PET is based on PET-only imaging machines.

●Technical approach – Glucose labeled with the positron emitter fluorine-18 (FDG) is taken up and initially metabolized in cells similarly to glucose. Once it is phosphorylated by hexokinase to FDG-6-phosphate, however, it cannot progress further through metabolic pathways and is trapped inside the cell. The greater metabolic demands of many neoplastic tissues and increased numbers of passive glucose transport gateways in the cell membrane facilitate concentration of FDG in neoplastic tissues such as melanoma [28].

While glucose uptake is sensitive for melanoma, it is not specific. Glucose uptake can also be seen in normal tissues such as the brain and liver, at sites of infection and inflammation, as well as concentrating in many other neoplastic diseases. Differentiating viable sites of melanoma from inflammatory changes at recent postsurgical sites or from tissue damage secondary to radiation therapy (RT) can be difficult, and therefore more specific tracers are being developed and investigated. Molecular probes being investigated include the use of radiolabeled alpha-melanocyte-stimulating hormone derivatives, very late antigen (VLA) and melanin targeting, as well as acetaminophen probes, among others [29-32].

PET cameras image the three-dimensional distribution of positron-emitting tracers in the body. When combined with CT imaging (PET/CT), the metabolic information from the PET scan is fused with the anatomic information from CT. As a result, PET/CT detects metastases more accurately than CT or PET alone [12,33-35].

There are multiple protocols for obtaining a CT scan during PET/CT. Some institutions prefer to do a full diagnostic CT scan as part of the PET/CT study, including contrast administration. This approach allows the patient to avoid returning for an additional imaging procedure at a separate date, but it results in a higher radiation dose for all patients undergoing the procedure. At many institutions, the CT portion of the PET/CT is a noncontrast, low-dose scan for attenuation correction and general localization of disease. Although there is radiation exposure associated with the addition of a radiotracer such as FDG, the total radiation dose for a PET/CT carried out with low-dose CT scan can be less than half that of a diagnostic CT scan [36-40] (see 'Radiation exposure' below). Because many patients referred for PET/CT have recently had a diagnostic CT, repeating a diagnostic quality CT scan may not be necessary.

●Using PET/CT to determine treatment response – Glucose and FDG uptake are markers for metabolism in many tissues; therefore, FDG PET/CT can not only be used to locate sites of disease through detection of elevated glucose metabolism in melanoma metastases, but can also be used to monitor changes in tumor metabolism following therapy [41,42].

Successful therapy can often be manifest as a decrease or cessation of glucose uptake in melanoma metastases. Several imaging criteria for determining PET response to therapy have been suggested, with the most commonly used being the European Organization for Research and Treatment of Cancer (EORTC) [43] criteria and the PET Response Criteria In Solid Tumors (PERCIST) [44]. Both schemata depend on changes in glucose uptake measured as standardized uptake values (SUVs).

While anatomic changes noted on CT are also important, they are often less sensitive and specific for measuring therapeutic effect, such as in patients with "pseudoprogression" on CT following immunotherapy, or early in treatment when there has not been sufficient time for tumor volume reduction to occur following tissue death [45-47]. Further details on immunotherapy response criteria are discussed separately. (See "Principles of cancer immunotherapy", section on 'Immunotherapy response criteria'.)

The introduction of immunotherapy using checkpoint inhibitors has particularly complicated the assessment of response to therapy with FDG PET/CT [41]. While decreased tumor metabolism (decreased FDG uptake) remains an important indicator of response to therapy, checkpoint inhibitors can cause increased FDG uptake secondary to local inflammatory changes (pseudoprogression) as well as increased uptake in patients with progression of disease that may be paradoxically stimulated with the checkpoint inhibitor (hyperprogression). An increase in FDG uptake following checkpoint inhibitor therapy, therefore, should be closely followed with additional imaging with FDG PET/CT to observe for further changes.

It should also be noted that FDG PET/CT may detect immune-related adverse events (irAEs) in patients on immunotherapy, such as colitis, hepatitis, and hypophysitis. (See "Toxicities associated with immune checkpoint inhibitors".)

SPECT/CT — Single-photon emission computed tomography (SPECT)/computed tomography (CT) provides anatomic information from CT that can be fused with physiologic information, such as lymphatic drainage patterns from Tc-99m-labeled lymphoscintigraphy agents for sentinel node imaging [48,49]. SPECT/CT is particularly valuable for revealing the depth of nodal uptake and identifying adjacent structures prior to sentinel lymph node excision. (See 'Lymph node evaluation' below.)

Examples of situations in which SPECT/CT may be particularly valuable include:

●Melanomas of the head and neck region – Lymphatic drainage in the head and neck is quite variable, and it is not unusual for the injection site to be in close proximity to the sentinel node. SPECT/CT can not only improve node detection accuracy, but also demonstrate anatomic information to assist the surgeon in planning the approach to lymphadenectomy, which may result in a net cost savings [50-52].

●Sentinel nodes in the inguinal region – In patients with melanoma of the lower extremities, sentinel node imaging often demonstrates several sites of tracer uptake in the inguinal region of the ipsilateral leg. In contrast to planar imaging, SPECT/CT can easily distinguish superficial inguinal nodes, which can be readily biopsied, from intrapelvic nodes [51].

●In-transit nodes – Focal uptake of tracer outside the typical node beds, such as in the back, along the arm, or in the chest wall, may be seen with in-transit nodes. In these cases, the location of the site of uptake can again be important in surgical planning. As an example, SPECT/CT imaging will more readily separate periscapular in-transit nodes from axillary nodes (image 1) [51].

Radiation exposure — For patients who are exposed to ionizing radiation as part of medical imaging studies, the risks of radiation exposure must be balanced with the benefit obtained by the imaging procedure. Inappropriate application of various imaging techniques may not always improve patient outcomes and may increase the cost of health care and the exposure to ionizing radiation [53].

These issues need to be considered whenever a discussion of radiographic imaging is undertaken. However, when imaging is used judiciously in patients with melanoma, the theoretical risks associated with radiation exposures are expected to be outweighed by the benefits gained from more accurate staging. (See "Radiation-related risks of imaging".)

APPROACH TO PATIENT IMAGING BASED ON DISEASE SITE — For patients who are suspected to have metastatic disease and those with known metastases, the approach to imaging should focus on symptomatic sites or areas most likely to be involved.

Melanoma commonly metastasizes to regional nodes and skin. Disseminated metastases are also often found in the lungs, brain, bowel, and a variety of other tissues [54]. The differential diagnosis for unsuspected lesions found anywhere on imaging studies in patients with known melanoma almost always includes melanoma. Although characteristic imaging features can help narrow the differential diagnosis, definitive diagnosis requires histologic examination.

There are limitations in available data regarding the diagnostic accuracy of computed tomography (CT), magnetic resonance imaging (MRI), or positron emission tomography (PET)/CT in staging and restaging of melanoma, as demonstrated by one Cochrane review [55]. However, these data reflect the problems behind doing large, long-term randomized controlled studies when imaging and therapy technology are advancing rapidly, rather than a lack of accuracy with imaging. Radiologic evaluation is therefore more likely to be more clinically useful if it is focused on disease site and symptoms, rather than exhaustive. It is important to recognize that in early-stage disease, false positive findings on imaging studies are more likely than true positive findings [56,57]. Specific recommendations on the role and extent of imaging studies in patients with newly diagnosed melanoma and following initial definitive therapy are discussed separately. (See "Staging work-up and surveillance of cutaneous melanoma".)

Lymph node evaluation — Regional lymph node involvement is the most important prognostic factor at presentation for patients without distant metastases. Determination of nodal status is also critical for decision-making regarding adjuvant therapy. Lymphoscintigraphy with sentinel lymph node (SLN) biopsy is the standard approach for assessment of regional lymph node involvement [58,59]. (See "Tumor, node, metastasis (TNM) staging system and other prognostic factors in cutaneous melanoma", section on 'Eighth edition AJCC TNM staging' and "Adjuvant and neoadjuvant therapy for cutaneous melanoma".)

Physical examination is an inaccurate predictor of nodal metastases; 20 percent of clinically node-negative patients have metastatic deposits, while 20 percent of those who are clinically node-positive have pathologically negative nodes. Microscopic disease associated with early lymphatic metastasis currently cannot be detected by any in vivo imaging modality, including fluorodeoxyglucose (FDG)-PET, CT, or nodal ultrasound [60-62].

Lymphoscintigraphy and subclinical adenopathy — SLN biopsy can provide important information that influences patient management and prognosis.

The SLN hypothesis maintains that tumor cells migrating from a primary tumor colonize the first (sentinel) node(s) receiving lymphatic drainage from the primary tumor site before involving other nodes. While cutaneous lymphatic drainage pathways follow a generally predictable route to specific node beds, the specific node to first receive lymphatic flow within each node bed (the "sentinel node") is difficult to determine using anatomic imaging techniques. In addition, the typical drainage pathways can vary substantially from person to person and region to region, with some areas of the skin draining to multiple lymph node beds [63].

The pattern of lymphatic drainage is most accurately demonstrated by lymphoscintigraphy, which uses a radioactive tracer injected alone or in combination with local injection of blue dye (methylene blue, isosulfan blue). The SLN biopsy can provide important information that directly influences patient management and prognosis. Intradermal injection of radioactive tracer around the primary tumor permits identification of sentinel node(s) in the majority of patients with the use of gamma camera imaging prior to surgery and a handheld gamma probe at the time of surgery. Once the site of sentinel node(s) is identified with the gamma camera, it may be marked on the skin by the imaging clinician before the patient is sent to surgery. The images should be available for the surgeon to review prior to surgery. The surgeon, with the aid of the images, skin markings, and a handheld gamma probe, can localize the sentinel nodes at the time of surgery (picture 1). If the sentinel node is negative for metastasis, the remaining regional nodes are very likely to be negative.

SLN biopsy procedure — Regional lymphatic assessment is typically performed at the same time as wide local excision of the primary tumor. The procedure begins with the intradermal injection of a radiolabeled tracer within 1.5 cm of the primary lesion. Proper intradermal injection is usually accompanied by the development of a tense wheal as the tracer is injected.

There are several tracers available for sentinel node scintigraphy, including nanocolloid human serum albumin, antimony sulfide colloid, sulfur colloid, and tilmanocept. The latter two (sulfur colloid and tilmanocept) are the only two agents approved for clinical use in the United States. Sulfur colloid is typically filtered through a 0.22 micron filter prior to injection. This tracer is a particulate that can enter the lymphatic channels and undergo phagocytosis and retention by cells in the first lymph node to which it is delivered. Tilmanocept is a mannose derivative that binds to CD206 receptors in macrophages and dendritic cells found in lymph nodes [64]. All tracers are typically labeled with technetium-99m (^99mTc), a gamma-emitting isotope with a half-life of six hours.

Because the tracer is injected intradermally, patients may experience an uncomfortable burning at the site of injection for several seconds. Tilmanocept has been shown to be less painful during injection than sulfur colloid [65]. Otherwise, the procedure is well tolerated by most patients. The tracer is subsequently picked up by the local lymphatic channels and is then deposited in the nodal tissue receiving lymphatic flow from the site of injection. The SLN is often detected within 10 to 30 minutes of injection.

Focal uptake of tracer can then be imaged using planar images with a standard nuclear medicine gamma camera, with single-photon emission computed tomography (SPECT) or with SPECT done concurrently with CT (SPECT/CT) for more precise localization of the site and depth of tracer uptake [49,66]. Typical images obtained are shown in the figures, using lymphoscintigraphy (image 2) and SPECT/CT (image 3).

When the primary lesion is located on the trunk, imaging should include both axillary and inguinal node beds because of the variability of lymphatic drainage. The area between the injection site and sentinel node bed(s) should also be surveyed for in-transit nodes. In-transit nodes can also be seen with injection of a primary lesion on the extremities. Crossover to contralateral node beds is rare (eg, right leg draining to left inguinal or left iliac nodes) but should be considered when obtaining images.

Care must be taken to image all nodal basins that may drain the site of the primary lesion, as SLNs in more than one nodal basin are common [67,68]. Unusual patterns of tracer concentration or unusual sites of nodal uptake should be directly communicated to the surgeon. Flow images obtained with the gamma camera can assist in providing information about the timing of the appearance of focal tracer accumulation. This can be particularly helpful if there is tracer contamination in the field of view. Nodal uptake tends to accumulate during dynamic images, whereas sites of inadvertent contamination should be present at the beginning of dynamic images and show no change in intensity.

Focal areas of tracer uptake may then be marked on the overlying skin by the nuclear medicine clinician for assistance with the initial guidance of the handheld gamma probe during surgery. Skin marking is particularly helpful for in-transit sites of uptake as well as sites of uptake in the inguinal/pelvic region, as superficial inguinal nodes can be marked separately from nodes deep to the pelvic fascia. Although surgery can be delayed for up to 24 hours following tracer injection [69], radioactive decay of the tracer and movement of the tracer through the SLN to secondary nodes are potential problems if surgery is delayed, particularly if smaller colloid particles or albumin are injected. If surgery is planned for the day after tracer injection and imaging, the nuclear medicine clinician should be notified. An increase in the injected dose of tracer and scheduling the patient for later in the afternoon will lessen the problems associated with radioactive decay of the tracer.

Visualizing the sentinel node with the gamma camera or locating it with the intraoperative probe can be difficult when the injection site is close to the draining node or node bed. This problem is often encountered with head and neck lesions. In these cases, injecting a smaller volume of injectate and/or a smaller amount of tracer activity close to the primary lesion (within 0.5 cm) can sometimes make resection of the injected material around the primary lesion easier and, subsequently, allow easier localization of the sentinel node with the intraoperative probe. SPECT/CT may also be particularly useful for separating the injection site from adjacent nodes.

While tilmanocept may be less painful during injection, ^99mTc can dissociate from tilmanocept more readily than from sulfur colloid. The dissociated radioisotope can then be excreted by the kidneys and end up in the bladder. Radioisotope in the bladder can make probe localization of inguinal nodes during surgery difficult. Some surgeons therefore prefer using ^99mTc sulfur colloid to ^99mTc tilmanocept for mapping nodal drainage from any primary site that may drain to inguinal or pelvic nodes.

In experienced hands, the procedural false-negative rate (the inability to identify a SLN that contains tumor) is typically 5 percent or less. Procedural false negatives should be distinguished from biologic false negatives, in which regional lymph node relapses are seen following biopsy of a true negative SLN.

Causes of false-negative studies using lymphoscintigraphy for SLN localization include inadequate pathologic examination of the SLN, poor tracer injection technique, imaging the wrong nodal basin, failure to detect in-transit nodes, not imaging all possible nodal basins, and complete replacement of the SLN with neoplastic disease, causing the injected tracer to completely bypass the infiltrated node [70,71].

SLN scintigraphy has been validated in patients with local biopsy or focal excision of the primary lesion only. While there remains concern that prior wide local excision can change the normal patterns of lymphatic drainage around a primary melanoma, in at least one series, prior wide local excision did not appear to adversely affect the results of lymphatic mapping and SLN biopsy [72].

The radiation dose associated with lymphoscintigraphy is trivial. Very small amounts of radioactivity are injected, and much of that activity is removed during surgery. When CT is added to SPECT (SPECT/CT), the CT is typically not done with contrast, is only of a limited area, and is usually done with a low-dose imaging protocol, delivering only a fraction of the dose of diagnostic CT scans.

Ultrasound — In patients with clinically negative regional lymph nodes and a positive sentinel lymph node biopsy, careful follow-up of the nodal basin with serial ultrasound may be pursued rather than further surgery. This approach is discussed in detail separately. (See "Evaluation and management of regional nodes in primary cutaneous melanoma", section on 'Subsequent management'.)

Ultrasound is not widely utilized in primary staging of melanoma because it relies on anatomic changes that are not seen in nodal metastases until the tumor burden is substantial. While the resolution of high-frequency ultrasound has shown promising results in superficial cervical chain nodes as well as inguinal nodes in thin patients, axillary, iliac, and other deeper node beds are not as amenable to high-frequency ultrasonographic examination. In addition, sentinel node scintigraphy can identify aberrant drainage pathways to unexpected node beds and in-transit nodes.

Ultrasound relies upon the detection of anatomic changes in nodes with metastatic disease. Nodes that demonstrate the loss of the normal fatty hilum can undergo fine needle aspiration. The ultrasonographic finding of peripheral hypervascularity in combination with other ultrasonographic findings increases the sensitivity of ultrasound for the diagnosis of nodal metastasis to approximately 82 percent, with a positive predictive value of 52 percent [73]. Ultrasound might prevent sentinel node biopsy procedures in patients that are destined to have a complete lymph node dissection. Although ultrasound is not generally considered a substitute for SLN biopsy, serial ultrasounds may be useful in some circumstances to detect nodal recurrence [55,74].

Lungs and mediastinum

●Computed tomography – Contrast-enhanced CT scanning is the most widely used radiographic method for evaluating intrathoracic lesions because of its sensitivity for detecting small pulmonary nodules (image 4) [75]. CT is also superior to a plain chest radiograph for demonstrating mediastinal and hilar adenopathy that often accompanies parenchymal lesions and/or the presence of lymphangitic spread, important issues for patients who have a potentially resectable, isolated pulmonary metastasis [76]. (See "Metastatic melanoma: Surgical management", section on 'Lung'.)

●PET/CT – PET/CT scanning is more sensitive than CT alone for detecting occult metastatic foci, including those in the lungs. FDG uptake in an anatomically benign-appearing lymph node can also raise the suspicion for metastatic involvement. Thus, PET/CT may be helpful in patients with an apparently isolated pulmonary metastasis prior to planned thoracotomy and resection of the metastasis [77]. (See 'FDG PET and PET/CT' above.)

Liver — The liver is a common site of metastatic disease, with the majority of patients showing metastatic disease at autopsy. When present, liver metastases are rarely solitary or isolated sites of disease. Hepatic metastases are more common in patients with choroidal or mucosal primaries, which more readily disseminate hematogenously. (See "Initial management of uveal and conjunctival melanomas".)

●PET/CT – Because glucose is normally stored in the liver, normal liver uptake of FDG diminishes the sensitivity of PET/CT for detection of liver metastases. However, as with brain imaging, unless there are laboratory test results or symptoms suggesting liver metastasis, whole-body PET/CT is often considered adequate for staging.

●CT versus MRI – If abnormalities are noted in the liver on either the PET or the noncontrast CT portion of the PET/CT scan, further workup with multiphase contrast CT or contrast-enhanced magnetic resonance imaging (MRI) can be helpful. The choice of CT or MRI for imaging the liver is often based on institutional practice. By the time metastatic disease to the liver manifests itself on PET/CT, through symptoms or by abnormal blood tests, there are usually multiple metastases, and multiphasic CT is adequate for localization and diagnosis. CT also has the advantage that it can be done more quickly than MRI and does not require the degree of patient cooperation that is required by MRI. If the number of hepatic metastases is important, such as in a patient that is being considered for partial hepatectomy, contrast-enhanced MRI is the most sensitive imaging study for the diagnosis of liver metastases [78].

Hepatic metastases may appear partially calcified or hemorrhagic and, if large, may contain areas of necrosis (image 5) [75]. The enhancement pattern following contrast injection may be uniform, inhomogeneous, or ring-like (image 6). Because liver metastases from melanoma are hypervascular, they are best evaluated by multiphase CT in the noncontrast and portal contrast phases [10]. Computed tomography arterial portography (CTAP) utilizes the predominant hepatic arterial blood supply of metastases to enhance lesion detection by increased differential tumor enhancement relative to the normal liver. The increased sensitivity of CTAP permits the detection of metastatic lesions less than 1 cm but requires catheter placement in the superior mesenteric artery and is, therefore, not practical in most patients [79,80].

While multiphase CT is the initial procedure of choice for anatomic imaging of the liver, MRI may be useful in distinguishing solitary lesions as either benign entities (eg, hemangiomas) or metastatic disease [81-83]. MRI is also superior for demonstrating vascular involvement and identifying additional hepatic lesions in the patient thought to have a single isolated metastasis. On MRI, metastatic lesions appear as low-signal areas on T1-weighted images and as moderately high-signal areas on T2-weighted images, while hemangiomas are characteristically well-circumscribed homogeneous lesions with low signal intensity (darker) on T1-weighted images and high intensity (brighter) on T2-weighted images. (See "Approach to the adult patient with an incidental solid liver lesion".)

Brain — The frequency of brain metastases is higher with melanoma than with most other malignancies. Central nervous system metastases are the second most common cause of mortality after pulmonary involvement [54]. The management of brain metastases in melanoma is discussed separately. (See "Management of brain metastases in melanoma".)

●MRI versus CT – MRI is the preferred imaging modality when brain metastases are suspected or need to be ruled out in nonemergent cases. MRI is significantly more sensitive than CT for the detection of metastatic disease (image 7) [84] and also provides information about possible involvement of the spinal cord and leptomeninges. When MRI is not available or when it is contraindicated, CT is an alternative. On CT, intracranial metastases may enhance uniformly or in a ring-like pattern. Similar patterns are seen on gadolinium-enhanced MRI studies (image 8).

Patients presenting acutely with symptoms of brain metastases can be imaged with CT as a first-line modality. CT is usually more readily available than MRI, and results may be obtained more quickly. Lesions big enough to cause symptoms are more likely to be detected on CT than smaller lesions not causing symptoms.

If a more complete evaluation of a patient with CT-positive brain metastases is needed, brain MRI may be more accurate than CT. CT brain imaging is not routinely done on all patients with stage IV melanoma, particularly asymptomatic patients; however, one study found brain lesions in as many as 12 percent of asymptomatic patients, and in the same study, brain metastases in asymptomatic patients were associated with a significantly poorer prognosis [85]. The impact of small asymptomatic brain metastases on the prognosis of patients receiving newer melanoma therapies has yet to be fully established.

If there is hemorrhage associated with the metastasis, MRI can not only document the presence of blood, but it can also characterize the approximate age of the hemorrhage by detecting signal changes associated with the breakdown products of hemoglobin and red blood cell membrane integrity. Even in the absence of hemorrhage, melanoma may show a brighter signal on T1 sequences due to the paramagnetic effects of melanin.

●FDG PET/CT – The primary substrate for brain metabolic activity is glucose. The result is intense FDG uptake in normally functioning brain cortical tissue. This obscures most melanoma metastases and makes FDG PET/CT insensitive for the diagnosis of metastatic disease to the brain. Fibroblast activation protein inhibitor (FAPI) and amino acid-based radiotracers have shown promise for PET/CT imaging of neoplastic metastasis to the brain, particularly in monitoring response to therapy in difficult cases where MRI of the brain is equivocal [86].

Bone metastases — Patients with melanoma presenting with pain that suggests a musculoskeletal origin should have imaging workup that is guided by the history of the symptoms and physical exam. Most bone metastases occur in the axial skeleton [87-89].

●Plain radiography and computed tomography – Localizing symptoms can guide plain film radiography, which may reveal degenerative changes, trauma, or possibly, lytic or blastic lesions, and therefore, may be sufficient for management. However, while inexpensive and readily available, plain films are not sensitive for early neoplastic disease. If plain films are negative and metastatic disease is clinically suspected, further imaging using other modalities should be guided by the likelihood of metastatic disease. For example, CT is more specific than plain radiographs in the evaluation of bony lesions and is particularly useful for the detection of purely lytic lesions, which may not be readily apparent by other imaging techniques, such as radionuclide bone scan [88]. MRI and PET/CT are more sensitive and specific for bone metastasis than both plain radiography and CT; however, MRI can be less readily available than CT [90].

●MRI – If the pain is localized and may be related to soft tissue injury, MRI is the procedure of choice, as it is sensitive and fairly specific for both soft tissue and bony lesions (image 9). MRI is the best modality for evaluating marrow signal abnormalities and accompanying features such as hemorrhage or soft tissue masses. However, MRI usually only evaluates a limited area of interest and has a low specificity.

●Nuclear imaging – If widespread metastatic disease is a concern, FDG PET/CT will provide accurate information about the site of interest as well as evaluate the extent of disease elsewhere in the torso [91-93]. PET/CT using either FDG or F-18 sodium fluoride (NaF) is generally considered to be the most sensitive and specific test for the diagnosis of occult melanoma bony metastases, although there is preliminary evidence suggesting whole-body PET/MRI may be more sensitive for occult malignant bone lesions [94]. F-18 sodium fluoride PET/CT has shown higher accuracy for detection of bone metastases when compared with conventional bone scintigraphy [95,96]. Addition of SPECT/CT to conventional bone scintigraphy, however, may elevate the accuracy of conventional bone scintigraphy to rival that of F-18 sodium fluoride PET/CT [97].

●Bone scan – Radionuclide bone scanning is a simple and sensitive test for screening the entire skeleton for metastases, even though the majority of lesions are lytic on plain film. Activity of osteoblasts at the site of a metastasis to bone can often be detected months before changes are seen on plain radiographs.

Despite its sensitivity, bone scanning is not routinely ordered in patients with melanoma unless symptoms suggesting bone metastases are present. The low specificity of this technique, combined with the low incidence of metastases to bone early in the disease, results in an unacceptable number of false-positive studies typically caused by benign processes such as trauma or degenerative joint disease [98-101].

Gastrointestinal tract — Metastatic lesions involving the gastrointestinal (GI) tract are relatively common and usually multiple. Melanoma is the solid tumor that metastasizes most frequently to the small bowel. This was illustrated by one series in which 110 of 2500 patients (4.4 percent) had GI involvement [102]. Of these, 35 percent had lesions affecting the small bowel. The colon, stomach, and esophagus were involved in 14, 7, and 1 percent of cases, respectively. (See "Metastatic melanoma: Surgical management", section on 'Gastrointestinal tract'.)

Metastatic melanoma involving the GI tract may present with anemia, overt bleeding, pain, obstruction, or intussusception. Radiographic evaluation is indicated in patients with melanoma who present with any of these symptoms or unexplained weight loss. However, GI metastases are often asymptomatic. In one series, less than 10 percent of patients with metastases to the bowel presented with pain, nausea, or vomiting [103].

Studies that may be useful to evaluate the GI tract include abdominopelvic CT scan, small bowel follow-through, and enteroclysis. Small bowel enteroclysis is the most useful anatomic imaging procedure for evaluating small bowel lesions, while an upper GI series is the best procedure for evaluation of symptoms referable to the esophagus and stomach. However, upper GI series imaging has been largely replaced by endoscopic evaluation, particularly when mucosal metastasis is suspected.

Endoscopic evaluation (fiberoptic or capsule) of bowel abnormalities detected on PET/CT may also be helpful. Fiberoptic endoscopy is more sensitive than capsule endoscopy and includes the ability to biopsy lesions but it can only access short sections of the proximal small bowel (during upper endoscopy) and the distal small bowel (during colonoscopy).

Melanoma may cause thickening of mucosal folds or multiple, small, submucosal lesions similar to metastatic deposits from lung or breast cancer, lymphoma, and Kaposi sarcoma. Submucosal lesions may appear rounded or polypoid [102]. If the center of the tumor ulcerates (as the tumor outgrows its blood supply), the lesion can have a bull's eye or target appearance on barium fluoroscopy. This manifestation is considered pathognomonic for melanoma metastases to the bowel [104].

If signs or symptoms of bowel lesions are present, the investigation should be guided by the patient presentation. If there is microscopic blood in the stool from a suspected colonic source, colonoscopy is very helpful for identifying the source. Biopsy or cautery, as needed, can be performed with colonoscopy. If there is pain or signs of obstruction, then CT or colonoscopy should be considered, depending on the presentation. Bowel obstruction caused by metastases encroaching on the bowel lumen is rare; obstruction is more commonly caused by intussusception [102].

While there are few published articles evaluating FDG-PET imaging of melanoma metastatic to the bowel, one report suggests that FDG-PET can detect bowel metastases that are missed by conventional anatomic imaging studies [105]. FDG can be normally seen in the small bowel and colon, particularly in patients taking metformin. Immunotherapy with checkpoint inhibitors can also cause colonic inflammatory changes that may take up FDG. These and other nonmalignant causes of tracer uptake can decrease the accuracy of PET/CT for the detection of bowel lesions. Bowel metastases may also be discovered incidentally during FDG PET/CT imaging for staging or monitoring response to therapy. Focal intense tracer uptake in the bowel with no adjacent segmental bowel uptake should raise suspicion for bowel metastasis. In this situation, large bowel abnormalities may be further evaluated using colonoscopy, and small bowel abnormalities may be evaluated using capsule endoscopy or barium radiography.

Skin and subcutaneous tissue — Involvement of skin or subcutaneous tissue may be the first sign of metastatic disease. In one report, 17 of 45 patients (38 percent) with CT evidence of metastatic melanoma had subcutaneous nodules shown on CT; in five (11 percent), they were the only site of metastatic disease [106]. These lesions are particularly conspicuous on CT scans because of the density of the metastatic nodules relative to the low attenuation of the surrounding subcutaneous fat (image 10).

However, imaging findings are rarely the first indicator of subcutaneous metastases, which can often be easily felt by the patient and clinician.

Muscle metastases — Rarely, melanoma metastases may invade muscle. On CT scanning, these metastases are hypervascular, and therefore typically have high attenuation relative to muscle following contrast injection. They may also present as low-attenuation lesions on CT imaging (image 11). PET/CT is more sensitive than CT for the diagnosis of muscle metastasis, but in one series, the false-positive rate was high for isolated findings. PET/CT was more accurate when multiple focal abnormalities were detected [107].

Cardiac involvement — Melanomas are more likely to metastasize to the heart than other solid tumors. Although most secondary pericardial tumors do not invade the myocardium, myocardial involvement is a classic feature of melanoma [108]. (See "Pericardial disease associated with cancer: Clinical presentation and diagnosis" and "Cardiac tumors", section on 'Secondary cardiac tumors'.)

Echocardiography is the primary imaging tool to establish the presence of a pericardial effusion. It is also useful to quantify the volume of an effusion and to evaluate its hemodynamic effects, particularly the presence of tamponade or constrictive pericarditis. (See "Constrictive pericarditis: Diagnostic evaluation" and "Cardiac tamponade" and "Differentiating constrictive pericarditis and restrictive cardiomyopathy".)

CT and MRI, although seldom used as a primary modality, also provide excellent information regarding pericardial effusion, especially when the fluid is loculated. Both techniques, which are usually performed for some other reason, are superior to echocardiography for determining whether an effusion is a transudate, exudate, or blood.

Large tumor masses in the pericardium can be recognized by most of the previously discussed imaging studies. In some cases, however, a hematoma, usually of traumatic origin, can be mistaken for neoplasia. Echocardiography, particularly transesophageal echocardiography complemented by ultrafast CT scanning and MRI imaging, can provide the degree of structural resolution necessary to ascertain the extent of local structural involvement.

As with the brain, the use of PET for imaging metastases to the myocardium is not practical because of the normal myocardial concentration of glucose.

Less common sites of metastatic involvement — Other less common sites of metastasis include the spleen, kidneys, adrenal glands, gallbladder, and breast.

●Spleen – The typical CT appearance of splenic metastases consists of single or multiple ill-defined, low-attenuation lesions, but they may also occasionally appear as well-defined cystic masses or target lesions. FDG uptake in splenic lesions can also raise the suspicion for metastatic disease.

●Kidney – Melanoma metastases to the kidney may similarly be single or multiple, and solid or cystic. It is unusual for renal metastases to calcify, and this finding should raise suspicion for another etiology, such as a primary renal cell carcinoma [109]. Although melanoma uncommonly presents as a perirenal mass, it was the most frequent cause of such masses in one large series [110]. Ultrasound may be useful to further characterize a renal mass as cystic or solid (more likely to be malignant) (image 12). Because FDG is normally excreted through the kidneys, FDG PET/CT is insensitive for metastases to the kidneys or elsewhere in the urinary tract. (See "Simple and complex kidney cysts in adults".)

●Adrenal glands – Isolated adrenal metastases are also uncommon in melanoma, but they may be frequently seen in patients with widespread metastases. The adrenals were involved in 50 percent of cases in an autopsy series [54]. In older autopsy series, adrenal metastases were often large and frequently bilateral [76]. Similar to their appearance at other metastatic sites, lesions may be solid or cystic. A retrospective review of 1180 melanoma patients with distant metastasis showed that 13 percent had adrenal metastases with a median survival of six months [111].

●Gallbladder – Melanoma metastases to the gallbladder are rare (1 to 2 percent of cases at autopsy), but they represent the most common hematogenously spread metastasis to this organ [54]. Gallbladder metastases are usually accompanied by liver metastases [112]. Metastatic lesions in the gallbladder appear as mucosal deposits that become polypoid as they grow. These deposits are particularly suspicious for malignancy when larger than 7 mm.

●Breast – Melanoma is a more common tumor to metastasize to the breast than most other neoplastic diseases, including lymphoma or leukemia. On mammography, melanoma metastases usually appear as multiple, dense, well-circumscribed masses without calcification.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Melanoma screening, prevention, diagnosis, and management".)

SUMMARY

●Extent of disease involvement – Melanoma is an aggressive cutaneous neoplasm that may spread to involve virtually any organ of the body. Dissemination may occur by direct extension from the primary site, by lymphatic spread to regional lymph nodes, or hematogenously. Various imaging modalities are critical for the proper staging and follow-up of patients with this disease. (See "Staging work-up and surveillance of cutaneous melanoma" and "Overview of the management of advanced cutaneous melanoma" and "Evaluation and management of regional nodes in primary cutaneous melanoma".)

●Evaluating regional lymph nodes – For patients with newly diagnosed melanoma, lymphoscintigraphy and sentinel lymph node biopsy are part of primary management for those with locoregional disease and no obvious metastases. (See 'Lymph node evaluation' above and "Evaluation and management of regional nodes in primary cutaneous melanoma", section on 'SLNB timing and technique' and "Evaluation and management of regional nodes in primary cutaneous melanoma".)

●Utility of FDG PET/CT – Besides localization and mapping the extent of metastatic disease, fluorodeoxyglucose (FDG) positron emission tomography (PET)/computed tomography (CT) has shown utility in monitoring disease response to therapy. However, the nonspecificity of FDG uptake should always be considered when interpreting sites of FDG uptake, particularly when clinical information suggests infectious/inflammatory changes may be responsible. (See 'FDG PET and PET/CT' above.)

●Approach to imaging based on disease site – For patients who are suspected to have metastatic disease and those with known metastases, the approach to imaging should focus on symptomatic sites or areas most likely to be involved. (See 'Approach to patient imaging based on disease site' above.)

•Total body burden of metastases – FDG PET/CT combines the anatomic imaging information from CT with the physiologic information that is derived from metabolic differences between tumor and normal tissue. FDG PET/CT, thus, is a preferred modality to define the total body burden of metastases, as well as to provide more specific information regarding specific sites, such as the lung, mediastinum, and abdomen. (See 'FDG PET and PET/CT' above.)

•Brain – Magnetic resonance imaging (MRI) is the preferred modality for identifying brain metastases, which are relatively common in patients with metastatic melanoma. (See 'Brain' above.)

•Liver – The preferred modality for the evaluation of possible liver metastases depends on institutional practice, indications for the test, and the ability of the patient to cooperate with imaging logistics. Both multiphasic CT and contrast-enhanced MRI can be used. Multiphasic CT is most commonly used and is often all that is needed. Contrast-enhanced MRI is more sensitive, which may be critical if hepatic resection is being considered. Liver metastases are present in more than one-half of patients during the course of metastatic disease. (See 'Liver' above.)

•Bone – Plain radiographs retain a role as the initial imaging modality to diagnosis bone metastases in symptomatic patients. CT, MRI, FDG PET/CT, or bone scan can also have a role when radiographs are not diagnostic. (See 'Bone metastases' above.)