Diagnosis of Stage IV Melanoma
The development of distant metastasis – stage IV disease – is associated with a dramatic decrease in the survival rate of patients who have melanoma. As seen with other tumor types, the success of treatment mainly depends on early diagnosis as well as initiation of optimal treatment. In the past decade, there has been dramatic expansion of the treatment armamentarium for patients with advanced melanoma with the development of two most important novel therapeutic strategies, immune checkpoint inhibitors and selective kinase inhibitors, which have led to significant improvement in survival and prognosis of metastatic melanoma patients. In this chapter, we provide a comprehensive overview of clinical aspects of the diagnosis of stage IV melanoma, including the timing and pattern of metastasis, clinical and imaging evaluation of distant metastases, and clinical application of biomarkers for diagnosis of metastatic melanoma. Also, various common and uncommon sites of metastasis of cutaneous melanoma are elucidated in detail.
KeywordsMelanoma Metastatic melanoma Organ-specific Radiology Biomarkers Brain metastasis
Metastatic melanoma is the most aggressive form of skin cancer, with the most compromised survival. Historically, once melanoma progressed to stage IV disease, it was usually incurable, and less than 10% of patients survived long term (Balch et al. 2001a). However, recent developments in immune-oncology and molecularly targeted therapy have translated into significant improvements in survival, with rates ranging from 28% to over 40% at 5 years in the most recent reports (Hamid et al. 2018; Long et al. 2018). In this chapter, we review various clinical aspects of the diagnosis of stage IV melanoma, including the timing of metastasis, patterns of metastasis, and clinical and imaging evaluation of distant metastases.
Timing of Distant Metastasis
Late recurrence, defined as the diagnosis of metastasis after a disease-free interval of 10 years or more, is a well-known phenomenon that is almost unique to melanoma. In a subset of patients who have thin, node-negative disease, melanoma recurs more than 10 years after the primary diagnosis (Jr et al. 1992; Crowley and Seigler 1990). Thus, especially for patients with thin melanomas, a 10-year disease-free interval may not be considered an absolute cure (Jr et al. 1992; Crowley and Seigler 1990; Faries et al. 2013). Further, in a study by Lo et al. (2018), 24.8% of patients initially diagnosed with thin melanoma (Breslow thickness less than or equal to 1 mm) died after 10 years of initial diagnosis.
Pattern of Metastasis
Common distant sites of metastatic melanoma
Clinical seriesa (%)
Autopsy seriesa (%)
Skin, subcutaneous tissue, and lymph nodes
Prognosis and Prognostic Factors
Patients with stage IV melanoma were historically considered to have a very poor prognosis, with survival typically measured in months rather than years (see also “Melanoma Prognosis and Staging”). Until recently, the duration of survival from the time that distant metastasis was documented was less than a year, with a median of approximately 8.5 months. The 1-year survival rate was 45%, although approximately 10% were estimated to live for 5 years or more (Gershenwald et al. 2003; Barth et al. 1995; Manola et al. 2000). However, recent studies involving immune checkpoint inhibitors and molecularly targeted agents have reported 1-year, 3-year, and 5-year survival rates exceeding 70%, 50%, and 40%, respectively (Schachter et al. 2017; Wolchok et al. 2017; Hamid et al. 2018). Multivariate analyses of prognostic factors have identified several independent factors that predict survival in this group, including the site and number of metastases (Manola et al. 2000; Unger et al. 2001; Balch et al. 2001b). Further, patient performance status, baseline tumor burden, serum LDH, and response to systemic therapy are associated with survival in stage IV patients treated with modern immunotherapy and targeted therapy (Joseph et al. 2018; Schadendorf et al. 2017; Robert et al. 2018). Several other studies have identified various prognostic factors for patients with stage IV melanoma, such as the site of the first metastases, number of metastatic sites, serum S100B, and duration of remission (Poo-Hwu et al. 1999; Balch et al. 1983; Barth et al. 1995; Manola et al. 2000; Sirott et al. 1993; Tarhini et al. 2009).
Sites of Distant Metastases
Number of Metastatic Sites
Patients with one metastatic site have a significantly improved survival rate compared with those with two or more distant sites (Balch et al. 1983; Gershenwald et al. 2003; Unger et al. 2001; Neuman et al. 2008). In an analysis of 200 patients with distant metastases who were treated at the University of Alabama, the number of metastatic sites was the most significant factor predicting survival in patients with distant metastases by single-factor analysis (Balch et al. 1983). The 1-year survival rate was 36% for patients with one metastatic site, 13% for patients with two sites, and less than 1% for patients with three or more sites. Within the single-site group, patients with metastases to the lung, skin, subcutaneous tissues, or distant lymph nodes had a better survival rate than patients with metastases to any other visceral site (Balch et al. 1983).
In contrast to this study, however, the number of metastatic sites was not a significant independent prognostic factor on multivariate analysis of the studies conducted at the John Wayne Cancer Institute (Barth et al. 1995). This suggests that certain sites of metastases have a dominant negative effect on survival. Patients whose initial site of metastasis was the liver or brain had a median survival of only 4 months compared with patients whose initial sites were the skin and/or lymph nodes, who had a median survival of 15 months (Barth et al. 1995). In addition, it should be noted that accurate information about the number of metastases depends, in part, on whether any diagnostic tests and which ones were performed to search for distant metastases. With CT and PET/CT imaging more routinely applied, it must be cautioned that small foci of tumor may be detected which are not detectable by physical exam alone. Thus, prognostic data based on physical exam or clinical evidence of stage IV metastatic disease may be more favorable if the stage IV disease is more limited and only detectable by imaging.
Elevated Serum Lactate Dehydrogenase
An elevated serum LDH level was among the most predictive independent factors for diminished survival in all published studies in which it was analyzed in a multivariate analysis, even after accounting for site and number of metastases (Keilholz et al. 1998, 2002; Eton et al. 1998; Deichmann et al. 1999; Balch et al. 2009). Thus 1- and 2- year overall survival rates for those stage IV patients in the 2008 AJCC/International Union Against Cancer melanoma staging database with a normal serum LDH level were 65% and 30%, respectively, compared with 32% and 18%, respectively, when the serum LDH level was elevated at the time of staging (p <0.0001). These findings led to the incorporation of LDH into the AJCC staging system for patients with metastatic disease (Balch et al. 2009). However, an elevated serum LDH level should be considered definitive only when at least two determinations are obtained more than 24 h apart. A single elevated serum LDH level can be false positive because of hemolysis or factors unrelated to melanoma. Additionally, current research suggests that metastatic melanoma patients with an elevated serum LDH at baseline or during treatment with immune checkpoint inhibitors have a poor prognosis (Kelderman et al. 2014; Weide et al. 2016).
Duration of Remission
The disease-free interval before the onset of distant metastasis was a significant prognostic factor in a multivariate analysis of studies conducted at the University of Alabama and the John Wayne Cancer Institute (Balch et al. 1983; Barth et al. 1995). The stage of disease preceding distant metastasis was also identified as an important prognostic factor in the John Wayne Cancer Institute analysis (Barth et al. 1995). For patients who had progressed directly from stage I or II disease, a disease-free interval of 34 months or longer was associated with prolonged survival, whereas for patients with a history of stage III melanoma, a disease-free interval of 18 months or longer was associated with prolonged survival.
Poor performance status has been identified as another negative prognostic factor in patients with stage IV melanoma (Gershenwald et al. 2003; Manola et al. 2000; Unger et al. 2001; Keilholz et al. 1998). In a study of 1362 patients with metastatic melanoma treated in eight Eastern Cooperative Oncology Group (ECOG) melanoma trials conducted over a 25-year period, Manola et al. (2000) found that on multivariate analysis, the following factors were associated with increased risk of death: ECOG performance status of 1 or more (relative risk [RR] = 1.49); metastatic disease in the gastrointestinal tract (RR = 1.49), liver (RR = 1.44), pleura (RR = 1.35), or lung (RR = 1.19); and number of metastatic sites (RR = 1.12).
Other Prognostic Factors
Other laboratory parameters, such as elevated alkaline phosphatase levels (Manola et al. 2000; Sirott et al. 1993; Eton et al. 1998), hypoalbuminemia (Sirott et al. 1993; Eton et al. 1998), and thrombocytosis (Manola et al. 2000; Sirott et al. 1993) were also identified as prognostic markers. The role of laboratory tests, including blood-based biomarkers, in the management of patients with high-risk or metastatic melanoma is discussed later in this chapter.
Clinical Evaluation of Metastasis
Extensive radiographic evaluation of patients with AJCC stage I, II, or III melanoma who are clinically free of disease rarely reveals radiologic evidence of metastases. There are emerging data indicating that imaging studies, including CT scans, commonly reveal metastases in the absence of symptoms, signs, or abnormal findings on physical examination or standard laboratory test results if deployed in a surveillance approach for high-risk patients. Park et al. reported in a group of 466 patients (328 of whom were stage III and 23 of whom were resected stage IV) that when systematic clinical, lab, and CT imaging surveillance were performed, the tumor progression found by patients, physicians, and CT imaging alone was 27%, 14%, and 59%, respectively (Park et al. 2017). Thus, imaging can detect many cases of recurrence not detected by physical exam.
History and Physical Examination
Many cases of advanced melanoma metastases produce symptoms or can be detected by physical examination. The hallmark of metastatic disease is a symptom complex that progresses in either intensity or frequency. A complete history and physical examination are the most important components of the initial diagnostic evaluation of patients for metastatic disease. Among 261 patients with stage II or III melanoma who were followed prospectively by the North Central Cancer Treatment Group, symptoms signaled a recurrence of melanoma in 99 (68%) of 145 patients who had a relapse (Weiss et al. 1995). Recurrences were detected on physical examination in an additional 37 asymptomatic patients (26%). Altogether, 94% of recurrences were detected by history and physical examination, an abnormal chest radiograph identified only 9 (6%) of 145 recurrences, and in none of the patients, abnormal laboratory findings were the sole indicator of recurrent disease. Mooney et al. (1997) assessed the impact of a surveillance program that used physical examination, blood tests, and chest radiographs on 1004 patients with AJCC stage I or II cutaneous melanoma. Physical examination detected 72% of recurrences, constitutional symptoms indicated 17% of recurrences, and chest radiographs revealed 11% of recurrences. Similarly, among the 373 patients followed in a surveillance program at the Yale Melanoma Unit, in addition to the 34 (44%) of 78 recurrences in patients who were first seen with symptoms, physical examination detected 25 patients with asymptomatic recurrences (32%) (Poo-Hwu et al. 1999). Thus 76% of recurrences were diagnosed by complete history and physical examination alone. As discussed above, data from Park and colleagues show that imaging with CT can detect many patients with stage IV melanoma before it is clinically apparent (Park et al. 2017).
In the follow-up of melanoma patients with a high risk of recurrence, laboratory tests are frequently used. Basic blood chemistry, including LDH and liver function tests, is readily available in most clinical laboratories and may be an important screening tool for metastatic melanoma. An isolated increase in the serum alkaline phosphatase or LDH level is presumptive evidence of metastatic disease (Amer et al. 1978; Finck et al. 1983). However, there is no universal agreement among clinicians at major cancer centers concerning the utility of these tests in screening for metastases in patients with melanoma. Perhaps this is because once the LDH is elevated, the prognosis of patients, in general, was considered poor, with a historic median survival of 4 to 6 months and less than 30% of patients alive at 1 year (Balch et al. 2001a; Keilholz et al. 2002). Several studies of LDH levels have clearly demonstrated that abnormal LDH levels reflect a high tumor load and therefore mainly reflect a late stage of metastatic disease (Balch et al. 2000). As the enzyme that catalyzes pyruvate into lactate, LDH can be a marker of cancer metabolic activity and increased tumor cell uptake of glucose and high dependence on the anaerobic glycolytic pathway (Feron 2009). LDH has been shown to be a negative prognostic marker, regardless of treatment received even with modern-age therapies (Long et al. 2018, 2016). The detection of early stage IV disease with the use of LDH as a screening marker is very unlikely.
In general, biomarkers are molecules (proteins, lipids, carbohydrates) found in tissue, blood, or other fluids of patients with tumors, and their expression profile is associated with the course of the malignant disease. Marker molecules are expressed either by tumor cells or by cells of the tumor environment. Thus, most biomarkers are found primarily in malignant tissues, but after active secretion or passive release with tumor cell destruction, they become detectable in body fluids, particularly in the blood (Utikal et al. 2007). The serologic parameters most widely used currently for early detection of a melanoma recurrence are melanocyte lineage/differentiation antigens such as S-100 beta (Bonfrer et al. 1998; Tarhini et al. 2009) and the so-called melanoma inhibitory activity (Bosserhoff et al. 1997). The advantage of these proteins is their relatively high specificity for melanoma, even though these molecules are also expressed by other, nonmalignant cells.
The term S-100 describes an acidic calcium-binding protein derived from bovine brain extract that is soluble at neutral pH in 100% (saturated) ammonium sulfate. S-100 beta blood measurements were primarily used to detect severe central nervous system damage before S-100 beta was demonstrated to be a useful serum marker for metastatic melanoma. Melanoma inhibitory activity was originally detected in melanoma cell culture supernate and was shown to have an important role in cell-matrix interaction and metastasis. Because S-100 beta demonstrated superiority in terms of a better sensitivity-specificity ratio compared with routine blood parameters, it found acceptance in the German and Swiss guidelines for the care of patients with melanoma (Dummer et al. 1995; Garbe et al. 2007). Dozens of clinical trials indicate that protein S-100 beta is an excellent prognostic marker for patients with stage III and stage IV disease, with a strong correlation between S-100 beta serum values and treatment outcomes. However, S-100 beta appears to have no value in staging of stage I and II disease or in the detection of micrometastases (Berking et al. 1999; Bonfrer and Korse 2001; Hauschild et al. 1999b, c). Studies comparing S-100 beta as a prognostic marker with LDH showed that S-100 beta is an earlier indicator of stage IV disease, whereas LDH is primarily a marker of late-stage disease, with a high tumor burden and poor prognosis (Hauschild et al. 1999a; Tarhini et al. 2009; Egberts et al. 2012).
Screening disease-free melanoma patients with S-100 beta as a routine serum marker showed, in several clinical trials, a higher sensitivity and specificity for several melanoma-related antigens compared with alkaline phosphatase, LDH, and reverse transcription-polymerase chain reaction (RT-PCR) (Garbe et al. 2003; Schlagenhauff et al. 2000). The sensitivity to detect new metastasis was 29% for protein S-100 beta, 22% for melanoma inhibitory activity, 2% for LDH, 17% for alkaline and phosphatase, and 24% for multimarker RT-PCR. The diagnostic accuracy was best for melanoma inhibitory activity (86%) and S-100 beta (84%); alkaline phosphatase (79%), LDH (77%), and RT-PCR (72%) demonstrated lower values (Garbe et al. 2003). Early on, immunoradiometric assays were used for the serum measurement of S-100 beta. Currently, four different assays for the determination of S-100 beta in the serum are marketed (Smit et al. 2005). Despite the many studies supporting its sensitivity and its acceptance in Germany and Switzerland, most countries, including the United States, have not incorporated S-100 beta testing into the routine metastatic surveillance of patients at high risk for recurrence of melanoma.
Several other serologic tumor markers for melanoma have been investigated (Lee 1980; Karjalainen 2001). At the John Wayne Cancer Institute, a serum enzyme-linked immunosorbent assay (ELISA) was developed to measure a circulating immune complex composed of TA90, a tumor-associated 90 kd glycoprotein antigen, and its immunoglobulin G antibody that is expressed in 72% of melanomas (Gupta and Morton 1992; Kelley et al. 2001). The serum TA90 ELISA assay test had a sensitivity of 77% in detecting recurrences of melanoma (Kelley et al. 1998). When this test was used in conjunction with a PET scan, the detection rate for occult melanoma lesions increased to 93% in 87 patients who had no clinical evidence of disease (Hsueh et al. 1998). A series of studies has evaluated the role of TA90 in predicting recurrence in patients with either thick primary tumors or stage III disease (Kelley et al. 1998, 2001; Chung et al. 2002; Faries et al. 2007). In addition, the efficacy of TA90 complex relative to melanoma inhibitory activity and S-100 beta has also been explored in 75 patients before they were given immunotherapy (Faries et al. 2007). Results showed that at least one marker became elevated before 41 (80%) of 51 recurrences. TA90-IC was the earliest elevated marker in 29 (57%), melanoma inhibitory activity in 11 (22%), and S-100 beta in 4 (8%). These markers were inconsistently able to predict patient survival. In addition, although the mean values for melanoma inhibitory activity and S-100 beta increased progressively with disease progression, TA90IC exhibited a parabolic curve, the significance of which was unclear.
The serum levels of various cytokines, such as interleukin-8 (Scheibenbogen et al. 1995) and interleukin-10 (Franzke et al. 1998; Dummer et al. 1995; Itakura et al. 2011), adhesion molecules (i.e., ICAM-I), (Franzke et al. 1998), and lipid-bound sialic acid (LASA-P) (Miliotes et al. 1996), also showed a correlation with either tumor stage or tumor burden, but these are not yet in routine clinical use, largely as a result of their low specificity for metastatic melanoma because abnormal serum levels were also found in several inflammatory conditions. Another interesting new biomarker for melanoma was named YKL-40. YKL-40 is a growth factor in connective tissue cells and stimulates migration of endothelial cells. Cancer cells, macrophages, and neutrophils secrete YKL-40 into the serum (Schmidt et al. 2006a). Its function in cancer is unknown, though it is hypothesized to play a role in cancer cell proliferation, angiogenesis, and protection from apoptosis (Johansen et al. 2006). In a clinical trial, YKL-40 was measured in serial serum samples from 110 patients with metastatic melanoma – before, during, and after treatment – and from 245 healthy control subjects. Elevated serum YKL-40 levels were an independent prognostic factor for poor survival. More recent clinical studies have demonstrated contrasting results in correlating serum YKL-40 with tumor stage as well as the prognostic potential of YKL-40 in predicting relapse and survival (Schmidt et al. 2006b; Díaz-Lagares et al. 2011; Egberts et al. 2012). More importantly, use of immunomodulatory drugs can potentially influence YKL-40 expression, and therefore its use as a biomarker is limited by problems of false-negative results during treatment (Krogh et al. 2010). Despite the progress to date, there is still a need for biomarkers with a higher specificity and sensitivity for early detection of melanoma metastasis.
Serum proteomic profiling is another innovative approach for biomarker discovery which involves high-throughput analysis of serum proteins to identify signature biomarker patterns specific for tumor type. The first study of 101 patients with early-stage melanoma and 104 with advanced-stage melanoma, which employed a matrix-assisted laser desorption/ionization (MALDI) protein chip technology and artificial neural networks, demonstrated a correct stage assignment in 88% of the serum samples from patients with different disease stages (Mian et al. 2005). Eighty percent of serum samples from patients with stage III disease were correctly assigned as progressors or non-progressors with the use of a random-sample, cross-validation statistical methodology. Eighty-two percent were correctly identified by MALDI, whereas only 21% were detected by serum S-100 beta measurements. In another study, increased expression of serum amyloid A (SAA), detected in the serum of melanoma patients using proteomic analysis, was associated with a poor prognosis and a high risk of progression (Findeisen et al. 2009). Carol and colleagues utilized surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF-MS) to discriminate serum samples of 30 melanoma and 24 non-cancer patients and demonstrated a good diagnostic accuracy of 98.1% (Caron et al. 2009). Several other approaches to proteomics have been utilized in the recent past to establish clinically relevant biomarkers, but to date the results have been relatively limited. Gene expression profiling has also shown promise as a prognostic biomarker in patients with operable melanoma, classifying tumors into low-risk or high-risk for distant metastasis. However, the data are still too immature to gain acceptance into national practice guidelines (Zager et al. 2018).
Detection of Cells in Messenger RNA Using PCR
Because dissemination of tumor cells through circulating blood is essential for the formation of distant metastases, it has been hypothesized that the presence of melanoma cells in the peripheral blood could serve as a marker for early dissemination of disease. The detection of circulating melanoma cells in the blood represents an appealing prognostic tool, but even after dozens of clinical studies, no consensus on this topic exists. Early studies attempted to detect circulating tumor cells in the bloodstream by standard cytologic techniques. However, the level of cancer cells in the peripheral blood is usually low (less than 1 in 106 cells) in patients who have limited metastatic disease.
Later, PCR technology was employed to enhance the sensitivity of tests used in the detection of the rare tumor cells circulating in the blood. Amplification of genetic sequences uniquely associated with melanoma would significantly improve the specificity and sensitivity of prospective markers, particularly at low tumor volumes. One advantage of RT-PCR is its potentially higher specificity and sensitivity compared with protein assays.
Initially Smith et al. (1991) demonstrated the ability to identify tyrosinase messenger RNA (mRNA) in blood samples from several patients with metastatic melanoma. Tyrosinase, a melanosomal protein, is the key enzyme involved in melanin biosynthesis. Currently, tyrosinase is the most commonly employed mRNA marker assayed in melanoma. In the early 1990s, researchers began to investigate the clinical utility of RT-PCR in the detection of tumor-associated markers in patients with various stages of melanoma (Brossart et al. 1993; Battayani and Grob 1995; Foss et al. 1995; Kunter et al. 1996; Pittman et al. 1996; Mellado et al. 1996). Despite these initially encouraging results, most of the reports demonstrate a disparity between PCR positivity and clinical staging (Gläser et al. 1997; Mellado et al. 1999; Reinhold et al. 1997). In patients who have disseminated melanoma, where tumor volume and circulating tumor cells are expected to be the greatest, positive PCR results have ranged from 0% to 100% (Brossart et al. 1993; Foss et al. 1995). Because of such discrepancies, the reliability of single-marker RT-PCR assays for the detection of circulating tumors has been questioned (Gläser et al. 1997; Mellado et al. 1999; Reinhold et al. 1997; Bostick et al. 1998; Brossart et al. 1995). The low fraction of positive findings in the peripheral blood of patients with melanoma might also reflect intermittent shedding of tumor cells. The limitations in sensitivity and specificity of the single-marker RT-PCR assay are also thought to be related to the heterogeneity of tumor marker expression among tumors, as well as within an individual tumor lesion or among multiple lesions in individual patients. With the use of a multimarker RT- PCR assay, including five antigens – tyrosinase, tyrosinase-related protein 1 (TRP-1), tyrosinase-related protein 2 (TRP-2), Pmel 17/gpl00, and Melan-A/MART-1 – Taback et al. (2001) demonstrated improved sensitivity over a single-marker approach in the detection of occult metastasis.
A systematic review and meta-analysis of 53 studies enrolling 5433 patients showed a correlation of circulating melanoma cells with lymph node metastasis, disease stage, and patients’ progression-free and overall survival (Mocellin et al. 2006). However, the rates of circulating melanoma cells were higher than expected in the early stages and, conversely, lower than expected in the metastatic setting. The authors interpreted the results as a consequence of low accuracy of currently available detection assays. Furthermore, the heterogeneity of the study results published so far warrants some caution when evaluating the favorable results reported in the pooled data (Mocellin et al. 2006). More recently, the sequencing of circulating tumor DNA (ctDNA) as a surveillance method has shown promise. Testing in the context of the adjuvant AVAST-M trial, a negative trial of bevacizumab, ctDNA levels of known BRAF and NRAS mutants were monitored after surgery. Patients with detectable ctDNA during follow-up had significantly worse metastasis-free and overall survival (Lee et al. 2018). In patients with metastatic melanoma, an analysis from four randomized trials of treatment with BRAF-MEK inhibitors reported undetectable ctDNA prior to treatment initiation to be associated with improved progression-free survival (Santiago-Walker et al. 2016). Overall, the clinical implications of serial ctDNA monitoring, especially in patients with metastatic melanoma, are still uncertain, and more work is needed to determine the appropriate platform for clinical use. Although these methodologies are promising in terms of determining prognosis, diagnosing metastatic disease, and assessing response to treatment, they cannot at the present time be used as a basis for clinical decision-making.
Diagnostic imaging is essential for the diagnosis of patients with stage IV disease, as well as for their more precise staging and for follow-up of those receiving treatment. With the improvement of treatment for patients with stage IV melanoma, the importance of systematic systemic imaging to assess treatment response continues to grow. A variety of imaging techniques are used to follow and evaluate patients with melanoma, including chest radiographs, regional nodal ultrasound imaging, CT scans, PET and PET/CT scans, and brain/spine and hepatic MRIs. The value of each of these procedures for surveillance of patients with early-stage melanoma is described (See also “Surveillance and Follow-up of Melanoma Patients”). The value of these tests for the diagnosis, staging, and follow-up of patients with stage IV disease will be discussed next. There is a wide variety of recommendations for imaging in stage I–III melanoma. However, in stage IV, follow-up imaging is often patient-specific, and the frequency and methods of imaging are dependent on the planned therapeutic interventions.
A key question for all diagnostic procedures is which patients should have a workup for systemic metastatic disease. With the use of sentinel node biopsy, detection of metastatic disease and low-volume disease in a single lymph node are not uncommon. Several studies have shown a low yield of positive diagnostic imaging tests for detection of systemic metastases after positive sentinel nodes are detected. For example, Miranda et al. (2004) found that less than 1% of imaging studies performed in such patients detected true metastatic disease. Similarly Gold et al. (2007) found less than a 4% yield of systemic metastases in a similar group of patients, but pointed out that patients with thick primary melanomas and metastatic disease in the sentinel nodes were at greatest risk for systemic metastases. More recently, Holtkamp et al. (2017) noted that metastasis was found in only 2 out of 143 melanoma patients with a positive sentinel lymph node biopsy using routine staging imaging with CT or PET/CT, and none of the 143 patients had a change in AJCC stage. There is always the chance that results will be false positive or indeterminate when imaging is employed for groups of patients at very low risk for metastases. An overarching question, clinically, is whether early detection of systemic metastatic disease makes a difference in patient management or outcomes. As therapies for stage IV melanoma continue to improve, the detection of lower tumor burden recurrence may be of increasing interest and relevance.
The first step, beyond physical exam, in determining the presence or absence of stage IV melanoma has historically been anatomic imaging. This approach is changing given the excellent sensitivity of PET/CT scans that combine the anatomic imaging of CT with the functional imaging of PET. Patients with suspected systemic metastases, based on history, physical examination, chest radiography, or laboratory testing, should be evaluated extensively because of the likelihood of detecting additional asymptomatic lesions that might alter patient prognosis or treatment selection. Such patients should be evaluated by means of MRI of the brain and CT or PET/CT scans of the chest and abdomen. A CT scan or a PET/CT scan of the pelvis is indicated for patients with a history of primary tumors below the waist or if there are symptoms suggestive of metastatic involvement. Other imaging studies should be ordered on the basis of symptoms (e.g., bone scan for patients with bone pain, wireless video capsule endoscopy for patients with iron deficiency anemia after negative initial evaluation on upper endoscopy and/or colonoscopy, or spine MRI for patients with back pain).
Hypervascular metastases of the liver are typically scanned by multiple-phase imaging, including the unenhanced phase, hepatic arterial phase, and portal venous phase. In general, imaging of the portal venous phase alone usually offers satisfactory detection of lesions. Greater lesion detection can be achieved by unenhanced images in addition to the portal venous phase images. The hepatic arterial phase does not seem to increase lesion detection. However, atypical melanoma metastases may require all phases for optimal viewing (Mellado et al. 1999). MRI may also be used to detect melanin as manifested by a high signal on T1-weighted images or to exclude hemangioma. Ultrasonography may be of value in assessing hepatic, splenic, or pancreatic lesions. Ultrasound images show lesions that are usually predominantly hypoechoic. If the lesions are of heterogeneous echogenicity, hemorrhage is suspected. Upper endoscopy and/or colonoscopy are indicated if there are signs or symptoms suggesting metastatic disease to the gastrointestinal tract (i.e., anemia, hemoccult-positive stools, cramping abdominal pain with bloating) followed by a wireless video capsule endoscopy if the initial evaluation is negative or non-diagnostic.
CT scans of the chest are useful for evaluating suspected pulmonary, pleural, or mediastinal metastases. However, even in patients who have known distant metastases (stage IV disease), these scans are best suited for cases in which the presence of pulmonary metastases would alter the treatment plan or when better definition of lesions is required before a patient is entered into a research protocol.
Although the traditional approach to detecting metastatic melanoma has been physical examination and anatomic imaging by means of chest radiography and CT, practice patterns have continued to evolve in the past several years as imaging techniques improve. It is increasingly recognized that PET imaging, especially PET combined with CT as PET/CT imaging, offers superior diagnostic accuracy compared with CT alone in detecting visceral metastases from patients with stage IV melanoma.
The rationale for PET imaging is that the vast majority of cancers have accelerated rates of glucose metabolism relative to normal tissues. This accelerated glycolysis is multifactorial in origin, but it was recognized early in the past century by Warburg and is often referred to as aerobic glycolysis (Kim and Dang 2006). This phenomenon has been targeted successfully for imaging by means of the glucose analog l8F-fluoro-2-deoxy-d-glucose (FDG). 18F is a positron emitter with approximately a 110-minute half-life and is produced using a medical cyclotron. FDG is accumulated in cancer cells through glucose transporters and is then phosphorylated to FDG-6-phosphate by hexokinase. The FDG-6 phosphate cannot be metabolized any further and is essentially trapped in the cancer cells and thus can be imaged (Torizuka et al. 1998).
Preclinical studies showed FDG accumulation in melanoma xenografts and animal models to be among the highest of any human cancer type (Kern 1991; Wahl et al. 1991). This suggested human imaging would be feasible with FDG using PET imaging techniques. The feasibility of PET imaging of melanoma in humans was demonstrated in 12 patients in 1993 at which time there were several cases of successful detection of otherwise undetectable tumor foci (Gritters et al. 1993).
PET is a technique that detects the in vivo distribution of positron emitters throughout the body with signal intensity proportional to the absolute accumulation of the radiotracer in the body. Although PET is not an anatomic imaging tool, it is one that is limited by physics and instrumentation. PET scanners in current clinical use have resolutions of approximately 4 to 8 mm, suggesting that lesions smaller than the resolution size will be difficult to detect. However, resolution in PET does not accurately describe the “smallest lesion” that can be detected. For example, lesions with very avid accumulation of FDG can be detected under the proper circumstances even if they are smaller than the physical “resolution” of the scanner. However, detection can be compromised if there are normal structures with intense radiotracer uptake near the tumor foci that lower the target/background uptake ratios and thus lesion detectability. Similarly, if a lesion is “moving” during the time of acquisition, such as a nodule in the base of the lungs, detection can be impaired. Thus, although modern PET scanners can survey the entire body, their overall detection capability for melanoma varies somewhat based on disease location. The low sensitivity of PET for very small-volume disease has been demonstrated in several studies evaluating the characteristics of nodal metastases detected by PET. Wagner et al. (2001) and Crippa et al. (2000) have shown that PET is generally quite insensitive for lesions less than 5 mm in diameter. FDG PET begins to reliably detect metastatic tumor in lymph nodes at volumes greater than approximately 80 mm3, but sensitivity falls rapidly below this. Thus efforts to use PET to detect primary lesions and low-volume nodal metastases have been associated with poor levels of sensitivity. By contrast, detection of more substantial foci of melanoma (greater than 5 mm in diameter) has been quite successful in many studies (Friedman and Wahl 2004; Steinert et al. 1995).
Many series report a greater sensitivity with FDG- PET scans compared with conventional radiographic studies for the detection of most metastases, except for those in the lungs (especially the base of the lung and the brain) (Friedman and Wahl 2004; Schwimmer et al. 2000). In a series of 100 patients with stage IV disease from the Sydney Melanoma Unit, 415 metastatic lesions were evaluated by means of PET and routine CT imaging (Damian et al. 1996). PET detected 93% of lesions. In 20 patients, PET detected 24 metastases up to 6 months earlier than conventional imaging or physical examination, consistent with the findings of Park et al. (2017). Furthermore, the selection of surgical and medical treatment was influenced by PET findings in 22 cases (Damian et al. 1996). PET without concurrent CT may complement routine imaging studies rather than replace them. In a series of 68 patients, PET detected fewer pulmonary, hepatic, and brain metastases but more lymph node and bone metastases than conventional CT imaging (Dietlein et al. 1999). In 76 patients with metastatic melanoma, Holder et al. (1998) showed a sensitivity of 94% for PET but only 55% for CT imaging.
A prospective evaluation of 106 whole-body FDG PET scans was reported from Duke University Medical Center in 95 patients with clinically evident stage III lymph node and/or in-transit melanoma (Tyler et al. 2000). A total of 234 areas were evaluated pathologically, 165 of which were confirmed histologically to be melanoma. PET identified 144 of the 165 areas of melanoma for a sensitivity of 87.3% with false-negative findings generally in subcentimeter disease. There were 39 areas of increased PET activity that were not associated with malignancy, for a 78.6% predictive value of a positive test. With the application of pertinent clinical information, the predictive value of a positive PET scan could be improved to 90.6%. This emphasizes the importance of appreciating that FDG uptake can occur in nonmalignant conditions, such as foci of infection/inflammation, and these must not be confused with active melanoma. These findings led to a change in the planned clinical management in patients after 16 of the 106 PET scans (15.1%) (Tyler et al. 2000).
PET/CT, which is now the norm in PET imaging as PET-only devices are no longer marketed for whole-body imaging, has been shown to be more sensitive than either PET or CT alone. In a retrospective study of 250 patients, PET/CT imaging detected significantly more visceral and non-visceral metastases than either PET alone or CT alone (98.7%, 88.8%, and 69.7%, respectively) (Reinhardt et al. 2006). In this series, PET/CT permitted more accurate staging of distant metastases (M) than either PET or CT alone (98.0% vs. 93.2% and 83.6%, respectively). PET/CT was also more accurate than CT for staging regional nodes (N) (98.4% vs. 86.4%). It should be noted that detection of small-volume nodal metastases is not particularly effective with PET/CT, but larger metastases and palpable nodes are quite well assessed with PET.
Similar improvements in imaging melanoma with PET/CT versus PET alone were reported in a prospective trial of 124 high-risk melanoma patients (Strobel et al. 2007a). In 53 of the 124 patients, metastases were found. In 46 of 53 patients with metastases, lesions had increased the FDG uptake. In seven patients with metastatic disease, metastases did have increased FDG uptake versus background (standardized uptake value [SUV] <1.5; n = 5) or had faint FDG uptake (SUV 2.5 and 2.9; n = 2), findings that were inconclusive with PET alone. These lesions were interpreted as metastases only with coregistered CT images. Sensitivity, specificity, and accuracy, respectively, of PET/CT for detection of metastases were good at 85%, 96%, and 91%, whereas those of PET/CT with dedicated CT interpretation were 98%, 94%, and 96% (p = 0.016). The authors concluded that dedicated analysis of coregistered CT images significantly improves the accuracy of integrated PET/CT for depiction of metastases in high-risk melanoma patients. It was noted in this report as well that false-negative results were seen in low-uptake lesions and also pulmonary metastases with PET alone, but these lesions were seen well on the CT component of PET/CT (Strobel et al. 2007a).
The Impact of PET/CT Imaging on Clinical Management
There is an increasing body of data showing that changes in management are often made based on the information provided by PET or PET/CT. In more than 25% of cases, when properly used, PET changed the way patients were managed (Falk et al. 2007). There are also data showing PET is of very limited value in assessing early-stage melanomas, because only rarely do changes in management occur as a result of the infrequency of detecting distant metastases (Clark et al. 2006). Forschner et al. evaluated the impact of FDG PET/CT on surgical management of 333 patients with stage III/IV melanoma, of whom surgical metastasectomy was planned in 107 patients (Forschner et al. 2017). Fifty-one percent of the patients evaluated had major changes in management as a result of the PET, notably with 32 patients moving to nonsurgical treatment due to disseminated metastases not allowing surgical resection. Schule et al. examined management changes in patients with stage III and IV melanomas, first based on CT and then with PET/CT data (Schüle et al. 2016). In about 50% of patients, PET/CT led to major management changes. Lewin et al. retrospectively evaluated the utility of PET/CT imaging in the detection of metastatic melanoma in stage III melanoma patients (Lewin et al. 2018). They found that in a group of 1170 patients, relapses were identified in 38% of patients and that 69% of the relapses were without symptoms. The positive predictive value of PET was lower in patients with stage IA melanoma, while higher in those with stage IIIC disease. A negative PET had about 80% NPV for short-term tumor recurrence. PET appears to be useful in monitoring treatment response, often paralleling but sometimes providing additional information to that obtained from serum assays such as S-100 (Strobel et al. 2007b). It does appear that for lesions that are visible on both PET and CT scans, the two techniques are comparable in assessing response, although the number of patients assessed is quite limited (Strobel et al. 2008). Using targeted therapy to BRAF mutant melanomas, McArthur et al. have shown that responses across lesions are relatively homogenous and rapid, often greater in percentage SUV change than change in tumor size (McArthur et al. 2012). As systemic treatments for melanoma improve, it is likely that PET will have an expanding role in management (Wong et al. 2017).
The introduction of checkpoint inhibitor therapies, either single or dual, has markedly improved the management of patients with metastatic melanoma. Anti CTLA-4 and anti PD-1 therapies of melanoma reactivate lymphocytes in tumors and can lead to tumor growth before response. This “pseudoprogression” of tumors can be very confusing, unless it is recognized, as it is not a sign of tumor progression. With CT, new response criteria have been developed for treatment assessment, the so-called IR RECIST criteria (Wolchok et al. 2009). Not uncommonly, new lesions can appear in the setting of response in other lesions. Wolchok described ipilimumab monotherapy resulting in four distinct favorable response patterns: (a) shrinkage in baseline lesions, without new lesions, (b) durable stable disease (in some patients followed by a slow, steady decline in total tumor burden), (c) response after an increase in total tumor burden, and (d) response in the presence of new lesions. All patterns were associated with favorable survival outcomes.
Thus, at centers where PET/CT is available, if stage IV metastatic disease is suspected, it is recommended that whole-body PET/CT be performed, along with a satisfactory-quality CT as part of the PET/CT, to achieve optimal diagnostic sensitivity for systemic metastases. Brain metastases are better detected by MRI, however, and therefore this should be part of the workup in high-risk patients. PET should not be used in low-risk patients, because the false-positive rate is too high and the technique becomes cost-ineffective.
Ultrasonographic techniques have high spatial resolution and deliver no ionizing radiation. Resolution and detection of deep structures are compromised, however, because of the absorption of sound waves by intervening tissues. Bone and air can markedly decrease the accuracy or render ineffective ultrasound methods for detecting deep-seated tumor foci. The use of ultrasound imaging is variable in different parts of the world, and its role in detecting metastatic disease is typically less in areas in which cross-sectional imaging such as MRI, PET, and CT is available. Ultrasound imaging does have high resolution and is perhaps best suited to detecting locoregional nodal metastases. In some settings ultrasound imaging is used to identify small suspicious lymph nodes, with confirmation that the nodes contain tumor by means of fine-needle aspiration biopsy (Schäfer-Hesterberg et al. 2007, 2011).
Brain MRI/Spinal MRI
MRI is probably the most sensitive tool for detecting brain metastases. Although CT is useful, the soft tissue resolution characteristics of MRI make it particularly valuable for detecting small lesions in the brain and spine. A variety of pulse sequences are used, and typically contrast material containing gadolinium is used to optimize sensitivity. In the past several years, it has been recognized that there is a risk of nephrogenic systemic fibrosis in patients with impaired renal function who received gadolinium in some form (Kuo et al. 2007). This is only a relative contraindication, but it must be considered in the choice of diagnostic agents. Nonetheless, the use of MRI is very important if metastases are suspected. Also, it has been appreciated that MRI signal characteristics, prominence on T2 signal, can aid in the detection of melanoma metastases and help distinguish them from other types of brain metastases (Gaviani et al. 2006). Although MRI can detect asymptomatic brain metastases in some patients, typically these are patients with stage IV disease who have been found to have disease in other organs. In patients who do not have stage IV disease, brain metastasis is very uncommon (Fogarty and Tartaglia 2006).
Although radionuclide imaging of the liver with 99mTc sulfur colloid, the brain with 99mTc diethylenetriamine pentaacetate (DPTA), and whole-body gallium-67 citrate scanning played a somewhat important role in the management of melanoma in the past, these scans are now seldom used because of the evolution of better diagnostic imaging approaches, notably PET/CT, CT, and MRI. Their use is not cost effective for routine screening or the targeted detection of occult metastatic melanoma because of their low diagnostic yield (Muss et al. 1979; Evans et al. 1980). However, the radionuclide bone scan remains a very sensitive whole-body tool for assessing the skeleton. Bone scans with 99mTc diphosphonates are far more sensitive than radiographs for detecting bone metastases. Bone metastases can be detected by MRI with good sensitivity as well, and the results are generally superior to those of a radionuclide bone scan, but the process of whole-body MRI is relatively slow and substantially more expensive. Radiographic studies of the bone and/or a radionuclide bone scan should be employed for specific symptoms or signs of bone disease. Other than MRI, or FDG PET/CT, a bone scan is probably the most sensitive test for skeletal metastatic disease, but a careful history and directed radiographs are necessary to ensure that areas of uptake do not represent old trauma or inflammation (Muss et al. 1979; Devereux et al. 1980). With bone scanning, very careful attention to patient history and appropriate use of correlative anatomic imaging are essential to ensure the suitable specificity of this method. Because many bone metastases begin in the marrow, there have been case reports of a greater sensitivity with PET than with bone scanning in detecting marrow involvement with melanoma (Aydin et al. 2005). When bone scan, MRI, and FDG PET were directly compared in a broad range of bone lesions in children, including few children having melanoma, sensitivities for lesion detection were 82%, 90%, and 71%, supporting the limited role of the bone scan and the importance of FDG PET (Daldrup-Link et al. 2001). Abe compared the bone scan with FDG PET and found comparable accuracies of bone scan (97%) and FDG PET (95%), with bone scan less sensitive for marrow lesions and the FDG PET more sensitive for marrow involvement with tumor (Abe et al. 2005).
Radiolabeled Monoclonal Antibodies
Although used investigationally in the United States and approved for a number of years in Europe, radiolabeled monoclonal antibodies do not have a significant role in the management of patients with melanoma. Essentially, they have been totally replaced, as far as radionuclide imaging is concerned, with PET/CT imaging with FDG, which has far better performance characteristics.
In summary, imaging of melanoma is essential to identify the full extent of stage IV metastatic disease. If imaging is applied in low-risk populations, a high false-positive rate will occur; however, applied in settings in which there is at least a moderate probability of metastatic disease (e.g., stage III melanoma that is not clinically apparent or known stage IV melanoma), the yield of imaging studies is higher, as is the change in management as a result of imaging. As imaging methods have improved, emerging data indicates that many patients who develop stage IV melanoma are asymptomatic, and the disease first becomes detectable by PET/CT or CT for visceral disease. For locoregional disease, ultrasound may be highly informative. In the past several years, it has become clear that PET/CT with the use of FDG is a very sensitive tool for detecting metastatic melanoma, but it is also clear that it has limitations, which are compensated for by the availability of other imaging approaches. The combination of PET/CT and MRI is required for a complete assessment of visceral and brain metastases of melanoma. Response monitoring can be performed by means of several methods, but since PET/CT is often the most sensitive imaging test, this is being used more frequently to assess response in stage IV disease. The PERCIST criteria can be applied to assess response, but appears to have limitations in assessing some aspects of response in patients receiving immune checkpoint therapy.
In patients who have symptoms, abnormal findings on physical examination or laboratory tests, and/or abnormal radiographic studies, a definitive diagnosis of metastatic melanoma can be accurately made only by biopsy. An excisional or needle biopsy is relatively easy to perform when the suspected metastasis is superficially located. More deeply situated lesions may also be diagnosed by a thin-needle biopsy (Rodrigues et al. 2000). In many circumstances, however, radiographic studies are sufficient to make a clinical diagnosis, especially if the metastases involve more than one site and the abnormality was absent on previous studies.
Cytologic examination of urine, sputum, or cerebrospinal, peritoneal, or pleural fluid or a bone marrow aspiration or biopsy examination also may yield a diagnosis of metastatic melanoma, especially when specific symptoms are referable to the relevant areas (Jaffer and Zakowski 2002; Khoddami 1993; Wasserstrom et al. 1982). The differential diagnosis often includes anaplastic or undifferentiated carcinoma and lymphoma. For example, in a patient who is a heavy cigarette smoker and has a history of a primary melanoma that was removed 5 years previously, an anaplastic lung lesion could be either a primary lung carcinoma or metastatic melanoma. Electron microscopy and special immunohistochemical stains are often useful in making the proper diagnosis (Michie et al. 1987; Jungbluth et al. 1999). Melanoma cells contain a unique organelle, the melanosome, that is involved in the biosynthesis of pigmentation and can be recognized by electron microscopy in most melanoma lesions.
Antibodies that recognize two melanoma antigens, S-100 and HMB-45, are frequently used to distinguish melanomas from non-melanocytic tumors (Cochran et al. 1982). In particular, HMB-45 is a molecular marker that is expressed in melanosomes and is encoded by a gene called gp100/pMel17 that can determine skin color (Ordóñez et al. 1988b). Another set of markers, including tyrosinase and tyrosinase-related proteins, are also expressed by melanosomes and can be useful for the diagnosis of melanocytic tumors (Itakura et al. 2008). Other markers are routinely used to identify carcinomas (cytokeratins) and lymphomas (immunoglobulin and B-cell antigens).
S-100 is expressed in melanocytic cells, including melanocytes, nevus cells, and primary and metastatic melanomas (Cochran et al. 1982, 1993). However, normal non-melanocytic cells, including neurons, pituicytes, Langerhans cells, dendritic cells, and macrophages, also express S-100. In addition, S-100 can be detected in breast tumors, salivary gland cancers, and other adenocarcinomas (up to 42% in some reports), as well as in peripheral nerve sheath tumors, liposarcomas, and cartilaginous tumors (Stroup and Pinkus 1988; Herrera et al. 1988; Daimaru et al. 1985). Thus S-100 is a marker that is quite sensitive (90%) but not specific for melanoma. Staining for S-100 has been used to diagnose difficult cases of primary melanoma, including amelanotic melanomas (Jundt et al. 1986; Gibson and Goellner 1988) and desmoplastic melanomas. However, its greatest use is in the diagnosis of undifferentiated metastatic tumors.
HMB-45 is expressed in the junctional component of melanocytic nevi, primary and metastatic melanoma cells, and fetal skin melanocytes (Gown and Vogel 1985). Adult epidermal melanocytes do not usually express HMB-45, but HMB-45-positive melanocytes can be detected near healing excision sites and over vascular lesions (e.g., hemangiomas) (Fogarty and Tartaglia 2006). In common melanocytic nevi, the deeper dermal cells are usually negative for HMB-45. In contrast, the intradermal component of dysplastic nevi and dermal nevus cells adjacent to a melanoma may be positive (Smoller et al. 1989). HMB-45 appears to be a more specific marker than S-100 (Ordóñez et al. 1988a, b) and is also sensitive for metastatic lesions (Ordonez et al. 1988a), including both melanotic and amelanotic metastases involving the central nervous system; however, its sensitivity for melanoma is lower than that of S-100 (Ordóñez et al. 1988b; Walts et al. 1988).
Melan-A, also termed as MART-1 (melanoma antigen recognized by T cells), is an antigen expressed in normal melanocytes of the skin and retina and is also expressed in all nevi and in most primary and metastatic melanoma (Busam et al. 1998). An anti-Melan-A murine monoclonal antibody, A103, has been tested on metastatic melanomas. The immunoreactivity of A103 was compared with that of HMB-45 in metastatic melanomas on formalin-fixed, paraffin-embedded tissues (Jungbluth et al. 1998). Of the 75 metastatic melanomas, 61 (81%) were A103 positive, and 56 (75%) were HMB-45 positive. Of the 19 HMB-45-negative lesions, eight were A103 positive; and of the 14 A103-negative lesions, three were HMB-45 positive. Eleven metastatic lesions, as well as two of ten primary melanomas, were dual negative. These negative cases consisted mainly of the spindle cell and desmoplastic melanoma variants. Of the positive cases, A103 showed homogeneous staining in a significantly higher proportion of cases than HMB-45 (72% vs. 52%). In addition, focal staining with less than 5% reactive tumor cells was seen more frequently in HMB-45 (12 of 56) than in A103 (5 of 61). Another study on sentinel lymph nodes also demonstrated decreased specificity of HMB-45 as compared to Melan-A for melanoma (Mahmood et al. 2002). These results indicate Melan-A to be one of the most sensitive markers in the diagnosis of metastatic melanoma.
SOX10 (Sry-related HMg-Box gene 10) is a nuclear transcription factor that plays an important role in differentiation of neural crest progenitor cells to melanocytes (Shakhova et al. 2012). SOX10 has been demonstrated to be a sensitive and specific marker of melanoma including spindle and desmoplastic subtypes (Mohamed et al. 2013). It is considered a reliable marker for supplementing traditionally used immunohistochemical stains. In a study to identify metastatic melanoma in sentinel lymph nodes, Willis BC et al. reported that SOX10 staining showed significantly greater staining intensity than S-100, HMB-45, and Melan-A (P = 0.000, 0.000, and 0.003, respectively) (Willis et al. 2015). Further, the percentage of tumor cells stained by SOX10 was significantly increased when compared with S-100, HMB-45, and Melan-A (P = 0.015, 0.000, and 0.001, respectively). Also, since SOX10 it is not expressed in dendritic cells in lymph nodes, it may be very useful for the detection of micrometastases in sentinel lymph nodes (Blochin and Nonaka 2009; Jennings and Kim 2011).
Over the past decade, tumor mutation profiling to inform targeted therapy has become routine practice in treatment decision-making for metastatic melanoma in the context of highly effective BRAF and MEK inhibitors. According to recommended guidelines, it is mandatory to perform mutation testing in advanced-stage melanoma (stage IIIC or IV) before the initiation of systemic treatment (Coit et al. 2016). Mutation testing can be carried out routinely on pre-treatment tumor biopsies or resection specimens. It is recommended that mutation testing should, at least, include all known activating BRAF mutations. For BRAF mutation analysis, many different methods are currently in use including conventional Sanger bidirectional sequencing, high-resolution melting (HRM) analysis followed by sequencing, pyrosequencing, and next-generation sequencing (NGS) with targeted gene panels (Ihle et al. 2014). Each BRAF genotypic test has its own inherit advantages and disadvantages that must be considered prior to using it to test a patient’s tumor sample. For fast turnaround time, as might be required for rapidly progressing patients with stage IV melanoma, immunohistochemistry (IHC) staining of BRAF V600E protein can also be performed where results are available in 24–48 h (Long et al. 2013). It has been observed that a clinical response to BRAF inhibitors can be evident within 1 day of initiation of therapy, and hence it is critical to rapidly detect positive cases to ensure timely initiation of BRAF-inhibitory treatment.
In recent years, several companion diagnostic tests have become available that are often co-developed with a drug or drug class and are recommended for the prescription of the specific drug therapy. The Cobas 4800 BRAF-V600 mutation test was one such test, which was developed as a companion diagnostic test for use in the clinical trials with vemurafenib and was used widely after vemurafenib was approved as a treatment for advanced melanoma. The Cobas test is a real-time PCR test that detects BRAF V-600E mutation with greater sensitivity and specificity than Sanger sequencing (Lopez-Rios et al. 2013). Another test is CE THxID™-BRAF diagnostic test based on real-time PCR technology, aimed at simultaneous detection of both BRAF V600E and V600K mutations in patients intended for possible treatment with dabrafenib and trametinib. The major advantages of real-time PCR are faster performance, better reproducibility, and lower cost compared with traditional genomic sequencing methods.
Sites of Distant Metastases
Signs and symptoms involving specific metastatic sites and diagnostic methods are described in detail in the following sections. The frequency of metastases to various sites, both clinically and at autopsy, is displayed in Table 1.
Skin, Subcutaneous Tissues, and Distant Lymph Nodes
Melanoma and carcinoma of the breast, colon, and lung are the most common primary solid tumors that develop cutaneous metastases (Lookingbill et al. 1993; Schwartz 1995; Rolz-Cruz and Kim 2008). In patients who have metastatic melanoma, the skin and subcutaneous tissues are the most common sites of distant metastasis, and metastases at these sites are often the first sign of hematogenous spread (Balch et al. 1983; Finck et al. 1983).
Skin and subcutaneous metastases are typically 0.5 to 2.0 cm in diameter and are readily detectable on physical examination. Metastases can be single or multiple and occur anywhere on the body. They are usually firm, round, and pigmented, but the pigmentation may not be visible if the nodule is located deep in the subcutaneous fat. They may be flat small papules on the skin but may be elevated with a nodular or fungating appearance. When the first subcutaneous nodule is detected, it may appear to be isolated, but several more nodules are often detected within a short time. In rare cases cutaneous metastases may have a zosteriform distribution (Itin et al. 1995; Chaudhary et al. 2013).
Soft tissue metastasis can be detected by PET/CT imaging in patients with melanoma more frequently than in those with other tumors. A study of 500 consecutive cancer patients showed soft tissue metastases in 4 (9.8%) of 41 patients with melanoma versus only 5 (1.1%) of 459 of patients with other malignancies (Nguyen et al. 2007). Although detection of these soft tissue lesions provided a biopsy site, in some patients the results of this test did not change either the staging or their treatment.
Distant lymph node metastases can occur in any anatomic region. Superficial nodal metastases are easily detected by physical examination. Adenopathy within the thorax can often be detected on a chest radiograph, with CT scans providing confirmation. Abdominal nodal metastases are usually detected by CT scans. Nodal metastases involving the retroperitoneum, pelvis, or mediastinum can become quite large and may give rise to symptoms by invasion or displacement of adjacent structures and tissues (Feldman and Kricun 1979).
Soft tissue metastases involving skeletal muscle have also been reported. A single-institution review of 118 patients with metastases involving skeletal muscle, alone or in conjunction with subcutaneous tissue, reported that the leading histologic diagnosis was melanoma, which occurred in 20 cases (16%) (Plaza et al. 2008). Other primary sites that were documented in more than 10% of cases included lung, breast, kidney, and colon and rectum. In 27% of patients the soft tissue metastasis was the initial manifestation of the disease, making diagnosis difficult. Furthermore, many of these tumors displayed histologic features that resembled soft tissue sarcoma, requiring the use of immunohistochemical stains to establish diagnosis.
Lung, Pleura, and Mediastinum
The lungs and pleura are the second most common initial sites of melanoma metastasis. Almost all patients who have disseminated melanoma develop metastases in the chest before they die. In a series of 7564 patients with melanoma who were seen at Duke Comprehensive Cancer Center between 1979 and 1990, a total of 945 patients (12%) developed pulmonary metastases (Harpole et al. 1992). In univariate analysis, the statistically significant predictors of pulmonary metastasis were those that predicted recurrence of primary melanoma in general. It is interesting to note that African-American descent appeared to be associated with early development of pulmonary metastasis.
Pulmonary metastases (especially small parenchymal nodules) are usually asymptomatic at the time of detection on chest radiographs. Eventually, however, they tend to cause one or more of the following symptoms: persistent cough, hemoptysis, dyspnea, or chest pain. An irritating, dry, nonproductive cough may progress to hemoptysis and coughing up clots of melanin-stained tumor (Braman and Whitcomb 1975). These symptoms are usually caused by small submucosal bronchial deposits that become enlarged and ulcerated. They may also be caused by one or more of the following: (1) a large extrabronchial deposit that compresses and then invades a major bronchus; (2) pleural effusion, hemothorax, or, less frequently, a pneumothorax; or (3) pleural or chest wall invasion by the metastases (Harpole et al. 1992; Gibbons and Devig 1978; Yeung and Bonnet 1977).
Factors affecting survival in patients with pulmonary metastases have been extensively reviewed. Petersen et al. (Petersen et al. 2007) reported 1720 patients with pulmonary metastasis from melanoma culled from the Duke University database. In this series, which was not restricted to patients who were initially seen with pulmonary metastases, median survival was 7.3 months from initial detection of pulmonary metastasis. Significant predictors of survival from the multivariate analysis included nodular histologic type, disease-free interval, number of pulmonary metastases, presence of extra-thoracic metastases, and the performance of pulmonary metastasectomy. Surgery appeared to be associated with a survival advantage of more than 12 months in patients with a disease-free interval longer than 5 years and of 10 months in patients without extrathoracic metastases, possibly supporting a therapeutic role for surgery in these select patient populations. In another series Neuman et al. (2007) analyzed the outcome of 122 patients with melanoma treated at Memorial Sloan-Kettering Cancer Center who had pulmonary metastasis as their initial site of metastatic disease. Median survival was 14 months, and 5-year survival was 8%. Factors predictive of improved survival were solitary pulmonary metastasis and the absence of extrapulmonary disease. Among treatment factors, metastasectomy also appeared to be an independent predictor of favorable outcome with a median survival of 40 months for those undergoing surgery versus 13 months for those not selected for this procedure. This difference was noted despite the fact that 88% of the patients undergoing surgery had a recurrence of melanoma at a median of 5 months after the surgery, implying that the improvement in outcome was more related to the biology of the disease than the impact of the surgery. The role of surgery in metastatic melanoma is discussed in more detail (see also “Role of Surgical Resection of Distant Melanoma Metastases”). Unfortunately response to systemic therapy did not correlate with a survival difference.
Brain and Spinal Cord
Melanoma is the third most common tumor to metastasize to the brain, with breast and lung cancers being the first and second, respectively (Zimm et al. 1981). At autopsy, cerebral metastases are identified in 36–54% of patients with advanced melanoma (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a). They are also common in clinical series (Balch et al. 1983; de la Monte et al. 1983; Budman et al. 1978). Brain metastases are one of the most common causes of death from melanoma, ranging from 20% to 54% of patients in different series (Lee 1980; de la Monte et al. 1983; Budman et al. 1978; Fife et al. 2004; Skibber et al. 1996). Among those with documented brain metastases, these lesions contribute to death in up to 95% of cases (Sampson et al. 1998). The brain is the initial site of metastasis in 12–20% of patients in different clinical series of stage IV melanoma and may be the only site of metastatic disease. However, the presence of brain metastasis is more commonly associated with evident widespread visceral disease (Amer et al. 1978; de la Monte et al. 1983).
In one series of patients with cerebral metastases of melanoma, the median interval between the diagnosis of the primary melanoma and the detection of the cerebral metastases was 22 months (Munoz et al. 1993); in another series it was found that 83% of patients developed their cerebral metastases within 5 years of diagnosis of the primary melanoma (Amer et al. 1978). However, rarely brain metastases may develop many years (18 or more) after resection of a primary melanoma. Approximately 25–40% of patients who have cerebral metastases will have only a solitary lesion (Amer et al. 1978; Gupta and Brasfield 1964a; Ewend et al. 1996; Bullard et al. 1981). In patients with brain metastasis, the cerebrum is involved most frequently (usually the frontal lobe), followed by the cerebellum, base of the brain, and spinal cord, correlating well with the areas that receive the highest amount of blood flow. The two hemispheres are involved with equal frequency. In a series of 30 patients who had cerebral melanoma metastases, the metastases were classified by size as measured by CT (Merimsky et al. 1992). Lesions between 1.1 and 4 cm were the most common. Lesions >4 cm in diameter were the least common and were usually solitary. Asymptomatic patients usually had metastases ≤1 cm in diameter, whereas those who had nonspecific complaints or neurobehavioral changes usually had metastases between 1.1 and 4 cm in size.
Although all patients with melanoma are at risk for metastasis to the brain, characteristics that are associated with increased risk of systemic metastases are generally the same as those that correlate with the subsequent development of brain metastasis. These include thick, nodular, and ulcerated primary tumors, head and neck areas, acral lentiginous or mucosal tumors, lymph node involvement, male gender, presence of other visceral disease, and elevated serum LDH levels at diagnosis (Sampson et al. 1998; Saha et al. 1994; Bedikian et al. 2011; Zakrzewski et al. 2011). Studies have demonstrated that hyperactivity of signal transducer and activity of transcription 3 (STAT3) play a critical role in promoting melanoma invasion and metastasis, and tumor STAT3 expression within primary melanomas is predictive of central nervous system metastasis (Niu et al. 2002; Xie et al. 2006). Based on these observations, various preclinical studies are testing novel strategies to target STAT3 expression in the tumor, and a trial evaluating the efficacy of WP1066, a novel inhibitor of STAT3 signaling, is currently ongoing for melanoma patients with brain metastases.
Headache and mental deficits are the most common symptoms in patients who have brain metastases. The headache characteristically begins as an early morning pain. As the disease progresses and intracranial pressure increases, the headache persists for a longer time during the day and becomes more severe. Headaches are usually generalized but may be slightly worse in the frontal or occipital regions and frequently can be associated with visual changes or early morning nausea. Seizures are more common in patients who have melanoma metastases than in those who have brain metastases of other tumors; approximately 25% of patients have seizures. Although seizures are more common later in the course of the disease (Amer et al. 1978; Munoz et al. 1993; Konstadoulakis et al. 2000), they are the presenting symptom in approximately 15% of cases (Hulick 1995). An unusual feature of cerebral metastases of melanoma is their propensity to hemorrhage, which is greater than that in cerebral metastases of other histologic types. Hemorrhage occurs in 33–50% of melanoma patients who have brain metastases (Weisberg 1985; Ginaldi et al. 1981). Ginaldi et al. (1981) observed CT evidence of hemorrhage in most of the 93 patients they examined, and the hemorrhagic component increased with increased size of the metastases. Lesions of the posterior fossa are associated with hydrocephalus in 33% of cases.
The most common physical sign of brain metastases is a focal neurologic defect (Amer et al. 1978; Pennington and Milton 1975). The presence of papilledema is a helpful sign, but its absence is not useful diagnostically. Occasional patients will have rapidly progressing audio-vestibular symptoms, including tinnitus, vertigo, and/or sensorineural hearing loss or facial nerve deficits; these findings are typical of metastasis to the cerebellopontine angle (Arriaga et al. 1995; Kingdom et al. 1993). Clinically and radiographically these lesions resemble acoustic neuromas. Diabetes insipidus attributable to meningeal, hypothalamic, or pituitary involvement may be the heralding sign in a few patients (Ten Bokkel Huinink et al. 2000). It is also common for patients to develop subarachnoid or intracerebral hemorrhages (Ginaldi et al. 1981; Bremer et al. 1978; Maiuri et al. 1985).
Historically, melanoma patients with brain metastases generally have had a poor prognosis. In two historical series totaling almost 1400 patients, the median survival was 4 months, and 1-year survival rates were 9% and 19%, respectively (Fife et al. 2004; Sampson et al. 1998). However, with the use of stereotactic radiosurgery (SRS) to treat brain metastasis as well as the success of targeted therapy and checkpoint inhibitor immunotherapy, there has been significant improvement in survival and quality of life for melanoma patients with brain metastases. Also, increased use of MRI aimed at detecting brain metastasis during initial screening may have contributed to improved control and prolonged overall survival. A recent study observed that the median overall survival from the time of diagnosis of melanoma brain metastasis was 22.7 months in patients diagnosed after 2011 as compared with 8.5 months and 7.5 months for patients diagnosed between 2009 and 2010 and 2000 to 2008, respectively (p = 0.002) (Sloot et al. 2018). Another recent retrospective study noted a median overall survival of 11 months, with 1- and 2-year overall survival rates of 50% and 27%, respectively, when SRS and immune or targeted therapy were combined for treatment of melanoma patients with brain metastases (Gaudy-Marqueste et al. 2017).
Favorable prognostic signs include the presence of a single brain metastasis, with no other visceral metastatic disease, an initial presentation with a metastasis to the brain and a normal serum LDH (Fife et al. 2004; Sampson et al. 1998; Davies et al. 2011; Eigentler et al. 2011). In contrast, multiple brain lesions, extensive visceral metastases, a primary lesion of the head and neck region, elevated serum LDH levels, or a greater number of neurological symptoms carry an unfavorable prognosis (Sampson et al. 1998; Zakrzewski et al. 2011).
Melanoma has a propensity to spread to the leptomeninges and leptomeningeal melanomatosis usually carries a bleak prognosis. Leptomeningeal involvement is found at autopsy in more than 50% of cases, yet it is demonstrated infrequently in contrast-enhanced CT scans (Munoz et al. 1993). The clinical presentation of meningeal metastases is varied and often nonspecific and includes headache, signs of meningeal irritation, and confusion. Motor dysfunction and cranial nerve palsies are also common. The development of unexplained sensorineural hearing loss, tinnitus, or vertigo resulting from eighth cranial nerve involvement is a frequent clinical presentation of leptomeningeal disease, as well cerebellopontine angle metastasis.
Other patients develop radicular pain involving the lower back or legs, bowel, and/or bladder dysfunction – that is, cauda equina syndrome – due to “drop” metastases involving the cauda equina. Although cauda equina involvement is the most common spinal manifestation of leptomeningeal disease, nerve roots in the thoracic and cervical spinal cord can also be involved. Cytologic examination of the cerebrospinal fluid is usually positive, although a centrifuged sample must be examined and multiple specimens taken on different days are frequently necessary (Amer et al. 1978). Even so, cerebrospinal fluid cytology is normal in up to 10% of patients in whom the diagnosis is confirmed at the time of autopsy (Chamberlain et al. 1990).
Gadolinium-enhanced MRI is the most sensitive imaging technique for the diagnosis of meningeal carcinomatosis, being positive in more than 70% of cases (Chamberlain et al. 1990; Chamberlain 1966; Sze et al. 1989). MRI is probably more sensitive than cytologic examination of a single cerebrospinal fluid specimen, although it is less specific because false-positive cytologic findings are rare. Therefore patients who have symptoms suggestive of leptomeningeal metastasis should be evaluated with a gadolinium-enhanced MRI. Typical MRI findings in the brain include thin, diffuse leptomeningeal contrast enhancement, multiple nodular deposits in the subarachnoid space, cerebellar folia, or cortical surface, and tumor masses, especially at the base of the brain, with or without hydrocephalus. Occasionally frank leptomeningeal involvement is not seen on MRI, but bulky subependymal disease or multiple small sulcal metastases suggest the diagnosis. The increasingly frequent use of gadolinium-enhanced MRI for staging of patients with metastatic melanoma has led to the detection of leptomeningeal disease in many patients in advance of symptoms developing. In patients with typical symptoms and no localizing findings on MRI, the diagnosis may be suggested by PET scan demonstrating diffusely diminished glucose utilization in an otherwise normal-appearing brain (Roelcke and Leenders 2001). In the spinal cord, MRI can show linear enhancement of the entire cord and linear or nodular enhancement of the cauda equina. Occasionally clumping of nerve roots at the cauda equina suggests the diagnosis, even if contrast enhancement is not seen.
Melanoma is one of the most common causes of metastatic disease involving the gastrointestinal tract (Liang et al. 2006). Although this diagnosis is made while the patient is still alive in only 10% of cases, gastrointestinal metastases have been observed in more than 50% of patients at postmortem examination (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a, b; Prakoso and Selby 2007). Among patients who have gastric metastases, melanoma is a common cause (Bognel et al. 1992; De Palma et al. 2006). The average time from diagnosis of primary cutaneous melanoma to gastrointestinal metastasis is around 52 months (Patel et al. 2014).
The anatomic distribution of melanoma metastases in the gastrointestinal tract has been shown in many studies (Reintgen et al. 1984; Goldstein et al. 1977). The small intestine is the most common site of gastrointestinal metastases (26–58% of all patients); the stomach (7–26%) and colon (14–28%) are less common sites. In unusual cases, gastrointestinal metastases involves the esophagus (3–9%) or anus (1%). Although solitary metastases occasionally occur in the gastrointestinal tract, multiple lesions in each affected organ and synchronous lesions in multiple organs are more common. Further, metastases usually occurs late in the course of disseminated disease and is associated with a poor prognosis (median survival of 4 to 6 months) (Liang et al. 2006). A surgical series from the John Wayne Cancer Institute, in which many patients had isolated gastrointestinal tract involvement, suggested that such disease might represent a uniquely favorable metastatic presentation (Barth et al. 1995). Furthermore, colonic metastases, although rarely detected clinically, in one study were reported to develop after an unusually long disease-free interval (median 7.47 years) and to be well palliated with surgery, with 1-year and 5-year survival rates of 37% and 21%, respectively (Tessier et al. 2003).
The most characteristic lesions are exophytic or polypoid submucosal nodules that may be umbilicated or have undergone central cavitation (Goldstein et al. 1977). Less characteristic are shallow ulcerative lesions that usually occur in the stomach and resemble gastric ulcers and solitary infiltrative lesions that occur most often in the small bowel and are associated with a more favorable prognosis (Reintgen et al. 1984). In approximately half of the cases, the metastases were pigmented, whereas the remainder were fleshy and amelanotic.
Gastrointestinal metastases are frequently not detected clinically and diagnosis is often made during surgery or endoscopy. This is because associated symptoms are generally absent, vague, or nonspecific. Gastric metastasis may manifest as asymptomatic iron deficiency anemia resulting from chronic indolent bleeding (Basagoiti et al. 1992). Because symptoms caused by early involvement of the gastrointestinal tract usually are vague and subtle, the physician should question the patient carefully about such symptoms. In patients with melanoma, persistent nonspecific complaints such as epigastric discomfort, nausea, anorexia and weight loss, or the unexplained development of anemia should lead to a suspicion of gastrointestinal metastasis. Dysphagia may indicate esophageal involvement (Wood and Wood 1975; Schneider et al. 1993). Large metastases are sometimes palpable on abdominal examination, or they can be visualized by sigmoidoscopy, colonoscopy, or endoscopic retrograde cholangiopancreatography. CT scans can frequently miss involvement because of difficulty in distinguishing between bowel and tumor. PET/CT imaging can be helpful in circumventing this obstacle. Capsule endoscopy can be used to detect the presence and extent of metastatic involvement more reliably than conventional investigations and should be considered in the workup of patients with suspected metastases that have remained undetected by other means (Prakoso and Selby 2007). Biopsy of suspected lesions following endoscopic or surgical intervention (laparotomy) often clinches the diagnosis. Special immunohistochemical stains, including S-100 and HMB-45, are particularly useful in supporting the diagnosis.
Acute and potentially catastrophic symptoms occur occasionally. The most common clinical manifestations are caused by (1) chronic bleeding, with anemia, anorexia, and weight loss; (2) acute bleeding, with hematemesis or melena; or (3) obstruction of the small bowel, with abdominal pain, nausea, and vomiting (Goodman and Karakousis 1981; Klausner et al. 1982; Fawaz and Hill 1983; Karakousis et al. 1974). Bowel obstruction was the immediate cause of death in 8% of patients in one autopsy series. Uncommon clinical presentations include bowel perforation (Goodman and Karakousis 1981; Klausner et al. 1982; Fraser-Moodie et al. 1976; Den Uil et al. 2014), peptic ulcer symptoms (Booth 1965; Goldman et al. 1977b), malabsorption, and intractable diarrhea (Goldstein et al. 1977). Occasionally hemoperitoneum can develop after hemorrhage from subperitoneal deposits (Klausner et al. 1982; Goldman et al. 1977a).
Gastrointestinal bleeding usually emanates from the small intestine or, less frequently, from the stomach, and it can occur intermittently. Serial stool guaiac tests should be performed in patients with anemia and those who report a change in bowel habits. Intussusception is a frequent cause of obstructive symptoms and other abdominal complaints (Fawaz and Hill 1983; Karakousis et al. 1974; Mucci et al. 2007; Butte et al. 2009). Intussusception is characterized by an insidious onset and usually follows a chronic or subacute course. The triad of abdominal cramps, nausea without vomiting, and abdominal distention was the most consistent symptom complex in two series (Fawaz and Hill 1983; Karakousis et al. 1974). The signs of a palpable abdominal mass, bloody stools, and abdominal tenderness are typically lacking in these patients. The symptoms of intussusception may be confounded if the patient also has symptoms that could be attributed to gastrointestinal toxicity from systemic therapy. Conservative measures can frequently relieve the intussusception, and surgical resection usually provides satisfactory short-term and occasional long-term palliation (Mucci et al. 2007).
Liver, Biliary Tract, and Spleen
Hepatic metastases were diagnosed in 10–20% of patients who had metastatic melanoma, in various clinical series, but they are present in most patients at autopsy (Patel et al. 1978; Lee 1980; Balch et al. 1983; Gupta and Brasfield 1964a; Amer et al. 1979). Liver metastases occur typically in the clinical setting of widespread metastatic disease, and liver-only metastasis is very rare. The prognosis of patients who have liver metastases is generally poor, with an average life span of only 4 to 6 months in historical studies (Balch et al. 1983, 2001b; Amer et al. 1979; Karakousis et al. 1983). However, emergence of various locoregional treatment options as well as systemic immunotherapy and targeted therapy options to reduce the burden of disease has improved the prognosis in the recent past. Symptoms and physical signs from early liver metastases are uncommon, and most patients are diagnosed by the initial screening tests or follow-up surveillance scans. Patients may experience anorexia and abdominal fullness with weight loss, followed within weeks by general lassitude and debility. Loss of appetite may precede a clinically palpable liver by at least 1 month. Conversely, patients may have palpable hepatomegaly and feel perfectly well. As the hepatic metastases progress, distressing nausea and vomiting may develop. Patients may also report a vague “dragging” sensation or “fullness” in the upper abdomen. Fever caused by hepatic necrosis, often accompanied by drenching sweats, is not uncommon. Severe acute abdominal pain in the right upper quadrant may develop because of hemorrhage into a necrotic hepatic metastasis. Clinical jaundice is not common until the condition is far advanced. Hepatic failure is a rare complication of liver metastases (Lesur et al. 1992). On physical examination, nodules on the anterior and inferior surfaces of the liver or generalized hepatomegaly may be palpated. However, most patients with liver metastases do not exhibit hepatomegaly.
Melanoma is one of the few tumors that metastasize to the spleen (Amer et al. 1979; Kamaya et al. 2006; Reccia et al. 2015), and it does so commonly. It may rarely cause splenomegaly but is more often diagnosed as an incidental finding on an abdominal CT scan or FDG PET/CT (Felix et al. 1976; Compérat et al. 2007; Görg and Hoffmann 2008; Sen et al. 2013). Most patients who have splenic metastases (up to 88%) have concomitant liver or pancreatic metastases. In patients with solitary splenic metastases, ultrasound-guided fine-needle biopsy of the spleen can be performed safely and is highly likely to reveal the underlying cause (Cavanna et al. 2007). Surgical metastasectomy for solitary splenic metastasis has been shown as an effective treatment option, associated with a favorable long-term survival outcome (De Roeck et al. 2017). Patients may have symptoms of massive hemorrhage from a ruptured splenic metastasis; reports of spontaneous splenic rupture are quite rare, however (Buzbee and Legha 1992).
Bone metastases occur infrequently in most clinical series (11–17%) but are observed more commonly in autopsy series (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a; Amer et al. 1979; Fon et al. 1981) (see Table 1). Skeletal metastases is usually seen in patients who have widespread metastatic disease, but occasionally these are the first evidence of disease recurrence (Fon et al. 1981; Stewart et al. 1978).
Approximately 80% of bone metastases involve the axial skeleton with the spine being the most common (Fon et al. 1981; Stewart et al. 1978; Spiegel et al. 1995; Gokaslan et al. 2000; Zekri et al. 2017). When metastases involve a vertebral body, compression fracture often results, leading to neurologic symptoms such as radicular back pain, paresthesia or paresis of the legs, or urinary retention (Zekri et al. 2017). Only about 10% of lytic lesions occur in weight -bearing bones and have the potential for pathologic fracture (Fon et al. 1981; Stewart et al. 1978). Occasionally metastasis will occur beneath an articular surface, causing a joint effusion, or it will occur as a solitary lesion in a toe or finger (Fon et al. 1981; Shenberger and Morgan 1982; Gelberman et al. 1978).
In asymptomatic patients with melanoma, particularly those who have stage I, II, or III disease, the yield of a bone scan is too low to justify its use as a screening procedure (Muss et al. 1979; Felix et al. 1976). Bone metastases are most frequently diagnosed in symptomatic patients, although occasionally they are seen incidentally on radiographs (e.g., rib metastases on routine chest radiograph) or CT scans or PET/CT. The pain caused by bone metastases is typically persistent, progressive, and localized. It is penetrating in nature and can be quite severe in intensity.
Bone metastases can be imaged radiographically or scintigraphically. The radionuclide bone scan has traditionally been the initial test for evaluating suspected bone metastases. Its remarkable sensitivity, which is reported to be 50–80% greater than that of radiographs alone, allows detection of skeletal lesions much earlier than skeletal radiography (Felix et al. 1976; Fon et al. 1981). Bone scan abnormalities are nonspecific, however, and they must be correlated with radiographic studies and the patient’s history (e.g., fractures, trauma, arthritis) to distinguish between benign and malignant lesions. Plain radiographs are relatively insensitive for detecting skeletal metastases of melanoma but remain useful for evaluating specific symptomatic regions. MRI represents the most sensitive method for localizing symptomatic bone lesions and evaluating the extent and stability of the involved bones. MRI can be particularly useful in evaluating the spine, pelvic bones, or skull. In some cases it may be necessary to perform a bone biopsy to establish the diagnosis before initiating treatment.
Some of the radiographic patterns of bone metastases have been described (Felix et al. 1976; Fon et al. 1981). Typical bone metastases from melanoma are medullary in location (91.6%) and osteolytic in nature (87.5%), and they provoke little, if any, bone formation. The margins of the lytic regions are usually ill defined. Bone resorption in melanoma metastases has been attributed to the actions of tumor-associated macrophages, which are capable of differentiating into osteoclasts under the influence of tumor cells and associated fibroblasts (Lau et al. 2006). Although bone metastases of melanoma cannot be easily distinguished from other osteolytic metastases by routine imaging, PET and PET/CT scans more frequently show increased glucose uptake in melanoma metastases than those from many other solid tumors and can be positive even in the setting of negative bone scans (Aydin et al. 2005).
Complications of bone metastases of melanoma include cortical erosion and destruction (46%), pathologic fractures (22%), and soft tissue involvement (12%). Hypercalcemia is also commonly observed complication in patients with evidence of multiple bone metastases (Attia et al. 2003). Atypical bone metastases of melanoma may have a mixed osteolytic-osteoblastic pattern (10%) but are rarely completely osteoblastic. Other atypical appearances include intense trabecular rarefaction without a discrete lesion (3%) or the presence of a sclerotic rim and periosteal reaction (12%). Atypical metastases may radiographically resemble osteogenic sarcoma. The radiologic patterns of the bone metastases change in response to treatment. Recalcification, sclerotic rim formation, and periosteal reactions are commonly observed in responding lesions (Fon et al. 1981).
The mean life span of melanoma patients who have bony metastases was found to be around 4 to 6 months in historical studies, and it was even shorter when other metastatic sites are present (Balch et al. 1983; Fon et al. 1981; Stewart et al. 1978). In a series of 114 patients who had clinically or radiographically evident spinal metastases compiled from a cohort of 7010 melanoma patients, the median survival was 86 days, and survival was reduced in patients who also had more than two additional metastatic sites (Spiegel et al. 1995). However, effective palliation can today be achieved in many patients with either systemic or local therapy.
Skeletal metastases may involve bone marrow, sometimes extensively in the vertebrae or sternum (Anner and Drewinko 1977; Einhorn et al. 1974). Fever and/or cytopenia may accompany bone marrow metastases. Rarely bone marrow involvement may be the initial presentation of the disease (Basu et al. 2002; Jain et al. 2007). Bone marrow aspiration yielded metastatic disease in 9% of patients in one series of patients with stage IV melanoma and was positive in 2 of 14 patients with stage III melanoma in another series (Einhorn et al. 1974). Studies of S-100 beta in bone marrow aspirates were unhelpful in distinguishing tumor-involved bone marrow from normal bone marrow and, in contrast to peripheral blood, demonstrated no prognostic value (Faye et al. 2008).
Kidneys and Urinary Tract
Metastatic melanoma is frequently found in the kidneys and urinary tract at autopsy. The kidneys contain metastatic melanoma in 35–48% of autopsied patients, the bladder in 13–18%, and the ureters and prostate in only 2–5% (Patel et al. 1978; Lee 1980; Nambisan et al. 1987; Neuman et al. 2007; Gupta and Brasfield 1964a). Metastases at these sites rarely cause clinically recognizable symptoms until the terminal stages of the disease. Historically, death usually occurred within 1 to 4 months after the clinical diagnosis was established. However, solitary metastases do occur occasionally and may be amenable to surgical treatment (Shimko et al. 2007). In addition, symptomatic metastases can sometimes be palliated with radiation therapy.
A detailed history, physical examination, and urinalysis can occasionally help to identify metastases involving these organs, especially in patients who have known stage IV disease. The most common symptom prompting a urinary tract investigation is gross or microscopic hematuria (Gupta and Brasfield 1964a; Amer et al. 1979). An abdominal or rectal mass, symptoms of urinary tract infection, melanuria, melanospermia (Smith et al. 1973), or azotemia may also be a presenting feature. Metastases in the renal pelvis may be large enough to cause obstruction with hydronephrosis or bleeding. Lumbar pain can be caused by large metastatic deposits in the kidney. Infiltration of the metastatic renal mass into the inferior vena cava has also be observed (Boughan et al. 2009). Bladder metastases tend to be more symptomatic than metastases at other sites. Patients usually have hematuria, dysuria, or urinary frequency (Tolley et al. 1975). Prostatic and urethral metastases can also cause hematuria, dysuria, or urinary hesitancy (Berry and Reese 1953).
CT scans or cystoscopy can be used to diagnose the metastases in most cases. Renal metastases usually present as small (3–10 mm) cortical nodules that are asymptomatic incidental findings on CT images. Occasionally they can present as solitary large masses mimicking primary renal carcinoma (Shimko et al. 2007). Metastases in the renal pelvis and ureters are subepithelial and can cause a confusing radiologic appearance with a single or multiple filling defects (Goldstein et al. 1974; Nakazono et al. 1975). Although patients who have bladder metastases commonly are seen with multiple subepithelial pedunculated or sessile lesions, solitary lesions can occur (Meyer 1974). The diagnosis can be made occasionally by cytologic examination of the urine demonstrating melanoma cells (Woodard et al. 1978). Definitive diagnosis can also be established by immunohistological staining of the renal biopsy specimen.
Heart and Pericardium
Although autopsy series showed involvement of the heart in 40–55% of patients, antemortem recognition of cardiac metastatic disease has been reported in less than 1% of historical clinical series (Glancy and Roberts 1968; Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a; Einhorn et al. 1974; Abraham et al. 1990). However, with prolongation of overall survival of metastatic melanoma patients and technical advances in imaging techniques, there has been more frequent detection of cardiac metastases in the recent past. The right side of the heart is affected more frequently than the left side (Tse et al. 2017; Zitzelsberger et al. 2017). Also, metastatic involvement usually occurs in the pericardium or myocardium (rarely in the endocardium or valves). Solitary metastatic lesions are rare (Abraham et al. 1990; Savoia et al. 2000). Cardiac metastatic disease is usually clinically silent. Asymptomatic massive involvement of the myocardium has been termed “charcoal heart” (Waller et al. 1980). Clinical signs and symptoms, when present, are nonspecific and include pericardial effusion with or without tamponade (Kutalek et al. 1985), superior vena cava syndrome (Emmot et al. 1987), cardiac arrhythmias including sustained ventricular tachycardia (Sheldon and Isaac 1991), congestive heart failure (Schneider et al. 1994), and right ventricular inflow or outflow obstruction (or both) because of intracavitary metastasis (Emmot et al. 1987). Almost all patients who have proven cardiac or pericardial metastases have fairly advanced disease with metastasis to multiple visceral organs and bear a poor prognosis with median survival of a few months. A recent retrospective study reported that 11 patients survived only 12 months after diagnosis of cardiac melanoma metastasis (Tse et al. 2017).
In symptomatic patients a definitive diagnosis should be pursued to exclude other, more treatable processes. No reproducible electrocardiographic patterns are helpful in diagnosing cardiac metastases. A two-dimensional echocardiogram and a cardiac CT scan with contrast enhancement are helpful for detecting intracardiac or pericardial metastases (Kutalek et al. 1985). Cardiac MRI can be particularly useful in defining the extent and location of cardiac involvement (Gindea et al. 1987). Pericardiocentesis is a useful diagnostic tool in patients who have a pericardial effusion demonstrable by echocardiography or CT and can be therapeutic in patients with symptoms of tamponade.
Tissue diagnosis of metastatic cardiac lesions is now typically obtained with fluoroscopic or transthoracic echocardiographically guided ventricular endomyocardial biopsy. Results have been excellent and morbidity has been low (Flipse et al. 1990). Transesophageal echocardiography has been used to guide transvenous biopsy (Malouf et al. 1996). Transesophageal echocardiography, in addition to providing excellent anatomic definition and location of the intracardiac mass, provides precise localization of the biopsy forceps and confirms that biopsy specimens are obtained from the correct structure, thus avoiding the potential for serious complications. In the modern era, treatment of cardiac melanoma metastases essentially involves systemic therapy, although extra attention must be given to the possibility of life-threatening arrhythmia or cardiac rupture during treatment (Licciardello et al. 2017).
Pancreatic metastases are commonly found in melanoma patients at autopsy (in 38–53% of reported cases) (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a; Einhorn et al. 1974). The metastases are usually multiple discrete lesions that rarely replace functioning glandular tissue or cause significant ductal obstruction. Metastases in the pancreas may or may not cause symptoms and are usually diagnosed as an incidental finding by CT, PET, or ultrasound imaging (Mehta and Omer 2017). However, pancreatic metastases may very occasionally appear as an abdominal mass, causing obstruction of the duodenum and/or biliary tract (Larsen et al. 2013) or gastrointestinal bleeding. Large pancreatic metastases can cause troublesome backache. Definitive diagnosis can be established using ultrasonography-assisted biopsy, with immunohistochemical staining of the biopsy specimen.
Peritoneum and Mesentery
Metastasis to the peritoneum and mesentery is a common finding at autopsy. It usually occurs in patients who have disseminated abdominal metastases and is almost always associated with bowel or hepatic metastases (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a; Einhorn et al. 1974; Levitt et al. 1982). Symptoms, which may be vague and nonspecific, include abdominal pain, nausea, or vomiting (Stryga et al. 2009). Severe symptoms may occur as a result of massive malignant ascites or bowel obstruction caused by bulky metastases. The diagnosis of large lesions can be made by CT, PET, or ultrasonography or as an incidental finding during a laparotomy (Goldstein et al. 1977; Levitt et al. 1982; McDermott et al. 1996). In patients who have peritoneal and/or mesenteric metastases, survival is typically short, and quality of life is frequently hampered by the need for repeated paracentesis with associated discomfort and malnutrition.
The adrenal glands are a frequent site of metastasis and are involved in 36–54% of patients at autopsy (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a). Adrenal metastases are often bilateral. It is unusual for them to cause symptoms, and if symptoms do occur, they are generally vague. Adrenal insufficiency has only rarely been documented but may occur more often than is suspected clinically because the symptoms are often attributed to the patient’s generalized disease (Srinivasan et al. 2016). Peritoneal hemorrhage from metastases within the adrenal gland occurs rarely, but it can be particularly problematic because the location of the adrenal glands limits physiologic tamponade of bleeding (Lam and Lo 2002). Perhaps as a consequence, adrenal metastasis has been reported as a cause of death in some patients (Seidenwurm et al. 1984). Characteristic CT appearances of adrenal metastases from melanoma include bilateral adrenal masses greater than 5 cm in diameter, with central or irregular areas of necrosis/hemorrhage (Rajaratnam and Waugh 2005). Caution has been urged in the use of fine-needle aspiration for diagnosis of incidentally discovered adrenal masses, because the results are rarely informative and the procedure can be hazardous (Quayle et al. 2007). In a retrospective review of 154 patients with adrenal metastases, the median survival was 6.4 months and was negatively affected by the presence of synchronous extra-adrenal metastases or an elevated LDH level (Mittendorf et al. 2008). Twenty-two patients underwent surgery, with 20 rendered disease-free either by adrenalectomy alone or by adrenalectomy with concomitant metastasectomy. Patients who underwent surgery had an improved survival compared with those managed nonoperatively (p <0.0001). Others have reported that adrenalectomy can be performed laparoscopically with minimal complications (Castillo et al. 2007). Nonetheless, considering the general poor prognosis of patients with adrenal metastases, it is recommended that surgical treatment be considered only in selected patients, such as those with limited extra-adrenal metastatic disease who can be rendered disease-free (Mittendorf et al. 2008).
The thyroid gland is involved with melanoma metastases in 25–39% of patients at autopsy (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a). Both lobes may be involved in a diffuse manner. Patients may be seen with an asymptomatic mass in the neck, but this is rarely an isolated finding (Bozbora et al. 2005; Srivatsa and Rhee 2009). Metastases rarely cause thyroid dysfunction (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a; Einhorn et al. 1974) Parathyroid involvement is infrequent, even in patients at autopsy, in whom the incidence is only 2–4% (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a).
The pituitary gland can also be involved with metastatic melanoma, although this is an infrequent occurrence (less than 5%) even in autopsy series (McCutcheon et al. 2007). Few cases of diabetes insipidus from metastatic melanoma have been reported (Amer et al. 1978, 1979; Gupta and Brasfield 1964a; Ten Bokkel Huinink et al. 2000). Identification of pituitary involvement by imaging alone is difficult because of inconsistent signal characteristics on MRI and confusion related to associated hemorrhage (McCutcheon et al. 2007). Although transsphenoidal resection can be successfully performed, patient survival is usually limited by the frequent development of additional central nervous system and systemic metastases.
Melanoma is one of the most common metastatic neoplasms involving the breast parenchyma (Majeski 1999; Pressman 1973; Arora and Robinson 1992; Lee 2007). The prevalence ranges from 2% to 6% in different series (Einhorn et al. 1974; Arora and Robinson 1992). In a small series of patients who had breast metastases of melanoma, 14 of 15 patients were premenopausal women, and the median age was 38 years. Five patients had bilateral breast involvement, and all of the patients had other sites of metastases at the time of diagnosis (Arora and Robinson 1992). It is not uncommon for these lesions to be associated with cutaneous or subcutaneous metastases at other sites.
Metastases to the breast are occasionally diagnosed clinically by palpation. Although all patients have abnormal findings on mammograms, the radiographic appearance of these lesions may closely resemble benign breast disease (Paulus and Libshitz 1982). Because metastatic tumors can simulate primary breast carcinoma on physical examination, mammography, and frozen section biopsy, it is important to distinguish between these two entities before embarking on radical breast surgery (Barker and Girling 1989). Histologic features as well as immunohistochemical stains (S-100, HMB-45, Melan-A) can distinguish melanoma from primary breast malignancies. Fine-needle aspiration with cytopathologic and immunocytochemical analysis can frequently confirm the diagnosis as well (Fulciniti et al. 2007).
Ovaries, Uterus, and Placenta
Melanoma can occasionally metastasize to the organs of the female reproductive tract. Ovarian metastases are observed at autopsy in 7–16% of cases (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a; Einhorn et al. 1974). They may be unilateral or bilateral (Fitzgibbons et al. 1987). They are rarely detected clinically but may present as a palpable abdominal mass (Einhorn et al. 1974; Gonzalez and Hammond 1983; Ulbright et al. 1984; Sbitti et al. 2011). Ovarian metastases may also cause an acute abdomen caused by rupture of an ovarian lesion (Silveira et al. 1977; Habek et al. 2012).
Uterine metastases are seen in 4–11% of autopsy cases and are usually not symptomatic (Patel et al. 1978; Lee 1980; Gupta and Brasfield 1964a). Uterine metastases most commonly involve the cervix or myometrium, and endometrial metastases are rare. However, patients with melanoma metastases involving the uterus may occasionally have abnormal vaginal bleeding (Bauer et al. 1984; Casey and Shapiro 1974) and such bleeding may be the first manifestation of metastatic disease (Jaffrey et al. 1976; Fambrini et al. 2008). An endometrial curettage usually yields the diagnosis (Takeda et al. 1978). In addition, melanoma has occasionally been reported to metastasize to the maternal placenta (Freedman and McMahon 1960; Anderson et al. 1989; Altman et al. 2003). However, involvement of the fetal aspects of the placenta or the fetus itself has very rarely been reported. The prognosis for the mother is usually extremely poor (Altman et al. 2003).
Testes and Penis
Melanoma was the most common cause of testicular metastases in one series, in which it was observed in 9 of 22 patients (Johnson et al. 1971). Metastases to the testes and penis were detected in approximately 5–7% of patients at autopsy (Lee 1980; Gupta and Brasfield 1964a; Hanash et al. 1969). Testicular metastases rarely cause any significant clinical problems, and patients usually have testicular enlargement which can be detected on physical examination. Ultrasound examination, CT, or MRI may be helpful in delineating the boundaries of the lesion, and definitive diagnosis can be established by immunohistochemical staining of biopsy specimens. In one patient at autopsy, metastatic melanoma was found to involve the corpus cavernosum of the penis (Paquin and Roland 1956).
Oral Cavity, Pharynx, and Larynx
Melanoma is the neoplasm that most frequently metastasizes to the larynx and surrounding areas (Ferlito et al. 1988; Gutfreund et al. 1995). Henderson and colleagues (Henderson et al. 1986) reviewed 54 patients who had melanoma metastatic to the upper aerodigestive tract and found that the most common sites (in decreasing order) were the tonsil, tongue, nasopharynx, larynx, and lip. Metastases to the oropharynx have also been reported (Henderson et al. 1986). Two thirds of the patients had evidence of metastatic disease elsewhere.
Metastases to this area often cause symptoms that may be misinterpreted as an intercurrent illness. Pain, airway obstruction, and bleeding were the most common symptoms in one study (Myall et al. 1983). In patients with a toothache or tooth abscess, metastases to the mandible or maxilla have been documented (Myall et al. 1983; Pliskin et al. 1976; Samit et al. 1978; Welch et al. 1985; Meyer and Shklar 1965). Melanoma metastases to the tongue, tonsils, or pharynx can simulate an inflammatory process or a sore throat, and metastases to the larynx may be the cause of hoarseness (Chamberlain 1966; Ferlito et al. 1988; Ashur et al. 1979; Sood et al. 1999; Freeland et al. 1979; Wood and Donegan 1983; Zegarelli et al. 1973). Physicians or dentists evaluating melanoma patients with such symptoms should have a high index of suspicion for metastatic disease. Many of these lesions can easily be seen by direct or indirect laryngoscopy.
Eye and Orbit
Metastasis involving the eye is present in only 0.5% of autopsy cases (Lee 1980; Gupta and Brasfield 1964a; Zakka et al. 1980). Usually present within the globe, the metastases can be bilateral (Font et al. 1967). When melanoma occurs within the globe, it most commonly involves the choroid or uvea (Font et al. 1967; de Bustros et al. 1985; Eide and Syrdalen 1990; Albert et al. 2007). Less commonly, metastatic melanoma can involve the iris (Font et al. 1967; de Bustros et al. 1985; Hirst et al. 1979), the retina (Robertson et al. 1981), or rarely the vitreous (Robertson et al. 1981; Murray 1983; Les Cole et al. 1986; Bowman et al. 1994; Breazzano and Barker-Griffith 2016). Metastases may also be present simultaneously in the eye and the fatty tissue within the orbit, or they may be present in the extraocular tissues alone (Eide and Syrdalen 1990; Vida and Binder 1978).
Ocular metastasis occurs an average of 3 years after the diagnosis of the primary melanoma (Font et al. 1967; de Bustros et al. 1985), although it has occurred as late as 10 years after diagnosis (Hirst et al. 1979). Metastases to the eye can be the initial presentation of metastatic melanoma, but this is uncommon. The symptoms of ocular melanoma metastasis include blurred vision, ocular pain, and redness of the eye. Other symptoms include “floaters” and pain, which can mimic a therapy-resistant uveitis. There may also be signs or symptoms of secondary glaucoma caused by angle closure from choroidal detachment, direct angle involvement, or obstruction of the trabecular meshwork by tumor or inflammatory cells.
Signs of metastatic disease include pupil dilation, fixation field defects, or pigment deposits in the uvea. The metastatic lesion may be visible on fundoscopic or slit-lamp examination. The diagnosis is usually made by direct visual examination, but a fine-needle aspiration may occasionally be required (de Bustros et al. 1985; Robertson et al. 1981). Ophthalmic ultrasonography can be used to define a tumor mass in the globe or orbit. CT or MRI of the orbits and head may also demonstrate associated tumors in the orbit or other sites within the head. The survival of patients with ocular melanoma metastases is generally quite poor, with a median reported survival of only 72 days.
Melanosis is an unusual complication of advanced melanoma that is characterized by generalized, diffuse pigmentation (Böhm et al. 2001). The skin diffusely becomes dark gunmetal gray or slate colored, and melanuria is often observed (Böhm et al. 2001; Konrad and Wolff 1974; Silberberg et al. 1968; Lerner and Moellmann 1993). Although melanosis does occasionally occur in patients who do not have detectable metastatic disease, most affected patients have widespread metastatic disease, in particular extensive liver involvement, and high tumor cell turnover. Considering these two features, it is not surprising that this presentation has an extremely poor prognosis. The cause of the skin and urine discoloration has been debated, but it is most likely attributable to deposition in dermal macrophages or excretion in the urine of free melanin pigment, resulting from tumor cell pigment incontinence, which has overwhelmed the capacity of the reticuloendothelial system to retain it (Murray et al. 1999; Adrian et al. 1981; Spremulli et al. 1983; Steiner et al. 1991). Aberrant melanocyte growth factor production may contribute to this phenomenon (Böhm et al. 2001; Steiner et al. 1991).
Surveillance in Patients Who Have Localized Melanoma
Several studies have been published that describe follow-up of patients who have been treated for melanoma (Romero et al. 1994). Uniform staging, evaluation, treatment, and surveillance schemes for all stages of melanoma have been proposed by the National Comprehensive Cancer Network in the United States. These guidelines, in general, allow substantial flexibility in follow-up, particularly with regard to blood tests and radiologic examinations. This flexibility reflects the limited clinical studies, and therefore the data available, on the utility of blood tests and follow-up radiologic examinations. Clinicians should choose follow-up schedules and perform tests at intervals within these guidelines, with frequencies determined by the patient’s clinical circumstance, risk of recurrence, anxiety, and the physician’s experience. These issues are discussed in more detail (see also “Surveillance and Follow-up of Melanoma Patients”).
- Albert JG, Gimm O, Stock K, Bilkenroth U, Marsch WC, Helmbold P (2007) Small-bowel endoscopy is crucial for diagnosis of melanoma metastases to the small bowel: a case of metachronous small-bowel metastases and review of the literature. Melanoma Res 17(5):335–338PubMedCrossRefPubMedCentralGoogle Scholar
- Anwar H, Sachpekidis C, Winkler J, Kopp-Schneider A, Haberkorn U, Hassel JC, Dimitrakopoulou-Strauss A (2018) Absolute number of new lesions on 18 F-FDG PET/CT is more predictive of clinical response than SUV changes in metastatic melanoma patients receiving ipilimumab. Eur J Nucl Med Mol Imaging 45(3):376–383PubMedCrossRefPubMedCentralGoogle Scholar
- Ashur H, Mizrahi S, Ben-Hur N, Dolberg L (1979) Metastatic malignant melanoma to the palatine tonsils: report of case. J Oral Surg (American Dental Association: 1965) 37(2):110–112Google Scholar
- Balch CM, Buzaid AC, Atkins MB, Cascinelli N, Coit DG, Fleming ID, Houghton A, Kirkwood JM, Mihm MF, Morton DL, Reintgen D, Ross MI, Sober A, Soong SJ, Thompson JA, Thompson JF, Gershenwald JE, McMasters KM (2000) A new American Joint Committee on Cancer staging system for cutaneous melanoma. Cancer 88(6):1484–1491PubMedCrossRefPubMedCentralGoogle Scholar
- Balch CM, Buzaid AC, Soong SJ, Atkins MB, Cascinelli N, Coit DG, Fleming ID, Gershenwald JE, Houghton A, Kirkwood JM, McMasters KM, Mihm MF, Morton DL, Reintgen DS, Ross MI, Sober A, Thompson JA, Thompson JF (2001a) Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol 19(16):3635–3648PubMedCrossRefPubMedCentralGoogle Scholar
- Balch CM, Soong S-J, Gershenwald JE, Thompson JF, Reintgen DS, Cascinelli N, Urist M, McMasters KM, Ross MI, Kirkwood JM (2001b) Prognostic factors analysis of 17,600 melanoma patients: validation of the American Joint Committee on Cancer melanoma staging system. J Clin Oncol 19(16):3622–3634PubMedCrossRefGoogle Scholar
- Bognel C, Lasser P, Zimmermann P (1992) Gastric metastases. Apropos of 17 cases. Ann Chir 5:436–441Google Scholar
- Bostick PJ, Chatterjee S, Chi DD, Huynh KT, Giuliano AE, Cote R, Hoon D (1998) Limitations of specific reverse-transcriptase polymerase chain reaction markers in the detection of metastases in the lymph nodes and blood of breast cancer patients. J Clin Oncol 16(8):2632–2640PubMedCrossRefPubMedCentralGoogle Scholar
- Cho SY, Lipson EJ, Im H-J, Rowe SP, Gonzalez EM, Blackford A, Chirindel A, Pardoll DM, Topalian SL, Wahl RL (2017) Prediction of response to immune checkpoint inhibitor therapy using early-time-point 18F-FDG PET/CT imaging in patients with advanced melanoma. J Nucl Med 58(9):1421PubMedPubMedCentralCrossRefGoogle Scholar
- Deichmann M, Benner A, Bock M, Jäckel A, Uhl K, Waldmann V, Näher H (1999) S100-Beta, melanoma-inhibiting activity, and lactate dehydrogenase discriminate progressive from nonprogressive American Joint Committee on Cancer stage IV melanoma. J Clin Oncol 17(6):1891–1896PubMedCrossRefPubMedCentralGoogle Scholar
- Dietlein M, Krug B, Groth W, Smolarz K, Scheidhauer K, Psaras T, Stützer H, Lackner K, Schicha H (1999) Positron emission tomography using 18F-fluorodeoxyglucose in advanced stages of malignant melanoma: a comparison of ultrasonographic and radiological methods of diagnosis. Nucl Med Commun 20(3):255–261PubMedCrossRefPubMedCentralGoogle Scholar
- Eigentler TK, Figl A, Krex D, Mohr P, Mauch C, Rass K, Bostroem A, Heese O, Koelbl O, Garbe C (2011) Number of metastases, serum lactate dehydrogenase level, and type of treatment are prognostic factors in patients with brain metastases of malignant melanoma. Cancer 117(8):1697–1703PubMedCrossRefPubMedCentralGoogle Scholar
- Eton O, Legha SS, Moon TE, Buzaid AC, Papadopoulos NE, Plager C, Burgess AM, Bedikian AY, Ring S, Dong Q, Glassman AB, Balch CM, Benjamin RS (1998) Prognostic factors for survival of patients treated systemically for disseminated melanoma. J Clin Oncol 16(3):1103–1111PubMedCrossRefPubMedCentralGoogle Scholar
- Ewend MG, Carey LA, Brem H (1996) Treatment of melanoma metastases in the brain. In: Seminars in surgical oncology, vol 6. Wiley Online Library, New York, pp 429–435Google Scholar
- Forschner A, Olthof S-C, Gückel B, Martus P, Vach W, la Fougère C, Nikolaou K, Keim U, Eigentler TK, Garbe C (2017) Impact of 18F-FDG-PET/CT on surgical management in patients with advanced melanoma: an outcome based analysis. Eur J Nucl Med Mol Imaging 44(8):1312–1318PubMedCrossRefPubMedCentralGoogle Scholar
- Franzke A, Probst-Kepper M, Buer J, Duensing S, Hoffmann R, Wittke F, Volkenandt M, Ganser A, Atzpodien J (1998) Elevated pretreatment serum levels of soluble vascular cell adhesion molecule 1 and lactate dehydrogenase as predictors of survival in cutaneous metastatic malignant melanoma. Br J Cancer 78(1):40–45PubMedPubMedCentralCrossRefGoogle Scholar
- Friedman KP, Wahl RL (2004) Clinical use of positron emission tomography in the management of cutaneous melanoma. In: Seminars in nuclear medicine, vol 4. Elsevier, New York, pp 242–253Google Scholar
- Garbe C, Leiter U, Ellwanger U, Blaheta HJ, Meier F, Rassner G, Schittek B (2003) Diagnostic value and prognostic significance of protein S-100β, melanoma-inhibitory activity, and tyrosinase/MART-1 reverse transcription-polymerase chain reaction in the follow-up of high-risk melanoma patients. Cancer 97(7):1737–1745PubMedCrossRefGoogle Scholar
- Garbe C, Hauschild A, Volkenandt M, Schadendorf D, Stolz W, Reinhold U, Kortmann R-D, Kettelhack C, Frerich B, Keilholz U, Dummer R, Sebastian G, Tilgen W, Schuler G, Mackensen A, Kaufmann R (2007) Evidence and interdisciplinary consense-based German guidelines: diagnosis and surveillance of melanoma. Melanoma Res 17(6):393–399PubMedCrossRefGoogle Scholar
- Gaudy-Marqueste C, Dussouil A, Carron R, Troin L, Malissen N, Loundou A, Monestier S, Mallet S, Richard M, Régis J (2017) Survival of melanoma patients treated with targeted therapy and immunotherapy after systematic upfront control of brain metastases by radiosurgery. Eur J Cancer 84:44–54PubMedCrossRefGoogle Scholar
- Gelberman RH, Stewart WR, Harrelson JM (1978) Hand metastasis from melanoma: a case study. Clin Orthop Relat Res 136:264–266Google Scholar
- Gershenwald JE, Balch CM, Soong SJ, Thompson JF (2003) Prognostic factors and natural history. In: Balch CM, Houghton AN, Sober SJ, Soong SJ (eds) Cutaneous melanoma, 4th edn. Quality Medical Publishing, St LouisGoogle Scholar
- Gershenwald JE, Scolyer RA, Hess KR, Sondak VK, Long GV, Ross MI, Lazar AJ, Faries MB, Kirkwood JM, McArthur GA, Haydu LE, Eggermont AMM, Flaherty KT, Balch CM, Thompson JF, for members of the American Joint Committee on Cancer Melanoma Expert Panel, the International Melanoma Database, Discovery Platform (2017) Melanoma staging: evidence-based changes in the American Joint Committee on Cancer eighth edition cancer staging manual. CA Cancer J Clin 67(6):472–492. https://doi.org/10.3322/caac.21409CrossRefPubMedPubMedCentralGoogle Scholar
- Gläser R, Rass K, Seiter S, Hauschild A, Christophers E, Tilgen W (1997) Detection of circulating melanoma cells by specific amplification of tyrosinase complementary DNA is not a reliable tumor marker in melanoma patients: a clinical two-center study. J Clin Oncol 15(8):2818–2825PubMedCrossRefGoogle Scholar
- Habek D, Marton I, Bauman R, Prka M, Bobić MV (2012) Ruptured ovarian metastatic malignant melanoma caused acute abdomen. Central Eur J Med 7(2):241–244Google Scholar
- Hamid O, Robert C, Daud A, Hodi FS, Hwu WJ, Kefford R, Wolchok JD, Hersey P, Joseph RW, Weber JS, Dronca RS, Mitchell TC, Patnaik A, Zarour HM, Joshua AM, Jensen E, Ibrahim N, Ahsan S, Ribas A (2018) 5-year survival outcomes in patients (pts) with advanced melanoma treated with pembrolizumab (pembro) in KEYNOTE-001. J Clin Oncol 36(suppl):abstr 9516. ASCO annual meeting, 2018CrossRefGoogle Scholar
- Ihle MA, Fassunke J, König K, Grünewald I, Schlaak M, Kreuzberg N, Tietze L, Schildhaus H-U, Büttner R, Merkelbach-Bruse S (2014) Comparison of high resolution melting analysis, pyrosequencing, next generation sequencing and immunohistochemistry to conventional Sanger sequencing for the detection of p. V600E and non-p. V600E BRAF mutations. BMC Cancer 14(1):13PubMedPubMedCentralCrossRefGoogle Scholar
- Jaffrey I, Lefkowitz L, Lestch S, Weiss D (1976) Metastatic melanoma presenting as abnormal uterine bleeding: A case report. Mount Sinai J Med, New York 43(2):180Google Scholar
- Joseph RW, Elassaiss-Schaap J, Kefford RF, Hwu WJ, Wolchok JD, Joshua AM, Ribas A, Hodi FS, Hamid O, Robert C, Daud AI, Dronca RS, Hersey P, Weber JS, Patnaik A, de Alwis DP, Perrone AM, Zhang J, Kang SP, Ebbinghaus SW, Anderson KM, Gangadhar T (2018) Baseline tumor size is an independent prognostic factor for overall survival in patients with melanoma treated with pembrolizumab. Clin Cancer Res. https://doi.org/10.1158/1078-0432.CCR-17-2386
- Kamaya A, Weinstein S, Desser TS (2006) Multiple lesions of the spleen: differential diagnosis of cystic and solid lesions. In: Seminars in ultrasound, CT and MRI, vol 27, No. 5, WB Saunders, pp 389–403Google Scholar
- Keilholz U, Martus P, Punt CJA, Kruit W, Mooser G, Schadendorf D, Liénard D, Dummer R, Koller J, Voit C, Eggermont AMM (2002) Prognostic factors for survival and factors associated with long-term remission in patients with advanced melanoma receiving cytokine-based treatments: second analysis of a randomised EORTC Melanoma Group trial comparing interferon-alpha2a (IFNalpha) and interleukin 2 (IL-2) with or without cisplatin. Eur J Cancer (Oxford, England: 1990) 38(11):1501–1511CrossRefGoogle Scholar
- Kelderman S, Heemskerk B, van Tinteren H, van den Brom RR, Hospers GA, van den Eertwegh AJ, Kapiteijn EW, de Groot JWB, Soetekouw P, Jansen RL (2014) Lactate dehydrogenase as a selection criterion for ipilimumab treatment in metastatic melanoma. Cancer Immunol Immunother 63(5):449–458PubMedPubMedCentralGoogle Scholar
- Lee RJ, Gremel G, Marshall A, Myers KA, Fisher N, Dunn JA, Dhomen N, Corrie PG, Middleton MR, Lorigan P, Marais R (2018) Circulating tumor DNA predicts survival in patients with resected high-risk stage II/III melanoma. Ann Oncol 29(2):490–496. https://doi.org/10.1093/annonc/mdx717CrossRefPubMedPubMedCentralGoogle Scholar
- Lerner A, Moellmann G (1993) Two rare manifestations of melanomas: generalized cutaneous melanosis and rapid solar induction of showers of small pigmented lesions. A critical review of the literature and presentation of two additional cases. Acta Derm Venereol 73(4):241–250PubMedPubMedCentralGoogle Scholar
- Lesur G, Bergemer A, Metges J, Barre O, Dupuy P, Le JP (1992) The metastatic liver: a rare cause of hepatocellular insufficiency. Ann Gastroenterol Hepatol 5:217–220Google Scholar
- Licciardello M, Giura M, Rozzo G, Marra E, Fava P, Astrua C, Contu V, Lista P, Fierro M, Quaglino P (2017) Arrhythmias in a patient with metastatic melanoma treated with targeted therapy and implantable cardioverter defibrillator. Br J Dermatol 177(2):584–587PubMedCrossRefPubMedCentralGoogle Scholar
- Long GV, Wilmott JS, Capper D, Preusser M, Zhang YE, Thompson JF, Kefford RF, von Deimling A, Scolyer RA (2013) Immunohistochemistry is highly sensitive and specific for the detection of V600E BRAF mutation in melanoma. Am J Surg Pathol 37(1):61–65. https://doi.org/10.1097/PAS.0b013e31826485c0CrossRefPubMedPubMedCentralGoogle Scholar
- Long GV, Grob JJ, Nathan P, Ribas A, Robert C, Schadendorf D, Lane SR, Mak C, Legenne P, Flaherty KT, Davies MA (2016) Factors predictive of response, disease progression, and overall survival after dabrafenib and trametinib combination treatment: a pooled analysis of individual patient data from randomised trials. Lancet Oncol 17(12):1743–1754. https://doi.org/10.1016/S1470-2045(16)30578-2CrossRefPubMedPubMedCentralGoogle Scholar
- Long GV, Eroglu Z, Infante J, Patel S, Daud A, Johnson DB, Gonzalez R, Kefford R, Hamid O, Schuchter L, Cebon J, Sharfman W, McWilliams R, Sznol M, Redhu S, Gasal E, Mookerjee B, Weber J, Flaherty KT (2018) Long-term outcomes in patients with BRAF V600-mutant metastatic melanoma who received dabrafenib combined with trametinib. J Clin Oncol 36(7):667–673. https://doi.org/10.1200/JCO.2017.74.1025CrossRefPubMedPubMedCentralGoogle Scholar
- Lopez-Rios F, Angulo B, Gomez B, Mair D, Martinez R, Conde E, Shieh F, Vaks J, Langland R, Lawrence HJ (2013) Comparison of testing methods for the detection of BRAF V600E mutations in malignant melanoma: pre-approval validation study of the companion diagnostic test for vemurafenib. PLoS One 8(1):e53733PubMedPubMedCentralCrossRefGoogle Scholar
- McArthur GA, Puzanov I, Amaravadi R, Ribas A, Chapman P, Kim KB, Sosman JA, Lee RJ, Nolop K, Flaherty KT (2012) Marked, homogeneous, and early [18F] fluorodeoxyglucose-positron emission tomography responses to vemurafenib in BRAF-mutant advanced melanoma. J Clin Oncol 30(14):1628–1634PubMedPubMedCentralCrossRefGoogle Scholar
- Mehta M, Omer E (2017) Isolated metastatic melanoma to the pancreas diagnosed by endoscopic ultrasonography: a case report. J Pancreas 18(2):144–146Google Scholar
- Mellado B, Colomer D, Castel T, Munoz M, Carballo E, Galan M, Mascaró JM, Vives-Corrons JL, Grau JJ, Estapé J (1996) Detection of circulating neoplastic cells by reverse-transcriptase polymerase chain reaction in malignant melanoma: association with clinical stage and prognosis. J Clin Oncol 14(7):2091–2097PubMedCrossRefPubMedCentralGoogle Scholar
- Mellado B, Gutierrez L, Castel T, Colomer D, Fontanillas M, Castro J, Estapé J (1999) Prognostic significance of the detection of circulating malignant cells by reverse transcriptase-polymerase chain reaction in long-term clinically disease-free melanoma patients. Clin Cancer Res 5(7):1843–1848PubMedGoogle Scholar
- Mian S, Ugurel S, Parkinson E, Schlenzka I, Dryden I, Lancashire L, Ball G, Creaser C, Rees R, Schadendorf D (2005) Serum proteomic fingerprinting discriminates between clinical stages and predicts disease progression in melanoma patients. J Clin Oncol 23(22):5088–5093PubMedCrossRefPubMedCentralGoogle Scholar
- Michie SA, Spagnolo DV, Dunn KA, Warnke RA, Rouse RV (1987) A panel approach to the evaluation of the sensitivity and specificity of antibodies for the diagnosis of routinely processed histologically undifferentiated human neoplasms. Am J Clin Pathol 88(4):457–462PubMedCrossRefPubMedCentralGoogle Scholar
- Munoz J, Vinolas N, Castro J, Graus F, Conde C, Castel T, Biete A, Tolosa E, Estapé J (1993) Cerebral metastasis of melanoma: study of 48 patients. Med Clin 101(18):684–687Google Scholar
- Petersen RP, Hanish SI, Haney JC, Miller CC, Burfeind WR, Tyler DS, Seigler HF, Wolfe W, D’Amico TA, Harpole DH (2007) Improved survival with pulmonary metastasectomy: an analysis of 1720 patients with pulmonary metastatic melanoma. J Thorac Cardiovasc Surg 133(1):104–110.e102PubMedCrossRefGoogle Scholar
- Pfannenberg C, Aschoff P, Schanz S, Eschmann SM, Plathow C, Eigentler TK, Garbe C, Brechtel K, Vonthein R, Bares R (2007) Prospective comparison of 18F-fluorodeoxyglucose positron emission tomography/computed tomography and whole-body magnetic resonance imaging in staging of advanced malignant melanoma. Eur J Cancer 43(3):557–564PubMedCrossRefGoogle Scholar
- Rajaratnam A, Waugh J (2005) Adrenal metastases of malignant melanoma: characteristic computed tomography appearances. J Med Imaging Radiat Oncol 49(4):325–329Google Scholar
- Rappeport E, Loft A, Berthelsen A, Von Der Recke P, Noergaard Larsen P, Mellon Mogensen A, Wettergren A, Rasmussen A, Hillingsoe J, Kirkegaard P (2007) Contrast-enhanced FDG-PET/CT vs. SPIO-enhanced MRI vs. FDG-PET vs. CT in patients with liver metastases from colorectal cancer: a prospective study with intraoperative confirmation. Acta Radiol 48(4):369–378PubMedCrossRefGoogle Scholar
- Reinhardt MJ, Joe AY, Jaeger U, Huber A, Matthies A, Bucerius J, Roedel R, Strunk H, Bieber T, Biersack H-J (2006) Diagnostic performance of whole body dual modality 18F-FDG PET/CT imaging for N-and M-staging of malignant melanoma: experience with 250 consecutive patients. J Clin Oncol 24(7):1178–1187PubMedCrossRefGoogle Scholar
- Robert C, Ribas A, Hamid O, Daud A, Wolchok JD, Joshua AM, Hwu WJ, Weber JS, Gangadhar TC, Joseph RW, Dronca R, Patnaik A, Zarour H, Kefford R, Hersey P, Zhang J, Anderson J, Diede SJ, Ebbinghaus S, Hodi FS (2018) Durable complete response after discontinuation of pembrolizumab in patients with metastatic melanoma. J Clin Oncol 36(17):1668–1674. https://doi.org/10.1200/JCO.2017.75.6270CrossRefPubMedGoogle Scholar
- Sachpekidis C, Larribere L, Pan L, Haberkorn U, Dimitrakopoulou-Strauss A, Hassel JC (2015) Predictive value of early 18 F-FDG PET/CT studies for treatment response evaluation to ipilimumab in metastatic melanoma: preliminary results of an ongoing study. Eur J Nucl Med Mol Imaging 42(3):386–396PubMedCrossRefGoogle Scholar
- Samit A, Falk H, Ohanian M, Leban S, Mashberg A (1978) Malignant melanoma metastatic to the mandible. J Oral Surg (American Dental Association: 1965) 36(10):816–821Google Scholar
- Santiago-Walker A, Gagnon R, Mazumdar J, Casey M, Long GV, Schadendorf D, Flaherty K, Kefford R, Hauschild A, Hwu P, Haney P, O'Hagan A, Carver J, Goodman V, Legos J, Martin AM (2016) Correlation of BRAF mutation status in circulating-free DNA and tumor and association with clinical outcome across four BRAFi and MEKi clinical trials. Clin Cancer Res 22(3):567–574. https://doi.org/10.1158/1078-0432.CCR-15-0321CrossRefPubMedGoogle Scholar
- Schachter J, Ribas A, Long GV, Arance A, Grob JJ, Mortier L, Daud A, Carlino MS, McNeil C, Lotem M, Larkin J, Lorigan P, Neyns B, Blank C, Petrella TM, Hamid O, Zhou H, Ebbinghaus S, Ibrahim N, Robert C (2017) Pembrolizumab versus ipilimumab for advanced melanoma: final overall survival results of a multicentre, randomised, open-label phase 3 study (KEYNOTE-006). Lancet. https://doi.org/10.1016/S0140-6736(17)31601-XCrossRefGoogle Scholar
- Schadendorf D, Long GV, Stroiakovski D, Karaszewska B, Hauschild A, Levchenko E, Chiarion-Sileni V, Schachter J, Garbe C, Dutriaux C, Gogas H, Mandala M, Haanen J, Lebbe C, Mackiewicz A, Rutkowski P, Grob JJ, Nathan P, Ribas A, Davies MA, Zhang Y, Kaper M, Mookerjee B, Legos JJ, Flaherty KT, Robert C (2017) Three-year pooled analysis of factors associated with clinical outcomes across dabrafenib and trametinib combination therapy phase 3 randomised trials. Eur J Cancer 82:45–55. https://doi.org/10.1016/j.ejca.2017.05.033CrossRefPubMedGoogle Scholar
- Schwimmer J, Essner R, Patel A, Jahan S (2000) A review of the literature for whole-body FDG PET in the management of patients with melanoma. Q J Nucl Med Mol Imaging 44(2):153Google Scholar
- Sen CA, Kargi A, Kaya V, Tanriverdi O (2013) Isolated and solitary splenic metastasis detected by positron emission tomography in a patient with malignant melanoma: case report and review of the literature. Contemp Oncol 17(2):214Google Scholar
- Shakhova O, Zingg D, Schaefer SM, Hari L, Civenni G, Blunschi J, Claudinot S, Okoniewski M, Beermann F, Mihic-Probst D, Moch H, Wegner M, Dummer R, Barrandon Y, Cinelli P, Sommer L (2012) Sox10 promotes the formation and maintenance of giant congenital naevi and melanoma. Nat Cell Biol 14(8):882–890. https://doi.org/10.1038/ncb2535CrossRefPubMedPubMedCentralGoogle Scholar
- Srivatsa S, Rhee MK (2009) Metastatic melanoma presenting as a thyroid nodule in a 38-year-old woman. Hosp Physician 45:39–41Google Scholar
- Strobel K, Dummer R, Steinert HC, Conzett KB, Schad K, Lago MP, Soyka JD, Veit-Haibach P, Seifert B, Kalff V (2008) Chemotherapy response assessment in stage IV melanoma patients – comparison of 18 F-FDG-PET/CT, CT, brain MRI, and tumormarker S-100B. Eur J Nucl Med Mol Imaging 35(10):1786PubMedCrossRefPubMedCentralGoogle Scholar
- Stryga M, Janas M, Szopiński P (2009) Melanoma metastases to the small intestine mesentery, greater omentum and parietal peritoneum-case report. Pol J Surg 81(4):203–207Google Scholar
- Tessier DJ, McConnell EJ, Young-Fadok T, Wolff BG (2003) Melanoma metastatic to the colon: case series and review of the literature with outcome analysis. Dis Colon Rectum 46(4):441–447. https://doi.org/10.1097/01.dcr.0000059657.64526.b6CrossRefPubMedPubMedCentralGoogle Scholar
- Wolchok JD, Hoos A, O’Day S, Weber JS, Hamid O, Lebbé C, Maio M, Binder M, Bohnsack O, Nichol G (2009) Guidelines for the evaluation of immune therapy activity in solid tumors: immune-related response criteria. Clin Cancer Res. https://doi.org/10.1158/1078-0432.CCR-09-1624PubMedCrossRefGoogle Scholar
- Wolchok JD, Chiarion-Sileni V, Gonzalez R, Rutkowski P, Grob JJ, Cowey CL, Lao CD, Wagstaff J, Schadendorf D, Ferrucci PF, Smylie M, Dummer R, Hill A, Hogg D, Haanen J, Carlino MS, Bechter O, Maio M, Marquez-Rodas I, Guidoboni M, McArthur G, Lebbe C, Ascierto PA, Long GV, Cebon J, Sosman J, Postow MA, Callahan MK, Walker D, Rollin L, Bhore R, Hodi FS, Larkin J (2017) Overall survival with combined nivolumab and ipilimumab in advanced melanoma. N Engl J Med. https://doi.org/10.1056/NEJMoa1709684PubMedCrossRefGoogle Scholar
- Zager JS, Gastman BR, Leachman S, Gonzalez RC, Fleming MD, Ferris LK, Ho J, Miller AR, Cook RW, Covington KR, Meldi-Plasseraud K, Middlebrook B, Kaminester LH, Greisinger A, Estrada SI, Pariser DM, Cranmer LD, Messina JL, Vetto JT, Wayne JD, Delman KA, Lawson DH, Gerami P (2018) Performance of a prognostic 31-gene expression profile in an independent cohort of 523 cutaneous melanoma patients. BMC Cancer 18(1):130. https://doi.org/10.1186/s12885-018-4016-3CrossRefPubMedPubMedCentralGoogle Scholar
- Zakrzewski J, Geraghty LN, Rose AE, Christos PJ, Mazumdar M, Polsky D, Shapiro R, Berman R, Darvishian F, Hernando E (2011) Clinical variables and primary tumor characteristics predictive of the development of melanoma brain metastases and post-brain metastases survival. Cancer 117(8):1711–1720PubMedCrossRefGoogle Scholar