Comments on and Illustrations of the EFSUMB CEUS Guidelines: Transabdominal and Endoscopic Ultrasound Features of Intrapancreatic Metastases and the Role of Multiparametric Imaging and EUS-Guided Sampling in Rare Pancreatic Tumors

Simple Summary Intrapancreatic metastases are rare. They could occur at the same time or years later after the initial diagnosis of the tumor. Sonography and endosonography with contrast enhanced techniques provide very good additional information for differential diagnosis from other tumors. The most common solitary metastasis is that of renal cell carcinoma, with good prognosis after surgical resection. The further procedure, and in other tumor entities, depends on the primary tumor. This requires confirmation by endosonographic guided sampling, with the collection of material for immunohistological examination. Abstract A definite pathologic diagnosis of intrapancreatic metastasis is crucial for the management decision, i.e., curative or palliative surgery versus chemotherapy or conservative/palliative therapy. This review focuses on the appearance of intrapancreatic metastases on native and contrast-enhanced transabdominal ultrasound and endoscopic ultrasound. Differences and similarities in relation to the primary tumor, and the differential diagnosis from pancreatic carcinoma and neuroendocrine neoplasms are described. The frequency of intrapancreatic metastases in autopsy studies and surgical resection studies will be discussed. Further emphasis is placed on endoscopic ultrasound-guided sampling to confirm the diagnosis.


Introduction
The World Federation for Ultrasound in Medicine and Biology (WFUMB) has published guidelines on the use of contrast-enhanced ultrasound (CEUS) for the evaluation of focal liver lesions [1][2][3], and the European Federation of Societies for Ultrasound in Medicine (EFSUMB) has published guidelines for the evaluation of non-hepatic indications [4,5]. The improved detection and characterization of common focal pancreatic lesions (FPL) such as pancreatic ductal adenocarcinoma (PDAC) and pancreatic neuroendocrine neoplasms (PanNENs) are the main topics of these guidelines. AFSUMB guidelines for the performance of contrast-enhanced harmonic endoscopic ultrasound (CH-EUS) in the evaluation of pancreatic and other lesions have been published [6]. CH-EUS is recommended for the characterization of pancreatic solid masses [4][5][6]. In recent years, conventional ultrasound (US) and CEUS features of less common FPL have also been described in detail, including autoimmune pancreatitis [7][8][9][10], pancreatic tuberculosis [11,12], pancreatic ascariasis [13,14], and pancreatic hydatid cysts [15,16]. Nearly 90% of pancreatic neoplasms in adults are represented by invasive PDAC and related subtypes, according to the WHO classification from 2019. Cystic and intraductal neoplasms account for 4-5%, PanNENs for 3-4%, and acinar cell carcinomas and other rare entities account for the remaining 2-3% [17,18]. On the other hand, intrapancreatic metastases are rarely diagnosed. If the patient has a prior history of a malignant tumor, the possibility of intrapancreatic metastasis should always be considered when a solid pancreatic lesion is discovered. This may occur years after the diagnosis of a primary tumor even when the malignancy has been resected or treated medically with no recent current evidence of disease elsewhere. Pancreatic metastases must be differentiated from PDAC, as well as from PanNENs, other solid pancreatic lesions such as focal autoimmune pancreatitis, solid pseudopapillary neoplasms, and other rare pancreatic tumors. In the majority of cases, it is necessary to consider whether differentiating these entities would change the treatment approach. This article addresses the appearance of intrapancreatic metastases on transabdominal ultrasound (US), endoscopic ultrasound (EUS), and contrast-enhanced techniques such as CEUS and CH-EUS.

Disease Frequency
Intrapancreatic metastases should always be considered as a differential diagnosis vs. the more common PDAC because their prevalence is higher than previously expected, and their discovery may require a totally different treatment approach compared to PDAC.
In a multicenter study of small pancreatic lesions up to 15 mm, the prevalence of PDAC was only about 40%, whereas alternative lesions dominated. Pancreatic metastases were diagnosed using EUS-guided sampling in 7% of all solid FPL in this cohort [24].
Intrapancreatic metastases typically occur either as a manifestation of extensive metastatic tumor disease or as an isolated pancreatic location.
In autopsy studies, intrapancreatic metastases were diagnosed in up to 6% of patients overall, and in 15% of tumor patients [25][26][27]. The frequency of pancreatic metastases depends on the location of the primary tumor (Table 1). In an autopsy study of 154 patients who died of exocrine pancreatic carcinoma, 12.4% had a history of or a concomitant other tumor disease [28]. In the case of an imaging diagnosis of a pancreatic tumor, the most important differential diagnosis in patients with previous extra pancreatic tumors is whether the tumor is a pancreatic metastasis or a primary pancreatic tumor (PDAC or PanNENs). Interestingly, the diagnosis of pancreatic metastases with EUS-guided sampling was the first manifestation of a malignant disease in 12.5% [21], 16% [29], 18.8% [30], and 50% of cases [31] in different studies. Lung carcinomas and gastric carcinomas were the most common primary tumors of intrapancreatic metastases in autopsy studies [27,32] (Table 2).
Intrapancreatic metastases account for only 1.6% of pancreatic resections performed for malignancy [33], with renal cell carcinoma (RCC) being the most common primary tumor [34][35][36][37][38][39][40][41][42][43] (Table 3). The distribution in studies of EUS-guided sampling is similar (Table 4). Surgical studies predominantly included patients with solitary intrapancreatic metastases and no other organ overt localization. The most common cause of isolated pancreatic metastases is RCC [35,44], but, in 36.4% of cases, RCC metastases develop in multiple pancreatic locations [42]. The most common intrapancreatic RCC metastases are those of the clear cell subtype [45]. The striking differences between the relative frequency of RCC metastases between surgical and autopsy cohorts may be explained by their rather good prognosis and often solitary occurrence, which makes these patients more likely to qualify for pancreatic resection than those with pancreatic metastases from other primary tumors [22,39,43,46,47].

Clinical Presentation
Approximately 50% of intrapancreatic metastases are not associated with symptoms and are detected by imaging during follow-up or in the context of imaging for the clarification of other complaints and findings [39,48,56]. Symptoms include abdominal pain and non-specific complaints. If the metastasis is localized to the head of the pancreas, jaundice due to biliary compression may occur [19,39,56]. A mild increase in CA 19-9 (>40 U/mL) was observed in 46% of patients with intrapancreatic metastases, whereas a moderate increase (>100 U/mL) was observed in 28.6% of patients [19]. Compared to PDAC and PanNENs, intrapancreatic metastases were less often associated with diabetes mellitus [19]. Elucidating a prior history of a malignant tumor is, of course, of the utmost importance in order to suspect the presence of intrapancreatic metastases.

Pathology
In autopsy studies, 55% of metastases were solitary, 25% multiple, and 20% diffuse. Thirty-three percent of metastases were not visible macroscopically [27]. This must be considered in the case of surgical resection. Therefore, it is recommended that intraoperative sonography should be performed when partial pancreatic resections are planned [48]. Pancreatectomy may be preferred over local resection procedures for tumors with better prognosis such as RCC [48,57].

US and EUS
Intrapancreatic metastases usually are hypoechoic, homogeneous or heterogeneous, and often well defined [22,58,59]. El Hajj et al. described hypoechoic lesions in 80%, mixed hypoechoic/anechoic in 14%, hyperechoic in 4%, and anechoic in 2% [60]. In total, 80% lesions were solid, 18% were solid and cystic (RCC, melanoma, small bowel neuroendocrine tumor [carcinoid], and gastric squamous cell carcinoma), and 2% were cystic (melanoma) [60]. These data agree with those of DeWitt et al. These authors reported well-defined borders in 46% of cases. They concluded that in the presence of a tumor history, a smooth-bordered hypoechoic lesion should be suspicious for the presence of metastasis [51]. Fusaroli et al. described all metastases in their series as hypoechoic and predominantly homogeneous. The borders were regular except for breast cancer metastases [61].
Lesion diameter may vary, with or without pancreatic duct involvement [63]. Of 23 patients with various intrapancreatic metastases, only 20% had pancreatic ductal dilation [64]. The lesions may be single or multiple [19,59,65]. Pancreatic metastases tend not to infiltrate into adjacent vessels [19]. Nevertheless, there may be exceptions ( Figure 1). RCC metastases are usually hypervascular. This can already be detected on color and power Doppler Imaging [66]. ally, cystic areas may also be present [63,64].
Lesion diameter may vary, with or without pancreatic duct involvement [63]. Of 23 patients with various intrapancreatic metastases, only 20% had pancreatic ductal dilation [64]. The lesions may be single or multiple [19,59,65]. Pancreatic metastases tend not to infiltrate into adjacent vessels [19]. Nevertheless, there may be exceptions ( Figure 1). RCC metastases are usually hypervascular. This can already be detected on color and power Doppler Imaging [66]. In strain elastography on EUS, the lesion is stiffer (e). EUS-FNA obtaining histologic material confirmed metastasis from a colon cancer.
Yuan et al. described melanoma metastases which were hypoechoic with clear borders. Surprisingly, the entire pancreas had a large volume, and the parenchyma was inhomogeneous with an uneven shape [63]. This may also be a ributed to morphologic changes in the parenchyma accompanying the metastases. At the same time, diffuse (macroscopically invisible) infiltration may also be present [27]. In strain elastography on EUS, the lesion is stiffer (e). EUS-FNA obtaining histologic material confirmed metastasis from a colon cancer. Surprisingly, the entire pancreas had a large volume, and the parenchyma was inhomogeneous with an uneven shape [63]. This may also be attributed to morphologic changes in the parenchyma accompanying the metastases. At the same time, diffuse (macroscopically invisible) infiltration may also be present [27]. Chou et al. described a pancreas with diffuse metastatic infiltration, seen as hypoechoic enlargement with hypervascularity in Doppler studies. This would be atypical for ductal adenocarcinoma and would suggest autoimmune pancreatitis as a differential diagnosis [66].
In an elastography study of small pancreatic lesions up to 15 mm, metastases were stiffer compared to surrounding pancreatic parenchyma in 59% of cases. Surprisingly, there was a soft elastography image in 41%. As a result, soft tissue findings on elastography imaging do not exclude metastasis. While a hypoechoic solid lesion that is soft or isoelastic on strain elastography would be compatible with PanNENs in addition to pancreatic metastasis, this finding almost certainly excludes the diagnosis of PDAC [67].
The typical features of intrapancreatic metastases compared with PDAC and PanNENs on B-mode US, Duplex US, Power Doppler US, and elastography are summarized in Table 5.
Intrapancreatic metastases can be hyperenhanced as well as isoenhanced and hypoenhanced in the arterial phase. RCC metastases are usually hyperenhanced [61,83] (Figures 2 and 3). The partially overlapping enhancement pa erns raise different ties especially with regard to discrimination of pancreatic metas When a lesion is isoenhanced, it must be differentiated from inflam the first place. Hypoenhanced lesions must be differentiated from mor-PDAC. Multiple lesions usually rule out PDAC. Taking into a sonomorphologic and enhancement pa erns in FPL, a careful tumor last decades may be the key to suspect a pancreatic metastasis. Chen contrast enhancement pa erns of FPL using CEUS. All five metastas small cell lung cancer (SCLC) showed hyperenhancement in the ar enhancement. While the metastasis from squamous cell lung cance lesions, showed a rapid washout in the venous phase, all RCC me perenhanced. On the other hand, none of the PanNENs showed a hancement in the venous phase. Only the RCC metastases and showed continuous hyperenhancement. The authors calculated a r ity of 80.0% and a high specificity of 94.2% for the diagnosis of pa continuous hyperenhancement in the venous phase. In a strict sens lated to RCC metastases. The data are applicable to percutaneous EUS. On the basis of a few cases, no generalization can be given. The EUS probes and SonoVue usually leads to early destruction of the co exception is the intrapancreatic accessory spleen with long lasting e and CH-EUS [84,85]. The very rare PEComa has also been descri marked and prolonged hyperperfusion, with a washout of the lesi The description of a hyperenhancement of RCC metastases in the v contrasts with the results of Fusaroli et al. in CH-EUS. Here, th showed a slow washout. Of course, these different observations m se ings of the ultrasound systems and CEUS-software used for CE described metastases from RCC as circumscribed hypoechoic lesi The partially overlapping enhancement patterns raise differential diagnostic difficulties especially with regard to discrimination of pancreatic metastases from PanNENs. When a lesion is isoenhanced, it must be differentiated from inflammatory processes in the first place. Hypoenhanced lesions must be differentiated from the most common tumor-PDAC. Multiple lesions usually rule out PDAC. Taking into account the overlap of sonomorphologic and enhancement patterns in FPL, a careful tumor history including the last decades may be the key to suspect a pancreatic metastasis. Chen et al. [83] studied the contrast enhancement patterns of FPL using CEUS. All five metastases, four RCC and one small cell lung cancer (SCLC) showed hyperenhancement in the arterial phase and early enhancement. While the metastasis from squamous cell lung cancer, like most malignant lesions, showed a rapid washout in the venous phase, all RCC metastases remained hyperenhanced. On the other hand, none of the PanNENs showed a continuous hyperenhancement in the venous phase. Only the RCC metastases and three benign lesions showed continuous hyperenhancement. The authors calculated a relatively low sensitivity of 80.0% and a high specificity of 94.2% for the diagnosis of pancreatic metastasis by continuous hyperenhancement in the venous phase. In a strict sense, this can only be related to RCC metastases. The data are applicable to percutaneous CEUS and not to CH-EUS. On the basis of a few cases, no generalization can be given. The use of high-frequency EUS probes and SonoVue usually leads to early destruction of the contrast agent. The only exception is the intrapancreatic accessory spleen with long lasting enhancement in CEUS and CH-EUS [84,85]. The very rare PEComa has also been described on CH-EUS with marked and prolonged hyperperfusion, with a washout of the lesion at a late stage [82]. The description of a hyperenhancement of RCC metastases in the venous phase of CEUS contrasts with the results of Fusaroli et al. in CH-EUS. Here, the RCC metastases all showed a slow washout. Of course, these different observations may be due to different settings of the ultrasound systems and CEUS-software used for CE-EUS [61]. Liang et al. described metastases from RCC as circumscribed hypoechoic lesions with evidence of vascularity on color Doppler Imaging. On CEUS, the metastases showed rapid inhomogeneous arterial hyperenhancement. No significant washout was observed in the venous phase. There was a necrotic area in the center of the lesion [86]. Figures 4 and 5 show the different appearance and contrast behavior of a pancreatic RCC metastasis and a pancreatic metastasis from a rectal carcinoma in the arterial and venous phases. Yuan et al. and Nakamura et al. described intrapancreatic metastases of malignant melanoma in CEUS as isoenhanced to slightly hypoenhanced in the arterial phase. In the venous phase, the lesions were hypoenhanced [63,87]. The mild hypoenhancement in the arterial phase and hypoenhancement in the venous phase are suspicious for a malignant lesion and make differentiation from PDAC difficult. However, the presence of multiple lesions is an argument against PDAC (Figures 6-8). Annular enhancement has been described for a colorectal pancreatic metastasis [88] (See Figure 1). vascularity on color Doppler Imaging. On CEUS, the metastases showed rapid inhomogeneous arterial hyperenhancement. No significant washout was observed in the venous phase. There was a necrotic area in the center of the lesion [86]. Figures 4 and 5 show the different appearance and contrast behavior of a pancreatic RCC metastasis and a pancreatic metastasis from a rectal carcinoma in the arterial and venous phases. Yuan et al. and Nakamura et al. described intrapancreatic metastases of malignant melanoma in CEUS as isoenhanced to slightly hypoenhanced in the arterial phase. In the venous phase, the lesions were hypoenhanced [63,87]. The mild hypoenhancement in the arterial phase and hypoenhancement in the venous phase are suspicious for a malignant lesion and make differentiation from PDAC difficult. However, the presence of multiple lesions is an argument against PDAC (Figures 6-8). Annular enhancement has been described for a colorectal pancreatic metastasis [88] (See Figure 1).  In venous phase, slow washout at 42 s (c) with heterogeneous pa ern at 1.14 min (d). EUS revealed additional multiple small hypoechoic lesions that had escaped percutaneous sonography. The lesions were highly hypoechoic, almost anechoic, smoothly confined, without evidence of vessels in the EUS duplex (e). Lesions were stiffer on strain elastography in EUS (f). The diagnosis of metastatic malignant melanoma was confirmed percutaneously, ultrasound-guided from the lymph nodes and by EUS-FNP from the pancreas. Fusaroli et al. studied 11 intrapancreatic metastases by CH-EUS with SonoVue. Contrast uptake, enhancement pa ern, and contrast washout were assessed. All RCC metastases showed hyperenhancement, homogeneous pa ern, and slow washout. The one lymphoma metastasis also showed contrast hyperenhancement with homogeneous pa ern but fast washout. The melanoma metastasis was isoenhanced, with a heterogeneous contrast pa ern and fast washout [61]. All other metastases (breast, ovarian, and colon cancer) and sarcoma metastases in the study of Fusaroli et al. were hypoenhanced with homogeneous or heterogeneous pa ern and fast or slow washout [61]. Hypoenhancement in CH-EUS-uptake makes differentiation from PDAC impossible [61]. Despite having the same organ of origin, intrapancreatic metastases with different histologic types may also appear differently in B-scan mode and contrast pa ern (see Figures 9 and 10). In venous phase, slow washout at 42 s (c) with heterogeneous pattern at 1.14 min (d). EUS revealed additional multiple small hypoechoic lesions that had escaped percutaneous sonography. The lesions were highly hypoechoic, almost anechoic, smoothly confined, without evidence of vessels in the EUS duplex (e). Lesions were stiffer on strain elastography in EUS (f). The diagnosis of metastatic malignant melanoma was confirmed percutaneously, ultrasound-guided from the lymph nodes and by EUS-FNP from the pancreas. In venous phase, slow washout at 42 s (c) with heterogeneous pa ern at 1.14 min (d). EUS revealed additional multiple small hypoechoic lesions that had escaped percutaneous sonography. The lesions were highly hypoechoic, almost anechoic, smoothly confined, without evidence of vessels in the EUS duplex (e). Lesions were stiffer on strain elastography in EUS (f). The diagnosis of metastatic malignant melanoma was confirmed percutaneously, ultrasound-guided from the lymph nodes and by EUS-FNP from the pancreas. Fusaroli et al. studied 11 intrapancreatic metastases by CH-EUS with SonoVue. Contrast uptake, enhancement pa ern, and contrast washout were assessed. All RCC metastases showed hyperenhancement, homogeneous pa ern, and slow washout. The one lymphoma metastasis also showed contrast hyperenhancement with homogeneous pa ern but fast washout. The melanoma metastasis was isoenhanced, with a heterogeneous contrast pa ern and fast washout [61]. All other metastases (breast, ovarian, and colon cancer) and sarcoma metastases in the study of Fusaroli et al. were hypoenhanced with homogeneous or heterogeneous pa ern and fast or slow washout [61]. Hypoenhancement in CH-EUS-uptake makes differentiation from PDAC impossible [61]. Despite having the same organ of origin, intrapancreatic metastases with different histologic types may also appear differently in B-scan mode and contrast pa ern (see Figures 9 and 10). Fusaroli et al. studied 11 intrapancreatic metastases by CH-EUS with SonoVue. Contrast uptake, enhancement pattern, and contrast washout were assessed. All RCC metastases showed hyperenhancement, homogeneous pattern, and slow washout. The one lymphoma metastasis also showed contrast hyperenhancement with homogeneous pattern but fast washout. The melanoma metastasis was isoenhanced, with a heterogeneous contrast pattern and fast washout [61]. All other metastases (breast, ovarian, and colon cancer) and sarcoma metastases in the study of Fusaroli et al. were hypoenhanced with homogeneous or heterogeneous pattern and fast or slow washout [61]. Hypoenhancement in CH-EUS-uptake makes differentiation from PDAC impossible [61]. Despite having the same organ of origin, intrapancreatic metastases with different histologic types may also appear differently in B-scan mode and contrast pattern (see Figures 9 and 10). In the study by Dietrich et al. using CEUS and/or CH-EUS, 61% of metastases were hyperenhanced, 11% were isoenhanced, and 28% were hypoenhanced [24]. The number of RCC was 42.9% [24], suggesting that metastases other than RCC may also be hyperenhanced.
Huang et al. analyzed five different CEUS pa erns in a CEUS-based nomogram for malignant and benign solid pancreatic lesions. Hypoenhancement in the venous phase was a feature of malignant lesions [89]. It should be noted that this does not seem to be true for intrapancreatic RCC metastases [83,86]. The different enhancement pa erns are summarized in Table 6.  In the study by Dietrich et al. using CEUS and/or CH-EUS, 61% of metastases were hyperenhanced, 11% were isoenhanced, and 28% were hypoenhanced [24]. The number of RCC was 42.9% [24], suggesting that metastases other than RCC may also be hyperenhanced.
Huang et al. analyzed five different CEUS patterns in a CEUS-based nomogram for malignant and benign solid pancreatic lesions. Hypoenhancement in the venous phase was a feature of malignant lesions [89]. It should be noted that this does not seem to be true for intrapancreatic RCC metastases [83,86]. The different enhancement patterns are summarized in Table 6.
If there is differential diagnostic evidence for a lesion in the pancreas that would change the procedure such as a suspected metastasis from another malignancy in the history, EUS-guided sampling is indicated [90]. Before deciding on an EUS-guided sampling, one should always consider if and how this would change the further management of the patient. If an extensive metastatic spread of a primary tumor with a poor prognosis is already present and known, no consequences will result from securing the histology of pancreatic metastases or another tumor. Then, the EUS-FNP of the pancreas metastasis is not necessary. However, tissue diagnosis should be performed prior to initiating specific oncologic treatment. This enables a reliable specific diagnosis, provides information on prognostic markers and molecular signatures, and is a prerequisite for the initiation of targeted, personalized therapy. EUS-guided sampling will influence the patient's subsequent course of management if it is to confirm metastasis when there has been remission of a previous tumor. The risks of EUS-guided sampling are low [91][92][93][94][95].
Surgical resection is not the treatment of choice for every patient with pancreatic metastases. In the study by DeWitt et al., only 12.5% of patients [51] underwent surgery, and in the study by Krishna et al., only 7.5% of patients with pancreatic metastases underwent surgery [19]. However, in patients with isolated RCC metastases, due to the high long-term survival rates, surgery should always be considered [41,42].
Fritscher-Ravens et al. described solitary pancreatic metastases without a known primary tumor in 50% of cases [31]. In the report of Ardengh et al., the primary tumor was initially not previously known in 16% of cases, and was only established by EUS-guided sampling of the pancreatic metastasis [29] (Figure 11).
In the study of El Hajj et al., EUS-guided sampling diagnosed pancreatic metastases in six patients (12%), with a previous negative result in a computed tomography scan of the abdomen [60]. In DeWitt's study, 4/24 (17%) previously had a negative computed tomography scan [51].
If there is differential diagnostic evidence for a lesion in the pancreas that would change the procedure such as a suspected metastasis from another malignancy in the history, EUS-guided sampling is indicated [90]. Before deciding on an EUS-guided sampling, one should always consider if and how this would change the further management of the patient. If an extensive metastatic spread of a primary tumor with a poor prognosis is already present and known, no consequences will result from securing the histology of pancreatic metastases or another tumor. Then, the EUS-FNP of the pancreas metastasis is not necessary. However, tissue diagnosis should be performed prior to initiating specific oncologic treatment. This enables a reliable specific diagnosis, provides information on prognostic markers and molecular signatures, and is a prerequisite for the initiation of targeted, personalized therapy. EUS-guided sampling will influence the patient's subsequent course of management if it is to confirm metastasis when there has been remission of a previous tumor. The risks of EUS-guided sampling are low [91][92][93][94][95].
Surgical resection is not the treatment of choice for every patient with pancreatic metastases. In the study by DeWi et al., only 12.5% of patients [51] underwent surgery, and in the study by Krishna et al., only 7.5% of patients with pancreatic metastases underwent surgery [19]. However, in patients with isolated RCC metastases, due to the high longterm survival rates, surgery should always be considered [41,42].
Fritscher-Ravens et al. described solitary pancreatic metastases without a known primary tumor in 50% of cases [31]. In the report of Ardengh et al., the primary tumor was initially not previously known in 16% of cases, and was only established by EUS-guided sampling of the pancreatic metastasis [29] (Figure 11).  Figure 11. Intrapancreatic metastasis of RCC in a male, 70 years old. In lesion on the pancreatic tail. Native EUS showed the lesion to be hypoec (a), the pancreatic duct was not dilated, duplex EUS showed good va the lesion (b). Elastographically, the lesion was stiffer (c). Anamnestica performed due to trauma-related hemorrhage. This had occurred 20 y was not explicitly aware of a tumor. EUS-FNP (19 G) was performed, very bloody and yielded no result. It was not until the repeat (19 G) th could be diagnosed, although a renal tumor was not known.
In contrast to PDAC, the diagnosis of intrapancreatic metas bination of cytomorphology and ancillary studies. Immunocyto ing cytological smears, and extensive immunohistochemical sta ing cell block or tissue core sections and are crucial for differen NENs, and other rare pancreatic tumors, as well as for specifi metastases (see Table 7). This must be taken into account when ment and preparation [22,45,53,54,59], especially when target tre In addition, for the pathologist, information about a tumor's h ( Figure 12).  . Intrapancreatic metastasis of RCC in a male, 70 years old. Incidental finding of a 16 mm lesion on the pancreatic tail. Native EUS showed the lesion to be hypoechoic and smoothly confined (a), the pancreatic duct was not dilated, duplex EUS showed good vascularization with vessels in the lesion (b). Elastographically, the lesion was stiffer (c). Anamnestically, a nephrectomy had been performed due to trauma-related hemorrhage. This had occurred 20 years previously. The patient was not explicitly aware of a tumor. EUS-FNP (19 G) was performed, the aspirates of which were very bloody and yielded no result. It was not until the repeat (19 G) that the metastasis of an RCC could be diagnosed, although a renal tumor was not known.
In contrast to PDAC, the diagnosis of intrapancreatic metastases is based on the combination of cytomorphology and ancillary studies. Immunocytochemistry is possible using cytological smears, and extensive immunohistochemical stains can be performed using cell block or tissue core sections and are crucial for differentiation from PDAC, PanNENs, and other rare pancreatic tumors, as well as for specific diagnosis in pancreatic metastases (see Table 7). This must be taken into account when planning tissue procurement and preparation [22,45,53,54,59], especially when target treatment is considered [96]. In addition, for the pathologist, information about a tumor's history is very important (Figure 12). EUS-TCB after EUS-FNA was performed in n = 2 with negative cytology. Diagnosis of RCC metastasis was confirmed in n = 1, n = 1 was again false negative. (c) Figure 11. Intrapancreatic metastasis of RCC in a male, 70 years old. Incidental finding of a 16 mm lesion on the pancreatic tail. Native EUS showed the lesion to be hypoechoic and smoothly confined (a), the pancreatic duct was not dilated, duplex EUS showed good vascularization with vessels in the lesion (b). Elastographically, the lesion was stiffer (c). Anamnestically, a nephrectomy had been performed due to trauma-related hemorrhage. This had occurred 20 years previously. The patient was not explicitly aware of a tumor. EUS-FNP (19 G) was performed, the aspirates of which were very bloody and yielded no result. It was not until the repeat (19 G) that the metastasis of an RCC could be diagnosed, although a renal tumor was not known.
In the study of El Hajj et al., EUS-guided sampling diagnosed pancreatic metastases in six patients (12%), with a previous negative result in a computed tomography scan of the abdomen [60]. In DeWi 's study, 4/24 (17%) previously had a negative computed tomography scan [51].
In contrast to PDAC, the diagnosis of intrapancreatic metastases is based on the combination of cytomorphology and ancillary studies. Immunocytochemistry is possible using cytological smears, and extensive immunohistochemical stains can be performed using cell block or tissue core sections and are crucial for differentiation from PDAC, Pan-NENs, and other rare pancreatic tumors, as well as for specific diagnosis in pancreatic metastases (see Table 7). This must be taken into account when planning tissue procurement and preparation [22,45,53,54,59], especially when target treatment is considered [96]. In addition, for the pathologist, information about a tumor's history is very important (Figure 12).

Figure 12.
Intrapancreatic metastasis of breast carcinoma in a female, 71 years old. Incidental finding of a 25 mm mass on the pancreatic body during ultrasound examination. On EUS, the lesion was mildly polycyclic but smoothly circumscribed. A tiny anechoic area was seen centrally. Vessels were demonstrable on duplex. A neuroendocrine tumor was suspected. In the EUS-FNP, this was not Figure 12. Intrapancreatic metastasis of breast carcinoma in a female, 71 years old. Incidental finding of a 25 mm mass on the pancreatic body during ultrasound examination. On EUS, the lesion was mildly polycyclic but smoothly circumscribed. A tiny anechoic area was seen centrally. Vessels were demonstrable on duplex. A neuroendocrine tumor was suspected. In the EUS-FNP, this was not confirmed at first. The medical history was taken again, and it was found out that a breast carcinoma had been treated many years ago. Knowing this history, additional immunohistological examinations were performed, which revealed the metastasis of the breast cancer.
A consensus of pathologists from two centers described that cytomorphology alone was sufficient to diagnose metastatic RCC, SCLC, and hepatocellular carcinoma (HCC). They reported that morphology was sufficient to diagnose metastatic melanoma, breast cancer, and colon cancer, but confirmatory immunocytochemistry was routinely used for these neoplasms. These experts also suggested that esophageal, gastric, and non-small cell lung cancer (NSCLC) metastases cannot be confirmed by cytomorphology alone; immunocytochemistry is mandatory for the diagnosis of metastasis [51,60].
When comparing intrapancreatic metastases with PDAC and PanNENs, cytology from EUS-guided sampling was less sensitive in metastases [21]. Therefore, other authors advocated for the use of histological material [62,64,97] by EUS-guided sampling. Histologically adequate material allows for an evaluation of the specimen with preserved architecture and enables the performance of immunohistochemistry, which is essential for the specific diagnosis of metastasis, and also for the performance of molecular pathology studies [62,64].
Because renal cell carcinomas are highly perfused, EUS-guided specimens may also be bloody (Figure 11).

Computed Tomography (CT) and Magnetic Resonance Imaging (MRI)
Well-defined tumor margins, hyperattenuation in the arterial phase, maximal tumor enhancement in the arterial phase, the absence of pancreatic duct dilatation, absence of upstream pancreatic atrophy, absence of vascular involvement, and absence of bile duct dilatation were the most important features on CT distinguishing intrapancreatic metastases from PDAC [98]. In contrast-enhanced CT and MRI, metastases were hyperenhancing, isoenhancing, or hypoenhancing with a homogeneous or heterogeneous contrast pattern. Larger metastases showed a rim enhancement [99]. Small hypervascularized metastases usually show homogeneous enhancement, whereas larger ones tend to show peripheral enhancement [99,100].
RCC metastases were hypointense on MRI compared with normal pancreatic tissue on T1-weighted pre-contrast images and were hyperintense on T2-weighted. In contrastenhanced CT scans, RCC metastases had the same density as normal pancreatic tissue. After contrast application, most RCC metastases had high enhancement compared to normal pancreatic tissue on both CT and MRI on arterial and venous phase images [99]. Another study described the marked hyperattenuation of RCC metastases in the arterial phase that decreased during the portal phase, and in the larger lesions was associated with central liquefied areas [101].
With contrast-enhanced CT imaging, the differential diagnoses for hypervascular lesions include PanNENs and metastases (RCC and medullary thyroid carcinomas) [100].
Colorectal metastases had the same features as renal cell carcinoma metastases on both pre-contrast MRI and CT images. After contrast application, the outer margin was isoenhancing to the pancreatic tissue, while the inner parts of the tumor were not enhanced. This was attributed to necrosis in the central area of the metastases [99]. Metastases from breast and lung carcinoma were always hypoattenuating and leiomyosarcoma metastases were inhomogeneous [101].
Single melanoma metastases are described on CE-CT with peripheral or rim enhancement [87,102].

Renal Cell Carcinoma
Pancreatic RCC metastases may occur many years after tumor nephrectomy. In the meta-analysis with 414 patients by Huang et al., the median interval from nephrectomy to the appearance of pancreatic metastases was 93.6 months (range 5-288 months) [39]. In another meta-analysis including 855 observations, an interval between nephrectomy and the detection of isolated pancreatic RCC metastases of 9.6 ± 6.5 years was calculated [42]. In a few cases, a significant delay of pancreatic metastasis after the manifestation of primary RCC of 21 years [55], 28 years [56], 29 years [60], and 36 years [103] was described.
According to meta-analytic data (n = 893 observations from case and cohort studies), the majority of isolated intrapancreatic RCC metastases are observed metachronously (92.6%) [42]. Although intrapancreatic RCC metastases are usually solitary, multiple pancreatic localizations are observed in 36.4% of cases with a mean number of 3.1 and the highest reported number of intrapancreatic manifestations was 15 in this patient group [33,42,59]. Diffuse infiltration is also possible [66]. Tumor localization in the right or left kidney does not affect the localization of the metastasis in the pancreas [35,42].
Patients with pancreatic metastases from renal cell carcinoma have a better prognosis and life expectancy than intrapancreatic metastases from other primary tumors [35,44,104].
While the median life expectancy of all surgically resected patients with pancreatic metastases in the patient population of Reddy et al. was 3.7 years, it was 4.8 years for RCC and only 0.9 years for intrapancreatic melanoma metastases. Patients with metastatic breast cancer or melanoma did not survive longer than 2 years, and no patient with colon cancer, lung cancer, or sarcoma survived more than 5 years [35]. Only the patients with metastatic Langerhans cell histiocytosis and the patients with a seminoma were alive >11 years after resection of the pancreas [35]. In another study, the median survival of all patients with pancreatic metastases was 4.4 years, and that of patients with pancreatic RCC metastases was 8.7 years [44]. This important observation was underscored by meta-analytic data [39,43]. While the 5-year survival of all patients with surgical therapy of intrapancreatic metastases was 50%, it was 70.4% for the subgroup of patients with RCC metastases [43]. Consistently, another meta-analysis also reported a 5-year overall survival of 72% for patients after resection of intrapancreatic metastases from RCC, which was significantly superior compared to patients operated on for non-RCC intrapancreatic metastases [39].
Sellner et al. calculated cumulative 5-year and 10-year survival rates of 75.7% and 47.3%, respectively, from 415 casuistically reported postoperative observations in RCC pancreatic metastases [42]. According to the meta-analyses of Adler et al., Huang et al., and Sellner et al., none of the following factors significantly influence treatment outcomes: singular or multiple localization of RCC metastases; synchronous or a metachronous occurrence; the interval to nephrectomy; size of metastases [33,39,41,43,46]. Nevertheless, Rodger et al., in their most current systematic review of 35 cohort studies, reported that a metachronous presentation and a longer disease free interval before the presentation of intrapancreatic metastases were significantly associated with better survival outcomes [105]. Interestingly, even patients with isolated intrapancreatic RCC metastases who have foregone active therapy have a relatively favorable 3-year survival rate of 56%, although this is significantly worse compared to patients undergoing curatively intended therapy [46]. Blanco-Fernandez et al. reported postoperative overall survival rates at 1, 3, and 5 years of 96%, 88%, and 83%, respectively [56]. In this respect, a surgical approach in intrapancreatic RCC metastases is favored. Tanis et al. reported the survival benefit of RCC patients with resections of pancreatic metastases. Patients with pancreatic resection had a 2-year survival rate of 80% versus 72% compared to those without, and at 5 years of 41% versus 14% [106]. New effective therapeutic approaches emerged in metastatic RCC with tyrosine kinase inhibitors (TKIs), mTor inhibitors, and immune checkpoint inhibitors [42]. Comparable results to surgery have been described with TKIs in isolated RCC pancreatic metastases [107]. In the study by Santoni et al., surgical resection did not improve survival compared with TKI therapy. However, complete tumor healing is possible only with surgical resection in a certain percentage of patients with pancreatic RCC metastases [107].
Preliminary results of EUS-guided radiofrequency ablation in pancreatic metastasis from RCC appear promising and were reported as feasible, safe, and effective in a small number of patients [108].

Lung Cancer
Isolated pancreatic metastases from lung cancer are extremely rare. The few reports available in the literature indicate that SCLC is the most typical histologic subtype that metastasizes to the pancreas [48,109,110]. In most cases, they are not resectable at the time of diagnosis because the disease has already progressed and spread. Overall, 10 cases of NSCLC and 1 case with SCLC had a median survival time of 19 months (range 6-24 months) in a meta-analysis [48]. In another recent meta-analysis with 23 patients, 34.8% of them had extra pancreatic metastases, and the mean overall survival was 17.65 months after pancreatic resection [111] (Figure 13).

Colorectal Carcinoma
In a meta-analysis of pancreatic resection for pancreatic metastases of colorectal cancer, recurrences occurred frequently, with a median survival of 21 months (range 5-105 months) [48]. Newer meta-analytic data based on 24 cases showed a 5-year-survival of 46% [39]. In all patients, symptoms (abdominal pain and obstructive jaundice) were improved after surgical resection of the metastases until relapse. There are no comparative data on the chemotherapy of pancreatic metastases without resection. It is discussed that surgical resection may be performed as part of a palliative therapeutic approach to relieve clinical symptoms. This should be decided in an interdisciplinary approach [48].

Malignant Melanoma
Intrapancreatic malignant melanoma metastases have a poor prognosis. In a metaanalysis of surgical resected intrapancreatic melanoma metastases, the median survival time of these patients was 10 months (range 3-25 months) [48]. Patients with intrapancreatic melanoma metastases did not benefit from surgical resection [35]. In a recent metaanalysis, cumulative survival at 1 year, 3 years, and 5 years was 71%, 35%, and 26%, respectively. The median survival was 24 months. Incomplete resection and concurrent extra pancreatic metastases were factors that negatively affected survival. In contrast, for solitary pancreatic metastases, the 1-year, 3-year, and 5-year survival rates were 76%, 43%, and 41%, respectively. The authors concluded that curative pancreatic resection may positively influence survival in selected patients [112].

Breast Cancer
Intrapancreatic metastases may become apparent after a long latency period (median 39.5 months, range 0-216) [48]. It has not been possible in pancreatic metastases from breast cancer to determine disease progression without surgical resection and to assess the true survival benefit after metastasectomy. However, in selected patients, surgical resection could play a palliative role in combination with chemotherapy, endocrine treatment, and radiotherapy in the multimodal treatment of metastatic breast carcinoma. This should be decided with a multidisciplinary approach [48].

Sarcoma
Metastatic sarcoma generally has a poor prognosis. The radical surgical resection of pancreatic sarcoma metastases is the main therapeutic option. In the surgical resection statistics, only individual cases are reported that do not allow generalization [32,34,35,37,40,48,49]. A few case reports are described [113][114][115][116]. Mostly, these are patients who have already undergone surgery for other metastases. After resection, there is

Colorectal Carcinoma
In a meta-analysis of pancreatic resection for pancreatic metastases of colorectal cancer, recurrences occurred frequently, with a median survival of 21 months (range 5-105 months) [48]. Newer meta-analytic data based on 24 cases showed a 5-year-survival of 46% [39]. In all patients, symptoms (abdominal pain and obstructive jaundice) were improved after surgical resection of the metastases until relapse. There are no comparative data on the chemotherapy of pancreatic metastases without resection. It is discussed that surgical resection may be performed as part of a palliative therapeutic approach to relieve clinical symptoms. This should be decided in an interdisciplinary approach [48].

Malignant Melanoma
Intrapancreatic malignant melanoma metastases have a poor prognosis. In a metaanalysis of surgical resected intrapancreatic melanoma metastases, the median survival time of these patients was 10 months (range 3-25 months) [48]. Patients with intrapancreatic melanoma metastases did not benefit from surgical resection [35]. In a recent meta-analysis, cumulative survival at 1 year, 3 years, and 5 years was 71%, 35%, and 26%, respectively. The median survival was 24 months. Incomplete resection and concurrent extra pancreatic metastases were factors that negatively affected survival. In contrast, for solitary pancreatic metastases, the 1-year, 3-year, and 5-year survival rates were 76%, 43%, and 41%, respectively. The authors concluded that curative pancreatic resection may positively influence survival in selected patients [112].

Breast Cancer
Intrapancreatic metastases may become apparent after a long latency period (median 39.5 months, range 0-216) [48]. It has not been possible in pancreatic metastases from breast cancer to determine disease progression without surgical resection and to assess the true survival benefit after metastasectomy. However, in selected patients, surgical resection could play a palliative role in combination with chemotherapy, endocrine treatment, and radiotherapy in the multimodal treatment of metastatic breast carcinoma. This should be decided with a multidisciplinary approach [48].

Sarcoma
Metastatic sarcoma generally has a poor prognosis. The radical surgical resection of pancreatic sarcoma metastases is the main therapeutic option. In the surgical resection statistics, only individual cases are reported that do not allow generalization [32,34,35,37,40,48,49]. A few case reports are described [113][114][115][116]. Mostly, these are patients who have already undergone surgery for other metastases. After resection, there is a 30-50% risk of recurrence. In a tertiary-referral hospital for soft-tissue sarcoma with 6744 new cases of soft-tissue sarcoma, 7 underwent duodenopancreatectomy for sarcoma metastases to the pancreatic head. The median survival was 21 months (range: 4 days to 86 months) [117].

Figure 14.
Merkel cell carcinoma. Pancreatic lesion 10mm in size without main pancreatic duct dilation was detected as a low echoic lesion (arrowhead) with B-mode EUS (left). CE-EUS detected a pancreatic lesion with hyper-intensity of enhancement (arrowhead) as compared to that of surrounding pancreatic tissue (right).

Surgical Resection of Intrapancreatic Metastasis
Whether pancreatic resection should be performed depends on the primary tumor and the patient's prognosis, and whether the metastasis is limited to the pancreas or whether other organ metastases are present. Resection can be performed with curative intent but may also be offered palliatively for symptom relief. Risks and benefits must be carefully evaluated in the individual situation. Depending on the location and number of metastases in the pancreas, surgical resection can be performed as partial duodenopancreatectomy, distal pancreatectomy, total duodenopancreatectomy, central pancreatectomy, and local tumor resection [42]. Depending on the primary tumor, extension, and tumor extent, systemic therapy must be considered.

Conclusions
When a patient with a history of tumor is diagnosed with a focal lesion in the pancreas, the possibility of metastasis must always be considered. The likelihood increases when multiple lesions are present. Metastases are in the majority of cases hypoechoic and are frequently well circumscribed. The surgical resection of pancreatic metastases is selectively based on the tumor stage and other sites of metastatic disease. At CEUS and CH-EUS, metastases are usually hypoenhanced. In contrast, RCC metastases are well vascularized and show hyperenhancement like PanNENs. For CEUS, there are case reports of the long-lasting enhancement of RCC metastases even in the venous phase. This was not true for PanNENs, but also not true for RCC metastases in CH-EUS. Melanoma metastases were initially isoenhanced to slightly hypoenhanced in case reports, and then also showed washout. Except for RCC metastases in CEUS, washout in the venous phase was a malig- Figure 14. Merkel cell carcinoma. Pancreatic lesion 10mm in size without main pancreatic duct dilation was detected as a low echoic lesion (arrowhead) with B-mode EUS (left). CE-EUS detected a pancreatic lesion with hyper-intensity of enhancement (arrowhead) as compared to that of surrounding pancreatic tissue (right).

Surgical Resection of Intrapancreatic Metastasis
Whether pancreatic resection should be performed depends on the primary tumor and the patient's prognosis, and whether the metastasis is limited to the pancreas or whether other organ metastases are present. Resection can be performed with curative intent but may also be offered palliatively for symptom relief. Risks and benefits must be carefully evaluated in the individual situation. Depending on the location and number of metastases in the pancreas, surgical resection can be performed as partial duodenopancreatectomy, distal pancreatectomy, total duodenopancreatectomy, central pancreatectomy, and local tumor resection [42]. Depending on the primary tumor, extension, and tumor extent, systemic therapy must be considered.

Conclusions
When a patient with a history of tumor is diagnosed with a focal lesion in the pancreas, the possibility of metastasis must always be considered. The likelihood increases when multiple lesions are present. Metastases are in the majority of cases hypoechoic and are frequently well circumscribed. The surgical resection of pancreatic metastases is selectively based on the tumor stage and other sites of metastatic disease. At CEUS and CH-EUS, metastases are usually hypoenhanced. In contrast, RCC metastases are well vascularized and show hyperenhancement like PanNENs. For CEUS, there are case reports of the long-lasting enhancement of RCC metastases even in the venous phase. This was not true for PanNENs, but also not true for RCC metastases in CH-EUS. Melanoma metastases were initially isoenhanced to slightly hypoenhanced in case reports, and then also showed washout. Except for RCC metastases in CEUS, washout in the venous phase was a malignancy criterion in both CEUS and CH-EUS. EUS-guided sampling is a valuable diagnostic step, as metastasis can be confirmed cytologically/histologically. This is especially true for the confirmation of recurrence in patients who have been in long-term remission. In some cases, metastases could be detected even with a negative CT. With EUS-guided sampling, the identification of the primary tumor can be made with a high degree of reliability. For the diagnostic workup, ancillary studies for immunocytochemistry are recommended in addition to cytomorphology. This requires material processing using a cellblock technique or by obtaining core biopsies by EUS-guided sampling with sufficient material for histological processing.
Management depends on the primary tumor, the number of metastases in the pancreas, and the overall tumor stage. Solitary RCC metastases have been resected with acceptable long-term outcomes. Surgical resections have also been reported in other tumor entities. This may be performed with curative intent. The overall condition of the patient must be weighed up together with the risks, such as postoperative morbidity or mortality, the options of modern chemotherapy and immunotherapy, and the expected survival benefit.