Lipid-Rich Variant of Pancreatic Endocrine Neoplasms
Rajendra Singh, MD,* Olca Basturk, MD,* David S. Klimstra, MD,¡ Giuseppe Zamboni, MD,¿ Runjan Chetty, MD,§ Sanaa Hussain, MD,¡ Stefano La Rosa, MD,” Asli Yilmaz, MD,7 Paola Capelli, MD,¿ Carlo Capella, MD,” Jeanette D. Cheng, MD,* and N. Volkan Adsay, MD*
Abstract: Most pancreatic endocrine neoplasms (PENs) show characteristic and well-recognized endocrine morphology; however, a lipid-rich pattern, which can present a diagnostic problem in biopsies, has been reported, mostly as individual cases. Some have been included in descriptions of the rare clear-cell variant associated with von Hippel-Lindau (VHL) syndrome. The histogenesis, clinicopathologic characteristics, and significance of this lipid-rich pattern have not been unraveled. In this study, 11 PENs exhibiting foamy, microvesicular cytoplasm were analyzed. In some cases, the nuclei were distorted by the vesicles, and the usual endocrine chromatin pattern was not evident. The growth pattern was relatively diffuse, with vague compartmentalization of the cells by a delicate vasculature; prominent nesting was noted in only 4 cases. Pathology reports indicated substantial diagnostic challenge in these cases; on biopsies, 1 case was originally diagnosed as adrenal cortical carcinoma, another as renal cell carcinoma, a third as solid-pseudopapillary tumor, and a fourth had a fine needle aspiration cytologic diagnosis of ade- nocarcinoma. All cases were chromogranin and synaptophysin positive. Electron microscopy in 3 cases confirmed the cytoplasmic vesicles to be lipid vacuoles. Neurosecretory granules were also evident. Clin- ically, as in conventional PENs, there appeared to be two distinct subsets: Two cases were familial or functional/syndromic (1 with VHL and the other with MEN-1 and glucagonoma syndrome) and occurred in younger adults (ages 41 and 47 years); the majority (n = 9) were nonfunctional/nonsyndromic and nonfamilial. The latter group was mostly represented by elderly males (mean age: 65 vs. 58 years in conventional sporadic PENs). Immunohistochemically, markers impli- cated in VHL-associated neoplasia, including HIF-1a, inhibin, and Melan-A (in clear-cell PENs) and MUC6 (in serous cystadenomas) were mostly negative in lipid-rich PENs (1 of 10, 1 of 10, 0 of 10 and 0 of 10, respectively). In conclusion, the lipid-rich pattern, remi- niscent of adrenal cortical cells, represents a distinct subset of PENs. It presents a diagnostic challenge for surgical pathologists, especially in biopsies. EM supports the name lipid-rich for this variant. The
From the *Department of Pathology, Harper Hospital and Karmanos Cancer Institute, Wayne State University, Detroit, MI; }Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, NY; #Department of Pathology, Verona University, Italy; §Department of Pathology, University Health Network and University of Toronto, Toronto, Ontario, Canada; Department of Pathology, University of Insubria, Varese, Italy; and {Department of Pathology, University of Calgary, Calgary, Alberta, Canada.
Reprints: N. Volkan Adsay, MD, Department of Pathology, Harper University Hospital, 3990 John Road, Detroit, MI 48201 (e-mail: adsayv@med. wayne.edu).
Copyright @ 2006 by Lippincott Williams & Wilkins
findings suggest that the pathogenesis of lipid-rich tumors may be different from the VHL-associated clear-cell variants of PENs.
Key Words: lipid-rich pancreatic endocrine tumors, VHL, clear cell (Am J Surg Pathol 2006;30:194-200)
P ancreatic endocrine neoplasms (PEN), also referred to as islet cell tumors, are characterized by relatively uniform cuboidal cells arranged in trabeculae, festoons, solid nests, or gland-like structures. The nuclei exhibit a distinctive chroma- tin pattern referred to as salt-and-pepper.5,13,30 In typical cases, the cytoplasm of PEN cells is eosinophilic to amphophilic and finely granular. Rarely, however, the cytoplasm may appear clear. This clear-cell change mostly has been reported in von Hippel-Lindau (VHL) syndrome-associated PENs, which are seen in 15% patients with VHL.5,9 Half of these are reported to be of the clear-cell type. Interestingly, clear-cell change in PENs is rarely reported in a non-VHL setting.2,3
In 1997, a different pattern was recognized by Ordonez and Silva in a PEN report as “islet cell tumor with vacuolated lipid-rich cytoplasm.”24 These lipid-rich PENs often pose a diagnostic problem in biopsies because it is a substantial deviation from the usual appearance of PENs. The histogen- esis, clinicopathologic characteristics, and the significance of these lesions remain unclear. Their relationship to clear-cell and VHL-associated PENs is also unknown.
We present 11 cases of PENs showing foamy, micro- vesicular cytoplasm, only 2 of which were associated with a familial syndrome. The clinicopathologic characteristics of these cases are described.
MATERIALS AND METHODS
Eleven cases of PENs characterized histologically by foamy, microvesicular cytoplasm were identified in the files of the authors’ institutions (Harper Hospital and Karamonas Cancer Institute of Wayne State University, Detroit MI; Memo- rial Sloan-Kettering Cancer Center, New York, NY; Verona University, Italy; University of Insubria, Varese, Italy; and University of Toronto, Toronto, Ontario, Canada) or their consultation files. The pathology material, including the slides and reports, were reviewed. The original diagnoses were noted along with any diagnostic difficulty identified in the reports. Where necessary, patients’ charts were reviewed for demo- graphic factors such as age, gender, location and size of tumor, and functionality status and to determine if there was any syndrome associated with the PEN.
Am J Surg Pathol . Volume 30, Number 2, February 2006
Immunohistochemical stains for chromogranin and synaptophysin were performed to verify the endocrine nature of these tumors. Markers previously documented in VHL- associated clear-cell tumors including HIF-1a,14 inhibin, 13 and Melan-A35 were also performed. MUC6,13 which has been found to be positive in VHL associated serous cystadenomas, was also done. For HIF-1a, the Catalyzed Signal Amplifica- tion system (Dako Corporation, Carpinteria, CA) was used. Briefly, slides were deparaffinized, rehydrated in graded alcohols, and placed in Tris-buffered saline solution. Antigen retrieval was achieved by autoclaving the slides for 10 minutes in citrate buffer (pH 6.0; Dako, Denmark). Endogenous peroxidase was blocked by incubating the slides in 0.3% H2O2 for 15 minutes. Sections were incubated for 15 minutes with normal serum from donor species of secondary antibody followed by an overnight incubation at room temperature with the primary antibody (anti-HIF-1 [clone H1alpha67], Novus Biologicals, Inc., Littleton, CO; 1:1000 dilution). Link anti- body was applied for 15 minutes. Sections were incubated for 15 minutes with streptavidin biotin complex and then another 15 minutes with an amplification reagent. Streptavidin- peroxidase solution was applied, and the sections were incubated for 15 minutes. Finally, peroxidase activity was developed with AEC (3-amino-9-ethylcarbazole) solution. In between the steps, slides were washed in Tris-buffered saline. Counterstaining was done with Mayer’s hematoxylin. A positive (renal carcinoma) and a negative control slide were always included in each immunostaining. For HIF-1a, only cytoplasmic staining was regarded as a positive stain as both tumors and negative controls showed evidence of nuclear and stromal background staining. For inhibin, MUC6, and Melan-A, the stains were performed using the streptavidin-biotin tech- nique. Primary antibodies for these markers, their source, dilu- tion, and antigen retrieval methods and positive controls for each are listed in Table 1. The negative control consisted of sections of colonic cancer, liver, kidney, prostate, and ovary in- cubated with nonimmune mouse serum without primary anti- body. Staining was confined to the cytoplasm of the positive control specimens for all three antibodies.
In 2 cases, ultrastructural examination was performed on tissue fixed in 2.5% glutaraldehyde. In another case, the tissue was obtained from formalin-fixed, paraffin-embedded speci- men. The samples were postfixed with osmium tetroxide,
stained with lead citrate, and examined under transmission electron microscopy.
RESULTS
Clinical Findings and Original Pathologic Workup
Nine of the 11 patients were males. Two of the cases were associated with familial conditions: one with VHL syndrome and another with MEN-1 syndrome. These latter 2 patients were 41 and 47 years old, respectively, whereas the average age of the nonfamilial group was 65 (range, 43-87 years). One patient had a previous history of renal cell carcinoma. Seven of the lesions were detected in the tail of the pancreas with the remaining four in the head region. The clinical findings of each case, including age, gender, location of tumor, size and other pertinent factors, are documented in Table 2.
In 4 of the cases, there were initial problems establishing the correct diagnosis. Case no. 9 (Fig. 1) was originally diag- nosed on a needle core biopsy as an “adrenal cortical neoplasm, either heterotopic or metastatic”; however, because of the absence of a radiologically demonstrable adrenal lesion and because the biopsy was indeed from a 9-cm pancreatic mass, the case was reevaluated. The tumor was noted to be negative for CD68, PAS, EMA, Mart-1, and S-100, strongly positive for neuroendocrine markers and vimentin, and focal positivity for keratin. Subsequently, serum hormone levels proved this lesion to be gastrin producing. Another patient (case no. 10) had a previous history of nephrectomy for renal cell carci- noma. He presented with a mass measuring 7.9 cm in greatest dimension in the left renal fossa, which was difficult to distinguish from the surrounding inferior spleen and pancre- atic tail region. Biopsy showed cells that were uniform in shape and size and contained foamy cytoplasm with intra- cytoplasmic vesicles. The cells were strongly positive for cytokeratin and negative for EMA, vimentin, and inhibin. Based on the histology and immunohistochemical profile, this was signed out as an epithelial lesion consistent with renal cell carcinoma. Subsequently, the tumor was found to be positive for neuroendocrine markers and correctly diagnosed as lipid- rich PEN. Another case (case no. 11) was sent to the authors with the diagnosis of solid pseudopapillary tumor of the
| TABLE 1. Immunohistochemical Stains | ||||||
|---|---|---|---|---|---|---|
| Antibody | Type | Source | Clone | Dilution | Antigen Retrieval | Positive Control |
| HIF1-& | Mouse monoclonal | Novus Biologicals, Littleton, CO | Hif1 alpha67 | 1:1000 | EDTA for 30 min | Endometrial cancer |
| Inhibin | Mouse monoclonal | Cell Marque, Hot Springs, AZ | Inhibin alpha igg2a | Prediluted | EDTA steam for 30 min | Normal ovary |
| MUC6 | Mouse monoclonal | Vector Lab, Burlingame, CA | Muc6 clh5 | 1:50 | Citrate buffer steam for 60 min | Stomach |
| Chromogranin A | Rabbit polyclonal | Dako Corp, Carpinteria, CA | 1:300 | EDTA for 20 min | Adrenal | |
| Synaptophysin | Mouse monoclonal | Biogenix, Bogata, Columbia | Snp88 | 1:40 | citrate buffer steam for 20 min | Brain |
| Melan-A | Mouse monoclonal | Dako Corp, Carpinteria, CA | Melan-A103 igg1, kappa | 1:50 | EDTA steam for 30 min | Melanoma |
| TABLE 2. Clinicopathologic Findings | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Case No. | Age (yr) | Sex | History/Presentation | Location | Size | % of MIB1 Staining | % of Lipid-Rich Component | EM | Function/Syndrome | EM | Original Diagnosis |
| 1 | 43 | M | Transsexual; heavy hormonal therapy | Head | 3 | 5 | 80 | ||||
| 2 | 59 | F | Preoperative and postoperative diarrhea | Head | 10 | 3 | 75 | ||||
| 3 | 55 | M | Pancreatic tail mass | Tail | 11 | 5 | 25 | Lipid | Lipid | ||
| 4 | 66 | M | Distal pancreatic mass with invasion to hilum of spleen | Tail | 7.9 | 5 | 40 | Lipid | Lipid | ||
| 5 | 55 | M | Distal pancreatic "cystic" lesion | Tail | 4.8 | 2 | 90 | Cytology: adenocarcinoma | |||
| 6 | 47 | M | Two pancreatic masses with glucagon production | Tail | 3.5 | 1 | 80 | Glucagon/MEN1 | |||
| 7 | 41 | F | Pancreatic tumor | Head | 2.5 | 3 | 90 | VHL | D/D: paraganglioma | ||
| 8 | 49 | M | 11-cm mass in pancreatic tail on abdominal echo | Tail | 11 | 3 | 75 | Lipid | Lipid | ||
| 9 | 87 | M | Mass in abdomen | Head | 9.0 | 2 | 80 | Adrenocortical lesion | |||
| 10 | 75 | M | Mass in postnephrectomy bed | Tail | 10 | 3 | 85 | Renal cell carcinoma | |||
| 11 | 70 | M | Well-circumscribed pancreatic mass on echo | Head | 2.1 | 1 | 80 | Solid pseudopapillary tumor | |||
pancreas. This was a well-circumscribed pancreatic mass measuring 2.1 cm in greatest dimension in a 70-year-old man. Immunoperoxidase stains revealed the tumor cells to be pos- itive for cytokeratin, vimentin, alpha-1-antitrypsin, and syn- aptophysin and weakly positive for chromogranin. The tumor cells were negative for mucin and glycogen. The fourth case
was referred to a tertiary cancer center. This was a 55-year-old man with a distal pancreatic cystic lesion, which on aspiration was interpreted as atypical cells consistent with adenocarci- noma. Still, another case (case no. 7) was diagnosed as a prob- able neuroendocrine tumor with the differential diagnosis of paraganglioma.
Macroscopic Findings
The size of the tumors varied from 2.1 to 11 cm (median size, 6.8 cm). All the lesions were confined to the pancreas except for one (case no. 4), which invaded into the spleen and another (case no. 10), which was described to be “adjacent to the postnephrectomy bed.” Three of the lesions were focally cystic, with the largest cystic area measuring 4.8 cm in diameter. The tumors were described to have yellow-tan color. Rare foci of hemorrhage and necrosis were noted in 3 of 11 and 2 of 11 cases, respectively.
Microscopic Findings
Microscopically, the growth pattern was relatively diffuse with only vague compartmentalization of the cells by a delicate vasculature (Fig. 2); prominent nesting was seen only in 4 cases (Fig. 3). Individual cells were small to medium in size and showed abundant foamy, microvesicular cytoplasm in at least 75% of the tumor cells. The cells closely resembled xanthoma cells, and in the context of sinusoidal vasculature, the overall pattern was highly reminiscent of cells of the zona fasciculate of the adrenal cortex. The nuclei were round with finely granular chromatin and inconspicuous nucleoli in some areas; however, in other regions, the nuclei were relatively small, pyknotic, and partially obscured or even scalloped by
@ 2006 Lippincott Williams & Wilkins
the intracytoplasmic vesicles (Figs. 3 and 4). The latter finding was evident at least focally in 3 of 11 cases. In these areas, the endocrine chromatin pattern (“salt-and-pepper”) was not evi- dent. Vascular invasion was identified in 3 cases, and 1 of the cases showed direct invasion to a lymph node. No perineurial invasion was identified. Nuclear atypia and notable mitotic activity (2/50 HPF, microscopic field area of 0.19635 mm2) was observed in 1 of 11 cases.
Immunohistochemical Findings
Immunohistochemically, all the tumors were positive for chromogranin (Fig. 5) and synaptophysin, some focally and some very strongly. None of the 10 specimens stained showed any positivity for MUC6 or Melan-A, whereas only 1 case
each was positive for HIF-1& (Fig. 6) or inhibin, which was not the patient associated with VHL. Positivity for MIB1 was very focal ranging from 1% to 5% (Table 2).
Ultrastructural Findings
Electron microscopic examination performed on 3 cases showed epithelial cells separated from the stroma by a base- ment membrane. There were multiple small, round homoge- nous dense-core neurosecretory granules providing ultrastructural evidence of neuroendocrine differentiation. The cytoplasm showed focally prominent lipid droplets as well as markedly dilated cisternae of endoplasmic reticulum, corresponding to the vesicles identified on hematoxylin and eosin sections (Fig. 7).
DISCUSSION
Lipid-Rich Variant Is a Distinct Subtype of PENs Prone to Misdiagnosis
This study documents that the lipid-rich variant of PEN, which was first recognized in 1997 by Ordonez and Silva in a case report titled “islet cell tumor with lipid rich vacuolated cytoplasm,“24 is a distinctive subtype of PEN. The foamy, microvesicular cytoplasm, which is confirmed in this study by ultrastructural examination as lipid-filled, is exceedingly un- usual for endocrine neoplasia outside the setting of adrenal or thyroid, and for this reason the lipid-rich variant of PEN is highly prone to misdiagnosis, especially in biopsy specimens. If the pathologist is not aware of this entity and has not been informed of the clinical setting, misdiagnosis is particularly likely because this variant often lacks many typical attributes of endocrine neoplasia. The salt-and-pepper chromatin pattern
may be absent, the growth pattern may be more diffuse than is typical of endocrine tumors, and the vascular pattern has a more sinusoidal quality, which imparts a histologic appearance virtu- ally indistinguishable from that of adrenal cortical neoplasia.22 Indeed, 1 case was misdiagnosed as an adrenal cortical tumor; and in another, the cells were thought to be the xanthomatous macrophages seen in solid-pseudopapillary tumor.
Because the vesicles forming the cytoplasm are “clear,” lipid-rich PENs have occasionally been included under the heading of “clear-cell PENs.” Clear-cell PENs, which seem to be substantially more common than the lipid-rich ones, were first recognized in the literature as a case report in 19838; and since then, sporadic cases have appeared in the literature, most often associated with VHL syndrome.5,9,20,21 In some studies, clear-cell change has been reported in 15% of all PENs.5,9 Determination of the origin of these clear-cell lesions is often a diagnostic dilemma as well. The differential diagnosis of such lesions arising in the pancreatic area may be both primary and secondary tumors that may be rich in lipid, glycogen, or mucin (Table 3). These include foamy-gland adenocarcinoma,2,18,26 low-grade mucinous adenocarcinoma, ductal adenocarcinoma with clear-cell features, serous adenomas, especially its solid variant,19,32 solid pseudopapillary tumor,7 and sugar tumor of the pancreas.36 In addition, metastatic malignant neoplasms such as renal cell carcinoma, adrenal tumors, steroid-secreting tumors of the ovary, and clear-cell hepatocellular carcinoma also enter in the differential diagnosis of these tumors.1,27,31 Indeed, based on the history of a “kidney tumor,” 1 of our patients was diagnosed as metastatic renal cell carcinoma for more than 3 years (although the cytoplasm was “foamy and microvesicular” rather than “clear” and renal cell carcinomas rarely have microvesicular cytoplasm). Immunohistochemistry and electron microscopy are useful adjuncts in elucidating the differential diagnosis of these tumors. Immunoreactivity for chromogranin and synaptophysin usually points to an endo- crine origin. CD68, which is positive in both macrophages and adrenal cortical tumors, is negative in endocrine tumors.5 Electron microscopy reveals the dense core neurosecretory granules typical of endocrine lesions.
Unlike Clear-Cell PENs, Lipid-Rich PENs Do Not Show Any Significant Association With VHL Syndrome
In 2001, Hoang et al reported clear-cell endocrine pan- creatic tumors to be a distinctive neoplasm of VHL syndrome and that this variant did not occur much outside the context of VHL.9,16 All 5 cases described were females with an average age of 43 years. Later, a handful of endocrine tumors of the pancreas consisting predominantly or entirely of clear cells, but not associated with VHL disease, were reported.24 Indeed, some of these may be the lipid-rich variant rather than clear- cell. For instance, 1 case involved a tumor of the left supra- adrenal region that initially was interpreted to be an adrenal cortical carcinoma; however, this case was later proven to arise in the pancreatic tail and was a “malignant islet cell tumor” producing vasoactive intestinal peptide.8 Five of the other cases have been described in a review article of clear-cell neoplasms of the endocrine system and thymus, without detailed descriptions of each case.34 Overall, most of the clear-cell
@ 2006 Lippincott Williams & Wilkins
| Tumors | Histology | Immunohistochemistry | |
|---|---|---|---|
| Positive | Negative | ||
| Pancreatic | |||
| Solid pseudopapillary tumor (with areas of foamy macrophages) | Pseudopapilla; eosinophilic globules; grooved nuclei with nuclear overlapping | Beta-catenin, Progesterone receptors | CHR |
| Sugar tumor | Solid growth pattern; large epithelioid cells with clear or eosinophilic granular cytoplasm | HMB45 | CHR, Keratins |
| Serous adenoma (esp. solid variant) | Centrally located small round nuclei with dense homogenous chromatin; clear cytoplasm | PAS, CAM5.2, EMA | HMB45, CHR |
| Ductal carcinoma (foamy gland type and clear cell type | Glands, infiltrative pattern, desmoplastic stroma | CK7 | CHR, SYN |
| Nonpancreatic | |||
| Renal cell carcinoma | Cells with clear or eosinophilic cytoplasm with a delicate vascular network | RCC marker, CD10, Vimentin | CHR, SYN |
| Adrenocortical tumors | Alveolar pattern, nuclear atypia | Vimentin, Inhibin, CD68 | CHR, SYN, Keratins |
| Steroid secreting tumors of ovary | Round or polyhedral cells with granular eosinophilic cytoplasm (with lipofuscin pigment) | Inhibin, Keratins, CD99 | CHR, SYN |
| Clear cell hepatocellular carcinoma Clear cell tumors (genitourinary type) | Cords; prominent nucleoli; clear cytoplasm Clear cells; hobnailing; nuclear atypia | Hepar-1, CEA(canalicular) Keratins, EMA | CHR, SYN CHR, SYN |
PENs reported in literature have been associated with VHL. In contrast, in this series of 11 cases with lipid-rich PENs, only 1 patient had clinical evidence of VHL.
Lipid-rich PENs also lack the immunophenotypic evi- dence of association with VHL. An interesting finding in our study is the absence of HIF-1& immunoexpression in PENs. VHL is caused by germline mutations of the VHL tumor suppressor gene located on chromosome 3p25.3 The major lesions in VHL include hemangioblastomas in the central nervous system and retina, clear-cell renal cell carcinomas, pheochromocytomas, pancreatic serous adenomas and PENs, epididymal cystadenomas, endolymphatic sac tumors, carci- noid tumors, and multiple cysts of the kidney, pancreas, and epididymis.28 The VHL protein (pVHL) is widely expressed in human tissues and appears to have several distinct functions as a gatekeeper protein.17 The principal demonstrated function of the pVHL is the negative regulation of hypoxia-inducible factor-1 (HIF-1) by hypoxia inducible transcription factors that form the HIF-1 transcription complex.11 Dysfunction of pVHL caused by the VHL gene abnormality results in the absence of HIF-1 degradation and consequent overproduction of growth factors such as vascular endothelial growth factor, platelet de- rived growth factor beta, and transforming growth factor-alpha.25 These lead to uncontrolled activation of membrane bound receptor kinases and tumor growth. HIF-1& up-regulation is seen in clear-cell tumors associated with VHL but is absent in PENs not associated with VHL, thus pointing to a different etiology for these lesions. In this study, HIF-1 & expression was detected in only 1 of 10 lipid-rich PENs in a case that was not associated with VHL. This absence of HIF-1& in the patient with VHL is rather intriguing and suggests that other pathways might be responsible for tumorigenesis.
Markers commonly detected in other VHL-associated clear-cell, nonendocrine tumors are also negative in lipid-rich
PENs. In a study by Kosmahl et al, 100% of serous cystadenomas of the pancreas associated with VHL showed positivity for MUC6, whereas 50% were positive for alpha- inhibin.º Inhibin, along with Melan-A, has also been shown to be positive in other clear-cell tumors such as those arising in the ovary. In a study by Sinkre et al,29 four endocrine pan- creatic tumors associated with VHL disease were diffusely im- munoreactive for inhibin, whereas the two classic carcinoids of the gallbladder, the two renal cell carcinomas associated with VHL, and 11 of the 13 sporadic endocrine pancreatic tumors, including one with clear-cell features, did not show reactivity for inhibin. In our study, only 1 case showed positivity for inhibin, whereas none of the cases showed any positivity for either MUC6 or Melan-A. This further corroborates that the lipid-rich lesions may not be related to alterations of VHL gene.
As in Other PENs, Lipid-Rich PETs Appear to Occur in Two Distinct Clinical Settings
Two of the 11 lipid-rich PENs were familial-associated or functional/syndromic (1 VHL and 1 glucagon-producing MEN-1) and both were seen in younger patients (ages, 41 and 47 years). In contrast, nonfunctional (nonsyndromic) and nonfamilial cases (n = 9) were seen predominantly in elderly males (mean age, 65 years, vs. 58 years in ordinary PENs). It appears that lipid-rich PENs, especially the more striking examples, are seen in substantially older patients and tend to have a larger size (average size, 6.8 cm) than conventional nonfunctional pancreatic endocrine tumors (average size reported as 4.7 cm).1º This raises the possibility that massive accumulation of lipid in these cells may be a senescence-related, degenerative phenomenon. Similar degenerative changes have been previously described in pheochromocytomas his- tologically mimicking an adrenal cortical tumor.33 The clinical
settings of clear-cell PENs seem to be slightly different in terms of the functional status of the two age groups. Those clear-cell PENs associated with VHL are usually nonfunc- tional, occur at a younger age, and are more common in females.9 In contrast in this study, lipid-rich PENs were not mostly associated with VHL and occurred in older males, as seen in 6 of our 8 cases, which are males older than 55 years.
In summary, this study documents that PENs charac- terized by microvesicular, foamy cytoplasm is a distinctive variant that is prone to misdiagnosis because it often lacks many characteristic attributes of endocrine neoplasia, it closely mimics foamy macrophages or adrenal cortical cells, and it can also be easily mistaken for “clear-cell” tumors. Ultrastructural studies confirm that the vesicles that fill the cytoplasm and sometimes even indent the nucleus are indeed lipid, and thus justify the term “lipid-rich” for this variant. Lipid-rich PENs are typically nonfunctional (9 of 11), and unlike their clear-cell counterparts, they do not appear to have a strong association with VHL disease. Furthermore, they also lack the immuno- phenotypic markers of VHL-associated tumors. That lipid-rich PENs occur predominantly in elderly patients raises the possibility of a degenerative process. The cause, mechanisms, and biologic significance of excessive lipid accumulation in these tumors ought to be analyzed further.
REFERENCES
1. Adsay NV, Andea A, Basturk O, et al. Secondary tumors of the pancreas: an analysis of a surgical and autopsy database and review of the literature. Virchows Arch. 2004;444:527-535.
2. Adsay V, Logani S, Sarkar F, et al. Foamy gland pattern of pancreatic ductal adenocarcinoma: a deceptively benign-appearing variant. Am J Surg Pathol. 2000;24:493-504.
3. Barghorn A, Komminoth P, Bachmann D, et al. Deletion at 3p25.3-p23 is frequently encountered in endocrine pancreatic tumours and is associated with metastatic progression. J Pathol. 2001;194:451-458.
4. Buchler P, Reber HA, Buchler M, et al. Hypoxia-inducible factor 1 regulates vascular endothelial growth factor expression in human pan- creatic cancer. Pancreas. 2003;26:56-64.
5. Chetty R, Asa SL. Pancreatic endocrine tumors: an update. Adv Anat Pathol. 2004;11:202-210.
6. Chetty R, Kennedy M, Ezzat S, et al. Pancreatic endocrine pathology in von Hippel-Lindau disease: an expanding spectrum of lesions. Endocr Pathol. 2004;15:141-148.
7. Goldstein J, Benharroch D, Sion-Vardy N, et al. Solid cystic and papillary tumor of the pancreas with oncocytic differentiation. J Surg Oncol. 1994; 56:63-67.
8. Guarda LA, Silva EG, Ordonez NG, et al. Clear cell islet cell tumor. Am J Clin Pathol. 1983;79:512-517.
9. Hoang MP, Hruban RH, Albores-Saavedra J. Clear cell endocrine pancreatic tumor mimicking renal cell carcinoma: a distinctive neoplasm of von Hippel-Lindau disease. Am J Surg Pathol. 2001;25:602-609.
10. Hochwald SN, Zee S, Conlon KC, et al. Prognostic factors in pancreatic endocrine neoplasms: an analysis of 136 cases with a proposal for low- grade and intermediate-grade groups. J Clin Oncol. 2002;20:2633-2642.
11. Kim WY, Kaelin WG. Role of VHL gene mutation in human cancer. J Clin Oncol. 2004;22:4991-5004.
12. Klimstra DS, Adsay NV. Benign and malignant tumors of the pancreas. In: Odze RD, Goldblum JR, Crawford JM, eds. Surgical Pathology of the GI Tract, Liver, Biliary Tract and Pancreas. Philadelphia: Saunders, 2003: 699-731.
13. Kosmahl M, Wagner J, Peters K, et al. Serous cystic neoplasms of the pancreas: an immunohistochemical analysis revealing alpha-inhibin,
neuron-specific enolase, and MUC6 as new markers. Am J Surg Pathol. 2004;28:339-346.
14. Krieg M, Haas R, Brauch H, et al. Up-regulation of hypoxia-inducible factors HIF-1alpha and HIF-2alpha under normoxic conditions in renal carcinoma cells by von Hippel-Lindau tumor suppressor gene loss of function. Oncogene. 2000;19:5435-5443.
15. Lack EE. Pathology of the Pancreas, Gallbladder, Extrahepatic Biliary Tract and Ampullary Region. New York: Oxford University Press, 2003: 323-373.
16. Libutti SK, Choyke PL, Bartlett DL, et al. Pancreatic neuroendocrine tumors associated with von Hippel Lindau disease: diagnostic and management recommendations. Surgery. 1998;124:1153-1159.
17. Lott ST, Chandler DS, Curley SA, et al. High frequency loss of heterozygosity in von Hippel-Lindau (VHL)-associated and sporadic pancreatic islet cell tumors: evidence for a stepwise mechanism for malignant conversion in VHL tumorigenesis. Cancer Res. 2002;62: 1952-1955.
18. Luttges J, Vogel I, Menke M, et al. Clear cell carcinoma of the pancreas: an adenocarcinoma with ductal phenotype. Histopathology. 1998;32: 444-448.
19. Mohr VH, Vortmeyer AO, Zhuang Z, et al. Histopathology and molecular genetics of multiple cysts and microcystic (serous) adenomas of the pan- creas in von Hippel-Lindau patients. Am J Pathol. 2000;157:1615-1621.
20. Mount SL, Weaver DL, Taatjes DJ, et al. Von Hippel-Lindau disease presenting as pancreatic neuroendocrine tumour. Virchows Arch. 1995; 426:523-528.
21. Musso C, Paraf F, Petit B, et al. Pancreatic neuroendocrine tumors and von Hippel-Lindau disease. Ann Pathol. 2000;20:130-133.
22. Namiq AL, Delcore R, Damjanov I. Pathologic quiz case: clear cell tumor of the pancreas associated with hypercalcemia. Arch Pathol Lab Med. 2003;127:241-242.
23. Nunobe S, Fukushima N, Yachida S, et al. Clear cell endocrine tumor of the pancreas which is not associated with von Hippel-Lindau disease: report of a case. Surg Today. 2003;33:470-474.
24. Ordonez NG, Silva EG. Islet cell tumour with vacuolated lipid-rich cytoplasm: a new histological variant of islet cell tumour. Histopathology. 1997;31:157-160.
25. Pugh CW, Ratcliffe PJ. The von Hippel-Lindau tumor suppressor, hypoxia-inducible factor-1 (HIF-1) degradation, and cancer pathogenesis. Semin Cancer Biol. 2003;13:83-89.
26. Ray S, Lu Z, Rajendiran S. Clear cell ductal adenocarcinoma of pancreas: a case report and review of the literature. Arch Pathol Lab Med. 2004;128: 693-696.
27. Ritter JH, Mills SE, Gaffey MJ, et al. Clear cell tumors of the alimentary tract and abdominal cavity. Semin Diagn Pathol. 1997;14:213-219.
28. Sano T, Horiguchi H. Von Hippel-Lindau disease. Microsc Res Tech. 2003;60:159-164.
29. Sinkre PA, Murakata L, Rabin L, et al. Clear cell carcinoid tumor of the gallbladder: another distinctive manifestation of von Hippel-Lindau disease. Am J Surg Pathol. 2001;25:1334-1339.
30. Solcia E, Capella C, Kloppel G. Tumors of the pancreas. In: Armed Forces Institute Pathology: Atlas of Tumor Pathology, vol. 20. Washington, DC: American Registry of Pathology, 1997:53-64.
31. Thompson LD, Heffess CS. Renal cell carcinoma to the pancreas in surgical pathology material. Cancer. 2000;89:1076-1088.
32. Tomsova M, Pohnetalova D. Serous microcystic adenoma of the pancreas: cytologic and histologic features. Cesk Patol. 2003;39:163-167.
33. Unger PD, Cohen JM, Thung SN, et al. Lipid degeneration in a pheochromoyctoma histologically mimicking an adrenal cortical tumor. Arch Pathol Lab Med. 1990;114:892-894.
34. Wick MR, Ritter JH, Humphrey PA, et al. Clear cell neoplasms of the endocrine system and thymus. Semin Diagn Pathol. 1997;14:183-202.
35. Yao DX, Soslow RA, Hedvat CV, et al. Melan-A (A103) and inhibin expression in ovarian neoplasms. Appl Immunohistochem Mol Morphol. 2003;11:244-249.
36. Zamboni G, Pea M, Martignoni G, et al. Clear cell ‘sugar’ tumor of the pancreas: a novel member of the family of lesions characterized by the presence of perivascular epithelioid cells. Am J Surg Pathol. 1996;20: 722-730.