—  SHORT COURSE #08  —

Mimics in Gastrointestinal Pathology

Case 4 - Desmoid Tumors and Other Gist Mimics

Alyssa Krasinskas, Jeffrey Goldsmith and Susan Abraham


Clinical History:
This 31-year-old woman presented with abdominal pain. By CT scan, she was found to have an 11 cm mass near the ileocecal valve.

Microscopic Features:
Grossly, there was a firm, white, circumscribed-appearing mass lesion that stretched the overlying ileal and colonic mucosa, which was otherwise normal. Sections showed a uniform low-power appearance, with spindled cells set in a stroma that was mostly collagenous. Areas of mildly myxoid stroma were also present focally. The spindled cells had wavy nuclei with tapered ends. There were occasional mitotic figures. By immunostaining, lesional cells were negative for both CD117 and CD34, with appropriate internal positive controls.

Discussion

Introduction:
In 1998, gastrointestinal stromal tumors (GISTs) were shown to harbor gain-of-function mutations in the KIT gene. [1] It is now known that ~80-85% of GISTs have KIT mutations, 5-10% have mutations of the platelet-derived growth factor receptor A (PDGFRA) gene, and the remaining 10% are wild-type. [2] In 2001, the use of imatinib mesylate (Gleevec®) to target activated KIT protein was reported to have a dramatic clinical benefit in a patient with metastatic GIST. [3] This potential to respond to targeted therapy has heightened the need to correctly diagnose these tumors and separate them from other soft tissue mimics. Not all GISTs are immunoreactive for KIT (CD117), not all tumors that express CD117 are GISTs, and there are a variety of mesenchymal tumors in the gastrointestinal tract or abdominal cavity that can overlap with GIST in their gross characteristics, histology, and immunophenotype (Table 1). This talk will review several of the lesions that most commonly cause confusion in the differential diagnosis of GIST.

Desmoid Tumors:
Desmoid tumors (mesenteric fibromatosis) are uncommon (only ~900 new cases occur each year in the U.S.), [4] but they can mimic GISTs because of their similar intra-abdominal location, size range, and gross features. In one study, 52% of desmoid tumors seen in consultation were initially diagnosed as GISTs by the referring pathologists. [5] Theoretically, the locations of the two tumors should allow for their distinction; desmoids arise in the mesentery or retroperitoneum, while GISTs arise in the bowel wall. However, the infiltrative border of desmoid tumors often results in their extension into the bowel wall, and some cases even show lamina propria involvement. A few desmoids even seem to arise in the muscularis propria. [6] Conversely, some GISTs are primary in the mesentery and omentum, or metastasize to these sites. [7]

The distinction between desmoid tumor and GIST is important for two reasons. First, ~10% of desmoid tumors arise in patients with familial adenomatous polyposis (FAP) and therefore screening for FAP should be considered in patients who develop desmoids. Second, the biologic behavior, treatment, and prognosis of these tumors are different. Unlike GISTs, desmoids do not metastasize but tend to recur and infiltrate within the abdominal cavity, potentially resulting in obstruction or even perforation of the bowel wall. [8] The tendency to recur after surgery is stronger in patients with FAP, where desmoid disease is the leading cause of death in those who have undergone prophylactic colectomy. [9]

The link between desmoids and FAP led to understanding the molecular abnormalities of these tumors, and also led to the elucidation of an immunophenotype that can be helpful in differentiating them from GIST. In FAP, desmoid tumors arise because of second-hit mutations of the adenomatous polyposis coli (APC) gene superimposed on germline APC mutations. Bi-allelic inactivation of APC results in cytoplasmic and nuclear accumulation of b-catenin protein, which in turn leads to increased cell proliferation rates. Some sporadic desmoids also arise through bi-allelic APC inactivation, [10] but more commonly APC is wild-type and they instead harbor activating mutations of the b -catenin gene that lead to b -catenin protein stabilization. [11] Regardless of whether APC or b -catenin is mutated, accumulation of b -catenin protein in neoplastic spindle cell nuclei can be demonstrated by immunohistochemistry in most (>90%) desmoids. b -catenin staining is negative in GISTs, sclerosing mesenteritis, and leiomyosarcomas and so can help to distinguish desmoid tumors from their mimics. [12, 13]

The use of CD117 immunostaining to differentiate desmoids from GISTs can be difficult. Various studies have reported CD117 staining in 0-75% of desmoids (Table 2), depending on the antibody used, its dilution, and antigen retrieval method. [6, 12, 14, 15, 16, 17, 18] Optimization of the immunostaining method has been shown to result in a very low level (only ~5%) of CD117 expression in desmoids. [17] Smooth muscle actin (SMA) and desmin are not helpful in distinguishing desmoids from GISTs. CD34, however, can be helpful since it is not expressed in desmoid tumors but is present 40%, 54%, 57%, 59%, 82%, and 94% of GISTs from the ileum/jejunum, duodenum, omentum/mesentery, colon, stomach, and anorectum, respectively. [7, 19, 20, 21, 22, 23]

Immunohistochemistry is not needed for the diagnosis of most desmoids, which can be diagnosed by recognition of several characteristic histologic features. Unlike GISTs, desmoids usually show histologically infiltrative borders even in cases where they appear grossly circumscribed, and they lack the necrosis or cystic degeneration that can be seen in GISTs. Desmoids are lesions of only low- to moderate-cellularity, composed of uniform spindle cells in a collagenous stroma; keloid-like collagen fibers may be present. They lack the high cellularity, nuclear atypia, atypical mitoses, skeinoid fibers, nuclear palisading, and epithelioid component of some GISTs. [5, 6] Their vascular pattern also differs. A typical low-power feature of desmoids is the presence of small muscular arteries and dilated, thin-walled veins with slight perivascular lymphocytes. Mitotic rate is not a helpful feature in distinguishing desmoids from GISTs. The mitotic index of desmoids is usually low (mean of 2-4 mitoses per 50 high power fields [HPFs]) but can reach 10-11 mitoses per 50 HPFs. [5, 6]

Unlike the situation in GISTs, size, mitotic index, and mucosal invasion in desmoids do not serve as prognostic features. Management is aimed at local control, since desmoids do not metastasize. Surgical resection has historically been the first-line approach to desmoid tumors but this might not be feasible in cases where a large portion of the small bowel is involved. Additionally, surgery itself is one of the risk factors for occurrence and recurrence of desmoids in patients with FAP, who can develop multiple tumors. [8] Because of the potential problems with surgery, and because clean margins are difficult to achieve with the infiltrative nature of these tumors, other approaches to management have been attempted. These include primary or adjuvant external beam radiation, cytotoxic and noncytotoxic chemotherapies, tamoxifen, and NSAIDs. [4] Even though desmoids do not have KIT mutations, treatment with imatinib was recently shown cause partial tumor regression or stabilization of disease in a minority of patients with advanced desmoid tumors. [24]

Smooth Muscle Tumors

Leiomyomas:
In the years before CD117 and CD34 immunostaining, many mesenchymal tumors called leiomyomas or cellular leiomyomas were actually low-grade GISTs. However, there are true leiomyomas of the colorectum (and rarely, small bowel) that differ from GISTs in their anatomical location, histology, immunophenotype, and biologic behavior. In fact, true leiomyomas are several times more common than colonic GISTs. [25] Colorectal leiomyomas do not tend to cause symptoms and nearly all are incidental findings. They can be discovered either during endoscopic screening for colorectal cancer, where they present as small polyps covered by an intact mucosa, or as incidental polyps in resection specimens. Their main importance is in correct classification so that they are not mistaken for GIST.

The largest study of colorectal leiomyomas is by Miettinen et al encompassing 88 tumors. [25] They range from 1 mm – 2.2 cm with a median diameter of 4 mm, and display a mild male predominance (2.4:1). They arise almost exclusively from the muscularis mucosae, which is in contrast to the muscularis propria origin of most GISTs. Therefore, leiomyomas are easy to biopsy and resect endoscopically. Histologically, leiomyomas contain plump smooth muscle bundles that merge with the muscularis mucosae. They are usually even less cellular than the adjoining muscularis mucosae. In distinction to GISTs, leiomyomas are always negative for CD117 and CD34 and are diffusely positive for muscle markers including SMA and desmin. Occasional leiomyomas demonstrate significant nuclear atypia akin to "symplastic" leiomyomas of the uterus. In Miettinen's series, 2 of 88 lesions showed cellular atypia, and one of these also had a single mitotic figure. Leiomyomas of the muscularis mucosae behave in a benign fashion, although follow-up to ensure complete endoscopic excision of "atypical" leiomyomas was recommended. [25]

True leiomyomas also commonly arise in the distal two-thirds of the esophagus and GE junction region. Leiomyomas are the most common intramural soft tissue tumor of the esophagus and outnumber esophageal GISTs by 2 to 3-fold. [26] Unlike colonic leiomyomas, esophageal leiomyomas usually arise in the muscularis propria and only rarely (~1%) in the muscularis mucosae. [27] They can lack the rounded contour of colorectal leiomyomas and instead can have lobulations, a serpiginous outline, or a horseshoe shape. [28] Esophageal leiomyomas can be minute ("seedling" leiomyomas of <1 mm – 1 cm are recognized in 22% – 46% of surgical resections for esophageal carcinomas [27, 29]) or giant (>10 cm). [30] Small leiomyomas are asymptomatic while larger tumors (>5 cm) can result in epigastric discomfort and dysphagia. [28]

The distinction between esophageal leiomyoma and GIST cannot be based on anatomical location or size, but instead requires evaluation of the morphology and immunophenotype. Grossly, leiomyomas have a firm consistency, white or tan color, and whorled cut surface that is similar to uterine leiomyomas. Histologically, they are low- to moderate cellularity with bland, pink spindle cells and only rare mitoses. 24% of esophageal GISTs are epithelioid rather than spindled [26] and they are more cellular than leiomyomas. Immunohistochemistry easily separates esophageal leiomyomas from low-grade GISTs. SMA and desmin are almost always globally positive in leiomyomas, but only 12.5% of esophageal GISTs show SMA positivity and only 19% are desmin positive. [26] Leiomyomas are always negative for CD34 and CD117. It is important not to over-interpret CD117 reactivity in mast cells or the rare leiomyoma that has a few entrapped CD117+ interstitial cells of Cajal. [26]

Esophageal leiomyomas are benign regardless of their size. In a 40-year review of esophageal smooth muscle tumors at Massachusetts General Hospital, there was no malignant degeneration nor any tendency toward recurrence. [28] In contrast, 53% of patients with esophageal GISTs in Miettinen's study died of disease, including all patients who had tumors larger than 10 cm. [26] When symptomatic, leiomyomas usually can be treated conservatively with enucleation. Open enucleation by thoracotomy has been the traditional approach, but more recently video-assisted thoracoscopic and laparoscopic approaches to enucleation are now low-morbidity alternatives. [31]

Rarely, intramural leiomyomas arise in the duodenum or anorectum. Most of these are asymptomatic but they can cause duodenal obstruction [20]or rectal bleeding. [23] Intramural leiomyomas of the duodenum and anorectum are similar grossly, microscopically, and immunophenotypically to esophageal leiomyomas, but have not been reported to reach the large sizes of some esophageal tumors. Sizes in the duodenum have ranged from 2 – 4 cm [20]and in the anorectum from 1.5 – 4 cm. [23]

Leiomyosarcomas:
GI leiomyosarcomas are much less common than either leiomyomas or GISTs. Only rare esophageal leiomyosarcomas have been described. Of 3 well-characterized esophageal leiomyosarcomas, all were large (9 – 16 cm) with mild nuclear pleomorphism and high mitotic rates (>50 mitoses/50 high power fields), and ultimately resulted in death. [26] One case report described a pedunculated intralumenal esophageal leiomyosarcoma arising from the muscularis mucosae, [32] but probably most esophageal leiomyosarcomas (like leiomyomas) originate in the muscularis propria.

In the remainder of the GI tract, leiomyosarcomas are generally larger than leiomyomas and they do not arise from the muscularis mucosae. Duodenal leiomyosarcomas range from 10 – 19 cm and most are transmural. [20] Colonic leiomyosarcomas range from 3 – 8 cm and tend to form polypoid intralumenal masses. [22] Leiomyosarcomas of the anorectum are smaller (2 – 5.5 cm) at presentation, and also tend to form polypoid intralumenal masses. [23] Ulceration of the overlying mucosa commonly occurs. Histologically, they are spindle cell tumors resembling smooth muscle. In contrast to leiomyomas and in contrast to many GISTs, the mitotic rate of gastrointestinal leiomyosarcomas is high, with >50 mitoses per 50 high power fields in most reported cases.

Leiomyosarcomas are uniformly positive for S MA, although occasionally the staining is focal. Desmin positivity is less consistent. Of 19 leiomyosarcomas from the esophagus, duodenum, colon, and anorectum studied by Miettinen and colleagues, only 12 (63%) were desmin positive and the labeling was often focal. [20, 22, 23, 26] All of Miettinen's cases were negative for CD117 and for CD34, with the exception of a single esophageal leiomyosarcoma which showed CD34 labeling in 40% of the tumor cells. All leiomyosarcomas studied for KIT mutations have been negative. [20, 22]

Gastrointestinal Schwannomas:
GI schwannomas are rare tumors that are often misdiagnosed as GISTs. Of 191 gastrointestinal mesenchymal tumors evaluated by Kwon and colleagues, only 12 (6.3%) were schwannomas. [33] Schwannomas are most common in the stomach, then colorectum, esophagus, and small intestine. Histologically they are distinctive. Unlike schwannomas in other sites, GI schwannomas typically lack well-defined nuclear palisading, Verocay bodies, xanthoma cells, and hyalinized vessel walls. Instead, the usual GI schwannoma is a moderately or highly cellular tumor containing interlacing bundles of spindle cells with wavy, tapered nuclei. "Benign nuclear atypia" is a characteristic finding consisting of nuclear variability and hyperchromasia. [34] Cell borders are often indistinct and the tapered nuclei appear to be interspersed among parallel collagen strands or in a myxoid stroma with inflammatory cells. Mitotic activity is rare, in most studies <1 mitosis per 50 high power fields [33, 34, 35] and never over 5 per 50 high power fields. [36]

Grossly, GI schwannomas can not reliably be distinguished from GISTs. Similar to GISTs, they are well-circumscribed but unencapsulated lesions and most involve the muscularis propria. Size is variable, from <1 cm up to 12 cm; among gastric schwannomas the mean size is 2.8 cm. [37] Schwannomas lack necrosis and they do not invade the lamina propria; however, mechanical ulceration of the overlying mucosa can occur. Cytologic features do not separate schwannomas from GISTs, especially since nuclear palisading is reported in one-third of gastric GISTs. [19] A low-power feature that should alert the pathologist to the presence of schwannoma is a peripheral lymphoid cuff, often containing germinal centers. This is a characteristic finding in GI schwannomas [33, 34, 35, 36, 37, 38] and is uncommon in GISTs. [19]

Immunostaining is necessary for an accurate distinction between schwannoma and GIST. GI schwannomas are strongly and diffusely S100 positive in both nuclei and cytoplasm. Positive staining for glial fibrillary acidic protein (GFAF) is seen in 63.6%, [35] 70%, [39] or 100% [33, 34, 36, 40] and can be focal. Schwannomas are negative for muscle markers, CD117, and CD34. In contrast, 91% of gastric GISTs stain for CD117, 82% for CD34, and <1% for S100. [19] Correctly diagnosing schwannoma is clinically important because these tumors are all benign. The behavior of GISTs, in contrast, is hard to predict and depends on tumor size, location, and mitotic activity. In the largest study of gastric GISTs, overall tumor-specific mortality was 17%. [19] GI schwannomas have no relationship to gastrointestinal autonomic nerve tumors (GANTs), a subtype of stromal tumors with distinctive ultrastructural features (neurosecretory granules, well-developed cell processes, and rudimentary cell junctions) but identical molecular characteristics and biologic behavior to GISTs. [2]

Patients with neurofibromatosis 2 (NF2) are at increased risk of CNS and soft tissue schwannomas, particularly bilateral vestibular schwannomas. However, none of the reported GI schwannomas has occurred in a patient with NF2. Lasota and colleagues studied GI schwannomas for NF2 mutations and loss of heterozygosity (LOH) at the NF2 locus on chromosome 22q. [39] Unlike conventional schwannomas – where inactivating NF2 mutations are commonly (40-60%) identified – GI schwannomas had no NF2 mutations and only a low rate (5%) of LOH at 22q. This underscores the differences (both morphologic and genetic) between GI schwannomas and conventional schwannomas.

Inflammatory Fibroid Polyps:
Inflammatory fibroid polyps (IFPs) are rare, benign tumors of unclear origin. They are most common in the stomach, particularly antrum and prepylorus, where they tend to form semipedunculated polyps covered either by normal mucosa or – when large – by a central erythematous depression, ulceration, or inflammatory cap. [41] IFPs also occur in the small bowel, particularly in the terminal ileum, where they can cause intussusception or obscure bleeding. Uncommonly, they arise in the colon, esophagus, or gallbladder. IFPs are usually based in the submucosa, but they often extend into the overlying mucosa and can occasionally be transmural. Microscopically, there is a proliferation of spindled and stellate stromal cells which tend to condense around blood vessels in whorled, perivascular cuffs. These stromal cells are embedded in a granulation tissue-type stroma or fibromyxoid stroma with numerous blood vessels and inflammatory cells. Prominent eosinophils are characteristic, but lymphocytes, lymphoid aggregates, plasma cells, macrophages, and mast cells are also present.

IFPs can be mistaken for GIST because of their spindle cell proliferation. Their abundant small vessels and inflammatory cells can be mistaken for nonspecific granulation tissue. In cases where morphology is not diagnostic, immunophenotyping can be helpful. Stromal cells of IFP are positive for CD34 [42, 43, 44, 45] – which is the most commonly used immunostain to confirm the diagnosis – as well as fascin, CD35, cyclin-D1, and calponin. [42] A minority stain for smooth muscle actin. [42, 44, 45] Positivity for CD34 can lead to confusion with GIST. However, IFPs are negative for CD117 [42, 45] although it is important not to over-interpret CD117+ mast cells.

Solitary Fibrous Tumors:
Solitary fibrous tumors (SFTs) are rare neoplasms that are derived from submesothelial cells. Although most common in the pleura, SFTs can occur in a wide variety of extrapleural locations including the mesentery and peritoneum. [46, 47] The fact that SFTs only rarely occur in the abdominal cavity can accentuate their tendency to be misdiagnosed as GIST, since pathologists can forget to include SFT in the differential diagnosis of a spindle cell tumor. Grossly, SFTs form circumscribed firm masses with pale, sometimes whorled, cut surfaces. The spindled cells are resemble fibroblasts growing in fascicular, storiform, or myxoid patterns. [46] Hyalinized collagen fibers are characteristic of SFTs. Features predicting malignancy are nuclear pleomorphism, >4 mitoses per 10 HPFs, and hemorrhage or necrosis. [48] Pleural and extrapleural SFTs are CD34 and CD99 positive, and negative for CD117. Immunostaining for CD99 does not help to distinguish SFTs from GISTs since 89% of GISTs are CD99 positive. [49] CD117 is the best marker to help make the distinction. However, relatively few immunophenotypic studies of SFTs have included abdominal cases. Miettinen and colleagues evaluated 5 abdominal SFTs and confirmed that all 5 were CD34+ and negative for CD117, SMA, desmin, and S100. [47]

Summary:
Remaining aware of the many different mesenchymal tumors of the GI tract and abdominal cavity is the first step in avoiding an over-diagnosis of GIST. Knowing the locations and histologic features of these different tumors is the second step, and in some cases (for example, leiomyomas of the muscularis mucosae, classic desmoid tumors, inflammatory fibroid polyps, or sclerosing mesenteritis) histology alone may be enough for a confident diagnosis. In many cases, immunophenotyping is needed as a third step to exclude/confirm GIST. Immunostaining only for CD117 is insufficient and a panel of immunostains is required (Table 3). Several studies have shown that a basic panel consisting of CD117, CD34, desmin, and S100 differentiates most mesenchymal tumors of the GI tract. [18, 47, 50] b -catenin can be added to the immunopanel when separation of desmoid from GIST is problematic. [18]

TABLE 1. Soft Tissue Lesions in the Differential Diagnosis of GIST
Lesion Type Usual Location
Discussed in handout
Leiomyoma (colorectal) Muscularis mucosae
Leiomyoma (esophagus/GEJ) Mural
Leiomyosarcoma Mural
Schwannoma Mural
Inflammatory fibroid polyp Submucosal
Desmoid Mesenteric
Solitary fibrous tumor Subserosal or peritoneal
Other
Benign fibroblastic polyp Mucosal
Ganglioneuroma Mucosal/rarely mural
Granular cell tumor Submucosal
Glomus tumor Submucosal
Neurofibroma Submucosal
Sarcoma (e.g., liposarcoma) Mesenteric
Sclerosing mesenteritis Mesenteric
Inflammatory myofibroblastic tumor Mesenteric/occasionally mural
Follicular dendritic cell tumor Mesenteric or mural

TABLE 2. CD117, CD34, Smooth Muscle Actin, and Desmin in Desmoid Tumors

Study CD117 (%) CD117 Antibody, Dilution, and Pretreatment CD34 (%) SMA (%) Desmin (%)
Miettinen [14] 0/14 (0%) Santa Cruz, 1:400, heat - - -
Leithner [15] 0/13 (0%) Dako, 1:1000, not specified - - -
Hornick [16] 1/20 (5%) Dako, 1:250, none - - -
Lucas [17] 1/19 (5.3%) Dako, 1:250, none - - -
Yamaguchi [18] 15/25 (60%) Dako, 1:50, heat 0/25 (0%) 21/25 (84%) 5/25 (20%)
Montgomery [12] 6/10 (60%) Dako, prediluted, steam 0/10 (0%) 5/9 (56%) 3/8 (37%)
Yantiss [6] 9/12 (75%) Dako, 1:30, heat 0/12 (0%) 9/12 (75%) -

TABLE 3. Immunohistochemistry for GIST and Other Mesenchymal Tumors of the GI Tract

Tumor CD117 CD34 Desmin S100 β-catenin
GIST* + (>90%) + (47-100%) - (<5%) - (1-8%) -
Leiomyoma - - + - -
Leiomyosarcoma - - + (70-100%) - -
Schwannoma - - - + -
Solitary fibrous tumor - + - - (0-24%) - (0-24%)
Desmoid - - +/- (19%) - + (90-100%)
Inflammatory fibroid polyp - + (>70%) - - N/A

*The staining pattern of GISTs depends upon the specific GI site. The numbers given above in parentheses for GIST and other mesenchymal tumors represent averages from refs 18, 47, and 50.

References:
  1. Hirota S, Isozaki K, Moriyama Y, et al. Gain-of-function mutations of c-kit in human gastrointestinal stromal tumors. Science. 1998;279:577-580.

  2. Rubin PB. Gastrointestinal stromal tumours: an update. Histopathology. 2006;48:83-96.

  3. Joensuu H, Roberts PJ, Sarlomo-Rikala M, et al. Effect of the tyrosine kinase inhibitor STI571 in a patient with a metastatic gastrointestinal stromal tumor. N Engl J Med. 2001;344:1052-1056.

  4. Hosalkar HS, Fox EJ, Delaney T, et al. Desmoid tumors and current status of management. Orthop Clin North Am. 2006;37:53-63.

  5. Rodriguez JA, Guarda LA, Rosai J. Mesenteric fibromatosis with involvement of the gastrointestinal tract. A GIST simulator: a study of 25 cases. Am J Clin Pathol. 2004;121:93-98.

  6. Yantiss RK, Spiro IJ, Compton CC, et al. Gastrointestinal stromal tumor versus intra-abdominal fibromatosis of the bowel wall. A clinically important differential diagnosis. Am J Surg Pathol. 2000;24:947-957.

  7. Miettinen M, Monihan JM, Sarlomo-Rikala M, et al. Gastrointestinal stromal tumors/smooth muscle tumors (GISTs) primary in the omentum and mesentery: clinicopathologic and immunohistochemical study of 26 cases. Am J Surg Pathol. 1999;23:1109-1118.

  8. Latchford AR, Sturt NJ, Neale K, et al. A 10-year review of surgery for desmoid disease associated with familial adenomatous polyposis. Br J Surg. 2006;93:1258-1264.

  9. Galiatsatos P, Foulkes WD. Familial adenomatous polyposis. Am J Gastroenterol. 2006;101:385-398.

  10. Alman BA, Li C, Pajerski ME, et al. Increased beta-catenin protein and somatic APC mutations in sporadic aggressive fibromatoses (desmoid tumors). Am J Pathol. 1997;151:329-334.

  11. Tejpar S, Nollet F, Li C, et al. Predominance of beta-catenin mutations and beta-catenin dysregulation in sporadic aggressive fibromatosis (desmoid tumor). Oncogene. 1999;18:6615-6620.

  12. Montgomery E, Torbenson MS, Kaushal M, et al. Beta-catenin immunohistochemistry separates mesenteric fibromatosis from gastrointestinal stromal tumor and sclerosing mesenteritis. Am J Surg Pathol. 2002;26:1296-1301.

  13. Bhattacharya B, Dilworth HP, Iacobuzio-Donahue C, et al. Nuclear beta-catenin expression distinguishes deep fibromatosis from other benign and malignant fibroblastic and myofibroblastic lesions. Am J Surg Pathol. 2005;29:653-659.

  14. Miettinen M. Are desmoid tumors KIT positive? Am J Surg Pathol. 2001;25:549-550.

  15. Leithner A, Gapp M, Radl, et al. Immunohistochemical analysis of desmoid tumors. J Clin Pathol. 2005;58:1152-1156.

  16. Hornick JL, Fletcher CD. Immunohistochemical staining for KIT (CD117) in soft tissue sarcomas is very limited in distribution. Am J Clin Pathol. 2002;117:188-193.

  17. Lucas DR, al-Abbadi M, Tabaczka P, et al. c-Kit expression in desmoid fibromatosis. Comparative immunohistochemical evaluation of two commercial antibodies. Am J Clin Pathol. 2003;119:339-345.

  18. Yamaguchi U, Hasegawa T, Masuda T, et al. Differential diagnosis of gastrointestinal stromal tumor and other spindle cell tumors in the gastrointestinal tract based on immunohistochemical analysis. Virchows Arch. 2004;445:142-150.

  19. Miettinen M, Sobin LH, Lasota JL. Gastrointestinal stromal tumors of the stomach. A clinicopathologic, immunohistochemical, and molecular genetic study of 1765 cases with long-term follow-up. Am J Surg Pathol. 2005;29:52-68.

  20. Miettinen M, Kopczynski J, Makhlouf H, et al. Gastrointestinal stromal tumors, intramural leiomyomas, and leiomyosarcomas in the duodenum. A clinicopathologic, immunohistochemical, and molecular genetic study of 167 cases. Am J Surg Pathol. 2003;27:625-641.

  21. Miettinen M, Kopczynski J, Sobin LH, et al. Gastrointestinal stromal tumors of the jejunum and ileum. A clinicopathologic, immunohistochemical, and molecular genetic study of 906 cases before imatinib with long-term follow-up. Am J Surg Pathol. 2006;30:477-489.

  22. Miettinen M, Sarlomo-Rikala M, Sobin LH, et al. Gastrointestinal stromal tumors and leiomyosarcomas in the colon. A clinicopathologic, immunohistochemical, and molecular genetic study of 44 cases. Am J Surg Pathol. 2000;24:1339-1352..

  23. Miettinen M, Furlong M, Sarlomo-Rikala M, et al. Gastrointestinal stromal tumors, intramural leiomyomas, and leiomyosarcomas in the rectum and anus. A clinicopathologic, immunohistochemical, and molecular genetic study of 144 cases. Am J Surg Pathol. 2001;25:1121-1133.

  24. Heinrich MC, McArthur GA, Demetri GD, et al. Clinical and molecular studies of the effect of imatinib on advanced aggressive fibromatosis (desmoid tumor). J Clin Oncol. 2006;24:1195-1203.

  25. Miettinen M, Sarlomo-Rikala M, Sobin LH. Mesenchymal tumors of muscularis mucosae of colon and rectum are benign leiomyomas that should be separated from gastrointestinal stromal tumors – a clinicopathologic and immunohistochemical study of eighty-eight cases. Mod Pathol. 2001;14:950-956.

  26. Miettinen M, Sarlomo-Rikala M, Sobin LH, et al. Esophageal stromal tumors. A clinicopathologic, immunohistochemical, and molecular genetic study of 17 cases and comparison with esophageal leiomyomas and leiomyosarcomas. Am J Surg Pathol. 2000;24:211-222.

  27. Abraham SC, Krasinskas AM, Hofstetter WL, et al. "Seedling" mesenchymal tumors (gastrointestinal stromal tumors and leiomyomas) are common incidental tumors of the esophagogastric junction. Am J Surg Pathol. (in press).

  28. Mutrie CJ, Donahue DM, Wain JC, et al. Esophageal leiomyoma: a 40-year experience. Ann Thorac Surg. 2005;79:1122-1125.

  29. Agaimy A, Wunsch PH. Sporadic Cajal cell hyperplasia is common in resection specimens for distal oesophageal carcinoma. A retrospective review of 77 consecutive surgical resection specimens. Virchows Arch. 2006;448:288-294.

  30. Cheng BC, Chang S, Mao ZF, et al. Surgical treatment of giant esophageal leiomyoma. World J Gastroenterol. 2005;11:4258-4260.

  31. Zaninotto G, Portale G, Costantini M, et al. Minimally invasive enucleation of esophageal leiomyoma. Surg Endosc. 2006; (Epub).

  32. Aiko S, Yoshizumi Y, Sugiura Y, et al. Pedunculated esophageal leiomyosarcoma: a case report. Dis Esophagus. 1998;11:263-267.

  33. Kwon MS, Lee SS, Ahn GH. Schwannomas of the gastrointestinal tract: clinicopathological features of 12 cases including a case of esophageal tumor compared with those of gastrointestinal stromal tumors and leiomyomas of the gastrointestinal tract. Pathol Res Pract. 2002;198:605-613.

  34. Prevot S, Bienvenu L, Vaillant JC, et al. Benign schwannoma of the digestive tract: a clinicopathologic and immunohistochemical study of five cases, including a case of esophageal tumor. Am J Surg Pathol. 1999;23:431-436.

  35. Hou YY, Tan YS, Xu JF, et al. Schwannoma of the gastrointestinal tract: a clinicopathological, immunohistochemical, and ultrastructural study of 33 cases. Histopathology. 2006;48:536-545.

  36. Miettinen M, Shekitka KM, Sobin LH. Schwannomas in the colon and rectum. A clinicopathologic and immunohistochemical study of 20 cases. Am J Surg Pathol. 2001;25:846-855.

  37. Daimaru Y, Kido H, Hashimoto H, et al. Benign schwannoma of the gastrointestinal tract: a clinicopathologic and immunohistochemical study. Hum Pathol. 1998;19:257-264.

  38. Arai T, Sugimura H, Suzuki M, et al. Benign schwannoma of the esophagus: report of two cases with immunohistochemical and ultrastructural studies. Pathol Int. 1994;44:460-465.

  39. Lasota J, Wasag B, Dansonka-Mieszkowska A, et al. Evaluation of NF2 and NF1 tumor suppressor genes in distinctive gastrointestinal nerve sheath tumors traditionally diagnosed as benign schwannomas: a study of 20 cases. Lab Invest. 2003;83:1361-1371.

  40. Sarlomo-Rikala M, Miettinen M. Gastric schwannoma – a clinicopathological analysis of six cases. Histopathology. 1995;27:355-360.

  41. Matsushita M, Hajiro K, Okazaki K, et al. Endoscopic features of gastric inflammatory fibroid polyps. Am J Gastroenterol. 1996;91:1595-1598.

  42. Pantanowitz L, Antonioli DA, Pinkus GS, et al. Inflammatory fibroid polyps of the gastrointestinal tract: evidence for a dendritic cell origin. Am J Surg Pathol. 2004;28:107-114.

  43. Wille P, Borchard F. Fibroid polyps of intestinal tract are inflammatory-reactive proliferations of CD34-positive perivascular cells. Histopathology. 1998;32:498-502.

  44. Hasegawa T, Yang P, Kagawa N, et al. CD34 expression by inflammatory fibroid polyps of the stomach. Mod Pathol. 1997;10:451-456.

  45. Ozolek JA, Sasatomi E, Swalsky PA, et al. Inflammatory fibroid polyps of the gastrointestinal tract: clinical, pathologic, and molecular characteristics. Appl Immunohistochem Mol Morphol. 2004;12:59-66.

  46. Levy AD, Rimola J, Mehrotra AK, et al. From the archives of the AFIP: benign fibrous tumors and tumorlike lesions of the mesentery: radiologic-pathologic correlation. Radiographics. 2006;26:245-264.

  47. Miettinen M, Sobin LH, Sarlomo-Rikala M. Immunohistochemical spectrum of GISTs at different sites and their differential diagnosis with a reference to CD117 (KIT). Mod Pathol. 2000;13:1134-1142.

  48. de Parrot M, Kurt AM, Robert JH, et al. Clinical behavior of solitary fibrous tumors of the pleura. Ann Thorac Surg. 1999;67:1456-1459.

  49. Shidham VB, Chivukula M, Gupta D, et al. Immunohistochemical comparison of gastrointestinal stromal tumor and solitary fibrous tumor. Arch Pathol Lab Med. 2002;126:1189-1192.

  50. Rossi G, Valli R, Bertolini F, et al. PDGFR expression in differential diagnosis between KIT-neagtive gastrointestinal stromal tumours and other primary soft-tissue tumours of the gastrointestinal tract. Histopathology. 2005;46:522-531.