Cole, Adam J. MD; Quick, Charles M. MD
Endometrial endometrioid adenocarcinoma (EEC) is the most common primary gynecologic malignancy in the United States, accounting for approximately 40,000 new cases each year.1 It typically arises in the setting of endometrial intraepithelial neoplasia or complex atypical hyperplasia, and has been linked to chronic estrogen exposure. Excess estrogen may result from ovarian dysfunction, estrogen-secreting tumors, nulliparity, unopposed estrogen (drugs), or obesity.1 A diagnosis of invasion in EEC can often easily be made in the presence of characteristic features including traditional desmoplastic myometrial invasion; gland crowding with loss of intervening myometrium; and complex epithelial papillary tufting with extensive necrosis.2,3 The 2010 FIGO staging system uses DOI and tumor spread to adjacent organs or tissues as staging criteria when evaluating uterine carcinoma, and thus the determination of DOI is of critical importance. The spectrum of morphologic patterns of invasion can often make the accurate assessment of DOI difficult.
Five different patterns of EEC myoinvasion have been described, including diffusely infiltrating (considered by many to be “traditional” irregular gland infiltration); broad front; adenoma malignum; adenomyosis (AM)-like; and microcystic, elongated, and fragmented (MELF) glands. Often, cases will contain more than 1 pattern of myoinvasion.4 Each of these patterns presents its own unique challenges when staging EEC cases. The goal of this review is to help the reader to better understanding the different patterns of invasion, and highlight the diagnostic pitfalls associated with each pattern.
INFILTRATING, IRREGULAR GLANDS
Infiltrating irregular glands (also known as the “single gland pattern”) characterizes what most pathologists conceive of as myometrial invasion. As described by Clement and Young,5 the invasive tumor takes the form of individually dispersed glands widely scattered throughout the myometrium. The glands usually occur singly or in groups of 3 or less, have irregular gland contours, and can be accompanied by a desmoplastic response, although this last feature is not required4 (Fig. 1). Assessing DOI is usually straightforward and is measured from the overlying endomyometrial junction to the deepest invasive gland.
This pattern of invasion is the most common type of invasion when referring to endometrioid endometrial adenocarcinoma. Ali et al3 noted that of 70 myoinvasive tumors, 62 (89%) showed a conventional destructive invasion pattern described as raggedly infiltrating malignant glands and a desmoplastic stromal reaction. Quick et al,4 in a large series of low-grade EEC, reported an incidence of 65% of this particular pattern of myoinvasion. It was also noted that this type of invasion was associated with higher stage, increased incidence of lymphovascular invasion (LVI) and a higher recurrence rate.4
“BROAD FRONT” OR “PUSHING”
Broad front, or pushing border, EEC has been previously defined as, “infiltration that was marked by a large swath of neoplastic glands that appear to push into the underlying myometrium with or without desmoplasia.”4 Quick and colleagues noted that of the 75 invasive cases of low-grade, low-stage invasive EEC that were examined, 16 (21%) were noted to have a broad front pattern (second only to diffusely infiltrating irregular glands, which comprised 65% of cases). A carcinoma that has overgrown the endometrium, but not invaded the myometrium, may accentuate an irregular endomyometrial junction, and thus cause confusion with the broad front pattern. Moderately to markedly irregular endomyometrial junctions are exceedingly common, occurring in up to 57% of hysterectomy specimens.3 In such cases, well circumscribed and rounded nests of tumor may bulge into the superficial myometrium, lending an appearance that is often misinterpreted as superficial myometrial invasion5 (Fig. 2). The presence of a linear tumor front, as opposed to an undulating contour, is a strong indicator of this type of pushing invasion (Fig. 3).
In cases of obvious broad front invasion, assessing the DOI should present little difficulty. There are, however, important exceptions, including a lack of uninvolved endomyometrial junction or “shoulder” on the slide; tumor involving AM; and the presence of an irregular endomyometrial junction. Ali et al3 noted that a desmoplastic response at the advancing edge is helpful when deciding if invasion is present. They further state that when evaluating DOI in a large, pushing tumor, the base of the adjacent uninvolved endomyometrial junction is used as the reference point from which to measure (ie, the previously referenced “shoulder”; Fig. 4). Without this shoulder, myometrial invasion may not be recognizable5; thus, when a pushing tumor is appreciated grossly, care should be taken to sample the adjacent normal endomyometrium, if present. Similar to AM-like invasion, the presence of endometrial stroma or uninvolved glands beneath the expansile front of tumor can be helpful in excluding invasion.
MICROCYSTIC, ELONGATED, AND FRAGMENTED GLANDS
The MELF pattern of invasion is characterized by neoplastic cells and glands that appear MELF. These glands are deceptive in their appearance and may be overlooked when assigning DOI.6 In their initial description, Murray and colleagues noted that this pattern was commonly associated with a fibro-myxoid stromal reaction with associated inflammatory cells (neutrophils and eosinophils). Frequently, the carcinomatous component contains individual tumor cells or tumor cell clusters. These tumor cells are often squamoid or eosinophilic (Figs. 5A, B). If present, glands are slit-like or microcystic (Fig. 5C). Although initially considered to be a rare pattern of invasion, the MELF pattern has been reported to vary from 7% to 16%.4,7 MELF is best appreciated at low power as a focal myxoid change within the myometrium, usually along the leading edge of the tumor. When these areas are noted, high power confirmation of eosinophilic tumor cells and fragmented or cystic glands admixed with inflammation confirms the presence of MELF. In tumors with multiple patterns of invasion, including MELF, the MELF component frequently represents the deepest extent of invasion; therefore, its recognition is essential to the correct measurement of DOI.
LVI has long been established as a prognostic indicator when stratifying patients with EEC, and tumors with the MELF pattern of myoinvasion have been identified as having an increased potential for LVI and vascular spread.7–9 McKenney and colleagues reported 3 cases of EEC with LVI that morphologically resembled histiocytes (Fig. 5D). They noted that all cases were low grade, had positive lymph nodes, and had at least focal areas of MELF. In each case, the individual neoplastic cells could be identified because of their relatively larger size, abundant eosinophilic cytoplasm, and intracytoplasmic vacuoles. Immunohistochemistry showed the cells to be positive for Ber-EP4 and cytokeratin, confirming the diagnosis. Their study described the following histologic features when assessing for possible LVI and vascular invasion in specimens with MELF:
* Retraction artifact typically has smooth, rounded contours, whereas true vascular invasion is more slit-like and angulated. Proximity to larger caliber blood vessels was more indicative of true LVI versus artifact.
* Intranodal mesothelial cells may resemble histiocyte-like tumor cells. They suggested comparing these cells morphologically to the cancer within the myometrium and using immunohistochemical positivity (Ber-EP4 in their study) to confirm diagnosis of metastatic tumor (Figs. 5E, F).
Subsequent studies have further described LVI and lymph node involvement in association with MELF. Pavlakis and colleagues examined 351 total abdominal hysterectomies with bilateral salpingo-oophorectomy (TAHBSO) specimens both with and without lymphadenectomy, performed for well or moderately differentiated EEC stages IA to II. They noted areas of MELF in 13% of cases with lymphadenectomy. Positive lymph nodes were identified in 7 of 13 MELF containing cases (54%) and in 6 of 86 of the MELF negative cases (7%).9 Stewart et al7 examined 170 cases of endometrial malignancy and described the presence of incidence of MELF in 27 (16%) cases. These cases were more likely to show at least focal mucinous differentiation and LVI. They concluded that MELF seems to be restricted to low-grade, myoinvasive EEC and is more common in tumors exhibiting focal mucinous differentiation.
Immunohistochemistry may be helpful when evaluating EEC cases with MELF. Several studies have highlighted the utility of using cytokeratins when assessing for DOI and/or lymph nodes. In their series of cases with histiocyte-like lymph node involvement, McKenney et al8 described more widespread nodal metastasis when utilizing Ber-EP4 staining. Additional publications have highlighted the utility of immunostaining as well.10–14 One such study compared 15 benign hysterectomy specimens to 26 hysterectomy specimens with EEC containing MELF. They noted that MELF was consistently and strongly positive for CK19. These results suggest CK19 as a useful marker in identifying and assessing DOI and subtle foci of LVI.12
The interesting possibility that MELF represents a type of epithelial-mesenchymal transition (EMT) has been proposed. Guarino and colleagues describe epithelial-mesenchymal transition as “the disruption of intercellular contacts and the enhancement of cell motility, thereby leading to the release of cells from the parent epithelial tissue. The resulting mesenchymal-like phenotype is suitable for migration and, thus, for tumor invasion and dissemination, allowing metastatic progression to proceed.”15 In their study they note that the down-regulation of E-cadherin is an important step in tumors gaining the ability to invade.15 Later studies analyzed a panel of immunohistochemical stains in EEC including: CK7, estrogen receptor/progesterone receptor, and E-cadherin.10 E-cadherin expression was markedly decreased in MELF, supporting the concept that MELF represents a form of EMT. In addition, it was noted that MELF frequently lost estrogen receptor and progesterone receptor expression, while displaying strong CK7 staining, which is more consistent with MELF representing an EMT as opposed to degenerative glandular changes.10 The epithelial-mesenchymal transition theory, backed by the immunohistochemical evidence helps to explain why this pattern of myoinvasion seems to be more vasotropic.
In an attempt to describe molecular aberrations in association with MELF, Stewart and colleagues examined 33 low-grade EEC cases with MELF for evidence of KRAS, BRAF, and MSI mutations. When compared to tumors with conventional invasion, MELF-containing tumors demonstrated more frequent KRAS mutations than MELF-negative tumors (45% vs. 30%), but the difference was not statistically significant. No BRAF mutations were identified in any of the cases. They concluded that KRAS and BRAF genes were not directly implicated in the development of MELF pattern of myoinvasion in EEC.16
ADENOMYOTIC-LIKE INVASION AND AM
AM is among the most common of gynecologic diseases, affecting anywhere from 10% to 70% of all hysterectomy specimens depending on the specific definition of AM that is used.17–19 AM may be seen in continuation with overlying endometrium through direct connection or a “tunnel,” resulting in an extremely irregular endomyometrial junction. The presence of AM can be problematic in the evaluation of DOI, and the presence of AM is also a common factor believed to contribute to interobserver variability.20
Coexistent AM and EEC in hysterectomy specimens occurs at rates ranging from 16% to 34%,21 and involvement of AM by EEC has been reported in up to 36% of hysterectomy specimens for EEC cases.20,21 Distinguishing between AM, invasive carcinoma (the so called adenomyotic-like invasion pattern), and noninvasive carcinoma involving AM can be quite difficult. In addition, patient-specific factors such as stromal atrophy in postmenopausal patients and exogenous hormone effect, can lead to the alteration or loss of endometrial stroma, further complicating this differential. Lack of nuclear atypia and mitotic figures, and rounded borders, are often indicators of AM.
Ismiil et al21 noted a dramatic increase in the rate of associated myoinvasion (91.3%) when EEC involves AM, compared to cases with AM without tumor involvement (64%). The reason for this increase remains unclear; however, it may be that AM simply represents an increase in the endomyometrial junction surface area, thus leading to more opportunity for invasion. In addition, it was noted that specimens with concurrent AM and EEC demonstrated a propensity for deeper invasion when compared to cases with no AM. Interestingly, they also found that lymphovascular involvement occurred at similar rates whether originating from a focus of AM involved by EEC, or from direct extension of tumor from the overlying native endometrium.
When confronted with a case in which AM is clearly involved by noninvasive carcinoma, these foci should not be used to determine DOI. The presence of benign endometrial glands, residual stroma, and smooth glandular contours are helpful histologic features that can exclude invasion (Fig. 6). In cases with AM involved by carcinoma, additional sections may be helpful in ruling out invasion, especially as these cases demonstrate an increased incidence of invasion.
AM-like invasion should not be confused with AM involved by EEC. AM-like invasion is characterized by a group of invasive EEC glands (usually 3 or more) that infiltrate the myometrium in irregular “islands.”4 The distinction between invasive AM-like EEC and AM involved by tumor can become increasingly difficult as the AM becomes distended with tumor, compressing (and obliterating) the remaining uninvolved endometrial glands and stroma. There are several histologic clues that one can use when trying to determine the nature of the glands in question:
* Is residual endometrial stroma or uninvolved glands present? If so, then this likely does not represent invasion and should not be used in determining DOI.
* Does the focus in question have smooth, rounded borders? If so, myoinvasion is less likely.
* Desmoplasia, if present, often includes associated inflammation and edema that can aid in the low-power identification of potential invasion.
* Are there adjacent foci of obvious AM or invasion? Often, the identification of these areas can be helpful in determining the nature of the glands in question (Fig. 7).
Attempts have been made to utilize immunostaining, namely CD10 staining of endometrial stroma, to aid in the distinction of AM-like invasion from AM involved by EEC. Nascimento et al22 evaluated CD10 for this purpose with mixed results. In their study, CD10 positivity was identified in every case of AM involved by noninvasive EEC; however, 52% of cases demonstrated positive CD10 staining surrounding areas of obvious invasion. The precise reason for the CD10 positivity remains unclear (epithelial-mesenchymal transition remains a possibility); however, regardless of the cause, care is advised when interpreting positive CD10 staining in AM involved by EEC.
Although it is not a pattern of invasion, serous carcinoma arising within AM is worthy of brief mention. Two series describing serous endometrial intraepithelial carcinoma arising in AM have been reported.23,24 Although seemingly rare, the incidence is likely underreported. Serous endometrial intraepithelial carcinoma is composed of glands (that can occur within AM) that are lined by highly atypical, cuboidal to columnar cells with abundant eosinophilic cytoplasm. As seen elsewhere, the nuclear features of serous in situ carcinoma are distinctive, consisting of round to oval nuclei with significant pleomorphism and prominent, nucleoli.
ADENOMA MALIGNUM (MINIMAL DEVIATION ENDOMETRIAL ADENOCARCINOMA)
Adenoma malignum adenocarcinoma (AMA) was first described by Gusserow25 in 1870 and is described as a mucinous type of extremely well differentiated adenocarcinoma involving the uterine cervix. This entity has since been extensively described involving the uterine corpus.1,26–31 Longacre and Hendrickson26 describe these glands as regular, round, and often widely spaced glands that invade the myometrium without any accompanying desmoplastic stromal or inflammatory response (Fig. 8). Because of the lack of desmoplasia and mass formation, these lesions often have no grossly identifiable features. These glands can invade widely, often involving the cervix and ovaries.27
AMA is the least common of the reported patterns of myometrial invasion. In a series of 75 cases, Mai et al30 reported no cases of this entity. Another series of 324 consecutive cases identified only a single case.4 In their series of 110 cases, Longacre and Hendrickson26 noted no increased risk of recurrence or mortality seen with AMA. Their results demonstrate that prognosis was driven by stage and histologic grade, as opposed to the presence of AMA. They proposed using the following histologic features to help diagnose AMA of the uterine corpus:
* The presence of focal cytologic atypia.
* Pseudostratification of glands.
* Focal loose, edematous stromal response (not desmoplasia).
* The presence of inflammatory cell infiltrate in and around problematic glands.
* Lymph vascular space invasion.
Kalyanasundaram et al27 reported 14 cases of AMA, some of which displayed aggressive features including: cervical stromal involvement (7 cases), bilateral ovarian involvement (2 cases), and serosal involvement (2 cases). Of the 14 cases, 12 were composed exclusively of EEC, 1 was mixed EEC and clear cell carcinoma, and 1 was exclusively serous carcinoma. Their report highlighted the more common differential diagnostic mimics of AMA, including: AM, deeply endocervical glands, mesonephric remnants, cervical tubo-endometrioid metaplasia, endocervical adenocarcinoma in situ, endocervical adenocarcinoma, and adenoma malignum of the cervix. In addition, these reports indicate another potential pitfall, which is that when well differentiated malignant glands are present in the endocervical stroma, an endometrial primary should be ruled out. This pattern of cervical involvement, described as “burrowing” by Clement and Young,5 may mimic an endocervical primary, and lead to misdiagnosis. Indeed, this entity should be considered whenever deep-seated, deceptively bland glands are encountered.
We have described the 5 distinct invasive patterns that may be seen in the context of endometrioid endometrial adenocarcinoma of the uterine corpus. In addition, associations with adverse outcome or difficulty in diagnosis and staging have been highlighted.
* The most common invasive pattern, diffusely infiltrating, is also associated with the worst prognosis.
* Broad pushing tumor fronts may appear noninvasive, especially in cases with no adjacent normal endometrium for basing depth assessment.
* MELF is easily overlooked and has a propensity to be associated with histiocyte-like LVI and lymph node involvement.
* Am-like invasion is difficult to discern from carcinoma involved AM and can affect staging, using DOI measurements, especially when these foci are located in the outer half of the myometrium.
* Adenoma malignum is exceedingly rare; however, it is also deceptively bland and may be overlooked.
Identification of each pattern promotes understanding of the unique challenges associated with each, especially with regard to assessing for DOI, extent of tumor spread, LVI, and nodal involvement.
1. Crum CP, Nucci MR, Lee KRCrum CP, Nucci MR, Lee KR. Adenocarcinoma, carcinosarcoma, and other epithelial tumors of the endometrium. Diagnostic Gynecologic and Obstetric Pathology. Elsevier Science Health Science Health Division. 2010;19:517–565
2. Brown L. Pathology of uterine malignancies. Clin Oncol. 2008;20:433–447
3. Ali A, Black D, Soslow RA. Difficulties in assessing the depth of myometrial invasion in endometrial carcinoma. Int J Gynecol Pathol. 2007;26:115–123
4. Quick CM, May T, Nucci MR, et al. Low grade, low stage endometrial adenocarcinoma: a clinicopathologic analysis of 324 cases focusing on frequency and pattern of myoinvasion. Int J Gynecol Pathol. 2012;31:337–343
5. Clement PB, Young RH. Endometrioid carcinoma of the uterine corpus: a review of its pathology with emphasis on recent advances and problematic aspects. Adv Anat Pathol. 2002;9:145–184
6. Murray SK, Young RH, Scully RE. Unusual epithelial and stromal changes in myoinvasive endometrioid adenocarcinoma: a study of their frequency associated diagnostic problems, and prognostic significance. Int J Gynecol Pathol. 2003;22:324–333
7. Stewart CJR, Brennan BA, Leung YC, et al. MELF pattern invasion in endometrial carcinoma: association with low grade, myoinvasive endometrioid tumors, focal mucinous differentiation and vascular invasion. Pathology. 2009;41:454–459
8. McKenney JK, Kong CS, Longacre TA. Endometrial adenocarcinoma associated with subtle lymph-vascular space invasion and lymph node metastasis: a histologic pattern mimicking intravascular and sinusoidal histiocytes. Int J Gynecol Pathol. 2004;24:73–78
9. Pavlakis K, Messini I, Vrekoussis T, et al. MELF invasion in endometrial cancer as a risk factor for lymph node metastasis. Histopathology. 2011;58:966–973
10. Stewart CJR, Little L. Immunophenotypic features of MELF pattern invasion in endometrial adenocarcinoma: evidence for epithelial-mesenchymal transition. Histopathology. 2009;55:91–101
11. Stewart CJR, Crook ML. Galactin-3 expression in uterine endometrioid adenocarcinoma: comparison of staining in conventional tumor glands and in areas of MELF pattern myometrial invasion. Int J Gynecol Pathol. 2010;29:555–561
12. Stewart CJR, Crook ML, Lacey J, et al. Cytokeratin 19 expression in normal endometrium and in low-grade endometrioid adenocarcinoma of the endometrium. Int J Gynecol Pathol. 2011;30:484–491
13. Stewart CJR, Crook ML, Doherty DA. Micro-anatomical variation in cellular proliferation in endometrial adenocarcinoma, and inverse correlation between Ki67 and cytokeratin 7 expression. Histopathology. 2010;57:46–54
14. Stewart CJR, Crook ML. CD147 (EMMPRIN) and matrix metalloproteinase-2 expression in uterine endometrioid adenocarcinoma. Pathol Res Pract. 2011;1:30–36
15. Guarino M, Rubinoa B, Ballabio G. The role of epithelial-mesenchymal transition in cancer pathology. Pathology. 2007;39:305–318
16. Stewart CJS, Amanuel B, Grieu F, et al. KRAS mutation and microsatellite instability in endometrial adenocarcinomas showing MELF-type myometrial invasion. J Clin Pathol. 2010;63:604–608
17. Seidman JD, Kjerulff KH. Pathologic findings from the Maryland Women’s Health Study: practice patterns in the diagnosis of adenomyosis. Int J Gynecol Pathol. 1996;15:217–221
18. Ferenczy A. Pathophysiology of adenomyosis. Hum Reprod. 1998;4:312–322
19. Bergholt T, Eriksen L, Jacobsen M, et al. Prevalence and risk factors of adenomyosis at hysterectomy. Hum Reprod. 2001;16:2418–2421
20. Hanley KZ, Dustin SM, Stoler MH, et al. The significance of tumor involved adenomyosis in otherwise low-stage endometriod adenocarcinoma. Int J Gynecol Pathol. 2010;29:445–451
21. Ismiil N, Rasty G, Ghorab Z, et al. Adenomyosis involved by endometrial adenocarcinoma is a significant risk factor for deep myometrial invasion. Ann Diagn Pathol. 2007;11:252–257
22. Nascimento AF, Hirsch MS, Cviko A, et al. The role of CD10 staining in distinguishing invasive endometrial adenocarcinoma from adenocarcinoma involving adenomyosis. Mod Pathol. 2003;16:22–27
23. Abushahin N, Zhang T, Chiang S, et al. Serous endometrial intraepithelial carcinoma arising in adenomyosis: a report of 5 cases. Int J Gynecol Pathol. 2011;30:271–281
24. Koshiyama M, Suzuki A, Ozawa M, et al. Adenocarcinomas arising from uterine adenomyosis: a report of four cases. Int J Gynecol Pathol. 2002;31:239–245
25. Gusserow ALS. Ueber sarcoma des uterus. Arch Gynakol. 1870;1:240–251
26. Longacre TA, Hendrickson MR. Diffusely infiltrative endometrial adenocarcinoma: an adenoma malignum pattern. Am J Surg Pathol. 1999;23:69–78
27. Kalyanasundaram K, Ganesan R, Perunovic B, et al. Diffusely infiltrating endometrial carcinomas with no stromal response: report of a series, including cases with cervical and ovarian involvement and emphasis on the potential for misdiagnosis. Int J Surg Pathol. 2010;18:138–143
28. Fujiwara M, Longacre TA. Low-grade mucinous adenocarcinoma of the uterine corpus: a rare and deceptively bland form of endometrial carcinoma. Am J Surg Pathol. 2011;35:537–544
29. Landry D, Mai KT, Senterman MK, et al. Endometroid adenocarcinoma of the uterus with a minimal deviation invasive pattern. Histopathology. 2003;42:77–82
30. Mai KT, Perkins DG, Yazdi HM, et al. Endometroid carcinoma of the endometrium with an invasive component of minimal deviation carcinoma. Hum Pathol. 2002;33:856–858
31. Ioffe OB. Recent developments and selected diagnostic problems in carcinomas of the endometrium. Am J Clin Pathol. 2005;124(suppl 1):S42–S51
© 2013 Lippincott Williams & Wilkins, Inc.