- Email: [email protected]
CD10 Immunostaining distinguishes atypical polypoid adenomyofibroma (atypical polypoid adenomyoma) from endometrial carcinoma invading the myometrium
Human Pathology (2008) 39, 1446–1453
www.elsevier.com/locate/humpath
Original contribution
CD10 Immunostaining distinguishes atypical polypoid adenomyofibroma (atypical polypoid adenomyoma) from endometrial carcinoma invading the myometrium Yoshihiro Ohishi MD, PhD a,⁎, Tsunehisa Kaku MD, PhD b , Hiroaki Kobayashi MD, PhD c , Shinichi Aishima MD, PhD a , Yoshihisa Umekita MD, PhD d , Norio Wake MD, PhD c , Masazumi Tsuneyoshi MD, PhD a a
Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan Division of Maternal Health, Department of Health Sciences, School of Medicine, Kyushu University, Fukuoka, Japan c Department of Obstetrics and Gynecology, Kyushu University Hospital, Fukuoka, Japan d Field of Oncology, Department of Tumor Pathology, Kagoshima University Graduate School of Medical and Dental Sciences, Kagoshima, Japan b
Received 1 November 2007; revised 12 February 2008; accepted 18 February 2008
Keywords: CD10; Atypical polypoid adenomyofibroma (atypical polypoid adenomyoma APA); Myoinvasive endometrial carcinoma; Diagnosis
Summary The major differential diagnostic problem presented by atypical polypoid adenomyofibroma (atypical polypoid adenomyoma) (APA), which usually affects young women, is the exclusion of welldifferentiated endometrial carcinoma invading the myometrium. This distinction, however, is of great clinical importance from the standpoint of treatment because reproductive conservation is feasible for patients with APA. Recently, CD10, known to be a marker of endometrial stromal cells, was reported to be also expressed in cells immediately surrounding the neoplastic glands invading the myometrium [Am J Surg Pathol 27 (2003) 786-789; Mod Pathol 16(1) (2003) 22-27]. However, CD10 expression in the myofibromatous component of APA has not been previously examined in the literature. We therefore decided to examine whether the CD10-immunostaining pattern in APA is different from that in myoinvasive carcinoma. Furthermore, we also attempted to obtain any histopathologic findings that may offer some insight regarding the histogenesis of APA. Seven cases of APA were immunostained for CD10 using curettage or polypectomy specimens, in addition to hysterectomy specimens in 1 case. Areas with more fibrotic rather than muscular stroma were focally observed in 4 cases. The pattern of staining was compared with hysterectomy specimens taken from 19 cases in which well- to moderately differentiated endometrioid adenocarcinoma had deeply invaded the myometrium (outer two thirds of the myometrium) but was not associated with adenomyosis. In 6 of 7 cases of APA, CD10 was never expressed in the myofibromatous stromal components. In 1 case of APA, the fascicles of fibrotic and muscular mesenchymal cells in the interglandular areas were focally and weakly positive for CD10. All 19 myoinvasive carcinomas expressed CD10 to some extent in cells immediately surrounding the neoplastic myoinvasive glands (fringe-like staining pattern). The proportion of the myoinvasive nests immediately surrounded by CD10-positive mesenchymal cells was as follows: mean, 74%; median, 80%; minimum, 5%; maximum, 100%. The fringe-like CD10-staining pattern was not observed in APA. Furthermore, we identified a gradual transformation from preexisting endometrial stromal cells (CD10 positive) into the
⁎ Corresponding author. Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan. E-mail address: [email protected] (Y. Ohishi). 0046-8177/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.humpath.2008.02.006
CD10 immunostaining distinguishes atypical polypoid adenomyofibroma
1447
typical myofibromatous stromal component (CD10 negative) of APA in 1 case. In conclusion, this study demonstrated differences in the CD10 immunoreactivity or immunostaining pattern between the stromal components of APA and myoinvasive endometrial carcinoma. This difference should lead to a more accurate diagnosis of APA (pseudo-myoinvasive lesion). Furthermore, the histogenesis of APA may perhaps be explained by “myofibromatous metaplasia” of the endometrial stromal cells. © 2008 Elsevier Inc. All rights reserved.
1. Introduction In 1981, for the first time, Mazur [1] introduced an uncommon focal, polypoid lesion of the uterus featuring the proliferation of irregular endometrial glands with squamous metaplasia embedded within a prominent cellular smoothmuscle stroma as “atypical polypoid adenomyoma.” Longacre et al [2] proposed the alternative term atypical polypoid adenomyofibroma as they demonstrated that the stroma in these lesions contained a mixture of smooth-muscle cells, fibrous tissue, and endometrial stromal cells. They also expanded the concept of APA to lesions which contain glands that are sufficiently complex to satisfy the criteria for well-differentiated adenocarcinoma (APA of low malignant potential) [2]. According to the clinicopathologic study evaluating a series of 55 patients with APA, the authors reported that 95% of the patients were premenopausal and almost half were nulliparas [2]. There is thus an understandably strong desire by many of these patients to preserve their reproductive capacity. The authors have concluded that reproductive conservation is feasible for patients with either APA or APA of low malignant potential [2]. Although the distinction of APA from myoinvasive endometrial carcinoma is important from the standpoint of treatment strategy (feasibility of reproductive conservation), it can at times be very difficult, especially for a pathologist who is not an expert in this field. However, previous immunohistochemical studies have demonstrated that the stromal components in APA and adenocarcinoma may show similar markers including smooth-muscle actin, desmin, estrogen receptor, progesterone receptor, and CD34, underscoring the possible limited utility of these stains in differential diagnosis [3]. CD10, a cell surface–neutral endopeptidase degrading various bioactive peptides, has been used as a diagnostic tool for precursor acute lymphoblastic leukemia [4,5], Burkitt lymphoma [6], and follicular lymphoma [7]. CD10 is also expressed in a variety of nonhematopoietic tumors [8]. In the field of gynecologic pathology, CD10 is well known as a diagnostic marker of neoplastic and nonneoplastic endometrial stromal cells [9]. Recently, in studies in which they attempted to distinguish between myoinvasive carcinoma and carcinoma involving adenomyosis by CD10 immunostaining, Srodon et al [10] and Nascimento et al [11] reported that most of myoinvasive endometrial carcinomas expressed CD10 in the cells immediately surrounding the myoinvasive neoplastic glands. We therefore thought that if such
characteristic CD10 staining in myoinvasive carcinoma is absent in the myofibromatous stromal component of APA, which can be readily misinterpreted as direct evidence of myoinvasive carcinoma, then it could be a useful marker to distinguish APA from true myoinvasive carcinoma. Hence, we decided to attempt to determine whether the CD10immunostaining pattern in the myofibromatous stromal component of APA is different from that in myoinvasive carcinoma. Furthermore, we also carefully looked for any areas where CD10 immunostaining could offer some insight with regard to the histogenesis of APA.
2. Materials and methods 2.1. Case selection: APA A total of 7 lesions featuring localized polypoid proliferation of irregular glands with various degrees of squamous or morular differentiation and cellular smooth muscle or hybrid smooth muscle/fibrous (myofibromatous) stroma were selected for this study (APAs: cases 1-7). The clinicopathologic features of the 7 cases are summarized in Table 1. As for case 2, both curettage and hysterectomy samples were available. In cases 1, 2, and 3, nearly all the Table 1
Clinicopathologic features of APA
Case Age Location Histology
Treatment
1
18
Corpus
APA
2
40
Corpus
APA
3 4
29 38
Corpus Corpus
APA b APA-LMP c
5 6
36 22
Corpus Cervix
APA-LMP APA
7
36
Corpus
APA-LMP
NED a (54 mo) Curettage and NED hysterectomy (13 mo) Curettage Unknown Curettage and NED hysterectomy (36 mo) Polypectomy Unknown Polypectomy NED (57 mo) Polypectomy Recurrent (20 mo) ⁎
a
Follow-up
Polypectomy
No evidence of disease. Atypical polypoid adenomyofibrom (atypical polypoid adenomyoma). c Low malignant potential. ⁎ The histopathologic diagnosis of the recurrent lesion is atypical hyperplasia with conspicuous squamous morule but without abundant myofibromatous stroma. b
1448 stromal components were composed of short interlacing fascicles of smooth muscle cells and a small amount of intercellular collagen fibers (muscular-rich stroma). In cases 4, 5, 6, and 7, areas of sclerotic or cellular fibrous tissue with interspersed muscle (fibrous-rich stroma) were focally observed in addition to the muscular-rich stroma seen in cases 1, 2, and 3. The proportion of fibrous-rich stroma was 20% in case 4, 10% in case 5, 30% in case 6, and 30% in case 7. Cases that contained markedly complex glands were designated as APA of low malignant potential (cases 4, 5, and 7). Complex glands in these 3 cases were observed within chunks of abundant fibromuscular stroma. Complex glands were characterized by closely spaced, branching glands with occasional prominent morular metaplasia separated by thin intersecting fascicles of fibromuscular stroma. None of the cases contained inflamed granulation tissue–like stroma, epithelial fragments which did not contain abundant fibromuscular stroma but which met the criteria for carcinoma, sarcomatous stromal components, a dense chronic inflammatory cell infiltrate affecting the gland-to-stroma interfaces, glandular epithelium with marked pleomorphism, prominent nucleoli, or abnormal mitotic figures. In case 2, the periphery (surface) of the polyp showed a small focus of preexisting endometrial stromal cells which gradually transform into the typical myofibromatous stromal component of APA.
2.2. Case selection: myoinvasive carcinoma Nineteen cases of well- to moderately differentiated endometrioid adenocarcinoma invading the outer two thirds of the myometrium were selected. Ages ranged from 35 to 89 (mean, 60; median, 55) years. Myoinvasion was not estimated at the cornual region. Cases of adenocarcinoma invading the superficial myometrium were excluded from this study to avoid those cases that presented interpretative problems due to tangential sectioning or focal involvement of an irregular, deep endomyometrial junction. Cases of adenocarcinoma associated with adenomyosis were also excluded from this study. The myometrial invasion was deemed to be present only when neoplastic glands were observed within the myometrium without interposed normal glands or normal endometrial stroma. In 13 of 19 cases, myoinvasive glands surrounded by desmoplastic/granulation tissue or dense inflammatory cell infiltrate, or both, were at least focally identified. In 12 of 19 cases, myoinvasive glands without stromal reaction were at least focally identified. In 6 of 19 cases, myoinvasive glands with stromal reaction and those without stromal reaction were both identified. To evaluate the undoubted myoinvasive lesions, all the sections were taken from hysterectomy specimens.
2.3. Immunohistochemistry All of the microscopic slides were reviewed, and one representative, formalin-fixed, paraffin-embedded tissue block
Y. Ohishi et al. was selected for immunoperoxidase staining in each case. Immunoperoxidase staining using monoclonal antibodies to CD10 (clone 56C6, Novocastra, Newcastle, UK) at a dilution of 1:100, desmin (clone D33, Dako, Glostrup, Denmark) at a dilution of 1:100, α-smooth muscle actin (SMA) (clone 1A4, Sigma, St Louis, MO) at a dilution of 1:5000, and CK903 (clone 34betaE12, Enzo Life Science, New York, NY) at a dilution of 1:50 was performed using the universal immunoperoxidase polymer method (ENVISION-kit, Dako) according to the manufacturer's instructions. For staining with anti-CD10 and anti-CK903 antibodies, microwave heating was processed for the purpose of antigen retrieval. When immunoreactive cells were specifically detected in cells immediately surrounding irregular glands, such a staining pattern was designated as “fringe-like staining pattern”.
3. Results 3.1. APA cases (n = 7) The myofibromatous mesenchymal cells in 6 of the 7 APAs were exclusively negative for CD10 (cases 1-5 and 7) (Fig. 1A,B), whereas the remaining one (case 6) showed very focal and weak CD10 positivity in the muscular and fibrotic fascicles of mesenchymal cells within the interglandular areas. In the 3 cases of APA of low malignant potential (cases 4, 5, and 7), myofibromatous mesenchymal cells surrounding markedly complex glands were exclusively negative for CD10. CD10-positive mesenchymal cells immediately surrounding the irregular glands (fringe-like staining pattern) were not observed. The morular epithelium within the APA lesions was diffusely and strongly positive for CD10 (Fig. 1A,B). Concomitant normal endometrial stromal cells within small fragments of normal endometrial mucosa separate from APA lesions were also positive for CD10. Sporadic columnar epithelium and luminal surfaces of irregular glands were also positive for CD10. The CD10positive morular epithelium was also positive for CK903. The myofibromatous mesenchymal cells in all 7 APAs were diffusely and strongly positive for SMA. The myofibromatous stromal component in 5 of the 7 cases was focally and weakly positive for desmin (cases 1 and 4-7), whereas that in the remaining 2 cases was negative for desmin (cases 2 and 3). Desmin staining was strongest in the broad intersecting bands of fibromuscular tissue and weaker in the more cellular, periglandular zones. In case 2, there was a gradual transformation of preexisting proliferative-phase endometrial stromal cells (CD10 positive, SMA negative, desmin negative) into typical myofibromatous stroma of APA (CD10 negative, SMA positive, desmin negative) (Fig. 2A-C). The immunoreactivity of CD10 decreased in proportion as the transformation progressed (Fig. 2A,B). The CD10-positive morular epithelium was also positive for CK903 (Fig. 2C). In case 2,
CD10 immunostaining distinguishes atypical polypoid adenomyofibroma
1449
Fig. 1 Histopathology and CD10 immunostaining of APA and myoinvasive carcinoma. A, APA (case 2, hysterectomy specimen) (H&E; x100). Irregular endometrial glands embedded in abundant myofibromatous stroma. The morular component is visible. B, APA (case 2, hysterectomy specimen) (CD10 immunostaining; x100). The myofibromatous component is exclusively negative for CD10. CD10-positive morular component is also visible. C, Myoinvasive carcinoma with stromal reaction (H&E; x100). Well-differentiated endometrioid adenocarcinoma invading the myometrium. Desmoplastic reaction is also visible. D, Myoinvasive carcinoma with stromal reaction (CD10 immunostaining; x100). The desmoplastic stromal cells immediately surrounding the invasive glands are strongly positive for CD10 (fringe-like staining pattern). E, Myoinvasive carcinoma without stromal reaction (H&E; x200). Invasive glands are situated within the preexisting stroma between fascicles of smooth muscle without stromal reaction. F, Myoinvasive carcinoma without stromal reaction (CD10 immunostaining; x200). Preexisting intermuscular mesenchymal cells immediately surrounding the invasive glands are strongly positive for CD10 (fringe-like staining pattern).
1450
Y. Ohishi et al.
Fig. 2 Interfaces between preexisting endometrial stromal cells and myofibromatous component of APA (case 2, curettage specimen; x200). A, Hematoxylin and eosin. Gradual transition of preexisting endometrial stromal cells (right side) to myofibromatous component (left side) is visible. B, CD10 immunostaining. The normal endometrial stromal cells and morular component are strongly positive for CD10. The immunoreactivity of CD10 in endometrial stromal cells gradually decreases as the transition to myofibromatous component progresses. C, CK903 immunostaining. CD10-positive morular component is positive for CK903, whereas the myofibromatous component is exclusively negative. D, SMA immunostaining. The myofibromatous component is strongly positive for SMA.
the immunohistochemical findings in curettage and hysterectomy specimens were concordant.
3.2. Myoinvasive carcinoma cases (n = 19) All 19 myoinvasive carcinomas expressed CD10 to some extent in the cells immediately surrounding the neoplastic myoinvasive glands (fringe-like staining pattern) (Fig. 1C-F). The proportion of the myoinvasive nests immediately surrounded by CD10-positive mesenchymal cells was as follows: mean, 74%; median, 80%; minimum, 5%; maximum, 100%. With regard to the areas of myoinvasive carcinoma with stromal reaction, the proportion of the myoinvasive nests immediately surrounded by CD10positive mesenchymal cells was as follows: mean, 64%; median, 70%; minimum, 5%; maximum, 100%. Review of the hematoxylin and eosin (H&E) slides indicated that the CD10-positive cells appeared to be nondescript and had been
interpreted as representing a desmoplastic response to the invasive carcinoma (Fig. 1C,D). Most of the myoinvasive glands and gland-to-stroma interfaces without CD10-positive mesenchymal cells were associated with varying degrees of lymphoid cell infiltrate. Occasional myoinvasive glands were heavily affected by dense lymphoid cell infiltrate, resulting in a jagged gland contour. With regard to the areas of myoinvasive carcinoma without stromal reaction, the proportion of the myoinvasive nests immediately surrounded by CD10-positive mesenchymal cells was as follows: mean, 84%; median, 85%; minimum, 50%; maximum, 100%. Review of the hematoxylin and eosin (H&E) slides indicated that the CD10-positive cells appeared to be preexisting intermuscular mesenchymal cells (Fig. 1E,F). The desmoplastic/granulation tissue stromal components surrounding carcinomatous glands were diffusely and intensely positive for SMA and comparable with that of normal myometrium. Preexisting intermuscular mesenchymal
CD10 immunostaining distinguishes atypical polypoid adenomyofibroma Table 2 Summary of the immunohistochemical findings of each of the components
CD10 CK903 SMA Desmin
Stroma of APA a
Stroma of Squamous Normal myoinvasive morule endometrial carcinoma stroma
− (6/7 cases) − + + (5/7 cases)
+ − + +
+ + − −
+ − − −
a Atypical polypoid adenomyofibroma (atypical polypoid adenomyoma).
cells surrounding carcinomatous glands were only focally positive for SMA. The desmoplastic/granulation tissue stromal components surrounding carcinomatous glands showed patchy and less intense desmin-positive cells in comparison with the normal myometrium. The pattern of desmin staining was most pronounced along the interglandular zones in a pattern of distribution that would be expected for bundles of residual myometrial smooth muscle. Preexisting intermuscular mesenchymal cells surrounding carcinomatous glands were negative for desmin. SMA staining was detected to some extent in the CD10positive cells immediately surrounding the neoplastic glands, whereas desmin staining was not detected in any of the CD10-positive cells immediately surrounding the neoplastic glands. The immunohistochemical findings of each of the components are summarized in Table 2.
4. Discussion The major differential diagnostic problem presented by APA is the exclusion of well-differentiated endometrial carcinoma invading the myometrium. This distinction, however, is of great clinical importance from the standpoint of treatment because reproductive conservation is feasible for patients with APA, which usually affects young women. In this study, we demonstrated a significant difference in CD10-immunostaining pattern in stromal components between APA and myoinvasive carcinoma; the myofibromatous stromal component of APA was completely negative for CD10 in most cases, whereas mesenchymal cells immediately surrounding the myoinvasive carcinoma were positive (fringe-like staining pattern), irrespective of there being any stromal reaction. An exceptional case of APA demonstrating patchy and weak CD10 immunostaining did not show the fringe-like CD10-immunostaining pattern that is characteristic for myoinvasive carcinoma. Although we should acknowledge that the 2 most important factors for a correct diagnosis are the recognition that myoinvasive carcinoma is rarely seen in biopsy/ curettage specimens and the careful examination of routine
1451
H&E slides [2], it can at times be very difficult to exclude myoinvasive carcinoma confidently when a pathologist other than a specialist encounters APA in an endometrial biopsy or curettage specimen. On such occasions, we believe that a complete lack of CD10 immunoreactivity or a lack of a fringe-like CD10 staining pattern in the myofibromatous stromal component may facilitate a correct diagnosis of APA to be made, thus preventing the patient from undergoing an unnecessary surgical procedure. However, we need to bear in mind the fact that cases of adenocarcinoma coexisting with APA have been reported, and this is an important topic [12-16]. The prognosis of most of these patients appeared to be favorable [13-16]. On the other hand, Longacre et al [2] have expanded the concept of APA to lesions that contain glands that are sufficiently complex to satisfy the criteria for carcinoma, and they have termed such lesions as APA of low malignant potential. The authors have also concluded that the prognosis of patients with APA of low malignant potential is favorable and that reproductive conservation is feasible [2]. However, Sugiyama et al [12] reported the case of a 50-year-old woman who had endometrioid carcinoma with a focal clear cell adenocarcinoma component arising from APA, and this case had an unfavorable prognosis. We feel that the terminology regarding cases of adenocarcinoma coexisting with APA is confusing, and a standardized terminology is required to permit the selection of an appropriate therapy. In addition, we would like to emphasize that “adenocarcinoma arising in an endometrial polyp with smooth muscle component,” a more common uterine lesion especially among elderly patients, needs to be carefully excluded when a diagnosis of “adenocarcinoma coexisting with APA” or “APA of low malignant potential” is made. We should also point out the fact that not all of the myoinvasive glands were immediately surrounded by CD10positive mesenchymal cells. Most of the myoinvasive glands and gland-to-stroma interfaces without CD10-positive mesenchymal cells were associated with varying degrees of lymphoid cell infiltrate. Occasional myoinvasive glands were heavily affected by dense lymphoid cell infiltrate, resulting in a jagged gland contour. It would seem that the CD10-positive mesenchymal cells responsible for the fringelike staining pattern had disappeared owing to the dense lymphoid cell infiltrate. When the irregular glands embedded within the abundant stromal component are affected by dense lymphoid cells, the lesion should be suspected of being invasive carcinoma, even though CD10-positive mesenchymal cells are absent, as such a finding is not a feature of APA. In this study, as we intended to clarify the differences between APA and undoubted myoinvasive carcinoma, tissue blocks of myoinvasive carcinoma for immunohistochemistry were taken from hysterectomy specimens. In the curettage specimens, it is very difficult to decide whether the carcinomatous glands embedded in the abundant stromal component represent true myoinvasion or intramucosal invasion. We also reviewed the intramucosal lesions of
1452 myoinvasive carcinomas of this study and found various degrees of the fringe-like CD10-immunostaining pattern. CD10-positive cells surrounding the glands appeared to represent both desmoplastic stromal cells and preexisting endometrial stromal cells in the intramucosal lesions. Thus, one should also bear in mind the fact that the fringe-like CD10-immunostaining pattern in the curettage specimen does not represent direct evidence of myoinvasion. The morular component of APA is strongly positive for CD10 as previously described in the literature [17], and this finding acts as an internal positive control for CD10 immunostaining. Strong CD10 immunoreactivity within the morular component with a complete lack of CD10 staining within the myofibromatous stromal component of the same specimen excludes a false-negative finding. On the other hand, one should also carefully interpret CD10-positive spindle cells as the morular component sometimes shows a prominent spindle morphology that may simulate stromal cells. CD10-positive, spindle-shaped morular epithelium should not be misinterpreted as CD10-positive stromal cells. On such occasions, CK903 immunostaining is useful, as the morular component is also positive for CK903, whereas the myofibromatous stromal component is negative. Although the precise nature of CD10-positive mesenchymal cells immediately surrounding myoinvasive carcinoma is unclear, it has been variously speculated in previous studies that these cells represent CD10-positive fibroblasts within the desmoplastic reaction, that they represent endometrial stroma accompanied by invasive carcinoma, that they have been introduced to differentiate along endometrial stromal lines, or that they show spurious expression of a CD10 epitope in another protein [10,11]. Lack or paucity of desmin expression in the CD10-positive cells in our study and previous studies [10,11] suggests that most of these CD10-positive cells do not originate from myometrial smooth muscle. Recently, it has been reported that CD10 expression is not specific for endometrial stroma and endometrial stromal tumors [18,19]. Furthermore, CD10-positive mesenchymal cells are also frequently observed within areas of the invasive growth front of invasive colorectal and breast carcinoma [20,21]. In this context, we think that CD10-positive mesenchymal cells appear to be unassociated with endometrial stromal differentiation and may represent invasive carcinoma-associated stroma, irrespective of the primary site. The immunohistochemical pattern of SMA and desmin in APA and myoinvasive carcinoma was consistent with previous reports [3,10,11,13,15,16]. We also failed to demonstrate significant differences in the SMA- and desmin-staining pattern between APA and myoinvasive carcinoma, consistent with a previous report [3]. The paucity of desmin immunoreactivity demonstrated in our study and previous studies [3,13,15,16] may be one of the characteristic features of myofibromatous mesenchymal cells in APA, and the nature of the myofibromatous stromal component in APA appears to be different from that of myometrium and
Y. Ohishi et al. uterine smooth muscle neoplasms that usually show diffuse and strong desmin staining. Contrary to the lack of CD10 positivity in the myofibromatous stromal component of APA, it has been recently reported that uterine smooth muscle neoplasms including cellular leiomyoma and leiomyosarcoma can also express varying degrees of CD10 positivity [22]. This also supports the hypothesis that the nature of myofibromatous stroma in APA is different from that of uterine smooth muscle neoplasms. The myofibromatous stromal component in APA also appears to differ from myofibroblastic tumor in that typical myofibroblasts with cytoplasmic basophilia and sharply tapered or pointed nuclei [23] are a minor component, in that a dense chronic inflammatory infiltrate is rarely observed, and in that typical smooth muscle cells with eosinophilic cytoplasm and blunt-ended nuclei are the major component. Of note, we identified a gradual transformation from preexisting proliferative-phase endometrial stromal cells into the typical myofibromatous stromal component of APA in case 2. The interface area between proliferative endometrial stromal cells (CD10 positive, SMA negative, desmin negative) and typical myofibromatous component (CD10 negative, SMA positive, desmin negative) was weakly positive for CD10 (Fig. 2A,B,D). Based on this finding, we speculated that the histogenesis of the myofibromatous stromal component of APA might be explained by “myofibromatous metaplasia” of the endometrial stromal cells. Decreased CD10 immunoreactivity of endometrial stromal cells in proportion to “myofibromatous metaplasia” may reflect the fact that “smooth muscle” or “myxofibrous” variants of endometrial stromal tumor showed decreased CD10 immunoreactivity [24]. We speculate that myofibromatous mesenchymal cells in most APAs lose CD10 immunoreactivity through a metaplastic process. In conclusion, a complete lack of CD10 expression or a lack of fringe-like CD10-staining pattern in the myofibromatous stroma of APA should lead to a more accurate diagnosis of APA. Furthermore, the histogenesis of the myofibromatous stromal component of APA could perhaps be explained by “myofibromatous metaplasia” of endometrial stromal cells. We hope that our preliminary conclusions will be substantiated in future studies using a large number of APA cases.
Acknowledgments We thank Dr Shogo Urabe (Department of Pathology, Oita Prefectural Hospital, Oita, Japan) for kindly providing the materials used in our research. We thank Miss Noriko Aoki and Miss Yoko Nozuka (Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan) and Miss Emiko Hori (Kyushu University Hospital, Fukuoka, Japan) for their excellent
CD10 immunostaining distinguishes atypical polypoid adenomyofibroma technical assistance. The English used in this manuscript was revised by Miss K. Miller (Royal English Language Centre, Fukuoka, Japan).
[12]
[13]
References [14] [1] Mazur MT. Atypical polypoid adenomyomas of the endometrium. Am J Surg Pathol 1981;5:473-82. [2] Longacre TA, Chung H, Rouse RV, Hendrickson MR. Atypical polypoid adenomyofibromas (atypical polypoid adenomyomas) of the uterus: a clinicopathologic study of 55 cases. Am J Surg Pathol 1996; 20:1-20. [3] Soslow RA, Chung MH, Rouse RV, Hendrickson MR, Longacre TA. Atypical polypoid adenomyofibroma (APA) versus well-differentiated endometrial carcinoma with prominent stromal matrix: an immunohistochemical study. Int J Gynecol Pathol 1996;15:209-16. [4] Greaves MF, Hariri G, Newman RA, Sutherland DR, Ritter MA, Ritz J. Selective expression of the common acute lymphoblastic leukemia (gp100) antigen on immature lymphoid cells and their malignant counterparts. Blood 1983;61:628-39. [5] Weiss LM, Bindl JM, Picozzi VJ, Link MP, Warnke RA. Lymphoblastic lymphoma: an immunophenotype study of 26 cases with comparison to T cell acute lymphoblastic leukemia. Blood 1986;67: 474-8. [6] Gregory CD, Tursz T, Edwards CF, Tetaud C, Talbot M, Caillou B, et al. Identification of a subset of normal B cells with a Burkitt's lymphoma (BL)-like phenotype. J Immunol 1987;139:313-8. [7] Stein H, Lennert K, Feller AC, Mason DY. Immunohistological analysis of human lymphoma: correlation of histological and immunological categories. Adv Cancer Res 1984;42:67-147. [8] Chu P, Arber DA. Paraffin-section detection of CD10 in 505 nonhematopoietic neoplasms. Frequent expression in renal cell carcinoma and endometrial stromal sarcoma. Am J Clin Pathol 2000;113:374-82. [9] McCluggage WG, Sumathi VP, Maxwell P. CD10 is a sensitive and diagnostically useful immunohistochemical marker of normal endometrial stromal and of endometrial stromal neoplasms. Histopathology 2001;39:273-8. [10] Srodon M, Klein WM, Kurman RJ. CD10 immunostaining does not distinguish endometrial carcinoma invading myometrium from carcinoma involving adenomyosis. Am J Surg Pathol 2003;27:786-9. [11] Nascimento AF, Hirsch MS, Cviko A, Quade BJ, Nucci MR. The role of CD10 staining in distinguishing invasive endometrial adenocarci-
[15] [16]
[17]
[18]
[19]
[20]
[21]
[22]
[23] [24]
1453
noma from adenocarcinoma involving adenomyosis. Mod Pathol 2003;16(1):22-7. Sugiyama T, Ohta S, Nishida T, Okura N, Tanabe K, Yakushiji M. Two cases of endometrial adenocarcinoma arising from atypical polypoid adenomyoma. Gynecol Oncol 1998;71:141-4. Lee KR. Atypical polypoid adenomyoma of the endometrium associated with adenomyomatosis and adenocarcinoma. Gynecol Oncol 1993;51:416-8. Young RH, Treger T, Scully RE. Atypical polypoid adenomyoma of the uterus. A report of 27 cases. Am J Clin Pathol 1986;86:139-45. Fukunaga M, Endo Y, Ushigome S, Ishikawa E. Atypical polypoid adenomyomas of the uterus. Histopathology 1995;27:35-42. Mittal KR, Peng XC, Wallach RC, Demopoulos RI. Coexistent atypical polypoid adenomyoma and endometrial adenocarcinoma. HUM PATHOL 1995;26:574-6. Chiarelli S, Buritica C, Litta P, Ciani S, Guarch R, Nogales FF. An immunohistochemical study of morules in endometrioid lesions of female genital tract: CD10 is a characteristic marker of morular metaplasia. Clin Cancer Res 2006;12:4251-6. Mikami Y, Hata S, Kiyokawa T, Manabe T. Expression of CD10 in malignant mullerian mixed tumors and adenosarcomas: an immunohistochemical study. Mod Pathol 2001;15(9):923-30. McCluggage WG, Oliva E, Herrington CS, McBride H, Young RH. CD10 and calretinin staining of endocervical glandular lesions, endocervical stroma and endometrioid adenocarcinomas of the uterine corpus: CD10 positivity is characteristic of, but not specific for, mesonephric lesions and is not specific for endometrial stroma. Histopathology 2003;43:144-50. Ogawa H, Iwaya K, Izumi M, et al. Expression of CD10 by stromal cells during colorectal tumor development. HUM PATHOL 2002;33: 806-11. Makretsov NA, Hayes M, Carter BA, Dabiri S, Gilks CB, Huntsman DG. Stromal CD10 expression in invasive breast carcinoma correlates with poor prognosis, estrogen receptor negativity, and high grade. Mod Pathol 2007;20:84-9. Oliva E, Young RH, Amin MB, Clement PB. An immunohistochemical analysis of endometrial stromal and smooth muscle tumors of the uterus: a study of 54 cases emphasizing the importance of using a panel because of overlap in immunoreactivity for individual antibodies. Am J Surg Pathol 2002;26(4):403-12. Meis-Kindblom JM, Enzinger FM. Fibrous tissue tumors. MosbyWolfe: Color Atlas of Soft Tissue Tumors; 1996. p. 1-44. Yilmaz A, Rush DS, Soslow RA. Endometrial stromal sarcomas with unusual histologic features: a report of 24 primary and metastatic tumors emphasizing fibroblastic and smooth muscle differentiation. Am J Surg Pathol 2002;26:1142-50.
www.elsevier.com/locate/humpath
Original contribution
CD10 Immunostaining distinguishes atypical polypoid adenomyofibroma (atypical polypoid adenomyoma) from endometrial carcinoma invading the myometrium Yoshihiro Ohishi MD, PhD a,⁎, Tsunehisa Kaku MD, PhD b , Hiroaki Kobayashi MD, PhD c , Shinichi Aishima MD, PhD a , Yoshihisa Umekita MD, PhD d , Norio Wake MD, PhD c , Masazumi Tsuneyoshi MD, PhD a a
Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan Division of Maternal Health, Department of Health Sciences, School of Medicine, Kyushu University, Fukuoka, Japan c Department of Obstetrics and Gynecology, Kyushu University Hospital, Fukuoka, Japan d Field of Oncology, Department of Tumor Pathology, Kagoshima University Graduate School of Medical and Dental Sciences, Kagoshima, Japan b
Received 1 November 2007; revised 12 February 2008; accepted 18 February 2008
Keywords: CD10; Atypical polypoid adenomyofibroma (atypical polypoid adenomyoma APA); Myoinvasive endometrial carcinoma; Diagnosis
Summary The major differential diagnostic problem presented by atypical polypoid adenomyofibroma (atypical polypoid adenomyoma) (APA), which usually affects young women, is the exclusion of welldifferentiated endometrial carcinoma invading the myometrium. This distinction, however, is of great clinical importance from the standpoint of treatment because reproductive conservation is feasible for patients with APA. Recently, CD10, known to be a marker of endometrial stromal cells, was reported to be also expressed in cells immediately surrounding the neoplastic glands invading the myometrium [Am J Surg Pathol 27 (2003) 786-789; Mod Pathol 16(1) (2003) 22-27]. However, CD10 expression in the myofibromatous component of APA has not been previously examined in the literature. We therefore decided to examine whether the CD10-immunostaining pattern in APA is different from that in myoinvasive carcinoma. Furthermore, we also attempted to obtain any histopathologic findings that may offer some insight regarding the histogenesis of APA. Seven cases of APA were immunostained for CD10 using curettage or polypectomy specimens, in addition to hysterectomy specimens in 1 case. Areas with more fibrotic rather than muscular stroma were focally observed in 4 cases. The pattern of staining was compared with hysterectomy specimens taken from 19 cases in which well- to moderately differentiated endometrioid adenocarcinoma had deeply invaded the myometrium (outer two thirds of the myometrium) but was not associated with adenomyosis. In 6 of 7 cases of APA, CD10 was never expressed in the myofibromatous stromal components. In 1 case of APA, the fascicles of fibrotic and muscular mesenchymal cells in the interglandular areas were focally and weakly positive for CD10. All 19 myoinvasive carcinomas expressed CD10 to some extent in cells immediately surrounding the neoplastic myoinvasive glands (fringe-like staining pattern). The proportion of the myoinvasive nests immediately surrounded by CD10-positive mesenchymal cells was as follows: mean, 74%; median, 80%; minimum, 5%; maximum, 100%. The fringe-like CD10-staining pattern was not observed in APA. Furthermore, we identified a gradual transformation from preexisting endometrial stromal cells (CD10 positive) into the
⁎ Corresponding author. Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan. E-mail address: [email protected] (Y. Ohishi). 0046-8177/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.humpath.2008.02.006
CD10 immunostaining distinguishes atypical polypoid adenomyofibroma
1447
typical myofibromatous stromal component (CD10 negative) of APA in 1 case. In conclusion, this study demonstrated differences in the CD10 immunoreactivity or immunostaining pattern between the stromal components of APA and myoinvasive endometrial carcinoma. This difference should lead to a more accurate diagnosis of APA (pseudo-myoinvasive lesion). Furthermore, the histogenesis of APA may perhaps be explained by “myofibromatous metaplasia” of the endometrial stromal cells. © 2008 Elsevier Inc. All rights reserved.
1. Introduction In 1981, for the first time, Mazur [1] introduced an uncommon focal, polypoid lesion of the uterus featuring the proliferation of irregular endometrial glands with squamous metaplasia embedded within a prominent cellular smoothmuscle stroma as “atypical polypoid adenomyoma.” Longacre et al [2] proposed the alternative term atypical polypoid adenomyofibroma as they demonstrated that the stroma in these lesions contained a mixture of smooth-muscle cells, fibrous tissue, and endometrial stromal cells. They also expanded the concept of APA to lesions which contain glands that are sufficiently complex to satisfy the criteria for well-differentiated adenocarcinoma (APA of low malignant potential) [2]. According to the clinicopathologic study evaluating a series of 55 patients with APA, the authors reported that 95% of the patients were premenopausal and almost half were nulliparas [2]. There is thus an understandably strong desire by many of these patients to preserve their reproductive capacity. The authors have concluded that reproductive conservation is feasible for patients with either APA or APA of low malignant potential [2]. Although the distinction of APA from myoinvasive endometrial carcinoma is important from the standpoint of treatment strategy (feasibility of reproductive conservation), it can at times be very difficult, especially for a pathologist who is not an expert in this field. However, previous immunohistochemical studies have demonstrated that the stromal components in APA and adenocarcinoma may show similar markers including smooth-muscle actin, desmin, estrogen receptor, progesterone receptor, and CD34, underscoring the possible limited utility of these stains in differential diagnosis [3]. CD10, a cell surface–neutral endopeptidase degrading various bioactive peptides, has been used as a diagnostic tool for precursor acute lymphoblastic leukemia [4,5], Burkitt lymphoma [6], and follicular lymphoma [7]. CD10 is also expressed in a variety of nonhematopoietic tumors [8]. In the field of gynecologic pathology, CD10 is well known as a diagnostic marker of neoplastic and nonneoplastic endometrial stromal cells [9]. Recently, in studies in which they attempted to distinguish between myoinvasive carcinoma and carcinoma involving adenomyosis by CD10 immunostaining, Srodon et al [10] and Nascimento et al [11] reported that most of myoinvasive endometrial carcinomas expressed CD10 in the cells immediately surrounding the myoinvasive neoplastic glands. We therefore thought that if such
characteristic CD10 staining in myoinvasive carcinoma is absent in the myofibromatous stromal component of APA, which can be readily misinterpreted as direct evidence of myoinvasive carcinoma, then it could be a useful marker to distinguish APA from true myoinvasive carcinoma. Hence, we decided to attempt to determine whether the CD10immunostaining pattern in the myofibromatous stromal component of APA is different from that in myoinvasive carcinoma. Furthermore, we also carefully looked for any areas where CD10 immunostaining could offer some insight with regard to the histogenesis of APA.
2. Materials and methods 2.1. Case selection: APA A total of 7 lesions featuring localized polypoid proliferation of irregular glands with various degrees of squamous or morular differentiation and cellular smooth muscle or hybrid smooth muscle/fibrous (myofibromatous) stroma were selected for this study (APAs: cases 1-7). The clinicopathologic features of the 7 cases are summarized in Table 1. As for case 2, both curettage and hysterectomy samples were available. In cases 1, 2, and 3, nearly all the Table 1
Clinicopathologic features of APA
Case Age Location Histology
Treatment
1
18
Corpus
APA
2
40
Corpus
APA
3 4
29 38
Corpus Corpus
APA b APA-LMP c
5 6
36 22
Corpus Cervix
APA-LMP APA
7
36
Corpus
APA-LMP
NED a (54 mo) Curettage and NED hysterectomy (13 mo) Curettage Unknown Curettage and NED hysterectomy (36 mo) Polypectomy Unknown Polypectomy NED (57 mo) Polypectomy Recurrent (20 mo) ⁎
a
Follow-up
Polypectomy
No evidence of disease. Atypical polypoid adenomyofibrom (atypical polypoid adenomyoma). c Low malignant potential. ⁎ The histopathologic diagnosis of the recurrent lesion is atypical hyperplasia with conspicuous squamous morule but without abundant myofibromatous stroma. b
1448 stromal components were composed of short interlacing fascicles of smooth muscle cells and a small amount of intercellular collagen fibers (muscular-rich stroma). In cases 4, 5, 6, and 7, areas of sclerotic or cellular fibrous tissue with interspersed muscle (fibrous-rich stroma) were focally observed in addition to the muscular-rich stroma seen in cases 1, 2, and 3. The proportion of fibrous-rich stroma was 20% in case 4, 10% in case 5, 30% in case 6, and 30% in case 7. Cases that contained markedly complex glands were designated as APA of low malignant potential (cases 4, 5, and 7). Complex glands in these 3 cases were observed within chunks of abundant fibromuscular stroma. Complex glands were characterized by closely spaced, branching glands with occasional prominent morular metaplasia separated by thin intersecting fascicles of fibromuscular stroma. None of the cases contained inflamed granulation tissue–like stroma, epithelial fragments which did not contain abundant fibromuscular stroma but which met the criteria for carcinoma, sarcomatous stromal components, a dense chronic inflammatory cell infiltrate affecting the gland-to-stroma interfaces, glandular epithelium with marked pleomorphism, prominent nucleoli, or abnormal mitotic figures. In case 2, the periphery (surface) of the polyp showed a small focus of preexisting endometrial stromal cells which gradually transform into the typical myofibromatous stromal component of APA.
2.2. Case selection: myoinvasive carcinoma Nineteen cases of well- to moderately differentiated endometrioid adenocarcinoma invading the outer two thirds of the myometrium were selected. Ages ranged from 35 to 89 (mean, 60; median, 55) years. Myoinvasion was not estimated at the cornual region. Cases of adenocarcinoma invading the superficial myometrium were excluded from this study to avoid those cases that presented interpretative problems due to tangential sectioning or focal involvement of an irregular, deep endomyometrial junction. Cases of adenocarcinoma associated with adenomyosis were also excluded from this study. The myometrial invasion was deemed to be present only when neoplastic glands were observed within the myometrium without interposed normal glands or normal endometrial stroma. In 13 of 19 cases, myoinvasive glands surrounded by desmoplastic/granulation tissue or dense inflammatory cell infiltrate, or both, were at least focally identified. In 12 of 19 cases, myoinvasive glands without stromal reaction were at least focally identified. In 6 of 19 cases, myoinvasive glands with stromal reaction and those without stromal reaction were both identified. To evaluate the undoubted myoinvasive lesions, all the sections were taken from hysterectomy specimens.
2.3. Immunohistochemistry All of the microscopic slides were reviewed, and one representative, formalin-fixed, paraffin-embedded tissue block
Y. Ohishi et al. was selected for immunoperoxidase staining in each case. Immunoperoxidase staining using monoclonal antibodies to CD10 (clone 56C6, Novocastra, Newcastle, UK) at a dilution of 1:100, desmin (clone D33, Dako, Glostrup, Denmark) at a dilution of 1:100, α-smooth muscle actin (SMA) (clone 1A4, Sigma, St Louis, MO) at a dilution of 1:5000, and CK903 (clone 34betaE12, Enzo Life Science, New York, NY) at a dilution of 1:50 was performed using the universal immunoperoxidase polymer method (ENVISION-kit, Dako) according to the manufacturer's instructions. For staining with anti-CD10 and anti-CK903 antibodies, microwave heating was processed for the purpose of antigen retrieval. When immunoreactive cells were specifically detected in cells immediately surrounding irregular glands, such a staining pattern was designated as “fringe-like staining pattern”.
3. Results 3.1. APA cases (n = 7) The myofibromatous mesenchymal cells in 6 of the 7 APAs were exclusively negative for CD10 (cases 1-5 and 7) (Fig. 1A,B), whereas the remaining one (case 6) showed very focal and weak CD10 positivity in the muscular and fibrotic fascicles of mesenchymal cells within the interglandular areas. In the 3 cases of APA of low malignant potential (cases 4, 5, and 7), myofibromatous mesenchymal cells surrounding markedly complex glands were exclusively negative for CD10. CD10-positive mesenchymal cells immediately surrounding the irregular glands (fringe-like staining pattern) were not observed. The morular epithelium within the APA lesions was diffusely and strongly positive for CD10 (Fig. 1A,B). Concomitant normal endometrial stromal cells within small fragments of normal endometrial mucosa separate from APA lesions were also positive for CD10. Sporadic columnar epithelium and luminal surfaces of irregular glands were also positive for CD10. The CD10positive morular epithelium was also positive for CK903. The myofibromatous mesenchymal cells in all 7 APAs were diffusely and strongly positive for SMA. The myofibromatous stromal component in 5 of the 7 cases was focally and weakly positive for desmin (cases 1 and 4-7), whereas that in the remaining 2 cases was negative for desmin (cases 2 and 3). Desmin staining was strongest in the broad intersecting bands of fibromuscular tissue and weaker in the more cellular, periglandular zones. In case 2, there was a gradual transformation of preexisting proliferative-phase endometrial stromal cells (CD10 positive, SMA negative, desmin negative) into typical myofibromatous stroma of APA (CD10 negative, SMA positive, desmin negative) (Fig. 2A-C). The immunoreactivity of CD10 decreased in proportion as the transformation progressed (Fig. 2A,B). The CD10-positive morular epithelium was also positive for CK903 (Fig. 2C). In case 2,
CD10 immunostaining distinguishes atypical polypoid adenomyofibroma
1449
Fig. 1 Histopathology and CD10 immunostaining of APA and myoinvasive carcinoma. A, APA (case 2, hysterectomy specimen) (H&E; x100). Irregular endometrial glands embedded in abundant myofibromatous stroma. The morular component is visible. B, APA (case 2, hysterectomy specimen) (CD10 immunostaining; x100). The myofibromatous component is exclusively negative for CD10. CD10-positive morular component is also visible. C, Myoinvasive carcinoma with stromal reaction (H&E; x100). Well-differentiated endometrioid adenocarcinoma invading the myometrium. Desmoplastic reaction is also visible. D, Myoinvasive carcinoma with stromal reaction (CD10 immunostaining; x100). The desmoplastic stromal cells immediately surrounding the invasive glands are strongly positive for CD10 (fringe-like staining pattern). E, Myoinvasive carcinoma without stromal reaction (H&E; x200). Invasive glands are situated within the preexisting stroma between fascicles of smooth muscle without stromal reaction. F, Myoinvasive carcinoma without stromal reaction (CD10 immunostaining; x200). Preexisting intermuscular mesenchymal cells immediately surrounding the invasive glands are strongly positive for CD10 (fringe-like staining pattern).
1450
Y. Ohishi et al.
Fig. 2 Interfaces between preexisting endometrial stromal cells and myofibromatous component of APA (case 2, curettage specimen; x200). A, Hematoxylin and eosin. Gradual transition of preexisting endometrial stromal cells (right side) to myofibromatous component (left side) is visible. B, CD10 immunostaining. The normal endometrial stromal cells and morular component are strongly positive for CD10. The immunoreactivity of CD10 in endometrial stromal cells gradually decreases as the transition to myofibromatous component progresses. C, CK903 immunostaining. CD10-positive morular component is positive for CK903, whereas the myofibromatous component is exclusively negative. D, SMA immunostaining. The myofibromatous component is strongly positive for SMA.
the immunohistochemical findings in curettage and hysterectomy specimens were concordant.
3.2. Myoinvasive carcinoma cases (n = 19) All 19 myoinvasive carcinomas expressed CD10 to some extent in the cells immediately surrounding the neoplastic myoinvasive glands (fringe-like staining pattern) (Fig. 1C-F). The proportion of the myoinvasive nests immediately surrounded by CD10-positive mesenchymal cells was as follows: mean, 74%; median, 80%; minimum, 5%; maximum, 100%. With regard to the areas of myoinvasive carcinoma with stromal reaction, the proportion of the myoinvasive nests immediately surrounded by CD10positive mesenchymal cells was as follows: mean, 64%; median, 70%; minimum, 5%; maximum, 100%. Review of the hematoxylin and eosin (H&E) slides indicated that the CD10-positive cells appeared to be nondescript and had been
interpreted as representing a desmoplastic response to the invasive carcinoma (Fig. 1C,D). Most of the myoinvasive glands and gland-to-stroma interfaces without CD10-positive mesenchymal cells were associated with varying degrees of lymphoid cell infiltrate. Occasional myoinvasive glands were heavily affected by dense lymphoid cell infiltrate, resulting in a jagged gland contour. With regard to the areas of myoinvasive carcinoma without stromal reaction, the proportion of the myoinvasive nests immediately surrounded by CD10-positive mesenchymal cells was as follows: mean, 84%; median, 85%; minimum, 50%; maximum, 100%. Review of the hematoxylin and eosin (H&E) slides indicated that the CD10-positive cells appeared to be preexisting intermuscular mesenchymal cells (Fig. 1E,F). The desmoplastic/granulation tissue stromal components surrounding carcinomatous glands were diffusely and intensely positive for SMA and comparable with that of normal myometrium. Preexisting intermuscular mesenchymal
CD10 immunostaining distinguishes atypical polypoid adenomyofibroma Table 2 Summary of the immunohistochemical findings of each of the components
CD10 CK903 SMA Desmin
Stroma of APA a
Stroma of Squamous Normal myoinvasive morule endometrial carcinoma stroma
− (6/7 cases) − + + (5/7 cases)
+ − + +
+ + − −
+ − − −
a Atypical polypoid adenomyofibroma (atypical polypoid adenomyoma).
cells surrounding carcinomatous glands were only focally positive for SMA. The desmoplastic/granulation tissue stromal components surrounding carcinomatous glands showed patchy and less intense desmin-positive cells in comparison with the normal myometrium. The pattern of desmin staining was most pronounced along the interglandular zones in a pattern of distribution that would be expected for bundles of residual myometrial smooth muscle. Preexisting intermuscular mesenchymal cells surrounding carcinomatous glands were negative for desmin. SMA staining was detected to some extent in the CD10positive cells immediately surrounding the neoplastic glands, whereas desmin staining was not detected in any of the CD10-positive cells immediately surrounding the neoplastic glands. The immunohistochemical findings of each of the components are summarized in Table 2.
4. Discussion The major differential diagnostic problem presented by APA is the exclusion of well-differentiated endometrial carcinoma invading the myometrium. This distinction, however, is of great clinical importance from the standpoint of treatment because reproductive conservation is feasible for patients with APA, which usually affects young women. In this study, we demonstrated a significant difference in CD10-immunostaining pattern in stromal components between APA and myoinvasive carcinoma; the myofibromatous stromal component of APA was completely negative for CD10 in most cases, whereas mesenchymal cells immediately surrounding the myoinvasive carcinoma were positive (fringe-like staining pattern), irrespective of there being any stromal reaction. An exceptional case of APA demonstrating patchy and weak CD10 immunostaining did not show the fringe-like CD10-immunostaining pattern that is characteristic for myoinvasive carcinoma. Although we should acknowledge that the 2 most important factors for a correct diagnosis are the recognition that myoinvasive carcinoma is rarely seen in biopsy/ curettage specimens and the careful examination of routine
1451
H&E slides [2], it can at times be very difficult to exclude myoinvasive carcinoma confidently when a pathologist other than a specialist encounters APA in an endometrial biopsy or curettage specimen. On such occasions, we believe that a complete lack of CD10 immunoreactivity or a lack of a fringe-like CD10 staining pattern in the myofibromatous stromal component may facilitate a correct diagnosis of APA to be made, thus preventing the patient from undergoing an unnecessary surgical procedure. However, we need to bear in mind the fact that cases of adenocarcinoma coexisting with APA have been reported, and this is an important topic [12-16]. The prognosis of most of these patients appeared to be favorable [13-16]. On the other hand, Longacre et al [2] have expanded the concept of APA to lesions that contain glands that are sufficiently complex to satisfy the criteria for carcinoma, and they have termed such lesions as APA of low malignant potential. The authors have also concluded that the prognosis of patients with APA of low malignant potential is favorable and that reproductive conservation is feasible [2]. However, Sugiyama et al [12] reported the case of a 50-year-old woman who had endometrioid carcinoma with a focal clear cell adenocarcinoma component arising from APA, and this case had an unfavorable prognosis. We feel that the terminology regarding cases of adenocarcinoma coexisting with APA is confusing, and a standardized terminology is required to permit the selection of an appropriate therapy. In addition, we would like to emphasize that “adenocarcinoma arising in an endometrial polyp with smooth muscle component,” a more common uterine lesion especially among elderly patients, needs to be carefully excluded when a diagnosis of “adenocarcinoma coexisting with APA” or “APA of low malignant potential” is made. We should also point out the fact that not all of the myoinvasive glands were immediately surrounded by CD10positive mesenchymal cells. Most of the myoinvasive glands and gland-to-stroma interfaces without CD10-positive mesenchymal cells were associated with varying degrees of lymphoid cell infiltrate. Occasional myoinvasive glands were heavily affected by dense lymphoid cell infiltrate, resulting in a jagged gland contour. It would seem that the CD10-positive mesenchymal cells responsible for the fringelike staining pattern had disappeared owing to the dense lymphoid cell infiltrate. When the irregular glands embedded within the abundant stromal component are affected by dense lymphoid cells, the lesion should be suspected of being invasive carcinoma, even though CD10-positive mesenchymal cells are absent, as such a finding is not a feature of APA. In this study, as we intended to clarify the differences between APA and undoubted myoinvasive carcinoma, tissue blocks of myoinvasive carcinoma for immunohistochemistry were taken from hysterectomy specimens. In the curettage specimens, it is very difficult to decide whether the carcinomatous glands embedded in the abundant stromal component represent true myoinvasion or intramucosal invasion. We also reviewed the intramucosal lesions of
1452 myoinvasive carcinomas of this study and found various degrees of the fringe-like CD10-immunostaining pattern. CD10-positive cells surrounding the glands appeared to represent both desmoplastic stromal cells and preexisting endometrial stromal cells in the intramucosal lesions. Thus, one should also bear in mind the fact that the fringe-like CD10-immunostaining pattern in the curettage specimen does not represent direct evidence of myoinvasion. The morular component of APA is strongly positive for CD10 as previously described in the literature [17], and this finding acts as an internal positive control for CD10 immunostaining. Strong CD10 immunoreactivity within the morular component with a complete lack of CD10 staining within the myofibromatous stromal component of the same specimen excludes a false-negative finding. On the other hand, one should also carefully interpret CD10-positive spindle cells as the morular component sometimes shows a prominent spindle morphology that may simulate stromal cells. CD10-positive, spindle-shaped morular epithelium should not be misinterpreted as CD10-positive stromal cells. On such occasions, CK903 immunostaining is useful, as the morular component is also positive for CK903, whereas the myofibromatous stromal component is negative. Although the precise nature of CD10-positive mesenchymal cells immediately surrounding myoinvasive carcinoma is unclear, it has been variously speculated in previous studies that these cells represent CD10-positive fibroblasts within the desmoplastic reaction, that they represent endometrial stroma accompanied by invasive carcinoma, that they have been introduced to differentiate along endometrial stromal lines, or that they show spurious expression of a CD10 epitope in another protein [10,11]. Lack or paucity of desmin expression in the CD10-positive cells in our study and previous studies [10,11] suggests that most of these CD10-positive cells do not originate from myometrial smooth muscle. Recently, it has been reported that CD10 expression is not specific for endometrial stroma and endometrial stromal tumors [18,19]. Furthermore, CD10-positive mesenchymal cells are also frequently observed within areas of the invasive growth front of invasive colorectal and breast carcinoma [20,21]. In this context, we think that CD10-positive mesenchymal cells appear to be unassociated with endometrial stromal differentiation and may represent invasive carcinoma-associated stroma, irrespective of the primary site. The immunohistochemical pattern of SMA and desmin in APA and myoinvasive carcinoma was consistent with previous reports [3,10,11,13,15,16]. We also failed to demonstrate significant differences in the SMA- and desmin-staining pattern between APA and myoinvasive carcinoma, consistent with a previous report [3]. The paucity of desmin immunoreactivity demonstrated in our study and previous studies [3,13,15,16] may be one of the characteristic features of myofibromatous mesenchymal cells in APA, and the nature of the myofibromatous stromal component in APA appears to be different from that of myometrium and
Y. Ohishi et al. uterine smooth muscle neoplasms that usually show diffuse and strong desmin staining. Contrary to the lack of CD10 positivity in the myofibromatous stromal component of APA, it has been recently reported that uterine smooth muscle neoplasms including cellular leiomyoma and leiomyosarcoma can also express varying degrees of CD10 positivity [22]. This also supports the hypothesis that the nature of myofibromatous stroma in APA is different from that of uterine smooth muscle neoplasms. The myofibromatous stromal component in APA also appears to differ from myofibroblastic tumor in that typical myofibroblasts with cytoplasmic basophilia and sharply tapered or pointed nuclei [23] are a minor component, in that a dense chronic inflammatory infiltrate is rarely observed, and in that typical smooth muscle cells with eosinophilic cytoplasm and blunt-ended nuclei are the major component. Of note, we identified a gradual transformation from preexisting proliferative-phase endometrial stromal cells into the typical myofibromatous stromal component of APA in case 2. The interface area between proliferative endometrial stromal cells (CD10 positive, SMA negative, desmin negative) and typical myofibromatous component (CD10 negative, SMA positive, desmin negative) was weakly positive for CD10 (Fig. 2A,B,D). Based on this finding, we speculated that the histogenesis of the myofibromatous stromal component of APA might be explained by “myofibromatous metaplasia” of the endometrial stromal cells. Decreased CD10 immunoreactivity of endometrial stromal cells in proportion to “myofibromatous metaplasia” may reflect the fact that “smooth muscle” or “myxofibrous” variants of endometrial stromal tumor showed decreased CD10 immunoreactivity [24]. We speculate that myofibromatous mesenchymal cells in most APAs lose CD10 immunoreactivity through a metaplastic process. In conclusion, a complete lack of CD10 expression or a lack of fringe-like CD10-staining pattern in the myofibromatous stroma of APA should lead to a more accurate diagnosis of APA. Furthermore, the histogenesis of the myofibromatous stromal component of APA could perhaps be explained by “myofibromatous metaplasia” of endometrial stromal cells. We hope that our preliminary conclusions will be substantiated in future studies using a large number of APA cases.
Acknowledgments We thank Dr Shogo Urabe (Department of Pathology, Oita Prefectural Hospital, Oita, Japan) for kindly providing the materials used in our research. We thank Miss Noriko Aoki and Miss Yoko Nozuka (Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan) and Miss Emiko Hori (Kyushu University Hospital, Fukuoka, Japan) for their excellent
CD10 immunostaining distinguishes atypical polypoid adenomyofibroma technical assistance. The English used in this manuscript was revised by Miss K. Miller (Royal English Language Centre, Fukuoka, Japan).
[12]
[13]
References [14] [1] Mazur MT. Atypical polypoid adenomyomas of the endometrium. Am J Surg Pathol 1981;5:473-82. [2] Longacre TA, Chung H, Rouse RV, Hendrickson MR. Atypical polypoid adenomyofibromas (atypical polypoid adenomyomas) of the uterus: a clinicopathologic study of 55 cases. Am J Surg Pathol 1996; 20:1-20. [3] Soslow RA, Chung MH, Rouse RV, Hendrickson MR, Longacre TA. Atypical polypoid adenomyofibroma (APA) versus well-differentiated endometrial carcinoma with prominent stromal matrix: an immunohistochemical study. Int J Gynecol Pathol 1996;15:209-16. [4] Greaves MF, Hariri G, Newman RA, Sutherland DR, Ritter MA, Ritz J. Selective expression of the common acute lymphoblastic leukemia (gp100) antigen on immature lymphoid cells and their malignant counterparts. Blood 1983;61:628-39. [5] Weiss LM, Bindl JM, Picozzi VJ, Link MP, Warnke RA. Lymphoblastic lymphoma: an immunophenotype study of 26 cases with comparison to T cell acute lymphoblastic leukemia. Blood 1986;67: 474-8. [6] Gregory CD, Tursz T, Edwards CF, Tetaud C, Talbot M, Caillou B, et al. Identification of a subset of normal B cells with a Burkitt's lymphoma (BL)-like phenotype. J Immunol 1987;139:313-8. [7] Stein H, Lennert K, Feller AC, Mason DY. Immunohistological analysis of human lymphoma: correlation of histological and immunological categories. Adv Cancer Res 1984;42:67-147. [8] Chu P, Arber DA. Paraffin-section detection of CD10 in 505 nonhematopoietic neoplasms. Frequent expression in renal cell carcinoma and endometrial stromal sarcoma. Am J Clin Pathol 2000;113:374-82. [9] McCluggage WG, Sumathi VP, Maxwell P. CD10 is a sensitive and diagnostically useful immunohistochemical marker of normal endometrial stromal and of endometrial stromal neoplasms. Histopathology 2001;39:273-8. [10] Srodon M, Klein WM, Kurman RJ. CD10 immunostaining does not distinguish endometrial carcinoma invading myometrium from carcinoma involving adenomyosis. Am J Surg Pathol 2003;27:786-9. [11] Nascimento AF, Hirsch MS, Cviko A, Quade BJ, Nucci MR. The role of CD10 staining in distinguishing invasive endometrial adenocarci-
[15] [16]
[17]
[18]
[19]
[20]
[21]
[22]
[23] [24]
1453
noma from adenocarcinoma involving adenomyosis. Mod Pathol 2003;16(1):22-7. Sugiyama T, Ohta S, Nishida T, Okura N, Tanabe K, Yakushiji M. Two cases of endometrial adenocarcinoma arising from atypical polypoid adenomyoma. Gynecol Oncol 1998;71:141-4. Lee KR. Atypical polypoid adenomyoma of the endometrium associated with adenomyomatosis and adenocarcinoma. Gynecol Oncol 1993;51:416-8. Young RH, Treger T, Scully RE. Atypical polypoid adenomyoma of the uterus. A report of 27 cases. Am J Clin Pathol 1986;86:139-45. Fukunaga M, Endo Y, Ushigome S, Ishikawa E. Atypical polypoid adenomyomas of the uterus. Histopathology 1995;27:35-42. Mittal KR, Peng XC, Wallach RC, Demopoulos RI. Coexistent atypical polypoid adenomyoma and endometrial adenocarcinoma. HUM PATHOL 1995;26:574-6. Chiarelli S, Buritica C, Litta P, Ciani S, Guarch R, Nogales FF. An immunohistochemical study of morules in endometrioid lesions of female genital tract: CD10 is a characteristic marker of morular metaplasia. Clin Cancer Res 2006;12:4251-6. Mikami Y, Hata S, Kiyokawa T, Manabe T. Expression of CD10 in malignant mullerian mixed tumors and adenosarcomas: an immunohistochemical study. Mod Pathol 2001;15(9):923-30. McCluggage WG, Oliva E, Herrington CS, McBride H, Young RH. CD10 and calretinin staining of endocervical glandular lesions, endocervical stroma and endometrioid adenocarcinomas of the uterine corpus: CD10 positivity is characteristic of, but not specific for, mesonephric lesions and is not specific for endometrial stroma. Histopathology 2003;43:144-50. Ogawa H, Iwaya K, Izumi M, et al. Expression of CD10 by stromal cells during colorectal tumor development. HUM PATHOL 2002;33: 806-11. Makretsov NA, Hayes M, Carter BA, Dabiri S, Gilks CB, Huntsman DG. Stromal CD10 expression in invasive breast carcinoma correlates with poor prognosis, estrogen receptor negativity, and high grade. Mod Pathol 2007;20:84-9. Oliva E, Young RH, Amin MB, Clement PB. An immunohistochemical analysis of endometrial stromal and smooth muscle tumors of the uterus: a study of 54 cases emphasizing the importance of using a panel because of overlap in immunoreactivity for individual antibodies. Am J Surg Pathol 2002;26(4):403-12. Meis-Kindblom JM, Enzinger FM. Fibrous tissue tumors. MosbyWolfe: Color Atlas of Soft Tissue Tumors; 1996. p. 1-44. Yilmaz A, Rush DS, Soslow RA. Endometrial stromal sarcomas with unusual histologic features: a report of 24 primary and metastatic tumors emphasizing fibroblastic and smooth muscle differentiation. Am J Surg Pathol 2002;26:1142-50.