A Clinicopathologic Analysis of Atypical Proliferative (Borderline) Tumors and Well-Differentiated Endometrioid Adenocarcinomas of the Ovary

Bell, Karen A. M.D.; Kurman, Robert J. M.D.

American Journal of Surgical Pathology:
Original Articles

Atypical proliferative (borderline) endometrioid tumors (APTs) and well-differentiated endometrioid carcinomas of the ovary constitute a spectrum of morphologically diverse proliferative tumors. There is currently no agreement on the criteria for distinguishing them. We report the clinicopathologic features of 56 proliferative endometrioid tumors focusing on the criteria for invasion, the clinical significance of microinvasion and cytologic atypia, and prognosis. Endometriomas, adenofibromas, adenosarcomas and moderately to poorly differentiated carcinomas were excluded, as were patients with concurrent endometrioid carcinoma of the endometrium. The tumors were classified as atypical proliferative tumor (APT) (33 tumors), APT with intraepithelial carcinoma (high-grade cytology in a tumor lacking stromal invasion) (three tumors), APT with microinvasion (invasion [lt]5 mm) (five tumors), and invasive carcinoma (invasion [ge]5 mm) (15 tumors). All tumors were confined to the ovary (stage I). In 50 patients, the tumor involved one ovary, and in three patients, the tumors were bilateral. The predominant growth pattern was adenofibromatous in 29 tumors and glandular or papillary in 27 tumors. In 8 (24[percnt]) of 41 APTs, areas of benign adenofibroma were identified, and in 13 (87[percnt]) of 15 carcinomas, areas of associated APT were identified. Stromal invasion was manifested by confluent glandular growth in all 15 invasive carcinomas and all tumors with microinvasion. Destructive infiltrative growth was also present in 2 (13[percnt]) of 15 carcinomas. Confluent glandular growth was the most common manifestation of stromal invasion and therefore served as the best criterion for the diagnosis of carcinoma. Squamous differentiation was observed in 24 tumors, and mucinous differentiation was seen in 20 tumors and was most often seen in APTs. Endometriosis was present in 14 patients with APTs and one patient with carcinoma. Four patients had hyperplasia or atypical hyperplasia of the endometrium. One patient with an APT had a concurrent peritoneal serous neoplasm. Twenty-one patients had available clinical follow-up. Twenty (95[percnt]) of 21 patients, including six with invasive carcinoma, two with microinvasion, one with intraepithelial carcinoma, and 11 with APT were alive with no evidence of disease with a mean follow-up of 47 months. One patient with carcinoma had recurrent tumor after 46 months and was alive 40 months after resection of the recurrent tumor. In this large series of proliferative endometrioid tumors, all were stage I and only one patient had a recurrence. Most carcinomas contained evidence of a precursor APT, and in some APTs, an associated benign adenofibroma was identified. Microinvasion or intraepithelial carcinoma occurred in 19[percnt] of APTs. This finding likely reflects the various stages of endometrioid carcinogenesis in the ovary. For clinical management, we suggest that these tumors be divided into two categories[mdash]APTs and well-differentiated carcinoma[mdash]because based on the available data, cytologic atypia and microinvasion appear not to affect the prognosis.

Author Information

From the Division of Gynecologic Pathology, Department of Pathology (K.A.B., R.J.K.) and the Department of Gynecology and Obstetrics (R.J.K.), The Johns Hopkins Hospital, Baltimore, Maryland, U.S.A.

Address correspondence and reprint requests to Robert J. Kurman, MD, Department of Pathology, The Johns Hopkins Hospital, 600 North Wolfe Street, Baltimore, MD 21287, U.S.A.; e-mail[colon] rkurman[commat]jhmi.edu

Article Outline

Approximately 20[percnt] of surface ovarian tumors are reported to show endometrioid differentiation, and most of these are carcinomas. 2,9,17,18,25,30,31,33,43,44,46,47,53,55 In contrast, atypical proliferative (borderline) endometrioid tumors are uncommon, with only approximately 125 reported cases. 2,5,11,19,23,24,27,29,33,41,43,53,54,57,59 Although first established as a separate category in the early 1970s by the International Federation of Gynecology and Obstetrics and the World Health Organization (WHO), there is still no agreement on the criteria for the diagnosis of borderline endometrioid tumors. 20,51 Most authorities follow the guidelines of the WHO, which include in this category, tumors composed of glands lined by endometrioid-type epithelium that is atypical or cytologically malignant but that lack obvious stromal invasion. Using this definition, the pathologist is faced with two significant problems. First, the presence of a dense fibromatous stroma, which is the rule rather than the exception in these tumors, makes the recognition of a stromal reaction, that is, evidence of invasion, difficult. Second, because tumors containing cells that are cytologically malignant, but show no evidence of stromal invasion, qualify for inclusion in this category, it is theoretically possible for a grade 3 (noninvasive) carcinoma to be classified as borderline. Indeed, Scully et al. 50 recommend that if an endometrioid tumor that lacks stromal invasion is cytologically malignant, it should be graded 1 to 3 according to the criteria used for grading endometrioid carcinomas of the uterine corpus.

Recognizing these limitations of the WHO classification, various investigators have attempted to address the range of histologic appearances seen in proliferative or borderline endometrioid tumors. 2,8,36,50,53 Bell et al. 2 studied adenofibromatous endometrioid lesions and divided them into benign, atypical, borderline, and malignant categories based on a combination of architectural and cytologic features. A later study by Snyder et al., 53 which encompassed all proliferative endometrioid tumors, resulted in a four-tiered system that included adenofibroma, proliferative endometrioid tumor (PET), endometrioid tumor of low malignant potential (ETLMP), and ETLMP with microinvasion. Although adequately addressing the range of histologic appearances observed in these tumors, these classifications are cumbersome, and because only a relatively small number of cases with clinical follow-up have been reported, they provide limited prognostic information. We therefore undertook an analysis of proliferative endometrioid tumors aimed at addressing the following issues[colon] the histologic criteria for identifying stromal invasion, the clinical significance of cytologic atypia and microinvasion, and the usefulness of the current classification systems for providing prognostic information that can assist in management and that also reflects our understanding of carcinogenesis. Because the critical cut-off point in the WHO classification is between tumors of borderline malignancy and frankly invasive carcinoma, our analysis excluded endometriomas, endometrioid adenofibromas, and moderately to poorly differentiated endometrioid carcinoma.

Back to Top | Article Outline


Case Selection

Cases of atypical proliferative endometrioid (borderline or low malignant potential) tumors (APTs) and well-differentiated endometrioid carcinoma of the ovary were identified from the surgical pathology files of The Johns Hopkins Hospital and the consultation files of one of the authors (R.J.K.). Ovarian lesions diagnosed as endometriotic cyst or endometrioma, endometrioid adenofibroma, adenosarcoma, and moderately to poorly differentiated endometrioid carcinoma were not included in this study. Cases with concurrent endometrioid adenocarcinoma of the endometrium were also excluded. Fifty-six tumors from 53 patients met the inclusion criteria and formed the basis for this study. Forty-five (85[percnt]) of the 53 cases were consultation cases from outside institutions.

Back to Top | Article Outline
Clinical and Pathologic Data

Patient age, presenting symptoms, tumor site, size, gross appearance, presence of surface involvement or rupture, and the surgical procedure performed were determined from the surgical pathology report. Microscopic description of the endometrium was based on review of the slides when available or was recorded from the pathology report. All patients underwent removal of the tumor and surgical exploration, and some underwent surgical staging consisting of peritoneal biopsies, omentectomy, or lymph node sampling. For patients who underwent surgical staging, the clinicopathologic stage at diagnosis was assigned using the International Federation of Gynecology and Obstetrics staging system. 39 The remaining patients were assigned a stage based on the findings at surgical exploration. Clinical follow-up and treatment information was obtained for 21 patients through The Johns Hopkins Tumor Registry and from referring pathologists and clinicians. The Social Security Death Index (SSDI) database was searched using social security number or name and date of birth for all patients.

Tumors were examined for the presence of invasion, and when present, the size of the largest invasive focus was measured. Two patterns of stromal invasion were recognized[colon] haphazardly arranged nests and glands infiltrating the stroma with or without a stromal reaction (Fig. 1) and a confluent architectural pattern characterized by markedly crowded glands without intervening stroma or the presence of a cribriform pattern (Fig. 2). Confluent cribriform growth requires the fusion of multiple glands without intervening stroma. A confluent or cribriform proliferation in a single gland does not qualify for invasion.

The predominant growth pattern of the tumor was classified as adenofibromatous or glandular and papillary. Adenofibromatous tumors contained a fibrous or cellular stroma containing nodular islands of glands often showing squamous metaplasia (Fig. 3). Tumors with a predominantly glandular or papillary pattern were often cystic and contained an intracystic proliferation of papillary and glandular growth with a less prominent stromal component (Fig. 4). The gland-to-stroma ratio in the predominantly adenofibromatous growth pattern ranged from low to high.

In APTs and carcinomas, identification of an underlying benign adenofibroma consisting of a discrete area or areas in the tumor with fibrous or cellular stroma containing widely spaced glands lined by a single layer of bland epithelium was noted. In carcinomas, areas of associated APT were identified. Other types of differentiation, including squamous, mucinous, tubal, and secretory, were recorded.

Nuclear grade was evaluated in the most atypical area using a three-tiered grading system similar to one that has been previously described. 4 Grade 1 nuclei resemble those in proliferative phase endometrium but with minimal nuclear enlargement, no significant hyperchromasia, and single small nucleoli (Fig. 5A). Grade 2 nuclei showed more nuclear enlargement and pleomorphism, coarser chromatin, and variably prominent nucleoli (Fig. 5B). Grade 3 nuclei were characterized by marked nuclear enlargement with vesicular chromatin and prominent nucleoli (Fig. 5C).

In tumors with a stromal component, the stroma was characterized as cellular or fibrous. Cellular stroma was composed of densely packed, enlarged, spindle-shaped fibroblastic cells with minimal intervening collagen. In contrast, tumors with fibrous stroma were paucicellular with scattered elongated spindle cells in a densely collagenous background. Stromal inflammation was evaluated as none, mild, moderate, or marked, and the type of inflammation (that is, acute, chronic, or mixed) was noted. The presence or absence of necrosis was evaluated for each tumor. The type of necrosis was further characterized as zonal, which consisted of confluent areas of infarcted tumor measuring at least one half of a low-power field, or as lumenal, which consisted of necrotic debris confined to the cyst cavity or glandular lumens. The presence of endometriosis in proximity to the tumor or elsewhere was also noted.

Back to Top | Article Outline
Tumor Classification

Tumors were classified into four diagnostic categories based on architectural and cytologic features. APT is characterized by a proliferation of glands lined by endometrioid-type epithelium. Glands may be densely packed with little intervening stroma or may be separated by more abundant, cellular or fibromatous stroma. At the low end of proliferation, APTs merge with adenofibromas, but in contrast to the latter, in which glands are widely separated by abundant stroma, the glands in APTs have more complex outlines and are more closely packed. Cytologic atypia in APTs is variable but does not exceed grade 2. APT with intraepithelial carcinoma is an APT with diffuse grade 3 nuclear atypia.

Atypical proliferative tumor with microinvasion is an APT with one or multiple foci of stromal invasion, consisting of destructive infiltrative growth or confluent glandular growth, with each single focus measuring less than 5 mm in maximum diameter.

Well-differentiated invasive carcinoma is an endometrioid tumor with a predominantly glandular or papillary architecture ([lt]5[percnt] solid) and with at least one focus of stromal invasion measuring 5 mm or more in maximum diameter, regardless of cytologic features.

The cut-off of 5 mm for separating microinvasion from frank invasion is arbitrary but was selected to maintain uniformity with the value used for ovarian mucinous tumors.

Back to Top | Article Outline


Atypical Proliferative Tumors
Clinical Findings

There were 33 tumors from 31 patients whose age ranged from 24 to 85 years (mean, 46 yrs; median, 42 yrs). Sixteen patients had a mass, four presented with pain, and in 11 patients, the tumor was an incidental finding or the presenting symptoms were not known.

Back to Top | Article Outline
Gross Features

The tumors involved the right ovary in 11 patients and the left ovary in 16 patients. The side was not specified in one patient, and the tumors were bilateral in three patients. The mean tumor size was 6.4 cm (range, 1[ndash]26 cm). Fifteen tumors were cystic and solid; eight were predominantly solid; and 10 were predominantly cystic. Fifteen tumors were tan, yellow, or pink; three were brown; and four were gray or white. The cyst fluid was hemorrhagic or green in eight cases, mucoid in four cases, and clear in five cases. Two tumors were noted to have ruptured intraoperatively. In three tumors, gross involvement of the ovarian surface was evident.

Back to Top | Article Outline
Microscopic Features

Twenty-one tumors showed a predominantly adenofibromatous pattern (Fig. 6), and 12 showed a glandular or papillary pattern (Table 1). In seven tumors, there was evidence of an underlying benign adenofibroma. Two patients also had a serous adenofibroma in the contralateral ovary. The tumor stroma was cellular in 19 cases, with periglandular cuffing characterized by increased stromal cellularity surrounding glandular elements in seven cases, and was fibrous in 13 cases, with periglandular cuffing in three cases. In one tumor, glands were densely packed with minimal intervening stroma. Stromal inflammation was uncommon; nine tumors showed mild to moderate chronic inflammation and one showed marked chronic inflammation. Extensive areas of necrosis with large zones of infarcted and hemorrhagic tumor were seen in 10 tumors. An additional 15 tumors had evidence of necrosis that was confined to glandular lumens or cysts. Most tumors contained grade 1 nuclei (23 of 33 [lsqb]70[percnt][rsqb]), and the remainder contained grade 2 nuclei (10 of 33 [lsqb]30[percnt][rsqb]). Squamous and mucinous differentiation were common and were identified in 16 and 15 tumors, respectively. Secretory and tubal differentiation were less common and occurred in nine and eight tumors, respectively. In 18 of the 31 patients, the endometrium could be evaluated. In 12 of these patients, slides of the endometrium were available for review. Results for the other six patients were obtained from the surgical pathology reports. The endometrium showed complex atypical hyperplasia in two patients and simple hyperplasia in one (Table 2). The remaining patients had proliferative endometrium (eight patients), atrophic endometrium (five patients), and secretory endometrium (two patients). Twelve patients had endometriosis, which in nine patients involved the ovary with the endometrioid tumor. In one of these patients, the tumor appeared to arise in an endometriotic cyst.

Back to Top | Article Outline
Treatment and Follow Up

Seventeen patients had a total abdominal hysterectomy (TAH) and a bilateral salpino-oophorectomy (BSO); 12 had a unilateral salpino-oophorectomy (USO); one had a BSO; and one had a cystectomy. Sixteen patients also underwent surgical staging. In all 31 patients, the tumors were confined to the ovary (stage I). In 28 patients, the tumor was unilateral and three patients had bilateral tumors. Four tumors had surface involvement; two ruptured intraoperatively; and one patient had positive peritoneal washings. In one of the patients with bilateral tumors, one of the APTs showed microinvasion. One patient with an APT was found to have implants involving several sites in the peritoneum. The implants were invasive implants of the type seen in association with serous tumors of the ovary; they did not resemble the patient's endometrioid ovarian tumor. The patient was therefore not considered to have advanced stage disease. This patient had undergone a TAH and BSO with surgical staging, and no evidence of an ovarian serous tumor was found. Two patients were treated with chemotherapy; six received no chemotherapy and information was unavailable for the remaining 23 patients. Clinical follow up was available for 11 patients, and all were alive with no evidence of disease at a mean of 48 months of follow up (Table 3). In addition to these 11 patients, the patient who had peritoneal implants is known to be alive at 76 months of follow up, but the details of her clinical course are unavailable. None of the patients in the SSDI database were known to have died.

Back to Top | Article Outline
APT With Intraepithelial Carcinoma
Clinical Findings

The three patients ranged in age from 54 to 78 years (mean, 62 yrs; median, 55 yrs). One patient presented with a mass, and the presenting complaint was not known for the other two patients.

Back to Top | Article Outline
Gross Features

All three tumors were unilateral. Two were left-sided and one was right-sided. The mean tumor size was 9.5 cm (range, 4.5[ndash]16.0 cm). Two tumors were cystic and solid, and one was solid. Two were tan and one was white. One cystic tumor contained greenish fluid. None had surface involvement and none were ruptured.

Back to Top | Article Outline
Microscopic Features

All three tumors showed the architectural features of an APT, with two being predominantly adenofibromatous and one being glandular and papillary (Table 1). An associated benign adenofibroma was not identified in any case. The stroma was cellular in all three cases, with one tumor showing periglandular cuffing and mild chronic inflammation. All three tumors contained cells with diffuse, marked cytologic atypia evidenced by nuclear enlargement and pleomorphism with large, prominent nucleoli (Figs. 7 and 8). None of the three tumors showed necrosis. Squamous differentiation was not observed, but one tumor showed focal mucinous differentiation and one showed tubal differentiation. According to the surgical pathology reports, the endometrium was atrophic in one patient, proliferative in one patient, and benign in one patient. None of the patients had endometriosis (Table 2).

Back to Top | Article Outline
Treatment and Follow Up

All three patients underwent a TAH and BSO and two were also surgically staged. All three tumors were confined to the ovary (stage I). One of these patients had positive peritoneal washings. The one patient with available detailed treatment information did not receive chemotherapy. Clinical follow up was available in one patient who was alive with no evidence of disease at 18 months (Table 3). None of the patients in the SSDI database were known to have died.

Back to Top | Article Outline
Atypical Proliferative Tumors With Microinvasion
Clinical Findings

The five patients ranged in age from 19 to 69 years (mean, 42 yrs; median, 38 yrs). Four patients presented with a mass and one with pain.

Back to Top | Article Outline
Gross Features

Four tumors were unilateral (two right-sided and two left-sided). One patient had bilateral tumors[colon] an APT involving the left ovary and an APT with microinvasion involving the right ovary. The mean tumor size was 8.4 cm (range, 5[ndash]12 cm). Three tumors were cystic, and two were cystic and solid. Three tumors were tan and two were brown. Four tumors contained hemorrhagic cyst fluid.

Back to Top | Article Outline
Microscopic Features

All five tumors had the overall architecture of an APT, with two being predominantly adenofibromatous and three being glandular and papillary (Table 1). In one tumor, an underlying benign adenofibroma was identified. All five tumors showed at least one focus of confluent glandular growth measuring less than 5 mm in greatest diameter. Three tumors contained multiple small foci of confluent glandular growth with the largest focus measuring 2, 3, and 4 mm. The other two tumors each showed a single focus of confluent glandular growth measuring 1 and 2 mm in greatest diameter. No areas of destructive infiltrative growth were identified in any of the tumors. The stroma was fibrous in four tumors and cellular in one. Three tumors had moderate to marked stromal inflammation. Three tumors contained large zones of necrosis, and two contained necrotic debris within cyst and gland lumens. The nuclear grade was grade 1 in three tumors and grade 2 in two others. Squamous differentiation was present in four tumors, mucinous and secretory in one, and mucinous in another. Three patients underwent concurrent endometrial sampling, which was available for review in one patient. In two cases, the endometrium was atrophic, and in one case, it was proliferative. Two patients had endometriosis (Table 2). In one of these patients, a focal area of complex atypical hyperplasia was identified in an area of peritoneal endometriosis.

Back to Top | Article Outline
Treatment and Follow Up

Three patients had a TAH and BSO, and two of them had surgical staging. Two patients had a USO, and one of these was surgically staged. Four tumors were confined to one ovary. One patient had bilateral APTs, one had microinvasion, and this patient also had positive peritoneal washings. Detailed treatment information was available for three patients; one was treated with chemotherapy and the other two were not. Clinical follow-up was known for two patients who had no evidence of disease at a mean of 24 months of follow up (Table 3). None of the patients in the SSDI database were known to have died.

Back to Top | Article Outline
Invasive Carcinoma
Clinical Findings

The 15 patients ranged in age from 32 to 70 years (mean, 55 yrs; median, 56 yrs). The most common presentation was a mass seen in seven patients. Two patients presented with increasing abdominal girth, one with pain, and in five patients, the tumor was an incidental finding or the presenting symptom was unknown.

Back to Top | Article Outline
Gross Features

All 15 tumors were unilateral, with eight tumors involving the right ovary and seven tumor involving the left ovary. The mean tumor size was 15 cm (range, 8[ndash]27 cm). Eight tumors were cystic and solid, five were predominantly solid, and two were predominantly cystic. Nine tumors were tan, yellow, or pink; two were brown; and two were gray or white. In seven of the cystic tumors, the cyst fluid was hemorrhagic or green. In two tumors, there was involvement of the surface of the ovary, and in one tumor there was evidence of rupture.

Back to Top | Article Outline
Microscopic Features

The diagnosis of carcinoma was based on the finding of confluent glandular growth exceeding 5 mm in 15 (100[percnt]) cases (Table 1). Destructive infiltrative growth was also seen in two (13[percnt]) cases. In tumors with a confluent pattern of invasion, areas of confluent growth frequently exceeded 1 cm measured on any one slide. In the two tumors with destructive infiltration, a space surrounded the infiltrative nests, but no significant desmoplastic stromal reaction was discernible (Fig. 9). A glandular or papillary architecture was seen in most carcinomas, with only four tumors showing a predominantly adenofibromatous growth pattern. In 13 (87[percnt]) tumors, areas of APT could be identified in association with the carcinoma. The stroma was fibrous in eight tumors and cellular in five tumors. In two tumors, stroma was barely discernible because of the marked glandular proliferation. Five tumors had mild to moderate stromal inflammation, and 10 showed no significant stromal inflammation. Eight tumors contained large zones of necrosis, two of which also contained necrotic material in cysts and gland lumens. One tumor showed necrosis only within cysts and gland lumens, and six tumors had no necrosis. The tumors showed a spectrum of nuclear atypia ranging from grade 1 in four tumors, to grade 2 in four tumors, and grade 3 in seven tumors. The two tumors with destructive infiltrative growth had grade 3 atypia. Four tumors showed squamous differentiation. Secretory, tubal, and mucinous differentiation were seen in one, one, and two cases, respectively. Thirteen patients (four with slides available for review) had concurrent endometrial sampling. The endometrium was proliferative in four patients, atrophic or inactive in four patients, secretory in two patients, benign in two patients, and had simple hyperplasia in one patient (Table 2). One patient had evidence of endometriosis involving the uterine serosa and parametria.

Back to Top | Article Outline
Treatment and Follow Up

Twelve patients underwent a TAH and BSO; two underwent a USO; and one underwent a BSO. All patients underwent surgical staging. All 15 tumors were confined to one ovary (stage I). Two tumors had involvement of the surface of the ovary and one ruptured intraoperatively. Detailed treatment information was available for five patients, of whom two received systemic chemotherapy and three did not. Clinical follow up was available for seven patients, six of whom had no evidence of disease, with a mean follow up of 44 months (Table 3). One patient had recurrent carcinoma involving the bowel 46 months after undergoing a TAH and BSO for a unilateral mass that had reportedly ruptured. The primary tumor was predominantly glandular and had diffuse grade 3 nuclear atypia. The diagnosis of carcinoma was based on the finding of large areas of confluent glandular growth. The serosal bowel recurrence was resected and was histologically similar to the ovarian primary tumor. The patient was alive 40 months after this recurrence but was lost to further clinical follow up.

Back to Top | Article Outline


In the WHO classification, endometrioid ovarian tumors are divided into benign, borderline, and malignant (carcinoma) categories, with a diagnosis of carcinoma based on the identification of destructive infiltrative stromal invasion, but the criteria for the identification of stromal invasion are not described in detail in the WHO classification. 48 In contrast, the criteria for the identification of stromal invasion in the endometrium, which separates complex atypical hyperplasia from well-differentiated carcinoma, are well established and include altered desmoplastic stroma, confluent glandular growth, and papillary growth pattern. 32 We reasoned that the same criteria for recognizing stromal invasion in the endometrium could be applied to ovarian endometrioid tumors. Application of these criteria to ovarian tumors, however, has limitations because the frequent adenofibromatous architecture of the ovarian tumors makes identification of a desmoplastic stromal reaction difficult. In addition, the tendency for many ovarian neoplasms to show a papillary architecture limits the interpretation of papillary growth as evidence of stromal invasion in the ovary. In most cases in this series, the diagnosis of carcinoma was based only on a complex glandular proliferation with confluent growth. This also has been the experience of Scully et al. 50 Confluent glandular growth was the only pattern of invasion seen in the tumor from the one patient who had a recurrence. Accordingly, confluent glandular growth appears to be the most common and reliable indicator of stromal invasion in endometrioid tumors. In the two carcinomas in this series that showed destructive infiltrative growth by virtue of haphazard glands infiltrating the stroma, recognition of an unequivocal stromal reaction was difficult because of the adenofibromatous nature of endometrioid tumors in the ovary. Thus, relying on the presence of this pattern as the sole diagnostic criterion for carcinoma would have resulted in an underdiagnosis of carcinoma, because only 13[percnt] of the carcinomas in our series had this type of invasion.

Although the criteria for the diagnosis of APTs vary among studies, a review of the literature identified 92 cases with available clinical follow up, which appear to correspond to what we have designated APT, APT with microinvasion, or intraepithelial carcinoma. 2,11,23,24,33,41,53 The data from these seven studies and the 14 patients with clinical follow up from the current series are summarized in Table 4. All 106 tumors were stage I at the time of diagnosis, and 105 (99[percnt]) of 106 patients had no evidence of recurrence or died of other causes, with a mean follow up ranging from 28 to 96 months. Among these 106 patients, there was one who was found to have recurrent disease 6 weeks after initial surgery. The tumor was histologically identical to her ovarian lesion and involved the vagina. 24 The vaginal tumor was not resected and the patient was alive with disease 6.4 years later. The stable clinical course suggests that the tumor was implanted or incompletely resected at the time of surgery rather than representing a true metastasis. Another patient who had an unusual clinical course was reported by Bell et al. 2 This patient underwent a USO for an atypical endometrioid adenofibroma and had an endometrioid carcinoma in her opposite ovary 3 years later. At the time of the study, she was currently free of disease 6 months after resection of the carcinoma. In view of the apparent absence of disease elsewhere, it is likely that this case represents a new independent carcinoma developing in the opposite ovary.

Among all the reported cases of atypical proliferative (borderline) endometrioid tumors, only two were described as having advanced stage disease at diagnosis, but clinical follow-up information was unavailable in both cases. 15,43 One of these patients had [ldquo]peritoneal implants[rdquo] of bland endometrioid epithelium outside of the ovary unassociated with hemorrhage or endometrial type stroma. 43 Endometrial stroma, however, is notoriously difficult to identify even in what is otherwise bona fide endometriosis. Accordingly, it seems likely that the bland endometrial tissue in this case was endometriosis rather than a metastasis from the ovarian tumor. The other case of an endometrioid adenofibroma of the ovary associated with endometriosis had a similar tumor involving the omentum. 15 It is likely that the omental tumor, like the ovarian tumor, developed independently from endometriosis. There was one patient in our series with peritoneal implants at the time of diagnosis. It is noteworthy that these implants showed no resemblance to the endometrioid tumor. Instead, the lesions were invasive implants of the type associated with serous borderline tumors. The patient had undergone a TAH and BSO, and no serous ovarian tumor was identified. The patient is known to be alive 76 months after the initial diagnosis, but detailed information regarding her clinical course was unavailable. It is possible that the patient had a primary peritoneal serous neoplasm. We are currently analyzing a group of similar tumors that we have classified as peritoneal micropapillary serous carcinoma (unpublished data). Thus, based on the current analysis and review of the literature, including those rare patients reported with extraovarian disease, there is no well-documented case of an advanced stage endometrioid APT, and no documented case with an adverse prognosis has been reported. Furthermore, in the current series, APTs including those with intraepithelial carcinoma and microinvasion are stage I and are also not associated with an adverse outcome, albeit with relatively short follow up.

A review of the literature of reported cases of endometrioid carcinomas shows that approximately 40[percnt] are stage I (Table 5) and many are low grade. 3,7,9,10,17,18,26,28,31,33,35,37,44,47,55,56,60 It should be noted that careful surgical staging was introduced in the early 1980s so that the stages reported in older studies are likely to be inaccurate. In addition, the criteria for histologic grading vary significantly among studies. We classified tumors that had predominantly glandular pattern with minimal solid areas as well-differentiated, regardless of cytologic atypia. Some reviewers might upgrade these tumors to moderately differentiated because of the cytologic atypia, as could be done in similar tumors arising in the endometrium. The reported 5-year survival for all stages in larger series of endometrioid carcinomas ranges from 38[percnt] to 70[percnt] (Table 6). 1,3,9,18,26,30,31,33[ndash]35,47,49,55,61 It should be emphasized, however, that invasive well-differentiated carcinomas are nearly always stage I and unilateral at presentation and their prognosis is particularly favorable (Table 7). 9[ndash]11,17,18,28,31,33,42,47,55,56,60 Of 29 reported patients with stage I, grade 1 tumors followed up for at least 5 years, only one patient died of disease after 24 months; the clinical details of this case were not reported. 9,17,31,33,42,55,56 Our series included only well-differentiated carcinomas that were all unilateral stage I tumors and six of seven patients with clinical follow up were alive with no evidence of disease at a mean follow up of 47 months. One patient with a stage I carcinoma had recurrent disease 46 months after initial surgery and was currently alive 40 months after this recurrence, but was subsequently lost to follow up. It is possible that the serosal bowel tumor represented a new primary tumor arising from endometriosis. The histologic similarity to the primary tumor, however, and lack of endometriosis in this patient favor a recurrence. Accordingly, based on our study and a review of the literature, stage I, well-differentiated endometrioid carcinomas have an excellent prognosis. Therefore, the important clinical cut-off point for stage I endometrioid tumors is the distinction of low-grade (grade 1) versus high-grade (grades 2 and 3) carcinomas, because the outcome of APTs and stage I, well-differentiated carcinoma is excellent and has a 5-year survival rate of nearly 100[percnt]. Finally, in view of the relatively uncommon occurrence of bilateral APTs and well-differentiated endometrioid carcinoma, the finding of bilateral tumors should raise the suspicion of metastases from the cervix, endometrium, or colon because they can closely simulate a primary ovarian endometrioid neoplasm.

This study also sheds light on the pathogenesis of endometrioid carcinoma of the ovary. In particular, the progression of endometrioid neoplasms appears to closely parallel tumor progression of ovarian mucinous neoplasia rather than serous carcinoma. 40 In the current study, 24[percnt] of the APTs, including tumors with microinvasion and intraepithelial carcinoma, showed evidence of an associated adenofibroma with multifocal areas of increased proliferation and architectural complexity. Similarly, 87[percnt] of the carcinomas also had foci of APT and areas of transition often observed. These observations have been made in previous studies and suggest a model of progression from benign endometriosis through atypical proliferative to malignant tumors similar to that seen with mucinous carcinomas. 2,19,37,59,60 Such transitions result in tumors that architecturally qualify as borderline tumors, but show marked cytologic atypia or foci of stromal invasion, necessitating the addition of diagnostic categories, including APT with intraepithelial carcinoma and APT with microinvasion, to cover the range of histologic appearances. In contrast, serous carcinomas frequently present as advanced stage, poorly differentiated tumors with no discernible evidence of a precursor lesion, suggesting that they develop de novo and only infrequently from a serous borderline tumor.

Although the cell of origin that gives rise to endometrioid neoplasms in the ovary is not known, several investigators have postulated an origin from endometriosis with subsequent neoplastic transformation. 3,9,10,19,21,22,45,46 Previous studies have reported concurrent endometriosis in approximately 20[percnt] of patients with endometrioid carcinoma and in approximately 35[percnt] of patients with borderline endometrioid tumors. 2,6,7,9[ndash]11,16[ndash]18,22,26,28,30,31,35,41[ndash]43,45[ndash]47,49,53,56,60 In approximately half the reported cases, the endometriosis was found in the same ovary containing the endometrioid neoplasm. Rare tumors, including one in this series, have been shown to arise in direct continuity from endometriosis, thereby providing further evidence for an origin from ectopic endometrial tissue. 7,9,10,17,44,46,47 Molecular analysis has also shown a common lineage in some ovarian carcinomas with adjacent endometriosis. 21 The data implicating the development of endometrioid tumors from endometriosis are strongly suggestive and highly provocative but do not completely rule out an origin from the surface epithelium of the ovary through a process of metaplasia and neoplastic transformation.

Another controversial area of investigation in endometrioid tumors of the ovary is the role of unopposed estrogenic stimulation in a manner analogous to that in uterine carcinogenesis. Examination of the endometrium from patients with endometrioid carcinoma of the ovary shows that approximately 10[percnt] to 15[percnt] patients have concurrent carcinoma involving the endometrium. 3,6,7,9,10,17,18,26,28,31,34,35,41,42,44,46,47,49,55,56,60 Based on the limited extent of the uterine tumors and favorable clinical outcome, many of these tumors are thought to represent separate primary lesions. 4,9,28,38,58 Molecular biologic studies have further supported the view that many of these patients have independent primary tumors, and it is conceivable that both neoplasms are related to systemic unopposed estrogenic stimulation. 12,14,52 Many endometrioid ovarian tumors, however, are not associated with concurrent endometrial hyperplasia or carcinoma. To ensure that all tumors in the current study were primary ovarian tumors and not metastatic foci from the uterus, cases with concurrent endometrial carcinoma were excluded. As was found in previous studies, evaluation of the endometrium in our cases showed no specific changes. 2,3,6,7,9[ndash]11,17,18,26,28,31,34,35,41,43,44,47,49,53,55,56,60 A minority of patients had hyperplasia, and the remainder were equally distributed among atrophic, proliferative, and secretory endometrium. Furthermore, none of the patients in this study with known clinical follow up, who did not initially undergo a hysterectomy, developed endometrial carcinoma. Thus, although endometrioid carcinomas of the uterus have been convincingly linked to excessive estrogenic stimulation, the findings in this study and others suggest that excess estrogen, at least systemically, plays a lesser role in endometrioid ovarian carcinogenesis.

In summary, we report the clinicopathologic features of 56 proliferative endometrioid ovarian tumors ranging from APTs to well-differentiated carcinomas. An argument can be made to divide these tumors into APT, APT with intraepithelial carcinoma, APT with microinvasion, and carcinoma because this categorization encompasses the wide range of histologic appearances seen in endometrioid tumors of the ovary and reflects the various stages of endometrioid carcinogenesis. From a managerial standpoint, however, the behavior of APTs, including the limited number of cases with intraepithelial carcinoma and microinvasion that have been studied, is benign after oophorectomy. Even well-differentiated carcinoma, when confined to the ovary, is rarely associated with an adverse outcome. Based on our findings and a review of the literature, we recommend that for clinical management, well-differentiated proliferative endometrioid tumors be classified into two categories[colon] APTs tumors and well-differentiated carcinoma. The terms borderline and low malignant potential are misleading because no well-documented case of such a tumor has been associated with an adverse outcome. The excellent prognosis of well-differentiated carcinomas suggests that surgery alone is adequate for management of these tumors, but closer clinical follow up than for patients with APTs is indicated because these tumors occasionally recur. Future clinicopathologic studies should evaluate criteria for distinguishing low-grade (grade 1) from high-grade (grades 2 and 3) endometrioid carcinoma because the literature suggests a more adverse outcome for patients with high-grade, stage I tumors. Based on current data, endometrioid APTs most likely arise from endometriosis that becomes more proliferative and atypical and results in the development of intraepithelial carcinoma, microinvasion, and eventually frankly invasive carcinoma. Future molecular biologic studies will undoubtedly clarify whether this is a valid model of endometrioid carcinogenesis.

Back to Top | Article Outline


1. Aure JC, Hoeg K, Kolstad P. Clinical and histologic studies of ovarian carcinoma[colon] long term follow-up of 990 cases. Obstet Gynecol 1971; 37[colon]1[ndash]9.
2. Bell DA, Scully RE. Atypical and borderline endometrioid adenofibromas of the ovary. A report of 27 cases. Am J Surg Pathol 1985; 9[colon]205[ndash]14.
3. Brescia RJ, Dubin N, Demopoulos RI. Endometrioid and clear cell carcinoma of the ovary. Factors affecting survival. Int J Gynecol Pathol 1989; 8[colon]132[ndash]8.
4. Burks RT, Sherman ME, Kurman RJ. Micropapillary serous carcinoma of the ovary. A distinctive low-grade carcinoma related to serous borderline tumors. Am J Surg Pathol 1996; 20[colon]1319[ndash]30.
5. Carter J, Atkinson K, Coppleson M, et al. A comparative study of proliferating (borderline) and invasive epithelial ovarian tumors in young women. Aust NZ J Obstet Gynaecol 1989; 29[colon]245[ndash]9.
6. Chew S, Tham KF, Ratnam SS. A series of ovarian clear cell and endometrioid carcinoma and their association with endometriosis. Singapore Med J 1997; 38[colon]289[ndash]91.
7. Curling OM, Hudson CN. Endometrioid tumors of the ovary. Br J Obstet Gynaecol 1975; 82[colon]405[ndash]11.
8. Czernobilsky B. Endometrioid neoplasia of the ovary[colon] a reappraisal. Int J Gynecol Pathol 1982; 1[colon]203[ndash]10.
9. Czernobilsky B, Silverman BB, Mikuta JJ. Endometrioid carcinoma of the ovary. A clinicopathologic study of 75 cases. Cancer 1970; 26[colon]1141[ndash]52.
10. DePriest PD, Banks ER, Powell DE, et al. Endometrioid carcinoma of the ovary and endometriosis[colon] the association in postmenopausal women. Gynecol Oncol 1992; 47[colon]71[ndash]5.
11. Eichhorn JH, Scully RE. Endometrioid ciliated-cell tumors of the ovary[colon] a report of 5 cases. Int J Gynecol Pathol 1996; 15[colon]248[ndash]56.
12. Emmert-Buck MR, Chuaqui R, Zhuang Z, et al. Molecular analysis of synchronous uterine and ovarian endometrioid tumors. Int J Gynecol Pathol 1997; 16[colon]143[ndash]8.
13. Fu YS, Stock RJ, Reagan JW, et al. Significance of squamous components in endometrioid carcinoma of the ovary. Cancer 1979; 44[colon]614[ndash]21.
14. Fujita M, Enomoto T, Wada H, et al. Application of clonal analysis. Differential diagnosis for synchronous primary ovarian and endometrial cancers and metastatic cancer. Am J Clin Pathol 1996; 105[colon]350[ndash]9.
15. Gaing AA, Kimble CC, Belmonte AH, et al. Invasive ovarian endometrioid adenofibroma with omental implants and collision with endometrial carcinoma. Obstet Gynecol 1988; 71[colon]440[ndash]4.
16. Garzetti GG, Ciavattini A, Goteri G, et al. Endometrioid carcinoma of the ovary. Retrospective study. Eur J Gynaecol Oncol 1993; 14[colon]51[ndash]5.
17. Gray LA, Barnes ML. Endometrioid carcinoma of the ovary. Obstet Gynecol 1967; 29[colon]694[ndash]701.
18. Grosso G, Raspagliesi F, Baiocchi G, et al. Endometrioid carcinoma of the ovary[colon] a retrospective analysis of 106 cases. Tumori 1998; 84[colon]552[ndash]7.
19. Hughesdon PE. Benign endometrioid tumors of the ovary and the M[uuml]llerian concept of ovarian epithelial tumors. Histopathology 1984; 8[colon]977[ndash]90.
20. International Federation of Gynecology and Obstetrics. Classification and staging of malignant tumors in the female pelvis. Acta Obstet Gynecol Scand 1971;50[colon]1[ndash]7.
21. Jiang X, Morland SJ, Hitchcock A, et al. Allelotyping of endometriosis with adjacent ovarian carcinoma reveals evidence of a common lineage. Cancer Res 1998; 58[colon]1707[ndash]12.
22. Jimbo H, Yoshikawa H, Onda T, et al. Prevalence of ovarian endometriosis in epithelial ovarian cancer. Int J Gynecol and Obstet 1997; 59[colon]245[ndash]50.
23. Kao GF, Norris HJ. Cystadenofibromas of the ovary with epithelial atypia. Am J Surg Pathol 1978; 2[colon]357[ndash]63.
24. Kao GF, Norris HJ. Unusual cystadenofibromas[colon] endometrioid, mucinous, and clear cell types. Obstet Gynecol 1979; 54[colon]729[ndash]36.
25. Katsube Y, Berg JW, Silverberg SG. Epidemiologic pathology of ovarian tumors[colon] a histopathologic review of primary ovarian neoplasms diagnosed in the Denver standard metropolitan statistical area, 1 July[ndash]31 December 1969 and 1 July[ndash]31 December 1979. Int J Gynecol Pathol 1982; 1[colon]3[ndash]16.
26. Klemi PJ, Gr[ouml]nroos M. Endometrioid carcinoma of the ovary. A clinicopathologic, histochemical and electron microscopic study. Obstet Gynecol 1979; 53[colon]572[ndash]9.
27. Kliman L, Rome RM, Fortune DW. Low malignant potential tumors of the ovary[colon] a study of 76 cases. Obstet Gynecol 1986; 68[colon]338[ndash]44.
28. Kline RC, Wharton JT, Atkinson EN, et al. Endometrioid carcinoma of the ovary[colon] retrospective review of 145 cases. Gynecol Oncol 1990; 39[colon]337[ndash]46.
29. K[oelig]rn J, Trop[eacute] CG, Abeler VM. A retrospective study of 370 borderline tumors of the ovary treated at the Norwegian Radium Hospital from 1970 to 1982. A review of clinicopathologic features and treatment modalities. Cancer 1993; 71[colon]1810[ndash]20.
30. Kottmeier HL. Clinical staging in ovarian carcinoma. In Gentil F, Junqueira C, eds. Ovarian Cancer, UICC Monograph Series, vol 11. New York, NY[colon] Springer-Verlag, 1968[colon]146[ndash]56.
31. Kurman RJ, Craig JM. Endometrioid and clear cell carcinoma of the ovary. Cancer 1972; 29[colon]1653[ndash]64.
32. Kurman RJ, Norris HJ. Evaluation of criteria for distinguishing atypical endometrial hyperplasia from well-differentiated carcinoma. Cancer 1982; 49[colon]2547[ndash]59.
33. Long ME, Taylor HC. Endometrioid carcinoma of the ovary. Am J Obstet Gynecol 1964; 90[colon]936[ndash]50.
34. Martin Jimenez A, Miralles Pi RM, Cabello J, et al. Endometrioid carcinoma of the ovary. A clinicopathologic study of 17 cases. Eur J Gynaecol Oncol 1994; 15[colon]96[ndash]100.
35. McMeekin DS, Burger RA, Manetta A, et al. Endometrioid adenocarcinoma of the ovary and its relationship to endometriosis. Gynecol Oncol 1995; 59[colon]81[ndash]6.
36. Norris HJ. Proliferative endometrioid tumors and endometrioid tumors of low malignant potential of the ovary. Int J Gynecol Pathol 1993; 12[colon]134[ndash]40.
37. Ordi J, Schammel DP, Rasekh L, et al. Sertoliform endometrioid carcinomas of the ovary[colon] a clinicopathologic and immunohistochemical study of 13 cases. Mod Pathol 1999; 12[colon]933[ndash]40.
38. Pearl ML, Johnston CM, Frank TS, et al. Synchronous dual primary ovarian and endometrial carcinomas. Int J Gynaecol Obstet 1993; 43[colon]305[ndash]12.
39. Pettersson F, ed. Annual Report on Gynecologic Cancer of FIGO, vol 20. Stockholm[colon] Panorama Press, 1988.
40. Riopel MA, Ronnett BM, Kurman RJ. Evaluation of diagnostic criteria and behavior of ovarian intestinal-type mucinous tumors. Atypical proliferative (borderline) tumors and intraepithelial, microinvasive, invasive and metastatic carcinomas. Am J Surg Pathol 1999; 23[colon]617[ndash]35.
41. Roth LM, Czernobilsky B, Langley FA. Ovarian endometrioid adenofibromatous and cystadenofibromatous tumors[colon] benign, proliferating and malignant. Cancer 1981; 48[colon]1838[ndash]45.
42. Roth LM, Liban E, Czernobilsky B. Ovarian endometrioid tumors mimicking sertoli and sertoli-leydig cell tumors. Sertoliform variant of endometrioid carcinoma. Cancer 1982; 50[colon]1322[ndash]31.
43. Russell P. The pathologic assessment of ovarian neoplasms. II[colon] the proliferating `epithelial' tumors. Pathology 1979; 11[colon]251[ndash]82.
44. Russell P. The pathologic assessment of ovarian neoplasms. III[colon] the malignant `epithelial' tumors. Pathology 1979; 11[colon]493[ndash]532.
45. Sainz de la Cuesta R, Eichhorn JH, Rice LW, et al. Histologic transformation of benign endometriosis to early epithelial ovarian cancer. Gynecol Oncol 1996; 60[colon]238[ndash]44.
46. Sampson JA. Endometrial carcinoma of the ovary, arising in endometrial tissue in that organ. Arch Surg 1925; 10[colon]1[ndash]72.
47. Schueller EF, Kirol PM. Prognosis in endometrioid carcinoma of the ovary. Obstet Gynecol 1966; 27[colon]850[ndash]8.
48. Scully RE, Sobin LH. Histological classification of ovarian tumors. In[colon]Histological typing of ovarian tumors, 2nd edition. New York, NY[colon] Springer, 1999[colon]4.
49. Scully RE, Richardson GS, Barlow JF. The development of malignancy in endometriosis. Clin Obstet Gynecol 1966; 9[colon]384[ndash]411.
50. Scully RE, Young RH, Clement PB. Endometrioid tumors. In[colon] Tumors of the ovary, maldeveloped gonads, fallopian tube, and broad ligament. Washington, DC[colon] Armed Forces Institute of Pathology, 1998[colon]107[ndash]40.
51. Serov SF, Scully RE, Sobin LH. International histologic classification and staging of tumors, no. 9. In[colon] Histologic typing of ovarian tumors. Geneva[colon] World Health Organization, 1973.
52. Shenson DL, Gallion HH, Powell DE, et al. Loss of heterozygosity and genomic instability in synchronous endometrioid tumors of the ovary and endometrium. Cancer 1995; 76[colon]650[ndash]7.
53. Snyder RR, Norris HJ, Tavassoli F. Endometrioid proliferative and low malignant potential tumors of the ovary. A clinicopathologic study of 46 cases. Am J Surg Pathol 1988; 12[colon]661[ndash]71.
54. Sugiyama T, Nishida T, Kataoka A, et al. A pregnant woman with clear cell adenocarcinoma of the ovary arising from endometriosis and with benign and borderline adenofibroma of the clear cell and endometrioid types. Eur J Obstet Gynecol Reprod Biol 1997; 72[colon]47[ndash]50.
55. Tidy J, Mason P. Endometrioid carcinoma of the ovary[colon] a retrospective study. Br J Obstet Gynaecol 1988; 95[colon]1165[ndash]9.
56. Tornos C, Silva EG, Ordonez NG, et al. Endometrioid carcinoma of the ovary with a prominent spindle-cell component, a source of diagnostic confusion. A report of 14 cases. Am J Surg Pathol 1995; 19[colon]1343[ndash]53.
57. Treilleux I, Godeneche J, Duvillard P, et al. Collagenous spherulosis mimicking keratinizing squamous metaplasia in a borderline endometrioid tumor of the ovary. Histopathology 1999; 35[colon]271[ndash]6.
58. Ulbright TM, Roth LM. Metastatic and independent cancers of the endometrium and ovary[colon] a clinicopathologic study of 34 cases. Hum Pathol 1985; 16[colon]28[ndash]34.
59. Ushijima K, Nishida T, Muraoka Y, et al. Ovarian endometrioid adenocarcinoma coexistent with benign and borderline endometrioid adenofibroma[colon] a case report. J Obstet Gynaecol Res 1998; 24[colon]161[ndash]6.
60. Young RH, Prat J, Scully RE. Ovarian endometrioid carcinoma resembling sex cord-stromal tumors. A clinicopathologic analysis of 13 cases. Am J Surg Pathol 1982[colon] 6[colon]513[ndash]22.
61. Zwart J, Geisler JP, Geisler HE. Five-year survival in patients with endometrioid carcinoma of the ovary versus those with serous carcinoma. Eur J Gynaecol Oncol 1998; 19[colon]225[ndash]8.

Endometrioid ovarian carcinoma; Endometrioid ovarian tumor; Borderline ovarian tumor; Atypical proliferative ovarian tumor; Low malignant potential; Endometriosis

[copy] 2000 Lippincott Williams [amp] Wilkins, Inc.