Epithelial/Stromal Tumors
The tumors in this category account for approximately two-thirds of all ovarian tumors and for about 90% of all ovarian cancers in the Western world. Epithelial ovarian tumors are heterogeneous neoplasms that are primarily classified according to cell type into serous, mucinous, endometrioid, clear cell, transitional, and squamous cell tumors. However, benign counterparts of these cells are not found in the normal ovary and their neoplastic occurrence has long been attributed to müllerian ‘neometaplasia’ of the ovarian surface epithelium (mesothelium), which derives from the celomic epithelium. Specifically, since the celomic epithelium gives rise to the müllerian ducts, it was proposed that, as the surface epithelium becomes malignant, it would acquire the morphologic features of the müllerian duct epithelium; i.e., serous (fallopian tube-like), endometrioid (endometrium-like), and mucinous (endocervical-like). This aberrant differentiation constitutes the basis for ovarian tumor classification. However, even if an origin from the surface epithelium cannot be excluded, there is now compelling evidence that a number of tumors thought to be primary ovarian cancers actually originate in other pelvic organs and involve the ovary secondarily. In fact, it has been discovered that high-grade serous carcinomas (HGSCs) may arise from precursor epithelial lesions in the distal fimbriated end of the fallopian tube, whereas endometrioid and clear cell carcinomas originate from ovarian endometriosis. The relative importance of the fallopian tube mucosa compared with the ovarian surface epithelium (mesothelium) in the genesis of high-grade serous ovarian cancers is still a subject of debate. On the other hand, it can be argued that tumors arising in endometriosis are ultimately of endometrial origin. Thus, in some cases, the term ovarian cancer may not be accurate, and it has been suggested that it should be replaced with the terms pelvic or peritoneal cancer. However, given the confusion that might follow in the literature, we think it best to keep the term ovarian cancer until the various possible origins of these diseases are known better.
Some epithelial ovarian tumors, particularly those in the serous category, may be exophytic, intracystic (endophytic), or both. The presence of exophytic growth is indicated by the word ‘surface,’ which is added to the designation, i.e., ‘serous surface papillary adenocarcinoma.’ In contrast, cystic tumors with intracystic growth are variably called ‘serous adenocarcinoma’ or described by the prefix cyst – before their name, i.e., ‘serous cystadenocarcinoma.’ Most surface epithelial–stromal neoplasms are predominantly epithelial, with only a minor component derived from the ovarian stroma. When gland-forming tumors have a predominant stromal component, the terms ‘adenofibroma’ and ‘cystadenofibroma’ (if grossly visible cysts are present) are used.
However, the subdivision of epithelial–stromal tumors that is most important from a clinical viewpoint is their classification into benign, borderline, and carcinoma forms, because it generally correlates with prognosis. This is done according to the amount of tumor cell proliferation, the degree of nuclear atypia, and the presence or absence of stromal invasion.
Borderline Tumors
Borderline tumors (also designated as tumors of low malignant potential) show histologic and cytologic features that are intermediate between those of clearly benign and clearly malignant tumors of the same cell type(s). They exhibit epithelial proliferation greater than that seen in their benign counterparts but an absence of destructive invasion of the stroma or confluent growth, and are associated with a much better prognosis, stage for stage, than that of ovarian carcinomas. Even in the absence of invasiveness within the ovary, borderline tumors of the serous type can either implant on peritoneal surfaces or be associated with independent foci of primary serous peritoneal neoplasia, and, rarely, invasion of the underlying tissues occurs in both circumstances. Exceptionally, tumors of borderline malignancy spread via lymphatics and blood vessels. In addition, a small number of these tumors are combined with or transform over time into obviously invasive carcinomas. Although favorable in the majority of cases, the biologic behavior of the borderline tumors differs from that of the obviously benign tumors of the same cell type(s). Therefore, the designation ‘tumors of borderline malignancy’ should be kept. Alternative terms such as proliferating , atypical , and atypical proliferative are misleading because they do not imply the malignant potential of a small but significant number of these tumors and discourage complete surgical staging and follow-up of the patients. Subdivision of the borderline group into benign and malignant, based on the presence of micropapillary architecture, is artificial since tumors with or without micropapillary pattern may rarely be associated with invasive peritoneal implants and poor outcome. Although the word ‘borderline’ may suggest uncertainty, it accurately describes the ambiguous histologic and biologic features of these neoplasms and remains the most appropriate term. Accordingly, it has been recommended by the World Health Organization (WHO) for the last four decades.
The distinction between borderline tumors and carcinomas is one of the most common problems in ovarian tumor pathology, yet the literature on borderline tumors is confusing, particularly with regard to their diagnostic features and treatment. Although the WHO has recommended the presence or absence of ‘obvious invasion’ of the stroma, most practitioners do not require obvious stromal invasion for the diagnosis of carcinoma if the epithelial cells are malignant cytologically. To promote terminology agreement, the WHO has proposed such tumors be classified as borderline with intraepithelial carcinoma . The extent of the carcinomatous epithelium should be noted in the pathology report.
Carcinomas
Malignant epithelial tumors (carcinomas) are the most common ovarian cancers, accounting for 90% of cases. Although traditionally referred to as a single entity, ovarian cancer is not a homogeneous disease but rather a group of diseases, each with different morphology and biologic behavior. Currently, based on histopathology, immunohistochemistry, and molecular genetic analysis, at least five main types of ovarian carcinomas are identified: HGSCs (70%), endometrioid carcinomas (10%), clear cell carcinomas (10%), mucinous carcinomas (3 %), and low-grade serous carcinomas (LGSCs; <5%). These tumors account for 98% of ovarian carcinomas, can be reproducibly diagnosed by light microscopy, and are inherently different diseases, as indicated by differences in epidemiologic and genetic risk factors, precursor lesions, patterns of spread, molecular events during oncogenesis, response to chemotherapy, and prognosis. The fact that one tumor type (HGSCs) accounts for over two-thirds of cases does not justify classifying ovarian carcinomas into only two pathogenetic types, lumping together the other four (endometrioid, clear cell, mucinous, and LGSCs) as ‘type I carcinomas.’ In fact, the latter tumors are clinically, morphologically, and molecularly distinct diseases that individually bear resemblance neither to HGSCs nor to each other. Thus, classifying ovarian carcinomas into only two pathogenetic types (‘types I and II’) appears to be artificial and simplistic.
Borderline Tumors
Borderline tumors (also designated as tumors of low malignant potential) show histologic and cytologic features that are intermediate between those of clearly benign and clearly malignant tumors of the same cell type(s). They exhibit epithelial proliferation greater than that seen in their benign counterparts but an absence of destructive invasion of the stroma or confluent growth, and are associated with a much better prognosis, stage for stage, than that of ovarian carcinomas. Even in the absence of invasiveness within the ovary, borderline tumors of the serous type can either implant on peritoneal surfaces or be associated with independent foci of primary serous peritoneal neoplasia, and, rarely, invasion of the underlying tissues occurs in both circumstances. Exceptionally, tumors of borderline malignancy spread via lymphatics and blood vessels. In addition, a small number of these tumors are combined with or transform over time into obviously invasive carcinomas. Although favorable in the majority of cases, the biologic behavior of the borderline tumors differs from that of the obviously benign tumors of the same cell type(s). Therefore, the designation ‘tumors of borderline malignancy’ should be kept. Alternative terms such as proliferating , atypical , and atypical proliferative are misleading because they do not imply the malignant potential of a small but significant number of these tumors and discourage complete surgical staging and follow-up of the patients. Subdivision of the borderline group into benign and malignant, based on the presence of micropapillary architecture, is artificial since tumors with or without micropapillary pattern may rarely be associated with invasive peritoneal implants and poor outcome. Although the word ‘borderline’ may suggest uncertainty, it accurately describes the ambiguous histologic and biologic features of these neoplasms and remains the most appropriate term. Accordingly, it has been recommended by the World Health Organization (WHO) for the last four decades.
The distinction between borderline tumors and carcinomas is one of the most common problems in ovarian tumor pathology, yet the literature on borderline tumors is confusing, particularly with regard to their diagnostic features and treatment. Although the WHO has recommended the presence or absence of ‘obvious invasion’ of the stroma, most practitioners do not require obvious stromal invasion for the diagnosis of carcinoma if the epithelial cells are malignant cytologically. To promote terminology agreement, the WHO has proposed such tumors be classified as borderline with intraepithelial carcinoma . The extent of the carcinomatous epithelium should be noted in the pathology report.
Carcinomas
Malignant epithelial tumors (carcinomas) are the most common ovarian cancers, accounting for 90% of cases. Although traditionally referred to as a single entity, ovarian cancer is not a homogeneous disease but rather a group of diseases, each with different morphology and biologic behavior. Currently, based on histopathology, immunohistochemistry, and molecular genetic analysis, at least five main types of ovarian carcinomas are identified: HGSCs (70%), endometrioid carcinomas (10%), clear cell carcinomas (10%), mucinous carcinomas (3 %), and low-grade serous carcinomas (LGSCs; <5%). These tumors account for 98% of ovarian carcinomas, can be reproducibly diagnosed by light microscopy, and are inherently different diseases, as indicated by differences in epidemiologic and genetic risk factors, precursor lesions, patterns of spread, molecular events during oncogenesis, response to chemotherapy, and prognosis. The fact that one tumor type (HGSCs) accounts for over two-thirds of cases does not justify classifying ovarian carcinomas into only two pathogenetic types, lumping together the other four (endometrioid, clear cell, mucinous, and LGSCs) as ‘type I carcinomas.’ In fact, the latter tumors are clinically, morphologically, and molecularly distinct diseases that individually bear resemblance neither to HGSCs nor to each other. Thus, classifying ovarian carcinomas into only two pathogenetic types (‘types I and II’) appears to be artificial and simplistic.
Benign and Borderline Serous Tumors
Serous tumors in both their benign and borderline forms show ciliated epithelial cells and other cell types resembling those of the fallopian tube. Although the less differentiated tumors lose the cytologic features of tubal epithelium, they usually exhibit distinctive patterns of growth, including fine and complex papillary structures with prominent cellular budding and glands with irregular slit-like lumens and solid sheets. Psammoma bodies and larger calcific deposits are usually found. Serous tumors, whether benign, borderline, or malignant, tend to be uniform throughout a given specimen, usually lacking the admixtures of these three subtypes so common in mucinous tumors. The Systematized Nomenclature of Medicine (SNOMED) classification of serous tumors is as follows:
- •
Benign serous tumors
- •
Serous adenofibroma, cystadenofibroma
- •
Serous surface papilloma
- •
Serous cystadenoma
- •
Serous papillary cystadenoma
- •
- •
Serous borderline tumor
- •
Serous carcinoma
- •
Epidemiology.
Benign Serous Tumors
Clinical Features
Benign serous tumors may occur at any age but are most common during the fifth decade. Usually, the tumors are asymptomatic and discovered incidentally during ultrasound investigation of another gynecologic disorder.
Macroscopic Features
Benign serous tumors are usually endophytic (cystadenomas), but may be exophytic (surface papillomas) or both. Serous cystadenomas are usually unilocular ( Figure 25.1 ) but may be multilocular, and have thin-walled cyst(s) ( Figure 25.2 ) filled with watery or occasionally thin mucinous fluid. They average about 10 cm in size. Customarily, a benign serous tumor is diagnosed only if the lesion is >1 cm in diameter. The cyst(s) may have a smooth inner surface or have polypoid excrescences ( Figure 25.3 ), which are firm if their stroma is fibrous and soft if it is edematous. Serous cystadenomas are bilateral in approximately 10–20% of cases ( Figure 25.2 ).
Serous surface papillomas appear as polypoid excrescences on the outer surface of the ovary. The surface tumors are often accompanied by underlying cystic components. Serous adenofibromas are hard, white, predominantly fibromatous tumors with small glands or cysts that contain clear fluid. When bilateral, their firm consistency may suggest the diagnosis of cancer.
Microscopic Features
The cysts and polypoid excrescences of benign serous tumors are typically lined by pseudostratified epithelium similar to that of the fallopian tube, which is frequently ciliated ( Figure 25.4 ). Tumors lined entirely by nonciliated cuboidal or columnar epithelium that resembles ovarian surface epithelium are also generally included in the serous category despite their indifferent appearance. The epithelial cells of benign serous tumors may secrete mucin, but, when it is present, it is usually confined to the lumens of cysts and apical portion of the cytoplasm of the epithelial cells. Mitoses are rare. There is no nuclear atypia. Psammoma bodies are infrequent (10%). Papillae when found are composed almost entirely of stroma ( Figure 25.5 ), which may be collagenous or markedly edematous. In adenofibromas and cystadenofibromas, glands and cysts are scattered within a predominantly fibromatous stroma. Otherwise typical benign cystic or surface serous tumors containing minor foci (<10%) consistent with borderline neoplasia (cell stratification and nuclear atypia) are kept in the benign category for clinical purposes.
Differential Diagnosis
Serous cystadenomas may be confused with rare examples of cystic struma ovarii ( Chapter 29 ), but the latter exhibits at least minor foci of identifiable thyroid follicles in the cyst’s outer wall. Rare rete cystadenomas may simulate serous cystadenomas but, in addition to their hilar location, often show a thick layer of smooth muscle and are lined by cells lacking cilia. Common peritoneal cysts, lined by mesothelial cells, often mimic serous cystadenomas. Serous surface papillary adenofibromas should be distinguished from the common small surface stromal proliferations sometimes seen in adult women. The latter lesions are typically multifocal, have a simple cuboidal epithelial lining, and do not form a mass lesion.
Macroscopic Features
Benign serous tumors are usually endophytic (cystadenomas), but may be exophytic (surface papillomas) or both. Serous cystadenomas are usually unilocular ( Figure 25.1 ) but may be multilocular, and have thin-walled cyst(s) ( Figure 25.2 ) filled with watery or occasionally thin mucinous fluid. They average about 10 cm in size. Customarily, a benign serous tumor is diagnosed only if the lesion is >1 cm in diameter. The cyst(s) may have a smooth inner surface or have polypoid excrescences ( Figure 25.3 ), which are firm if their stroma is fibrous and soft if it is edematous. Serous cystadenomas are bilateral in approximately 10–20% of cases ( Figure 25.2 ).
Serous surface papillomas appear as polypoid excrescences on the outer surface of the ovary. The surface tumors are often accompanied by underlying cystic components. Serous adenofibromas are hard, white, predominantly fibromatous tumors with small glands or cysts that contain clear fluid. When bilateral, their firm consistency may suggest the diagnosis of cancer.
Microscopic Features
The cysts and polypoid excrescences of benign serous tumors are typically lined by pseudostratified epithelium similar to that of the fallopian tube, which is frequently ciliated ( Figure 25.4 ). Tumors lined entirely by nonciliated cuboidal or columnar epithelium that resembles ovarian surface epithelium are also generally included in the serous category despite their indifferent appearance. The epithelial cells of benign serous tumors may secrete mucin, but, when it is present, it is usually confined to the lumens of cysts and apical portion of the cytoplasm of the epithelial cells. Mitoses are rare. There is no nuclear atypia. Psammoma bodies are infrequent (10%). Papillae when found are composed almost entirely of stroma ( Figure 25.5 ), which may be collagenous or markedly edematous. In adenofibromas and cystadenofibromas, glands and cysts are scattered within a predominantly fibromatous stroma. Otherwise typical benign cystic or surface serous tumors containing minor foci (<10%) consistent with borderline neoplasia (cell stratification and nuclear atypia) are kept in the benign category for clinical purposes.
Differential Diagnosis
Serous cystadenomas may be confused with rare examples of cystic struma ovarii ( Chapter 29 ), but the latter exhibits at least minor foci of identifiable thyroid follicles in the cyst’s outer wall. Rare rete cystadenomas may simulate serous cystadenomas but, in addition to their hilar location, often show a thick layer of smooth muscle and are lined by cells lacking cilia. Common peritoneal cysts, lined by mesothelial cells, often mimic serous cystadenomas. Serous surface papillary adenofibromas should be distinguished from the common small surface stromal proliferations sometimes seen in adult women. The latter lesions are typically multifocal, have a simple cuboidal epithelial lining, and do not form a mass lesion.
Serous Borderline Tumors
Clinical Features
Serous borderline tumors (SBTs) account for one-fourth to one-third of the non-benign serous tumors. They are most common in the fourth and fifth decades, with an average patient age of 42 years. Although often asymptomatic, the tumor may sometimes present with abdominal enlargement and pain due to rupture or torsion. Approximately 70% are confined to one or both ovaries (stage I) at the time of diagnosis; the remaining tumors have spread within the pelvis (stage II) or upper abdomen (stage III). Only rare cases have extended beyond the abdomen (stage IV) at the time of presentation. One-third of the stage I tumors are bilateral.
Macroscopic Features
Serous borderline tumors have one or more cysts that are lined by polypoid excrescences and closely packed finer papillae (endophytic growth) ( Figure 25.6 ). The cysts contain thick mucinous fluid, which does not necessarily indicate the mucinous nature of the tumor. The polypoid excrescences contain fibrous or edematous stroma. Serous borderline cystadenofibromas are firm, white, fibromatous tumors with a cystic component of variable size. In almost 50% of SBTs the papillary growth covers the outer surface of the ovary (serous surface papillary borderline tumor). Frequently, both exophytic and endophytic papillary components are present ( Figure 25.7 ). The typical gross features of serous carcinomas, such as friability, hemorrhage, and necrosis, are not seen in SBTs.
Microscopic Features
SBTs show stromal polypoid excrescences, glands, and papillae lined by stratified cuboidal to columnar epithelial cells and ciliated cells resembling those of the fallopian tube ( Figures 25.8–25.11 ). Larger hobnail cells with ample eosinophilic cytoplasm ( Figure 25.12 ) and mesothelial-like cells may also be present. Eosinophilic cells are increased when microinvasion has occurred. The lining epithelial cells show mild to moderate nuclear atypia (rarely severe) and mitoses are rare; therefore, the histologic pattern usually differs from that of carcinoma in situ of other organs. Psammoma bodies are found in about one-fourth of cases. The three most important diagnostic features are (1) arborizing papillae that form increasingly smaller branches ending in clusters of epithelial cells that appear to be detached from the stroma (hierarchical pattern of branching), (2) varying degree of nuclear atypia, and (3) absence of ‘frank’ stromal invasion or solid sheets of tumor with a cribriform pattern.
Several changes within SBTs may result in overdiagnosis of carcinoma. The pathologist’s main problems while dealing with SBTs are listed below.
- •
Pseudoinvasion
- •
Self-implantation
- •
Mesothelial cell hyperplasia
- •
Micropapillary pattern
- •
Microinvasion
- •
Peritoneal implants
- •
SBT in lymph nodes
- •
SBT of the peritoneum.
Pseudoinvasion, Autoimplants, and Mesothelial Cell Hyperplasia
Serous borderline tumors often show a relatively complex proliferation of glands and papillae ( Figure 25.10 ). The glands often invaginate into the stroma and, particularly when sectioned tangentially, may appear as if they have invaded the stroma. This pseudoinvasion differs, however, from the destructive stromal invasion of a carcinoma. The stroma about the pseudoinvasion is similar to the stroma elsewhere and the glands have an orderly distribution. In carcinoma, it forms a desmoplastic stroma typically and the neoplastic glands have a more disorderly arrangement. Exceptionally, an SBT implants on itself and may exhibit focal desmoplastic stroma. Autoimplants are sharply circumscribed desmoplastic plaques usually on the outer surface but occasionally on the inner (cystic) surface of the tumor, resembling noninvasive desmoplastic implants on extraovarian peritoneum ( Figure 25.13 ). Although the term autoimplant suggests that this lesion might arise by detachment of exophytic SBT and subsequent reattachment to itself, occurrence of this phenomenon has not been confirmed. Autoimplants seem to lack clinical significance. Sometimes, SBTs are partly covered by fibrous adhesions associated with mesothelial hyperplasia , which may be erroneously interpreted as surface borderline tumor or even carcinoma. The mesothelial cells, however, lack nuclear atypia and are typically arranged in a linear fashion ( Figure 25.14 ).
Micropapillary Pattern
Rarely, SBTs show an exuberant micropapillary (or focal cribriform) proliferation without destructive stromal invasion ( Figure 25.15 ). There is a filigree pattern of highly complex micropapillae growing in a nonhierarchical fashion from fibrous stalks ( Figure 25.16 ) composed of stratified nonciliated cuboidal cells with high nuclear to cytoplasmic ratio. Nuclear atypia is only mild with occasional higher grade nuclei present ( Figure 25.17 ). Mitotic figures are rare and no abnormal forms are seen. In contrast to the typical SBT, which usually shows variable degrees of cell proliferation and nuclear atypia, the micropapillary or cribriform SBT ( Figure 25.18 ) exhibits a homogeneous and marked degree of cell proliferation and uniform mild nuclear atypia ( Figure 25.17 ). Most micropapillary tumors contain areas of typical SBT, indicating the former’s probable origin. When the micropapillary architecture exceeds 5 mm in greatest dimension, it has been recommended to segregate these tumors from the less proliferative SBTs, claiming that they are likely to progress to invasive carcinoma, particularly as invasive peritoneal implants. Subsequently, it was required that the ‘micropapillae’ should be five times as long as wide ( Figure 25.16 ).
It has been shown that bilaterality, ovarian surface growth, and advanced stage (noninvasive implants) are more common features of micropapillary SBTs than of typical SBTs, but a strong association of the former tumors with invasive implants and poor outcome has been inconsistent. Although in four studies micropapillary SBTs were more frequently associated with invasive implants, micropapillary architecture did not have a significant adverse effect on survival when controlled for implant type. Stage II–III micropapillary SBTs with noninvasive peritoneal implants had the same favorable prognosis as typical SBTs ( Table 25.1 ); thus, from a prognostic viewpoint, micropapillary SBTs are closer to typical SBTs than to carcinomas. However, although largely noninvasive, serous tumors with prominent micropapillary architecture may contain areas of frank stromal invasion ( Figure 25.19 ); therefore, extensive sampling is indicated in these cases.
| Burks et al. | Seidman and Kurman | Eichhorn et al. | Prat and de Nictolis | Longacre et al. | Park et al. | Avril et al. | |
|---|---|---|---|---|---|---|---|
| Total cases | 10 | 11 | 19 | 13 | 23 | 33 | 2 |
| Cases with invasive implants | 6 b | 10 b | 3 | 1 | 5 | 5 | 0 |
| Chemotherapy a | 8/10 (80%) | 5/9 (56%) | 5/17 (29%) | 9/12 (75%) | — | 10/33 (30%) | 1 |
| 5 year follow-up | 5 | 10 | 10 | 9 | 23 | — | 2 |
| Alive, no tumor | 0 | 3 | 6 | 8 | 17 | 29 | 2 |
| Alive, recurrent tumor in contralateral oophorectomy | 0 | 0 | 2 | 1 | 0 | 3 | 0 |
| Alive, progressive residual/recurrent tumor | 2 b | 3 b | 0 | 1 c | 0 | 0 | 0 |
| Died of tumor | 2 | 4 | 2 | 0 | 5 | 1 | 0 |
| Died of unrelated causes | 1 | 0 | 0 | 0 | 1 | 0 | 0 |
| Follow-up <5 years | 5 | 1 | 5 | 3 | 0 | — | 0 |
| Lost to follow-up | 0 | 0 | 3 | 1 | 0 | 0 | 0 |
| Recent cases | 0 | 0 | 1 | 0 | 0 | 0 | 0 |
a Number of patients/number of patients with postoperative treatment information (%).
b Including ‘implants with micropapillary’ pattern but without recognizable invasion of underlying tissue.
The micropapillary pattern ( Figures 25.16 and 25.17 ), almost always associated with typical SBT, most likely represents a degree of epithelial proliferation intermediate between SBT and LGSCs. By analogy with cervical neoplasia, the difference between typical SBT and micropapillary SBT is like between cervical intraepithelial neoplasia (CIN) 2 and CIN 3. Although the micropapillary pattern in combination with other clinical and pathologic features (i.e., advanced stage, invasive implants, and microinvasion) may be associated with increased risk of disease progression, taken individually, micropapillarity is not a specific predictor of adverse prognosis. Almost all patients dying of recurrent tumor had invasive peritoneal implants, which are the key feature associated with a poor prognosis.
Microinvasion
Approximately 10% of otherwise typical SBTs contain one or more discrete foci of stromal microinvasion, which are made up of single epithelial cells or small clusters of such cells with abundant eosinophilic cytoplasm ( Figures 25.20 and 25.21 ). Often, the cell nests appear surrounded by clefts that separate the epithelium from the stroma ( Figures 25.22 and 25.23 ). These microscopic foci (arbitrarily defined as not exceeding 10 mm 2 in area, or 3 mm maximum linear dimension) are typically unassociated with a significant stromal reaction and are easily overlooked on routinely stained sections. Cytokeratin (CK) stains help to visualize the invasive epithelial cells ( Figure 25.24 ). Occasionally, foci of lymphatic invasion may be present. Several patients with these tumors have been pregnant at the time of diagnosis.
SBTs with microinvasion are associated with a higher frequency of bilaterality, exophytic ovarian surface growth, and advanced stage than typical SBTs lacking microinvasion, yet the rate of invasive implants is similar in both tumor groups. At least five studies ( Table 25.2 ) have confirmed that SBTs with or without microinvasion have a similar prognosis. However, two recent reports suggested that microinvasion may represent a risk factor for disease progression that is independent of stage and implant status. SBTs with microinvasion ( Figures 25.20 and 25.21 ) should be distinguished from SBTs with microinvasive carcinoma. The latter shows ‘destructive’ invasion of the stroma by malignant-appearing cells and desmoplastic response ( Figure 25.22 ). The lesion resembles an invasive peritoneal implant or LGSC. Lymphatic invasion is common and can be demonstrated by D2-40 immunoreaction.
| Tavassoli | Bell and Scully | Nayar et al. | Kennedy and Hart | Prat and de Nictolis | McKenney et al. | |
|---|---|---|---|---|---|---|
| Total cases | 18 | 21 | 7 | 4 | 20 | 60 |
| Advanced stage | 6 | 2 | — | 1 | 9 | 26 |
| Invasive implants/distant metastasis | 1 | 1 | 0 | 0 | 1 | 2 |
| Lymph node metastasis | 1 | 1 | 0 | 0 | 2 | 7 |
| Chemotherapy | 4 | 3 | — | 3 | 4 | 9 |
| Radiotherapy | 1 | 0 | — | 0 | 1 | 4 |
| 5 year follow-up (or until death) | 5 | — | 6 | 3 | 11 | — |
| Alive, no tumor | 4 | — | 6 | 3 | 9 | 41 |
| Alive, progressive residual/recurrent tumor | 0 | 1 | 0 | 0 | 1 | 2 |
| Died of tumor | 1 | 0 | 0 | 0 | 1 * | 7 |
| Follow-up <5 years | 13 | — | 1 | 1 | 5 | — |
| Lost to follow-up | 0 | 4 | 0 | 0 | 4 | 10 |
Immunohistochemistry
SBTs show positive immunoreaction for CK7, epithelial membrane antigen (EMA), and OC-125. WT1 is weakly positive. Some cases show focal positivity for CK20. Most cases express estrogen and progesterone receptors (ERs/PRs).
p53 is usually negative but can be focally positive; however, focal nuclear reaction for p53 does not correlate with TP53 mutations.
Peritoneal Implants
One controversial aspect of SBTs is their association in ~30–40% of cases with peritoneal implants. Implants are found more frequently in patients with tumors that have an exophytic component than in those that do not. They rarely present as bulky disease ( Figure 25.25 ), and, in most cases, are either microscopic or small macroscopic (≤1–2 cm) lesions. Their histologic appearance may vary greatly, ranging from foci of benign glandular epithelium resembling endosalpingiosis, to noninvasive papillae, plaques, or nodules of borderline epithelium and stroma, to invasive implants resembling an LGSC. Endosalpingiosis, typified by glands, cysts, and occasionally papillae with psammoma bodies, is a benign peritoneal lesion frequently associated with ovarian SBTs ( Figure 25.26 ). It may rarely be the substrate for the development of peritoneal SBTs or carcinomas, but its presence should not change the stage of a synchronous ovarian SBT.
The peritoneal implants of SBTs have been classified histologically into noninvasive and invasive types with the former further subdivided into epithelial and desmoplastic subtypes. The epithelial subtype of noninvasive implants shows papillary proliferations of atypical epithelial cells resembling those of the ovarian SBT ( Figure 25.27 ); they are typically present on the surface of the peritoneum ( Figure 25.28 ) or in smoothly contoured subperitoneal invaginations and exhibit little or no stromal reaction. In contrast, the desmoplastic subtype of noninvasive implant is largely composed of a stromal proliferation, which is plastered upon serosal surfaces or invaginations between lobules of omental fat ( Figures 25.25B–D , 25.29 , and 25.30 ). The stromal reaction surpasses quantitatively the epithelial component of the implant. In late lesions, small glands and papillae lined by atypical serous cells as well as psammoma bodies are entrapped by dense fibroblastic tissue that is often infiltrated by acute and chronic inflammatory cells. Early implants show necrosis with surface fibrin deposition and hemorrhage ( Figure 25.30 ). Invasive implants, which represent approximately 12% of the cases, manifest a disorderly infiltration of normal tissues, such as the omentum; in contrast to the well-defined limits of the desmoplastic implants, the invasive implants exhibit irregular borders ( Figure 25.31 ). They usually show a greater epithelial cell population ( Figure 25.32 ) and resemble histologically a low-grade serous adenocarcinoma ( Figure 25.33 ); marked cytologic atypia may be present. Implants should be sampled extensively since noninvasive and invasive implants may coexist. Also, some implants of serous carcinomas may be noninvasive and simulate the desmoplastic noninvasive implants of SBTs. The former implants, however, usually contain highly atypical epithelial cells.
At least 11 studies of SBTs with peritoneal implants ( Table 25.3 ) have clearly demonstrated that separation of the implants into invasive and noninvasive subtypes carries important prognostic implications. The rare tumors that were fatal were mostly those with invasive peritoneal implants.
| Noninvasive Implants | Invasive Implants | |
|---|---|---|
| McCaughey et al. | 2/13 | 4/5 |
| Bell DA et al. | 3/50 | 5/6 a |
| de Nictolis et al. | 0/10 | 4/9 |
| Kennedy and Hart | 1/25 b | 0/1 c |
| Seidman and Kurman | 1/51 | 2/3 |
| Gershenson et al. | 6/73 | 6/39 |
| Eichhorn et al. | 0/30 | 2/3 |
| Bell KA et al. | 2/29 d | 6/31 e |
| Prat and de Nictolis. | 0/34 | 3/6 f |
| Longacre et al. | 2/85 | 5/14 g |
| Park et al. | 0/14 | 1/5 |
| Avril et al. | 0/8 | 0/0 |
| Total | 17/422 | 38/122 |
a One patient died of tumor subsequent to her report.
b Transformation into serous carcinoma.
d One patient was alive with progressive disease (AWPD).
e Thirteen patients were AWPD.
f The other three patients were AWPD.
g One patient was AWPD and one died of leukemia but had persistent tumor.
Despite the poor prognosis associated with invasive implants when they actually invade the underlying tissues, their diagnostic criteria remain problematic. Truly invasive implants are rare lesions; i.e., from a population exceeding 3 million people, only six patients with stage III SBTs with invasive implants were encountered during a 17 year period. Some investigators, however, have reported higher rates of invasive implants (up to 35% in one series) but a lack of correlation between their presence and adverse prognosis. Such a discrepancy could be explained by the inclusion of desmoplastic noninvasive implants in the category of invasive implants.
Although the use of more liberal histologic criteria (i.e., solid epithelial nests surrounded by clefts and micropapillary architecture) increased the proportion of invasive implants in a recently reported series to more than 50%, association with death due to progressive disease was less significant than by three other groups using more stringent diagnostic rules. In fact, one of the latter studies confirmed that solid epithelial nests surrounded by clefts and micropapillary architecture are found more often in invasive than noninvasive implants; in that series, obvious destructive invasion was the only feature of the peritoneal implants specifically associated with poor outcome. As stated earlier, the presence of epithelial cell clusters surrounded by clefts or halos is a characteristic feature of stromal microinvasion (which is associated with a favorable prognosis; Figure 25.23 ). In contrast, an abundant epithelial component ( Figures 25.31 and 25.32 ), whether micropapillary or not, favors the diagnosis of invasion. Noninvasive desmoplastic implants typically contain an excess of granulation tissue-like stroma and minimal epithelial component in the form of small cellular nests or glands ( Figures 25.25B–D , 25.29 , and 25.30 ). The finding of single isolated cells in the stroma is not enough histologic evidence to diagnose the implant as invasive; such isolated cells can also be found in approximately one-third of the noninvasive desmoplastic implants.
The characteristic features of noninvasive epithelial implants, noninvasive desmoplastic implants, and invasive implants can be summarized as follows.
Noninvasive Epithelial Implants
- •
Papillary proliferations resembling ovarian SBT
- •
On peritoneal surface or in smoothly contoured invaginations
- •
Well-defined margins
- •
Little or no stromal reaction
- •
Psammoma bodies
Noninvasive Desmoplastic Implants
- •
Stromal proliferation plastered on serosal surfaces or between lobules of omental fat
- •
Well-defined margins
- •
Excess of granulation tissue-like stroma (>50%) and minimal epithelial component
- •
Hemorrhage, necrosis (early changes), fibrosis, and calcification (late changes)
Invasive Implants
- •
Disorderly distributed glands destructively invading normal tissue
- •
Poorly defined irregular margins
- •
Abundant epithelial component
- •
Dense collagenous stroma without significant inflammation
- •
Resemble LGSCs
- •
Aneuploidy
Immunohistochemical studies have shown that invasive implants show loss of calretinin-positive mesothelial cells and CD34-positive fibroblasts, which are preserved in noninvasive implants. Smooth muscle actin-positive myofibroblasts are present in both invasive and noninvasive implants.
Somatic Genetics
Recent molecular genetic studies indicate that SBTs and HGSCs represent separate pathogenetic entities. SBTs with and without micropapillary pattern frequently display B-RAF/K-RAS mutations but rarely mutant TP53 . In contrast, B-RAF/K-RAS mutations are rare in HGSCs, but TP53 mutations occur in approximately 60% of cases (see later). B-RAF/K-RAS mutations also occur in the epithelium of cystadenomas adjacent to SBTs, suggesting that they precede the development of the latter tumors. K-RAS mutations are also present in the glandular inclusions (endosalpingiosis) in the peritoneum and lymph nodes of patients with ovarian SBT. RNA expression profiles of ovarian serous tumors have disclosed that the mitogenic MAP kinase pathway is activated in SBTs; however, activation of downstream genes involved in extracellular matrix degradation is absent, suggesting an uncoupling of both events. Gene expression profiling has also shown prominent expression of wild p53, cyclin-dependent kinase inhibitor p21 , and other p53-modulated genes in SBTs, suggesting that this signaling pathway may play an important role in the distinct phenotype associated with this tumor. Although the profiles for invasive LGSCs did not contain the enhanced p53 signaling activity observed in SBTs, the former tumors were aligned with the SBTs instead of HGSCs. These findings provide additional proof that typical SBTs, SBTs with micropapillary pattern, and invasive LGSCs are closely related neoplasms and different from usual HGSCs. The rare invasive LGSCs with micropapillary pattern ( Figure 25.19 ; which also have B-RAF/K-RAS mutations) probably represent SBTs with infiltrative stromal invasion greater than microinvasion.
Clonality studies have examined whether multiple, synchronous, or metachronous SBTs (found at different sites in the abdominal cavity) arise as a result of spread from a single ovarian site, or whether such deposits are polyclonal representing independent primary tumors. Two studies based on X chromosome inactivation analysis support a multifocal origin of bilateral and advanced SBTs. However, tumor-related changes may interfere with X chromosome inactivation and this method poorly assesses clonality. In contrast, loss of heterozygosity (LOH) is an irreversible genetic event acquired during tumorigenesis. Its weakness, however, is that the absence of informative markers and the failure to detect LOH underestimates the frequency of clonality.
Recently, clonality of invasive and noninvasive peritoneal implants and lymph node deposits has been investigated by genome-wide allelotyping and B-RAF/K-RAS mutation analysis in 10 patients using 23 microsatellite markers. Concordant LOH for 1–5 microsatellite markers was found in all five informative cases in all tumor sites. In addition, identical K-RAS and B-RAF mutations were detected in four and two cases, respectively ( Figure 25.34 ). These findings strongly support the metastatic nature of noninvasive and invasive peritoneal implants and lymph node deposits.
Histogenesis
Ovarian SBTs have traditionally been thought to arise from the ovarian surface epithelium (mesothelium), from the müllerian epithelium within the ovary (surface epithelial inclusion cysts), or from endosalpingiosis. Recently, however, and based on the more frequent finding of papillary tubal hyperplasia and secretory cell outgrowths lacking PAX2 expression in the fallopian tubes of patients with SBTs compared to controls, a direct origin from the tubal epithelium with secondary ovarian involvement has been suggested.
Differential Diagnosis
SBTs are only rarely confused with other neoplasms. Their distinction from LGSCs has already been discussed. SBTs differ from endocervical-like (müllerian) mucinous borderline tumors by the lack of both intracellular mucin and a neutrophilic infiltrate (see later). SBTs may also occur as a component of mixed müllerian borderline tumors. Retiform Sertoli–Leydig cell tumors ( Chapter 28 ) can have areas indistinguishable from SBTs on routine staining, but occur in a younger age group, have other distinctive features such as long ribbons of sex cord-type cells, and other patterns of Sertoli–Leydig cell tumors. These tumors, like most sex cord tumors, are often reactive for α-inhibin.
Biologic Behavior
The overall outcome of SBTs is extremely favorable. For patients with stage I tumors, the risk of recurrence or the development of a second SBT has been estimated to be only 5–10%. The tumors rarely recur beyond 10 years.
Most SBTs maintain their microscopic features and indolent clinical behavior, and usually do not progress over the years to frankly invasive carcinoma. However, rare cases of late recurrence in the form of carcinoma have been described. In a more recent study, progression of SBT to LGSC occurred in 6–7% of patients late in the course of the disease and was associated with poor prognosis. Almost 80% occurred in patients without prior history of invasive implants. Of eight cases of SBT with malignant behavior, all had ovarian surface involvement, three exhibited focal stromal invasion greater than microinvasion, and only two were SBT with micropapillary pattern. Although exceptional, malignant transformation of SBT to HGSC may also occur early or following multiple recurrences ( Figure 25.35 ).
Although exophytic SBTs are more often associated with synchronous peritoneal implants than intracystic SBTs, the presence of ovarian surface involvement does not necessarily predict progression of disease. As previously stated, most fatal SBTs have had invasive implants.
Clinical Features
Serous borderline tumors (SBTs) account for one-fourth to one-third of the non-benign serous tumors. They are most common in the fourth and fifth decades, with an average patient age of 42 years. Although often asymptomatic, the tumor may sometimes present with abdominal enlargement and pain due to rupture or torsion. Approximately 70% are confined to one or both ovaries (stage I) at the time of diagnosis; the remaining tumors have spread within the pelvis (stage II) or upper abdomen (stage III). Only rare cases have extended beyond the abdomen (stage IV) at the time of presentation. One-third of the stage I tumors are bilateral.
Macroscopic Features
Serous borderline tumors have one or more cysts that are lined by polypoid excrescences and closely packed finer papillae (endophytic growth) ( Figure 25.6 ). The cysts contain thick mucinous fluid, which does not necessarily indicate the mucinous nature of the tumor. The polypoid excrescences contain fibrous or edematous stroma. Serous borderline cystadenofibromas are firm, white, fibromatous tumors with a cystic component of variable size. In almost 50% of SBTs the papillary growth covers the outer surface of the ovary (serous surface papillary borderline tumor). Frequently, both exophytic and endophytic papillary components are present ( Figure 25.7 ). The typical gross features of serous carcinomas, such as friability, hemorrhage, and necrosis, are not seen in SBTs.
Microscopic Features
SBTs show stromal polypoid excrescences, glands, and papillae lined by stratified cuboidal to columnar epithelial cells and ciliated cells resembling those of the fallopian tube ( Figures 25.8–25.11 ). Larger hobnail cells with ample eosinophilic cytoplasm ( Figure 25.12 ) and mesothelial-like cells may also be present. Eosinophilic cells are increased when microinvasion has occurred. The lining epithelial cells show mild to moderate nuclear atypia (rarely severe) and mitoses are rare; therefore, the histologic pattern usually differs from that of carcinoma in situ of other organs. Psammoma bodies are found in about one-fourth of cases. The three most important diagnostic features are (1) arborizing papillae that form increasingly smaller branches ending in clusters of epithelial cells that appear to be detached from the stroma (hierarchical pattern of branching), (2) varying degree of nuclear atypia, and (3) absence of ‘frank’ stromal invasion or solid sheets of tumor with a cribriform pattern.
Several changes within SBTs may result in overdiagnosis of carcinoma. The pathologist’s main problems while dealing with SBTs are listed below.
- •
Pseudoinvasion
- •
Self-implantation
- •
Mesothelial cell hyperplasia
- •
Micropapillary pattern
- •
Microinvasion
- •
Peritoneal implants
- •
SBT in lymph nodes
- •
SBT of the peritoneum.
Pseudoinvasion, Autoimplants, and Mesothelial Cell Hyperplasia
Serous borderline tumors often show a relatively complex proliferation of glands and papillae ( Figure 25.10 ). The glands often invaginate into the stroma and, particularly when sectioned tangentially, may appear as if they have invaded the stroma. This pseudoinvasion differs, however, from the destructive stromal invasion of a carcinoma. The stroma about the pseudoinvasion is similar to the stroma elsewhere and the glands have an orderly distribution. In carcinoma, it forms a desmoplastic stroma typically and the neoplastic glands have a more disorderly arrangement. Exceptionally, an SBT implants on itself and may exhibit focal desmoplastic stroma. Autoimplants are sharply circumscribed desmoplastic plaques usually on the outer surface but occasionally on the inner (cystic) surface of the tumor, resembling noninvasive desmoplastic implants on extraovarian peritoneum ( Figure 25.13 ). Although the term autoimplant suggests that this lesion might arise by detachment of exophytic SBT and subsequent reattachment to itself, occurrence of this phenomenon has not been confirmed. Autoimplants seem to lack clinical significance. Sometimes, SBTs are partly covered by fibrous adhesions associated with mesothelial hyperplasia , which may be erroneously interpreted as surface borderline tumor or even carcinoma. The mesothelial cells, however, lack nuclear atypia and are typically arranged in a linear fashion ( Figure 25.14 ).
Micropapillary Pattern
Rarely, SBTs show an exuberant micropapillary (or focal cribriform) proliferation without destructive stromal invasion ( Figure 25.15 ). There is a filigree pattern of highly complex micropapillae growing in a nonhierarchical fashion from fibrous stalks ( Figure 25.16 ) composed of stratified nonciliated cuboidal cells with high nuclear to cytoplasmic ratio. Nuclear atypia is only mild with occasional higher grade nuclei present ( Figure 25.17 ). Mitotic figures are rare and no abnormal forms are seen. In contrast to the typical SBT, which usually shows variable degrees of cell proliferation and nuclear atypia, the micropapillary or cribriform SBT ( Figure 25.18 ) exhibits a homogeneous and marked degree of cell proliferation and uniform mild nuclear atypia ( Figure 25.17 ). Most micropapillary tumors contain areas of typical SBT, indicating the former’s probable origin. When the micropapillary architecture exceeds 5 mm in greatest dimension, it has been recommended to segregate these tumors from the less proliferative SBTs, claiming that they are likely to progress to invasive carcinoma, particularly as invasive peritoneal implants. Subsequently, it was required that the ‘micropapillae’ should be five times as long as wide ( Figure 25.16 ).
It has been shown that bilaterality, ovarian surface growth, and advanced stage (noninvasive implants) are more common features of micropapillary SBTs than of typical SBTs, but a strong association of the former tumors with invasive implants and poor outcome has been inconsistent. Although in four studies micropapillary SBTs were more frequently associated with invasive implants, micropapillary architecture did not have a significant adverse effect on survival when controlled for implant type. Stage II–III micropapillary SBTs with noninvasive peritoneal implants had the same favorable prognosis as typical SBTs ( Table 25.1 ); thus, from a prognostic viewpoint, micropapillary SBTs are closer to typical SBTs than to carcinomas. However, although largely noninvasive, serous tumors with prominent micropapillary architecture may contain areas of frank stromal invasion ( Figure 25.19 ); therefore, extensive sampling is indicated in these cases.
| Burks et al. | Seidman and Kurman | Eichhorn et al. | Prat and de Nictolis | Longacre et al. | Park et al. | Avril et al. | |
|---|---|---|---|---|---|---|---|
| Total cases | 10 | 11 | 19 | 13 | 23 | 33 | 2 |
| Cases with invasive implants | 6 b | 10 b | 3 | 1 | 5 | 5 | 0 |
| Chemotherapy a | 8/10 (80%) | 5/9 (56%) | 5/17 (29%) | 9/12 (75%) | — | 10/33 (30%) | 1 |
| 5 year follow-up | 5 | 10 | 10 | 9 | 23 | — | 2 |
| Alive, no tumor | 0 | 3 | 6 | 8 | 17 | 29 | 2 |
| Alive, recurrent tumor in contralateral oophorectomy | 0 | 0 | 2 | 1 | 0 | 3 | 0 |
| Alive, progressive residual/recurrent tumor | 2 b | 3 b | 0 | 1 c | 0 | 0 | 0 |
| Died of tumor | 2 | 4 | 2 | 0 | 5 | 1 | 0 |
| Died of unrelated causes | 1 | 0 | 0 | 0 | 1 | 0 | 0 |
| Follow-up <5 years | 5 | 1 | 5 | 3 | 0 | — | 0 |
| Lost to follow-up | 0 | 0 | 3 | 1 | 0 | 0 | 0 |
| Recent cases | 0 | 0 | 1 | 0 | 0 | 0 | 0 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
