—  ARTHUR PURDY STOUT SOCIETY OF SURGICAL PATHOLOGISTS   —

Ovarian Tumors in Children and Young Women: Selected Important Issues in Differential Diagnosis


Robert H. Young
Massachusetts General Hospital
Boston, MA


I begin this presentation by considering the frequency of the various categories of ovarian tumors (and tumors within each category) during each of the first three decades and when indicated within parts of individual decades. These comments are aimed at highlighting one important aspect of gonadal pathology, namely, that although the diagnosis ultimately rests with the light microscopic findings, it strengthens the approach to the differential diagnosis to be aware of the statistical likelihood of any diagnosis. I will then consider three tumors of the young that have all been highlighted in the last three decades. I will conclude by noting briefly the issues associated with ovarian masses encountered during pregnancy. Two comprehensive reviews of ovarian tumors in childhood and adolescence were undertaken in recent times by Dr. Ernest Lack and colleagues [1, 2] . I have selected from his comprehensive list of references a selected group that I believe represents a good synopsis of the literature on this topic [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22] and have extended the time frame of coverage until the end of the third decade. The reason for this is that, although there are obviously exceptions, it is only when someone is over 30 that they start to progressively be at risk for most of the common ovarian neoplasms, both benign and malignant.

There are marked differences in the frequency of ovarian neoplasms in the first three decades compared with later years and then, to a lesser degree, within each of the first three decades. In these decades there is a much higher percentage of germ cell and sex cord-stromal tumors because of the lesser frequency of surface epithelial neoplasms and great rarity of metastatic tumors. This is borne out by the over 50 years experience at the Children's Hospital in Boston (even allowing for that consideration being only of the first two decades); 70% of the tumors were germ cell and 13% sex cord-stromal [1]. One-third of the germ cell tumors were in the primitive category, a much greater percentage than seen beyond 30 years. The surface epithelial tumors accounted for 16% of the neoplasms. Tumor-like lesions accounted for almost one-third of ovarian masses. This exceeds the figure for later years because of the relative high frequency of follicle cysts in the first two decades, which has two time periods, the neonatal and perimenarchal, when these lesions are common.

The first decade has a unique aspect related to tumor-like lesions, namely, the contribution to the numbers of ovarian masses provided by the occurrence, just noted, of follicle cysts in the neonatal period as a response to maternal stimulation. Sixteen of the 76 tumor-like lesions in the Children's Hospital series were in this category. These cysts are larger than follicle cysts seen later (except for those related to pregnancy), averaging 8.3 cm. in the Children's Hospital series. Their size makes them prone to torsion with significant symptomatology as a result.

Other pertinent comments about the first decade are as follows. Dermoid cysts are rare in the first two years but become progressively more common thereafter. Primitive germ cell tumors are seen but are much commoner in the next two decades. As an example of this in the Children's Hospital Series, only one of 8 patients with dysgerminoma was in the first decade [1]. Surface epithelial tumors of any type are exceptionally rare. Granulosa cell tumors, particularly of the juvenile type are seen occasionally, although it should be remembered that adult granulosa cell tumors are also seen in the young. Sertoli-Leydig cell tumors are seen occasionally and the retiform subtype accounts for a greater percentage than in any other decade except the second. The sex cord tumor with annular tubules, both Peutz-Jeghers Syndrome associated and unassociated with the syndrome, is rarely encountered in the first decade, being much commoner in the second and third [23], indeed being as common in those two decades as at any other time. Two, to date, unique, ovarian sex cord tumors associated with Peutz-Jeghers Syndrome, each responsible for sexual precocity, occurred in the first decade [24]. The small cell carcinoma of hypercalcemic type is rare before 10 years, only three examples being known, the youngest a 14-month old [25]. Miscellaneous other tumors such as primaries sarcomas are seen but rarely. Within the family of sarcomas, as one might expect, the rhabdomyosarcomas of embryonal and alveolar type, particularly the former, which have a skew to young individuals in general, are seen occasionally in the first three decades. Of the thirteen ovarian rhabdomyosarcomas reported by Nielsen et al [26], one was in a 7-year old, one in a 10-year old, and one in a 14-year old. Three of the remaining patients were in their 20's. Two of the tumors were alveolar, the remainder, embryonal. One infantile hemangioendothelioma has been reported in a neonate [27] but vascular sarcomas become commoner in the third decade. One tumor with a predilection for children, the neuroblastoma, accounts for the majority of the rare cases of ovarian metastasis in children and may also be seen in young adults. In the largest series of cases looking at spread to the ovary in children, eight of the 14 tumors were neuroblastomas [28]. Seven were primary in the adrenal gland and one in the posterior mediastinum. The other tumors were accounted for by three rhabodmyosarcomas, primary in the ethmoid sinus, right occipital region and left side respectively, and individual examples of Ewing's sarcoma, rhabdoid tumor of the kidney and a carcinoid tumor of the lung. Follicle cysts are rare after the first six months and other tumor-like lesions are rare at any time in this decade. A corpus luteum cyst does rarely occur, however, in neonates [29].

The second decade is noteworthy for the acceleration of the frequency of all the primitive germ cell tumors as shown by the average age of occurrence of all of them being circa 18 - 19 years. One rare but important monodermal teratoma, the primitive neuroectodermal tumor [30], is seen much more often in the second than in the first decade. The commoner monodermal tumors, struma ovarii and carcinoid, are uncommon in the first three decades. Surface epithelial tumors remain uncommon although measurably more frequent than in the first decade. Virtually the only cell types seen are serous and mucinous. In the Children's Hospital series serous slightly outnumbered mucinous. In our consultation material mucinous tumors are much more numerous. This is accounted for presumably by the fat that the serous tumors are almost always benign whether in the mucinous group, tumors in the borderline or low-grade carcinoma categories are encountered with some frequency and are more likely to provoke a consultative opinion. At least one mucinous tumor in a teenager had foci of anaplastic carcinoma [1]. We have seen one low-grade endometrioid carcinoma in a teenager [2] but this and other surface epithelial carcinomas in teenagers are a great rarity. The small cell carcinoma of hypercalcemic type begins to be seen with an accelerated pace once children enter the second decade and this tumor is only somewhat less common in the second, compared to the third decade. Metastatic carcinomas which are virtually unheard of in the first decade are seen occasionally in the teenage years with well documented examples of Krukenberg tumor and even metastatic colonic carcinoma (we have seen one of the latter in a 12-year-old) being documented. Lymphoma and leukemia involving the ovary is occasionally seen in the second decade and even before having no great predilection for any of the first three decades. Ovarian involvement by the intra-abdominal desmoplastic small round cell tumor with divergent differentiation is commonest in the teenage years, the three original cases of this type reported by us being in the mid teens [31] and subsequent experience has borne out a peak in this decade. As this is a time when the small cell carcinoma of hypercalcemic type is being seen with increased frequency, it is important that these two tumors be carefully distinguished.

One significant aspect of the third decade, particularly when compared to the prior two, is the appreciably greater frequency of surface epithelial tumors, particularly as the decade moves along. For example, serous borderline tumors, which are exceptionally rare before 20 are seen with some frequency in the third decade although there pace of occurrence accelerates even more after the age of 30. Of all tumors with any malignant potential in the surface epithelial category, serous borderline tumors and mucinous borderline tumors account for the majority encountered in the third decade. Mucinous carcinomas are occasionally seen, but less frequently than borderline tumors and serous carcinomas remain exceptionally rare in the third decade. Indeed, the extent to which they are dwarfed by serous borderline tumors is worthy of comment and in this age group one should always favor a borderline tumor rather than serous carcinoma unless evidence for the latter is incontrovertible. Germ cell tumors of all types are seen with slightly greater frequency than in the second decade. The exception to that comment is secondary malignant change in teratomas, and monodermal teratomas, which remain rare until patients pass beyond 30 years. However, the validity of the comment just made not withstanding, one can actually rarely see this phenomenon in patients younger than 30, although almost never younger than 20. Three of 19 cases of squamous cell carcinoma arising in dermoid reported by Pins et al [32] were in patients of 21, 25, and 30 years of age. One other patient in his series with a squamous cell carcinoma and associated endometriosis was a 29-year-old and another patient without an association with either a dermoid cyst or endometriosis was a 27-year-old.

The third decade sees a peak in the frequency of Sertoli-Leydig cell tumors, which occur at an average age of 25 years [33]. Seventy-five percent of all Sertoli-Leydig cell tumors occur in the first three decades. Granulosa cell tumors of both adult and juvenile types continue to be seen with some frequency in this decade. Steroid cell tumors are also seen in this decade and even in the first; about 7% of these tumors occur in prepubertal girls [34]. Pure stromal tumors, except the sclerosing stromal tumor remain rare; the latter has an average age of 27 years and is commoner in the third decade than any other. Enigmatically, the rare primary angiosarcoma of the ovary appears to be commoner in the third decade [35]. Three of 11 cases of granulocytic sarcoma of the female genital tract reported relatively recently involved one or both ovaries of patients from 13 to 30 years of age [36]. These neoplasms must be distinguished in the conventional manner from malignant lymphoma and other neoplasms.

I shall now consider individually three important and morphologically interesting tumors of the young.

Juvenile Granulosa Cell Tumor
Dr. Robert Scully became alerted to this neoplasm in the early 1970's when he noted that in young females granulosa cell tumors often had alarming microscopic features. Although these atypical features are occasionally observed in tumors from older women, their strong predilection for young females led to his designating them juvenile granulosa cell tumors (JGCTs).

On gross examination, the tumors have a spectrum of appearances similar to those of the adult granulosa cell tumor [37, 38, 39] . Microscopic examination typically reveals diffuse sheets or nodules of cells with abundant eosinophilic cytoplasm interrupted to varying degrees by follicles. The follicles may be round to oval but more typically vary in shape and size and often contain mucicarminophilic, eosinophilic or basophilic material. The nuclei of the granulosa cells appear more immature than those in the adult type of tumor. They are more hyperchromatic and frequently exhibit considerable mitoic activity; nuclear grooves are rare. Up to 15% of the tumors, show marked nuclear atypia [37].

The major differences between the adult and juvenile forms of granulosa cell tumor are cytologic, specifically the hyperchromatic, mitotically active nuclei without grooves in the juvenile form. Another important difference is the typical luteinization of the cells in the juvenile tumor. In adult granulosa cell tumors, one may see luteinized cells, but they are usually a minor component and only very rare tumors in this category are extensively luteinized [40]. Finally, the nature of the follicles in the JGCT differs in many cases from those in the adult form of neoplasm in that they exhibit much more variability in size and shape.

The differential diagnosis of the JGCT includes various other neoplasms. As malignant germ cell tumors are common in the age group in which the JGCT is most often encountered, and may even be functioning, the immature nuclei of the JGCT may lead to the misdiagnosis of a yolk sac tumor or embryonal carcinoma. However, the follicle formation which is seen at least focally in the great majority of JGCTs is diagnostic with one caveat. Cysts in the polyvesicular vitelline variant of yolk sac tumor can potentially be misconstrued as cystic follicles of a JGCT. Cellular stroma between the cysts of the polyvesicular tumor and foci of typical yolk sac tumor both may be helpful. Paradoxically, the extreme nuclear atypia of some JGCTs exceeds that of a yolk sac tumor.

The JGCT is sometimes misinterpreted as a thecoma, because of the absence or rarity of follicles in some specimens, the usual presence of abundant cytoplasm in the neoplastic cells and the occasional predominance of theca cells. Thorough sampling to demonstrate follicles and the performance of reticulum stains to help recognize the granulosa cell nature of at least some of the tumor cells are important in establishing the correct diagnosis. Also, thecomas lack significant mitotic activity in almost all the cases, occur before 30 years of age in less than 10% of the cases, and very rarely occur in children. An important differential diagnosis is with the small cell carcinoma of hypercalcemic type, an issue discussed below.

Rarely, the JGCT is confused with carcinomas of surface epithelial type. Tumors with marked atypia may be misinterpreted as undifferentiated carcinomas if foci of follicle differentiation are overlooked. In some cases, the tubulocystic variant of clear cell carcinoma is suggested when the follicles are lined by cells with features of hobnail cells. The absence of other patterns of clear cell carcinoma, the young age of the patient, the presence of follicles and focal areas of more typical JGCT should help resolve this problem. In some JGCTs with prominent follicles a pseudopapillary appearance may be encountered leading to a superficial resemblance to transitional cell carcinoma. Small follicles in the lining at the base of the pseudopapillae exclude transitional cell carcinoma. A metastatic tumor that may be confused with a JGCT is malignant melanoma because in some of the latter follicle-like spaces in a neoplasm composed of cells with abundant eosinophilic cytoplasm imparts a striking resemblance to a JGCT. We have not seen the basophilic mucicarminophilic fluid that is often present in the follicles of JGCTs in the follicle-like spaces of metastatic melanoma. When the diagnosis of JGCT is being considered in a patient over 20 years of age, the diagnosis of malignant melanoma should be excluded. Immunohistochemical staining for both S-100 protein and HMB-45, and electron microscopic examination, should establish the diagnosis if other clinical and routine pathologic features do not.

Small Cell Carcinoma of Ovary, Hypercalemic Type
It is appropriate to discuss this tumor next because it is often follicle forming, like the juvenile granulosa cell tumor. This distinctive ovarian cancer is the most common form of undifferentiated ovarian carcinoma in women under 40 years of age. It was recognized by Dr. Scully in the early 1970's and first presented in detail by him in his first fascicle and in 1982 "officially" described [41]. It was the unusual occurrence in young women of a small cell carcinoma that was associated with paraneoplastic hypercalcemia that led it to being recognized as a distinct entity. The patients have ranged from 14 months to 43 (mean, 24) years of age [42]. Most patients present with signs and symptoms related to an abdominal or pelvic mass, but rarely the clinical presentation is related to the hypercalcemia. Occasional familial cases have been encountered. In the largest series in the literature the stage of the tumor was IA in 33%, 1B in 1%, 1C in 16%, stage II in 5%, stage III in 43%, and stage IV in 1% [42]. This carcinoma has a dismal prognosis. The overall survival rate is approximately 16%; the corresponding figures are 33% for Stage 1A tumors and 7% for tumors >stage IA [42]. Favorable prognostic factors in the largest study included an age >30 years, a normal preoperative serum calcium, bilateral oophorectomy, tumors <10 cm, no large cell component, and the administration of adjuvant radiotherapy [42]. No chemotherapy has proven to be of significant consistent benefit in the management of patients with this highly malignant tumor.

The tumors are almost always unilateral, usually large, solid, soft and white. This appearance is similar to that of dysgerminoma and lymphoma and may contribute to confusion with these neoplasms as considered below. Necrosis and hemorrhage are conspicuous in many small cell carcinomas. On microscopic examination, the most common pattern is a more or less diffuse arrangement of small, closely packed cells with scant cytoplasm but evidence of their epithelial nature is provided by the focal presence of small nests, cords, and clusters of cells. An important feature that is seen in about 80% of the tumors is follicles that vary from small to large but are more often the latter. They are usually round to oval and contain eosiniphilic fluid. The tumor cells have small nuclei containing single nucleoli and, despite the aggressive nature of this tumor, are relatively monotonous in appearance with the exception noted below. Mitotic figures are numerous.

Variant features may be seen which complicate the interpretation in many cases. The commonest of these is the presence in about half of the cases of a component of large cells with abundant eosinophilic cytoplasm, a pattern for convenience referred to as the "large cell variant." This pattern is quite distinctive in many cases because of the frequent presence of a prominent stroma which often is myxoid. Overall the small cell carcinoma has relatively unimpressive fibrous stroma in most cases but it is occasionally focally relatively prominent. The large cells may have eccentric nuclei and dense globular cytoplasm. Nucleoli are usually more prominent in foci of the large cell variant. Rarely some cells in the small cell carcinoma have clear cytoplasm. This may suggest a dysgerminoma but the architectural and cytologic difference between the two tumors are so striking that simply considering small cell carcinoma should enable it to be diagnosed when its features are indeed present. Gross similarities mentioned earlier may enhance confusion, however, and it should also be remembered that the dysgerminoma may also be associated with hypercalcemia [43]. Another intriguing finding in about 10% of these tumors is minor foci of mucinous epithelium that typically stands out sharply from the background sea of small cells. Occasionally, however, the mucinous epithelium is less conspicuous and more cytologically atypical, and may merge with the small cells. Rarely signet-ring type cells are seen.

Electron microscopical examination has so far failed to reveal any specific features to identify the cell type of the tumor. McMahon and Hart [44] found the most consistent and striking finding to be the presence of abundant dilated rough endoplasmic reticulum, forming large vesicles filled with homogeneous, granular (proteinaceous) material of variable density. They considered these ultrastructural features to be of diagnostic value in the differential diagnosis of this tumor. Ultrastructural examination in some cases has shown numerous whorls of microfilaments to account for the globular cytoplasm. Immunohistochemical studies have also not helped identify the cell of origin but have simply confirmed an epithelial nature [45, 46, 47] . Evidence for a sex cord derivation is lacking.

This tumor is often confused with a granulosa cell tumor of either adult or juvenile types. However, the cells of the small cell carcinoma do not have the characteristic pale grooved nuclei of the adult granulosa cell tumor and the mitotic rate in the small cell carcinoma far exceeds that encountered in adult granulosa cell tumors. Distinction from the juvenile form of granulosa cell tumor is generally easy because the cells of the small cell carcinoma usually lack the abundant eosinophilic cytoplasm that is an almost invariable feature of the cells of the juvenile granulosa cell tumor. Even in cases of small cell carcinoma in which there are cells with abundant eosinophilic cytoplasm, distinction can usually be made because such cells are most often a focal finding and, in addition, they differ in appearance from the cells of the juvenile granulosa cell tumor, because of the dense, sometimes globular cytoplasm; the cells in a JGCT rarely have this appearance. When they do it is helpful that in cases of small cell carcinoma containing cells with abundant cytoplasm, there is generally a more disorderly architecture than seen in the juvenile granulosa cell tumor. Finally, with regard to distinction from both forms of granulosa cell tumor, it is often helpful that the small cell carcinoma has spread beyond the ovary at presentation, something which would be unusual for either variant of granulosa cell tumor.

Because their peak frequency is at about the same time, they may be grossly indistinguishable, and both may cause hypercalcemia, the dysgerminoma and small cell carcinoma may be confused as noted earlier. Although the small cell carcinoma may have prominent fibrous stroma it differs in quality generally from the delicate fibrous septa of the dysgerminoma and although, like any tumor, there may be some non-specific inflammatory cell infiltrate, the classic lymphocytic infiltrate that tends to hug the septa, typical of dysgerminoma, is lacking in the small cell carcinoma. The squared off nuclei of dysgerminoma are not seen in the small cell carcinoma, although this feature, and others for that matter, are not as overt in cases of dysgerminoma that are poorly preserved. If necessary, the typical immunostaining profile of dysgerminoma for placental-like alkaline phosphatase and OCT 3/4 will be helpful.

The differential diagnosis of the small cell carcinoma is also with other small cell malignant tumors that may involve the ovaries. Only those that characteristically occur in the same age range as the small cell carcinoma, or if not, have in our experience caused an interesting morphologic dilemma will be considered here. These include primary primitive neuroectodermal tumors, metastatic neuroblastoma, malignant lymphoma and leukemia, metastatic round cell sarcomas, metastatic malignant melanoma, and the intra-abdominal desmoplastic small round cell tumor. Even large cell carcinomas can be in the differential in cases with a significant component of the so-called large cell variant and this issue will also be briefly considered.

Neuroblastoma and primitive neuroectodermal tumors occur in the same age group as the small cell carcinoma. In cases of neuroblastoma, the presence of Homer-Wright rosettes is diagnostic, although such structures may be few in number. Appreciation of fibrillar material may also be important in the diagnosis in these cases and is even more crucial in cases of the primitive neuroectodermal tumor. In the latter tumor the cells are even smaller than in the small cell carcinoma and lack the distinctive follicles and other patterns. Rarely immunostains may be called upon to assist in the differential [47].

Malignant lymphoma or leukemia occasionally presents as a unilateral or bilateral ovarian mass and enters the differential diagnosis of the small cell carcinoma. The cells of these neoplasms may grow in a variety of patterns, such as insular, cord-like, and alveolar. The distinctive cytologic features of the malignant lymphoid or leukemic cells, however, differ from those of the cells of the small cell carcinoma, and the frequent follicle-like spaces of the latter are helpful in this differential. Immunostains will obviously provide aid if necessary.

Melanoma metastatic to the ovary is rarely composed of small cells with scanty cytoplasm and follicle-like spaces that produces a pattern similar to that of small cell carcinoma [48]. These tumors, however, also usually contain nevoid aggregates of cells that suggest melanoma. The history of a primary malignant melanoma and immunohistochemical staining of the ovarian tumor for S-100 protein and HMB-45 may be important in establishing the diagnosis and immunohistochemistry was decisive in one case of primary malignant melanoma of the ovary we have seen in which the cells were small and in which there were follicle-like spaces.

The most recently described small cell tumor that may involve the ovary and potentially be confused with the hypercalcemic form of small cell carcinoma is the intra-abdominal desmoplastic small round cell tumor (DSRCT) [31]. The ovarian involvement is much more often bilateral in cases of DSRCT. Although the hypercalcemic tumor may have an insular pattern, it is generally absent or only focal, whereas in the DSRCT discrete nests of tumor cells are characteristic. Additionally, the latter tumor has, as its name implies, a typically prominent desmoplastic stroma, which is either absent, or only a focal finding in the hypercalcemic small cell carcinoma. The nests in the DSRCT often have a basaloid appearance, which is not a feature of the hypercalcemic tumor and its cells, although small, are more reminiscent of those of the pulmonary form of small cell carcinoma than the hypercalcemic tumor. Finally, if any difficulty in the distinction of the DSRCT and hypercalcemic small cell carcinoma remains on the basis of the examination of routinely stained sections, immunohistochemical examination will be diagnostic as the staining of the DSRCT for desmin is not a feature of the hypercalcemic tumor.

The possibility of a small cell carcinoma may be raised in cases of metastatic alveolar rhabdomyosarcoma, because the follicle-like spaces of the small cell carcinoma may be simulated by cystic changes in the alveolar spaces of the rhabdomyosarcoma [49]. However, the distinctive and prominent alveolar pattern of rhabdomyosarcoma is not a feature of small cell carcinoma. The distinctive giant cells that are seen in some rhabdomyosarcomas have not been encountered in the hypercalcemic small cell carcinoma. One Ewing's sarcoma metastatic was initially thought possibly to represent a small cell carcinoma [50].

Finally, as alluded to earlier, the differential diagnosis of the small cell carcinoma, enigmatically, may actually be that of a large cell carcinoma when the tumor is predominantly the so-called large cell variant. The evaluation of large cell malignant tumors of the ovary is a topic on to itself, which can only be briefly considered here. It can bring up such diverse malignant lesions as malignant steroid cell tumor, malignant melanoma, anaplastic carcinoma arising in a mucinous tumor, anaplastic carcinoma arising in a dermoid, and of course, the neoplasm of interest to us here. Although there is a great temptation to rush to immuno in this circumstance, it is important to emphasize that good old-fashioned thorough sampling is just as likely to be helpful, if not more so. Certainly finding of focal evidence of a dermoid cyst or an underlying mucinous cystic neoplasm may be very helpful in capturing the diagnosis of two of the entities just mentioned. Although mucinous epithelium may be seen in association with a small cell carcinoma, it is minor foci that has not to date simulated the picture of a parent mucinous cystic tumor. Also, although the possibility exists of a small cell carcinoma being entirely of the large cell variant, I do not believe we have seen to date a case that did not at least somewhere have minor foci of the typical small cell morphology, which rules out all the lesions just considered except melanoma and that is certainly one area where admittedly immunohistochemistry is most appropriately something that should be called upon.

Retiform Sertoli-Leydig Cell Tumor
Sertoli-Leydig cell tumors (SLCTs) are classified into five major categories: well differentiated, of intermediate differentiation, poorly differentiated, heterologous and retiform. The last variant is of interest to us here because of its particular tendency to occur in the young (average age 15 years} and its intriguing morphology [51, 52, 53] . They account for about 10% of SLCTs. Clinically they are less often androgenic than other variants of SLCT, having these signs or symptoms only about 20% of the time. That observation is of relevance in that androgenic manifestations tend to have the diagnosis of Sertoli-Leydig cell tumor come to mind in a oung person; conversely their absence and unusual morphology may lead to the diagnosis being overlooked. Retiform tumors are often soft and spongy, cystic with edematous intraluminal polypoid excrescences or combinations thereof.

Retiform SLCTs are so named because they are characterized microscopically by growth patterns that simulate those of the rete testis. The basic pattern is an irregular network of elongated, often slit-like tubules and cysts which often contain papillae. The papillae may be short and rounded or blunt, often containing hyalinized cores, or larger with fibrous or edematous cores. Cysts may be markedly dilated with eosinophilic secretion imparting a struma-like appearance in some cases. The tubules are usually lined by a single layer of cuboidal epithelial cells with round to oval nuclei although stratification is conspicuous in some cases. The cytoplasm is typically scanty and mitotic activity variable, but sometimes marked. Similar cells typically line the papillae and cysts, but large cysts may be lined by flattened cells. The stromal component varies from moderately cellular fibrous tissue (which is occasionally focally hyalinized) to markedly edematous (accounting for the soft, spongy consistency) to very cellular, immature mesenchymal tissue which may show heterologous differentiation.

Retiform SLCTs may be misinterpreted as a variety of other tumors. Because of the young age of the patient and presence of papillae in many cases the diagnosis of yolk sac tumor is often considered. It should be remembered that like many ovarian tumors the retiform SLCT may contain hyaline bodies. Resemblance to a serous borderline tumor may be imparted by the presence of small papillary clusters in the cyst lumens. The clefts, papillae and a complex branching pattern associated with cellular stratification and atypicality may suggest a serous adenocarcinoma. The admixture of retiform tubules with immature mesenchymal tissue which may show heterologous differentiation can suggest the diagnosis of a malignant mixed mesodermal tumor. The presence in most cases of typical foci of SLCT, although sometimes very small in amount, and other clinical and pathologic features of the tumors should enable the diagnosis of a retiform SLCT to be established. Any pattern of SLCT may co-exist with the retiform pattern but typically the retiform foci merge with long, thick ribbons of immature cells consistent with Sertoli cells. When a tumor is entirely retiform, knowledge of the existence of this distinctive subtype of SLCT is essential in enabling the pathologist avoid a serious error in diagnosis.

Issues Related to Ovarian Tumors Discovered During Pregnancy
As the era from menarche through the end of the third decade accounts for a significant component of the reproductive era in females, it is obvious that during this time, by happenstance, a number of ovarian neoplasms will develop, or at least be discovered, in patients who are pregnant. Furthermore, pregnancy itself is responsible occasionally for the development of some specific tumor-like lesions. In this brief summary I will comment upon four issues, namely 1. Tumor-like lesions associated with pregnancy; 2. Alterations in sex cord stromal tumors due to pregnancy which may hinder their microscopic recognition; 3. Non-specific changes in ovarian neoplasms associated with pregnancy; and 4. Ovarian tumors with functioning stroma related to pregnancy.

When a patient is known to be pregnant that a mass lesion of the ovary, either cystic or solid, might be non-neoplastic should always be considered assuming appropriate gross and microscopic findings. Otherwise completely benign lesions may result in overly aggressive treatment with unfortunate consequences upon future fertility and increased morbidity. Two lesions may mimic cystic neoplasms, hyperreactio lutealis (multiple luteinized follicle cysts) and the large solitary luteinized follicle cyst of pregnancy and the puerperium [54]. It is helpful that the first lesion is bilateral in contrast to many cystic neoplasms that might be in the differential diagnosis and certainly when a frozen section is done on a cystic lesion of the ovaries from a pregnant patient, careful search for the distinctive features of follicle cysts and other changes of the intervening tissue that are seen in hyperreactio lutealis, such as edema, should be sought. When a massive unilocular smooth walled cyst is encountered in pregnancy, the large solitary follicle cyst should be considered. In contrast to most cystic granulosa cell tumors it has a lining of cells with copious eosinophilic cytoplasm and rather characteristically, focal bizarre nuclei. The solid tumor-like lesion of pregnancy is the pregnancy luteoma and it is a time honored adage that the diagnosis of steroid cell tumor should not be made in pregnancy unless a pregnancy luteoma has been ruled out. It is helpful in many cases that the pregnancy lutetoma is bilateral and multi-nodular, the nodules typically being beefy red in contrast to the yellow appearance of many steroid cell tumors, although should be acknowledged that some steroid cell tumors may be beefy and red. A well-known trap is the brisk mitotic activity of many pregnancy luteomas. The fourth tumor-like lesion is the incidentally discovered microscopic process known as granulosa cell tumorlets. They are likely to cause clinical mischief inasmuch even if they are considered neoplasms, erroneously, it would be tiny microscopic ones of no clinical consequence [55].

As the early reproductive era sees a peak in incidence of Sertoli-Leydig cell tumors and the majority of juvenile granulosa cell tumors and a sizable number of adult granulosa cell tumors occur in this age range, it is obvious that occasional patients with one of these neoplasms will encountered while she is pregnant. A pitfall is that these neoplasms may show considerable intercellular edema, or be extensively luteinized, both features which may obscure their characteristic morphologic features and render them, on average, more difficult to diagnose than similar neoplasms encountered outside of the setting of pregnancy [56]. Perhaps the most specific trap is that the edema may produce a pattern which simulates, at least to some degree, the reticular pattern of yolk sac tumor and we have certainly seen cases of sex cord stromal tumors erroneously considered to be yolk sac tumors during pregnancy, for this reason. Although thorough sampling is important at any time, it is particularly so when evaluating a perplexing mass lesion from a pregnant patient because sometimes only minor areas will show diagnostic features of one or another form of sex cord stromal tumor, making it evident that that is the diagnosis.

Any ovarian neoplasm from a pregnant patient is more likely to undergo hemorrhagic infaraction and rupture than outside the setting of pregnancy. This may result in a clinically dramatic presentation and also, because of the infarction, serves to make diagnostic foci often limited and is a further contribution to diagnostic difficulty. Many references on ovarian tumors in pregnancy are available in the paper that is Reference 56 of this handout.

Finally, yet another mischievous feature of ovarian tumors in pregnancy is their much greater tendency to exhibit stromal luteinization, which is often striking, and may be associated with dramatic androgenic manifestations. This may result in a clinician being convinced the patient has the most well-known androgenic neoplasm of young females, the Sertoli-Leydig cell tumor, and when encountering a tubular pattern in an ovarian tumor of a young virilized female, the pathologist quite reasonably wonders about Sertoli-Leydig cell tumor. However, in some cases of Krukenberg tumor a striking tubular pattern simulates the sertoliform pattern of a Sertoli-Leydig cell tumor and the luteinization of the stroma may be misconstrued as the Leydig cell component of a Sertoli-Leydig cell tumor. A number of the tubular Krukenberg tumors of the ovary described by us some years ago were misdiagnosed as Sertoli-Leydig cell tumors for the reasons just summarized [57].

References

  1. Lack EE, Goldstein DP. Primary ovarian tumors in childhood and adolescence. Current Problems in Obstetrics and Gynecology 7:4-90, 1984.

  2. Lack EE, Young RH, Scully RE. Pathology of ovarian neoplasms in childhood and adolescence. Pathol Annu 27(pt.2): 281-356, 1992.

  3. Abell MR, Johnson VJ, Holtz F. Ovarian neoplasms in childhood and adolescence. I. Tumors of germ cell origin. Am J Obstet Gynecol 92:1059-1081, 1965.

  4. Adelman S, Benson CD, Hertzler JH. Surgical lesions of the ovary in infancy and childhood. Surg Gynecol Obstet 141:219, 1975.

  5. Bower, RJ, Adkins JC. Surgical ovarian lesions in children. Ann Surg 47:474, 1981.

  6. Breen JL, Maxson WS. Ovarian tumors in children and adolescents. Clin Obstet Gynecol 20:607-623, 1977.

  7. Breen JL, Neubecker RD. Ovarian malignancy in children, with special reference to the germ-cell tumors. Ann NY Acad Sci 142:658, 1967.

  8. Ehren IM, Mahour G, Isaacs J Jr. Benign and malignant ovarian tumors in children and adolescents. A review of 63 cases. Am J Surg 147:339, 1984.

  9. Groeber WR. Ovarian tumors during infancy and childhood. Am J Obst Gynecol 86:1027-1035, 1963.

  10. Jereb B, Golouh R, Havlicek S. Ovarian cancer in children and adolescents: A review of 15 cases. Med Pediatr Oncol 3:339, 1977.

  11. Junaid JA. Ovarian neoplasms in children and adolescents in Ibadan, Nigeria. Cancer 47:610-614, 1981.

  12. La Vecchia C, Morris HB, Draper GJ. Malignant ovarian tumours in childhood in Britain, 1962-78 Br J Cancer 48:363, 1983.

  13. Li FP, Fraumeni JR Jr, Dalager N. Ovarian cancers in the young. Epidemiologic observations. Cancer 32:969-972, 1973.

  14. Lindfors O. Primary ovarian neoplasms in infants and children. A study of 81 cases diagnosed in Finland and Sweden. Ann Chir Cynaecol Fin. 177:1, 1971.

  15. Lucraft HH. Ovarian tumours in children--A review of 40 cases. Clin Radiol 30:279, 1979.

  16. Mahour GH, Woolley MM, Landing BH. Ovarian teratomas in children. A thirty-three-year experience. Am J Surg 132:587-589, 1976.

  17. Moore JG, Schifrin BS, Erez S. Ovarian tumors in infancy, childhood and adolescence. Am J Obst Gynecol 99:913-922, 1967.

  18. Norris HJ, Jensen RD. Relative frequency of ovarian neoplasms in children and adolescents. Cancer. 30:713-719, 1972.

  19. Towne BH, Mahour GH, Woolley MM. Issacs H Jr. Ovarian cysts and tumors in infancy and childhood. J Pediatr Surg 10:311-320, 1975.

  20. Shawis RN, Gohary AEL, Cook RCM. Ovarian cysts and tumours in infancy and childhood. Ann Royal Coll Surg Engl 67:17, 1985.

  21. Szamborski JJ. Ovarian tumors in childhood and adolescence. Histological types and malignancy with reference to the age and menstrual activity of patients. Neoplasma. 25:493, 1978.

  22. Thompson JP, Dockerty MB, Symmonds RE, Hayles AB. Ovarian and paraovarian tumors in infants and children. Am J Obstet Gynecol 97:1059, 1967.

  23. Young RH, Welch WR, Dickersin R, Scully RE. Ovarian sex cord tumor with annular tubules. Review of 74 cases including 27 with Peutz-Jeghers syndrome and four with adenoma malignum of the cervix. Cancer 50:1384-1402, 1982

  24. Young RH, Dickersin GR, Scully RE. A distinctive ovarian sex cord-stromal tumor causing sexual precocity in the Peutz-Jeghers syndrome. Am J Surg Pathol 7:233-243, 1983.

  25. Florell SC, Bruggers CS, Matlak M, Young RH, Lowichik A. Ovarian small cell carcinoma of the hypercalcemic type in a 14 month old: The youngest reported case. Med Pediatr Onc 32:304-307, 1999.

  26. Nielsen GP, Oliva E, Young RH, Rosenberg AE, Prat, J, Scully RE. Primary ovarian rhabdomyosarcoma: A report of 13 cases. Int J Gynecol Pathol 17:113-119, 1998.

  27. Prus D, Rosenberg AE, Blumenfeld A, Udassin R, Ne'eman Z, Young RH, Ariel I. Infantile hemangioendothelioma of the ovary: a monodermal teratoma or a neoplasm of ovarian somatic cells? Am J Surg Pathol 21:1231-1235, 1997.

  28. Young RH, Kozakewich HPW, Scully RE. Metastatic ovarian tumors in children: A report of 14 cases and review of the literature. Int J Gynecol Pathol 12:8-19, 1993.

  29. Miles PA, Penney LL. Corpus luteum formation in the fetus. Obstet Gynecol 61:525-529, 1983.

  30. Kleinman GM, Young RH, Scully RE. Primary neuroectodermal tumors of the ovary. A report of 25 cases. Am J Surg Pathol 17:764-778, 1993.

  31. Young RH, Eichhorn JH, Dickersin GR, Scully RE. Ovarian involvement by the intra-abdominal desmoplastic small round cell tumor with divergent differentiation. A report of three cases. Hum Pathol 23:454-464, 1992.

  32. Pins, MR, Young RH, Daly WJ, Scully RE. Primary squamous cell carcinoma of the ovary. Report of 37 cases. Am J Surg Pathol 20:823-833, 1996.

  33. Young RH, Scully RE. Ovarian Sertoli-Leydig cell tumors. A clinicopathological anaylsis of 207 cases. Am J Surg Pathol 9:543-569, 1985.

  34. Hayes MC, Scully RE. Ovarian steroid cell tumor, not otherwise specified. A clinicopathologic analysis of 63 cases. Am J Surg Pathol 11:835-845, 1987.

  35. Nielsen GP, Young RH, Prat J, Scully, RE. Primary Angiosarcoma of the Ovary. A report of seven cases and review of the literature. Int J Gynecol Pathol 16:378-382, 1997.

  36. Oliva E, Ferry JA, Young RH, Prat J, Srigley RJ, Scully RE. Granulocytic sarcoma of the female genital tract: A clinicopathologic study of 11 cases. Am J Surg Pathol 21:1156-1165, 1997.

  37. Young RH, Dickersin GR, Scully RE. Juvenile granulosa cell tumor. A clinicopathological analysis of 125 cases. Am J Surg Pathol 8:575-596, 1984.

  38. Lack EE, Perez-Atayde AR, Murthy ASK, Goldstein DP, Crigler JF, Vawter GF. Granulosa theca cell tumors in premenarchal girls. A clinical and pathologic study of ten cases. Cancer 48:1846-1854, 1981.

  39. Zaloudek C, Norris HJ. Granulosa tumors of the ovary in children. A clinical and pathologic study of 32 cases. Am J Surg Pathol 6:503-512, 1982.

  40. Young RH, Oliva E, Scully RE. Luteinized adult granulosa cell tumors of the ovary. A report of four cases. Int J Gynecol Pathol 13:302-210, 1994.

  41. Dickersin GR, Kline IW, Scully RE. Small cell carcinoma of the ovary with hypercalcemia. A report of 11 cases. Cancer 49:188-197, 1982.

  42. Young RH, Oliva E, Scully RE. Small cell carcinoma of the ovary, hypercalcemic type: A clinicopathologic analysis of 150 cases. Am J Surg Pathol 18:1102-1116, 1994.

  43. Fleischhacker DS, Young RH. Dysgerminoma of the ovary associated with hypercalcemia. Gynecol Oncol 52:87-90, 1994.

  44. McMahon JT, Hart WR. Ultrastructural analysis of small cell carcinoma of the ovary. Am J Clin Pathol 90:523-529, 1988.

  45. Aguirre P, Thor AD, Scully RE. Ovarian small cell carcinoma - Histogenetic considerations based on immunohistochemical and other findings. Am J Clin Pathol 92:140-149, 1989.

  46. Riopel MA, Perlman PJ, Seidman JD, et al. Inhibin and epithelial membrane antigen immunohistochemistry assist in the diagnosis of sex cord-stromal tumors and provide clues to the histogenesis of hypercalcemic small cell carcinomas. Int J Gynecol Pathol 17:46-53, 1998.

  47. McCluggage WG, Oliva E, Connolly LE, McBride, HA, Young RH. An immunohistochemical analysis of ovarian small cell carcinoma of hypercalcemic type. Int J of Gynecol Pathol 23:330-336, 2004.

  48. Young RH, Scully RE. Malignant melanoma metastatic to the ovary: A clinicopathologic analysis of 20 cases. Am J Surg Pathol 15:849-860, 1991.

  49. Young RH, Scully RE. Alveolar rhabdomyosarcoma metastatic to the ovary. A report of two cases and discussion of the differential diagnosis of small cell malignant tumors of the ovary. Cancer 64:899-904, 1989.

  50. Young RH, Scully RE. Sarcomas metastatic to the ovary: A report of 21 cases. Int J Gynecol Pathol 9:231-252, 1990.

  51. Young RH, Scully RE. Ovarian Sertoli-Leydig cell tumors with a retiform pattern: A problem in histopathologic diagnosis. A report of 25 cases. Am J Surg Pathol 7:755-771, 1983.

  52. Roth LM, Slayton RE, Brady LW, Blessing JA, Johnson G. Retiform differentiation in ovarian Sertoli-Leydig cell tumors. A clinicopathologic study of six cases from a gynecologic oncology group study. Cancer 55:1093-1098, 1985..

  53. Talerman A. Ovarian Sertoli-Leydig cell tumor (androblastoma) with retiform pattern. A clinicopathologic study. Cancer 60:3056-3064, 1989.

  54. Clement PB, Scully RE. Large solitary luteinized follicle cyst of pregnancy and puerperium: A clinicopathological analysis of eight cases. Am J Surg Pathol 4:431-438, 1980.

  55. Clement PB, Young RH, Scully RE. Ovarian granulosa cell proliferations of pregnancy. A report of nine cases. Hum Pathol 19:657-662, 1988.

  56. Young RH, Dudley AG, Scully RE. Granulosa cell, Sertoli-Leydig cell and unclassified sex cord-stromal tumors associated with pregnancy: a clinicopathological analysis of thirty-six cases. Gynecol Oncol 18:181-205, 1984.

  57. Bullon A, Arseneau J, Prat J, Young RH, Scully RE. Tubular Krukenberg tumor. A problem in histopathologic diagnosis. Am J Surg Pathol 5:225-232, 1981.