Atypical Endometriosis: a Clinicopathologic Study of 163 Cases

DISCUSSION

The incidence of AEM is not well known. The percentage of ovarian AEM of EM reported in the literature has varied between 1.2% to 12.2% or even more^[1,3,6]. Most of the AEMs reported previously were based on a relatively small sample size (less than 300 cases). In this series, the AEM cases were from EM lesions at 6,820 sites of 3,724 patients. This may be the largest group both in the number of AEMs and the EM foci up to now. In this study, using strict criteria to re-evaluate the samples of all of the 6,820 EM lesions, the statistical data showed that the rate of AEM was 6.8% (161/2,363 cases) in the ovarian EM, including 7 cases in bilateral ovaries. There have been only a few reports of extraovarian AEMs in the literature^[7]. We found 4 cases of extraovarian AEM occurring in the fallopian tube, cervix and uterine serosa respectively. This supplemented the data for the location of AEMs outside of the ovary. However, we did not find any AEM lesions in 2,506 EM foci, including the pelvic peritoneum, periuterine ligament, intestines, and the urinary tract, even though they are also the common EM sites. Our data indicate that AEMs occur mainly in the ovary, and occasionally in other locations. The distribution of AEMs was not uniformly found in various sites of the pelvic cavity, and the quantity of AEM was disproportionate to the number of EM foci.

AEMs derived from common EM lesions, kept the characteristic features of EM with endometrioid epithelium, stroma, and hemosiderin, or hemofuscin, and present with significant pleomorphism of epithelial cells, having identical features of tumor cells. The AEMs had some features of both EM and tumors, but also differed from them. The endometrioid tissue of the AEMs in most cases of this group became thin, usually showing atrophy, or a configuration without hormonal influence. The AEMs with “ectopic endometrial hyperplasia” were not evident in our series. The component of epithelium and stroma of AEMs was commonly disproportional. In this study, 63% of the AEM lesions had only endometrioid surface epithelial lining of the endometriotic cysts, with a few gland components found underneath it. The endometrioid stroma was not abundant and seemed less developed than in common EMs. In addition, AEMs have significant cellular atypia, with epithelial buds or pullulation in 78% of the lesions and a papillary structure in 44% of the lesions on microscopy. However, the severe inflammatory background was unconspicuous in most AEMs, and was ill defined as an inflammatory reaction. Fibrotic-collagen might be present in both common EMs and AEMs. In this group, more than 90% of the AEM foci have fibrotic-collagen, which was often thick, and this thickness extended over the endometrioid stroma. Atypical epithelia cells may be situated on the surface and overlay fibrotic and scar-like tissue, or even be within it. This means that the EM lesions might have been present for a long time. The phenomena of hormonal inactivity, significant cellular atypia, and scarring present in AEMs seemed different from physiological changes of the eutopic endometrium. The EM lesions that have experienced damage, repair and scarring changes over a long period of time, may have been exposed to some harmful, or carcinogenic factor released during this process. This might cause the development of EMs into AEMs and the final onset of tumorigenesis. The hormonal action may not be the major factor in the developmental of EMs to AEMs and of EMs, or AEMs into tumors.

The frequency of AEMs associated with tumors is not clear. Most of AEMs associated with a tumor occurred in the ovary, rarely in the pelvic, or gastrointestinal tract and or other places^[7-9]. In our study, 15.7% of ovarian AEMs coexisted with a tumor, and if benign tumors are excluded, it was 14.9% of ovarian AEMs having a borderline and malignant tumor. These data indicated that the ovarian AEMs have a relative higher potential for tumorigenesis and cancer formation. It has been reported in the literature that 0.7 to 2.5% of EMs have a risk of developing into malignant tumors. We support the theory that AEMs smay be a precancerous lesion^[9-12]. The tumors of the ovary accompanying AEMs, were mainly epithelial types. They may be benign, borderline or malignant, with the ratio increasing from benign to malignancy i.e. 11%, 33% and 55%.

Endometrioid and clear cell histology account for 93% of malignent tumors. The borderline and benign tumors in our series were serous, mucinous, and seromucous types originating from the mullerian. The phenomena of AEMs and EMs occurring with tumors not confined in malignant types were described in the literature and confirmed in our cases^[4,11]. The category of EMs associated cancer might extend to the EM-associated tumors, and the investigation of this hypothesis should be conducted with extensive testing. The relationship between AEMs and tumors may be direct, or indirect^[13-15].

In our study, most cases of AEMs associated with a tumor showed a direct transition from the AEMs epithelium to a tumor, especially for malignant tumors, but also were found nearby, or on both sides of the cyst wall. The epithelium of the tumor might have a consistent or different type compared to AEM. Epithelial metaplastic changes were found in 63% of ovarian EM reported by Fukunaga and Ushigome^[16]. In our series, we observed metaplastic changes in half of the AEMs, and eosinophilic and ciliated cell metaplasia was the most common type, identified by Fukunaga etal. and other authors. But both of these changes perhaps were not the true metaplasia^[17], as ciliated cells normally are found along the surface epithelium of the endometrium. We observed that AEMs were also often formed in the surface-type epithelium. Eosinophilic cell changes may be present in several types of cytoplasmic transformations. It is also a frequent feature of atypical cells. How to interpret these changes and whether or not the metaplasic changes of AEMs played an importent role in the tumor genesis is not known.

In our group, most AEM patients were at a reproductive age, with a median age of 40 years. The symptoms and gross appearance of AEMs resembled common EMs. Varing degrees of cellular atypia are common in EM lesions. It is important for the pathologist to recognize, or identify the characteristic features of AEMs, especially the significant cellular atypia, with cellular buds and papillary structure, which are different from an inflammatory reaction. Abundant inflammatory cells, slight atypia, and regular epithelium indicate a reaction change. Residual EM lesions must be recognized in EM-related ovarian cancer, which display a three-layer structure including endometrioid epithelium and stroma, as well as fibrotic-collagen often mixed with hemosiderin. The characteristic features of the three-layer structure both in EMs and AEMs was not found in the non-EM-derived epithelial tumor.

Since AEM lesions may be limited to local areas, they can be found in the white colored “aging ”part or scar-like tissue in EM cysts, and should not be overlooked in sampleing. In the examinations of these tumors, special attention should be paid to whether or not there is a trace of scar tissue in EMs and AEMs, and should be included in the diagnosis and identification of the AEM into routine practice.

In summary, we observed 163 cases of AEM lesions, which have some features of both EM and tumors, but differed from them. AEMs have a relatively high risk of becoming malignant. Except for malignant tumors, borderline and benign tumors can also coexist with AEM. AEMs may play an important role in the case of EM-associated tumors.