Wonderful post. We share a curiosity about this. I agree about the "retrograde-menstrual-flow theory" as being hogwash and regarded it as being about as "scientific" as UFO/alien visitations from the 50s/60s. More biologically sound is the idea about metaplasia. But what (to me, anyway, is more intriguing is how the metaplastic epithelium induces the stroma around it to become endometrial-like stroma or is the "stroma" just another metaplastic change. Also, what of the association between endometriosis and a range of neoplastic lesions?
Mark raises some interesting issues in his short comment. If the readers of this blog will indulge me, I will set off on a brief excursion into some pathologic esoterica. Under the microscope, endometriosis consists of endometrial glands surrounding by endometrial-like stroma. The question arises whether this stroma represents a metaplasia of the pre-existing garden-variety stroma induced by the endometrial glands. Although Mark does not raise the issue, it's also possible, I think, that the columnar endometrial epithelial cells undergo another type of metaplasia into spindled stromal cells. Such a suggestion is a form of heresy in that many of us were taught that mesenchymal cells are inert and fundamentally different than the dynamic epithelial cells. In other words, epithelial cells don't morph into stromal cells or stromal-like spindled cells.
Mark also raises the issue of whether some forms of endometriosis are pre-cancerous. I did a little snooping around and came up with the following passage from Clinical Obstetrics and Gynaecology:
For several decades, endometriosis has been suspected of playing a role in the aetiology of ovarian cancer....Epidemiological evidence from large-cohort studies confirms endometriosis as an independent risk factor for ovarian cancer. Further circumstantial evidence for this link was found in the common risk factors for ovarian cancer and endometriosis. These risk factors influence retrograde menstruation and endometriosis in the same positive or negative way. Based on data in the literature, the prevalence of endometriosis in epithelial ovarian cancer has been calculated to be 4.5, 1.4, 35.9, and 19.0% for serous, mucinous, clear-cell and endometrioid ovarian carcinoma, respectively. The risk of malignant transformation in ovarian endometriosis was calculated at 2.5% but this might be an underestimate. In addition, some authors described atypical endometriosis in a spatial and chronological association with ovarian cancer. Finally, molecular studies have detected common alterations in endometriosis and ovarian cancer. These data suggest that some tumours, especially endometrioid and clear-cell carcinomas, can arise from endometriosis. Moreover, endometriosis-associated ovarian cancer represents a distinct clinical entity, with a more favourable biological behaviour, given a lower stage distribution and better survival than non-endometriosis-associated ovarian cancer.
To close the loop, here's a link to an article about atypical endometriosis (see: Atypical endometriosis: a clinicopathologic study of 163 cases). So, some forms of endometriosis are much more than a nuisance -- they can rarely set the stage for an ovarian cancer, although one that behaves less aggressively than other types.