Endometriosis Then and Now: A 100-Year Journey Around Pathogenesis and Clinicopathologic Associations

Endometriosis Then and Now: A 100-Year Journey Around Pathogenesis and Clinicopathologic Associations Full access Sudarshana Roychoudhury From the Department of Pathology and Laboratory Medicine, Donald and Barbara Zucker School of Medicine at Hofstra/Northwell, Hempstead, New York (Roychoudhury) Search for other papers by Sudarshana Roychoudhury in Nearly a century has passed since a seminal review article was published in 1928 in Archives of Pathology on ectopic endometriosis, presenting the dominant pathogenetic theories of the era: the serosal theory suggesting that endometriosis would arise by metaplasia of the peritoneal mesothelium; the metastatic theory implying the role of uterine lymphovascular invasion by endometrial tissue; and the implantation theory. Since then endometriosis has been the subject of numerous studies due to the associated morbidity and its significant impact on patients’ quality of life and health care costs. Objective.— To provide an overview of our evolving understanding of pathogenesis and clinicopathologic features of endometriosis. Data Sources.— Review of the literature and personal experience of the authors. Conclusions.— The prevailing pathogenetic theory of endometriosis remains the retrograde menstruation and implantation theory proposed by Sampson in the 1920s. Contemporary studies have focused on the additional role of inherited genetic components, a supportive microenvironment, and molecular genetic alterations to improve the clinical classification, prognostication, and therapeutic options for patients with endometriosis. It has been nearly 100 years since the seminal review article by Victor Jacobsen was published in 1928 in Archives of Pathology on ectopic endometriosis.1 Jacobsen thoroughly presented and critically analyzed the prevailing pathogenetic theories of the era: the serosal theory suggesting that endometriosis would arise by metaplasia of the peritoneal mesothelium,2–4 the metastatic theory implying the role of uterine lymphovascular invasion by endometrial tissue,5,6 and the implantation theory.7–9 In the latter, Sampson7–9 proposed that menstrual blood could flow backward (“retrograde menstruation”) through the fallopian tubes, carrying viable endometrial cells that are capable of implanting on peritoneal surfaces. In his 1921 article describing 23 patients with hemorrhagic ovarian cysts, he hypothesized that in some cases the implants would originate from ruptured endometriotic cysts. Sampson also described intestinal involvement in several patients—intestinal adenomas of endometrial type—in 1922,10 and in 1925 introduced the term endometriosis.11 In the 1920s several investigators successfully proved the concept of autotransplantation in various animal models.12–16 Jacobsen was able to demonstrate implantation of scraped endometrial mucosa, first in the peritoneal cavity, and later in the intestinal tract and urinary bladder of rabbits12,15 and in the pelvic peritoneum of Rhesus monkeys.14 O’Keefe and Crossen13 transplanted endometrium to the ovaries, mesoappendix, and rectus muscle in dogs. In 1927 Schwarz16 described his experimental observations of endometriosis in the cesarean section scar of guinea pigs. Almost 100 years later, the most accepted contemporary pathogenetic explanation of endometriosis continues to be the retrograde menstruation and implantation theory.17 Other current theories—coelomic metaplasia theory18 and Müllerian remnant theory19—are thought to be responsible for only a small subset of endometriotic lesions involving the ovaries and midline structures (uterosacral ligament and cul de sac), respectively. During the past several decades the important role of inherited genetic components and a supportive microenvironment with increased angiogenic and inflammatory factors have also been recognized.20 Endometrial stromal cells show increased adhesion due to altered integrin profiles, while decreased immunity prevents cells from attacking endometriotic deposits.21,22 Molecular analysis of deep endometriosis revealed frequent somatic mutations, including known cancer driver mutations in ARID1A (AT-rich interaction domain 1A), PIK3CA (phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha), KRAS (Kirsten rat sarcoma viral oncogene homolog), and PPP2R1A (protein phosphatase 2 scaffold subunit A alpha) in a subset of patients.23 While the precise role of these molecular alterations in the disease pathogenesis is not entirely clear, they may provide an opportunity to improve the clinical classification and prognostication of endometriotic lesions. EPIDEMIOLOGY AND CLINICAL FEATURES The prevalence of endometriosis varies across populations and diagnostic techniques, with uncertain true prevalence due to laparoscopic diagnosis requirements.20,24 A 5- to 8-year delay between symptom onset and diagnosis affects prevalence estimates.25 Endometriosis is very common, affecting approximately 10% of women of reproductive age globally, totaling 190 million women.26,27 Its prevalence ranges from 2% to 11% in asymptomatic women, 5% to 50% in infertile women, and 5% to 21% in women hospitalized for pelvic pain.26 A systematic review found a prevalence of 4.4% among the general population, varying from 1.4% in Europe to 15.4% in Asia.28 The Rochester study reported a 0.3% annual incidence in women aged 15 to 49 years.29 Extrapelvic endometriosis is typically diagnosed at 34 to 40 years of age, while pelvic endometriosis typically occurs in a younger age group (25–30 years).30 Key risk factors for endometriosis include early menarche, long menstrual cycles, nulliparity, and family history.25 Genetic studies show heritable variations, with twin studies estimating 51% heritability.31 A meta-analysis revealed 27 significant genetic loci accounting for 2.2% of disease variance.32 Additional factors include skin pigmentation, sun exposure, alcohol, oral contraceptives, and environmental exposures like polychlorinated biphenyls and dioxin.25 Laparoscopy is the gold standard for the diagnosis of endometriosis, yet only 50% of biopsies from suspected areas show microscopic confirmation, likely resulting from underdiagnosis due to the subtlety or focality of microscopic findings at least in some of the cases.33 Additional causes of discrepant laparoscopic visual versus histopathologic diagnosis may include inadequate sampling and false-positive clinical assessment due to other entities, that is, endosalpingiosis, mesothelial hyperplasia, fibrosis, hemosiderin deposition, suture granulomas, inflammatory changes, residual carbon from previous ablations, and rare incidental benign or malignant neoplastic deposits.33 Clinical Classification and Staging Endometriosis is classified into 5 clinical subtypes by the extent and location of lesions34 (Table 1). For clinical assessment of severity of disease, the revised American Fertility Society classification35 is the most widely used (Table 2). SITES OF INVOLVEMENT AND THEIR CLINICOPATHOLOGIC SIGNIFICANCE Endometriosis may occur in pelvic and less frequently extrapelvic locations. Pelvic endometriosis commonly affects ovaries, fallopian tubes, pelvic peritoneum, uterine ligaments, and rectovaginal septum.36 Less common pelvic sites include the cervix, vagina, uterine serosa, bladder, and rectum.36 Extrapelvic endometriosis may occur in the gastrointestinal tract, kidneys, lungs, umbilicus, abdominal wall, diaphragm, and even the central nervous system.37 Gastrointestinal and Pancreatobiliary Tract Endometriosis affects the gastrointestinal tract in 15% to 37% of patients with pelvic endometriosis38 (Figure 1, A and B). The sigmoid colon is most frequently involved (65% of cases), followed by the rectum, ileum, appendix, cecum, stomach, and transverse colon.39,40 While typically microscopic, lesions can form large masses causing intestinal obstruction and mimicking colonic neoplasms.39 Histopathologic changes in adjacent tissues may resemble inflammatory or infective etiologies, rectal ulcer, or neoplastic lesions, delaying accurate diagnosis.39,41,42 Rare cases of hepatic,43,44 pancreatic,45 and gallbladder endometriosis46 have also been documented, presenting with abdominal pain or biliary tract obstruction. Genitourinary Tract Genitourinary tract endometriosis affects up to 6% of women with endometriosis.47 The bladder is most commonly affected (85% of cases), followed by the ureters, while kidney and urethral involvements are rare.47 Bladder endometriosis presents with urinary frequency, dysuria, or hematuria, while ureteral involvement may cause urinary obstruction and hydronephrosis.37 Other Less Common Sites of Endometriosis Cutaneous endometriosis comprises 0.4% to 4% of extragenital cases, with 30% to 40% occurring in the umbilicus.48,49 First described by Villar in 1886, umbilical endometriosis is also known as Villar’s nodule.50,51 Primary cutaneous endometriosis develops spontaneously in 30% of patients, while secondary cutaneous endometriosis occurs in scar tissue after abdominal or pelvic surgery, including cesarean delivery, through endometrial cell implantation.52 It typically presents with cyclical pain and bleeding.49 Thoracic endometriosis syndrome (TES) involves endometrial tissue in the thoracic cavity, affecting the pleura, lung parenchyma, and airways,53 and usually presents as catamenial pneumothorax and hemothorax, hemoptysis, and pulmonary nodules, with symptoms occurring during menstruation.53,54 TES diagnosis is challenging due to its rarity and requires exclusion of other causes.54 Rarely, endometriosis has also been reported in pelvic and abdominal lymph nodes,55,56 soft tissues,57,58 bone,59 spleen,60 and even in the brain61,62 and spinal cord,63 presenting with headaches and gait disturbances, back pain, radiculopathy, motor-sensory deficits, and loss of bladder/bowel control, varying with menstruation. Endometriosis at these distant sites suggests that at least some endometriotic lesions may result from lymphatic or venous vascular dissemination as hypothesized by Halban5 and Sampson.6 HISTOPATHOLOGIC FEATURES Endometriotic foci are composed of variable proportions of endometrial glands and stroma, the latter containing characteristic small arterioles and often clusters of hemosiderin-laden macrophages (Figure 1, C). The endometrial glandular epithelium may be atrophic or may follow the typical changes of the menstrual cycle.40 The endometriotic stroma consists of small uniform cells with scant cytoplasm and dark nuclei with small arterioles and/or stromal hemorrhage. Depending on the phase of the menstrual cycle the stroma may show progestin effect or breakdown, similar to menstrual endometrium.40 Several alterations may obscure the normal appearance of endometrial stroma. These include histiocytic infiltrate, fibrosis, elastosis, and myxoid and decidual change.40,64,65 In early lesions, histiocytic infiltrate shows granular, eosinophilic, or foamy cytoplasm (pseudoxanthoma cells).40 Advanced lesions contain pigmented histiocytes with lipofuscin, appearing as gray-brown cytoplasmic granules that stain with periodic acid–Schiff–diastase (PAS-D).40 Cases with extensive stromal histiocytic infiltrate have been termed xanthomatous endometriosis in the literature.40 Rarely, cholesterol clefts and giant cell reaction may also be seen (cholesteatomatous endometriosis).66 Fibrosis, initially composed of stellate fibroblasts, can progress to hyalinized tissue with calcifications in chronic lesions. Stromal decidual changes may occur during pregnancy or as a result of progesterone therapy (Figure 1, D). The glandular epithelium may show metaplastic changes, including ciliated (tubal), eosinophilic, hobnail, squamous, mucinous, and clear cell types, similar to those found in eutopic endometrium. In pregnant women, the endometrial glands are inactive, show cystic dilatation, and are lined by low cuboidal cells, which may be mistaken for endothelium or mesothelium, leading to ectopic decidua diagnosis.40 Cases with Arias-Stella reaction have also been documented.67 In postmenopausal patients, the diagnosis is challenging due to gland atrophy and diminished or fibrotic stroma. Exogenous hormonal treatment may also result in similar atrophic features.68 Immunohistochemical stains—CD10, WT1, and estrogen receptor (ER)—may help identify endometrial stromal cells and distinguish them from inflammatory cells69,70 (Figure 1, A; inset). Positive PAX8 (paired box 8) and ER immunostaining may be used to differentiate endometrial glandular epithelium from cystic mesothelial-lined structures (Figure 1, B; inset). However, PAX8 positivity has also been reported in the female pelvic peritoneum,71 thus additional mesothelial markers may be applied for this differential diagnosis. In addition, multiple deeper-level sections have been shown to increase the diagnostic accuracy and lower the false-negative rate in peritoneal biopsies.72 Based on this study, routine deeper-level sections (×3) should be considered as the standard practice in peritoneal biopsies from patients with clinical suspicion of endometriosis and no histologic evidence of endometriosis on the initial sections. Furthermore, if only representative sections were initially submitted from a larger peritoneal sample, the remaining specimen also needs to be processed for microscopic evaluation. In cases with diagnostic uncertainty an immunohistochemical panel to include CD10, ER, WT1, and PAX8 may also be considered to enhance the diagnostic yield. Morphologic Variants The term stromal endometriosis has been used for lesions that are composed exclusively of endometrial stroma, without overt glandular structures, and may pose a diagnostic challenge.73 It may appear as microscopic nodules (“micronodular stromal endometriosis”)74 beneath the pelvic peritoneum or omentum, less frequently in the ovary and cervix,75 and may mimic lymphoid aggregates under scanning magnification (Figure 2, A). Polypoid endometriosis forms polypoid masses, morphologically similar to eutopic endometrial polyps.76,77 It may affect colonic mucosa reaching up to 14 cm in size, potentially causing intestinal obstruction, or may be seen protruding into the lumen of cystic ovarian endometriosis.76 Pathogenetically, polypoid endometriosis has been linked to hormonal therapy and tamoxifen treatment.78,79 The differential diagnosis includes extrauterine Müllerian adenosarcoma, which can be distinguished by its characteristic periglandular stromal hypercellularity (periglandular cuffing), stromal atypia, and irregular (leaflike) glandular architecture with intraglandular stromal papillae. Stroma-only or stroma-predominant endometriotic foci may also mimic low-grade endometrial stromal sarcoma, although the latter typically forms a larger mass lesion with overt invasive growth pattern, including lymphovascular invasion. In some cases, regressing endometriotic foci may present as multiple peritoneal nodules consisting of central necrosis surrounded by pseudoxanthomatous cells and hyalinized stroma (descriptively termed necrotic pseudoxanthomatous nodule)80,81 (Figure 2, B), which may be confused with granulomatous infections or neoplastic processes. Endometriosis can also be associated with inflammatory and reactive changes,82–84 including significant mesothelial hyperplasia.85 The reactive mesothelial cells may form tubules, nests, and cords, or may appear as single cells in a fibrotic stroma with a pseudo-infiltrative pattern, mimicking adenocarcinoma. Liesegang rings—acellular, eosinophilic, PAS-positive structures—may also be found in endometriotic cysts and should not be mistaken for foreign material or microorganisms.86 Atypical Endometriosis The term atypical endometriosis encompasses both cytological and architectural atypia and has been reported in 1.7% to 4.4% of all endometriotic lesions.87,88 Cytological atypia, focal or multifocal, of the endometrial glandular epithelium is often found in endometriotic cysts. The epithelial cells are enlarged, polygonal with varied eosinophilic cytoplasm and nuclei showing mild to severe pleomorphism, hyperchromasia with smudged chromatin, and/or prominent nucleoli (Figure 2, C and D). The atypical cell lining is usually monolayered, with possible focal stratification and papillary proliferation, and stromal/epithelial neutrophilic infiltrate is commonly present.40 Cytological atypia is often attributed to reactive changes from blood and hormonal influence.24 Architectural atypia is characterized by crowded, atypical proliferation of endometrial glands, morphologically similar to that seen in atypical hyperplasia of eutopic endometrium.89 In the presence of atypical endometriosis thorough sampling is necessary to rule out an adjacent carcinoma. Atypical endometriosis has been pathogenetically linked to ovarian borderline tumors and endometrioid and clear cell carcinomas.87–89 Several recent studies have found shared molecular alterations, including PTEN (phosphatase and tensin homolog), KRAS, ATM (ataxia-telangiectasia mutated), ARID1A, and CTNNB1 (catenin beta 1) mutations, between atypical endometriosis and adjacent endometriosis-associated neoplasia.88,90–92 The prevalence of synchronous/subsequent tubo-ovarian neoplasia (both borderline tumors and carcinomas) has been reported in up to 25% of atypical endometriosis.88 Endometriosis-Associated Neoplasia In his 1928 article Jacobsen1 discussed the merits of regarding endometriosis as a neoplasia given its ability to invade other tissues, and also drew attention to the possibility of malignant neoplastic transformation, which had been described by Sampson93 a few years earlier. Recent epidemiologic studies have found that women with endometriosis have a 2- to 3-fold increased risk of ovarian epithelial tumors, primarily clear cell carcinoma (60%–70%) and endometrioid carcinoma (30%–40%).94–96 Less common endometriosis-associated tumors include borderline endometrioid and seromucinous tumors, low-grade endometrioid stromal sarcoma, and adenosarcoma.97–99 Endometriosis-associated malignant tumors most frequently occur in the ovaries, possibly due to the special microenvironment related to hormone levels, oxidative stress, cytokines, and other immunologic factors.100 Rarely, endometrioid carcinoma may develop from intestinal endometriotic foci and may mimic a colonic primary. The presence of associated endometriosis, absence of mucosal involvement, and frequent squamous differentiation typically help establish the correct diagnosis, although immunohistochemical workup is also often utilized for the precise diagnosis. CONCLUSIONS Endometriosis has been a focus of intense investigations and has fascinated gynecologists and pathologists for more than a century. As a chronic disease with potential severe impact on patients’ quality of life due to dysmenorrhea, pelvic pain, dyspareunia, and infertility, it can impose significant economic burden through medical costs and lost work.101 While the field has seen numerous advances in the past decades, the prevailing pathogenetic theory of endometriosis remains the retrograde menstruation and implantation theory proposed by Sampson7,9 in the 1920s. Contemporary studies have focused on the role of inherited genetic components, a supportive microenvironment, and molecular genetic alterations to improve the clinical classification and prognostication, and advance and expand the therapeutic options for patients with endometriosis. A and B, Endometriosis involving the gastrointestinal tract. A, Endometriosis in the appendiceal muscularis propria with atrophic glands and scant surrounding endometrial stroma, highlighted by CD10 immunostain (inset). B, Endometriosis involving the colonic mucosa in an endoscopic biopsy. PAX8 immunostain highlights the endometrial glandular epithelium (inset). C, Deep infiltrating endometriosis in periadnexal pelvic fibroadipose tissue. D, Endometriosis with stromal decidual change and attenuated glandular epithelium (hematoxylin-eosin, original magnification ×100 [A through D]; original magnification ×100 [insets A and B]). Figure 2. A, Nodular stromal endometriosis involving the pelvic peritoneum. B, Regressing endometriosis may appear as a necrotic pseudoxanthomatous nodule. C, Endometriotic cyst of the ovary shows a single layer of endometrial epithelial lining, surrounded by a thin rim of inflamed endometrial stroma. D, Ovarian endometriotic cyst with epithelial atypia (atypical endometriosis) (hematoxylin-eosin, original magnifications ×100 [A and C], ×40 [B], and ×200 [D]). Contributor Notes Corresponding author: Natalia Buza, MD, Yale University School of Medicine, 310 Cedar Street LH 108, PO Box 208023, New Haven, CT 06520-8023 (email: [email protected]). The authors have no relevant financial interest in the products or companies described in this article.