A History of Endometriosis

8. Sampson’s Theory of Implantation Endometriosis

Ronald E. Batt1

(1)

State University of New York at Buffalo, Buffalo, New York, USA

Abstract

From a cursory examination of the literature, one might conclude that research and authorship on adenomyomas passed seamlessly and directly from Cullen to Sampson in 1921; Cullen’s last contribution appearing in Volume two and Sampson’s first article in Volume three of the Archives of Surgery.1 Such was not the case. Unlike the clear and seamless continuity from von Recklinghausen to Cullen, the transfer of authority from Cullen to Sampson was mediated and complicated. Cullen remained the undisputed authority on uterine adenomyomas, but not of extrauterine adenomyomas. In the sense that Sampson postulated a novel theory of pathogenesis to explain the many extrauterine adenomyomas described by Cullen – and in that sense only – there was continuity: mutual interest in the same subject, but the torch had not been passed as it had from von Recklinghausen to Cullen. In every other sense, the transition was marked by discontinuity.

From Cullen to Sampson Via Casler

Cullen, Thomas Stephen.

Three cases of subperitoneal pedunculated adenomyoma.

Archives of Surgery 1921;2:443–454.

Sampson, John Albertson.

Perforating hemorrhagic (chocolate) cysts of the ovary.

Archives of Surgery 1921;3:245–323.

From a cursory examination of the literature, one might conclude that research and authorship on adenomyomas passed seamlessly and directly from Cullen to Sampson in 1921; Cullen’s last contribution appearing in Volume two and Sampson’s first article in Volume three of the Archives of Surgery.1 Such was not the case. Unlike the clear and seamless continuity from von Recklinghausen to Cullen, the transfer of authority from Cullen to Sampson was mediated and complicated. Cullen remained the undisputed authority on uterine adenomyomas, but not of extrauterine adenomyomas. In the sense that Sampson postulated a novel theory of pathogenesis to explain the many extrauterine adenomyomas described by Cullen – and in that sense only – there was continuity: mutual interest in the same subject, but the torch had not been passed as it had from von Recklinghausen to Cullen. In every other sense, the transition was marked by discontinuity.

Sampson had been casually interested in extrauterine adenomyomas for over a decade. He burst into print in 1921 with his first theory of peritoneal implantation from internally menstruating ovaries2which was beyond doubt directly inspired by the case of the externally menstruating ovary of De Witt Casler.3 Furthermore his research was facilitated by tissue sections and tissue blocks of hemorrhagic cysts of the ovary that Sampson had requested and received from Emil Novak, instructor in clinical gynecology at Johns Hopkins Medical School.

Some years later, Novak wrote a historical note: “The report of the first case of aberrant ovarian endometrium by Russell in 1899 didn’t cause a ripple of general interest, and the case was complacently accepted as representing an interesting but rare instance of misplaced muellerian rests. Following this an occasional instance of this sort was encountered in laboratories of gynecological pathology, but no one appreciated its significance until 1921, when Sampson published his first paper on the subject. I well recall that for a year or two before this I had gotten letters from him asking me to send him sections or blocks of any available hemorrhagic lesions of the ovary, and I wondered what he had in the back of his head. We all found out when he published his first classical paper in 1921, establishing the frequency and importance of endometriosis as both a pathological and clinical entity.– Emil Novak, M.D.”4

This discontinuity was not a “paradigm shift” of Kuhnian proportions, but a normal scientific progression. The editors of Thomas S. Kuhn’s essays, from 1970 to 1998, explain Kuhn’s position on normal science, a position that he had arrived at just before his death.5 “It is only because individuals working in a common research tradition are able to arrive at differing judgments concerning the degree of seriousness of the various difficulties they collectively face that some of them will be moved individually to explore alternative (often - as Kuhn likes to emphasize - seemingly nonsensical) possibilities, while others will attempt doggedly to resolve the problems within the current framework. The fact that the latter are in the majority when such difficulties first arise is essential to the fertility of scientific practices. For, usually, the problems can be resolved, and eventually are. In the absence of the requisite persistence to find those solutions, scientists would not be able to home in as they do, on those rarer but crucial cases in which efforts to introduce radical conceptual revision are fully repaid.”6

Undoubtedly, Cullen’s research on extrauterine adenomyomas contributed to the fund of knowledge that Sampson possessed when he grasped the significance of the “case of the menstruating ovary.”7 In one giant intuitive leap, Sampson bounded from Casler’s ovarian-uterus that “menstruated externally” into the vagina to perforating hemorrhagic (chocolate) cysts of the ovary that “menstruated internally” into the pelvis and their relation to pelvic adenomas of endometrial type. Inspired, he secured a ready supply of hemorrhagic ovarian cysts from Emil Novak. In less than 2 years, Sampson completed his study and presented his first theory of the pathogenesis of pelvic adenomas of endometrial type from perforating hemorrhagic (chocolate) cysts of the ovary in 1921.8 Two years earlier Cullen had issued an implicit invitation for others to investigate uterine mucosa in the ovary.9 This released Sampson from all ethical concern that he might be trespassing on the sovereign terrain of his former teacher in medical school and residency at Johns Hopkins University.

In the title of “Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (‘Adenomyoma’ of the uterus, rectovaginal septum, sigmoid, etc.),” one may discern a definite shift in emphasis from adenomyoma to ovary. The ovary became the organ of primary interest.10 Terminology changed. Adenoma replaced adenomyoma. The adjective “endometrial” replaced the phrase “of uterine mucosa.” Chocolate evoked the commonplace descriptive words introduced into pathology by Rokitansky. Analysis replaced description. Living pathology observed by the surgeon displaced the primacy of morbid pathology in the laboratory. Descriptive pathology gave way to pathogenesis as the prime focus of research.

Sampson recalled in 1921 that it was by observing the pelvic findings in a single case in 1910 that his “attention was first directed to the dense peritoneal adhesions which may result from the escape of the contents of these [perforating hemorrhagic] cysts.”11 He had operated his very first case on May 8, 1910, a patient with a frozen pelvis. Here is how Sampson described Case 1: “Perforating hemorrhagic cysts of both ovaries; multiple leiomyomas of the uterus; dense adhesions in the “cul-de-sac” uniting the anterior rectal wall to the supravaginal portion of the cervix and the lower portion of the uterus. The induration of the anterior rectal wall was so great as to simulate malignancy.”12 Unfortunately, he did not examine the ovarian cysts or the posterior wall of the uterus microscopically.

His second case, operated on March 27, 1912, also presented with a frozen pelvis.13 This case provided Sampson with a new insight. He recalled that he realized for the first time “the association between these [perforating hemorrhagic (chocolate)] cysts and ‘adenomyomas’ of the posterior uterine wall with adhesions between it and the rectum.”14 Rephrased, in 1912 Sampson stated he associated chocolate cysts of the ovary with adenomyomas and adhesions of the posterior wall of the uterus, the rectum and the rectovaginal septum. He saw the association, but not the pathophysiologic relationship; he had too little information to postulate pathogenesis.

Sampson operated his third patient with “much less extensive” disease on April 26, 1912.15 He noted that he would have “overlooked” his third case, had he not operated his second patient just 1 month before.16 He said he did not “observe another similar case” until June 13, 1918. Sampson confessed he “undoubtedly overlooked many” other cases in the interim.17 This 1918 case was his first patient with histologic evidence. “The hemorrhagic cyst was lined by low cuboidal and columnar epithelium, the columnar predominating. There was a vascular subepithelial stroma, containing evidence of old and recent hemorrhages and an occasional gland, which resembled a uterine gland.”18 As had Rokitansky in 1860, Sampson reported finding endometrial glands and stroma in the ovary. During 1918 and 1919, Sampson observed perforated hemorrhagic cysts “many times and studied them more from the standpoint of gross pathology, basing the microscopic studies solely on the routine examination of the material in the pathologic laboratory.”19 In 1920, subsequent to his insight from Casler’s case and mindful of Casler’s microscopic studies, Sampson began systematic microscopic studies in his own laboratory.20 He stated: “It was only this last year that I fully realized the true relation between these cysts and pelvic adenoma of the endometrial type and that the pelvic adhesions were often associated with or were in large part due to this adenomatous growth.”21

In 1921, Sampson acknowledged his debt to DeWitt B. Casler when he wrote the following lines: “Of physiologic interest, it is to be noted that the adenoma of the endometrial type developing in the ovary and arising in the portion of the pelvis as the result of the escape of the hemorrhagic contents of the ovary may be the seat of periodic hemorrhage, i.e., they may be ‘menstruating organs.’”22 Sampson recounted Casler’s menstruating ovary: “Casler, in 1919, reported an unusual case in which a patient menstruated through the vagina after a conservative hysterectomy in which one ovary was saved…at the second operation, 4 years later, the enlarged ovary was removed and it was found to contain cavities lined by ‘normal uterine mucosa.’”23

Sampson described the living pelvic pathology observed at surgery. He found perforated hemorrhagic cysts of the ovary adherent to adjacent structures. When the ovary was mobilized, the cyst ruptured and spilled its chocolate contents. Sampson reasoned “this rupture arises from reopening a previous perforation which has been sealed by the organ or structure to which the cyst has become adherent at the site of the perforation, or the cyst is torn in freeing it.”24 The ovarian cysts were usually between 2 and 4 cm. in diameter and were often bilateral. In Sampson’s experience, the site of perforation was always located on the “lateral surface of the ovary or on its free border; [he had] never found it on the mesial surface.”25 He noted that “adhesions are also found in other portions of the pelvis and especially in the culdesac, and these adhesions are apparently the result of the escape of the contents of the cyst.”26 Sampson found the most extensive development of adenoma of endometrial type in the culdesac. These were variously described as ranging from a “quiescent…localized thickening” to a “diffuse [invasive] growth involving the posterior surface of the supravaginal portion of the cervix, the posterior uterine wall, the floor of the culdesac and the anterior wall of the rectum, all of which may be adherent to one another.”27 The adenomatous disease of “endometrial type” and the associated pelvic adhesions had the same anatomic distribution, except when the endometrial disease invaded beyond the confines of the rectovaginal pouch of Douglas.

Sampson noted that invasion variously occurred posteriorly into the rectum or sigmoid colon; laterally into the broad ligament; anteriorly into the uterus, supravaginal cervix, or vaginal wall; and even through the entire thickness of the vagina into the posterior fornix.28 In Sampson’s own words the disease: “may grow down between the rectum and vagina, forming an ‘adenomyoma of the rectovaginal septum’ and may penetrate the vagina and appear in the posterior vaginal vault. In other cases, they may extend through the wall of the rectum or sigmoid.”29 What Sampson, Cullen, and other investigators referred to as “adenomyoma of the rectovaginal septum” were lesions formed by the obliteration of the rectovaginal pouch of Douglas. Put more precisely, the adenomyomatous disease obliterated the rectovaginal pouch of Douglas forming a “pseudo-rectovaginal septum” located above or cranial to the true anatomic rectovaginal septum of Denonvilliers. Sampson and Cullen were not referring to invasion of the anatomical rectovaginal septum of Denonvilliers.30

Sampson offered a modification of Emil Novak’s classification of ovarian hematomas.31 He suggested that a “fourth variety of ovarian hematoma…one lined wholly or in part by “endometrial tissue” should be added to Novak’s ovarian follicular, stromal, and corpus luteum hematomas.32 Sampson used the terms “tissue of endometrial type”33 and ‘endometrial tissue’”34 in place of Cullen’s term “uterine mucosa.” This was the first time that Sampson used the term “endometrial tissue.” I believe this was the first step toward his coining the term endometriosis.

One can see Sampson thinking deeply about the pathophysiology of endometrial tissue within the ovary and the histologic standard of comparison. “In the study of ovarian hematomas to determine whether or not the lining of the hematomas is of endometrial type, it would seem preferable to use as our standard of comparison not normal endometrium but ectopic endometrium in which there is a cyst (hematoma) formation due to the retention of ‘menstrual blood,’ similar to the condition in ovarian hematomas. We have abundant opportunity to study the variations in the appearance of the uterine mucosa in the hemorrhagic cysts or cavities, so often found in uterine ‘adenomyomas.’ These should be our standards of comparison in the study of ovarian hematomas because the physical conditions are similar.”35

By 1921, Sampson had developed his first theory of pathogenesis of pelvic adenomas and associated adhesive disease; both resulted from chocolate fluid containing endometrial tissue that spilled into the pelvis from ruptured hemorrhagic ovarian cysts. Next, Sampson pondered the pathogenesis of these perforated ovarian cysts. He cited the research of Runge36 and Wolf,37 who made serial sections through the site of rupture and “demonstrated the ‘epithelialization’ of ovarian hematomas by the invasion of ‘surface epithelium of the ovary’ through the opening caused by the rupture.”38 In other words, Runge and Wolf postulated secondary invasion of ovarian hematomas – be they follicular, stromal, or corpus luteum hematomas. Sampson interpreted their work: “It may be possible that, following the rupture of the hematoma or whatever structure preceded the secondary invasion, misplaced epithelium of endometrial type was present in the periphery of the ovary at this site and this epithelium was stimulated to become invasive and reline the cavity of the hematoma.”39 However, in the sentence just preceding this last statement, Sampson had speculated: “If these cysts are of endometrial type and if their epithelial lining arises from the invasion of the surface epithelium of the ovary through the place of rupture, we must conclude that a metaplasia of the epithelium occurs, by which it may not only assume the histologic picture of endometrial tissue but may even function as such.”40 Without reference to the serosal metaplasia theory of Iwanoff or the peritoneal metaplasia theory of Meyer,41 Sampson seems to have tentatively accepted metaplasia of the ovarian surface cells as the pathogenesis of perforating hemorrhagic cysts of the ovary, contingent on the validity of his assumptions stated above. Sampson gave us insight into his thoughts and his reservations. “In most of the specimens which I have examined it has been impossible to determine the exact nature of these cysts before the initial rupture. They may have been endometrial cysts at the start; or they may have resulted from an abnormal condition of a follicle by which a hematoma arose in a graafian, or atretic follicle; or possibly following ovulation, an abnormal corpus luteum developed, due to the invasion of the epithelial tissue as above mentioned. With my present knowledge, I prefer to mention these possibilities rather than made definite statements which later may prove to be incorrect.”42

Sampson divided 26 cysts that he had examined microscopically into three groups.43 Groups one and two represent different developmental stages of cysts lined with “misplaced atypical endometrial tissue.” The third group consisted of “tissue of endometrial type…present in pockets in the periphery of the ovary about the perforation.” He pondered the idea of ovarian cysts lined by “misplaced endometrial tissue” as having a “definite life history” characterized by passage through “various stages of growth or development to be followed by various stages of retrogression by which smaller cysts may ‘disappear’ as they are apparently rare after the menopause.”44 Once again Sampson avoided premature conclusions: “I hesitate to state what I believe their life history to be because I am not sure that I am correct.”45 But Sampson could not leave the subject without commenting further. “Both the hematoma and the clefts or pockets are apparently part of the same process. The stimulus which causes the epithelial invasion of the hematoma or the development of an ‘endometrial’ cyst also apparently causes the development of the clefts or pockets.”46 At this point he could not define the stimulus.

Twenty pages into this essay47, Sampson first employed the word “implanted,” an essential element of the ancient seed and soil botanical metaphor that one day would make his theory of retrograde uterine menstruation so compelling and memorable.48 “The ‘menstrual’ blood escaping [from the ovary] into the peritoneal cavity may carry with it some of the epithelium lining the cyst cavity, or similar tissue may escape from the endometrial pockets in the periphery of the ovary about the perforation. This epithelium may become implanted in the culdesac or other portions of the pelvis and there give rise to other foci of ‘endometrial’ tissue.”49

Sampson then turned to the subject of endometrial pelvic adhesions, the cause of which he began studying in 1918.50 He described an uncontrolled experiment by Dr. George S. Graham in 1918. Graham injected material “obtained aseptically” from ovarian cysts into the peritoneal cavity of eight rabbits and peritoneal adhesions resulted.51 Sampson said he had planned to write a paper in 1920 “describing these cysts and the adhesions resulting from them…At the time I believed that these cysts were ‘endometrial’ hematomas and that the adhesions arose from the escape of ‘menstrual’ blood into the peritoneal cavity.”52 He still believed so in 1921.

In Sampson’s judgment, when endometrial cysts ruptured, the spill was usually small and rarely caused symptoms.53 However, symptomatic or not, the spilled cystic contents caused irritation and adhesions about the ovary and in recesses in the pelvis. Sampson cited the work of Savage54 who believed that the adhesions “arose from a reactive inflammation due to the escape of the hemorrhagic contents of the cyst [unassociated with any] gross evidence of tubal inflammatory disease.”55 Hedley found similar evidence of inflammation and lack of evidence of tubal inflammatory disease.56 Sampson described the inflammatory host response to the irritation from spilled cyst contents.57 Based on 13 cases “in which tissue involved in the adhesions was examined histologically and adenoma was found,” Sampson described four adhesion patterns and locations of adenoma resulting from the rupture of endometrial cysts of the ovary and the spilling of their contents into the pelvic cavity.58

Pattern one: “extensive adhesions…obliterating …the culdesac: with adenoma of the endometrial type invading the cervical and uterine tissues and probably also…the rectum.”59

Pattern two: “adhesions between the uterus and rectum with multiple discrete invasions of the posterior uterine wall by adenoma of the endometrial type.”

Pattern three: “Adhesions in the normal peritoneal folds associated with the development of adenoma of endometrial type.”

Pattern four: “Discrete nodules of adenoma in the wall of the rectum and sigmoid.”60

Tellingly, Sampson documented that endometrial adenomas existed without adhesions, but he did not attempt to explain this apparent anomaly.

The crux of Sampson’s paper in the Archives of Surgery revolves around the answer to his own question: “What is primary?” His answer is interesting – not least – because it has been forgotten by nearly everyone. “The study of my material has convinced me that the ovary is the primary site.”61 By that Sampson meant the ovary was the source of all the pelvic adenoma of endometrial type and of all the associated pelvic adhesive disease. He believed the extension of the disease could be “growth by continuity” from the ovary or by “implantation of epithelial cells carried with the contents of the cyst or from the epithelial clefts and pockets in the ovary.”62 In 1921, Sampson used a powerful analogy to express his conviction that the ovary was the primary source of endometrial implants; he suggested that they were analogous to the implantation of cancer in the cul-de-sac from rupture of cancerous ovarian cysts.63

Sampson was particularly interested in “blebs or small peritoneal cysts, some containing blood [which] histologically are cystomas of endometrial type.”64 He offered three modes of pathogenesis to explain the presence of these blebs on the serosa of the uterus; “first, from the deposit of epithelial cells from the ovary on the peritoneum and their subsequent development into a cyst…; second, a small cyst or dilated gland may have been detached from the ovary…and may be implanted on the peritoneum and then increase in size, and third, the ‘specific’ irritation of the contents of the ovary caused a metaplasia of the peritoneal endothelium giving rise to a cyst of endometrial type.”65 Though not an advocate of the theory of metaplasia, but in fairness to his readers, Sampson quoted Lockyer, who was an advocate. 66 “Lockyer expresses the following views of the advocates of this theory: ‘Heterotopy of serosal epithelium is the probable explanation of the existence of the epithelial spaces and cysts in most of the extra-uterine swellings found between the rectum and genital tract,’ and again he states that ‘it has also been conclusively shown that the connective tissue which surrounds the ‘endothelial’ inclusions can be excited to hyperplasia which causes it to assume the characteristic histological features of the stroma of the uterine mucosa.’”67

Specifying that “a large percentage, and possibly all, of the ovarian hematomas reported in this paper were of endometrial type,”68 Sampson postulated the pathogenesis of bowel endometriosis, specifically endometriosis of the sigmoid colon: “Apparently some of the contents escaping from the ovarian hematoma had carried with them some of its epithelium, which was deposited on the surface of the sigmoid and later invaded it.”69 This is an important statement; Sampson essentially believed that living, viable endometrial cells from the ovarian hematoma were capable of implanting on the serosa of the sigmoid colon, invade and cause an “adenoma of endometrial type.”70 So saying, Sampson reinforced the principle message of this essay; that the ovary is the primary source of pelvic adenomas of endometrial type and the associated pelvic adhesions. After saying that “a large percentage…of the ovarian hematomas…were of endometrial type,” Sampson seemed to offer a circular argument: “The most important evidence that these ovarian hematomas and clefts or pockets are of endometrial type is the secondary development of adenoma of endometrial type in the tissue or structure which have become ‘infected’ by material escaping from them.”71

Sampson considered the pathogenesis of uterine adenomyomas. He accepted “invasion from ‘within’ the uterus…and…invasion from ‘without’ the uterus, [because] histologically the two tumors are identical.”72 However, he expressed uncertainty regarding the existence of a third group of uterine adenomyomas. “I do not know whether or not there is a third group of ‘adenomyoma’ arising from misplaced endometrial tissue in the uterine wall or by invasion from ‘without’ from other sources than the ovary.”73 Granting that ovarian hematomas may be the source of secondary endometrial adenomas in the pelvis, Sampson asked a provocative question. “Are all ectopic pelvic adenomyomas of endometrial type secondary to a similar condition in the ovary? I cannot answer this question.”74

Despite the explanatory power of his ovarian theory, Sampson’s question indicates he was not satisfied that it gave a comprehensive explanation. For example, Sampson “operated on two patients with pelvic adenoma of endometrial type without gross evidence of hematoma of the ovaries.”75 He considered two possible sources of the endometrial tissue he found on the surface of the ovary; either it represented “remains of an ‘endometrial’ hematoma in which nearly complete retrogression had occurred or ‘endometrial’ pockets or clefts which had functioned, namely, had menstruated, and the secondary pelvic adenomas had arisen from them.”76 Then, Sampson made two remarkable statements that reveal that he was thinking deeply of an alternative hypothesis, perhaps to complement rather than replace his theory of the primacy of ovarian hematomas as the source of adhesions and pelvic adenomas of endometrial type. The first statement was that: “It is also possible that their origin was independent of the ovaries.”77 The second statement is the more remarkable: “There is such a great variation both in the invasiveness of the pelvic adenomyomas and also in their finer histologic appearance that one wonders whether they are all the same.”78

In this first essay, Sampson analyzed the clinical features of his first 23 patients from which he fashioned a durable clinical picture of moderate, severe, and extensive endometriosis. He observed endometriotic disease in women aged 30 years to menopause, which also happens to be the ages at which uterine myomas are prevalent.

Given that Sampson observed mostly moderate, severe, and extensive disease at laparotomy, one can better appreciate the clinical picture he assembled. Sampson’s observations gained such an aura of authority that physicians soon came to regard endometriosis as a disease of the fourth and fifth decades of a woman’s life. Endometriosis in adolescents was not considered seriously until gynecologists began to investigate pelvic pain in adolescents by laparoscopy in the 1970s and 1980s. Sampson estimated he encountered “perforating hemorrhagic cysts of the ovary” in about 10% of his operations for relief of pelvic disease.79 He encountered both primary and secondary infertility; “the shortest time that had elapsed since childbirth in any of these cases was 5 years.”80 Sampson believed that the “extent and situation” of the adhesions and other pelvic diseases were the most important factors in “the origin of subjective symptoms.”81 In his cases, adhesions were always present. They resulted from the “irritating action” of spilled material produced during the “menstrual” cycle of the perforated hemorrhagic cysts of the ovary.82 Sampson specified that these cysts were more common in women with primary infertility, but were “apparently rare in women who have had salpingitis.”83

Then Sampson made a highly significant observation. “In not a single instance of the 23 cases reported in this paper was there any gross evidence of a recent or an old inflammatory disease of tubal origin; the fimbriated extremities of the tubes in all cases appeared normal and whatever adhesions were present about the tubes were of extratubal origin, that is, from the contents of the cyst.”84 He concluded that “the condition under consideration” was the cause of fixed retroversion or retroflexion of the uterus observed in ten of his cases. The cysts were usually small, often bilateral, and between 2 and 4 cm. in diameter, the largest being 9 cm. Note that Sampson did not consider cysts and adhesions to be invasive disease when he discussed the patients’ symptoms.

Twelve of the 17 symptomatic patients sought relief for pain, three for increasing constipation, and one each for uterine bleeding and sterility.85 The amount of pain and abnormal bleeding were quite variable, and often absent.86 Sampson concluded: “It is very difficult to decide whether or not these cysts cause profuse or irregular (too frequent) menstruation.”87 As for pelvic pain, he concluded that “there is usually nothing characteristic about the pain present in this condition nor is there necessarily any relation between the extent of the adhesions and the severity of the pain.”88 Later Sampson modified this statement saying, “if painful menstruation results, it is of the acquired variety, of recent development and may be progressive in severity.”89 Constipation is sometimes worse at the menses and “the symptoms of the advanced cases with narrowing of the lumen of the intestine are similar to those of intestinal obstruction due to other causes such as malignancy.”90

Within his first 23 cases, Sampson encountered many of the typical presentations of pelvic endometriosis, some of which he illustrated. A transverse section of the pelvic structures and surrounding pelvic and vertebral bones, with an ovarian hematoma adherent to the pelvic side wall.91 An illustration of a sagittal section of the pelvis depicts a retroflexed uterus adherent to endometriosis of the rectovaginal pouch of Douglas [RVPD] overlying the posterior fornix of the vagina.92 A sagittal view of the pelvis with complete obliteration of the RVPD showing a large ovarian endometrioma adherent to the rectum posteriorly and the upper posterior uterine fundus anteriorly.93The sagittal section of the pelvis with an “adenomyoma” that has invaded the rectum and posterior wall of the uterus with complete obliteration of the RVPD but has not invaded the posterior vaginal fornix.94 Complete obliteration of the RVPD with the rectum densely adherent to the supravaginal portion of the cervix up to the insertion of the uterosacral ligaments.95 There are good illustrations of an intact surgical specimen consisting of the complete uterus with both tubes and ovaries showing “perforating hemorrhagic cysts of both ovaries with discrete ‘adenomyomas’ of the posterior wall of the uterus.”96 A sagittal section of the uterus and posterior vaginal fornix showing an “adenoma of endometrial type” invading the posterior vaginal fornix and the supravaginal portion of the uterine cervix.97 A cross section of the uterine cervix and ovaries illustrating “centripetal disease” with both ovaries and the rectum densely adherent to the cervix with complete obliteration of the RVPD.98 “Multiple hemorrhagic peritoneal ‘blebs’ (adenomyomas)” on the posterior aspect of the uterus.99 A low power photomicrograph of one of the blebs.100

In describing his physical examination of such patients, Sampson stated that “palpatory finding in the culdesac [RVPD], when present, [were] the most characteristic physical signs.”101 Palpatory findings would include nodules in the uterosacral ligaments and rectum, adenomatous lesions penetrating into the posterior fornix of the vagina, as well as fixed retroflexed or retroverted uteri. Paraphrase simply cannot capture the experience of this gifted physician, only his own description of what he felt on examination and of his assessment of those findings will do justice to him. “The physical findings vary greatly…the uterus is often retroflexed or retroverted and adherent and the degree of adenomatous growth in the culdesac varies greatly in character in different cases. When slight, it is impossible to detect it. The involvement may be localized or diffuse. If localized, the area of induration may be flat or nodular, in the median line just behind the cervix, or laterally in the region of the uterosacral ligaments. The induration is usually low down, but occasionally may be higher up. Sometimes it is best detected on vaginal palpation and at other times felt best through the rectum.”102 Sampson commended Lockyer’s description of “extensive involvement of the rectogenital space by adenomyoma,”103 but no one ever described the physical findings better than Sampson.

The clinical picture – both history and physical findings – established by Sampson in 1921, with its analogy to “implantation” and the adhesive pattern of ovarian malignancy, not only captured the visual imagination of physicians but became fixed in textbooks for the next 50 years until modified by the addition of signs and symptoms, found at laparoscopy, that accompanied earlier stages of the disease. In a subsequent publication Sampson enhanced his argument regarding implantation adenomas of endometrial type originating from ovarian hematomas of endometrial type by analogizing them to pelvic implantations from ovarian carcinoma: “Implantation carcinoma of the various organs and structures of the peritoneal cavity is well recognized both by pathologists and clinicians and likewise the important part played by ovarian carcinoma as a source of these implantations. Implantation adenomas of endometrial type are analogous to those of carcinoma.”104

In 1921, Sampson believed that there were two pathologic conditions to treat, “the one present in the ovary or ovaries, and the secondary adhesions in the pelvis which are often associated with an adenoma of the endometrial type, the latter varying greatly in the degree to which it has invaded the tissues and organs involved.”105 Note that Sampson emphasized the ovarian cyst(s) and adhesions and that invasion was secondary in his thinking at this point. He believed that, as a rule, albeit with possible exceptions, “tissues of endometrial type” would stop growing and “actually atrophy” once the menopause was established.106 While Sampson respected the patients’ desires and performed conservative surgery, preserving ovarian and reproductive function in some cases, he realized that these patients might require a second surgery. To emphasize the risks involved in conservative surgery, Sampson wrote: “Casler’s case…is an extreme example of the possible results of conservative surgery in these cases.”107By his reference to Casler and his statements that follow one may infer that Sampson believed that removal of the ovaries and the disease was the treatment of choice: “To remove the pelvic adenoma and disregard the ovarian condition would be to leave the original growth behind, and furthermore, the persistence of the ovarian function might increase the growth of secondary pelvic adenomas not removed.”108 Then returning to the analogy to ovarian malignancy, Sampson stated further in italics: “Certainly we would not sanction the surgical judgment of the operator who removed the secondary peritoneal implantations of ovarian papilloma or cancer and did not remove the primary ovarian tumor. The conditions are analogous except that fortunately the adenoma of endometrial type is only rarely sufficiently invasive to cause serious damage to the parts involved.”109

Likely learning from the complications of sigmoid colon and rectal surgery experienced by Cullen, Sampson stated: “I have never resorted to the extremely radical operations, as in cancer of the uterine cervix, and even in these operations it may be impossible to remove all of the adenomatous growth.”110 To put his statement in proper context, one must remember that Sampson had extensive experience in the surgical treatment of cervical cancer.111Nonetheless, where possible Sampson avoided operating adenoma of the rectum, stating: “In the radical operations which I have employed, I have removed the entire uterus with both ovaries, and in freeing the uterus from the rectum, I have purposely kept close to the uterus, undoubtedly sometimes leaving adenoma on the rectal wall. In freeing the cervix laterally, I have kept close to it, in one case intentionally leaving adenoma in the broad ligament because it was too extensive to remove (Case 12).”112 Sampson qualified this statement by saying that only more experience would reveal whether this was the proper choice. Again, one must consider the circumstances under which Sampson operated. The average remaining life expectancy of a 30 year-old woman in 1921 was 38.15 years compared to 50.1 years in 2004.113 In other words, if a woman 30 years of age in 1921 had a hysterectomy and removal of both ovaries, on average she would be menopausal for 38.15 years. Estrogen hormone replacement therapy had not been invented. In fact, the estrogen hormone had not been identified, let alone isolated as a treatment modality. Sampson relied on surgical menopause to atrophy rectal and broad ligament endometriosis and other endometriotic deposits he left behind.

Given that “there seems to be a great variation in the degree of ‘invasiveness’ of the secondary pelvic deposits which is often evident in the individual case at operation,” Sampson did individualize his surgical treatment once the abdomen was open.114 Unfortunately, he had no reliable means to assess the extent of invasion before operating. Hence, his patients could not know before surgery whether Sampson could be conservative or if he would find it necessary to remove both ovaries. Judging from the following statement, we may assume that Sampson did discuss with his patients how he would proceed at surgery: “My present plan is to employ ovarian conservatism (excising the portion of the ovary or ovaries involved) or removing only the apparently diseased ovary in patients who desire to have the ovarian function maintained but only if the invasion of the pelvic tissues by the adenoma is slight.”115Then Sampson expressed his reservations about conservative surgery. “I am anxiously waiting to see whether the end-results will justify this stand. I am inclined to believe that ovarian conservatism is a rather dangerous experiment.”116 Sampson summarized his treatment recommendations in 1921: “In all other cases, either when ovarian conservation is not strongly desired or when the pelvic growth is apparently actually invasive, I believe that all ovarian tissues should be removed and as much as possible of the pelvic growth with it.”117

Sampson summarized his thinking in this classic paper by offering six data as evidence that perforating hemorrhagic cysts of the ovary are hematomas of endometrial type. (1) “These hematomas, as the uterine mucosa, manifest their ‘activity’ during the menstrual life of the patient. (2) Histologically, the epithelial lining of the ovarian hematomas is similar to that of the uterine hematomas, due to the retention of ‘menstrual’ blood, often present in ‘adenomyoma’ of the uterus. (3) Periodic hemorrhages occur in the ovarian hematomas which are similar in gross and histologic appearance to that of menstruating endometrium. (4) The ‘chocolate’ contents of the ovarian hematomas resemble old menstrual blood. (5) In two patients operated on at the time of the menstrual period, one the day that menstruation was due (Case 13), and the other the last day of menstruation (Case 19), the histologic changes in the ovarian ‘endometrial’ tissue corresponded to the phase of the menstrual cycle indicated by the menstrual history of the patient. (6) The fact that material escaping from the ovarian hematomas may give rise to the development of adenoma of endometrial type in the tissues thus soiled is further proof that these hematomas contain ‘endometrial’ tissue.”118

Having delivered a paper of such importance, Sampson ended with the provocative conclusion: “I cannot state that these ovarian hematomas of endometrial type are the only cause of ectopic pelvic adenomas.”119

Indeed, when Sampson published that statement, James C. Janney, a gynecologist in Brookline, Massachusetts, was in the process of reviewing 4,853 pathological specimens at the Free Hospital for Women. He found three cases of uterine tissue in the ovary which he explained as rare anomalies of development from embryonic rests.120 However, prior to publication, Janney read Sampson’s article in the Archives of Surgery of October 1921. He appended a note referring to a key element of Sampson’s article that uterine tissue in the ovary was not a rare phenomenon after all. “Note: Since the preparation of this paper, several cases of the same sort have been reported by Sampson in a paper read before the American Gynecological Society, which appeared in Archives of Surgery, October 1921. The discovery of endometrium in the wall of a large proportion of the so-called ‘hemorrhagic cysts of the ovary’ would seem to bear out the opinion expressed in this [Sampson’s] paper that the condition under discussion is not of such rare occurrence as the scarcity of reported cases would suggest.”121

Sampson presented his first theory to the American Gynecological Society on June 3, 1921. In his closing remarks to the discussion that followed this paper Sampson replied: “In answer to the questions which have been asked, I believe the growth is primary in the ovary, and not in the adenomyoma, and that the extension is from the ovary to the uterus and not from the uterus to the ovary.”122

A Historical Note by Sampson

“When these endometrial hematomas or cysts were described by me in 1921, I was not aware that they had been previously recognized and described. Three years later I found that Pick had described them in 1905 and had designated them adenoma or cystomas endometroides ovarii. Pick suggests that these cysts may be the same as Rokitansky’s cystosarcoma adenoids ovarii uterinum, described by the latter in his textbook of pathological anatomy published in 1861. Should anyone’s name be attached to these ovarian cysts, it should be Pick’s or Rokitansky’s, not mine.”123

Sampson acknowledged that Ludwig Pick had confirmed the likelihood that Rokitansky described ovarian endometriosis in 1861. Pick thought that Cystadenoma or Adenoma endometroides ovarii was a good name for chocolate ovarian cysts with the structure of the endometrium of the uterine corpus that he observed. He also thought it likely that Cystadenoma or Adenoma endometroides ovarii was identical with Cystosarcoma adenoids ovarii uterinum described by Rokitansky in 1861.124 (Italics added)

After his presentation at the American Gynecological Society, Sampson seized upon his second insight while doing further research; the powerful second theory of transtubal retrograde menstruation and implantation. However, in his rush to publish Perforating hemorrhagic (chocolate) cysts of the ovary, Sampson had sent the manuscript to the Archives of Surgery before he submitted the same manuscript to the editor of the Transactions of the American Gynecological Society. This delay gave him an opportunity to insert a historical note – the announcement of his second theory.

A Historical Announcement by Sampson

In a duplicate publication of Perforating Hemorrhagic (Chocolate) Cysts of the Ovary published in the Transactions of the American Gynecological Society for the year 1921, Sampson inserted an announcement of historic proportions, an announcement that is nowhere else to be found. “Note: Since sending in this paper for publication in the Archives of Surgery, I have had the opportunity to study material from more cases and wish to add this note. I described in this paper small hemorrhagic elevations in the ovaries of patients operated upon during the menstrual period which histologically proved to be due to hemorrhage about or into a space lined by tissue of endometrial type. I stated that I believed that they might develop into perforating hemorrhagic cysts and also might furnish the epithelium for the relining of follicular hematomas which might rupture near them. In my later studies I have found that these small ovarian hematomas may develop into larger ones, i.e., they may give rise to perforating hemorrhagic cysts. I have not as yet been able to demonstrate the relining of ruptured follicular hematomas from this source. The presence of ciliated epithelium in both the ovarian hematomas and also in the pelvic adenomas strengthens the implantation theory of the origin of the latter and weakens the serosal theory. The luteal like lining of many of these cysts could be explained as the result of hemorrhage in the walls of an endometrial cyst with consequent loss of the surface epithelium instead of the remains of a follicular hematoma.

Two possible sources of the origin of these small tubules or cysts of endometrial type in the ovary present themselves: first, congenital and second, acquired from the implantation of epithelium escaping from the tube during menstruation and its subsequent invasion of the ovary. Epithelium escaping from the tubes during menstruation and its subsequent implantation in the pelvis might also be a source of pelvic adenoma of endometrial type other than from perforating hemorrhagic cysts of the ovary.125 (Italics added)

Close on the heels of a full explication of his first theory, Sampson had announced his second theory of pathogenesis, that of retrograde menstruation; menstrual endometrium escaping through the fallopian tubes to implant and invade the ovary and pelvis. This second theory was not meant to replace the first theory; Sampson presented both theories as valid, as complementary. However, with the second theory Sampson explained the origin of ovarian adenoma of endometrial type and of perforating hemorrhagic cysts of the ovary. Both the first theory and the second theory explained the origin of pelvic adenoma of endometrial type. However, whereas the first theorybetter explained the dense pelvic adhesions as originating from the intensely irritating chocolate contents of the perforating hemorrhagic cysts of the ovary, the second theory not only explained the origin of ovarian adenomas of endometrial type and of perforating hemorrhagic cysts of the ovary, it also explained the origin of pelvic adenomas of endometrial type – including deeply invasive adenomas – when both ovaries were normal.

Taken together, Sampson’s first theory and his second theory fulfill the purpose of any scientific theory…explanation and control as explained by K. Codell Carter: “Thus we see that systematic explanations of disease phenomena depend on accepting the Distinguishability Hypothesis. There is another way of making this point. The purposes of any scientific theory are explanation and control. In a theory of disease, achieving either purpose requires universal necessary causes, and such causes depend on etiological characterizations. But etiological characterizations are possible only if the causes, in terms of which different diseases are defined are themselves distinct. Thus the formulation of etiological definitions (and hence the adequacy of any theory of disease) requires distinguishability.”126

In short, according to Sampson’s theories, the proximate cause of pelvic adenomas and adenomyomas of endometrial type was endometrial tissue derived from the ovary – the first theory – or from the uterine endometrium – the second theory; with the latter being the dominant theory. Again, K. Codell Carter clarifies: “The introduction of etiological definitions simply made what had been specific remotecauses into new proximate causes: whenever possible the essence of each disease became infection by a certain organism (say, tubercle bacilli) [endometrium] instead of a morbid alteration (for example, the formation of tubers) [adenomas and adenomyomas]. And the new proximate causes achieved in spades the same benefits as the old ones. Instead of merely explaining symptoms in terms of lesions (as the pathologists had done), researchers were now able to explain the lesions themselves, (and thereby symptoms and everything else lesions could ever explain) as well as many clinical and epidemiological facts that the pathologists could never begin to account for...The new proximate causes also provided much more effective targets for therapy and prophylaxis than did internal lesions”.127

Explication of Sampson’s Second Theory

In a paper presented to the Harvard Medical Society on February 14, 1922, Sampson explained his second theory of implantation adenomas of endometrial type, an explanation often overlooked because it was buried within the text of an article whose title only alluded to the treasure within.128

Implantation Adenomas of Endometrial Type

On February 14, 1922 Sampson awakened the community of medical scholars at Harvard University to a new disease – ovarian hematomas of endometrial type and implantation adenomas of endometrial type – when he presented a review of the subject at a meeting of the Harvard Medical Society at the Peter Bent Brigham Hospital.129 The first fruits of that teaching moment came from the pen of Joe Vincent Meigs.130 Meigs, a gynecologic oncologist inspired by Sampson’s scholarship, not only became an authority on endometriosis, but also spearheaded an interest in endometriosis that led to important contributions by him, his students, and his colleagues at Harvard, contributions that have continued without interruption into the twenty-first century.

Sampson opened his lecture praising Lockyer’s excellent monograph of 1918 that presented a substantial review of uterine and extrauterine adenomyomas. The first ten pages of this well-illustrated article, printed in the Boston Medical and Surgical Journal, recapitulated much of Sampson’s first paper in the Archives of Surgery with supplementary historical references.131 Chiari’s salpingitis isthmica nodosa of 1887 was now called “adenomyoma of the tube.” Sampson paid tribute to Cullen, his colleague and former professor at Johns Hopkins. “In 1895, Cullen described his first case of adenomyoma, and through his writing on this subject with their superb illustrations, he, more than anyone else, has demonstrated that the generally recognized adenomyoma of the uterus arises from an invasion of the uterine mucosa into the wall of the uterus. The origin of certain forms of adenomyoma of the uterus and the tube by the invasion of the mucosa lining their cavities is an established fact.” Sampson credited Baraban in 1891 and Pilliet in 1894 with the observation that adenomyoma resulted from mucosal invasion. Though recorded before Cullen, these observations had rested quietly in the medical literature until Cullen emphasized the importance of mucosal invasion in his 1896 rebuttal to von Recklinghausen’s theory of Wolffian rests.132

Then Sampson turned his attention briefly to adenomyomas that invaded the sigmoid colon, rectum, and “also those situated between the rectum and vagina; the latter are known as adenomyoma of the rectovaginal septum.” He considered the latter “the most interesting ones clinically.”133 This was followed by an excellent review of the theories of pathogenesis in which he singled out Iwanoff’s serosal theory as the most interesting, citing Lockyer’s review of its many supporters.134 So said, Sampson distanced himself from Iwanoff’s theory. “Cilia may sometimes be found on the epithelium lining the ovarian hematomas of endometrial type and likewise in the implantation adenomas. This latter fact weakens the serosal theory of the origin of ectopic adenomas of endometrial type and strengthens the implantation theory.”135 In a statement designed to raise his audience’s expectations, Sampson said: “I consider an ovarian hematoma with perforation as a frequent source of implantation adenoma of endometrial type, but possibly not the only source – as will be discussed later.”136

“Later” came after seven more pages of audience preparation and 17 magnificent illustrations. “The data which I have been able to obtain suggest that tubal and uterine epithelial cells may, under certain circumstances (as an abnormal menstruation with a back flow through the tube), be expelled from the fimbriated end of the tube and lodge on the surface of the ovary. They may become imbedded in the tissues of the ovary and, true to their type, form glands and tubules which actually invade the ovary. The process is analogous to that which results from the implantation of epithelial cells on the peritoneum from the perforation of ovarian hematomas of endometrial type, as described in the previous and also in this communication.”137 (Italics added)

Sampson casually inserted an observation that anticipated by 2 years Josef Halban’s theory of lymphatic dissemination of adenomatous tissue.138 “Adenoma is sometimes found invading the lymph vessels from these implantations (of adenomas of endometrial type), and metastases may occur from this source and explain the origin of similar growths found in the groin. I have seen a similar invasion of a lymph vessel in a primary ‘adenomyoma’ of the tube and believe that they also may occur in primary ‘adenomyoma’ of the uterus.”139

In this, his first presentation to a general academic audience, Sampson wished to leverage acceptance of his second theory of retrograde menstruation on acceptance of his more intuitive first theory of the pathogenesis of implantation adenomas of endometrial type resulting from perforating hemorrhagic cysts of the ovary. More simply put; Sampson believed his audience could more easily imagine retrograde menstruation through the relatively large perforation in the wall of a hemorrhagic cyst of the ovary than they could envision retrograde menstruation of viable endometrial fragments through the tiny proximal isthmica portion of the fallopian tube. To facilitate his audience’s (and later readers’) understanding, he presented circumstantial evidence in support of his second theory.140 “It was interesting to note the character of the implantations when there was no gross evidence of an ovarian hematoma with perforation. They were usually smaller and not as widely distributed as those generally found in the pelvis associated with ovarian hematomas with evidence of perforation. They also often presented a little different histological picture. The implantations apparently derived from the perforated ovarian hematoma are usually more active and rapidly growing. I believe that implantations from both sources may have been present in some specimens. These latter observations are, to me, the most convincing evidence that the ovarian hematomas may arise from tubal or uterine epithelium escaping from the tube (a possible result of internal menstruation).”141(Italics added)

Toward the end of his presentation, Sampson chose the powerful seed and soil metaphor to help his audience to understand that implantation adenomas “may develop wherever this epithelium falls on suitable ‘soil’.”142 He concluded his presentation with speculation regarding the origin of benign and malignant serous, endometrial and mucinous cysts of the ovary.143

Endometrial Hematomas of the Ovary

Joe Vincent Meigs was a house officer in 1921, the year Sampson published Perforating hemorrhagic (Chocolate) Cysts of the Ovary in the Archives of Surgery.144 Meigs considered it the “foremost contribution to gynecology and gynecological pathology in recent years,” a prescient observation that introduced his own first paper on the subject published just 5 months after Sampson spoke to the Harvard Medical Society.145 In possibly the first endorsement of Sampson’s first theory of the pathogenesis of adenomas of endometrial type from perforated hemorrhagic cysts of the ovary, Meigs declared:

“Sampson believes, and we believe has proved, that adeno-leiomyomata of the fallopian tube, of the round ligament, of the posterior wall of the uterus, of the posterior surface of the broad ligament, of the sigmoid, and also of the small intestine and appendix, are in some instances the results of implants from these cysts…His article gives an explanation for the many cases of severe pelvic inflammation, the etiology of which has been so obscure and puzzling heretofore.”146

So said, Meigs wanted to confirm Sampson’s work. He searched through the records of William P. Graves of Boston and the records of the Free Hospital for Women on Pond Avenue, in Brookline, Massachusetts for cases of endometrial hematomas of the ovary. Fortunately since 1903 – concurrent with the birth of surgical pathology – both Graves and the Free Hospital preserved “in formalin all gross specimens removed at operation.”147 In this retrospective case series, Meigs identified 16 cases of ovarian endometrial hematomas with pertinent history on each patient. He cut and stained fresh histologic tissue sections from the preserved specimens. To add further credence to his study, Meigs sought confirmation of the microscopic diagnosis of endometrial hematoma of the ovary “in nearly every case” from Frank B. Mallory, pathologist at the Boston City Hospital.148 Meigs described the histology in some detail with analysis that emphasized the presence and importance of hemorrhage and hemosiderin. “The endometrial-like tissue is found in various places in the ovary and in various forms, but usually it is near the perforation, and may be in the form of glands composed of columnar epithelium with an undifferentiated connective tissue stroma. Other specimens show a cuboidal, columnar, or cylindrical epithelium resting upon a thin layer of cellular tissue containing many blood vessels, and extravasated new and old blood. Still others have one of the above types of epithelium resting upon a thick layer of fibrous tissue, or even directly upon the ovarian tissue itself. The most important finding is the presence of hemorrhage, and especially the signs of old hemorrhage in a cyst wall lined with epithelium, which is not stratified. The sign of old hemorrhage is blood pigment (hemosiderin) chiefly in endothelial leucocytes, the presence of blood pigment, meaning that hemorrhage has occurred at some previous time. The blood remains in the tissue beneath the epithelium, because it cannot escape as does hemorrhage in the uterus or in the kidney, which have a passage to the outside of the body…The blood cannot escape, so it changes to blood pigment, and is later gathered up by endothelial leucocytes whose function it is to remove foreign bodies of this type.”149

Meigs confirmed all of Sampson’s findings except tubal patency. In his series, the fallopian tubes were normal in nine cases [cases 4,5,6,10,11,12,13,15,16]; the condition of the tubes was not described in one case [case 14]; one case had bilateral chronic salpingitis with open fimbriae [case 3]; and in five cases both tubes were closed [cases 1,2,7,8,9].150

He listed the various theories describing the etiology of endometrial hematomas. Contrary to his enthusiastic endorsement of Sampson’s first theory, Meigs did not endorse any of the theories of pathogenesis for endometrial hematomas of the ovary. He enumerated the following theories, evidencing no preference: (1) metaplasia of ovarian germinal epithelium; (2) Sampson’s second theory due to the implantation of endometrium reaching the ovary by way of the Fallopian tube – a theory Meigs heard directly from Sampson at the Peter Bent Brigham Hospital on February 14, 1922; and (3) the embryological theory expressed by WW Russell in 1899. With reference to the latter, Meigs discussed a paper by Janney from the Free Hospital for Women that antedated publication of Sampson’s second theory. According to Meigs, Janney postulated that:

“Uterine tissue found in the ovaries may be due to a developmental defect. The funnel, the earliest beginning of the müllerian duct, may develop on the medial instead of the lateral side of the tubal area from which the ovary arises. Tissue may become mixed when these two areas are close together and thus tissue capable of forming endometrium may be included in the ovary, later in life.”151

It is interesting to quote directly from Janney for the details. He reviewed the collection of human embryos at Harvard Medical School and consulted with Drs. Bremer and Begg of the Department of Embryology at Harvard Medical School. Dr. Bremer believed that ovarian tissue in the ovary “might be explained on the supposition of an accessory aberrant müllerian duct bud which was included in the ovary.”152 “Dr. Begg suggested that the proximity of the ‘anlagen’ of the ovary and tube in embryonic life was so great that there might be some critical period of embryonic development at which it would be possible for a tissue mixture to take place.”153

Having first consulted Drs. Bremer and Begg, Janney reviewed the Harvard collection of human embryos and made observations pertinent to the developmental theory of müllerian – endometriotic – rests. “In reviewing this series of embryos I have found cases where the funnel, instead of forming on the lateral side of the tubar area, forms on the medial side, and is thereby brought into a position much closer to the sex gland, a fact which is suggestive in view of the supposition advanced by Dr. Begg.

The funnel, which is the earliest beginning of the müllerian duct, is developed usually on the lateral side of the tubar area and separated from the genital areas by a fissure and the whole width of the tubar ridge, a distance of approximately 0.5–0.75 mm. In one case both funnels opened on the median side of the tubar area. In two other embryos one funnel opened on the medial side, and in a fourth the funnel opening seemed to extend from the lateral through to the median side of the tubar area. This unusual median position of the funnel reduces the distance between the genital gland and the müllerian duct very materially. In the cases which showed this variation the actual distances between the funnel and the genital gland at the nearest point were in four cases less than 0.1 mm (about 0.0875 mm) and in the fifth instance less than 0.2 mm (about 0.175). It is not unreasonable to suppose that tissue mixtures between the tubal and ovarian tissue could take place in cases where the distance separating the two tissues is less than 0.1 mm. I must note here that in the cases where this condition was present three of the embryos were too young to distinguish the sex with certainty. The genital glands were still undifferentiated. In the other two cases testes cords were present and the embryo was probably male. I do not feel that it would argue impossibly of such an explanation of uterine tissue in the ovary, had these been definitely male embryos, for there is nothing to show that such variations could not occur in the female. The most it could argue, I think, is that such an anomaly might someday be found in the testis also.”154

Theorists may wish to recall Janney’s caveat featuring Mark Twain. “Embryology may be suggestive but hardly conclusive, for the reason that suggestive appearances in an embryo can never be proved to be the early stages of a condition found in adults unless all of the steps can be demonstrated which seems unlikely in a condition of this rarity. Only if we are favored, like Mark Twain to the extent of seeing St. Peter’s skull at the age of seventeen and again in another museum at the time of his death can we hope to bring forth actual proof by the aid of embryology.”155

In 1951, Faulconer confirmed Janney’s hypothesis. He demonstrated in specimens from the Department of Embryology of the Carnegie Institution of Washington that the müllerian groove in human embryos “is not constant in relation to the cranial end of the Wolffian body, but is subject to much variation. The specimens examined revealed groove formation on all surfaces of the Wolffian body. These sites of invagination are designated as (1) dorsolateral, (2) lateral, (3) ventral, and (4) ventrolateral.”156 In 1998, Ludwig analyzed the embryology of the müllerian duct with respect to the pathogenesis of the Mayer-Rokitansky-Küster-Hauser [M-R-K-H] syndrome. He found that the müllerian duct develops independently of the coelomic epithelium above the mesonephros.157 Ludwig’s theory explains the pathogenesis of M-R-K-H syndrome as well as the pathogenesis of ectopic endosalpingiosis in the thorax: mediastinal paravertebral müllerian cystic endosalpingiosis, also known as Hattori cysts.158

Joe Vincent Meigs, the young clinician from Boston, appears to have been the first investigator to emphasize endometriosis-associated infertility problems. He did so by offering his own theory to explain the preponderance of ovarian endometrial hematomas in nulliparous women 30 years of age and older and their relation to infertility. “Women who bear children usually marry before 30, the age at which these cysts are first found…Provided the cyst has progressed to any extent the woman married after 30 will often be sterile. If, on the other hand, there is a microscopically small or non-adherent cyst present, and the young woman marries and has a child, the cyst may atrophy and perhaps even disappear in the stage of lactation atrophy which the genital tract frequently undergoes during the nursing period…We believe the explanation of …why they do not cause more frequent damage [at the younger age] …to be that the cysts develop very slowly. This is suggested by the fact that they usually do not appear until the patient is over 30 years of age; in other words, they are not found until 13 to 19 years of menstrual life have passed.”159

Meigs recorded that 5 of his 16 patients were between 20 and 30 years of age; 2 under age 25 years of age.160 In succeeding years, Meigs would elaborate on his theory. He would influence many young couples to marry at an early age and have children before the threat of sterility from endometrial hematomas of the ovary. Based on his assumption that the growth is probably very slow and seldom endangers life, Meigs recommended conservative surgery; but only “if all the endometrial tissue [could] be removed.” If however, all the endometrial tissue could not be removed, he recommended radical surgery; standard radical surgery for benign disease at that time meant supravaginal hysterectomy and removal of both tubes and ovaries. In 1924 Sampson would respond to some of the ideas in Meig’s paper when he addressed the life history of ovarian hematomas.161

Intestinal Adenomas of Endometrial Type

In this paper, Sampson displayed both his powers of observation and his pedagogical technique. He repeatedly asked questions of his audience and readers in the subtext of his illustrations. Then giving his readers and audience time for reflection, he answered the questions in the main text that followed. Sampson explained the delay in his appreciating the frequency of perforating hemorrhagic cysts of the ovary. He stated that he had failed to recognize the early stages of development and the later stages of regression and so overlooked many cases.162 Likewise, in a manner reminiscent of Cullen, he experienced a long learning curve before he routinely recognized intestinal adenomas of endometrial type. Sampson described and illustrated various disease patterns associated with intestinal adenomas, their laterality and asymmetry, and the intimate association with ovarian hematomas of endometrial type. However, he did not address so-called adenomyomas of the rectovaginal septum, or endometrial adenomas that penetrated into the posterior vaginal fornix. Nonetheless, interspersed within the text, Sampson provided insights into the pathogenesis of intestinal adenomas as well as the pathogenesis of the obliterated rectovaginal pouch of Douglas.

For the first time Sampson laid out the essential features of his second theory – his theory of retrograde menstruation, implantation, and invasion and the origin of ovarian hematomas of endometrial type.163I believe Sampson’s crucial insight for the second theory that he announced in the Transactions of the American Gynecological Society for the year 1921 came when he observed tubal epithelium in a minority of his cases of sigmoid adenomas of endometrial type. Invariably, in this paper Sampson gives precedence to tubal epithelium when referring to tubal and uterine epithelium. I believe Sampson’s giving precedence to tubal epithelium is important if subtle clue that his observation of tubal epithelium sparked the insight to his second theory.164

Sampson’s second theory as to the origin of these ovarian hematomas and also their relation to endometrial implantation is based on the following data. Sampson specified: “The ovarian hematomas are of endometrial type as shown by their structure, function (reaction to menstruation) and their endometrial implantations. They are rarely found in women under 30 years of age. If of developmental origin, we would expect to find them in younger women. They develop during the menstrual life of the patient in a period when tubal and uterine epithelium might escape from the fimbriated end of the tube and become deposited on the surface of the ovary just as peritoneal implantations arise from perforation of ovarian hematomas. (Italics added) In 49 cases of perforated ovarian hematomas which I have studied, the tubes were apparently patent in all, suggesting that this avenue for this source of implantation was open. These hematomas usually develop on the lateral and the under surfaces of the ovary, the portion of the ovary most likely to be soiled by material escaping from the lumen of the tube, as well seen in the ovarian adhesions found in pelvic inflammatory disease of gonorrheal origin, and they are also often bilateral. (Italics added) In 37 cases of ovarian hematomas of endometrial type with perforation, in which I have studied microscopically the tissues involved in the adhesions apparently resulting from escape of the contents of the cyst, adenoma of endometrial type was found in all but one specimen. On the other hand, in three cases of typical ovarian hematomas of endometrial type without any evidence of perforation, adhesions were not present in the pelvis and there was not any gross evidence of implantation adenomas; the pelvis was examined very carefully in each instance. In the cases of perforated ovarian hematoma with implantation, the extent of the implantation usually varied with the size of the hematoma and apparent size of the perforation. The larger the hematoma and the greater the size of the perforation the more extensive the distribution of the implantations.”165

It was not until near the end of my research when I read Sampson’s summary paper, The development of the implantation theory for the origin of peritoneal endometriosis published in 1940166 that I realized my analysis of the origin of Sampson’s second theory was incorrect.167 Sampson’s crucial insight for the second theory came not when he observed tubal epithelium in a minority of his cases of sigmoid adenomas of endometrial type. In 1940, Sampson recalled the crucial insight: “The detection of peritoneal endometriosis with and without ovarian involvement led to the second step in the development of the implantation theory. This consisted of strong circumstantial evidence indicating that bits of Müllerian tissue, derived from both the uterine and the tubal mucosa and carried by menstrual blood escaping through patent tubes into the peritoneal cavity, could become implanted on various pelvic structures including the ovaries, and the resulting perforating hemorrhagic ovarian cysts are only spectacular foci in the secondary spread of endometriosis.”168

Adenomas of endometrial type [endometriosis] have characteristics of pelvic inflammatory disease and of cancer. Having drawn a first analogy between the location of ovarian adhesions of endometrial type and the location of ovarian adhesions resulting from pelvic inflammatory disease (see italics above), Sampson drew a second analogy; that between implantations of endometrial type and implantations resulting from cancer. “Intestinal adenomas of endometrial type are implantation growths, similar in many ways to those arising from a rupture or perforation of a malignant (carcinomatous) ovarian cyst.”169 Based on his first 12 cases, Sampson made important observations that have stood the test of time and provided further circumstantial evidence to support his theory of implantation adenomas. “The portions of the intestinal tract most frequently involved are those usually found in the pelvis; as the sigmoid,170 rectum, appendix and terminal loop of the ileum.”171 He noted the asymmetry and laterality exhibited by intestinal adenomas. In six of eight cases of adenomas of the rectum and sigmoid colon the ovarian hematoma, “with evidence of a previous perforation, was situated in the left ovary.” In all four cases of appendiceal adenomas “a similar hematoma was situated in the right ovary in all four.”172

Even from such a small case series, Sampson recognized three characteristic patterns: (1) “Surface and superficial implantations”; (2) “implantations developing between folds of peritoneum and other adherent structures (pocketed implantations), best seen in the culdesac between the posterior wall of the uterus and the rectum, which are often fused together”; and (3) “the deep invasion of the underlying structure or organ. The tubules worm their way into the tissues of the intestine; and this is often associated with a marked hypertrophy of the surrounding connective tissue and muscle.”173 Sampson chose the expressive biological metaphor “worm”174 instead of the military metaphor “invade” to express the less aggressive nature of benign intestinal adenoma of endometrial type compared to cancer. He continued: “Many varieties of endometrial tissue and its derivatives may be found, including glands and tubules with and without a characteristic endometrial stroma, dilated tubules, miniature uterine cavities,175 hematomas, and the invasion of lymph vessels by endometrial polyps.”176

Sampson fashioned a “syndrome rarely furnished by any other condition,” by gathering symptoms and signs of women between the ages of 30 and the menopause who complained of “acquired dysmenorrhea or recent increase in menstrual pain…disturbance of intestinal function during menstruation,” and who on examination were found to have “a small adherent ovarian cyst or adherent ovary and palpatory findings [tenderness or nodules] in the culdesac.”177 Sampson noted that the majority of intestinal implants were insignificant but potentially invasive. But some become important when they interfere with the function of the intestine causing “marked constipation, partial obstruction, painful bowel movements and pressure sensations in the rectum during the menstrual period.”178 Sampson described the pathology and pathophysiology that leads to intestinal obstruction, once again analogous to carcinoma. “The disturbance of the function is then a mechanical one, as in carcinoma, namely, that of obstruction, and the obstruction in endometrial adenoma may be due to three factors: first, the constriction of the lumen of the bowel by growth and especially by the marked hypertrophy of the tissues surrounding the adenoma; second, by kinking the intestine, and third, by the accumulation of menstrual blood in the adenoma causing hematomas. The symptoms of obstruction may be more marked during the menstrual period, as at that time more blood may escape into the adenoma situated in the wall of the intestine…The implantation begins with the deposit of epithelium on the peritoneal surface of these structures. This epithelium sinks into the underlying tissues, and, true to its type, forms glands and tubules as shown in the previous communication. In some instances, a localized growth of endometrial mucosa arises like a polyp which may be sessile, or pedunculated, simulating the polyps found in the uterine cavity. In other instances the tubules invade the underlying tissue with very little evidence of the growth on the surface. The epithelium originally implanted may ‘die out’ or it may be covered with adhesions so that in some of the older lesions it may be impossible to determine the exact site of the original implantation. The tubules often burrow through the tissues in many directions and the portal of entry, if still present, may only be determined by cutting many sections, or better still by cutting serial sections.”179

Sampson addressed the issue of pathogenesis of intestinal adenomas when he stated that “probably the principal source” of these benign intestinal implantation adenomas of endometrial type was “epithelium escaping from an ovarian hematoma of endometrial type which has perforated”: this accorded with his first theory of pathogenesis.180 Affirming that in his experience the “majority” of intestinal adenomas of endometrial type arose from perforated ovarian hematomas, Sampson interjected his second theory to explain a minority of cases. “There is the possibility that some of them may have arisen from tubal and uterine epithelium escaping through the fimbriated end of the tube, independent of an ovarian hematoma with perforation.”181

Sampson addressed the issue of pathogenesis a second time when he asked the fundamental question: “How does this epithelium of endometrial type reach the ovary? Is it of developmental origin from the inclusion of epithelium of the müllerian or the Wolffian ducts or is it acquired during adult life?”182 Note how the debate between von Recklinghausen and Cullen lingered in the background, and had to be acknowledged before Sampson provided his evidence to the contrary. He affirmed that this epithelium of endometrial type was müllerian and acquired during adult life by employing his second theory, now expanded to include retrograde shedding of tubal epithelium as well as retrograde shedding of uterine menstrual endometrium. “The evidence which I have, at present, suggests that it is usually (possibly always) acquired from the implantation on the surface of the ovary of tubal or uterine epithelium escaping through the fimbriated end of the tube and possibly from tubal fimbriae in contact with the ovary….I consider the ovary as an intermediary host, hotbed or incubator, which may impart increased vigor and virulence to this epithelium, so that when it escapes from the ovary it may be more virulent (malignant) and invasive than before the hematoma developed and the perforation occurred. It may not be an essential intermediary host, for it is possible that pelvic implantations may arise from tubal and uterine epithelium escaping from the tube; and also implantation from both sources may be present in the same case.”183 (Italics added) Elsewhere, Sampson opined similarly: “The most natural conception of their origin would be that they arise from developmentally misplaced müllerian epithelium (Russell) or from the invasion of tubal epithelium from the fimbriae in contact with the ovary. Tubules are sometimes present in the hilum of the ovary which are apparently of Wolffian duct origin and these might be considered as a source of these hematomas. The data which I have been able to obtain suggest that tubal and uterine epithelial cells may, under certain circumstances (as an abnormal menstruation with a backflow), be expelled from the fimbriated end of the tube and lodge on the surface of the ovary.”184

However, Sampson favored the perforated ovarian adenoma of endometrial type as intermediate host because he believed incubation in the ovarian adenoma imparted added vigor or virulence to the endometrial tissue implanted there from the fallopian tube as well as accounting for the “wideness of distribution” of intestinal implants. Sampson argued: “Implantation carcinoma of the various organs and structures of the peritoneal cavity is well recognized both by pathologists and clinicians and likewise the important part played by ovarian carcinoma as a source of these implantations. Implantations adenomas of endometrial type are analogous to those of carcinoma.”185Sampson would not credit the serosal/coelomic metaplasia theory.186

The operative treatment of intestinal adenomas of endometrial type was unsettled in 1922. At surgery upon detecting an intestinal lesion, Sampson carefully examined all pelvic structures for other evidence of implantation adenomas of endometrial type, especially searching for any signs of a perforated ovarian hematoma of any size. “If evidence found indicates an adenoma of endometrial type [and not carcinoma] I do not disturb the intestinal lesion, except as it may be easily removed for histologic study, but deal with the pelvic organs as their condition requires.”187 When surgical menopause was acceptable to his patients, Sampson preferred to ignore the intestinal lesions and perform a total hysterectomy and remove both tubes and ovaries, believing that the surgical menopause resulting would accomplish the same good results that he had observed with natural menopause.188 Sampson made no mention of any adverse effects of surgical menopause. We may assume that Sampson believed spontaneous and surgical menopause to be benign physiological states. As of 1922, Sampson had encountered only one postmenopausal patient with evidence of “an undoubted ovarian hematoma of endometrial type, with its associated implantation adenomas.”189

As had Lockyer, Cullen, and others, Sampson misdiagnosed his first case of adenomyoma of the sigmoid colon (February 10, 1909) as a carcinoma and performed a segmental resection, the first of only two such resections he performed for bowel adenomas between 1909 and 1922.190 Sampson went on to describe the patterns of pelvic disease associated with intestinal adenomas of endometrial type. He illustrated four cases observed from the point of view of the surgeon: (1) a case of implantation adenoma of endometrial type involving the sigmoid colon with obliteration of the rectovaginal pouch of Douglas,191 (2) a case of ‘left frozen pelvis’ associated with partial obliteration of the rectovaginal pouch of Douglas,192 (3) another case of ‘left frozen pelvis,’ with sparing of the right tube and ovary,193and (4) a case of a completely frozen pelvis, with centripetal adherence of all structure to the uterus.194 Sampson also illustrated three cases of obliteration of the rectovaginal pouch of Douglas. In each case, giant sagittal sections were cut through the uterus, vagina, and adherent rectum showing obliteration of the rectovaginal pouch of Douglas.195 Sampson attributed the pathogenesis of obliteration of the rectovaginal pouch of Douglas to implantation adenoma of endometrial type and associated endometrial adhesive disease, not to müllerian rests or uterine mucosa “springing from” the back of the cervix or uterus as Cullen had postulated when describing the pathogenesis of adenomyomas of the rectovaginal septum.

Footnotes

1

Cullen TS. Three cases of subperitoneal pedunculated adenomyoma. Archives Surgery 1921;2:443–454. Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:245.

2

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323.

3

Casler DB. A unique, diffuse uterine tumor, really an adenomyoma, with stroma, but no glands. Menstruation after complete hysterectomy due to uterine mucosa in remaining ovary. Transactions of the American Gynecological Society. 1919;44:69–84. It is unknown when Sampson first heard of Casler’s case. He could have learned from Cullen’s presentation at the annual meeting of the New York State Medical Society at Syracuse, New York on May 7, 1919, or have read Cullen’s article in the August 1919 issue of the American Journal of Obstetrics and Diseases of Women and Children. On the other hand Sampson may have heard Casler present at the American Gynecological Society or read Casler’s article in the Transactions of the American Gynecological Society, 1919;44:69–84. What is known beyond doubt, Casler’s case of the externally menstruating ovary directly inspired Sampson’s hypothesis of the internally menstruating ovary. Casler precipitated an experiment of nature when he placed a cigarette drain in the vaginal cuff after hysterectomy. The drain created a fistula, an artificial channel connecting the left ovary and the vagina. When, 4 years later, an ovarian endometrioma ruptured – it ruptured into the fistulous tract and “menstruated” into the vagina, instead of rupturing into the abdomen and menstruating into the abdomen as perforating hemorrhagic (chocolate) cysts containing endometrial tissue are wont to do. It was Casler’s unique case that furnished the critical data that stimulated Sampson’s imagination and allowed him to make the intuitive leap from external ovarian “menstruation” to internal ovarian “menstruation” from a perforated hemorrhagic (chocolate) cyst into the abdomen, and the final intuitive leap to the deposition of endometrial tissue into the pelvis and the formation of pelvic adhesions.

4

Novak, Emil. A Note on the History of Endometriosis. Undated. Current Medical Digest, page 52. Reference obtained from the Sampson Archives at the Albany Medical College. Year and volume are not available.

5

James Conant and John Haugeland, “Editor’s Introduction.” in Kuhn, Thomas S. The Road Since Structure: Philosophical Essays, 1970–1993, with an Autobiographical Interview [Chicago, IL: University of Chicago Press, 2000], 3.

6

James Conant and John Haugeland, “Editor’s Introduction.” in Kuhn, Thomas S. The Road Since Structure: Philosophical Essays, 1970–1993, with an Autobiographical Interview [Chicago, IL: University of Chicago Press, 2000], 1. “In The Structure of Scientific Revolutions, as nearly everyone knows, Thomas Kuhn argued that the history of science is not gradual and cumulative but rather punctuated by a series of more or less radical ‘paradigm shifts.’ What is less well known is that Kuhn’s own understanding of how best to characterize these episodes itself underwent a number of significant shifts.” Thomas S. Kuhn, “The Natural and the Human Sciences,” in Kuhn, Thomas S. The Road Since Structure: Philosophical Essays, 1970–1993, with an Autobiographical Interview, edited by James Conant and John Haugeland [ Chicago, IL: University of Chicago Press, 2000], 221. In this essay, Kuhn admitted that he seldom uses the term paradigm shift, “having totally lost control of it.”

7

Casler DB. A unique, diffuse uterine tumor, really an adenomyoma, with stroma but no glands. Menstruation after complete hysterectomy due to uterine mucosa in remaining ovary. Transactions American Gynecological Society 1919;44:69–84.

8

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323. See also: Sampson JA. Perforating hemorrhagic (Chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Transactions American Gynecological Society 1921;46:162–241.

9

Cullen TS. The distribution of adenomyomas containing uterine mucosa. Am J Obstetrics and Diseases of Women and Children 1919;180:130–138:135–6. Cullen commented on cases of uterine mucosa in ovaries reported by Russell, Norris, and Casler and implicitly invited others to investigate these lesions when he stated: “In due time a sufficient number of such cases will undoubtedly be reported and then we shall be able to give a composite picture of both the clinical course and of the histological changes that occur in this most unusual group of cases.” Cullen repeated this invitation in 1920. Cullen TS. The distribution of adenomyomas containing uterine mucosa. Archives of Surgery 1920;1:215–283:264. Cullen commented on the microscopic features seen in “a photomicrograph that I have had made from one of Dr. Schwarz’s sections…It is a beautiful example of an ovary containing miniature uterine cavities….From the foregoing it is evident that in due time a sufficient number of cases will undoubtedly be reported, and then we shall possibly be able to give a composite picture of both the clinical course and of the histologic changes that occur in this most unusual group of cases.”

10

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:248. “Of physiologic interest, it is to be noted that the adenoma of endometrial type developing in the ovary and arising in the portion of the pelvis as the result of the escape of the hemorrhagic contents of the ovary may be the seat of periodic hemorrhages, i. e., they may be ‘menstruating organs’.”

11

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:249. Sampson made a point of recording his early observations in 1910 and 1912, precise to the exact date he operated each of these early cases, May 8, 1910 and March 27, 1912, as well as the exact dates he operated his later patients.

12

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:291.

13

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:291. Sampson described Case 2. “Perforating hemorrhagic cysts of both ovaries; “adenomyoma” of the posterior wall of the uterus, adherent to and invading the anterior wall of the rectum.”

14

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:250. “On March 27 of that year [1912] I removed an ‘adenomyomatous’ uterus in which the ‘adenomyoma’ had apparently extended through the posterior uterine wall and had invaded the anterior wall of the rectum…On section, the ‘adenomyoma’ was apparently not connected with the uterine mucosa. Bilateral perforating hemorrhagic cyst of the ovary were present.”

15

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:292. Sampson described Case 3. “Perforating hemorrhagic cysts of both ovaries; adherent retroflexed uterus; adenoma of endometrial type invading the posterior wall of the uterus and uniting it with the anterior wall of the rectum.”

16

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:250.

17

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary Archives of Surgery 1921;3:245–323:250. For some reason Sampson seems to have skipped Case 4, operated October 11, 1917, and Case 5, operated February 7, 1918, in his story, but report them in his chronology of cases on pages 293 and 294, respectively. I believe the explanation lies with absence of satisfactory histology of the ovaries, none was attempted in Case 4, and in Case 5 only one microscopic section was taken from the ovarian cyst which “showed a cyst with its wall lined by low and cuboidal epithelium.” No sections were taken from the posterior wall of the uterus. In Case 6 operated June 13, 1918, Sampson observed: “Perforating hemorrhagic cyst of the right ovary; adherent retroflexed uterus; ‘adenomyoma’ of posterior uterine wall; gall stones.”

18

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:295.

19

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:250–51.

20

As the title explicitly states, Sampson built this essay around gross pathology of the ovarian hematomas. He realized the full significance of chocolate cysts when he found they were partially or wholly lined by “tissue of the endometrial type” (page 247). Recognizing that ovarian cysts of this type might be “menstruating organs” (page 248), Sampson began systematic microscopic examinations of hemorrhagic ovarian cysts supplied by Emil Novak and then began microscopic examinations of his own material starting with Case 9 operated at the Albany Hospital on March 17, 1920 and continuing to his last case in this report, Case 23 operated on April 18, 1921.

21

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:251.

22

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:248. See Casler DB. A unique, diffuse uterine tumor, really an adenomyoma, with stroma, but no glands. Menstruation after complete hysterectomy due to uterine mucosa in remaining ovary. Transactions American Gynecological Society 1919;44:69–84. In 1919 Casler provided information decisive to the formation of Sampson’s initial theory of pathogenesis of pelvic endometriosis and pelvic endometriotic adhesions. Casler published a case report of an ovarian cyst which on microscopic examination was “made up almost entirely of uterine tissue.” Even more emphatically, Casler wrote of: “a large uterine growth of the ovary…the entire cyst, or uterine cavity, as it really is, is lined throughout by a single layer of tall columnar epithelium of the uterine type, and in places cilia can be made out.” Then Casler explained how regular “menstruation” could happen every month after total hysterectomy. “It is a natural process then that the uterine glands in the ovarian cyst should take on the active work of the uterus and maintain menstruation regularly.” In sum, Casler had presented a case of ovarian menstruation through the vagina: external ovarian menstruation. I believe Casler’s morphologic description and his physiologic reasoning planted the seed that led Sampson to imagine “internal ovarian menstruation.” I believe Sampson’s initial theory of the pathogenesis of pelvic endometriosis and pelvic endometriotic adhesions from perforating hemorrhagic (chocolate) cysts of the ovary was inspired by Casler’s observations.

23

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:255.

24

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:251–252.

25

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:253.

26

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:252.

27

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:248. I prefer the descriptive anatomical term rectovaginal pouch of Douglas to cul-de-sac which means literally – a blind diverticulum.

28

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:249.

29

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:249.

30

Late in the twentieth century, Dr. Dan Martin and Professor Philippe R. Koninckx found that only in about 1 in 450 cases of deeply invasive endometriosis of the rectovaginal pouch of Douglas did the lesion actually invade through the floor of the rectovaginal pouch into the true anatomic rectovaginal septum of Denonvilliers. Personal communication from Dan Martin to Ronald E. Batt, October 20, 2007.

31

Novak E. Hematomata of the ovary including corpus luteum cysts. Bulletin Johns Hopkins Hospital 1917;28:349–354.

32

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:254.

33

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:247.

34

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:254.

35

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:255–256.

36

Runge E. Ueber die Veranderunger der Ovarien bei Syncytralen Tumoren und Blasenmole; Zugleich ein Beitrag zur Histogenese. Arch f. Gynak. 1903;69:33–70.

37

Wolf EH. Ueber Haematoma Ovarii. Arch f. Gynak 1908;84:211–243.

38

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:256–257.

39

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:257.

40

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:257.

41

Cuthbert Lockyer, Fibroids and Allied Tumours (Myoma and Adenomyoma): Their Pathology, Clinical Features and Surgical Treatment [London: Macmillan and Company, 1918]. In 1898, N.S. Iwanoff published his theory that glandular cystic spaces in fibromyomas originated by an ingrowth of overlying serosa. [Iwanoff NS. “Drusiges cystenhaltiges Uterusfibromyom compliciert durch Sarcom und Carinom.” Monatsschr fur Geb und Gynak 1898; Bd. vii: S. 295.] Iwanoff claimed that, in a paper published previously in Russia, he had demonstrated microscopically that glandular structures in an adenomyoma were derived from the serosal epithelium. (Lockyer pp 292–3) Iwanoff also believed the carcinoma he observed within the adenomyoma resulted from malignant changes within the adenomyoma. Accordingly he labeled the lesion Adeno-fibromyoma cysticum sarcomatodes carcinomatosum. This important paper of Iwanoff has been accepted as the origin of the theory of coelomic metaplasia. Later, Robert Meyer introduced the analogous concept of “epithelial heterotopy.” “Meyer showed as many other observers since have done, that epithelial heterotopy or displacement can occur in the serosa as well as in the mucosa.” (Lockyer: 293) Lockyer stated his own position [his italics] of the pathogenesis of extrauterine endometriosis: “Heterotopy of serosal epithelium is the probable explanation of the existence of the epithelial spaces and cysts in most of the extrauterine swellings found between the rectum and genital tract.” (Lockyer: 295) Lockyer contended that many reliable observers – without the possibility of doubt – had proved by “repeated investigations” the transformation of flattened “so-called ‘endothelium’ of the peritoneum” to be transformed into cylindrical and columnar epithelium under the excitation of inflammation or the influence of pregnancy. (Lockyer: 295, 299. Lockyer quotes from Klages R. Zeitschr fur Geb und Gynak 1912; Bd. lxx: S. 858. “that the transition of flat peritoneal epithelium into cubical or cylindrical can occur, has been repeatedly proved, and notably so by Opitz and Robert Meyer.” Lockyer states that “Opitz had found that where the peritoneum lies in natural folds, as it does at the tubal angles, the initial condition already exists for the down-growth of epithelial elements.”) Such a positive statement from an authority of the caliber of Lockyer serves to demonstrate the strength of the theory of coelomic metaplasia and its hold on medically sophisticated investigators of the World War I era. The theory of coelomic metaplasia was well established before Sampson began his investigations and would remain a powerful alternative to his theory of pathogenesis during his lifetime.

42

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:258–259.

43

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:259–264.

44

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:261–262.

45

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:262.

46

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:263.

47

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:263.

48

Michael Worboys, Spreading Germs: Disease Theories and Medical Practice in Britain, 1865–1900 [Cambridge, UK: Cambridge University Press, 2000], 6.

49

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary Archives of Surgery 1921;3:245–323:264.

50

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:263.

51

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:263.

52

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:266.

53

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:264.

54

Savage S. Hematoma of the ovary and its pathological connection with the ripening and retrogression of the graafian follicle. Brit Gynaec J 1906;l21:285–305.

55

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:265.

56

Hedley JP. Hematoma of the ovary with report of 18 cases. J Obstet Gynec Brit Empire 1910;18:293–311.

57

In 1943, James Robert Goodall would describe in detail all the host responses to endometriosis.

58

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:246. Sampson operated 14 cases between May 1, 1920 and May 1, 1921. The incidence was 14 cases in 178 operations. What first drew Sampson’s attention to the importance of perforating hemorrhagic (chocolate) cysts of the ovary was not only their frequency (p. 246), but more importantly the “nature of the adhesions resulting from the escape [of very irritating … chocolate] contents into the peritoneal cavity.” (pp. 245–6) The adhesive pattern reflected the effect of gravity within the pelvis: the “most extensive and densest adhesions [were] usually found in the culdesac uniting the supravaginal portion of the cervix and lower portion of the posterior wall of the uterus to the bottom of the culdesac and the anterior rectal wall.” (p. 246.)

59

It is not by accident that Sampson placed the most severe disease in Pattern One. Most cases were severe.

60

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:268.

61

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary Archives of Surgery 1921;3:245–323:269.

62

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:269.

63

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:270.

64

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:271–272. See Figure 68, page 314 and Figure 70, page 316. Figure 70 (Case 22) is a low power photomicrograph of a bleb “lined by columnar cells resting on a cellular stroma” and filled with menstrual debris.

65

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:316. Figure 70 (Case 22).

66

Cuthbert Lockyer, Fibroids and Allied Tumours (Myoma and Adenomyoma): Their Pathology, Clinical Features and Surgical Treatment [London: Macmillan and Company, 1918], 295, 296.

67

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:272.

68

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:274.

69

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:274.

70

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary Archives of Surgery 1921;3:245–323:274.

71

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:275.

72

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:276.

73

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:276.

74

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:276.

75

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:277.

76

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:277.

77

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:277.

78

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:278.

79

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:319.

80

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:279.

81

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:280.

82

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:319.

83

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:280.

84

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:281.

85

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:284.

86

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:281–282.

87

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:283.

88

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:284.

89

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:285.

90

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:285.

91

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:247.

92

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:253.

93

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:259.

94

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:272.

95

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary> Archives of Surgery 1921;3:245–323:279.

96

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:289, 300 and 308. See Figure 43 (Case 17) on page 289, Figure 54 (Case 19) on page 300, and Figure 60 (Case 12) on page 306.

97

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:298.

98

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:301.

99

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:314.

100

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:316.

101

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:285.

102

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:285–286.

103

Cuthbert Lockyer, Fibroids and Allied Tumours (Myoma and Adenomyoma): Their Pathology, Clinical Features and Surgical Treatment [London: Macmillan and Company, 1918].

104

Sampson JA. Intestinal adenomas of endometrial type: their importance and their relation to ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary). Archives of Surgery 1922;5:217–280;224–225.

105

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:287.

106

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:287.

107

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:287–288.

108

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:287.

109

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:287–288.

110

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:289.

111

Clement, PB. History of Gynecologic Pathology IX: Dr. John Albertson Sampson. International Journal of Clinical Pathology 2001;20:86–101. In the bibliography of Sampson’s works following this scholarly biographical essay, Clement lists many articles of Sampson that pertain to radical surgery for cervical cancer and to the pathology of cervical cancer.

112

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:289–290.

113

U. S. Department of Health and Human Services Centers for Disease Control and Prevention: National Center for Health Statistics, National Vital Statistics System. United States Life Tables, 2004. National Vital Statistics Reports: Volume 56, Number 8, December 28, 2007, Page 30. Table 11. Life expectancy by age, race, and sex: Death-registration states, 1900–1902 to 1919–1921, and United States, 1929–1931 to 2004.

114

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:290.

115

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:290.

116

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:290–291.

117

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:291.

118

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323:322–323.

119

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Archives of Surgery 1921;3:245–323:323. Nowhere in this paper did Sampson mention transtubal menstrual dissemination of endometrial tissue into the peritoneal cavity.

120

Janney JC. Report of three cases of a rare ovarian anomaly. Am J Obstet Gynecol 1922;Feb:173–187.

121

Janney JC. Report of three cases of a rare ovarian anomaly. Am J Obstet Gynecol 1922;Feb:173–187: 187.

122

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary. Am J Obstet Gynecol 1921: 2:526–33. Note the discussion of his paper delivered at the American Gynecological Society meeting was published in the American Journal of Obstetrics and Gynecology while his paper was published in the Archives of Surgery in 1921 and also in the Transactions of the American Gynecological Society in 1921, in the later instance with the addition of a historical note of great importance.

123

Sampson JA. Heterotopic or misplaced endometrial tissue. Am J Obstet Gynecol 1925;10:649–664:655. Pick L. Arch f Gynaek 1905;lxxvi:251–275. Sampson was referring to Meigs who initiated the use of the term “(Sampson’s cyst),” and undoubtedly others. See Meigs JV. Endometrial hematomas of the ovary. Boston Med Surg J 1922;clxxxvii:1–13:10, 12.

124

Pick L. Arch f Gynaek 1905;lxxvi:251–275:261–262. “Ein- oder mehrfache im Ovarium verstreute Cysten mit syrupos-blutigem, chocoladenbraunem oder rothlichem Inhalt und schleimhautahnlicher pigmentirter Auskleidug weisen makroskopisch auf diese Form des Adenomas, das ein Adenoma oder Cystadenoma ovarii vom Bau des Endometrium corporis uteri darstellt und kaum einen treffenderen Namen erhalten kann als den eines Adenoma endometroides ovarii).1” [Footnote 1] “Vielleicht ist diese Geschwulstform identisch mit dem alten Rokitansky-schen Cystosarcoma adenoides ovarii uterinum. Lehrb. D. pathology. Anatom. III. Aufl. Bd. III. 1861. Wien. S. 423, S431.” On examination of a copy of Rokitansky’s contribution: Lehrbuch der Pathologischen Anatomie 1855–61. III:475–490, reference to page S 431 is actually a reference to a running heading at the top of page 475 and reads “Rokitansky Uterusdrüsen – Neubildung. 1. c. (S. 431).” Since the copy I possess runs from pp. 475–490, I do not have the earlier pages which would contain the reference to S. 23.

125

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Transactions of the American Gynecological Society 1921;46:162–241:235–6.

126

K. Codell Carter, The Rise of Causal Concepts of Disease: Case Histories [Burlington, VT: Ashgate, 2003], 106. See also pages 199 and 200. “Between about 1830 and 1880, medicine reorganized itself around the concept of universal necessary causes…The etiological research programme [achieved an] enormous increase in explanatory power.”

127

K. Codell Carter, The Rise of Causal Concepts of Disease: Case Histories, 107.

128

Sampson JA. Ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) and implantation adenomas of endometrial type. Boston Medical and Surgical Journal 1922;186:445–56.

129

Sampson, JA. Ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) and implantation adenomas of endometrial type. Boston Med Surg J 1922;186:445–456.

130

Meigs, JV. Endometrial hematomas of the ovary. Boston Med Surg J 1922:187:1–13.

131

Sampson JA. Boston Med Surg J 1922;186:445–456.

132

Sampson JA. Ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) and implantation adenomas of endometrial type. Boston Med Surg J 1922;186:445.

133

Sampson JA. Boston Med Surg J 1922;186:445–456;445.

134

Sampson JA. Boston Med Surg J 1922;186:445–456:446.

135

Sampson JA. Boston Med Surg J 1922;186:445–456:447.

136

Sampson JA. Boston Med Surg J 1922;186:445–456:448.

137

Sampson JA. Boston Med Surg J 1922;186:445–456:455.

138

Halban J. Hysteroadenosis metastatica. (Die lymphogene Genese der sog. Adenofibromatosis heterotopica.) Wiener klinische Wochenschrift 1924;37:1205–6.

139

Sampson JA. Ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) and implantation adenomas of endometrial type. Boston Med Surg J 1922;186:445–456:448.

140

Sampson JA. Boston Med Surg J 1922;186:445–456:455.

141

Sampson JA. Boston Medical and Surgical Journal 1922;186:445–456.:456.

142

Sampson JA. Ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) and implantation adenomas of endometrial type. Boston Medical and Surgical Journal 1922;186:445–456:456.

143

Sampson JA. Boston Med Surg J 1922;186:445–456:456.

144

Sampson JA. Perforating hemorrhagic (chocolate) cysts of the ovary: their importance and especially their relation to pelvic adenomas of endometrial type (“Adenomyoma” of the uterus, rectovaginal septum, sigmoid, etc.) Archives of Surgery 1921;3:245–323.

145

Meigs, JV. Endometrial hematomas of the ovary. Boston Med Surg J 1922:187:1–13:1

146

Meigs, JV. Boston Med Surg J 1922:187:1–13:2.

147

Meigs, JV. Boston Med Surg J 1922:187:1–13:2.

148

Meigs, JV. Boston Med Surg J 1922:187:1–13:6.

149

Meigs, JV. Boston Med Surg J 1922:187:1–13:1. Endothelial leucocytes are part of the innate immune system operative at birth. See: Richardson AC, Carpenter MW. Inflammatory mediators in gestational diabetes mellitus. Obstet Gynecol Clin N Am 2007;34:213–224:216.

“The immune system defends its host against both external threats, such as bacterial infection, viral infection; physical injury, and internal threats such as malignant transformation. The immune system has historically been divided into two parts: the innate and adaptive. They are separated purely for descriptive purposes and are not mutually exclusive of one another. From an evolutionary standpoint, the innate immune system predates the adaptive immune system. The innate immune system is considered to be the ‘first-line’ of defense against microbes or tissue damage. The adaptive immune system is activated by the innate immune system and responds to antigens to which the organism has already been exposed, thereby providing the ability to mount a more effective response.”

150

Meigs, JV. Endometrial hematomas of the ovary. Boston Med Surg J 1922:187:1–13.

151

Meigs, JV. Boston Med Surg J 1922:187:1–13:3.

152

Janney JC. Report of three cases of a rare ovarian anomaly. Am J Obstet Gynecol 1922;Feb:173–187:180.

153

Janney JC. Report of three cases of a rare ovarian anomaly. Am J Obstet Gynecol 1922;Feb:173–187:181.

154

Janney JC. Am J Obstet Gynecol 1922;Feb:173–187:182–184.

155

Janney JC. Am J Obstet Gynecol 1922;Feb:173–187:187.

156

Faulconer RJ. Observations on the origins of the müllerian groove in human embryos. Contrib Embryol 1951;229:161–164:161.

157

Ludwig KS. The Mayer-Rokitansky-Küster syndrome. An analysis of its morphology and embryology. Part II: embryology. Arch Gynecol Obstet 1998;262:27–42.

158

Batt RE. Mhawech-Fauceglia P, Odunsi K, Yeh J. Pathogenesis of mediastinal paravertebral müllerian cysts of Hattori: developmental endosalpingiosis-müllerianosis. Int J Gynecol Pathol 2010;29:546–561.

159

Meigs, JV. Endometrial hematomas of the ovary. Boston Med Surg J 1922:187:1–13:4, 3.

160

Meigs, JV. Boston Med Surg J 1922:187:1–13:4.

161

Sampson JA. Benign and malignant endometrial implants in the peritoneal cavity, and their relation to certain ovarian tumors. Surg Gynecol Obstet 1924;38:287–311.

162

Sampson JA. Intestinal adenomas of endometrial type: their importance and their relation to ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) Archives of Surgery 1922;5:217–280:218.

163

Sampson JA. Intestinal adenomas of endometrial type. Archives of Surgery 1922;5:217–280:261. For the first time Sampson refers to his “theory.” “This theory as to the origin of these ovarian hematomas and also their relation to endometrial implantations is based on the following data.”

164

Sampson JA. Intestinal adenomas of endometrial type. Archives of Surgery 1922;5:217–280.

165

Sampson JA. Intestinal adenomas of endometrial type. Archives of Surgery 1922;5:217–280:261–262. Regarding size of ovarian hematomas of endometrial type, see pages 217–8. “The size of these hematomas was described in the previous paper as being usually between 2 and 4 cm. in diameter, occasionally less than 2 cm. and also occasionally larger than 4 cm. … I would modify the foregoing statement in regard to the size by adding that they are often so small and inconspicuous that they may be easily missed both at the time of the operation and in the pathology laboratory.”

166

Sampson JA. The development of the implantation theory for the origin of peritoneal endometriosis. Am J Obstet Gynecol 1940;40:549–557.

167

The historian analyzing prospectively runs into similar problems encountered by the clinician and scientist analyzing prospectively.

168

Sampson JA. Am J Obstet Gynecol 1940;40:549–557:555.

169

Sampson JA. Intestinal adenomas of endometrial type: their importance and their relation to ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) Archives of Surgery 1922;5:217–280:277.

170

Sampson JA. Intestinal adenomas of endometrial type: their importance and their relation to ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) Archives of Surgery 1922;5:217–280:240. See Figure 27 (case 8) on page 240. This is a good illustration of the kinking of the sigmoid colon cause by implantation adenomas of the sigmoid colon.

171

Sampson JA. Archives of Surgery 1922;5:217–280:277. See also: 218, 225. See Figure 59 (Case 12) This excellent illustration shows the ileum adherent to an adenoma of the posterior uterine fundus. The caption gives Sampson’s interpretation of the pathogenesis of the lesion. “My interpretation of the etiology of this condition is as follows. At the previous operation, 4 years ago, some of the epithelium lining the hemorrhagic cyst of the left ovary became implanted on the posterior surface of the uterus and other portions of the pelvic contents and developed into implantation adenomas of endometrial type. The ileum became adherent to the implantation on the posterior surface of the uterus and was superficially invaded by it. The uterine wall was invaded to a much greater extent as indicated.”

172

Sampson JA. Archives of Surgery 1922;5:217–280:278.

173

Sampson JA. Archives of Surgery 1922;5:217–280:278. Hypertrophy of surrounding tissues is a good example of the host response to the intrusion of the “tubules” of the adenoma.

174

Sampson JA. Intestinal adenomas of endometrial type: their importance and their relation to ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) Archives of Surgery 1922;5:217–280. See Figure 42 (Case 2) on page 251. “Section of the wall of the sigmoid which was excised. It shows a typical adenoma of endometrial type. The implantation apparently began on the peritoneal surface, possibly through an epiploic appendage, indicated by the arrow, and invaded the subserosa and then wormed its way through the muscularis forming hematomas.” See also Figure 46 (Case 1) on page 254. “Section of the wall of the sigmoid showing an adenoma of endometrial type invading it…The adenoma first invaded the subserosa and then wormed its way through the muscularis forming a hematoma.”

175

Sampson JA. Intestinal adenomas of endometrial type: their importance and their relation to ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) Archives of Surgery 1922;5:217–280:252. See Figure 43 (Case 2) “A small uterine cavity found in a section of the wall of the sigmoid.”

176

Sampson JA. Archives of Surgery 1922;5:217–280:278. See Figure 44 (Case 2) on page 252. “Endometrial polyp invading a lymph vessel in the submucosa of the sigmoid. These polyps are frequently found in implantation adenomas of endometrial type.” See also Figure 45 (Case 2) on page 253. “Implantation adenoma (a) on the surface of the broad ligament and invading a lymph vessel in the broad ligament. Two endometrial polyps (e, p) are shown, the large one came from the adenoma (b) of this illustration. The invasion of the lymph vessels by these adenomas suggests that they may metastasize through these channels and offers one explanation for the appearance of adenoma in the groin.” Here Sampson has identified endometrial invasion of lymph vessels 2 years before Halban published his theory of lymphatic metastases in 1924.

177

Sampson JA. Archives of Surgery 1922;5:217–280:279.

178

Sampson JA. Archives of Surgery 1922;5:217–280:243–4.

179

Sampson JA. Archives of Surgery 1922;5:217–280:227–9.

180

Sampson JA. Archives of Surgery 1922;5:217–280:279.

181

Sampson JA. Archives of Surgery 1922;5:217–280:279.

182

Sampson JA. Intestinal adenomas of endometrial type: their importance and their relation to ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) Archives of Surgery 1922;5:217–280:280. See also Figure 10 (case 4) page 226. “I believe that this adenoma of the uterus did not arise from the direct invasion of the uterine mucosa of the uterine cavity or from developmental inclusions of müllerian epithelium or from a metaplasia of the peritoneal mesothelium; but I believe it arose from the implantation of epithelium from the lining of a hemorrhagic cyst of the ovary which had perforated.”

183

Sampson JA. Archives of Surgery 1922;5:217–280:280. Note how often Sampson mentions tubal epithelium. See text page 258. “The data which I have been able to obtain suggest that tubal and uterine epithelial cells may, under certain circumstances (as an abnormal menstruation with a backflow), be expelled from the fimbriated end of the tube and lodge on the surface of the ovary.”

184

Sampson JA. Archives of Surgery 1922;5:217–280:258–9.

185

Sampson JA. Archives of Surgery 1922;5:217–280:224–225. See also Figure 9 (Case 4) page 226. “Photomicrograph of an implantation adenoma (of endometrial type) on the surface of the left tube…Histologically it resembles normal endometrium. It is analogous to the implantation carcinoma shown in Figure 2.” (page 220).

186

Sampson JA. Archives of Surgery 1922;5:217–280:242. See Figure 29 (Case 8) on page 242. “Implantation adenoma invading the wall of the uterus. Photomicrograph of a portion of the uterine wall through one of the pits. The arrow indicates the bottom of the pit between the adhesions. The adenoma on the surface of the uterus is here shown invading the wall of the uterus.” See also Figure 30 (Case 8) page 242. “Implantation adenoma on the surface of the uterus. The photomicrograph shows a polypoid condition of the endometrium lining the bottom of a wide pit which has been exposed by freeing the uterus from the adherent sigmoid colon.” Finally, see Figure 60, page 271. “Adenoma of endometrial type of the posterior uterine wall and superficially invading the wall of the ileum which is fused to the uterus at this place…It only superficially invaded the wall of the intestine but has extensively invaded the uterine wall, giving rise to a typical so-called adenomyoma of the uterus, not arising from the direct invasion of the uterine mucosa from the uterine cavity or from the developmental inclusions of müllerian epithelium in the uterine wall or from a metaplasia of the peritoneal mesothelium but from the implantation of endometrial epithelium from the epithelia lining of a perforated hemorrhagic cyst of the ovary (of endometrial type), as probably the majority of the ectopic pelvic adenomyomas (of endometrial type) shown in this and the previous communication arose.”

187

Sampson JA. Archives of Surgery 1922;5:217–280:279. Ibid: 249.

188

Sampson JA. Intestinal adenomas of endometrial type: their importance and their relation to ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) Archives of Surgery 1922;5:217–280:279.

189

Sampson JA. Archives of Surgery 1922;5:217–280:217.

190

Sampson JA. Archives of Surgery 1922;5:217–280:219–220.

191

Sampson JA. Archives of Surgery 1922;5:217–280:222. See Figure 4 (Case 4).

192

Sampson JA. Archives of Surgery 1922;5:217–280:230. See Figure 14 (case 5). See also Ronald E. Batt, “Conservative and complete operations by laparotomy,” in Text and Atlas of Female Infertility Surgery, ed. Robert B. Hunt, 3rd ed. [St. Louis, MO: Mosby, 1999], 412–439:424–5. “As reproductive surgeons treat increasingly severe cases of endometriosis they encounter the left-frozen pelvis. A large ovarian endometrioma is fused to the broad ligament and ureter, with the oviduct sandwiched between the ovary and broad ligament, or adherent to tubal or antimesenteric border of the ovary, and the whole completely enveloped by sigmoid colon in obliterative adhesive disease.”

193

Sampson JA. Archives of Surgery 1922;5:217–280:239. See Figure 26 (case 8) page 239.

194

Sampson JA. Intestinal adenomas of endometrial type: their importance and their relation to ovarian hematomas of endometrial type (perforating hemorrhagic cysts of the ovary) Archives of Surgery 1922;5:217–280:234 See Figure 19 (case 10) on page 234. The completely frozen pelvis with centripetal adherence of all pelvic organs to the uterus is the most devastating form of pelvic adenomata of endometrial type from the viewpoint of the infertile couple. Ronald E. Batt, “Abdominopelvic diagnostic laparoscopy,” in Text and Atlas of Female Infertility Surgery, ed. Robert B. Hunt, 3rd ed. [St. Louis, MO: Mosby, 1999], 372–385:376–377. Complex disease patterns: (1) Partial and complete obliteration of the rectovaginal pouch. “Partial obliteration of the rectovaginal pouch usually indicates deep, nodular, invasive disease of one uterosacral ligament and sometimes of pararectal or rectal tissue adherent to it. Complete obliteration usually is associated with deep, nodular, invasive disease of both uterosacral ligaments, posterior cervix, and rectum…Centrifugal pattern: The predominant pattern of endometriosis is centrifugal, with adnexa adherent laterally to posterior broad ligaments with or without obliterative disease of the rectovaginal pouch … Centripetal: The centripetal pattern is a less common but more severe form of disease. The uterus is retroflexed, retroverted, and adherent to itself and to the rectum, with complete obliteration of the rectovaginal pouch and adherence of both ovaries to the posterior uterus and side of the rectum…Left frozen pelvis: This pattern is characterized by a large ovarian endometrioma fused to the broad ligament over the ureter, with the oviduct adherent between the ovary and broad ligament or adherent to the tubal pole and antimesenteric border of the ovary, and the whole enveloped by the sigmoid colon in dense obliterative adhesions. It develops spontaneously when the sigmoid colon envelops the left adnexa to contain chocolate debris from repeated ruptures of left ovarian endometriomas. It also may develop in response to surgical intervention. In both instances, the left adnexa often is damaged irreparably. Intravenous pyelogram is recommended to detect partial or complete obstruction of the left ureter.,,, Complete frozen pelvis: Frozen pelvis represents the most complex and severe pattern. It is the final expression of aggressive endometriosis, impaired host immune defenses, and often numerous attempts at medical and surgical treatment. The patient’s health may be threatened.”

195

Sampson JA. Archives of Surgery 1922;5:217–280. See Figure 18 (Case 5), page 233. Sagittal section of the myomatous uterus and adjacent pelvic structures indicating the condition present prior to the operation. The adenoma of endometrial type is shown fusing the cervix to the rectum and superficially invading these structures. Also illustrated is a sagittal section of a hematoma of the left ovary. See also Figure 25 (Case 10) page 238. “Condition prior to the operation, as seen in sagittal section of the uterus and adjacent structures. The implantation adenoma lodging and growing in the culdesac has invaded both the uterus and the rectum fusing these parts, and has extended downward between the rectum and the vagina to the right of the cervix (the perforated hematoma was in the right ovary) forming a tumor which could be distinctly felt before operation both on vaginal and rectal palpation.” See also Figure 41 (Case 2) on page 250.