You're using an outdated browser. Please upgrade to a modern browser for the best experience.
Please note this is a comparison between Version 1 by Marwan Habiba and Version 2 by Jason Zhu.
Many challenges remain towards realizing a satisfactory classification of adenomyosis. A major issue is the disagreement on the definition of adenomyosis when using histology as a gold standard. A uniformly agreed reporting system may be the first step towards a classification that takes symptoms into account.
- Adenomyosis
- Histopathology
- Classification
- Uterus
1. Defining Adenomyosis
In 1925, Frankl proposed the term ‘adenomyosis’ to distinguish the condition [1][17]. Since then, there have been numerous attempts to reach a consensus on how to define and classify adenomyosis.
1.1. Histopathology
Early definitions focused on the presence of endometrial tissue within the myometrium in hysterectomy specimens[2][18]. According to Lockyer [3][19], the first detailed description of the lesion researchers today identify as adenomyosis was made by Babeș (Victor Babesius) who presented a case of an intramural ‘myoma’ containing cysts lined with ‘low cubical epithelium derived from embryonic germs’ in 1882 [4][20]. Oskar Frankl [1][17] distinguished mucosal invasion within the myometrium and used the term ‘adenomyosis uteri’. Bird [5][10] defined adenomyosis as “the benign invasion of endometrium into the myometrium, producing a diffusely enlarged uterus which microscopically exhibits ectopic non-neoplastic, endometrial glands and stroma surrounded by the hypertrophic and hyperplastic myometrium”. This definition remains widely accepted. The classic diagnosis of adenomyosis relied on the identification of heterotopic endometrial glands and stroma within the myometrium. Myometrial hypertrophy, although itself difficult to define, is often viewed as relevant to gross or microscopic identification [6][21]. Whilst most discussions about adenomyosis and its diagnosis focus on the more common diffuse form, less commonly, adenomyosis can feature as a localized lesion within the myometrium (localized or discrete from).
The distinction between normal and adenomyosis has been debated over the years. Early writings viewed the condition as very rare and included descriptions of grossly enlarged and distorted uteri, but this was followed by a rapid rise in reported cases accompanied by calls for more strict diagnostic criteria and warnings against over-diagnosis [7][22].
One important aspect at the core of the debate is the almost universal irregularity at the endometrial myometrial interface. This poses a challenge in defining the cut-off point of what is normal. The notion of endometrial invasion refers to the vertical depth of endometrium presence below the mucosa and into the myometrium. The second significant challenge stems from the often-patchy distribution of adenomyosis within the myometrium. This means that the frequency of diagnosis can be related to the number of histological sections examined. This can be labor intensive and, as such, may not be pursued in routine clinical practice as it has no prognostic value once the uterus is removed [5][10].
As will be discussed below, definitions will remain subjective unless identified features can be linked to specific clinical manifestations. This is challenging because the symptoms of adenomyosis are not pathognomonic. The uncertainty is reflected in the wide variation in reported cases of adenomyosis in hysterectomy specimens. This can be as high as nine-fold between reporting histopathologists [8][13]. It is also important to recognize that symptoms may not correlate with the depth of invasion or the extent of disease. Minimal disease or ‘adenomyosis sub-basalis’ which may be excluded by definitions that adopt a more conservative cut-off point, has been linked to symptoms [5][9][10,23]. This creates additional complexity in research in adenomyosis as it raises the question of the appropriateness of both the study and the chosen control groups. Most published literature adopted the customs of the local histopathology laboratory as a reference point. These included the depth of microscopic high or low power fields, as a proportion of the myometrial thickness or, rarely, in terms of millimeters below the endometrial myometrial junction [7][22]. Given the high degree of uncertainty when applying the ‘gold standard’ definition, it is perhaps surprising that some of the studies that employed image-based diagnosis have claimed a high degree of accuracy. Myometrial hypertrophy that can account for uterine symmetrical enlargement linked to adenomyosis provided an additional feature to the clinical impression.
There is little research on the correlation between symptoms and the depth or extent of adenomyosis. The little evidence that is currently available suggests that dysmenorrhea is linked to the depth and extent of myometrial involvement, but that heavy periods are more common in women with less invasive disease [5][10]. All of this adds weight to the argument that histopathology should aim to report the actual depth of gland involvement rather than the presence or absence of disease. Reliance on histology meant that adenomyosis could, till recently, be only diagnosed in uteri removed surgically, most often in women who are symptomatic. Cullen referred to menstrual abnormalities and pain as the ‘expected’ symptoms and went on to state that the diagnosis could be readily made clinically including by his assistant [10][24]. Infertility is now recognized to be linked to adenomyosis.
1.2. Imaging
A new dimension to the diagnosis of adenomyosis came with the introduction of ultrasound and magnetic resonance imaging (MRI). Early efforts date back to the use of Gray Scale ultrasonography [11][25]. Hricak et al. [12][26] demonstrated the specific appearance on MRI of inner myometrial junctional zone (JZ), positioned between the endometrium and the outer myometrium, but did not link this to adenomyosis. Lee et al. [13][27] reported the identification of adenomyosis using MRI in hysterectomy specimens. Mark et al. [14][4] demonstrated the use of MRI in differentiating adenomyosis from fibroids. It has been pointed out that histological diagnosis relies on the detection of the ectopic glandular elements within the myometrium (the ‘adenosis’ component), whereas imaging relies on aberration in the appearance of the muscle (the ‘myosis’ component).
Non-invasive diagnosis offers an opportunity for the study of the natural history of adenomyosis, but efforts in this field will necessarily be hampered because of the limitation of resolution and because of the lack of agreement on the histological diagnostic ‘gold standard’. The recognition of increased myometrial JZ thickness as a marker of adenomyosis provided an important clue to diagnosis. However, in line with similar features, the cut-off point for diagnosis will necessarily be rooted on a balanced probability of detection.
- (a) Magnetic resonance imaging (MRI)
MRI showed promise for the non-invasive classification and interest in its use has grown over the last three decades. Critical in MRI diagnosis is the thickness and appearance of the JZ. JZ thickness >12 mm has been considered as indicative of adenomyosis [15][28], but this cut-off point is not universally agreed. Additional features that can aid the diagnosis include diffuse, low-intensity areas accompanied by tiny high-intensity spots seen subjacent to the endometrium [16][29]. Increased JZ thickness is used as an indirect indicator of adenomyosis. However, the genesis of this differential density is uncertain and the transition from the inner to the outer myometrium has been shown to be gradual with no distinct transitional point [17][30]. This may explain the lack of agreement on a cut-off point for the diagnosis [18][31].
Sensitivity and specificity of MRI seem satisfactory, but interpreting images needs to consider a number of variables including age, phase of the menstrual cycle, gravidity, parity, hormonal status, previous uterine surgery and uterine contractions. In addition, the JZ is not measurable in 20–30% of women [19][32].
- (b) Ultrasound imaging
Features of adenomyosis were reported using transabdominal ultrasound (TAS) [20][21][33,34], following the first report by Walsh et al. [11][25]. However, it was transvaginal scanning (TVS) that enabled better characterization [22][23][35,36] and proved to be more reliable [24][37]. TVS was also reported to be accurate for the diagnosis of leiomyoma and for combined adenomyosis and leiomyoma, but not to be specific for the diagnosis of adenomyosis only when compared to histopathology [25][38].
Brosens et al. [24][37] reported that ill-defined myometrial heterogeneity is a better predictor of adenomyosis compared to uterine morphometry. Further advances enabled more detailed characterization of the diagnostic criteria [26][27][39,40]. The diagnostic features of adenomyosis on ultrasound include: A globally enlarged uterus, a symmetrically enlarged uterus; cystic myometrial lesions surrounded by a hyperechoic halo; inhomogeneous, irregular myometrial echo texture, indistinct areas with increased or reduced echogenicity; subendometrial lines and buds, myometrial hypoechoic linear striations radiating with thin acoustic shadows not arising from echogenic foci or leiomyoma; indistinct endometrial–myometrial border, diffuse minimal vascularity within the myometrium; the question mark sign or a retroflexed uterus where the cervix is directed forward.
The probability of adenomyosis in a woman with heavy bleeding and positive ultrasound features is 68.1% and the probability of adenomyosis after a normal transvaginal ultrasound scan is 10%. The sensitivity and specificity for symptomatic women are 84.3% and 82.3%, and for all women undergoing hysterectomy are 81.1% and 85.1% [28][41].
A further development has been the introduction of three-dimensional (3D) ultrasound. Features linked to adenomyosis on 3D-TVS are: (i) Maximum Junctional Zone thickness (JZmax) ≥ 8 mm; (ii) myometrial asymmetry; (iii) hypoechoic myometrial striations. Exacoustos et al. [29][42] compared features of adenomyosis detectable on two- (2D) and three-dimensional (3D) TVS and correlated these with histopathologic features in the JZ and the outer myometrium. The presence of myometrial cysts was the most specific, and heterogeneous myometrium was the most sensitive 2D-TVS feature. The 3D-TVS markers linked to high sensitivity and the best accuracy were a JZ difference ≥ 4 mm and JZ infiltration and distortion. However, the metanalysis by Andres et al. [30][43] reported no improvement in overall accuracy using 3D-TVS compared with 2D-TVS.
Champaneria et al. [31][44] undertook a systematic review and meta-analysis of published articles that compared the diagnostic accuracy of TVS or MRI and that used histological diagnosis as the gold standard comparator. They analyzed three studies that reported on the use of MRI and six studies that reported on the use of TVS. The pooled sensitivity and specificity of TVS was 72% (95% CI: 65–79%) and 81% (95% CI: 77–85%), respectively. TVS had a positive likelihood ratio of 3.7 (95% CI: 2.1–6.4) and a negative likelihood ratio of 0.3 (95% CI: 0.1–0.5). The pooled sensitivity and specificity for MRI were 77% (95% CI: 67–85%) and 89% (95% CI: 84–92%) respectively. MRI had a positive likelihood ratio of 6.5 (95% CI: 4.5–9.3), and a negative likelihood ratio of 0.2 (95% CI: 0.1–0.4). Reported studies, however, used different cut-off points for histological diagnosis and included different cohorts of women which is reflected in the variations in the incidence of adenomyosis. Furthermore, little account is provided of the impact of concomitant pathology on the diagnosis [7][22].
The role of ultrasound is less clear when it comes to identifying the extent of adenomyosis. Sonography and histopathology concurred in only 57% of cases when assessing the depth of presence of endometrium within the myometrium and in only 23% of cases when assessing the degree of involvement (i.e., the volume of the uterine muscle tissue affected by the disease) and lesion density [32][45]. Available literature does not provide an indication of the diagnostic value of each of the above features and diagnostic accuracy remains operator dependent.
The Morphological Uterus Sonographic Assessment (MUSA) Group issued a consensus statement on terms, definitions, and measurements to be used when describing sonographic features of the myometrium [33][46]. It requires a detailed description of any lesions identified by ultrasound, not merely a positive or negative diagnosis. This involves description of disease location in the uterine wall affected (anterior, posterior, left lateral, right lateral, fundal); whether the lesion is focal or diffuse; determination of the presence of myometrial cysts; the degree of myometrial depth involvement (limited to the inner portion, invading the body of the uterus, reaching the serosa); the volume of the uterus affected (<25%, 25–50%, >50%); and size of the lesions [34][47]. This approach has considerable potential in enabling the understanding of symptoms and outcomes of treatment. However, how widely it will be adopted remains to be seen. As mentioned above, reliable imaging-based diagnosis is not widely available and inter-observer reproducibility remains a challenge. Several additional factors, such as patient’s age, phase of the menstrual cycle, parity, previous uterine surgery or associated pathology and uterine contractions can affect ultrasound features.
2. Classifying Adenomyosis
Over the years, a number of proposals have been put forwards to classify adenomyosis. Early efforts used histological criteria with a focus on mapping the depth of endometrial invasion within the myometrium, both to define the extent of the disease and to distinguish a cut-off point for diagnosis. An important driver was to avoid over-diagnosis. The distinction between adenomyosis and the normal irregularity at the interface between the endometrium and the myometrium remains a challenge. This is particularly the case because of the difficulty researchers encounter when attempting to establish a correlation between histological features and symptoms. It is perhaps not surprising that researchers have proposed different cut-off points. On the other hand, because hysterectomy is usually performed on symptomatic women, a stringent cut-off point risks under-diagnosis.
The introduction of imaging-based diagnosis enabled a global view of the uterus and provided a less labor-intensive opportunity for disease mapping compared to histological diagnosis. A distinction, however, needs to be made between disease mapping based on anatomical location of lesions and disease categorization rooted on symptom severity or prognostic indices. Each of these has its own challenges. Most published literature reports on the accuracy of diagnosis based on whether adenomyosis is present or absent, rather than on the accuracy of each individual feature, location or cut-off point for diagnosis. The severity of symptoms attributed to adenomyosis has not been shown to be predictable based on its topography within the myometrium. A contrary view was expressed by Weiss et al. [35][52], who argued that adenomyosis is a normal variant rather than a disease entity. This was based on their study of 137 patients who underwent a hysterectomy, including 48% with adenomyosis. Those with and without adenomyosis had the same prevalence of fibroids, endometriosis, abnormal bleeding and chronic pelvic pain [35][52]. The latter point is important, but not unique to adenomyosis. It is well recognized that the extent of endometriosis and the size of fibroids do not always correlate with symptom severity.
Another feature that has attracted debate with regards to the classification of adenomyosis is the relation between the ‘classic’ adenomyosis of the uterus and lesions in the pouch of Douglas and the uterovesical pouch. These exhibit histological similarities to adenomyosis. The debate about their pathogenesis and whether they originate from spread of uterine disease to surrounding tissues or spread of endometriosis into the myometrium has important implications when it comes to classification. Similarly, affections of the outer myometrium may represent invasion from endometriosis lesions implanted on the serosa. Understanding the pathophysiology can also have other implications for classification. Adenomyosis may represent a spectrum of diseases with the unifying feature of ectopic endometrium and stroma within the myometrium [36][57].
One consideration for classification of adenomyosis is the place of JZ hyperplasia and whether it is pathognomonic of adenomyosis. Rasmussen et al. [37][64] attempted to classify the disease when confined to the inner myometrial and JZ regions into three separate ultrasound-based categories: (i) Adenomyosis of the inner myometrium; (ii) junctional zone disease, characterized by a serrated appearance of the JZ; and (iii) linear junctional zone [37][64]. While the existence and importance of JZ hyperplasia is not uniformly agreed, it was defined as partial or diffuse thickening of the JZ from 8 mm and over to less than 12 mm in the absence of additional imaging signs of adenomyosis [38][9]. Some investigators suggested that adenomyosis could be divided into different categories, based on morphology and location of the lesion. The term ‘diffuse adenomyosis’ is applied to lesions that affect at least one myometrial wall and could be symmetric or asymmetric. A further subdivision into three categories according to the depth of involvement reaching less than one-third, less than two-thirds, or greater than two-thirds of the myometrium has also been suggested [38][9]. This is distinct from ‘focal adenomyosis.’ Finally, an ‘adenomyoma’ is represented by a myometrial mass with indistinct margins of primarily low-signal intensity on T2-weighted MRI sequences [38][39][40][9,59,67] and it could be solid or cystic; it is commonly located in the mid-myometrium and rarely protrudes into the endometrial cavity or under the serosa [39][59]. A new variant, ‘external adenomyosis’ (anterior or posterior), has recently been introduced for lesions found adjacent to the uterine serosa, being significantly associated with pelvic endometriosis.
The International Endometrial Tumor Analysis (IETA) Group was established to classify the sonographic features of endometrial and intrauterine lesions [40][67]. The consensus criteria were given the acronym MUSA (Morphological Uterus Sonographic Assessment) [33][46]. The features proposed for adenomyosis by the group are: the location of the disease (anterior, posterior, left lateral, right lateral, fundal); a classification of the lesions (focal or diffuse); the presence or absence of intra-lesion cysts; the involvement of the myometrium (limited to the inner portion, invading the body of the uterus, reaching the serosa); the extent of the disease (affecting <25%, 25–50%, >50% of the uterine volume); and the size of the lesions [34][47].
Recently, Exacoustos et al. [41][68] studied the relation between an ultrasound-based disease classification and symptoms. Women with ultrasound diagnosis of diffuse adenomyosis were older and had heavier menstrual bleeding compared to those with focal disease, but there were no statistically significant differences in the severity of dyspareunia or dysmenorrhea. Focal disease was associated with a higher percentage of infertility. Overall, there was no direct correlation between ultrasound depictions of the extent of the disease and symptoms [41][68]. The authors speculate that this may be related to co-existent pathology, or to a true lack of correlation between symptoms and disease extent. However, larger studies are needed before definitive conclusions can be reached.