Adenomyosis Diagnostic Features and Clinical Impact Connie O. Rees
ADENOMYOSIS: Diagnostic Features and Clinical Impact
2 The work of this thesis was funded in part by a grant by GE Healthcare, the Catharina Onderzoeksfonds, Stichting Vrouw en Onderzoek Catharina Ziekenhuis. Financial support for the printing of this thesis was kindly provided by Wetenschapsbureau Catharina Ziekenhuis, Endometriose Stichting, NVOG Werkgroep Gynaecologische Endoscopie, Erbe Nederland, Gedeon Richter Benelux, Medical Dynamics, Chipsoft and Goodlife Pharma. Cover Design and Layout: Designed using DALL-E ® and Canva ® Printing: Ridderprint Bv. ISBN: 978-90-386-5991-6 All rights reserved. No parts of this book may be reproduced in any form by any means without prior permission of the author. A catalogue record is available in the Eindhoven University of Technology and Ghent University Library.
3
4 Adenomyosis: Diagnostic Features and Clinical Impact PROEFSCHRIFT ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven en de Universiteit Gent en de, op gezag van de rector magnificus prof.dr. S.K. Lenaerts en prof. dr. R. van de Walle voor een commissie aangewezen door het College voor Promoties, in het openbaar te verdedigen op 2 April, 2024 om 16 uur door Connie O. Rees geboren te Tokyo, Japan
5 Dit proefschrift is goedgekeurd door de promotoren en de samenstelling van de promotiecommissie is als volgt: Voorzitter: prof.dr.ir. M.C. Beurden (TU/e) prof. dr. B. Heindyrckx (Ghent University) Promotoren: prof.dr.ir. M. Mischi prof. dr. S. Weyers (Ghent University Hospital) Copromotor: prof.dr. B.C. Schoot Promotiecommissieleden dr. K. Isaacson (Harvard School of Medicine) prof.dr. V. Mijatovic (Amsterdam UMC) prof.dr.ir. H. Wijkstra prof. dr. T. Hamerlynck (Ghent University) Adviseur: prof.dr. H.A.A.M. Van Vliet (Catharina Ziekenhuis Eindhoven) Het onderzoek of ontwerp dat in dit proefschrift wordt beschreven is uitgevoerd in overeenstemming met de TU/e Gedragscode Wetenschapsbeoefening
6
Table of Contents Part I Summary and introduction Summary Nederlandse Samenvatting Chapter 1 Introduction and thesis outline Part II Diagnosis of Adenomyosis on MRI Chapter 2 Objective measures of Adenomyosis on MRI and their Diagnostic Accuracy: a systematic review and metaanalysis Chapter 3 Prediction of Adenomyosis Diagnosis based on MRI Chapter 4 Prediction of Adenomyosis Diagnosis based on MRI: an external validation study Part III Effect of Adenomyosis on Uterine Contractile Function Chapter 5 Quantitative ultrasound imaging and characterisation of uterine peristaltic waves Chapter 6 Normal uterine contractile activity in healthy women throughout the menstrual cycle: an exploration of reference values using a novel quantitative 2D TVUS speckle tracking method Chapter 7 The influence of uterine abnormalities on uterine peristalsis in the non-pregnant uterus: a systematic review Chapter 8 Quantitative Ultrasound Measurement of Uterine Contractility in Adenomyotic versus Normal Uteri : A Multi-centre Prospective study Part IV Effect of Adenomyosis on Fertility Outcomes Chapter 9 Women with combined adenomyosis and endometriosis on MRI have worse IVF/ICSI outcomes compared to
8 adenomyosis and endometriosis alone: A matched retrospective cohort study Chapter 10 MRI Markers of Adenomyosis Severity Associated with Worse IVF/ICSI Outcomes Part V Effect of Adenomyosis on Obstetric Outcomes Chapter 11 The ADENO study: ADenomyosis and its Effect on Neonatal and Obstetric outcomes: a retrospective population-based study Part VI Discussion Chapter 12 General discussion and future perspectives References Part VII Supplementary Material Appendices per Chapter List of Author’s Publications Conference Abstracts and Presentations Acknowledgements Curriculum Vitae
9 GENERAL MESSAGES OF THIS THESIS: - Adenomyosis can be reliably diagnosed on MRI when additionally taking clinical symptoms into account - Adenomyosis disrupts uterine contractile function which may explain worse fertility and obstetric outcomes - Adenomyosis is a subtype of endometriosis and functions as part of a spectrum of the same disease - Severe adenomyosis in combination with endometriosis is associated with worse fertility outcomes - Adenomyosis is associated with a higher risk of adverse obstetric outcomes
Summary and Introduction PART I
11
12 SUMMARY: Adenomyosis is a common and potentially debilitating benign gynaecological condition characterised by the infiltration of endometrial tissue and stroma into the uterine myometrium. Associated symptoms are dysmenorrhoea, heavy menstrual bleeding, and subfertility, which can greatly impact quality of life. Adenomyosis often occurs in conjunction with endometriosis and is considered part of the spectrum of the same disease. The reported prevalence of adenomyosis is still uncertain, with estimates ranging from as low as 5% up to 85%. The lack of uniform diagnostic criteria for adenomyosis, and the fact that up to a third of women are asymptomatic (or do not present themselves to a clinician), exacerbates this problem. Historically, adenomyosis was only diagnosed on histopathology after hysterectomy at the end of a woman’s fertile phase of life. However, modern advances in non-invasive imaging have shown that adenomyosis is present in younger nulliparous women. With this knowledge, adenomyosis is being linked not only to clinical symptoms, but also to infertility and obstetric complications. Non-invasive and accurate diagnosis, along with clarity into the impact adenomyosis has on fertility and pregnancy, is thus crucial. Hence, we assessed the effect of adenomyosis on uterine contractile and reproductive function, starting with an exploration into its non-invasive diagnosis by way of MRI, and ending with its influence on fertility and obstetric outcomes. First, we conducted a literature study on the existing MRI-based diagnostic criteria for adenomyosis and carried out a meta-analysis into their reported diagnostic accuracy. We surmised that whilst a wide variety of MRI-based parameters have been used and reported, no uniformly applied criteria exist. Moreover, the individual diagnostic accuracy of each parameter is unclear. Based on the criteria described in the literature, we developed a multivariate diagnostic predictive tool for adenomyosis diagnosis using a retrospective cohort of patients with histologically confirmed adenomyosis. We also were able to validate this method using an external patient cohort. One promising element of MRI-based diagnosis parameter is the uterine anatomical ‘junctional zone’, which coordinates rhythmic uterine contractions,
13 or peristalsis. Previous research has suggested that abnormal uterine contractility affects fertility, and may thus represent a causal link between adenomyosis and infertility. Therefore, we introduced a novel quantitative speckle-tracking method on trans-vaginal ultrasound for the quantitative analysis of the uterine contractile function and investigated its potential to predict the success of in-vitro fertilisation (IVF) treatment. We then suggested a set of reference values for normal uterine contractile function in a prospective cohort of healthy women using this dedicated speckle tracking algorithm. Next, we investigated existing literature for the reported effect that uterine abnormalities (including adenomyosis, leiomyomas and congenital uterine anomalies) have on uterine contractile function. We found that whilst knowledge is scarce, uterine contractile behaviour seems to be affected by uterine pathology. Finally, uterine contractility features of healthy women were compared to women with adenomyosis. We observed significant differences in uterine contractility across the menstrual cycle between women with adenomyosis versus those without. To explore if fertility outcomes are significantly affected in women with adenomyosis, we investigated IVF/ICSI (intra-cytoplasmatic sperm injection) outcomes of a retrospective cohort of patients with MRI-diagnosed adenomyosis, endometriosis or both, compared to matched male infertility controls. We found that women with both adenomyosis and endometriosis have significantly fewer live births compared to the control group. We further report that adenomyosis patients with following characteristics have worse IVF/ICSI outcomes compared to male infertility controls: combined endometriosis, a larger (relative) junctional zone, and the presence of myometrial cysts. Finally, we investigated if the presence of adenomyosis was associated with adverse obstetric outcomes. To do so, we carried out a retrospective analysis of Dutch population-level data looking at obstetric outcomes in women with histologically diagnosed adenomyosis over a period of 23 years. Women with proven adenomyosis demonstrated a higher prevalence of a wide range of adverse obstetric outcomes including: hypertensive disorders, a higher rate of caesarean sections, more small-for-gestational-age children and failure to progress in labour.
14 Overall, we show that adenomyosis can be reliably diagnosed non-invasively on MRI using a variety of parameters, and confirm its effect on the whole spectrum of uterine (reproductive) function.
15 Nederlandse Samenvatting Adenomyose is een veel voorkomende en vaak invaliderende goedaardige gynaecologische aandoening. Het wordt gekenmerkt door infiltratie van endometriumweefsel en stroma in het myometrium van de baarmoeder. Geassocieerde symptomen zijn: dysmenorroe, hevige menstrueel bloedverlies en onvruchtbaarheid, die allen de kwaliteit van leven sterk kunnen beïnvloeden. Adenomyose komt vaak voor in combinatie met endometriose en wordt vaak beschouwd als onderdeel van het spectrum van dezelfde ziekte. De gerapporteerde prevalentie van adenomyose is nog steeds onzeker, met schattingen variërend van zo laag als 5% tot zelfs 85%. Het gebrek aan uniforme diagnostische criteria voor adenomyose, en het feit dat tot een derde van de vrouwen mogelijk asymptomatisch is (of zich niet meldt bij een arts), verergert dit probleem. Historisch gezien werd adenomyose alleen gediagnosticeerd op histopathologie na hysterectomie aan het einde van de vruchtbare levensfase van een vrouw. Moderne ontwikkelingen op het gebied van niet-invasieve beeldvorming hebben echter aangetoond dat adenomyose ook aanwezig is bij jongere vrouwen. Met deze kennis wordt adenomyose niet alleen in verband gebracht met klinische symptomen zoals pijn en hevig bloedverlies, maar ook met onvruchtbaarheid en obstetrische complicaties. Niet-invasieve en nauwkeurige diagnose, samen met duidelijkheid over de impact die adenomyose heeft op vruchtbaarheid en zwangerschap, is dus cruciaal. Daarom onderzochten we het effect van adenomyose op de contractiele en reproductieve functie van de uterus, beginnend met een verkenning van de niet-invasieve diagnose door middel van MRI, en eindigend met de mogelijke invloed op vruchtbaarheid en obstetrische uitkomsten. Allereerst hebben we een literatuurstudie uitgevoerd naar de bestaande (op MRI gebaseerde) diagnostische criteria voor adenomyose en een metaanalyse uitgevoerd naar hun gerapporteerde diagnostische nauwkeurigheid. We stelden vast dat er weliswaar een grote verscheidenheid aan MRIgebaseerde parameters is gebruikt en gerapporteerd, maar dat er geen uniform toegepaste criteria bestaan. Bovendien is de individuele diagnostische nauwkeurigheid van elke parameter onduidelijk. Op basis van de in de literatuur beschreven criteria ontwikkelden we vervolgens een multivariaat
16 diagnostisch predictie model voor de diagnose van adenomyose met behulp van een retrospectief cohort van patiënten met histologisch bevestigde adenomyose. We konden dit model tevens extern valideren met behulp van een apart patiënten cohort. Een veelbelovend diagnostische parameter op basis van MRI is de anatomische 'junctionele zone' van de baarmoeder, die de ritmische samentrekkingen van de baarmoeder, of peristaltiek, coördineert. Eerder onderzoek heeft suggereert dat abnormale contractiliteit van de baarmoeder de vruchtbaarheid beïnvloedt en dus een oorzakelijk verband kan vormen tussen adenomyose en onvruchtbaarheid. Daarom introduceerden we een nieuwe kwantitatieve speckle-tracking methode op transvaginale echografie voor de kwantitatieve analyse van de contractiliteit van de baarmoeder. We onderzochten we het potentieel van een nieuwe kwantitatieve speckle-tracking methode op transvaginale echografie om het succes van in-vitrofertilisatie (IVF) behandelsucces te voorspellen. We hebben daarbij een reeks referentiewaarden voorgesteld voor normale baarmoeder contractiliteit in een prospectief cohort van gezonde vrouwen met behulp van dit speckle trackingalgoritme. Vervolgens hebben we de bestaande literatuur onderzocht op het gerapporteerde effect dat afwijkingen aan de baarmoeder (waaronder adenomyose, leiomyomen en aangeboren baarmoederafwijkingen) hebben op de contractiele functie van de baarmoeder. Tot slot werden kenmerken van de uteruscontractiliteit van gezonde vrouwen vergeleken met die van vrouwen met adenomyose. We zagen significante verschillen in uteruscontractiliteit tijdens de menstruatiecyclus tussen vrouwen met adenomyose en vrouwen zonder adenomyose. Om te onderzoeken of vruchtbaarheidsuitkomsten daadwerkelijk significant worden beïnvloed bij vrouwen met adenomyose, onderzochten we IVF/ICSI (intracytoplasmatische sperma-injectie) uitkomsten van een retrospectief cohort van patiënten. Deze patiënten hadden MRI-gediagnosticeerde adenomyose, endometriose of beide, en vergeleken deze met gematchte controle met mannelijke onvruchtbaarheid. We vonden dat vrouwen met zowel adenomyose als endometriose significant minder levendgeborenen hadden in vergelijking met de gezonde controlegroep van veronderstelde normale vrouwen. Vervolgens werden individuele MRI-kenmerken van adenomyose bij IVF/ICSI-patiënten in verband gebracht met de IVF/ICSI-uitkomsten. We
17 rapporteren dat adenomyose patiënten met de volgende kenmerken slechtere IVF/ICSI-uitkomsten hebben in vergelijking met mannelijke onvruchtbaarheidscontroles: gecombineerde endometriose, een (relatief) grotere junctionele zone, en de aanwezigheid van myometriale adenomyotische cysten. Tot slot onderzochten we of de aanwezigheid van adenomyose geassocieerd was met ongunstigere obstetrische uitkomsten. Het aanwezig zijn van afwijkend spierweefsel van de baarmoeder zou immers ook effect op de zwangerschap en de bevalling kunnen hebben. Hiertoe voerden we een retrospectieve analyse uit van gegevens op Nederlands populatieniveau waarbij we keken naar obstetrische uitkomsten bij vrouwen met histologisch gediagnosticeerde adenomyose over een periode van 23 jaar. Vrouwen met bewezen adenomyose vertoonden een hogere prevalentie van een breed scala aan ongunstige obstetrische uitkomsten, waaronder: hypertensieve aandoeningen, een hoger percentage keizersneden, meer kleine kinderen voor de zwangerschapsduur en een langere duur van de bevalling. Over het geheel genomen laten we zien dat adenomyose betrouwbaar op MRI beelden gediagnosticeerd kan worden aan de hand van verschillende parameters, en bevestigen we het negatieve effect van adenomyose op het hele spectrum van de baarmoeder(voortplantings)functie.
CHAPTER 1: General Introduction
19
20 General Introduction to Adenomyosis Adenomyosis is a prevalent and potentially debilitating benign gynaecological condition characterised by the infiltration of endometrial tissue and stroma into the uterine myometrium. The most commonly associated symptoms are dysmenorrhoea, heavy menstrual bleeding, and subfertility, which can greatly impact quality of life (1,2). The prevalence of adenomyosis as a whole remains unknown and debated, ranging from estimates as low as 5% (3) to as high as 85% (4). This uncertainty stems for a large part from the fact that there exists no consensus on the diagnostic criteria of adenomyosis, and that up to a third of women with adenomyosis are asymptomatic and/or may never present themselves to a clinician (5–7). It is thought that adenomyosis originates in the so-called ‘junctional zone’ (JZ) between the endometrial and myometrial layer of the uterus (see Figure 1.1), where subsequent thickening and irregularity of the JZ characterises adenomyosis. The identification and evaluation of the JZ hence plays a crucial role in the diagnosis and recognition of adenomyosis. Whilst it is commonly accepted that irregularities in the JZ are indicative of adenomyosis, the exact pathophysiological mechanisms behind the disease are still open to discussion. Generally speaking, there are three theories as to the aetiology of adenomyosis: - De novo metaplasia of Mullerian remnants in the myometrium - Tissue injury and active repair (8–10), whereby chronic (micro-) trauma to the JZ and endometrium leads to invagination of endometrium into the myometrium, e.g. due to uterine surgery for example - ‘From outside to inside’ invasion of endometriosis cells into the uterine myometrium (11,12) Adenomyosis Subtypes and Classifications: Adenomyosis is generally accepted to have three subtypes: focal, diffuse and cystic. Focal adenomyosis is concentrated in one area of the myometrium, generally as a single lesion characterised by (heterogenous) thickening of the JZ. Diffuse adenomyosis affects most of the uterus, and results in an enlarged and globular uterus with generally irregular JZ borders.
21 Finally, cystic adenomyosis, widely regarded as the rarest subtype, involves cystic adenomyomas in the myometrium, which contain endometrial tissue (13). Figure 1.1 shows examples of these subtypes on MRI. Figure 1.1: Examples of MRI subtypes (on MRI), A. Focal adenomyosis concentrated in the posterior wall (red arrow) B. Diffuse adenomyosis affecting the whole uterine wall C. Cystic adenomyosis with islands of endometrial tissue in the myometrium (largest cystic focus circled in red). (Images taken from own study populations). Whilst the aforementioned subtypes are most commonly reported in the literature, some authors (14–18) have further made the distinction between adenomyosis of the inner myometrium (AIM) and adenomyosis of the outer myometrium (AOM). Here, the inner myometrium refers to the inner one third of the myometrium closest to the endometrium, and the outer myometrium refers to the outer two thirds (see Figure 1.2). Similarly, AIM can also be referred to as adenomyosis interna, and AOM as adenomyosis externa. This classification system can also be combined with the former, where an adenomyosis lesion could also be described as ‘focal adenomyosis of the outer myometrium’ (FAOM) (12) for example.
22 Figure 1.2: Examples of adenomyosis subtypes, A. Adenomyosis of the inner myometrium of the posterior wall (AIM), B. (Focal) Adenomyosis of the outer myometrium (FAOM) on the anterior wall, in continuum with an endometriosis lesion of the bladder. (Images taken from own study populations). Diagnosing adenomyosis Accurate adenomyosis diagnosis remains challenging as there is no consensus on diagnostic criteria, and it often differs between regions, hospitals and clinicians (19). In the associated condition endometriosis, it is known that the diagnostic delay is an average of nine years (20,21). It is unknown how long this is for adenomyosis, but it is likely to be similar, if not longer. Another element of the disease that impedes easy diagnosis is the fact that adenomyosis often coexists with other (benign) diseases; namely leiomyomas and endometriosis. Leiomyomas particularly can hamper accurate diagnosis as they are much easier to recognise on various imaging techniques (and thus may distract from adenomyosis lesions), and also are associated with similar clinical symptoms (22–24). In fact, concomitant leiomyomas are reported in up to 50% of adenomyosis patients (25). Similarly, adenomyosis and endometriosis are also often found together, with adenomyosis sometimes being referred to as ‘endometriosis interna’. Opinions are divided as to if adenomyosis and endometriosis are separate diseases, or subtypes of the same disease (8,10,19,26).
23 Histopathological diagnosis of adenomyosis The gold standard for adenomyosis diagnosis remains histopathology. Indeed, adenomyosis was first identified in hysterectomy specimens by von Rokitansky in the 1860’s as ‘cystosarcoma adenoids uterinum’ or ‘adenomyoma’, and eventually given its present name of ‘uterine adenomyosis’ by Frankl in 1925 (27). On histology, most often (and most accurately) after hysterectomy, adenomyosis is recognised as myometrial invasion of endometrial stroma, surrounded by myometrial hyperplasia. The minimum distance required for invasion into the myometrium to be considered as indicative of adenomyosis is debated (19). Some experts propose a range of 1-4mm, others at least one third of the total myometrial thickness. The common theme is that there must be recognisable endometrial tissue deeper than the endo-myometrial junction (7). An example of adenomyosis on histopathology is shown in Figure 1.3. The aforementioned criteria for adenomyosis are thus not uniformly classified, and different pathologists and hospital tend towards different definitions (19). The most commonly reported definitions for adenomyosis lesions are however, as follows: • At least one low power field from (an irregular) endo-myometrial junction, or • 1 to 2.5 mm below the basal layer of endometrium, or • Deeper than 25% of the overall myometrial thickness Another limitation of histological diagnosis is that accurate diagnosis is highly reliant on the method of tissue sampling and the number of slices taken at hysterectomy, and subsequently the size of the lesion. There is thus a potential for small adenomyosis lesions to be missed (19).
24 Figure 1.3. Example of adenomyosis on histology Whilst hysterectomy is most commonly used as the standard for adenomyosis diagnosis, adenomyosis can also be diagnosed histologically using biopsy, generally in the context of hysteroscopy. Biopsy diagnosis can be a useful alternative for histopathological diagnosis over hysterectomy, especially in the context of patients wishing to preserve their fertility, or otherwise wishing to avoid hysterectomy. The challenge of variable topographic distribution of adenomyosis across the uterus is multiplied in biopsy-driven adenomyosis diagnosis however. The hysteroscopist must biopsy exactly the correct area, and in cases without visibly suspect adenomyosis lesions during the procedure, a negative diagnosis by no means excludes the presence of adenomyosis. Especially in women of fertile age, whereby hysterectomy would be preferable to avoid, there is thus still a need for an accurate diagnostic tool. The logical way in which to realise this, is by way of diagnosis via noninvasive imaging techniques.
25 Non-invasive diagnosis - Ultrasound In daily gynaecological practice, adenomyosis is often first suspected (alongside presence of corresponding symptoms) on the basis of the aspect of the uterus on trans-vaginal ultrasound (TVUS). Characteristics indicative of adenomyosis are generally described according to the MUSA (Morphological Uterus Sonographic Assessment (28)) criteria (see Figure 1.4). The reported sensitivity, specificity, and Area Under the Curve (AUC) of TVUS are (with 95% CI) 74% (SD 68%–79%), 76% (SD 71%–79%), and 0.7 respectively (29). See Figure 1.5 for illustrative examples on TVUS. Figure 1.4. Sonographic characteristics of Adenomyosis according to the Morphological Uterine Sonographic Assessment (MUSA) Criteria of Van den Bosch et al. 2015 (30) Despite the possibility to fairly accurately diagnose adenomyosis on standard TVUS, there are several limitations that should be considered. One problem posed by diagnosing adenomyosis by TVUS is in that it is susceptible to interobserver variability (29,31–34), which often depends on the experience and expertise of the ultrasound technician or physician (35). As such, specifically in cases of mild or atypical adenomyosis, adenomyosis is easily overlooked. It
26 can also be difficult to distinguish between adenomyosis and other benign uterine pathologies such as leiomyomas. In fact, the sensitivity of TVUS for adenomyosis has been reported to be as low as 33% in the added presence of leiomyomas (36). One form of TVUS diagnosis that is also in continuous development in the context of (benign) uterine disorders is 3-dimensional TVUS (3D-TVUS). 3DTVUS is generally able to better assess the JZ and the overall structure of the endometrium in comparison to its 2-dimensional counterpart (19). Furthermore, due to being able to visualise several planes at once (coronal, sagittal and transverse), uterine asymmetry resulting from adenomyosis is more easily visualised. Because of these advances, 3D TVUS has a high reported sensitivity and specificity (95% CI), of 84% (77%–89%) and 84% (77%–89%) respectively (29). Figure 1.5. Adenomyosis examples on TVUS. A. Globular enlargement of the uterus with fan-shaped shadowing. B. Globular enlargement of the uterus with scattered hyperintense myometrial cystic foci. C. Uterus with asymmetric uterine wall thickening with scattered echogenic foci. (Images taken from own study populations). - Magnetic Resonance Imaging (MRI) Adenomyosis can be difficult to distinguish from other (benign or malignant) uterine disorders such as leiomyomas or carcinomas. For this reason, diagnosing adenomyosis using MRI has become preferred and is considered reliable, especially when the presentation is atypical on TVUS (29,32,37). MRI is not considered the first-in line diagnostic method however, due the higher associated cost and lesser availability than TVUS. On imaging, the uterus is made up of three distinct layers (or ‘zones’), each with slightly different imaging characteristics (i.e. signal intensity) on MRI (38).
27 The first is the inner lining of the uterus, or endometrial zone, followed by the junctional zone (JZ), and then the myometrial zone. On T2-weighted images, the endometrium normally exhibits high signal intensity, whereas the surrounding JZ exhibits relatively low signal intensity. The remainder of the myometrium is generally characterised by intermediate signal on T2 imaging (39,40). On T1 images, it is also possible to visualise potential haemorrhagic areas within the uterine structure, such as those sometimes seen in adenomyomas (41). As adenomyosis is thought to arise from the disruption and thickening of the JZ, the standard MRI imaging techniques used are thus T1- and T2- weighted MRI, which are most effective in identifying potential anomalies in the JZ. Unlike with TVUS, which has the MUSA criteria, there are no accepted classification or diagnostic criteria when evaluating an MRI for presence of adenomyosis. While much has been reported about typical, atypical, direct and indirect MRI manifestations of adenomyosis (17,33,42), the recognition of these features often still depends on the professional analysing the images. Several different classifications of adenomyosis have been suggested based on these differing criteria (17,43), but few objective guidelines for adenomyosis exist. The most widely reported objective measures are based on the appearance of JZ, with a thickness of over 12mm, a difference > 5mm between maximum and minimum JZ diameter, and a ratio of >40% of JZ to the full myometrium thickness thought to signify adenomyosis (34,36,40,44). These criteria are still not widely accepted however, with the 12mm cut-off value recently coming into question (45). This highlights the need for further characterisation of adenomyosis on MRI. Evaluation of the JZ can also be problematic as its thickness changes during the menstrual cycle and is affected by hormonal fluctuations, which can limits the ability to distinguish between ‘normal’ JZ and adenomyosis foci (42,46–48). Furthermore, uterine contractions during MRI can give a false impression of (focal) adenomyosis (42). Other examples of how adenomyosis can be objectified on MRI include uterine volume, uterine wall thickness or the volume or size of (suspected) adenomyotic foci (4,33,40,49,50). Some studies have also attempted to correlate MRI characteristics of adenomyosis with clinical outcomes, such as dysmenorrhoea, treatment response or obstetric outcomes (4,49,51–54), but it is unclear if
28 there is a direct relationship. This is a promising and clinically relevant query that needs further exploration as it would then become possible to predict (risk of) certain outcomes on the basis of non-invasive imaging. There are thus a wide range of parameters on MRI that can be used to characterise and visualise adenomyosis, such as junctional zone thickness, myometrial signal intensity and uterine size (33). However, many of them have not been specifically investigated for diagnostic accuracy versus histopathology, and little is likewise known about how they may correlate to clinical outcomes (32). Despite attempts to create (imaging-based) classification systems for adenomyosis (16,17), there currently exists no widely used tool for prediction of adenomyosis diagnosis. Adenomyosis treatment Once accurately diagnosed, adenomyosis can be (symptomatically) treated fairly effectively using a variety of strategies. . The choice of strategy is the result of type of complaints and clinical issues (e.g. pain, dysmenorrhea, abnormal bleeding, dyspareunia, fertility, age, uterine preservation wishes). All treatment options have a reported positive effect on quality of life, uterine and lesion size, symptoms (e.g. dysmenorrhoea) and fertility to differing extents (7,55). Medical treatment In clinical practice, initial treatment options for adenomyosis often involve symptomatic (pain or bleeding) relief (NSAID’s or tranexaminic acid) and/or hormonal therapy (56,57). Hormonal therapies range from systemic combined oral contraceptive pills (COCs) and progesterone-only pills (POPs) to more localised levonorgestrel intra-uterine devices (Lng-IUDs). All of these hormonal options have the goal of relative suppression of endometrium buildup and stimulation of endometrial atrophy, with subsequent reduction of adenomyosis-related symptoms. As with associated endometriosis, GnRH agonists or antagonists, selective oestrogen and/or progesterone modulators (SERMs and SPRMs) can also be employed in cases resistant to other forms of hormonal therapy (56,58).
29 Embolisation and radiofrequency ablation In recent years, more focus is also being laid on minimally invasive treatments such as uterine artery embolisation (UAE) or high-frequency ultrasound (HiFU) ablation strategies (59,60). Both of these methods, though not available to, or suitable for, all adenomyosis patients, also show promise with regards to symptom and lesion reduction whilst being uterus-sparing (61). Surgical treatment If the patient wishes to avoid hormonal therapies, or there are relevant contraindications, surgical therapy by way of local excision of adenomyotic lesions, or a hysterectomy can be indicated. With the ongoing development of a wide array of medicinal therapies, hysterectomy is becoming seen as a lastresort option in women without an active child-wish. Surgical excision of smaller localised lesions can also be an option in women wishing to retain reproductive function (62), but is only feasible in certain cases, and is generally only done in centres with the relevant expertise. Without total hysterectomy, and especially without concomitant hormonal therapy, it is likely, however, that the disease will return over time (63). Effect of adenomyosis on uterine (contractile function) Embryologically, the uterine myometrium is made up two types of tissue: the neometrium and the archimetrium. The neometrium makes up the outer two layers of the myometrium, and is generally thought to be the driving force behind strong uterine contractions, as associated with contractions during labour and pregnancy. The archimetrium, the innermost layer of myometrium bordering on the endometrium (which includes the JZ), is linked to uterine peristaltic contractions outside of pregnancy (64). In the 1990’s and early 2000’s, it was established that normal uterine contractile function follows a distinct pattern throughout the menstrual cycle (65–69). At the start of the cycle, during menstruation, uterine contractions travel mostly from the fundus-to-cervix direction with a relatively low frequency. Subsequently, in the follicular phase, uterine peristalsis travels from the cervix-to-fundus direction, with increasing contraction frequency until the periovulatory (late follicular) phase. After ovulation, during the luteal phase, the uterus enters a relatively quiescent state with the lowest frequency of uterine contractions. The characteristics of these contractions proved of
30 importance for fertility when examined in in-vitro fertilisation (IVF) populations (70,71). One theory as to the aetiology of adenomyosis, and also the pathophysiological mechanism behind its symptoms, relates to how adenomyosis affects the JZ and thereby uterine contractile function. Chronic peristaltic contractions could induce repeated (micro) trauma to the JZ, causing inflammation which in turn leads to locally increased oestrogen production, which stimulates myometrial activity and inducing a vicious cycle of chronic hyperperistalsis (61). Over a lifetime, this cycle leads to gradually worsening adenomyosis with increasing age (63). This theory has been corroborated by longitudinal studies that have reported a directly proportional relationship between JZ width, adenomyosis severity and age. (46,72–74). Several studies have investigated uterine contractility in endometriosis and adenomyosis, in an attempt to corroborate the theory that uterine contractions are abnormal in these diseases (75,76). It is theorised that with a disruption in the JZ, such as by adenomyosis, uterine contractility is likewise impaired, leading to so-called dysperistalsis (76,77) and consequently, adenomyosisrelated symptoms. Despite these studies showing promising results, the methods used to assess uterine contractility have been subjective, complex, and/or relatively user- and patient-unfriendly. If uterine contractility (in normal and abnormal uteri) could be objectively and reliably assessed, a concrete conclusion could be made regarding how, and if, uterine contractility is affected in adenomyosis patients.
31 Effect of adenomyosis on fertility In the previous sections, we have established how adenomyosis behaves, is diagnosed, and how it may disrupt uterine contractile function. One other crucial aspect of uterine function (arguably the uterine function) that adenomyosis has been reported to affect is fertility. With the advent of improving imaging techniques, adenomyosis has also been more frequently diagnosed in younger, nulliparous women, and is being increasingly causally linked to sub- or in-fertility and adverse pregnancy outcomes (78–81). Adverse reproductive outcomes in adenomyosis could occur for several reasons. First, JZ disruption in adenomyosis is thought to cause dysfunction in uterine peristaltic movement, and thereby inhibits both sperm transport and embryo implantation (75,82,83). Further evidence exists showing a different expression of factors involved in embryo implantation such as osteopontin and HoxA10 in adenomyosis, as well as increased endometrial free radicals and inflammatory cytokines (84,85) which may impact embryo development. In the context of infertility, several studies have investigated specifically how adenomyosis affects IVF outcomes. The majority of studies seem to show a negative influence of adenomyosis on IVF, with a recent retrospective study by Sharma et al. (86) showing a significantly reduced pregnancy rate after IVF in women with signs of adenomyosis on TVUS. A study by Ballester et al. (87) with endometriosis patients showed a particularly large contrast, with IVF pregnancy rates of 19% versus 82% respectively for patients with and without adenomyosis. A study examining adenomyosis characteristics on TVUS also reported a direct relationship between severity of adenomyosis and assisted reproductive technologies (ART) outcomes (80). These findings have been corroborated in meta-analyses (88,89), with Vercellini et al., reporting that adenomyosis reduces likelihood of pregnancy in ART by 28%. Conversely, studies have also been published that cannot show an association of adenomyosis with IVF failure (90–93), including in a meta-analysis (94). Thus, there still remains a lack of consensus regarding the extent to which adenomyosis affects fertility and IVF.
32 The junctional zone and infertility As junctional zone dysfunction and changes are thought be associated with infertility, several studies have specifically investigated if changes in the JZ can be correlated with IVF outcomes. While these studies are heterogenous in design and study population, they all seem to show a potential link between thickened JZ and infertility. For instance, Meylaerts et al. (95) reported that infertile women have a thicker mean JZ on MRI than healthy nulliparous women. Similarly, El Gaber et al.(96) showed that women with recurrent implantation failure had a thicker JZ (on TVUS) than women with unexplained infertility. Kunz et al.(97) also linked a thickened JZ to reduced oocyte quality amongst IVF or intracytoplasmic sperm injection (ICSI) patients. Maubon et al. (98) and Piver et al. (99) suggested more specifically that a JZ of over 7mm can be used to predict IVF failure with 97% accuracy. Several of these studies included women with endometriosis, and suggested that a thicker junctional zone (and thereby potential adenomyosis) could also be independently related to IVF failure in the context of endometriosis (97,98,100). Adenomyosis on MRI and infertility Despite promising results separately suggesting a link between MRI characterisation of the JZ and infertility, and adenomyosis on MRI and infertility, few studies have investigated these two aspects in conjunction with each other. Some evidence does exist suggesting a link between adenomyosis severity and fertility, for instance that women with diffuse adenomyosis do seem to have worse fertility outcomes than those with focal adenomyosis (101). More detailed studies characterising adenomyosis on MRI in the context of infertility are needed. We thus propose to carry out a detailed exploration into the characterisation of adenomyosis and MRI in the context of infertility, and to attempt to correlate this concretely to IVF outcomes. Effect of adenomyosis on pregnancy outcomes In the foregoing section, we established that it is now becoming accepted that presence of adenomyosis affects a woman’s chances of becoming pregnant, and increases her chance of miscarriage. More and more current literature also denotes however, that the risk of complications continue into later pregnancy as well.
33 Adenomyosis and pre-term birth and miscarriage One of the most commonly reported adverse obstetric outcomes that has been associated with adenomyosis is pre-term birth (PTB, delivery prior to 37 weeks gestation). A recently published systematic review and meta-analysis based on 4 studies investigating pregnancies in women with adenomyosis calculated an odds ratio (OR) of 2.74 for PTB. The calculated OR for miscarriage was even higher at 3.40 (81). Taken individually, the reported OR’s for PTB in women with adenomyosis have ranged from 1.96 (102) to as high as 24.53 (103). While most studies were conducted in women with and without adenomyosis, several investigated obstetric outcomes in women with endometriosis, stratified by the presence or absence of adenomyosis. These studies showed mixed results (86,104–106). A meta-analysis did however report higher OR’s for PTB in adenomyosis compared to only endometriosis patients (1.47 vs. 3.09) (107). Due to infertility often accompanying adenomyosis, a large proportion of women with adenomyosis end up going on to have pregnancies facilities with assisted reproductive technologies (ART). This is a potentially confounding factor in adverse obstetric outcomes as there is a higher risk of preterm birth in IVF pregnancies. Several studies did however correct for mode of conception, and still reported a statistically significant higher odds ratio for preterm birth in women with adenomyosis (101,103,104,108,109). Adenomyosis and placental function In addition to impacting fertility, recent studies also suggest that adenomyosis gives an increased risk for adverse obstetric outcomes (50,107,110). This is possibly due to altered trophoblast invasion and spiral artery remodelling (111,112), linking it to conditions such as placental disorders, hypertensive disorders of pregnancy (HDPs) and foetal growth restriction (FGR) and smallfor-gestational age (SGA) infants. In accordance with this theory, significantly higher OR’s have been reported by the majority of studies investigating obstetric outcomes in adenomyosis. A meta-analysis of eleven studies calculated a composite OR of 3.90 for SGA (81). The studies that did not report significant outcomes did still report a trend of a higher risk of SGA in adenomyosis (86,106). A few studies have also been published linking adenomyosis and hypertensive disorders such as preeclampsia (PE), with an OR of 7.87 being suggested in one meta-analysis (81). One study also
34 suggested that women with adenomyosis specifically have a higher risk of late-onset preeclampsia(4). Evidence also exists which introduces a correlation between adenomyosis and placental malposition (i.e. placenta praevia) and placental abruption (101,104,106,108). Adenomyosis and neonatal outcomes No studies thus far have been able to prove a significant association between adenomyosis and adverse neonatal outcomes (not including low birth weight, LBW). The few studies investigating outcomes such as low Apgar score or umbilical artery pH did not report significant results (50,105). Despite most the published studies reporting evidence to associate adenomyosis with obstetric complications there are a few common weaknesses to these studies which may limit their generalisability. First, they all have relatively small sample sizes, with the largest study including 245 women with adenomyosis (101), and most studies including 50-60. Furthermore, all studies used trans-vaginal ultrasound diagnosis (or 2 only self-reported diagnosis (101,110)) as the diagnostic method of choice for adenomyosis, which is not considered the golden standard of adenomyosis diagnosis. It can be argued therefore that one cannot be certain that the previously carried out studies truly included the correct population. In addition, no studies have thus far been conducted investigating this question in a Dutch population, with very few being carried out in a large (European) population in general. Further common issues with these studies are heterogenous inclusion/exclusion criteria, and presence of various confounding factors such as most studies only investigating fertility outcomes in IVF/ICSI patients. So, due to the scarce and still relatively conflicting evidence, women with adenomyosis are not generally considered to qualify as having high-risk pregnancies, and no guidelines exist for the management of pregnant women diagnosed with adenomyosis. More large-scale studies are needed to yield unambiguous results to inform clinical practice and management of these women.
35 Thesis Objectives: In summary, there are a number of knowledge gaps inhibiting the complete understanding of the diagnosis and clinical impact of adenomyosis. Transvaginal ultrasound (TVUS) faces limiations in inter-observer variability and difficulty distinguishing adenomyosis from other uterine pathologies. MRI is preferred for atypical cases but lacks widely accepted criteria. The importance of accurate diagnosis is crucial, especially given adenomyosis’s association with infertility. Studies and expert opinion does suggest a potential link between adenomyosis and aberrant uterine contractile function, fertility and adverse obstetric outcomes. Consensus on the extent of adenomyosis’ impact on fertility and contractile function remains elusive however, emphasising the need for further research. We have identified the following objectives for this thesis: - What are the objective parameters of adenomyosis on MRI and what is their potential for the accurate diagnosis of adenomyosis on using MRI? - How can we objectively assess uterine contractile function in the nonpregnant uterus, and is it affected by the presence of adenomyosis? - To what extent does adenomyosis affect fertility outcomes in the context of IVF? - Does adenomyosis affect obstetric and neonatal outcomes?
36 Thesis Outline: In this thesis, we assess the effect of adenomyosis on the whole spectrum of uterine contractile and reproductive function, starting with an exploration into non-invasive diagnosis by way of MRI, and ending with its influence on obstetric outcomes. The thesis is divided into dedicated parts assessing each of the thesis objectives as described above. Part II: Chapter Two of this thesis focusses on the existing MRI-based diagnostic criteria for adenomyosis and their reported diagnostic accuracy from available literature. Based on the resultant described criteria, we then developed a multivariate diagnostic tool for adenomyosis diagnosis based on MRI and clinical parameters in Chapter Three. In Chapter Four we externally validate this prediction model using a separate patient cohort. Part III: In Chapter Five we develop a new 2D TVUS method employing speckle tracking and explore its use for assessing uterine contractile function and its potential for predicting IVF success. In Chapter Six, using this new method, we suggest a set of reference values for normal uterine contractile function in a prospective cohort of healthy women with normal uteri. In Chapter Seven, we then investigate existing literature for the effect that uterine abnormalities (including adenomyosis, leiomyomas and congenital uterine anomalies) have on uterine contractile function. We then used the same cohort of normal women to compare uterine contractility features to women with adenomyosis in Chapter Eight to assess how and if uterine contractility is affected. Part IV: Chapter Nine explores the IVF/ICSI outcomes of patients with adenomyosis, endometriosis or both compared to match male infertility controls. Subsequently, in Chapter Ten individual MRI characteristics of adenomyosis in IVF/ICSI patients are associated with IVF/ICSI outcomes. Part V: After assessing the effect of adenomyosis on fertility in the previous chapters, Chapter Eleven focusses on the association of adverse obstetric outcomes in histologically diagnosed adenomyosis patients at the Dutch population-level.
37
Diagnosis of Adenomyosis on MRI PART II
39
CHAPTER 2: Objective measures of adenomyosis on MRI and their diagnostic accuracy—a systematic review & metaanalysis Authors: Connie O. Rees Joost Nederend Massimo Mischi Hubertus A. A. M. van Vliet Benedictus C. Schoot The content of this chapter has been published in Acta Obstetrica et Gynecologica Scandinavica, 2021; vol. 100: pages 1377–1391. DOI: 10.1111/aogs.14139
41
42 ABSTRACT: Objectives: To systematically review literature on how adenomyosis can be objectively quantified on MRI in a scoping manner, and review the diagnostic performance of these characteristics compared to histopathological diagnosis. Additionally, to summarize correlations between measures of adenomyosis on MRI and clinical outcomes. Materials & Methods: We searched databases Pubmed, Embase and Cochrane for relevant literature up to April 2020 according to PRISMA guidelines. We included studies that objectively assessed adenomyosis on MRI, and separately assessed studies investigating the diagnostic performance of MRI versus histopathology for inclusion in a meta-analysis. The QUADAS-2 tool was used for risk of bias, with many studies showing an unclear or high risk of bias. Results: 80 studies were included, of which 14 assessed the diagnostic performance of individual MRI parameters, with four included in the metaanalysis of diagnostic accuracy. Common MRI parameters were: junctional zone (JZ) characteristics such as maximal JZ thickness; pooled sensitivity 71.6 (95% CI 46.0 – 88.2), specificity 85.5% (52.3 – 97.0) , JZ differential; pooled sensitivity 58.9% (95% CI 44.3 – 72.1), specificity 83.2% (95% CI 71.3 – 90.8) and JZ to myometrial ratio; pooled sensitivity 63.3% (95% CI 51.9 – 73.4), specificity 79.4% (95% CI 42.0 – 95.4), adenomyosis lesion size , uterine morphology (pooled sensitivity 42.9% (95% CI 15.9 – 74.9), specificity 87.7%, (95%CI 37.9 – 98.8)) and changes in signal intensity (e.g. presence of myometrium cysts; pooled 59.6% (95% CI 41.6 – 75.4) and specificity of 96.1% (95% CI 80.7 – 99.3). Other MRI parameters have been used for adenomyosis diagnosis however their diagnostic performance is unknown. Few studies attempted to correlate adenomyosis MRI phenotype to clinical outcomes. Conclusions: A wide range of objective parameters for adenomyosis exist on MRI; however, in many cases their individual diagnostic performance remains uncertain. JZ characteristics remain the most widely used and investigated with acceptable diagnostic accuracy. Specific research is needed into how these objective measures of adenomyosis can be correlated to clinical outcomes
43 KEYWORDS: Adenomyosis, Uterus, Endometriosis, Magnetic Resonance Imaging, Diagnosis, Non-invasive Imaging ABBREVIATIONS: MRI: magnetic resonance imaging; JZ: junctional zone; JZmax; maximal junctional zone; JZmin: minimal junctional zone; JZ Diff: junctional zone differential; TVUS: transvaginal ultrasound; SROC: summary receiver operator curve; SI: signal intensity; LSI: low signal intensity; HSI: high signal intensity; ADC: apparent diffusion coefficient; ; DWI: diffusion weighted imaging; DTI: diffusion tensor imaging;
44 INTRODUCTION: Adenomyosis is a prevalent and potentially debilitating gynaecological condition characterized by dysmenorrhoea and heavy menstrual bleeding. Adenomyosis is thought to arise from and lead to disruptions in the uterine ‘junctional zone’ (JZ) between the uterine endometrium and myometrium. With the advent of improving imaging techniques, adenomyosis has been more frequently diagnosed in younger, nulliparous women. Along with greatly affecting their quality of life, it is also increasingly linked to sub- or infertility and adverse pregnancy outcomes (81,85,107). The relationship between (the extent of) adenomyosis and these clinical outcomes remains largely unknown. One barrier to elucidating the relationship of adenomyosis (severity) to clinical outcomes is the accurate diagnosis of adenomyosis. Despite continuing advances in 2D and 3D transvaginal ultrasound imaging, MRI is generally considered to be the most consistently accurate in the diagnosis of adenomyosis (29,32); however, there is still no accepted classification system or a set of diagnostic criteria to evaluate adenomyosis on MRI. While much has been reported about typical, atypical, direct and indirect MRI manifestations of adenomyosis (33,42) , the recognition of these features often still depends on the experience and expertise of the radiologist and/or gynaecologist. Furthermore, it is still disputed which of the wide range of features reported is the most accurate. This makes it difficult to assess the true diagnostic accuracy of MRI for adenomyosis as different centres and physicians may use different criteria. If adenomyosis could be noninvasively and objectively quantified (e.g. on MRI), the burden of disease could be correlated with various clinical outcomes, such as symptom severity, therapy response, or fertility outcomes. Similarly, potential changes in adenomyosis could be more easily followed over a patient’s lifetime, or during their menstrual cycle (33,34,47,113). To the best of our knowledge, there is currently no comprehensive overview describing the quantitative analysis of adenomyosis on MRI imaging. The objectives of this review are thus as follows: the primary objective is to evaluate the diagnostic accuracy of MRI features for adenomyosis versus histopathology, with secondary objectives being to (1) summarize in a
www.ridderprint.nlRkJQdWJsaXNoZXIy MTk4NDMw