Visualising uterine microcirculation using contrast-enhanced ultrasound for differentiating between healthy controls and participants with adenomyosis: rationale and design of the UteroVue Adenomyosis study

Eva J.E. de Bock: Writing – review & editing, Writing – original draft, Visualization, Project administration, Methodology, Conceptualization. Nicole B. Burger: Writing – review & editing, Writing – original draft, Methodology, Conceptualization. Naem Donner: Writing – review & editing, Methodology. Jan Hein T.M. van Waesberghe: Writing – review & editing, Methodology. Ferenc I. Kandi: Writing – review & editing, Methodology. Catarina Dinis Fernandes: Writing – review & editing, Methodology. Massimo Mischi: Writing – review & editing, Methodology. Judith A.F. Huirne: Writing – review & editing, Methodology, Funding acquisition, Conceptualization. Lynda J.M. Juffermans: Writing – review & editing, Writing – original draft, Methodology, Funding acquisition, Conceptualization.

A prospective cross-sectional study is conducted at the department of Obstetrics and Gynaecology of the Amsterdam University Medical Centre, location Academic Medical Centre, Amsterdam, the Netherlands. This is a tertiary referral centre for adenomyosis. The study was registered with the Central Committee on Research Involving Human Subjects (registration number NL-007190) and approved by the medical ethics committee of the Amsterdam UMC (reference number: 2023.0341). This study consists of participant recruitment and study group allocation, and ultrasound examinations with qualitative and quantitative analysis. In addition, clinically indicated MRI scans and uterine specimens from hysterectomies are collected and analysed when available. Participants are recruited via the Benign Gynaecology outpatient clinic of the Amsterdam University Medical Centre and via flyering. Two groups are sought: (1) healthy controls, and (2) participants with evident adenomyosis based on ultrasound [24] . The flow diagram is presented in Fig. 1 . All participants provide informed consent. Fig. 1 Flow chart of participants. B-mode = brightness mode; MV-flow = microvascular flow imaging; CEUS = contrast-enhanced ultrasound; MUSA = Morphological Uterus Sonographic Assessment. * Contraindications for SonoVue™ are known allergy to SonoVue™ or any of its components, a history of any clinically unstable cardiac condition, a history of severe cardiac rhythm disorders within the last 7 days, severe pulmonary hypertension, uncontrolled systemic hypertension, and respiratory distress syndrome. Flow chart of participants. B-mode = brightness mode; MV-flow = microvascular flow imaging; CEUS = contrast-enhanced ultrasound; MUSA = Morphological Uterus Sonographic Assessment. * Contraindications for SonoVue™ are known allergy to SonoVue™ or any of its components, a history of any clinically unstable cardiac condition, a history of severe cardiac rhythm disorders within the last 7 days, severe pulmonary hypertension, uncontrolled systemic hypertension, and respiratory distress syndrome. We include premenopausal participants ≥ 18 years of age. Exclusion criteria for all participants are: previous hysterectomy; current pregnancy; intrauterine device in situ (to prevent shadowing artifacts on CEUS); dominant uterine fibroid(s) hindering adequate assessment of adenomyosis; endometrial polyp(s); (suspected) uterine malignancy; congenital uterine anomaly; uterine arteriovenous malformation; use of angiogenesis inhibitor; insufficient proficiency in both Dutch and English to provide informed consent; or a contraindication for contrast agent SonoVue™ (Bracco Suisse SA, Geneva, Switzerland). Study group allocation is performed after ultrasound evaluation. Healthy controls are defined according to the Delphi consensus study [25] . They may have a single subendometrial line, some hyperechoic islands, and an irregular or interrupted junctional zone, but must not show any subendometrial buds, myometrial cysts, myometrial asymmetry, a globular uterine shape, fan-shaped shadowing, or translesional vascularity. Participants with adenomyosis are stratified according to extent of adenomyosis: either mild (<25% affected myometrium), or moderate to severe (≥25% affected myometrium) [23] . These strata of participants with adenomyosis are analysed separately. During the study visit, the medical history of each participant is recorded, including medication use, allergies, obstetric and surgical history, menstrual cycle day and gynaecological symptoms. To conduct the ultrasound exam of the uterus, a HERA W10 ultrasound machine (Samsung Medison Co. Ltd., Seoul, Republic of Korea) and an EV2–10A transvaginal probe are used. Conventional ultrasound examinations are performed and consist of two-dimensional (2D) and 3D brightness mode (B-mode) and power Doppler imaging, 2D microvascular flow imaging (MV-flow) using MV-Flow™, and 2D strain elastography. To administer the contrast agent, peripheral intravenous access is secured. Before starting the CEUS scan, the probe is stabilised in the midsagittal plane using a custom-made probe fixture. The participant receives an intravenous 2.4-mL contrast agent bolus (SonoVue™) followed by a 10-mL saline flush. A 5-min 2D dual acquisition of CEUS and B-mode is performed. While maintaining the same probe position, 2D B-mode, power Doppler, MV-flow and strain elastography images are acquired to enable comparison of different ultrasound techniques within the same mid-sagittal plane. Then, a second bolus of 2.4 mL contrast agent and a consecutive 10-mL saline flush are administered, and a 3-minute 3D CEUS acquisition is performed. To allow for post-acquisition motion correction, 3D B-mode volumes are obtained both immediately before and after the 3D CEUS acquisition. The acquired B-mode, power Doppler, MV-flow, elastography and CEUS scans are qualitatively evaluated by two sonographers specialised in adenomyosis (E.B. and N.B.) by consensus. B-mode, power Doppler and MV-flow scans are analysed according to the recommendations of the MUSA group [26] . Additionally, the percentage of the myometrium that is affected by adenomyosis is estimated. Elastography scans are evaluated as proposed by Stoelinga et al. [27] . CEUS scans are assessed as previously described [22] . All evaluated features are described in Table 1 . Table 1 Ultrasound evaluation of the uterus. Feature Ultrasound technique Description 

 General Size uterus: length, height, width B-mode Measurement [cm] Diagnosis B-mode Adenomyosis/Fibroid/Caesarean scar defect/Endometrial polyp/(Suspected) malignancy/Congenital uterine anomaly/Uterine arteriovenous malformation/None 

 Myometrium Subendometrial lines and buds B-mode Yes/No/Cannot be determined Hyperechoic islands B-mode Yes/No/Cannot be determined Myometrial cysts B-mode combined with power Doppler and MV-flow Yes/No/Cannot be determined Fan-shaped shadowing B-mode Yes/No/Cannot be determined Irregular junctional zone B-mode Yes/No/Cannot be determined Interrupted junctional zone B-mode Yes/No/Cannot be determined Asymmetrical myometrial thickening B-mode Yes/No/Cannot be determined Globular uterine shape B-mode Yes/No/Cannot be determined Vascularity Power Doppler and MV-flow Absent/Minimal/Moderate/Abundant Percentage of myometrium affected by adenomyosis B-mode combined with power Doppler and MV-flow Estimation [%] Uniformity of elasticity Strain elastography < 50% uniform/≥ 50% uniform Order of wash-in CEUS Order of outer, middle and inner myometrium Predominant intensity of enhancement CEUS Non-enhanced/Hypo-enhanced/Iso-enhanced/Hyper-enhanced Perfusion pattern CEUS Homogeneous/Heterogeneous Non- to hypo-enhanced regions at peak enhancement CEUS Yes/No/Cannot be determined Hyper-enhanced regions at peak enhancement CEUS Yes/No/Cannot be determined Architecture of vessels in myometrium CEUS Orderly/Chaotic Enlarged lumen of vascular arcade CEUS Yes/No/Cannot be determined Enlarged lumen of vessels in outer myometrium, excluding the vascular arcade CEUS Yes/No/Cannot be determined Enlarged lumen of vessels in middle and inner myometrium, excluding the vascular arcade CEUS Yes/No/Cannot be determined 

 Endometrium Endometrial bilayer thickness B-mode Measurement [cm] Echogenicity B-mode Hypo-echoic/Iso-echoic/Hyper-echoic/Three-layer pattern/Non-uniform/Cannot be determined Predominant intensity of enhancement CEUS Non-enhanced/Hypo-enhanced/Iso-enhanced/Hyper-enhanced Perfusion pattern CEUS Homogeneous/Heterogeneous Enlarged lumen of vessels in endometrium CEUS Yes/No/Cannot be determined B-mode = brightness mode; MV-flow = microvascular flow imaging; CEUS = contrast-enhanced ultrasound. Ultrasound evaluation of the uterus. B-mode = brightness mode; MV-flow = microvascular flow imaging; CEUS = contrast-enhanced ultrasound. To enable quantitative analysis of the CEUS scans, the sonographer manually delineates regions of interest in the myometrium and endometrium. In participants with a caesarean scar defect [28] , the lower boundary of the myometrial region of interest is placed 5 mm fundal to the proximal part of the caesarean scar defect, to exclude the defect and its surrounding tissue from analysis. Time-intensity curve parameters are calculated for each region of interest using VueBox (Bracco Suisse SA, Geneva, Switzerland). In addition, we analyse microbubble dispersion, which characterises the degree of spreading of microbubbles through the uterine vasculature. To derive dispersion parameters, we adapt the CUDI algorithm to the uterus. For each pixel, we assess the similarity between its measured time intensity curve and those of pixels within a surrounding ring. This ring has an inner radius of 1 mm and an outer radius of 2.5 mm, and its size is optimised to capture early angiogenic changes while accounting for the resolution of the ultrasound machine. Similarity is measured in terms of correlation , spectral coherence , mutual information , and conditional entropy [29] . The resulting spatiosimilarity parametric maps reflect the convective-dispersion characteristics of the complete vascular network [30] . Finally, radiomic analysis is performed to characterise the spatial distribution of CUDI parameters, providing additional insights related to heterogeneity in angiogenesis. For the subset of participants with a clinically indicated MRI scan of the uterus, the MRI scan is collected if it was conducted either before CEUS or within 9 months after CEUS. One radiologist specialised in gynaecology (J.W.) assesses the MRI scans on features of adenomyosis ( Table 2 ) [31] , defining adenomyosis as present when the junctional zone thickness is ≥ 12 mm [32] . For analysis, the collected MRI scans obtained prior to CEUS on which the radiologist diagnoses adenomyosis are included, since adenomyosis, once present, is considered irreversible ( Table 3 ). But, MRI scans conducted before CEUS that do not show adenomyosis are only included if performed within 9 months prior to CEUS. This limits the chance of including MRI scans of uteri that developed adenomyosis between MRI and CEUS scanning. The relationship between MRI and CEUS results is assessed. Table 2 Magnetic resonance imaging evaluation of the uterus. Feature Description Maximum thickness junctional zone Measurement [mm] Foci of high signal intensity on T2-weighted scans Yes/No Foci of high signal intensity on T1-weighted scans Yes/No/T1-weighted scan unavailable If thickness junctional zone ≥ 12 mm, characteristics of adenomyosis: − Distribution − Location − Affected myometrial layers Diffuse/Focal Anterior/Posterior/Fundal/Left lateral/Right lateral/Global Inner/Middle/Outer Table 3 Maximum time interval between contrast-enhanced ultrasound (CEUS) exam and magnetic resonance imaging (MRI), and between CEUS exam and hysterectomy. Diagnosis: adenomyosis* Diagnosis: no adenomyosis MRI MRI performed before or within 9 months after CEUS MRI performed from 9 months before CEUS to 9 months after CEUS Histopathology Hysterectomy performed within 9 months after CEUS No time limit *Diagnosis by radiologist or pathologist Magnetic resonance imaging evaluation of the uterus. Distribution Location Affected myometrial layers Maximum time interval between contrast-enhanced ultrasound (CEUS) exam and magnetic resonance imaging (MRI), and between CEUS exam and hysterectomy. *Diagnosis by radiologist or pathologist For the subset of participants that undergo a hysterectomy after CEUS, histological specimens are collected at eight locations: fundus, uterotubal junction, isthmus, and between isthmus and uterotubal junction, bilaterally. Samples are processed into histological slides, and assessed by a pathologist specialised in gynaecopathology (N.D.). Slides are histochemically stained with haematoxylin and eosin (H&E). The histopathological diagnostic criterion for adenomyosis is defined as ectopic endometrial glands and stroma at > 2 mm into the myometrium or involving > 1/3 of myometrial thickness in at least one histological slide [33] . Uterine specimens are only included if hysterectomy occurred within 9 months after CEUS ( Table 3 ). This ensures that adenomyosis was already present at the time of CEUS. Histopathology slides are matched to corresponding regions on CEUS scans and their relationship is evaluated. The primary outcomes are quantitative CEUS perfusion and dispersion parameters. The secondary outcomes are qualitative CEUS characteristics and the relationship between CEUS findings and findings on conventional ultrasound, MRI and histopathology. Given the exploratory nature of this study, a formal sample size calculation is not performed. Based on the large inter-patient variation in CEUS findings in our pilot study of 10 participants with adenomyosis [19] , we aim to include 81 participants via the outpatient clinic and 81 via flyering. Anticipating a dropout rate of 10% due to incidental findings of other uterine disorders, inadequate image quality, or missed appointments, we expect to recruit approximately 180 participants to achieve the desired number of participants. To assess differences between healthy controls and participants with adenomyosis, qualitative CEUS characteristics and quantitative CEUS perfusion and dispersion parameters are compared. Continuous variables are analysed using t -tests or non-parametric tests. Categorical variables are assessed using chi-squared tests. A p -value < 0.05 is considered statistically significant. A sensitivity analysis is performed to assess the impact of including participants with a caesarean scar defect. If their inclusion substantially affects the results, they are excluded from the final analysis.

This protocol describes the first large study to apply quantitative CEUS analysis in adenomyosis. By performing CEUS in both healthy controls and participants with adenomyosis, and combining qualitative and quantitative analysis, this study aims to identify distinguishing microcirculatory features for ultimately improving diagnostics of adenomyosis. A limitation of this study is that study groups are not age-matched. Since adenomyosis becomes more prevalent with increasing age [34] , [35] , [36] , the effect of uterine aging on CEUS findings should be evaluated in future studies. Another potential limitation is the anticipated lack of histopathological confirmation of the diagnosis via hysterectomy specimen in most participants. In the past, hysterectomy was frequently performed in patients with heavy menstrual bleeding and/or dysmenorrhea. However, in recent years, uterine-sparing treatments have become increasingly available, and are often preferred. Although histopathology has long been regarded as the reference standard for adenomyosis, its accuracy is heavily dependent on the number of uterine sections that are examined. Increasing the number of examined sections from three to six nearly doubles the detection rate, indicating that routine clinical histopathology may fail to detect adenomyosis if insufficient tissue is examined [37] . Growing evidence indicates that adenomyosis represents a heterogeneous spectrum that cannot be fully captured by the extent of myometrial involvement on conventional ultrasound [19] , [38] . CEUS may provide additional insights into disease characterisation, as each disease subtype may show different perfusion patterns. Findings from this study may therefore contribute to the refinement of ultrasound features of adenomyosis, and the identification of disease subtypes.

Heavy menstrual bleeding, painful menstruations and subfertility can profoundly impact quality of life. One possible underlying cause of these symptoms is adenomyosis: a benign condition characterised by ectopic growth of endometrial tissue within the myometrium [1] , [2] , [3] , [4] , [5] , [6] . Adenomyosis is challenging to diagnose, due to the mostly diffuse blend of ectopic endometrium within the healthy myometrium, as well as little awareness of this condition. As a result, there is considerable variation in the reported prevalence of adenomyosis, ranging from 7% to 67%, depending on diagnostic modality and study population [3] , [4] , [5] , [6] , [7] , [8] , [9] . Improving care for patients with symptomatic adenomyosis begins with accurately establishing a diagnosis. Currently, the two diagnostic imaging modalities for adenomyosis are transvaginal two-dimensional (2D) ultrasound with Doppler techniques and Magnetic Resonance Imaging (MRI). Ultrasound is the first-line modality, as MRI has limited availability and long waiting lists, especially for high-resolution scanners, as well as high costs [10] , while its diagnostic performance is similar to that of ultrasound [11] . However, neither offers adequate diagnostic performance [11] , especially in the hands of non-experts [12] . This seems to be the cause for the diagnostic delay for adenomyosis averaging 11 years, during which patients consult about eight healthcare professionals [13] . By developing better diagnostics, this delay may be shortened, enabling earlier appropriate disease management. In the search for improved diagnostics, we may need to turn our gaze towards features that cannot be revealed by current imaging techniques. Such a feature may be the uterine microvasculature. Histological studies revealed differences in the uterine microvasculature between adenomyotic and healthy myometrium [14] , [15] , [16] . Thus, detection of abnormal microvasculature may help us to recognise adenomyosis. Doppler ultrasound techniques and MRI are unable to image the microvasculature. An ultrasound technique that can visualise the microvasculature is contrast-enhanced ultrasound (CEUS). This technique uses intravenously injected encapsulated gas-filled microbubbles to visualise and quantify blood flow (perfusion). CEUS recordings can be quantified over time by extracting time-intensity curves, from which perfusion parameters can be obtained, e.g. rise time , fall time , wash-in/-out area under the curve . This offers unique information on the complete vascular network, down to the smallest blood vessels [17] . While CEUS is already routinely used in clinical evaluation of other organs, such as liver, kidneys, heart, spleen, pancreas, bowel, bladder, lymph nodes and prostate [18] , its potential in gynaecology has not been widely explored. Recently, our pilot study on the feasibility of uterine CEUS showed that CEUS can provide new insights into adenomyosis [19] . In ten patients with adenomyosis, CEUS showed diverse myometrial perfusion patterns, suggesting a heterogenous disease spectrum. Our pilot study supports the need for further investigation of the potential of CEUS for diagnosing adenomyosis with a larger sample size. Beyond the quantitative time-intensity curve analysis performed in our pilot study, a method for further quantitative assessment is contrast ultrasound dispersion imaging (CUDI), which has shown promise in prostate cancer diagnosis [20] , [21] . This technique analyses the convective-dispersion kinetics of microbubble transport through the microvasculature, which reflects the microvascular architecture and can therefore aid in the detection of angiogenesis. Together with its feasibility in the gynaecological outpatient setting [22] , accessibility, and relatively low costs compared with MRI, CEUS may have considerable clinical value as a complementary technique following conventional ultrasound. By offering an alternative diagnostic pathway, CEUS could potentially accelerate diagnosis and reduce the need for MRI scanning. Initially, the study described in this protocol focuses on moderate to severe adenomyosis, defined as involvement of ≥ 25% of the myometrium [23] , based on Morphological Uterus Sonographic Assessment (MUSA) features [24] . Next, the study expands towards the analysis of mild adenomyosis (<25% affected myometrium). In this explorative cross-sectional prospective study, we aim to describe qualitative CEUS characteristics and use quantitative CEUS perfusion and dispersion parameters to differentiate between healthy controls and participants with (moderate to severe) adenomyosis.

The authors have nothing to declare.