A Novel Semi-Quantified Method for Grading Sonographic Severity of Adenomyosis and its Relation With Clinical Symptoms: An Observational Study

Of the 216 participants who met the inclusion criteria, 69 were excluded due to hormonal therapy use and 3 due to other reasons (1 presence of dominant myoma and 2 breastfeeding). Another 24 participants were excluded due to missing ultrasound data, caused by technical issues during the transfer of ultrasound files to the storage drive, and with another 5 participants the 3D ultrasound quality was poor. This resulted in 115 participants, of whom 80 completed the questionnaire (Figure  3 ). Table  1 shows the baseline characteristics of the included participants. Participants had a mean age of 38 years, 29.6% (n = 34) were nulliparous, and 35.7% (n = 41) had a previous spontaneous miscarriage. Forty percent (n = 46) had previous gynecological surgery. Furthermore, 20.9% (n = 24) of the participants had uterine myomas and 26.5% (n = 30) had endometriosis. Twenty‐five percent (n = 20) of the participants demonstrated an elevated self‐BAT score (Table  2 ). Dysmenorrhea was reported by 95.2% (n = 100) of the participants (median NRS 8.0), and 61.5% (n = 67) reported heavy menstrual bleeding. Only 48 participants completed a PBAC, of which 60.4% (n = 29) had an elevated PBAC score. The overall median PBAC was 190.5 (103.3–334.8) (Table  2 ). To address outcome‐specific missing data, baseline characteristics of participants included in the analyses for dysmenorrhea and for those with an available PBAC score are presented separately in Table  S1 . These characteristics were similar to those of the total study cohort (n = 115). Flow chart of inclusions and questionnaire. General Characteristics of the Study Population Prospectively Collected During the Outpatient Consultation Abbreviations: BMI, body mass index; NSAID, non‐steroidal anti‐inflammatory drugs; P, percentile; SD, standard deviation. Missing values: N = 106 (92.2% available). Numbers do not add up as patients can have had both vaginal and cesarean delivery. Based on the ultrasound assessment during the outpatient clinic consultation. Based on patient history or the ultrasound assessment during the outpatient clinic consultation. Menstrual Related Results Collected Prospectively During the Outpatient Clinic Visit and Retrospectively via Questionnaire Abbreviations: Hb, hemoglobin; NRS, Numeric Rating Scale; P, percentile; PBAC, pictorial blood loss analysis chart; SD, standard deviation; Self‐BAT, Self‐Administered ISTH‐Bleeding Assessment Tool. Missing values: N = 37 (84.1% available). N = 27 (64.3% available). N = 110 (95.7% available). N = 48 (41.7% available). N = 75 (65.2% available). Jonckheere‐Terpstra test showed a statistically significant relation between higher dysmenorrhea pain scores (NRS) ( P  < .01), higher PBAC scores ( P  < .01), lower Hb values ( P  = .03), and higher severity categories of the real‐time ultrasound evaluation (Table  3 ). The adjusted analyses for the correlation between real‐time ultrasound and clinical outcomes were only statistically significant for PBAC and subjective blood loss (Tables  4 and S2, analyses adjusted for myoma, medication to reduce menstrual blood loss, and increased self‐BAT). However, the unadjusted analyses also showed a correlation between dysmenorrhea and Hb value with severe real‐time evaluation compared to mild (Tables  S2 and S3 ). Additionally, the mean number of XI‐VOCAL slices increased across real‐time severity categories: 7.43 ± 2.40 slices for mild, 8.23 ± 2.92 for moderate, and 12.06 ± 4.15 for severe adenomyosis; supporting a positive relationship between real‐time assessment and XI‐VOCAL assessment of adenomyosis. Correlations Between Adenomyosis Severity Assessed by Real‐Time Ultrasound and Clinical Outcomes Abbreviations: Hb, hemoglobin; NRS, Numeric Rating Scale; PBAC, pictorial blood loss analysis chart; XI‐VOCAL, Extended Imaging for Volume Calculation. Adjusted Linear and Logistic Regression of Correlation Between Real‐Time Ultrasound Assessment of Adenomyosis and Symptoms and Hemoglobin Levels Note : * P  < .05, ** P  < .01. Abbreviations: CI, confidence interval; NRS, Numeric Rating Scale; OR, odds ratio. Adjusted for uterine surgery (caesarean, laparoscopic or laparotomic myoma enucleation), endometriosis. Adjusted for endometriosis. Adjusted for myoma, medication to reduce menstrual blood loss (non‐steroidal anti‐inflammatory drugs, tranexamic acid), increased score of the Self‐Administered ISTH‐Bleeding Assessment Tool (>5). Adjusted for myoma, medication to reduce menstrual blood loss (non‐steroidal anti‐inflammatory drugs, tranexamic acid), increased score of the Self‐Administered ISTH‐Bleeding Assessment Tool (>5), use of iron supplements (oral or injections). The adjusted analysis of the PBAC score showed a statistically significant positive relation with XI‐VOCAL counting (analyses adjusted for myoma, medication to reduce menstrual blood loss, and increased self‐BAT). The variables dysmenorrhea, chronic abdominal/pelvic pain, dyspareunia, and subjective assessment of blood loss did not demonstrate a statistically significant correlation with XI‐VOCAL slice count (Tables  5 and S4). Unadjusted results are shown in Table  S5 and also demonstrated that the statistically significant association with PBAC score remained (Figure  4 ), as well as a significant association between the XI‐VOCAL counting and hemoglobin value (Tables  S4 and S5 , Figure  5 ). Sensitivity analyses for the unadjusted and adjusted models are presented in Tables  S6 and S7 , respectively. Adjusted Linear and Logistic Regression of Correlation Between XI‐VOCAL Counting and XI‐VOCAL Categories Adenomyosis Assessment and Symptoms and Hemoglobin Levels Note : * P  < .05, ** P  < .01. Abbreviations: CI, confidence interval; NRS, Numeric Rating Scale; OR, odds ratio; VS, versus; XI‐VOCAL, Extended Imaging for Volume Calculation. Adjusted for uterine surgery (caesarean, laparoscopic or laparotomic myoma enucleation), endometriosis. Adjusted for endometriosis. Adjusted for type 0–5 myoma, medication to reduce menstrual blood loss (non‐steroidal anti‐inflammatory drugs, tranexamic acid), increased score of the Self‐Administered ISTH‐Bleeding Assessment Tool (>5). Adjusted for type 0–5 myoma, medication to reduce menstrual blood loss (non‐steroidal anti‐inflammatory drugs, tranexamic acid), increased score of the Self‐Administered ISTH‐Bleeding Assessment Tool (>5), use of iron supplements (oral or injections). Unadjusted correlation between XI‐VOCAL counting and PBAC score. Unadjusted correlation between XI‐VOCAL counting and Hb value. XI‐VOCAL categories and the clinical findings per severity category are presented in Table  6 . The mean PBAC score demonstrated an increasing trend as the XI‐VOCAL categories assessment became more severe. This finding was confirmed by a statistically significant Jonckheere‐Terpstra test ( P  150) did not display a tendency relative to the XI‐VOCAL categories severity of adenomyosis. Similarly, the data did not show a clear association between the degree of adenomyosis and NRS scores for dysmenorrhea, chronic abdominal/pelvic pain, and dyspareunia (Table  6 ). XI‐VOCAL category severe was positively correlated with the PBAC score in the adjusted analysis compared to mild XI‐VOCAL ( P  < .01) (Tables  5 and S8 , Figure  6 , analyses adjusted for myoma, medication to reduce menstrual blood loss, and increased self‐BAT). XI‐VOCAL categories were not correlated with any of the other outcomes (Tables  5 and S8 ). Unadjusted results were in line with the adjusted results and can be found in Table  S5 . Results of the corresponding sensitivity analyses can be found in Table  S6 (unadjusted) and Table  S7 (adjusted), which showed no statistically significant correlations, closely matching the findings of the original analyses. Comparison of Clinical Outcomes Across Adenomyosis Severity Assessed by XI‐VOCAL Categories Abbreviations: Hb, hemoglobin; NRS, Numeric Rating Scale; PBAC, pictorial blood loss analysis chart; XI‐VOCAL, Extended Imaging for Volume Calculation. Subjective assessment during outpatient clinic visits. Boxplot of unadjusted PBAC scores by XI‐VOCAL categories assessment (mild 1–5 slices, moderate 6–10 slices, severe 11–20). Significant correlation compared to mild XI‐VOCAL. In assessing the consistency between adenomyosis‐related symptoms, XI‐VOCAL results, and real‐time sonographic severity assessments by clinicians, the analysis showed that the real‐time assessment of severe vs mild adenomyosis was significantly correlated with the PBAC score ( β 590.009), which was consistent with XI‐VOCAL categories ( β 627.363). Additionally, XI‐VOCAL counting showed a significant correlation ( β 53.152), a result that was consistent despite XI‐VOCAL counting being a continuous variable, whereas PBAC and XI‐VOCAL categories were ordinal variables. However, real‐time assessments of adenomyosis severity were significantly correlated with the subjective assessment of blood loss, whereas no such correlation was found for the XI‐VOCAL counting or categories (Tables  4 and 5 ).

The real‐time ultrasound severity showed a statistically significant correlation with the adjusted PBAC and subjective blood loss. A strong correlation was found between XI‐VOCAL counting and XI‐VOCAL categories and adjusted PBAC. However, none of the ultrasound severity assessments correlated with adjusted dysmenorrhea, chronic abdominal/pelvic pain, and dyspareunia. Prior studies highlight the need for a uniform classification system for assessing adenomyosis with ultrasound imaging, yet no universally accepted or validated system exists. 6 , 8 , 34 , 35 , 36 Previous studies attempted to correlate the number of sonographic features with clinical symptoms, 37 , 38 finding indirect features most associated with pain. 8 , 37 However, these studies did not assess the spatial distribution of features (a specific region or diffusely distributed throughout the uterus) and assigned equal weight to all features. Another study proposed a scoring system based on affected myometrial layer, adenomyosis type, and thickness of the affected area but found no significant link to symptoms. 39 , 40 This scoring system, focused on junctional zone thickness, is controversial due to cycle phase variations and peristalsis and assumed measurement feasibility in diffuse adenomyosis. 41 To avoid equally weighing adenomyosis features and better account for the degree of myometrial involvement, this study employed the most recent subjective classification system, including direct and indirect ultrasound features of adenomyosis proposed by MUSA consensus group. 8 The correlation found between ultrasound findings and severity of symptoms contrasts with histological studies, which reported no correlation between heavy menstrual bleeding and adenomyosis extent. 42 However, a correlation between foci number (distinct lesions per histological section) and severe dysmenorrhea was found, and penetration depth (ratio of adenomyotic invasion depth to total myometrial thickness) was linearly related to pain score. 42 , 43 Similarly, research on endometriosis shows no correlation between severity stage and pain, suggesting that factors such as chronic inflammation or sensory nerve presence, not visible on ultrasound, may contribute to pain. 44 The observed 25% prevalence of elevated self‐BAT scores indicating an increased bleeding tendency aligns with prior Dutch research showing 29% of participants with heavy menstrual bleeding had an underlying bleeding disorder, irrespective of gynecological abnormalities. 45 XI‐VOCAL counting was significantly correlated with more objective assessments of menstrual blood loss (PBAC score and hemoglobin). Stronger correlations between real‐time assessment and clinical outcomes than between XI‐VOCAL and clinical outcomes suggest that symptom severity may influence subjective evaluations. This is evident as real‐time assessment significantly correlated with subjective blood loss (moderate vs. mild: OR = 3.423 [95% CI 1.054–11.121]; severe vs. mild: OR = 5.476 [95% CI 1.141–26.283]) and dysmenorrhea severity in unadjusted analyses (β = 1.434 [95% CI 0.101–2.767]), whereas XI‐VOCAL showed no significant correlations. However, real‐time assessments during outpatient visits may provide additional, dynamic insights, such as sliding viscera. The sliding sign can provoke pain in patients with pelvic floor hypertonia or intra‐abdominal adhesions, which may explain pain not directly related to adenomyosis on ultrasound. This highlights the added value of real‐time evaluations. In contrast, XI‐VOCAL counting provides a standardized sonographic evaluation of disease severity, offering a standardized alternative for research and potential clinical use. Though not yet implemented in clinical settings, XI‐VOCAL could help assess adenomyosis severity, evaluate disease progression, thereby improving prognosis accuracy. Moreover, by distinguishing mild, moderate, and severe adenomyosis, it may enable more personalized therapeutic strategies and improve patient counseling. This may also benefit fertility care. The secondary objective was supported, with both XI‐VOCAL and real‐time assessments showing strong correlations with objective measures of blood loss, such as the PBAC score. While real‐time assessments offered stronger correlations with subjective outcomes, both methods have distinct value: XI‐VOCAL provides an objective evaluation of adenomyosis severity, while real‐time assessments offer dynamic, clinically relevant insights. The fact that participants without self‐reported dysmenorrhea did not provide their NRS scores may partly explain the high NRS median and the absent correlation with adenomyosis severity. To address this, missing NRS values were imputed as zero for participants who reported no pain. The results of these analyses (Tables  S6 and S7 ) are highly consistent with the primary results, with similar effect estimates and confidence intervals. This suggests that our main findings were not substantially influenced by selective symptom reporting. Still, prospective studies with systematic symptom scoring would further strengthen them. The high median NRS score for dysmenorrhea may have masked a correlation with XI‐VOCAL, as most participants reported severe pain, likely due to the tertiary referral center setting. The study sample predominantly comprised individuals seeking specialized consultations for adenomyosis and endometriosis. Other pain‐related factors (e.g., cesarean section, laparoscopic or open myomectomy, presence of endometriosis) were adjusted for in the analyses, but bias may remain due to the complex, multifactorial (including mental health and smoking) and not fully elucidated nature of pain. 46 , 47 , 48 , 49 , 50 This study used a novel semi‐quantified sonographic method to assess adenomyosis severity, enabling blinded interpretation of uterine volume independent of clinical symptoms to reduce bias. While blinding is essential for research, clinical use would likely involve symptom knowledge. Still, the standardized 3D ultrasound evaluation and redefined MUSA criteria enhance recognition, diagnosis, and consistency. Covariates were adjusted for, and the self‐BAT was used to account for underlying bleeding disorders, often overlooked despite their high prevalence in heavy menstrual bleeding. 45 , 51 Significant correlations between real‐time ultrasound assessment (subjective) and XI‐VOCAL (standardized) with the PBAC score, even after adjustment for bleeding tendency, support their independent value in quantifying adenomyosis. Additionally, the study linked adenomyosis measurements with core symptomatic outcomes, including pain, bleeding, reproductive outcomes, and haematology. 52 XI‐VOCAL counting required ~2 min per patient, and scored 9/9 for feasibility in a previous pilot study, 13 indicating potential for outpatient use and as a reference for future tools, such as AI‐based approaches. Limitations include the partially retrospective PBAC assessment with 58% missing data and hemoglobin values available for only 65% of participants, as laboratory testing was based on clinical indication rather than protocol. Also, excluding patients receiving hormonal medication may have introduced bias, as those with more severe symptoms are more likely to use such medication. Additionally, XI‐VOCAL was only evaluated by experienced ultrasound examiners with specific expertise in the evaluation of adenomyosis, and performance with less‐experienced sonographers remains untested, limiting external validity and generalizability. 53 Also, assessing interobserver reliability using stored video/3D volumes may overestimate reproducibility compared to real‐time scanning. Ideally, this would be assessed in real time, but a second examination was considered too burdensome. Moreover, although adenomyosis diagnosis was based on ≥1 direct feature, recent research has shown poor intra‐ and interobserver reproducibility of these features on ultrasound (Kappa 0.27 and 0.13), raising concerns about diagnostic consistency. 54 Additionally, differentiating myoma size would be useful, as it affects clinical outcomes. 55 Finally, no log of patients declining participation was kept, limiting insight into inclusion bias. We recommend a prospective cohort study with a representative sample from various healthcare settings, covering a wide range of adenomyosis severity, including fertility and asymptomatic patients. The objective is to investigate the correlation between sonographic severity, assessed by real‐time ultrasound and XI‐VOCAL, and symptoms across a broader, more generalizable patient population. Potential influencing factors such as uterine fundus involvement and volume or layer‐specific involvement should be explored. Additionally, ferritin levels, alongside hemoglobin, should be included to provide a more comprehensive understanding of the relationship between blood loss and adenomyosis severity. While this study highlighted a high prevalence of subfertility and recurrent miscarriages in adenomyosis patients, the relationship between adenomyosis severity and fertility outcomes remains unexamined; previous research identified predictors of clinical pregnancy but overlooked the degree of myometrial involvement. 56 Finally, techniques such as elastography, contrast‐enhanced ultrasound, and microvascular flow may also add value. 57 , 58 , 59

XI‐VOCAL appears to be a promising offline sonographic technique for evaluating the severity of adenomyosis, as it offers a more standardized approach compared to clinicians' real‐time ultrasound assessments during outpatient visits. There seems to be a correlation between XI‐VOCAL and more objective measures of heavy menstrual bleeding, yet a comprehensive cohort study is necessary to investigate whether there is a relationship between the degree of adenomyosis and dysmenorrhea, chronic abdominal/pelvic pain, and dyspareunia.

A prospective observational cohort study was performed at the outpatient clinic of a tertiary referral center for gynecological care between 2021 and 2024. This referral center is specifically specialized in adenomyosis, endometriosis, and uterine niches, and patients were referred either from secondary care or from primary care in case of a clearly tertiary care indication. The local ethical committee approved the study (number 2021.0035) which prospectively registered ( NCT06117410 ). A previous pilot study, which aimed to develop the most feasible method to standardize the assessment of adenomyosis severity with the best inter‐ and intra‐observer agreement, was executed before the current study's design. 13 All women with adenomyosis, defined as the presence of at least one direct MUSA feature on TVUS performed on a Samsung WS80A or HeraI10 machine, who visited the outpatient clinic of a tertiary referral center for gynecological care, were included consecutively after obtaining informed consent. Exclusion criteria were the presence of dominant myoma(s) (diameter greater than 5 cm or more than three myomas), (suspected) uterine or cervical malignancy, active pelvic inflammatory disease, current pregnancy, breastfeeding, (peri)menopausal status or current use of hormonal medication (such as oral contraceptives, levonorgestrel‐releasing intra‐uterine devices, ulipristal and gonadotrophin‐releasing hormone agonists and antagonists), or missing or inadequate ultrasound imaging. Data were prospectively gathered from patients who visited the outpatient clinic between May 2021 and July 2024, with retrospective data collection via an online questionnaire in cases of missing information. Each patient underwent a standardized assessment, which is part of our routine clinical practice, and included a detailed review of their medical, obstetric, and gynecological history, as well as their current gynecological symptoms and medication use. Pain symptoms were evaluated using a Numeric Rating Scale (NRS) if pain was reported. Menstrual blood loss was evaluated by patients' subjective assessment as light, normal, or heavy, and they were additionally asked to report the presence of blood clots (Supporting Information Form S1). This was followed by a physical examination and a standardized transvaginal sonographic examination by one of six experienced gynecologists. The physical and sonographic examinations were not only standardized for the purpose of this study but also part of our routine clinical protocol. As part of the standardized transvaginal ultrasound, the presence of adenomyosis was assessed using the MUSA criteria. The study used two ultrasound machines: a Samsung WS80A machine with a V5‐9 intra‐cavity transducer and a Samsung HeraI10 machine with an EV2‐10A intra‐cavity transducer (Samsung, Seoul, South Korea). Standardized predefined settings were used for each machine to ensure consistency in image acquisition; however, sonographers were allowed to adjust the gain and depth as needed to optimize image quality for individual patients (Figure  S1 ). Only patients who were seen in one of the two consultation rooms equipped with these ultrasound systems were eligible for inclusion to ensure availability of 3D transvaginal sonography and consistency in image quality. A clinical standardized protocol, including power Doppler, was followed for both 2D and 3D transvaginal sonographic examinations. These examinations as well as the standardized reporting of the findings are part of our clinical routine and were systematically stored for every patient who visits our department and focused on capturing specific sonographic findings, including uterine size, the presence of intracavitary abnormalities, myomas, signs of endometriosis, and the assessment of adenomyosis using the MUSA criteria (Figure  1 ). Specifically, a sagittal still image and video were obtained both with and without Doppler. A 3D volume acquisition was then performed. In the transverse plane, a still image and a sweep video from cervix to fundus were recorded. Both ovaries were assessed individually, as well as the presence of sliding signs of the bladder and bowels. Finally, any additional abnormalities were evaluated. The affected myometrium was estimated and reported as proposed by MUSA: mild (50% affected). 14 This will be referred to as real‐time severity. A pilot study, in which this method was used as a secondary outcome, demonstrated good inter‐observer agreement (Fleiss' κ  = 0.647, P  < .001), supporting its reproducibility in clinical practice. 13 To assess for endometriosis, the International Deep Endometriosis Analysis (IDEA) protocol was followed for every patient as part of routine clinical care. This included systematic evaluation of the uterus and adnexa, assessment of soft markers (such as site‐specific tenderness and ovarian mobility), examination of the real‐time ultrasound sliding sign, and assessment of deep endometriosis nodules in both the anterior and posterior compartments. Clinical suspicion of endometriosis based on these findings was noted in the case report form. 15 Noteworthy, the imaging and its storage are performed for every patient visiting our department; however, for this study, only the ultrasound scans of patients who met the inclusion criteria were reviewed and assessed. Systematic real‐time assessment at the outpatient clinic. As part of routine care, hemoglobin level was assessed in case of heavy menstrual bleeding. In the absence of reported real‐time severity due to protocol deviation, another specialized gynecologist (RL) assessed the stored ultrasound video clips and classified the condition as mild, moderate, or severe. The medical history, clinical characteristics, sonographic assessment, and laboratory data were pseudonymized, stored securely and filled out in an electronic case report form (eCRF) by the researchers, based on data collected prospectively during clinical consultations and recorded in the patient's medical records. However, because data collection took place during clinical consultations, prioritization of patient care sometimes led to incomplete documentation. Although symptoms were routinely discussed during the outpatient visit, it was not always documented whether specific symptoms were explicitly absent. To improve data completeness, a follow‐up questionnaire was sent retrospectively to all patients who met the inclusion but not exclusion criteria and had signed the informed consent (Supporting Information Form S2), including reminders of the questionnaire. Participants were first asked whether they could recall their symptoms at the time of the outpatient visit. If so, they continued with the questionnaire. If not, they were allowed to proceed only if their symptoms remained stable since that visit; otherwise, the questionnaire was discontinued to minimize recall bias. This questionnaire included the following subjects, all mandatory, and was sent between several weeks to 1.5 years after the initial outpatient visit: Pain: the presence of dysmenorrhea, pelvic pain, dyspareunia, and corresponding pain score using a numerical rating scale (NRS from 0 to 10) when the symptoms were present (definitions can be found in Figure  S1 ). 16 , 17 Patients who did not report the presence of pain were not asked to report a NRS pain score. Menstrual cycle: The frequency, duration of the menstrual cycle and bleeding, episodes of intermenstrual or postcoital bleeding, and the presence of blood clots were recorded. Subjective amount of menstrual blood loss using a Likert scale (little, normal, heavy). Pictorial Blood Loss Assessment Chart (PBAC) of the subsequent menstrual cycle after inclusion in the study. In instances where the PBAC was not returned, only participants who reported that their menstrual bleeding pattern remained consistent compared to the bleeding pattern at the inclusion date (i.e., date of clinical examination and ultrasound) were requested to complete the PBAC. Conversely, if the participants could not recall the details or the bleeding had changed, the PBAC could not be completed to reduce the potential for recall bias. A PBAC score exceeding 150 was considered an elevated score, given the associated sensitivity of 78–83% and specificity of 77–88% with this threshold. 18 , 19 The validated Dutch Self‐Administered ISTH (International Society on Thrombosis and Hemostasis)‐Bleeding Assessment Tool (self‐BAT), which is based on the ISTH‐BAT, was completed to assess the severity of bleeding symptoms and identify mild bleeding disorders with high sensitivity and negative predictive value. 20 , 21 , 22 An abnormal bleeding score for women has been defined as greater than 5. 23 Pain: the presence of dysmenorrhea, pelvic pain, dyspareunia, and corresponding pain score using a numerical rating scale (NRS from 0 to 10) when the symptoms were present (definitions can be found in Figure  S1 ). 16 , 17 Patients who did not report the presence of pain were not asked to report a NRS pain score. Menstrual cycle: The frequency, duration of the menstrual cycle and bleeding, episodes of intermenstrual or postcoital bleeding, and the presence of blood clots were recorded. Subjective amount of menstrual blood loss using a Likert scale (little, normal, heavy). Pictorial Blood Loss Assessment Chart (PBAC) of the subsequent menstrual cycle after inclusion in the study. In instances where the PBAC was not returned, only participants who reported that their menstrual bleeding pattern remained consistent compared to the bleeding pattern at the inclusion date (i.e., date of clinical examination and ultrasound) were requested to complete the PBAC. Conversely, if the participants could not recall the details or the bleeding had changed, the PBAC could not be completed to reduce the potential for recall bias. A PBAC score exceeding 150 was considered an elevated score, given the associated sensitivity of 78–83% and specificity of 77–88% with this threshold. 18 , 19 The validated Dutch Self‐Administered ISTH (International Society on Thrombosis and Hemostasis)‐Bleeding Assessment Tool (self‐BAT), which is based on the ISTH‐BAT, was completed to assess the severity of bleeding symptoms and identify mild bleeding disorders with high sensitivity and negative predictive value. 20 , 21 , 22 An abnormal bleeding score for women has been defined as greater than 5. 23 The frequency, duration of the menstrual cycle and bleeding, episodes of intermenstrual or postcoital bleeding, and the presence of blood clots were recorded. Subjective amount of menstrual blood loss using a Likert scale (little, normal, heavy). Pictorial Blood Loss Assessment Chart (PBAC) of the subsequent menstrual cycle after inclusion in the study. In instances where the PBAC was not returned, only participants who reported that their menstrual bleeding pattern remained consistent compared to the bleeding pattern at the inclusion date (i.e., date of clinical examination and ultrasound) were requested to complete the PBAC. Conversely, if the participants could not recall the details or the bleeding had changed, the PBAC could not be completed to reduce the potential for recall bias. A PBAC score exceeding 150 was considered an elevated score, given the associated sensitivity of 78–83% and specificity of 77–88% with this threshold. 18 , 19 The validated Dutch Self‐Administered ISTH (International Society on Thrombosis and Hemostasis)‐Bleeding Assessment Tool (self‐BAT), which is based on the ISTH‐BAT, was completed to assess the severity of bleeding symptoms and identify mild bleeding disorders with high sensitivity and negative predictive value. 20 , 21 , 22 An abnormal bleeding score for women has been defined as greater than 5. 23 The 3D sonographic images were assessed offline using a stand‐alone software for XI‐VOCAL (Samsung Medison 5D Viewer software). Severity was assessed in 20 parallel transversal planes of the stored 3D volumes. Counting the number of affected slices of the total of 20 assessed slices between fundus and cervix with at least one direct MUSA feature (Figure  2 ) was called XI‐VOCAL counting. XI‐VOCAL counting was also categorized (XI‐VOCAL categories) into mild (0–5 affected slices), moderate (6–10 affected slices), and severe (11–20 affected slices). The categories were based on the estimated percentages of affected myometrium in the categories of the subjective assessment of the MUSA consensus group. 10 For details of the predefined criteria in the assessment of adenomyosis see Figure  S1 . The formal feasibility of the XI‐VOCAL method, in terms of technique, time, and interpretation, was previously evaluated, 13 in which it achieved a maximum score (9/9) indicating good feasibility for clinical implementation. In the same study, interobserver reliability for XI‐VOCAL counting was good (ICC = 0.733, P  < .001). One gynecological sonographer (LMT, 5 years of experience) assessed the 3D sonographic images blinded for severity of symptoms to ensure unbiased evaluation. XI‐VOCAL counting, severity presented in ratio of affected slices (n/20) (Trommelen et al. 13 ). The sample size was determined through a power analysis for a multiple regression model, assuming a small to medium effect size ( f 2  = 0.1) and a significance level of 0.05 ( α ). This F ‐test–based calculation only applied to the continuous outcome variables. To meet the assumption of normality of the residuals, this was explicitly tested and confirmed. Assuming eight predictor variables (with adenomyosis severity as the key variable of interest), 91 participants were required to achieve a power of 0.80. Considering a non‐response rate of 20%, 115 participants were needed. Statistical analyses were performed using IBM SPSS Statistics Version 28, and a P ‐value of <.05 was considered statistically significant. Descriptive statistics were given for the study demographics. Jonckheere‐Terpstra test was used to investigate the trend between the XI‐VOCAL categories and symptoms and between the real‐time severity scores and symptoms. To examine the correlation of real‐time severity, XI‐VOCAL counting, and XI‐VOCAL categories with clinical outcomes, as well as linear and logistic regression analyses, were performed. Potential covariates were predefined based on logic and what is known in the literature. Specifically, the PBAC score and subjective assessment of blood loss analyses were adjusted for FIGO classification type 0–5 myoma (based on ultrasound), 24 , 25 , 26 , 27 medication to reduce menstrual blood loss (nonsteroidal anti‐inflammatory drug [NSAID], tranexamic acid), 28 and increased self‐BAT score. 29 The same variables and the use of iron supplements (tablets or injections) 30 were used to correct the Hb values. For the outcomes of dysmenorrhea and chronic abdominal/pelvic pain, the analyses were adjusted for previous uterine surgery (cesarean and laparotomic or laparoscopic myomectomy), 31 , 32 and surgically proven or ultrasound suspected endometriosis (based on patient history or ultrasound). 33 Dyspareunia was adjusted for endometriosis. 33 These covariates were addressed using multivariable linear or logistic regression to ensure the observed relationships were not biased by these factors. Additionally, sensitivity analyses were performed to assess the robustness of the findings. Specifically, for participants who reported no pain symptoms and therefore did not receive an NRS score, an NRS score of zero was imputed. The adjusted analyses using these imputed values were compared with the original analyses.

Data S1. Supporting Information. Data S2. Supporting Information.