Diagnostic accuracy of transvaginal ultrasound for adenomyosis using consensus-based direct and indirect sonographic signs: prospective cohort study

Objective To evaluate the diagnostic performance of transvaginal ultrasound (TVS) for adenomyosis using consensus‐based direct and indirect ultrasound signs, with histopathology as the reference standard.

This prospective diagnostic accuracy study included women undergoing TVS examination at a tertiary referral center (Hospital San Juan de Dios, Santiago, Chile) between 1 December 2022 and 1 December 2023. Women aged ≥ 18 years who underwent hysterectomy within 6 months after TVS examination were enrolled consecutively. Cases with malignant histopathological findings were excluded. Ultrasound signs of adenomyosis were assessed according to the revised Morphological Uterus Sonographic Assessment consensus. Histopathology served as the reference standard. Diagnostic performance measures for individual ultrasound features and multivariable models were calculated.

Among 2967 women evaluated using TVS, 345 (11.6%) underwent hysterectomy within 6 weeks and constituted the study cohort. Histopathology confirmed adenomyosis in 155 (44.9%) women. Individual ultrasound features showed high specificity but limited sensitivity. Among direct signs, echogenic subendometrial lines showed the best overall balance (sensitivity, 53.5% (95% CI, 45.4–61.6%); specificity, 94.7% (95% CI, 90.5–97.4%)). Individual indirect signs showed specificity > 97% but low sensitivity. A model combining direct ultrasound signs showed high specificity (93.7% (95% CI, 89.2–96.7%)) but limited sensitivity (60.0% (95% CI, 51.8–67.8%)); overall diagnostic accuracy was 78.6% (95% CI, 73.8–82.8%). The addition of indirect ultrasound signs into the model did not provide meaningful incremental benefit.

Our findings highlight the importance of distinguishing between direct and indirect ultrasound features, with direct signs providing the most consistent diagnostic value, while indirect signs should be interpreted as supportive rather than definitive. © 2026 International Society of Ultrasound in Obstetrics and Gynecology.

adenomyosis, diagnostic accuracy, MUSA consensus, transvaginal ultrasound

Adenomyosis is a common gynecological disorder defined by the presence of ectopic endometrial glands and stroma within the uterine myometrium, resulting in hypertrophy and hyperplasia of the adjacent myometrial tissue 1 . The reported prevalence varies widely, ranging from 5% to 70%, mainly owing to the historical lack of standardized diagnostic criteria 2 . Traditionally, diagnosis has relied on histopathological findings obtained from hysterectomy specimens, which has limited accurate estimation of the prevalence of the disease and the assessment of it in women of reproductive age 3 . In current clinical practice, transvaginal ultrasound (TVS) is considered the first‐line imaging modality for the evaluation of adenomyosis because of its wide availability, low cost and non‐invasive nature. Several retrospective studies have demonstrated that TVS provides adequate diagnostic performance, with reported sensitivity and specificity of approximately 78% and 87%, respectively 4 , 5 , 6 . Beyond its association with pelvic pain and abnormal uterine bleeding, adenomyosis has been recognized increasingly for its impact on fertility and reproductive outcomes 7 , 8 , 9 . Nirgianakis et al. 10 investigated the role of adenomyosis in women undergoing assisted reproductive technology and reported a significant association with a lower pregnancy rate (odds ratio (OR), 0.69 (95% CI, 0.51–0.94)) and a higher rate of miscarriage (OR, 2.17 (95% CI, 1.25–3.79)) in women with ultrasound findings suggestive of adenomyosis. To improve diagnostic reproducibility, the ultrasound criteria for adenomyosis were standardized by expert sonographers as part of the Morphological Uterus Sonographic Assessment (MUSA) consensus, published in 2015 and subsequently updated in 2019 11 , 12 . More recently, a Delphi consensus of Harmsen et al. 13 refined the definitions of uterine morphological features associated with adenomyosis and classified them into direct and indirect criteria according to their relationship with the underlying pathophysiology. Direct signs indicate the presence of ectopic endometrial tissue within the myometrium, whereas indirect signs reflect secondary changes, such as muscular hypertrophy (globular uterus) or ultrasound artifacts, including acoustic shadowing. The aim of this study was to evaluate the diagnostic performance of TVS for adenomyosis, using recently defined direct and indirect ultrasound features, and to assess whether specific combinations of these features could improve overall diagnostic accuracy.

This prospective cohort study was conducted at a tertiary referral center (Hospital San Juan de Dios, Santiago, Chile). All women undergoing TVS between 1 December 2022 and 1 December 2023 were evaluated consecutively for the presence or absence of ultrasound features suggestive of adenomyosis. Women aged ≥ 18 years undergoing TVS who subsequently underwent hysterectomy within 6 months were eligible for inclusion. Patients were enrolled consecutively during the study period. Cases with malignant histopathological findings at hysterectomy were excluded. Ethical approval was granted by the Hospital San Juan de Dios Ethics Committee (approval number: 08064/2023), and all patients provided written informed consent to participate in the study. Ultrasound examination Targeted assessment for adenomyosis was performed systematically during each TVS examination, evaluating direct and indirect ultrasound features according to the definitions proposed by the revised MUSA consensus 13 . Direct ultrasound features included hyperechogenic islands, echogenic subendometrial lines and myometrial cysts (Figure 1). Indirect ultrasound features included asymmetrical myometrial wall thickening, globular uterus, translesional vascularity, fan‐shaped shadowing, irregular endometrial–myometrial junction and interrupted endometrial–myometrial junction (Figure 2). In addition, the presence of heterogeneous myometrium and the ‘question mark’ sign, as described by Di Donato et al. 14 , were assessed (Figure 3). TVS examinations were performed using a Voluson S8 ultrasound system (GE Healthcare, Zipf, Austria) equipped with a 4–9‐MHz transvaginal probe. All scans were conducted by one of two authors (C.M., E.C.), both highly experienced gynecological sonographers classified as Level III operators according to the European Federation of Societies for Ultrasound in Medicine and Biology (EFSUMB) criteria 15 . Histopathological assessment Patients were followed prospectively to identify those who subsequently underwent hysterectomy for any indication. Histopathological examination of the uterus, obtained within a maximum interval of 6 months after the TVS examination, served as the reference standard for the diagnosis of adenomyosis. Hysterectomy specimens were processed according to the institutional pathology protocol for benign uterine disease. Representative full‐thickness sections of the anterior and posterior uterine walls (from the endometrium to the serosa) were obtained routinely, along with additional sections from the uterine isthmus and cervix. In total, a minimum of four representative histopathological sections per uterus were evaluated systematically, with additional sampling performed in the presence of macroscopically abnormal areas such as fibroids or other suspicious lesions. All specimens were examined by experienced gynecological pathologists who were blinded to the ultrasound findings. Adenomyosis was defined as the presence of endometrial glands and stroma within the myometrium, clearly separated from the endometrial–myometrial junction. As adenomyosis may present as a focal condition, it is acknowledged that histopathological sampling may not detect all cases. However, systematic representative sampling from different uterine regions was performed in all cases to reduce the risk of underdiagnosis. Histopathological findings were compared with the preoperative TVS assessment. Statistical analysis Statistical analysis was performed using R software version 4.2.3 (R Foundation for Statistical Computing, Vienna, Austria) and RStudio version 1.2.5033 (Posit PBC, Boston, MA, USA). Demographic, clinical and ultrasound characteristics were compared between patients with and without histopathologically confirmed adenomyosis. Ultrasound findings were compared with histopathological findings, which served as the reference standard, to calculate diagnostic accuracy measures, including sensitivity, specificity, positive predictive value (PPV), negative predictive value, overall diagnostic accuracy and OR, all reported with their corresponding 95% CI. Data are presented as n (%), mean ± SD or median (interquartile range (IQR)), as appropriate. Categorical variables were analyzed using the chi‐square test. Parametric continuous variables were analyzed using Student's t‐test for comparison of means, with the Mann–Whitney U‐test used as a non‐parametric alternative where appropriate. Multivariable logistic regression models were fitted using maximum likelihood estimation with a logit link function to evaluate the association between ultrasound features and histopathologically confirmed adenomyosis. Regression coefficients, adjusted ORs, 95% CIs and P‐values are reported for individual predictors. Model discrimination was assessed using classification performance at a prespecified probability threshold of 0.5, as conventionally applied in logistic regression models, and agreement between model‐based classification and histopathology was evaluated using Cohen's kappa coefficient. To assess the robustness of this threshold, the optimal cut‐off was additionally estimated using the Youden index derived from receiver‐operating‐characteristics (ROC)‐curve analysis. The incremental diagnostic value of indirect ultrasound features beyond that of direct ultrasound features was evaluated by comparing a combined multivariable model including both direct and indirect ultrasound features with a model including direct ultrasound features alone, using likelihood ratio testing and Akaike's Information Criterion (AIC). A parsimonious modeling approach was favored when no significant improvement in model fit was observed. All statistical tests were two‐sided. P < 0.05 was considered statistically significant.

During the study period, a total of 2967 women underwent TVS at our center. Among these, 456 (15.4%) women presented with at least one direct or indirect ultrasound feature suggestive of adenomyosis. Of all the women evaluated using TVS during the study period, 345 (11.6%) subsequently underwent hysterectomy for any indication within 6 months after the initial TVS examination, and constituted the study cohort. The most common indication for hysterectomy was abnormal uterine bleeding (155/345 (44.9%)), followed by endometrial pathology (71/345 (20.6%)), adnexal pathology (50/345 (14.5%)), pelvic organ prolapse (25/345 (7.2%)), uterine fibroids not associated with abnormal uterine bleeding (21/345 (6.1%)), cervical pathology (18/345 (5.2%)) and other indications (5/345 (1.4%)). Histopathological examination confirmed adenomyosis in 155 (44.9%) women, while 190 (55.1%) women showed no histopathological evidence of adenomyosis. Demographic and clinical characteristics of the study cohort are summarized in Table 1. Women with histopathologically confirmed adenomyosis were slightly younger (mean age, 48.6 ± 9.3 vs 52.4 ± 11.9 years; P = 0.05) and more frequently premenopausal (116/155 (74.8%) vs 96/190 (50.5%); P < 0.01) compared with women without adenomyosis. Clinical symptoms were assessed only in premenopausal women (n = 212). In this subgroup, women with adenomyosis more frequently reported abnormal uterine bleeding (95/116 (81.9%) vs 62/96 (64.6%); P < 0.01), dysmenorrhea (77/116 (66.4%) vs 31/96 (32.3%); P < 0.01) and chronic pelvic pain (33/116 (28.4%) vs 17/96 (17.7%); P = 0.01) than women without adenomyosis. Uterine fibroids were present in 194/345 (56.2%) of the study population, with no significant difference between women with and those without adenomyosis (83/155 (53.5%) vs 111/190 (58.4%); P = 0.42). A significant difference in overall uterine volume was observed between women with and those without adenomyosis (median uterine volume, 99.0 (IQR, 65.0–146.0) cm3 vs 88.0 (IQR, 47.3–201.5) cm3; P < 0.01). Given that uterine fibroids represent the main confounding factor influencing uterine size, uterine volume was also analyzed specifically in women without fibroids to isolate the effect of adenomyosis on uterine enlargement. In this subgroup, median uterine volume remained significantly higher in women with adenomyosis compared to those without adenomyosis (76.5 (IQR, 57.5–119.0) cm3 vs 61.0 (IQR, 35.0–90.0) cm3; P < 0.01). Table 1. | Characteristic | Total (n = 345) | Adenomyosis (n = 155) | No adenomyosis (n = 190) | P | |---|---|---|---|---| | Age (years) | 50.9 ± 11.1 | 48.6 ± 9.3 | 52.4 ± 11.9 | 0.05 | | Body mass index (kg/m2) | 30.9 ± 6.33 | 31.08 ± 6.24 | 30.77 ± 6.39 | 0.65 | | Nulliparous | 31 (9.0) | 13 (8.4) | 18 (9.5) | 0.87 | | Premenopausal | 212 (61.5) | 116 (74.8) | 96 (50.5) | < 0.01 | | Presence of uterine fibroids | 194 (56.2) | 83 (53.5) | 111 (58.4) | 0.42 | | Uterine volume (cm3) | |||| | Overall | 92.0 (55.0–176.0) | 99.0 (65.0–146.0) | 88.0 (47.3–201.5) | < 0.01 | | Patients without fibroids* | 68.0 (42.5–103.0) | 76.5 (57.5–119.0) | 61.0 (35.0–90.0) | < 0.01 | | Clinical symptoms† | |||| | Abnormal uterine bleeding | 157/212 (74.1) | 95/116 (81.9) | 62/96 (64.6) | < 0.01 | | Dysmenorrhea | 108/212 (50.9) | 77/116 (66.4) | 31/96 (32.3) | < 0.01 | | Chronic pelvic pain | 50/212 (23.6) | 33/116 (28.4) | 17/96 (17.7) | 0.01 | Data are presented as mean ± SD, n (%), median (interquartile range) or n/N (%), as appropriate. Adenomyosis status was defined according to histopathological findings. Analysis restricted to patients without uterine fibroids (total, n = 151: adenomyosis, n = 72; no adenomyosis, n = 79). Clinical symptoms were analyzed only in premenopausal women (n = 212). The diagnostic performance of individual ultrasound features for the detection of adenomyosis in the overall cohort is detailed in Table 2. Among direct ultrasound features, echogenic subendometrial lines showed the best overall balance, with a sensitivity of 53.5% (95% CI, 45.4–61.6%), specificity of 94.7% (95% CI, 90.5–97.4%) and diagnostic accuracy of 76.2% (95% CI, 71.4–80.6%). Myometrial cysts demonstrated high specificity (98.4% (95% CI, 95.5–99.7%)) and PPV (95.5% (95% CI, 87.5–99.1%)), with a sensitivity of 41.3% (95% CI, 33.5–49.5%). Hyperechogenic islands were infrequent but highly specific (specificity, 100% (95% CI, 98.1–100%)), resulting in a low sensitivity of 14.2% (95% CI, 9.1–20.7%) despite a PPV of 100% (95% CI, 84.6–100%). Table 2. | Ultrasound signs | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | Accuracy (%) | |---|---|---|---|---|---| | Direct ultrasound signs | ||||| | Myometrial cysts (n = 67) | 41.3 (33.5–49.5) | 98.4 (95.5–99.7) | 95.5 (87.5–99.1) | 67.3 (61.4–72.8) | 72.8 (67.7–77.4) | | Echogenic subendometrial lines (n = 93) | 53.5 (45.4–61.6) | 94.7 (90.5–97.4) | 89.2 (81.1–94.7) | 71.4 (65.4–76.9) | 76.2 (71.4–80.6) | | Hyperechogenic islands (n = 22) | 14.2 (9.1–20.7) | 100 (98.1–100) | 100 (84.6–100) | 58.8 (53.2–64.2) | 61.4 (56.1–66.6) | | Indirect ultrasound signs | ||||| | Asymmetrical myometrial wall thickening (n = 35) | 21.3 (15.1–28.6) | 98.9 (96.2–99.9) | 94.3 (80.8–99.3) | 60.6 (54.9–66.1) | 64.1 (58.7–69.1) | | Globular uterus (n = 58) | 34.8 (27.4–42.9) | 97.9 (94.7–99.4) | 93.1 (83.3–98.1) | 64.8 (59.0–70.3) | 69.6 (64.4–74.4) | | Translesional vascularity (n = 18) | 11.0 (6.5–17.0) | 99.5 (97.1–100) | 94.4 (72.7–99.9) | 57.8 (52.2–63.2) | 59.7 (54.3–64.9) | | Fan‐shaped shadowing (n = 38) | 22.6 (16.3–30.0) | 98.4 (95.5–99.7) | 92.1 (78.6–98.3) | 60.9 (55.2–66.4) | 64.3 (59.0–69.4) | | Irregular endometrial–myometrial junctional zone (n = 63) | 36.1 (28.6–44.2) | 96.3 (92.6–98.5) | 88.9 (78.4–95.4) | 64.9 (59.0–70.5) | 69.3 (64.1–74.1) | | Interrupted endometrial–myometrial junctional zone (n = 25) | 14.8 (9.6–21.4) | 98.9 (96.2–99.9) | 92.0 (74.0–99.0) | 58.8 (53.1–64.2) | 61.2 (55.8–66.3) | | Other ultrasound signs | ||||| | Heterogeneous myometrium (n = 244) | 83.9 (77.1–89.3) | 40.0 (33.0–47.3) | 53.3 (46.8–59.7) | 75.2 (65.7–83.3) | 59.7 (54.3–64.9) | | Question mark sign (n = 15) | 9.0 (5.0–14.7) | 99.5 (97.1–100) | 93.3 (68.1–99.8) | 57.3 (51.7–62.7) | 58.8 (53.4–64.1) | Data in parentheses are 95% CI. NPV, negative predictive value; PPV, positive predictive value. Indirect ultrasound features were generally characterized by high specificity and limited sensitivity. Asymmetrical myometrial wall thickening and globular uterus showed a specificity of 98.9% (95% CI, 96.2–99.9%) and 97.9% (95% CI, 94.7–99.4%), respectively, with a sensitivity of 21.3% (95% CI, 15.1–28.6%) and 34.8% (95% CI, 27.4–42.9%), respectively. Other indirect signs, including translesional vascularity and fan‐shaped shadowing, demonstrated very high specificity (> 98%) but low sensitivity (< 25%). Among other ultrasound features, heterogeneous myometrium was highly sensitive (sensitivity, 83.9% (95% CI, 77.1–89.3%)) but had low specificity (40.0% (95% CI, 33.0–47.3%)) for the detection of adenomyosis, whereas the question mark sign showed very high specificity (99.5% (95% CI, 97.1–100%)) but low sensitivity (9.0% (95% CI, 5.0–14.7%)). A subgroup analysis restricted to premenopausal women (n = 212) showed a similar diagnostic pattern, with a modest increase in sensitivity and a slight reduction in specificity on average across the different ultrasound signs. Detailed diagnostic performance of individual ultrasound features in this subgroup is provided in Table S1. Multivariable logistic regression models combining ultrasound features showed diagnostic performance comparable to that of individual signs, with only a modest increase in sensitivity and similar specificity (Table 3). The model based on direct ultrasound features achieved a sensitivity of 60.0% (95% CI, 51.8–67.8%), specificity of 93.7% (95% CI, 89.2–96.7%) and overall diagnostic accuracy of 78.6% (95% CI, 73.8–82.8%). The addition of indirect ultrasound features to the model did not result in a significant improvement in performance. Likelihood‐ratio testing comparing the direct‐sign model with the combined direct‐ and indirect‐sign model confirmed that inclusion of indirect ultrasound features did not significantly improve model fit (χ2 = 5.99, degrees of freedom = 6, P = 0.425). Similarly, the direct‐sign model had a lower AIC than the combined model (340.5 vs 346.5), supporting its selection as the more parsimonious model. Table 3. | Model | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | Accuracy (%) | Cohen's κ (%) | |---|---|---|---|---|---|---| | Direct ultrasound signs* | 60.0 (51.8–67.8) | 93.7 (89.2–96.7) | 88.6 (80.9–94.0) | 74.2 (68.1–79.6) | 78.6 (73.8–82.8) | 55.0 (47.0–64.0) | | Indirect ultrasound signs† | 55.5 (47.3–63.5) | 94.7 (90.5–97.4) | 89.6 (81.7–94.9) | 72.3 (66.3–77.8) | 77.1 (72.3–81.4) | 52.0 (43.0–61.0) | | Other ultrasound signs‡ | 83.9 (77.1–89.3) | 40.0 (33.0–47.3) | 53.3 (46.8–59.7) | 75.2 (65.7–83.3) | 59.7 (54.3–64.9) | 23.0 (14.0–31.0) | | Combined direct + indirect signs§ | 61.3 (53.1–69.0) | 94.2 (89.9–97.1) | 89.6 (82.2–94.7) | 74.9 (69.0–80.3) | 79.4 (74.8–83.6) | 57.0 (49.0–66.0) | Data in parentheses are 95% CI. Includes myometrial cysts, echogenic subendometrial lines and hyperechogenic islands. Includes asymmetrical myometrial wall thickening, globular uterus, translesional vascularity, fan‐shaped shadowing, irregular endometrial–myometrial junctional zone and interrupted endometrial–myometrial junctional zone. Includes heterogeneous myometrium and the question mark sign. Includes all direct and indirect ultrasound features. κ, Cohen's kappa coefficient; NPV, negative predictive value; PPV, positive predictive value. ROC‐curve analysis showed similar discriminatory ability for the direct‐sign model (area under the ROC curve (AUC), 0.780 (95% CI, 0.7381–0.8217)) and the combined model (AUC, 0.788 (95% CI, 0.7468–0.8298)). The optimal probability thresholds identified using the Youden index were essentially identical to the prespecified threshold of 0.5 for both models (0.499 and 0.504), indicating that ROC‐curve analysis supported the use of the conventional classification threshold. A regression coefficient table for the multivariable model including direct and indirect ultrasound features is provided in Table S2. In the adjusted model, which included only ultrasound features as predictors, myometrial cysts and echogenic subendometrial lines remained independently associated with adenomyosis, with adjusted ORs of 12.04 (95% CI, 2.96–72.86; P = 0.002) and 5.17 (95% CI, 1.78–16.27; P = 0.003), respectively. Indirect ultrasound features did not demonstrate significant independent associations after adjustment. Among women with histopathologically confirmed adenomyosis, it was detected on preoperative TVS in 95/155 (61.3%) cases, while 60/155 (38.7%) cases showed no ultrasound features suggestive of adenomyosis (Table S3). Women with ultrasound‐detected adenomyosis were younger (median age, 46 (IQR, 43–49) years vs 52 (IQR, 45–64) years; P < 0.01), more frequently premenopausal (86/95 (90.5%) vs 30/60 (50.0%); P < 0.01) and were more likely to report abnormal uterine bleeding (81/86 (94.2%) vs 24/30 (80.0%); P = 0.04) and dysmenorrhea (72/86 (83.7%) vs 13/30 (43.3%); P < 0.001). Uterine volume in patients without fibroids was significantly greater in women with ultrasound‐detected adenomyoisis compared to those without (median, 93.0 (IQR, 64.2–130.0) cm3 vs 63.0 (IQR, 24.8–81.8) cm3; P < 0.01). Heterogeneous myometrium was present in 95/95 (100%) of ultrasound‐detected cases and in 35/60 (58.3%) of cases that were not detected on ultrasound. Representative examples of false‐negative and false‐positive cases are shown in Figure S1.

In this prospective diagnostic accuracy study, we evaluated the performance of TVS for the diagnosis of adenomyosis using updated, consensus‐based ultrasound criteria, with histopathology as the reference standard. Our findings confirm that TVS is a highly specific but moderately sensitive tool for the detection of adenomyosis, particularly when individual ultrasound features are considered in isolation. Importantly, our findings suggest that combining direct ultrasound features into a single model does not substantially improve the diagnostic accuracy compared with individual signs, and the addition of indirect features into the combined model does not provide a meaningful incremental benefit. TVS is currently recommended as the first‐line imaging modality for the diagnosis of adenomyosis because of its wide availability, patient acceptability and relatively low cost. Notably, recent meta‐analyses of Tellum et al. 16 and Alcázar et al. 17 have shown that TVS and magnetic resonance imaging have comparable diagnostic performance for adenomyosis, supporting the role of TVS as the initial imaging approach in clinical practice. Moreover, according to the meta‐analysis of Andres et al. 18 , which included five studies and 568 patients, the pooled sensitivity and specificity of two‐dimensional TVS using combined imaging features were 83.8% and 63.9%, respectively. In that analysis, heterogeneous myometrium showed the highest sensitivity, whereas globular uterus demonstrated the highest specificity. Our findings are consistent with these results: heterogeneous myometrium exhibited high sensitivity but low specificity, supporting its role as a screening feature, while globular uterus showed very high specificity but limited sensitivity, reinforcing its limited standalone diagnostic value. Variability in diagnostic performance has also been reported in individual studies. Maudot et al. 19 , in a retrospective cohort of 242 women with histopathologically confirmed adenomyosis, reported moderate sensitivity and high specificity, highlighting the limitations of relying on isolated ultrasound signs. Similarly, Krentel et al. 20 reported that combining multiple ultrasound signs improved diagnostic accuracy compared with individual features. In line with these observations, we found that individual ultrasound signs were generally highly specific but insufficiently sensitive. Combining multiple features resulted in only modest improvements in diagnostic performance, particularly for direct signs, which we regard as the primary sonographic features in the diagnosis of adenomyosis. To date, there are no published prospective studies specifically evaluating the diagnostic performance of TVS using the direct and indirect features proposed in the 2022 revised MUSA consensus 13 . In a retrospective validation study including 26 women undergoing hysterectomy, Maghsoudlou et al. 21 reported very high sensitivity but low specificity for composite direct signs, and moderate sensitivity and specificity for composite indirect signs. Although these findings support the diagnostic relevance of direct features, the retrospective design and limited sample size restrict their generalizability. Our prospective design and larger cohort provide complementary evidence supporting the role of direct ultrasound features as the main contributors to diagnostic performance. In our multivariable analysis, models incorporating direct ultrasound features achieved a more balanced diagnostic profile, with improved accuracy and agreement with histopathology compared with models incorporating indirect or ‘other’ ultrasound features only. The lack of substantial improvement after adding indirect features into the combined direct‐ultrasound‐features model may be explained by their reactive nature and very low sensitivity. Although indirect signs are highly specific, they are infrequently present and may also be observed in other uterine conditions, limiting their incremental value in predictive models. Several factors influence the sonographic detection of adenomyosis and may partly explain the moderate sensitivity observed. Age and menopausal status are particularly relevant: adenomyosis is more frequently identified in premenopausal women. In our cohort, women with histopathologically confirmed adenomyosis that was not detected on ultrasound were older and more often postmenopausal than those with ultrasound‐detected adenomyosis (Table S3), supporting this observation. To further explore this, we performed a subgroup analysis restricted to premenopausal women, in whom adenomyosis is more clinically relevant. In this subgroup, diagnostic performance showed a higher sensitivity with a slight reduction in specificity, while maintaining a similar overall profile. These findings suggest that ultrasound features may be more readily identifiable in premenopausal women, and support the applicability of our results in this population. False‐positive and false‐negative cases highlight the limitations of ultrasound diagnosis. False negatives may occur in subtle or diffuse disease, particularly in postmenopausal women, in whom myometrial atrophy may reduce feature visibility. The lower prevalence of symptoms among women with ultrasound‐undetected adenomyosis may also contribute to underdiagnosis, as the absence of typical clinical features may reduce the level of suspicion and the likelihood of identifying subtle ultrasound findings. False positives may arise from conditions that mimic adenomyosis, such as uterine fibroids (especially with degenerative changes), non‐specific myometrial heterogeneity or focal calcifications. In addition, transient myometrial contractions may cause temporary asymmetrical myometrial wall thickening, leading to misinterpretation of indirect ultrasound features and an increased risk of overdiagnosis. Uterine fibroids represent a major confounding factor. In our cohort, fibroids were present in 56.2% of patients, a prevalence comparable with that reported in previous studies 20 , 22 , emphasizing their potential impact on ultrasound interpretation. Uterine volume is not assessed routinely as a diagnostic parameter for adenomyosis, probably owing to multiple confounders, particularly fibroids. By restricting the analysis to women without fibroids, we observed consistently higher uterine volumes in those with adenomyosis, with values above approximately 80 cm3 more frequently observed in affected patients. This supports uterine volume as a potential complementary marker in selected contexts, although it should be interpreted cautiously and not considered diagnostic in isolation. The main strengths of this study lie in its prospective design, the systematic application of updated consensus‐based criteria and the use of histopathology as the reference standard. TVS examinations were performed by experienced gynecological sonographers, ensuring standardized acquisition and interpretation. An important limitation of the study is that the population was restricted to women undergoing hysterectomy, which may have introduced spectrum bias. This cohort probably represents patients with more severe or symptomatic disease, particularly those with abnormal uterine bleeding or pelvic pain, and may overrepresent diffuse or advanced forms of adenomyosis. In addition, hysterectomy is more frequently performed in older women, which may limit the applicability of our findings to younger or reproductive‐age populations, in whom adenomyosis may present with different phenotypes, including infertility‐related manifestations. Therefore, caution is warranted when extrapolating these results to less symptomatic patients or those managed conservatively. In addition, although examinations were performed by experienced operators, interobserver reproducibility was not formally assessed, and diagnostic performance may differ in settings with varying levels of operator expertise. In conclusion, TVS remains the first‐line imaging modality for the evaluation of suspected adenomyosis. When applied using consensus‐based criteria, it provides clinically meaningful diagnostic information. Our findings highlight the importance of distinguishing between direct and indirect ultrasound features, with direct signs offering the most consistent diagnostic value, while indirect signs should be interpreted as supportive rather than definitive findings. Supporting information ACKNOWLEDGMENT We wish to acknowledge Dr Nelson Burgos Siegmund for his invaluable support and guidance throughout the development of this and other research endeavors. DATA AVAILABILITY STATEMENT The data that support the findings of this study are available from the corresponding author upon reasonable request.

- 1. Gordts S, Grimbizis G, Campo R. Symptoms and classification of uterine adenomyosis, including the place of hysteroscopy in diagnosis. Fertil Steril. 2018;109(3):380‐388.e1. [DOI] [PubMed] [Google Scholar] - 2. Struble J, Reid S, Bedaiwy MA. Adenomyosis: a clinical review of a challenging gynecologic condition. J Minim Invasive Gynecol. 2016;23(2):164‐185. [DOI] [PubMed] [Google Scholar] - 3. Habiba M, Benagiano G. Classifying adenomyosis: progress and challenges. Int J Environ Res Public Health. 2021;18(23):12386. [DOI] [PMC free article] [PubMed] [Google Scholar] - 4. Bazot M, Daraï E. Role of transvaginal sonography and magnetic resonance imaging in the diagnosis of uterine adenomyosis. Fertil Steril. 2018;109(3):389‐397. [DOI] [PubMed] [Google Scholar] - 5. Dason ES, Maxim M, Sanders A, et al. Guideline No. 437: diagnosis and management of adenomyosis. J Obstet Gynaecol Can. 2023;45(6):417‐429.e1. [DOI] [PubMed] [Google Scholar] - 6. Janicas C, Cunha TM. Adenomyosis at a glance: an integrated review of transvaginal ultrasound and mr imaging findings. Curr Probl Diagn Radiol. 2023;52(5):412‐417. [DOI] [PubMed] [Google Scholar] - 7. Moawad G, Kheil MH, Ayoubi JM, Klebanoff JS, Rahman S, Sharara FI. Adenomyosis and infertility. J Assist Reprod Genet. 2022;39(5):1027‐1031. [DOI] [PMC free article] [PubMed] [Google Scholar] - 8. Vercellini P, Viganò P, Bandini V, Buggio L, Berlanda N, Somigliana E. Association of endometriosis and adenomyosis with pregnancy and infertility. Fertil Steril. 2023;119(5):727‐740. [DOI] [PubMed] [Google Scholar] - 9. Wang L, Zheng H, Weng C, Su Q, Yang F, Wang F. Ultrasound‐defined adenomyosis subtypes and their impact on outcomes following frozen embryo transfer: a propensity score‐matched cohort study. Ultrasound Obstet Gynecol. 2025;66(6):786‐794. [DOI] [PubMed] [Google Scholar] - 10. Nirgianakis K, Kalaitzopoulos DR, Schwartz ASK, et al. Fertility, pregnancy and neonatal outcomes of patients with adenomyosis: a systematic review and meta‐analysis. Reprod Biomed Online. 2021;42(1):185‐206. [DOI] [PubMed] [Google Scholar] - 11. Van den Bosch T, Dueholm M, Leone FP, et al. Terms, definitions and measurements to describe sonographic features of myometrium and uterine masses: a consensus opinion from the Morphological Uterus Sonographic Assessment (MUSA) group. Ultrasound Obstet Gynecol. 2015;46(3):284‐298. [DOI] [PubMed] [Google Scholar] - 12. Van den Bosch T, de Bruijn AM, de Leeuw RA, et al. Sonographic classification and reporting system for diagnosing adenomyosis. Ultrasound Obstet Gynecol. 2019;53(5):576‐582. [DOI] [PubMed] [Google Scholar] - 13. Harmsen MJ, Van den Bosch T, de Leeuw RA, et al. Consensus on revised definitions of Morphological Uterus Sonographic Assessment (MUSA) features of adenomyosis: results of modified Delphi procedure. Ultrasound Obstet Gynecol. 2022;60(1):118‐131. [DOI] [PMC free article] [PubMed] [Google Scholar] - 14. Di Donato N, Bertoldo V, Montanari G, Zannoni L, Caprara G, Seracchioli R. Question mark form of uterus: a simple sonographic sign associated with the presence of adenomyosis. Ultrasound Obstet Gynecol. 2015;46(1):126‐127. [DOI] [PubMed] [Google Scholar] - 15. Education and Practical Standards Committee, European Federation of Societies for Ultrasound in Medicine and Biology . Minimum training recommendations for the practice of medical ultrasound. Ultraschall Med. 2006;27(1):79‐105. [DOI] [PubMed] [Google Scholar] - 16. Tellum T, Nygaard S, Lieng M. Noninvasive diagnosis of adenomyosis: a structured review and meta‐analysis of diagnostic accuracy in imaging. J Minim Invasive Gynecol. 2020;27(2):408‐418.e3. [DOI] [PubMed] [Google Scholar] - 17. Alcázar JL, Vara J, Usandizaga C, Ajossa S, Pascual MÁ, Guerriero S. Transvaginal ultrasound versus magnetic resonance imaging for diagnosing adenomyosis: a systematic review and head‐to‐head meta‐analysis. Int J Gynaecol Obstet. 2023;161(2):397‐405. [DOI] [PubMed] [Google Scholar] - 18. Andres MP, Borrelli GM, Ribeiro J, Baracat EC, Abrão MS, Kho RM. Transvaginal ultrasound for the diagnosis of adenomyosis: systematic review and meta‐analysis. J Minim Invasive Gynecol. 2018;25(2):257‐264. [DOI] [PubMed] [Google Scholar] - 19. Maudot C, Vernet T, Debras E, Fernandez H, Capmas P. Diagnostic accuracy study of sonography in adenomyosis: a study of current practice. J Gynecol Obstet Hum Reprod. 2023;52(7):102604. [DOI] [PubMed] [Google Scholar] - 20. Krentel H, Keckstein J, Füger T, et al. Accuracy of ultrasound signs on two‐dimensional transvaginal ultrasound in prediction of adenomyosis: prospective multicenter study. Ultrasound Obstet Gynecol. 2023;62(5):739‐746. [DOI] [PubMed] [Google Scholar] - 21. Maghsoudlou P, Chatroux LR, Ajao MO, King LP, Groszmann YS. Diagnostic accuracy of ultrasound for adenomyosis‐a blinded, retrospective, single‐expert validation study of the 2022 MUSA guidelines. J Minim Invasive Gynecol. 2025;32(11):996‐1001.e1. [DOI] [PubMed] [Google Scholar] - 22. Naftalin J, Hoo W, Nunes N, Holland T, Mavrelos D, Jurkovic D. Association between ultrasound features of adenomyosis and severity of menstrual pain. Ultrasound Obstet Gynecol. 2016;47(6):779‐783. [DOI] [PubMed] [Google Scholar] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.