Prevalence and ultrasonographic characterization of newly detected findings suggestive of deep and ovarian endometriosis in postmenopausal women undergoing routine gynecological examination

Fabio Barra: Conceptualization, Methodology, Investigation, Data curation, Formal analysis, Writing—original draft, and Writing—review and editing. Umberto Perrone: Investigation, Data curation, and Formal analysis. Anjeza Xholli: Investigation, Data curation, and Writing—review and editing. Franco Alessandri: Investigation and Resources. Maria Grazia Centurioni: Investigation, Resources, and Writing—review and editing. Stefano Bogliolo: Investigation and Data curation. Alberto Izzotti: Methodology and Writing—review and editing. Claudio Gustavino: Supervision, Resources, and Writing—review and editing. Angelo Cagnacci: Methodology, Supervision, and Writing—review and editing. Simone Ferrero: Conceptualization, Methodology, Supervision, Project administration, and Writing—review and editing. All authors critically reviewed the manuscript, approved the final version, and agree to be accountable for all aspects of the work.

The study protocol was approved by the local ethics committee (N. CET—Liguria: 412/2024—DB id 14211; last vers. approval February 10th, 2025) and conducted in accordance with the Declaration of Helsinki. This study was registered in ClinicalTrials.gov ( NCT07503938 ). The study was reported in accordance with the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) statement (Document S1 ). Written informed consent was obtained from all participants. All data were anonymized before analysis to ensure patient confidentiality.

No external funding was received for this study.

The participant selection process is detailed in Figure  S1 . Overall, 492 postmenopausal participants met the inclusion criteria and were included in the final analysis. The mean age of the study population was 60.2 ± 6.1 years, and the median time since menopause was 8.0 years (interquartile range 4.3–12.9). A larger proportion of patients reported a history of dysmenorrhea ( n  = 165; 33.5%) or chronic pelvic pain ( n  = 125; 25.4%) during reproductive life (Table  1 ). Infertility history was assessable in 488 participants; four participants who had never attempted conception were classified as not assessable for this variable. Demographic characteristics of the study population. Data available for 488 participants; infertility history was not assessable in four participants. Dyspareunia was assessable in 437 participants; 55 participants who were not currently sexually active were classified as not assessable. At the time of assessment, current pelvic pain was reported by 70 participants (14.2%), deep dyspareunia by 53 (12.1%), superficial dyspareunia by 75 (17.2%), and vaginal dryness by 213 (43.3%). Regarding postmenopausal hormonal exposure, 55 participants (11.2%) were currently receiving systemic HRT and 22 (4.5%) had used it previously. One participant had discontinued HRT in the past and subsequently restarted treatment; therefore, overall systemic HRT exposure involved 76 participants (15.4%). Local vaginal estrogen or dehydroepiandrosterone exposure was reported by 56 participants (11.4%). Additional demographic and clinical characteristics of the study population are reported in Tables  1 and S1 . Hypoechoic nodular thickenings in the posterior compartment and/or adnexal cystic formations suggestive of endometriosis were identified in 40 participants (8.1%; 95% CI 5.9%–10.9%). Of these, 36 participants (7.3%; 95% CI 5.2%–10.0%) had posterior compartment nodules suggestive of DE (Figures  1 and 2 ). Their anatomical distribution of nodules included the lower or upper rectum ( n  = 15), the rectosigmoid junction or sigmoid colon ( n  = 6), the uterosacral ligaments and uterine torus ( n  = 19), and the rectovaginal septum or posterior vaginal fornix ( n  = 4); seven participants (19.4% of those with posterior nodules) harbored more than one posterior lesion. A reduced or absent posterior sliding sign was observed in 75 participants (15.2%). Among these, 20 patients (26.7%) also had at least one posterior hypoechoic nodule. (A) A hypoechoic lesion (asterisk), showing more regular margins and focal thickening of the anterior rectal wall; a maximum thickness of 6 mm was recorded, with no ultrasonographic evidence of (sub)mucosal layer involvement. (B) Concomitant left ovarian cyst with ultrasonographic features suggestive of an atypical endometrioma, characterized by isoechoic, non‐uniform internal echoes, deviating from the typical “ground glass” appearance, and without relevant intralesional vascularization on Doppler assessment. (C) Another hypoechoic lesion is visible in the posterior compartment, showing a flattened morphology and an indistinct lesion–tissue interface with a gradual hypoechoic transition into the surrounding tissues. The lesion involves the left uterosacral ligament (*) and extends into the left posterolateral parametrium, where it assumes a tapering configuration. (D) Cranial extension of the lesion reaches the lateral parametrial plane (arrows) up to the uterine vessels, which appear not infiltrated. The adjacent ovarian cyst appears fixed, supporting the presence of adhesions within the posterior compartment. From a morphological perspective, most intestinal nodules displayed a plaque‐like hypoechoic pattern (mean largest diameter 14 ± 6.8 mm). A maximum thickness <7 mm was recorded in 18 of 21 lesions (85.7%), and margins were judged regular in 17 of them (81.0%). In 16 nodules (76.2%), there was no ultrasonographic evidence of submucosal layer involvement, whereas the lesion–tissue interface appeared indistinct, with a gradual, subtly hypoechoic transition into the surrounding tissue in 14 lesions (66.7%) (Figure  1 ). One additional intestinal lesion was considered indeterminate at the ultrasonographic examination and was not classified among the 21 intestinal lesions suggestive of DE. This lesion showed pronounced vascularization on color Doppler, a maximum thickness > 10 mm, extensive involvement of the submucosal and mucosal layers, and clinically relevant rectal bleeding (Figure  3C ). Because these findings raised concern for an alternative diagnosis, further investigation was undertaken. Colonoscopy revealed a rectal adenocarcinoma, which was treated with neoadjuvant radiotherapy followed by surgical resection. (A) An ovarian cyst with ultrasonographic features suggestive of an atypical endometrioma is visible, characterized by homogeneous low‐level internal echoes and an intracystic echogenic concretion, with detectable mild vascularization on Doppler assessment. The cyst appears adherent to the left uterosacral ligament, where a hypoechoic nodule suggestive of DE is identified (*). The lesion shows a flattened morphology and indistinct, subtly hypoechoic margins, extending along the left posterior parametrium to the anterior wall of the upper rectum (arrows). The cyst had been persistently present for 5 years and was associated with a CA125 level of 40.4 U/mL, prompting surgical management; histopathology ultimately confirmed a serous borderline tumor confined to the ovary (FIGO stage I). (B) Another ovarian cyst with features suggestive of an atypical endometrioma, showing a mixed internal echotexture with a partially anechoic component (*) and non‐uniform low‐level echoes, deviating from the typical “ground glass” appearance. The cyst content appears mildly heterogeneous and layered, without internal vascularization on Doppler assessment. (C) Rectal adenocarcinoma incidentally visualized during transvaginal ultrasound as a posterior rectal wall lesion. The lesion showed features distinct from the predominantly plaque‐like nodules suggestive of DE observed in this postmenopausal cohort, including a bulky hypoechoic morphology, marked rectal wall thickening, disruption of the normal layered bowel wall architecture, and extension beyond the muscularis propria, also infiltrating the recto‐vaginal septum (*). In the presence of rectal bleeding, colonoscopic evaluation was performed and confirmed rectal adenocarcinoma. Ultrasonographic features suggestive of adenomyosis were identified in 32 participants (6.5%). The most common direct sign was the presence of hyperechoic myometrial islands ( n  = 25; 78.1%). Among indirect signs, asymmetry between the anterior and posterior uterine walls was observed in 12 participants (37.5%), while an interrupted junctional zone was documented in 10 (31.3%). Adnexal assessment revealed six ovarian cysts with ultrasonographic features suggestive of endometriomas, including typical and atypical appearances (1.2%). Two of these participants also had posterior compartment nodules suggestive of DE. One cyst exhibited the classic unilocular “ground‐glass” appearance and was managed conservatively. The remaining five cysts showed atypical features, including multilocularity, heterogeneous internal echoes, focal papillary excrescences, or localized wall thickening (Figures  3 and 4 ). Serum CA125 levels were measured in three of these cases and were within the normal range in two participants, while a modest elevation above the conventional cut‐off was documented in one patient (40.4 U/mL). In view of postmenopausal status, atypical morphology, and, in one case, CA125 elevation, three participants were referred for surgical management. Histopathological examination confirmed benign endometriosis‐related pathology in two cases, whereas one lesion initially considered suggestive of an atypical endometrioma at the index ultrasonographic examination was subsequently diagnosed as a FIGO stage I serous borderline tumor confined to the ovary (Figure  3A ). (A) Bilocular ovarian cyst with ultrasonographic features suggestive of an atypical endometrioma (37 × 32 mm), characterized by homogeneous low‐level internal echoes and the presence of an avascular intracystic hyperechogenic concretion (*). (B) Orthogonal scan of the posterior compartment showing a low‐profile, plaque‐like hypoechoic lesion (23 × 5 mm) with flattened morphology and indistinct, subtly hypoechoic margins. The lesion extends along the left uterosacral ligament (*) and tapers into the left posterior parametrium (arrows). It appears non‐bulky, with preserved layered anatomy of the adjacent bowel wall and no ultrasonographic evidence of submucosal or mucosal involvement. Other incidental gynecological findings at ultrasound included uterine fibroids ( n  = 158; 32.1%), congenital uterine malformations ( n  = 8; 1.6%), and simple ovarian cysts <3 cm ( n  = 23; 4.7%), three of which (0.6% overall) required surgical management due to atypical morphology in accordance with postmenopausal risk stratification. In the post hoc blinded second‐reader analysis, 112 archived examinations were independently reviewed, including all 40 participants with posterior nodules and/or endometrioma‐like ovarian cysts and 72 randomly selected participants originally negative for both findings. For posterior nodules, the second reader confirmed 33 of 36 originally positive examinations and classified 75 of 76 originally negative examinations as negative, corresponding to an overall agreement of 96.4% and a Cohen's κ of 0.92 (95% CI 0.84–1.00). For these ovarian cysts, five of six originally positive examinations were confirmed, and all 106 originally negative examinations were classified as negative, corresponding to an overall agreement of 99.1% and a Cohen's κ of 0.90 (95% CI 0.72–1.00) (Table  S2 ). When the population was stratified according to time since menopause, the prevalence of hypoechoic nodular thickenings progressively decreased across categories, from 13.9% in participants within 5 years of menopause to 4.5% in those 6–10 years from menopause and 4.3% in those more than 10 years beyond menopause ( p for trend = 0.001). In univariate logistic regression analysis (Table  2 ), a history of dysmenorrhea (OR 2.68, 95% CI 1.35–5.33; p  = 0.005) and chronic pelvic pain (OR 2.55, 95% CI 1.28–5.09; p  = 0.008) during reproductive life were significantly associated with the presence of posterior hypoechoic nodules. Participants within 5 years of menopause onset showed markedly increased odds of nodular findings compared with those more than 10 years beyond menopause (OR 3.61, 95% CI 1.55–8.42; p  = 0.003). A history of infertility was not significantly associated with posterior nodular findings (OR 2.14, 95% CI 0.96–4.79; p  = 0.063). Among participants for whom dyspareunia was assessable, current deep dyspareunia was not significantly associated with posterior nodules (OR 1.76, 95% CI 0.69–4.49; p  = 0.239). Current pelvic pain and dyschezia were also not significantly associated with them. Univariable logistic regression analyses of factors associated with posterior compartment hypoechoic nodules in postmenopausal participants. Note : Outcome: presence of ≥1 posterior compartment hypoechoic nodule (yes/no). Abbreviations: HRT, hormone replacement therapy; yr, years. Excluding for diagnosing/treating endometriosis. In the multivariable logistic regression model (Table  3 ), being within 5 years of menopause remained independently associated with posterior hypoechoic nodules (adjusted OR 3.25, 95% CI 1.35–7.83; p  = 0.009). A history of dysmenorrhea (adjusted OR 2.20, 95% CI 1.05–4.60; p  = 0.036) and chronic pelvic pain (adjusted OR 2.36, 95% CI 1.11–5.00; p  = 0.026) during reproductive life also retained independent significance. In contrast, a history of infertility, systemic HRT exposure, current pelvic pain, and current dyschezia were not independently associated with posterior nodular findings (all p  > 0.05). Firth penalized logistic regression yielded consistent results, with time since menopause ≤5 years, history of dysmenorrhea, and history of chronic pelvic pain remaining significantly associated with posterior nodules (Table  S3 ). Multivariable logistic regression analyses of factors associated with posterior compartment hypoechoic nodules in postmenopausal participants. Abbreviations: HRT, hormone replacement therapy; yr, years. No clear associations were observed in exploratory univariable analyses of ovarian cysts suggestive of endometriomas or adenomyosis features. Given the limited number of outcome events, particularly for these ovarian cysts, adjusted analyses were not considered sufficiently reliable for interpretation.

The present study provides the first systematic characterization of ultrasonographic findings suggestive of endometriosis in a large cohort of postmenopausal participants undergoing routine gynecological evaluation. Ultrasound findings suggestive of endometriosis were identified in 8.1% of participants, including posterior compartment nodules suggestive of DE in 7.3%. This prevalence is higher than historical estimates of postmenopausal endometriosis, which were derived primarily from isolated case reports, small surgical series, and incidental autopsy findings. 11 , 12 , 23 Previous gynecological surveillance and ultrasound examinations may not have systematically assessed endometriosis, potentially leaving subtle posterior compartment lesions unrecognized. From a morphological perspective, the nodules detected in our cohort showed several features that appeared to differ from the classic ultrasonographic appearance of DE described in reproductive‐age populations. Rather than presenting predominantly as solid hypoechoic nodules with spiculated or star‐shaped contours and acoustic shadowing, 6 , 24 the lesions observed in our postmenopausal cohort more commonly exhibited a flattened, plaque‐like configuration, limited Doppler vascularity, and relatively regular contours, often accompanied by an indistinct lesion–tissue interface. This combination of ultrasonographic features may be compatible with a greater fibrotic component or tissue remodeling; however, ultrasound morphology and Doppler vascularity alone cannot establish the histological composition or biological activity of these lesions. This interpretation therefore remains hypothesis‐generating, although it is consistent with previous histopathological observations suggesting increased fibrotic changes in postmenopausal endometriosis. 25 With regard to anatomical distribution, DE in reproductive‐age populations most frequently involves the uterosacral ligaments (52.7%–69.2%), rectovaginal septum and vagina (12.0%–16.2%), bowel (9.9%–22.7%; rectosigmoid/rectum), and, less frequently, the bladder (6.3%–6.4%). 5 , 26 A similarly predominant involvement of posterior pelvic structures was observed in our cohort, in which all detected nodules were located within the posterior compartment. Our multivariable analysis (Table  3 ) offers additional insights. First, we observed an association between posterior hypoechoic nodules and time elapsed since menopause. The prevalence of nodular findings was highest within the first 5 years after menopause and declined sharply thereafter, with no further substantial decrease beyond this early postmenopausal window (approximately 14% vs. 4%–5%). In line with this descriptive pattern, participants within 5 years of menopause had approximately threefold higher adjusted odds of nodular findings than those more than 10 years beyond menopause. This cross‐sectional pattern indicates that the frequency of detectable posterior nodular findings differs according to time since menopause but should not be interpreted as evidence of within‐person lesion regression, remodeling, thinning, or reduced detectability over time. Because different participants were examined at each postmenopausal interval, the observed association may also reflect differences in chronological age, reproductive‐age disease phenotype, previous treatments or surgery, hormonal exposures, or other unmeasured cohort characteristics. Progressive tissue remodeling or reduced ultrasonographic conspicuity may represent possible biological explanations, but these remain speculative and cannot be demonstrated by the present study. Accordingly, the relative contributions of time since menopause and chronological age cannot be disentangled from these cross‐sectional data. The dissociation between historical and current symptomatology represents another salient finding. A history of dysmenorrhea and chronic pelvic pain during reproductive life was independently associated with postmenopausal nodular findings, whereas current pelvic pain and dyschezia were not independently associated with posterior nodules. Current deep dyspareunia was likewise not associated with posterior nodules in univariable analysis restricted to participants for whom dyspareunia was assessable. Taken together, these findings suggest that posterior nodular abnormalities may frequently be detected in the absence of concurrent pain symptoms, without implying that they represent a biologically distinct lesion phenotype. Nevertheless, in postmenopausal women, pelvic and genital symptoms are frequently influenced by age‐related and hypoestrogenic changes—including vaginal atrophy, urinary dysfunction, and pelvic floor disorders—which may obscure or outweigh any symptom contribution from residual endometriotic lesions. 27 , 28 Importantly, this interpretation remains inferential: histological confirmation was not routinely obtained because it was not clinically indicated in most cases, and our study therefore provides an initial descriptive characterization of postmenopausal ultrasound findings suggestive of DE rather than histologically confirmed disease. Consequently, some degree of diagnostic misclassification cannot be excluded. The lack of an independent association between nodular findings and HRT in our cohort warrants careful interpretation. Symptom exacerbation or lesion growth has been described in postmenopausal women receiving HRT—particularly unopposed estrogen—although absolute risks of recurrence and malignant transformation are difficult to quantify, and the impact of HRT on these outcomes remains uncertain. 28 Although no statistically significant association was detected between systemic HRT exposure and posterior nodules, the effect estimate was imprecise and does not exclude a clinically meaningful association. The limited number of exposed participants and the cross‐sectional design preclude definitive conclusions regarding subtle effects on lesion persistence or growth. Furthermore, combining current and past exposure and different HRT regimens may have attenuated regimen‐specific associations. Moreover, the primary endpoint of this study was structural rather than a dynamic assessment of hormonally driven activity. As such, our findings should not be interpreted as evidence of hormonal neutrality, but rather as an absence of detectable structural differences within the limits of the available data. A single case in our cohort underscores the risk of misclassification in postmenopausal individuals, particularly when hypoechoic nodules are detected in “red‐flag” sites such as the rectum or bladder, where malignancies may mimic endometriotic morphology. 29 , 30 In our series, a rectal nodule detected at TVS was ultimately diagnosed as rectal adenocarcinoma (with no evidence of endometriosis), highlighting the need for caution when interpreting posterior compartment lesions in this age group. Although TVS is not a first‐line tool for colorectal cancer, it may incidentally identify rectal or rectosigmoid masses during pelvic assessment and, in selected cases, prompt further diagnostic work‐up. 31 Crucially, alarm symptoms—including rectal bleeding, iron‐deficiency anemia, or new‐onset bowel habit changes—mandate dedicated colorectal evaluation and should not be attributed to presumed endometriosis. In this context, ultrasound should be framed as an opportunistic trigger rather than a substitute for organized screening pathways (FOBT/FIT and colonoscopy). A similar caveat can be raised theoretically for the urinary bladder: focal bladder wall thickening may mimic anterior compartment endometriosis on ultrasound yet represent urothelial malignancy, particularly when accompanied by macroscopic hematuria. 32 Given the relative rarity of bladder involvement in this setting, we did not observe this scenario in our cohort; however, its potential clinical implications in postmenopausal women support heightened vigilance and a low threshold for urological assessment when suspicious imaging features or relevant symptoms are present. Adnexal findings in our cohort provide a complementary perspective. Five of the six ovarian cysts initially considered suggestive of endometriomas exhibited atypical ultrasonographic features. Previous studies have described age‐related variations in endometrioma morphology and highlighted the reduced specificity of the classic ground‐glass appearance after menopause. 15 , 16 In our series, one surgically treated adnexal lesion was diagnosed as a FIGO stage I serous borderline ovarian tumor (Figure  3A ), representing a sonographic diagnostic mimic of an atypical endometrioma rather than evidence of endometriosis‐associated malignant transformation. This finding underscores the importance of careful evaluation of atypical ovarian cysts after menopause. Given that endometriosis is a recognized precursor background for endometrioid and clear‐cell ovarian carcinoma, persistent or atypical adnexal lesions after menopause warrant careful evaluation. 33 Serum tumor markers (CA125, HE4) and multiparametric risk models (e.g., IOTA ADNEX) may support risk stratification, although their performance has not been explicitly validated in asymptomatic postmenopausal individuals with newly detected small lesions resembling endometriomas with atypical morphology. Accordingly, meticulous morphological assessment, contextual interpretation of biomarkers, and short‐interval follow‐up or surgical management remain essential. 34 , 35 This study has several strengths, including its prospective design, relatively large sample size, and the use of a standardized, endometriosis‐oriented TVS protocol applied by an expert operator with systematic recording of lesion morphology and soft markers, supported by detailed clinical phenotyping and multivariable analysis. However, some limitations must be acknowledged. Participants with a previous surgical or radiological diagnosis of endometriosis were excluded, which reduces ascertainment bias and enables a focused evaluation of incidentally detected, previously unrecognized postmenopausal lesions, but limits generalizability and may underestimate the overall prevalence of postmenopausal endometriosis sequelae. Accordingly, our prevalence estimate should be interpreted as the prevalence of newly detected ultrasound findings suggestive of DE and/or ovarian endometriomas among participants without a previous diagnosis of endometriosis, rather than as the overall prevalence of postmenopausal endometriosis. Histologic confirmation was not obtained because surgery was rarely clinically indicated; therefore, lesion characterization remains presumptive and based on imaging findings and clinical context. Because histological verification was available only in selected clinically indicated cases, some degree of diagnostic misclassification cannot be excluded, particularly after menopause, when residual endometriosis, fibrotic lesions of different etiologies, and occasionally malignant disease may share overlapping ultrasonographic features. Nevertheless, all examinations were performed according to standardized IDEA criteria, 6 and lesions displaying atypical morphology underwent appropriate clinical investigation and, when indicated, histological assessment. The additional morphological characterization was performed post hoc on archived ultrasound images, rather than using prospectively predefined descriptors, and should therefore be considered exploratory. Although all examinations were acquired by a single expert operator, the blinded review of archived images demonstrated high interobserver agreement for the identification of posterior nodules and ovarian cysts suggestive of endometriomas. However, this assessment did not involve independent image acquisition, limiting generalizability across operators and settings. Furthermore, the original operator was aware of participants' clinical histories, potentially introducing observer bias. The cross‐sectional design did not permit evaluation of lesion evolution over time or causal inferences regarding symptom trajectories and hormonal exposure. The frequency and characteristics of previous ultrasound examinations were not systematically recorded, precluding assessment of whether individual lesions had been previously overlooked or had become detectable only later. Reproductive‐age symptoms, including pain history, were collected retrospectively and may therefore be subject to long‐term recall bias. Finally, because participants were recruited from regional outpatient gynecology clinics, selection bias cannot be excluded, and extrapolation of these findings to the general postmenopausal population should be made with caution. Future multicenter studies incorporating multiple operators, independent real‐time image acquisition, external blinded review, and longitudinal follow‐up are needed to assess the reproducibility and evolution of these findings across different settings.

This study provides the first systematic ultrasound characterization of findings suggestive of DE and ovarian cysts suggestive of endometriomas in postmenopausal women without a previous diagnosis of endometriosis. Posterior nodules were frequently plaque‐like and poorly vascularized and were associated with a history of reproductive‐age pain, but not independently with current pelvic pain or dyschezia. Their histological nature and biological activity cannot be determined from ultrasound alone. Atypical imaging features or alarm symptoms require careful evaluation to exclude alternative diagnoses, including malignancy.

Endometriosis is a chronic, estrogen‐dependent inflammatory disease characterized by the presence of endometrium‐like tissue outside the uterus, affecting an estimated 6%–10% of reproductive‐age women. 1 , 2 Although its pathogenesis is multifactorial—encompassing retrograde menstruation, hormonal dysregulation, immune dysfunction, and genetic susceptibility—estrogen is the principal driver of ectopic endometrial proliferation and inflammatory activity. 3 During reproductive years, systemic ovarian estrogen sustains lesion activity and symptoms such as pelvic pain, dysmenorrhea, dyspareunia, and infertility. 4 Endometriosis most commonly affects pelvic structures but may also occur at extrapelvic sites. 5 Transvaginal ultrasound (TVS) has become the first‐line modality for detection and mapping of ovarian and deep endometriosis (DE), and the standardized IDEA (International Deep Endometriosis Analysis Group) approach has demonstrated high diagnostic performance in premenopausal patients with suspected disease. 6 , 7 Ultrasonographic assessment of superficial endometriosis is also evolving, although its detection remains less established than that of ovarian endometriosis and DE, and dedicated standardized criteria have only recently been proposed and require further validation. 8 Because of its hormonal dependence, endometriosis has long been regarded as a condition confined to the reproductive years. Menopause, characterized by ovarian senescence and declining systemic estrogen levels, is typically associated with symptom remission, and many previously affected individuals experience a substantial reduction in pain and discontinuation of hormonal treatment. 3 However, this view may be overly simplistic: morphological alterations driven by chronic inflammation and fibrosis are unlikely to regress immediately in parallel with hormonal withdrawal and may persist for years as structural residues of the disease, potentially causing pelvic pain even in the absence of a cyclical menstrual pattern. In addition, localized estrogen biosynthesis within endometriotic implants, extragonadal estrogen production in adipose tissue, and exogenous hormonal exposure through hormone replacement therapy (HRT) have been proposed as mechanisms that could theoretically sustain some degree of postmenopausal disease activity and, albeit rarely, contribute to malignant transformation, although their true clinical impact remains uncertain and is still a matter of debate. 9 In 2022, the ESHRE guideline introduced a dedicated chapter on postmenopausal endometriosis, while simultaneously acknowledging that available evidence is scarce, of low quality, and insufficient to define the natural history of the disease after menopause. 10 Most of what is known derives from case reports, small surgical series, or incidental autopsy findings, and population‐based estimates of a 2%–5% prevalence likely underestimate the true burden, as many lesions remain clinically silent. 11 , 12 , 13 Furthermore, women currently in their sixth or seventh decade lived their reproductive years in an era of limited awareness and lower diagnostic accuracy—particularly before widespread access to high‐resolution TVS—raising the possibility that a substantial proportion of postmenopausal findings may represent long‐standing, previously unrecognized sequelae of endometriosis. Despite this, the ultrasonographic appearance of endometriosis in postmenopausal women has never been systematically described, even by operators with extensive expertise in endometriosis imaging. Consequently, we lack robust data on the prevalence, anatomical distribution, and morphological characteristics of ultrasound findings suggestive of endometriosis in this population. The aim of this study was therefore to describe, in a large cohort of postmenopausal participants undergoing routine gynecological examination, the prevalence and ultrasonographic aspects of newly detected findings suggestive of DE and ovarian endometriosis, and to explore their clinical correlates.

The authors declare no conflicts of interest.

This was a prospective, observational study that included participants assigned female sex at birth who had undergone spontaneous menopause and consecutively attended two regional gynecology outpatient clinics for routine gynecological examinations over a 24‐month period (ASL‐4 Liguria, Metropolitan Area of Genoa, Italy and AOM—IRCCS Ospedale Policlinico San Martino, Genoa, Italy) between June 2023 and June 2025. The recruitment period was established before study initiation to ensure inclusion of a sufficiently large consecutive cohort for prevalence estimation. Menopausal status was defined by at least 12 consecutive months of spontaneous amenorrhea not attributable to pregnancy, lactation, medical therapy, or other pathological causes, or, when clinically required, by biochemical evidence consistent with menopause based on a serum Follicle‐Stimulating Hormone (FSH) level > 25 IU/L. For the purposes of this study, “routine gynecological examination” referred to a scheduled outpatient gynecological assessment performed irrespective of the presence or absence of current gynecological symptoms; therefore, the study population was not restricted to asymptomatic participants undergoing screening. Exclusion criteria were a previous radiological (including ultrasonographic) and/or surgical diagnosis of endometriosis, a history of hysterectomy and/or bilateral oophorectomy (any route), or active pelvic malignancy. Participants were also excluded from the analysis when the examination was not feasible or not interpretable for predefined technical reasons, including inability to tolerate probe insertion (e.g., severe introital stenosis, vaginismus, or inability to undergo TVS in participants who had never been sexually active). Current or prior use of HRT was recorded but was not considered an exclusion criterion. At the gynecological visit, each participant completed a structured questionnaire specifically designed for this study. The questionnaire collected detailed demographic, reproductive, and clinical information, including age, body mass index (BMI), parity, smoking status, medical comorbidities, and time since menopause. Gynecological history was recorded with particular attention to pain‐related symptoms during reproductive life, including dysmenorrhea, chronic pelvic pain, dyspareunia, and dyschezia, as well as prior pelvic or abdominal surgery. A history of dysmenorrhea and chronic pelvic pain was assessed by asking participants whether these symptoms had occurred at any time during reproductive life. No minimum retrospective threshold for symptom intensity, duration, or frequency was applied. Previous infertility was recorded when participants reported ≥12 months of unsuccessful conception attempts and/or a previous clinical diagnosis of infertility and/or the use of medically assisted reproduction, at any time during reproductive life. Participants who reported never having attempted conception were classified as not assessable for infertility and were excluded from analyses involving this variable. Information on menopausal status and hormonal exposure was collected in detail, including current or previous use of systemic HRT and local vaginal hormonal therapy. Current symptoms were assessed at the time of the visit using an 11‐point numeric rating scale (0 = no pain, 10 = worst imaginable pain) for pelvic pain, deep and superficial dyspareunia, and dyschezia. Clinically relevant pain was prespecified as a score ≥ 3. Participants who were not currently sexually active were classified as not assessable for dyspareunia and were excluded from analyses involving dyspareunia. Participants were also asked about gastrointestinal symptoms, including changes in bowel habits, abdominal bloating, and unexplained rectal bleeding. Lower urinary tract symptoms were systematically recorded as well, including urinary frequency, urgency, dysuria, and nocturia, as well as any history of recurrent urinary tract infections or macroscopic hematuria. For each participant, questionnaire responses and relevant clinical information retrieved from medical records (including comorbidities, medications, and previous imaging or surgical reports) were coded using predefined categories and entered into a dedicated electronic database. Data entry procedures incorporated internal consistency checks to minimize transcription errors and ensure uniformity across variables. All participants underwent a standardized TVS examination using either a Resona IV ultrasound system (Mindray, Shenzhen, China) or a Voluson S8 ultrasound system (GE HealthCare, Chicago, IL, USA), equipped with transvaginal transducers operating within a 5–9 MHz frequency range. All examinations were conducted by a single operator with extensive expertise in TVS and endometriosis imaging, ensuring consistency in image acquisition and lesion characterization. Color Doppler assessment was performed using a pulse repetition frequency (PRF) of 0.3–0.6 kHz, a low wall‐filter setting, and gain adjusted to the highest level that did not generate background color noise to optimize sensitivity for low‐velocity flow. Doppler vascularity was assessed qualitatively, and no quantitative vascularity threshold was applied. Ultrasound examinations were performed in the context of routine clinical consultations; therefore, the operator was not blinded to the participants' clinical history, symptoms, or questionnaire responses. In the postmenopausal period, the baseline scan included systematic evaluation of uterine size and morphology, endometrial thickness and echogenicity, myometrial texture, adnexal structures and ovarian remnants (when visualized), and exclusion of pelvic masses or free fluid. The bladder, ureters, and rectosigmoid region were also routinely inspected to exclude non‐gynecologic pelvic pathology. The examination incorporated the first three steps of the IDEA consensus for the identification of indirect signs of DE. 6 First, the uterus and adnexa were evaluated for ultrasonographic direct and indirect features of adenomyosis—recorded according to the revised MUSA criteria 14 —as well as for ovarian cysts with ultrasonographic features suggestive of endometriomas, including typical and atypical appearances. Typical features included unilocular morphology and homogeneous ground‐glass echogenicity. Atypical presentations were considered when cysts exhibited internal echoes suggestive of endometriotic content, even if heterogeneous, together with non‐classical morphological features such as multilocularity, solid components, or papillary projections, as previously described in older women. 15 , 16 This classification was established at the index ultrasound examination and reflected an initial sonographic suspicion rather than a definitive diagnosis. Second, soft markers of pelvic adhesions and possible DE were assessed. Ovarian mobility and tenderness were evaluated by exerting gentle pressure between the ovary and uterus: reduced or absent gliding suggested adhesions, including the presence of “kissing ovaries.” The distal third of both ureters was identified and scrutinized for dilatation, as previously described, 17 complementing this evaluation by abdominal ultrasonography to assess for hydronephrosis through inspection of calyceal dilatation, renal pelvis configuration, and cortical thickness. Third, the status of the pouch of Douglas was investigated using the posterior sliding sign. A negative sliding sign was defined as the absence of physiological movement between the rectum and the posterior uterine wall or cervix during probe‐induced mobilization, suggesting potential adhesion or obliteration of the pouch of Douglas related to posterior compartment DE. 18 Similarly, the anterior sliding sign was assessed to identify reduced mobility between the bladder wall and the anterior uterine surface, indicative of adhesions at the vesicouterine fold. 19 Following this preliminary assessment, a systematic evaluation of pelvic compartments (IDEA step 4) was carried out. DE was identified as endometriotic lesions extending on or beneath the peritoneal surface, without applying a minimum depth threshold, as recently defined. 2 The pelvic posterior compartment, including the rectosigmoid, uterosacral ligaments, rectovaginal septum, and posterior vaginal fornix, was carefully examined for hypoechoic nodules or focal thickening. Normal rectal wall layers can be visualized on TVS: the anterior rectal serosa is seen as a thin hyperechoic line; the muscularis propria is hypoechoic, with the longitudinal smooth muscle and circular smooth muscle separated by a faint, thin hyperechoic line; the submucosa is hyperechogenic; and the mucosa is hypoechoic (Figure  1B ). Bowel DE usually appears on TVS as a thickening of the hypoechoic muscularis propria or as hypoechoic nodules, with or without hyperechoic foci. Rectovaginal lesions were identified as nodules located below the peritoneal reflection along the lower margin of the posterior cervix, whereas vaginal involvement manifested as thickening or distortion of the posterior fornix. The uterosacral ligaments, representing the most accessible component of the posterior parametrium on TVS, were inspected in transverse and longitudinal planes for hypoechoic thickening or focal nodularity. The other parametrial regions were evaluated according to recent standardized evidence. 20 , 21 Superficial endometriosis was not investigated, as the study protocol was designed specifically for DE and ovarian endometriosis and did not prospectively apply the recently proposed IDEA criteria for ultrasonographic assessment of superficial disease. 8 (A) Ultrasonographic sagittal view of the uterus showing features consistent with menopausal involution, including a thin endometrium and mild asymmetry between the anterior and posterior uterine walls. (B) A plaque‐like hypoechoic rectal lesion (*) characterized by a flattened morphology, a relatively regular contour, and an indistinct lesion–tissue interface with a gradual hypoechoic transition into the surrounding tissues. The dashed line outlines the ultrasonographically visible extent of the lesion. The adjacent rectal wall layers remain recognizable, with no evidence of submucosal or mucosal involvement. (C) A rectal nodule (*), detected in another participant, showing a similar plaque‐like hypoechoic configuration and relatively regular contour. (D) Further view of the lesion shown in panel C, demonstrating tapering extension toward the left posterior parametrium and involvement of the left uterosacral ligament (arrows). C, cervix; PP, posterior parametrium; PVF, posterior vaginal fornix; RL, rectal lumen; RML, rectal muscular layer; RMU, rectal mucosal layer; RSM, rectal submucosal layer; RVS, rectovaginal septum; RW, rectal wall; T, torus uterinus; USL, uterosacral ligament; VM, vaginal muscular layer. Lesions suggestive of DE were first identified and anatomically classified according to the standardized IDEA criteria, independently of the additional morphological characteristics described below. Once a lesion was considered suggestive of DE, its ultrasonographic appearance was further characterized to explore potentially distinctive features in this postmenopausal population. These additional morphological characteristics were not prospectively predefined before recruitment but were assessed post hoc through retrospective review of the archived ultrasound images of examinations showing findings suggestive of DE. For each nodule, we assessed maximum thickness, margin regularity, internal echotexture, color Doppler vascularity, relationship with adjacent structures, and, in the case of intestinal lesions, the presence or absence of apparent involvement of the individual bowel‐wall layers, as already proposed. 22 These additional characteristics were used for descriptive purposes and were not diagnostic criteria for DE. A plaque‐like configuration referred to a flattened, broad‐based hypoechoic thickening extending along the involved structure rather than forming a rounded or bulky nodule. The contour of the lesion was described as regular when it appeared smooth and non‐spiculated. An indistinct lesion–tissue interface was defined as a gradual, poorly demarcated transition between the lesion and surrounding tissues; these two features were therefore not mutually exclusive. Lesions whose ultrasonographic appearance was considered indeterminate because an alternative diagnosis could not be reasonably excluded were not classified as DE for analytic purposes and underwent additional diagnostic assessment as clinically indicated. Lesions extending across more than one anatomical structure were classified as separate sites to maintain consistency in anatomical mapping. All examinations and representative images were digitally archived for subsequent review and quality control. As an additional post hoc agreement analysis performed in response to peer review, all the archived examinations showing findings suggestive of DE and ovarian endometriosis and a randomly selected sample of negative examinations were independently reviewed by a second sonographer with expertise in endometriosis imaging. Examinations were anonymized and presented in random order. Therefore, the second reader was blinded to participants' clinical characteristics, including the original ultrasound classification. The reader independently assessed the presence or absence of posterior nodules suggestive of DE and ovarian cysts suggestive of endometriomas and, when present, their anatomical location and principal morphological characteristics. The study was conceived primarily as a descriptive prevalence study. Available literature suggested a prevalence of postmenopausal endometriosis of approximately 2%–5%, 13 although robust estimates derived from systematic ultrasound screening studies were lacking. Therefore, sample size considerations were based on prevalence estimation rather than hypothesis testing. Using the upper limit of prevalence estimates (5%), a confidence level of 95%, and a desired precision of ±2%, the minimum required sample size was estimated to be approximately 457 participants. The prevalence of ultrasound findings suggestive of endometriosis was calculated as the proportion of participants presenting with at least one finding suggestive of DE and/or an endometrioma‐like ovarian cyst at the initial ultrasonographic examination, with exact 95% confidence intervals (Clopper–Pearson method). Lesions considered indeterminate at the initial (index) examination because an alternative diagnosis could not be reasonably excluded were not classified as DE for analytic purposes and underwent further diagnostic assessment as clinically indicated. Regression analyses used the presence of posterior compartment hypoechoic nodules suggestive of DE as the outcome. Because the primary objective of the study was prevalence estimation, regression analyses were considered exploratory. Associations between clinical variables and posterior compartment hypoechoic nodules were first assessed using univariable logistic regression. The multivariable model included time since menopause (≤5 years, 6–10 years, and >10 years as the reference category), history of dysmenorrhea, chronic pelvic pain and infertility, systemic HRT exposure, current pelvic pain, and current dyschezia. Historical variables were coded as present/absent, HRT as any past or current systemic exposure versus none, and current pain symptoms according to the prespecified numeric rating scale ≥3 threshold. Covariates were selected on clinical grounds and entered simultaneously into the model, without automated or univariable p‐value‐based selection. Given the limited number of outcome events, adjusted estimates were considered exploratory, and Firth penalized logistic regression using the same covariates was performed as a sensitivity analysis. Analyses involving infertility were restricted to participants with assessable infertility history, and analyses involving dyspareunia to currently sexually active participants; no imputation was performed. Interobserver agreement for the presence of posterior nodules and endometrioma‐like cysts was assessed using overall percentage agreement and Cohen's κ with 95% confidence intervals. Results are reported as odds ratios (ORs) with 95% confidence intervals (CIs), and a two‐sided p ‐value < 0.05 was considered statistically significant. Multicollinearity was assessed using the variance inflation factor, with values <2 considered acceptable. All analyses were performed using SPSS Statistics v25.0 (IBM Corp., Armonk, NY, USA).

Document S1 . STROBE Statement—Checklist of items that should be included in reports of cross‐sectional studies. Figure S1. Flow diagram of participant recruitment, eligibility assessment, exclusions, and final study population. Table S1. Baseline characteristics of participants with and without posterior hypoechoic nodules in postmenopause. Table S2 . Blinded second‐reader agreement for ultrasonographic findings suggestive of endometriosis. Table S3 . Firth penalized logistic regression analysis of factors associated with posterior compartment hypoechoic nodules in postmenopausal participants.