Application of deep learning model in the sonographic diagnosis of uterine adenomyosis

Application of deep learning model in the sonographic diagnosis of uterine adenomyosis | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Application of deep learning model in the sonographic diagnosis of uterine adenomyosis Diego Raimondo, Antonio Raffone, Anna Chiara Aru, Matteo Giorgi, and 10 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-2176240/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Purpose: To evaluate the diagnostic performance of Deep Learning (DL) machine for the detection of adenomyosis on uterine ultrasonographic images and compare it to intermediate ultrasound skilled trainees. Methods: Prospective observational study conducted between 1st and 30th April 2022. Transvaginal ultrasound (TVUS) diagnosis of adenomyosis was investigated by an experienced sonographer on 100 fertile-age patients. Videoclips of the uterine corpus were recorded and sequential ultrasound images were extracted. Intermediate ultrasound skilled trainees and DL machine were asked to make a diagnosis reviewing uterine images. We evaluated and compared the accuracy, sensitivity, positive predictive value, F1- score, specificity and negative predictive value of the DL model and the trainees for adenomyosis diagnosis. Results: Accuracy of DL and intermediate ultrasound skilled trainees for the diagnosis of adenomyosis were 0.51 (95% CI, 0.48-0.54) and 0.70 (95% CI, 0.60-0.79), respectively. Sensitivity, specificity and F1-score of DL were 0.43 (95% CI, 0.38-0.48), 0.82 (95% CI, 0.79-0.85) and 0.46 (0.42-0.50), whereas intermediate ultrasound skilled trainees had sensitivity of 0.72 (95% CI, 0.52-0.86), specificity of 0.69 (95% CI, 0.58-0.79) and F1-score of 0.55 (95% CI, 0.43-0.66). Conclusion: In this preliminary study DL model showed a lower accuracy but a higher specificity in diagnosing adenomyosis on ultrasonographic images compared to intermediate skilled trainees. artificial intelligence deep learning adenomyosis endometriosis trainee ultrasound Figures Figure 1 1. Introduction Adenomyosis is a benign gynecological disease described by the presence of endometrial glands and stroma within the myometrium, as well as reactive hyperplasia and hypertrophy of the muscular layer [1]. Adenomyosis can cause symptoms like heavy menstrual bleeding, dysmenorrhea and infertility [2–6]. Pathological examination of myometrial specimen remains the gold standard for the diagnosis of adenomyosis which has an estimated prevalence ranging from 21–36% among hysterectomized women. However, only a small and selected percentage of symptomatic women with adenomyosis undergoes hysterectomy, therefore the real prevalence of the disease is underestimated [3, 7–9]. Conversely, transvaginal ultrasound (TVUS) is the method of choice for the non-invasive diagnosis of adenomyosis because it is cheap and easily accessible [10–12]. Unfortunately, although standardization of terminology for the description of myometrium with Morphological Uterus Sonographic Assessment (MUSA) allows universal recognition and assessment of typical adenomyotic ultrasound features [2–3, 13], TVUS is an operator-dependent technique, with adequate diagnostic performance and inter-operator reproducibility only if it is performed by expert sonographers [10, 14–15]. Recently, the need to improve efficiency in all clinical settings using technological advances led to the development of powerful instruments such as artificial intelligence (AI) [16]. AI is defined as the use of several complex algorithm-based applications that can solve problems by simulating human cognitive functions, including data learning and processing, problem solving and decision making [17]. Machine learning (ML) and deep learning (DL) can be accounted among the newest developed technologies in this area. Deep learning is a subfield of machine learning, able to consistently add new data with self-learning ability, thus increasing the performance of the application itself and able to find correlations that humans cannot [18]. AI applicability for healthcare purposes has already been partially investigated and has shown promising results in several medical fields, including gynecology. In particular, AI in gynecological studies was tested for several tasks on medical images, including discriminating malignancy or benignity of ovarian masses, diagnosing cervical cancer, staging endometrial cancer, or diagnosing rectosigmoid endometriosis [17, 19–20]. To the best of our knowledge, no study has ever assessed the accuracy of DL in the diagnosis of adenomyosis using TVUS. Therefore, the aims of this study are to first evaluate the diagnostic performance of DL in the diagnosis of adenomyosis on uterine ultrasonographic images and compare it to that of intermediate ultrasound skilled trainees. 2. Materials And Methods 2.1. Study Protocol and Selection Criteria This was a proof-of-concept, single-center, observational, prospective, cohort study, carried out in a tertiary academic centre. The whole study followed an a priori protocol previously drawn up according to the STROBE guidelines and checklist, which were also followed for reporting the study [21]. We enrolled all consecutive women referring for gynecological ultrasound to the Division of Gynecology and Human Reproduction Physiopathology, IRCCS Azienda Ospedaliero-Universitaria di Bologna, 40138 Bologna, Italy, from April 1, 2022 to April 30, 2022. A priori defined exclusion criteria were the following: age less than 18 years old, virgo status, ongoing or recent (less than 6 months) pregnancy, suspicion of gynecological malignancy, previous hysterectomy, menopausal status, coexistence of adenomyosis and fibroids at TVUS. 2D gray scale mode TVUS uterine images were extracted from ultrasound video clips by an automatic system to develop an end-to-end DL model for the classification of uterine images. Accuracy metrics of this DL model in diagnosing adenomyosis were calculated and compared with those of intermediate skilled TVUS trainees. 2.2. Study Outcomes Primary outcome was the diagnostic performance of the DL machine for the diagnosis of adenomyosis at 2D gray scale mode TVUS images. Secondary outcomes were the following: - the comparison between the accuracy of the DL machine and that of the intermediate skilled TVUS trainees for the diagnosis of adenomyosis; - the comparison of recall (sensitivity), precision (positive predictive value, PPV), F1- score (harmonic mean of precision and sensitivity), specificity and negative predictive value (NPV) between the DL model and the intermediate skilled trainees for the diagnosis of adenomyosis. 2.3. Patient assessment For each patient, anamnestic and clinical data were acquired, as follows: age, body mass index (BMI), parity, history of infertility, previous endometriosis surgery, moderate-to-severe pain symptoms defined as numerical rating scale (NRS) equal or superior to 5 [22], heavy menstrual bleeding referred to as pictorial blood loss analysis chart ≥100 [23], and use of hormonal therapy. 2.4. Ultrasound Details Voluson E8 ultrasound machine (GE Healthcare, Zipf, Austria) with a 4-9 MHz volumetric vaginal probe was used for all acquisitions. Ultrasound scans were obtained with patients in a modified lithotomic position. During 2D gray scale mode TVUS examination, an expert sonographer classified uteruses in three groups: homogenous myometrial echogenicity, fibroids or adenomyosis. Adenomyosis was diagnosed when two or more of the following sonographic criteria were present: globular uterus appearance, asymmetrical thickening of uterine walls, hypoechogenic myometrial cysts, hyperechoic islands, fan-shaped shadowing, echogenic subendometrial lines and buds, junctional zone irregolarities [2-3, 8]. Otherwise, uterine fibroids were diagnosed as well-defined round lesions of the myometrium, frequently with shadows at the edge or an internal fan-shaped shadow [2, 24]. For each patient, presence of deep endometriotic lesions and endometrioma was also investigated according to the IDEA consensus [25, 26]. 2.5. Deep Learning (DL) An end-to-end DL model was developed for the classification of uterine images. Sequential ultrasound images including uterine corpus and cervix were extracted from ultrasound video clips by an automatic system. These ultrasound images were used for the construction, validation and testing of the DL system. The dataset of ultrasound video clips was divided in a random and balanced way into three parts: training (30%), validation (30%) and testing set (40%). The training set was used to train the network by teaching it the parameters of the models. Two architectures were considered: ResNet and Vgg. Among these, Vgg13, Vgg19, ResNet 18 and ResNet 34 models were used. The validation set was used for early stopping, which saves the network weights at the point of best performance, and for optimizing the hyper-parameters. The hyper-parameters used were as follows: - pre-trained and un-trained networks on Microsoft Common Objects in Context; - Batch size: 8-16-32; - Patience: 3-5-10; L1 and L2 regularizations have been implemented in the network. To find the best combination of hyperparameters, Tree Parzen Estimator (TPE) was used as a sampler and Successive Halving Pruner (SHP) as a pruner. To reduce over-fitting, data augmentation was also applied, which generates additional training models using random image transformations. To this end, the captured images were extracted with the resolution reduced from 300x300 pixels to 224x224 pixels, random horizontal flips and vertical flips were employed and Gaussian blur was applied. The test set was used to independently assess the generalization error for the final models chosen. Diagnostic performance of each DL models was acquired. 2.6. Diagnostic performance of trainees For each patient, uterine images were acquired for storage using short video clips (8-10 sec). The uterus (cervix and corpus) was filmed in a sagittal plane and with a lateral left-to-right movement of the probe, using a grayscale mode. Videoclips were downloaded in Mp4 format from the hospital electronic database system and then de-identified prior to get reanalyzed by three intermediate ultrasound skilled trainees. These trainees were 4th year residents in O&G with intermediate ultrasound skills (consisting of more than 500 gynecologic ultrasound cases) doing their postgraduate studies in endometriosis management [27]. The trainees blinded to clinical data were asked separately to make their own diagnosis reviewing uterine images of testing set. Diagnostic performance of each trainee was acquired. 2.7. Statistical analysis Numerical variables were summarized as mean ± SD or median (95% CI); categorical variables were summarized as counts and percentages. Chi-squared test, Fisher's exact test and variance analysis were used for comparison of categorical and numerical variables, where appropriate. To compare the performance of the best DL model with that of the best trainee in diagnosing adenomyosis, accuracy, sensitivity, positive predictive value (PPV), F1- score (harmonic mean of positive predictive value and sensitivity), specificity and negative predictive value (NPV) were calculated. Analyses were conducted using Stata 15 software (StataCorp. 2017. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC). The significance level was set at 5%. 2.8. Ethical Statement and Informed Consent The study protocol received approval by the local Ethics Committee (114/2022/Oss/AOUBo). All patients signed an informed consent before entering the study, and all data were anonymized. 3. Results During the study period, 100 eligible patients were enrolled. Ultrasound diagnosis by expert operator were as follows: 45 patients with homogeneous echogenicity of the myometrium, 30 with fibroids and 25 women with adenomyosis. Baseline and clinical characteristics are summarized in Table 1 . Mean (± SD) age and BMI of the study sample were 35.4 ± 8.0 years and 22.5 ± 2.5 kg/m², respectively. There was no significant difference in terms of baseline data among the three study groups, except for age, rate of spontaneous delivery, and heavy menstrual bleeding, which were higher in the fibroids group, while previous surgery for endometriosis was more frequent in the adenomyosis group. Table 1 Baseline characteristics of the study sample and of the three types of diagnosis by expert sonographers Variable All ( n = 100) Diagnosis by expert sonographers p-value Adenomyosis Fibroids Homogenous echogenicity ( n = 25) ( n = 30) ( n = 45) Age, years 35.4 ± 8.0 35.1 ± 7.1 42.1 ± 6.2 31.1 ± 6.5 < 0.001 * BMI, kg/m² 22.5 ± 2.5 22.0 ± 1.8 22.9 ± 2.3 22.5 ± 2.9 0.395 Spontaneous delivery 24 (24%) 5 (20%) 12 (40%) 7 (16%) 0.045 * Caesarean section 9 (9%) 2 (8%) 4 (13%) 3 (7%) 0.608 Infertility 22 (22%) 4 (16%) 7 (23%) 11 (24%) 0.700 Previous surgery for endometriosis 21 (21%) 9 (36%) 7 (23%) 5 (11%) 0.046 * Hormonal therapy 37 (37%) 14 (56%) 10 (33%) 13 (29%) 0.070 Moderate to severe pain symptoms (NRS equal or superior to 5) Dysmenorrhea 21 (21%) 8 (32%) 2 (7%) 11 (24%) 0.053 Chronic pelvic pain 8 (8%) 2 (8%) 1 (3%) 5 (11%) 0.574 Dyspareunia 19 (19%) 9 (36%) 4 (13%) 6 (13%) 0.063 Heavy menstrual bleeding 11 (11%) 3 (12%) 8 (27%) 0 (0%) < 0.001 * Coexistence of endometriosis at TVUS Deep endometriosis 11 (11%) 4 (16%) 3 (10%) 4 (9%) 0.655 Endometrioma 27 (27%) 9 (36%) 7 (23%) 11 (24%) 0.501 *P-value ≤ 0.05. All characteristics were expressed as number (percentage). Abbreviations: BMI: body mass index; kg, kilograms; m, meters; NRS: numerical rating scale; TVUS: transvaginal ultrasound. Sonographic signs suggestive for adenomyosis in the “adenomyosis group” diagnosed by expert operator are reported in Table 2 . “Globular uterus” was the most frequent sonographic sign (72%), followed by “asymmetrical thickening of uterine walls” and “fan shaped shadowing” (60%). Table 2 Sonographic features of adenomyosis in the adenomyosis group (25 patients), assessed by an expert sonographer Characteristics Prevalence n (%) Globular uterus 18 (72%) Asymmetrical thickening 15 (60%) Fan-shaped shadowing 15 (60%) Myometrial Cysts 12 (48%) Junctional zone irregularities 8 (32%) Hyperechoic islands 7 (28%) Echogenic subendometrial lines and buds 7 (28%) Question mark sign 7 (28%) All characteristics were expressed as number (percentage) After the application of data augmentation, number of uterine images were as follows: - training set: n = 1645 homogeneous echogenicity, n = 1071 fibroids, n = 836 adenomyosis; - validation set: n = 481 homogeneous echogenicity, n = 336 fibroids, n = 252 adenomyosis; - testing set: n = 495 homogeneous echogenicity, n = 359 fibroids, n = 336 adenomyosis. Confusion matrix of the DL for the testing set is shown in Fig. 1 . As reported in Table 3 , accuracy of DL and intermediate ultrasound skilled trainees for the diagnosis of adenomyosis were 0.51 (95% CI, 0.48–0.54) and 0.70 (95% CI, 0.60–0.79), respectively. Sensitivity, specificity and F1-score of DL were 0.43 (95% CI, 0.38–0.48), 0.82 (95% CI, 0.79–0.85) and 0.46 (0.42–0.50), whereas intermediate ultrasound skilled trainees had sensitivity of 0.72 (95% CI, 0.52–0.86), specificity of 0.69 (95% CI, 0.58–0.79) and F1-score of 0.55 (95% CI, 0.43–0.66). Table 3 Diagnostic performance of intermediate ultrasound skilled trainees and DL in the diagnosis of homogeneous echogenicity, fibroid and adenomyosis in the testing set Variable Adenomyosis Fibroids Homogeneous echogenity Intermediate ultrasound skilled trainees DL Intermediate ultrasound skilled trainees DL Intermediate ultrasound skilled trainees DL Sensitivity 0.72 (0.52–0.86) 0.43 (0.38–0.48) 0.63 (0.46–0.78) 0.57 (0.52–0.62) 0.58 (0.44–0.72) 0.53 (0.49–0.57) Specificity 0.69 (0.58–0.79) 0.82 (0.79–0.85) 0.82 (0.72–0.89) 0.73 (0.70–0.76) 0.84 (0.74–0.94) 0.71 (0.68–0.74) PPV 0.44 (0.30–0.59) 0.49 (0.43–0.55) 0.61 (0.43–0.76) 0.47 (0.42–0.52) 0.74 (0.59–0.89) 0.57 (0.52–0.62) NPV 0.88 (0.77–0.94) 0.78 (0.75–0.81) 0.84 (0.74–0.91) 0.80 (0.77–0.83) 0.71 (0.60–0.82) 0.68 (0.65–0.71) Accuracy 0.70 (0.60–0.79) 0.51 (0.48–0.54) 0.77 (0.67–0.84) 0.68 (0.65–0.71) 0.72 (0.63–0.81) 0.64 (0.61–0.67) F1-score 0.55 (0.43–0.66) 0.46 (0.42–0.50) 0.61 (0.49–0.72) 0.52 (0.48–0.56) 0.65 (0.52–0.78) 0.70 (0.67-0-73) Values are expressed as median (95%, CI). Abbreviations: PPV, positive predictive value; NPV, negative predictive value; DL, deep learning. 4. Discussion 4.1 Main findings and interpretation Despite recently AI has gained popularity in the field of medical imaging and has increased its applications in gynecology, no study has ever used this tool in the diagnosis of uterine adenomyosis. Therefore, this study can be considered a proof-of-concept for this issue. The development of standardized and shared diagnostic criteria for the ultrasound diagnosis of adenomyosis has made it possible to achieve adequate levels of accuracy (around 83%), but only in the hands of experienced sonographers or after an intensive and supervised training of O&G trainees [9, 27]. Recently, deep learning (DL) based on artificial neural networks with representation learning has been adopted to help operators to untangle among differential diagnoses. In the present study, the DL model showed a low accuracy in the diagnosis of uterine adenomyosis (51%). This observation may reflect the complexity of the disease. Indeed, adenomyosis is a heterogeneous disease that can have several phenotypes, varying per extension (diffuse, focal or adenomyoma) and location (internal myometrial or junctional invasion) within the myometrium [4, 11]. As a secondary finding, the accuracy of intermediate ultrasound skilled trainees (70%) resulted higher than that of the DL. Moreover, these trainees showed a higher sensitivity (72% vs 43%), but a lower specificity (69% vs 82%) compared to those of the DL. This over-diagnosis could be explained by the tertiary center setting in which frequency of adenomyosis is estimated higher than general population and the offline assessment of uterine images instead of personal execution of TVUS. Conversely, the DL model showed a higher specificity, being more effective in identifying healthy uteruses, with low false positive values. Indeed, the DL model could be a useful tool to exclude adenomyosis where it is not present and disprove the over diagnosis of less experienced operators, avoiding unnecessary second-level examinations or over treatment cases. 4.2. Limitations A limitation of the study is the small sample size and the monocentric design of the study. Larger multicentric studies might be needed to better evaluate the potential clinical aid of AI in the diagnosis of adenomyosis. Although the lack of histological confirmation of adenomyosis may be considered another limitation of the study, pathological examination is unethical in patients without any indication for surgery. On the other hand, to date, an experienced sonographer must be considered an adequate alternative to histological diagnosis in women who are asymptomatic or have not completed their reproductive plan. Impossibility to fully investigate the JZ by using 3D-TVUS examination and to evaluate translesional vascularity through Power Doppler mode may have affected the diagnostic performance of the expert sonographer firstly and then that of the trainees and the DL machine. 5. Conclusions In this proof-of-concept study, DL model achieved a low diagnostic performance for the detection of adenomyosis. Compared to intermediate skilled trainees, DL showed lower accuracy but higher specificity in diagnosing adenomyotic uteri. In order to improve the diagnostic performance of the DL in diagnosing adenomyosis, future research might be focalized to specific training of the DL machine on the recognition of each of ultrasound criteria suggestive for adenomyosis. Moreover, more studies are needed to evaluate any improvement of DL performance adding other sonographic signs (i.e. translesional vascularity using Power Doppler and junctional zone thickness or irregularities at 3D TVUS) and/or clinical data (i.e. presence and severity of pain symptoms and uterine tenderness). Declarations Acknowledgments: No acknowledgement to provide. Ethics approval: The study was conducted in accordance with the Declaration of Helsinki and its later amendments, and approved by the local Ethics Committee of Azienda Ospedaliero-Universitaria di Bologna (protocol code 114/2022/Oss/AOUBo and date of approval 24/03/2022). Funding: The authors did not receive support from any organization for the submitted work. Data Availability: The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the need of privacy maintenance of patients. Authors’ Contribution Statement: Conceptualization and design of the study: DR and ES; Methodology: GC, MG, AR, AT, AM and ACA; Deep learning model development (software), FAG and MGCAC; Validation, AR, AT, AM, PC and RS; Formal Analysis, FAG and LL; Investigation, IG; Data Curation IG; Writing – Original Draft Preparation, MG and ACA; Writing – Review & Editing, DR, GC, LL, AM, PC, AT and AR; Supervision, AM, PC, LL, GC, RS; Project Administration, DR. All authors have approved the final version of the manuscript. Informed Consent: Written informed consent has been obtained from the patients to publish this paper. Disclosure of Potential Conflict of Interes t : The authors have no relevant financial or non-financial interests to disclose. References Cunningham, R.K., Horrow, M.M., Smith, R.J., Springer, J.: Adenomyosis: A Sonographic Diagnosis. RadioGraphics (2018). https://doi.org/10.1148/rg.2018180080 . van den Bosch, T., Dueholm, M., Leone, F.P.G., Valentin, L., Rasmussen, C.K., Votino, A., Van Schoubroeck, D., Landolfo, C., Installé, A.J., Guerriero, S., 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). https://doi.org/10.1002/uog.14806 . Van den Bosch, T., de Bruijn, A.M., de Leeuw, R.A. Dueholm, M., Exacoustos, C., Valentin, L., Bourne, T., Timmerman, D., Huirne, J.A.F.: Sonographic classification and reporting system for diagnosing adenomyosis. Ultrasound Obstet. Gynecol. (2019). https://doi.org/10.1002/uog.19096 . Exacoustos, C., Morosetti, G., Conway, F., Camilli, S., Martire, F.G., Lazzeri, L., Piccione, E., Zupi, E.: New Sonographic Classification of Adenomyosis: Do Type and Degree of Adenomyosis Correlate to Severity of Symptoms? J. Minim. Invasive Gynecol. (2020). https://doi.org/10.1016/j.jmig.2019.09.788 . Harada, T., Khine, Y.M., Kaponis, A., Nikellis, T., Decavalas, G., Taniguchi, F.: The Impact of Adenomyosis on Women’s Fertility. Obstet. Gynecol. Survey. (2016). https://doi.org/10.1097/OGX.0000000000000346 . Zannoni, L., Ambrosio, M., Raimondo, D., Arena, A., Del Forno, S., Borghese, G., Paradisi, R., Seracchioli, R.: Question Mark Sign and Transvaginal Ultrasound Uterine Tenderness for the Diagnosis of Adenomyosis. J. Ultrasound Med. (2020). https://doi.org/10.1002/jum.15237 . Naftalin, J., Hoo, W., Pateman, K., Mavrelos, D., Holland, T., Jurkovic, D.: How common is adenomyosis? A prospective study of prevalence using transvaginal ultrasound in a gynaecology clinic. Hum. Reprod. (2012). https://doi.org/10.1093/humrep/des332 . di Donato, N., Montanari, G., Benfenati, A., Leonardi, D., Bertoldo, V., Monti, G., Raimondo, D., Seracchioli, R.: Prevalence of adenomyosis in women undergoing surgery for endometriosis. Eur. J. Obstet. Gynecol. Reprod. Biol. (2014). https://doi.org/10.1016/j.ejogrb.2014.08.016 . Exacoustos, C., Zupi, E.: A new era in diagnosing adenomyosis is coming. Fertil. Steril. (2018). https://doi.org/10.1016/j.fertnstert.2018.07.005 . Andres, M.P., Borrelli, G.M., Ribeiro, J., Baracat, E.C., Abrão, M.S., Kho, R.M.: Transvaginal Ultrasound for the Diagnosis of Adenomyosis: Systematic Review and Meta-Analysis. J. Minim. Invasive Gynecol. (2018). https://doi.org/10.1016/j.jmig.2017.08.653 . 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). https://doi.org/10.1016/j.jmig.2019.11.001 . Liu, L., Li, W., Leonardi, M., Condous, G., Da Silva Costa, F., Mol, B.W., Wong, L.: Diagnostic Accuracy of Transvaginal Ultrasound and Magnetic Resonance Imaging for Adenomyosis. J. Ultrasound Med. (2021). https://doi.org/10.1002/jum.15635 . Harmsen, M.J., Van den Bosch, T., de Leeuw, R.A., Dueholm, M., Exacoustos, C., Valentin, L., Hehenkamp, W.J.K., Groen-man, F., De Bruyn, C., Rasmussen, C., et al.: Consensus on revised definitions of Morphological Uterus Sonographic Assess-ment (MUSA) features of adenomyosis: Results of modified Delphi procedure. Ultrasound Obstet. Gynecol. (2022). https://doi.org/10.1002/uog.24786 Rasmussen, C.K., Van den Bosch, T., Exacoustos, C., Manegold-Brauer, G., Benacerraf, B.R., Froyman, W., Landolfo, C., Condorelli, M., Egekvist, A.G., Josefsson, H., et al.: Intra- and Inter‐Rater Agreement Describing Myometrial Lesions Using Morphologic Uterus Sonographic Assessment: A Pilot Study. J. Ultrasound Med. (2019). https://doi.org/10.1002/jum.14971 . Lazzeri, L., Morosetti, G., Centini, G., Monti, G., Zupi, E., Piccione, E., Exacoustos, C.: A sonographic classification of adeno-myosis: Interobserver reproducibility in the evaluation of type and degree of the myometrial involvement. Fertil. Steril. (2018). https://doi.org/10.1016/j.fertnstert.2018.06.031 . Shen, J., Zhang, C.J.P., Jiang, B., Chen, J., Song, J., Liu, Z., He, Z., Wong, S.Y., Fang, P.H., Ming, W.K.: Artificial Intelligence Versus Clinicians in Disease Diagnosis: Systematic Review. JMIR Med Inform. (2019). https://doi.org/10.2196/10010 Guerriero, S., Pascual, M., Ajossa, S., Neri, M., Musa, E., Graupera, B., Rodriguez, I., Alcazar, J.L.: Artificial intelligence (AI) in the detection of rectosigmoid deep endometriosis. Eur J Obstet Gynecol Reprod Biol. (2021). https://doi.org/10.1016/j.ejogrb.2021.04.012 . Akkus, Z., Cai, J., Boonrod, A., Zeinoddini, A., Weston, A.D., Philbrick, K.A., Erickson, B.J.: A Survey of Deep-Learning Applications in Ultrasound: Artificial Intelligence-Powered Ultrasound for Improving Clinical Workflow. J Am Coll Radiol. (2019). https://doi.org/10.1016/j.jacr.2019.06.004 . Sone, K., Toyohara, Y., Taguchi, A., Miyamoto, Y., Tanikawa, M., Uchino-Mori, M., Iriyama, T., Tsuruga, T., Osuga, Y.: Application of artificial intelligence in gynecologic malignancies: A review. J Obstet Gynaecol Res. (2021). https://doi.org/10.1111/jog.14818 . Christiansen, F., Epstein, E.L., Smedberg, E., Åkerlund, M., Smith, K., Epstein, E.: Ultrasound image analysis using deep neural networks for discriminating between benign and malignant ovarian tumors: comparison with expert subjective assessment. Ultrasound Obstet Gynecol. (2021). https://doi.org/10.1002/uog.23530 . von Elm, E., Altman, D.G., Egger, M., Pocock, S.J., Gøtzsche, P.C., Vandenbroucke, J.P., STROBE Initiative.: The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies. Ann. Intern. Med. 2007, 147, 573–577. Erratum in Ann. Intern. Med. (2008). https://doi.org/10.7326/0003-4819-147-8-200710160-00010 . Bourdel, N., Alves, J., Pickering, G., Ramilo, I., Roman, H., Canis, M.: Systematic review of endometriosis pain assessment: how to choose a scale? Hum Reprod Update. (2015). https://doi.org/10.1093/humupd/dmu046 . Higham, J.M., O'Brien, P.M., Shaw, R.W.: Assessment of menstrual blood loss using a pictorial chart. Br J Obstet Gynaecol. (1990). https://doi.org/10.1111/j.1471-0528.1990.tb16249.x . Freytag, D., Günther, V., Maass, N., Alkatout, I.: Uterine Fibroids and Infertility. Diagnostics (Basel). (2021). https://doi.org/10.3390/diagnostics11081455 . Guerriero, S., Condous, G., van den Bosch, T., Valentin, L., Leone, F.P., Van Schoubroeck, D., Exacoustos, C., Installé, A.J., Martins, W.P., Abrao, M.S., et al.: Systematic approach to sonographic evaluation of the pelvis in women with suspected endometriosis, including terms, definitions and measurements: a consensus opinion from the International Deep Endometriosis Analysis (IDEA) group. Ultrasound Obstet Gynecol. (2016). https://doi.org/10.1002/uog.15955 . Exacoustos, C., De Felice, G., Pizzo, A., Morosetti, G., Lazzeri, L., Centini, G., Piccione, E., Zupi, E.: Isolated Ovarian Endometrioma: A History Between Myth and Reality. J Minim Invasive Gynecol. (2018). https://doi.org/10.1016/j.jmig.2017.12.026 . Indrielle-Kelly, T., Fischerova, D., Hanuš, P., Frühauf, F., Fanta, M., Dundr, P., Lavu, D., Cibula, D., Burgetova, A.: Early Learning Curve in the Assessment of Deep Pelvic Endometriosis for Ultrasound and Magnetic Resonance Imaging. Biomed Res Int. (2020). https://doi.org/10.1155/2020/8757281 . Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-2176240","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":145209588,"identity":"5922798f-917d-4162-971f-068aa6d98451","order_by":0,"name":"Diego Raimondo","email":"","orcid":"","institution":"IRCCS Azienda Ospedaliero-Universitaria di Bologna","correspondingAuthor":false,"prefix":"","firstName":"Diego","middleName":"","lastName":"Raimondo","suffix":""},{"id":145209589,"identity":"d9452c56-dae2-4ea8-b2fd-31513f688112","order_by":1,"name":"Antonio Raffone","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7klEQVRIiWNgGAWjYDAD9uYDQLICiJmZGwioZYZQPMcSgOQZkAAjKVoY20BMAlr4Z/cffFzAcE+eh4352IOP82qj+duBWn5UbMOpReLOYWbjGQzFhj1sbOmGM7cdz51xmLGBsefMbdzW3Ehmk+ZhSGDcL99jJs277VhuA1ALM2Mbbi3yUC32PWw8ZtJ/5xzLnU9IiwFUSyJYC2NDTe4GQloMbyQbG/MYJCQD/ZIm2XPsQO5GoJaD+PwidyPx4WOeigTbHmCISfyoqcudd/7wwQc/KvB4H+I8OOswmDxAQD0KqCNF8SgYBaNgFIwQAAD9U1Milfqi+AAAAABJRU5ErkJggg==","orcid":"","institution":"University of Bologna","correspondingAuthor":true,"prefix":"","firstName":"Antonio","middleName":"","lastName":"Raffone","suffix":""},{"id":145209590,"identity":"5d8b974d-6fa4-444b-9737-07b9131ca357","order_by":2,"name":"Anna Chiara Aru","email":"","orcid":"","institution":"University of Bologna","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"Chiara","lastName":"Aru","suffix":""},{"id":145209591,"identity":"21a58e5d-79fe-44ff-abd6-3d4a3aebd688","order_by":3,"name":"Matteo Giorgi","email":"","orcid":"","institution":"University of Siena","correspondingAuthor":false,"prefix":"","firstName":"Matteo","middleName":"","lastName":"Giorgi","suffix":""},{"id":145209592,"identity":"380e6778-b5c9-49e0-a493-fd4f0d845a07","order_by":4,"name":"Ilaria Giaquinto","email":"","orcid":"","institution":"Morgagni – Pierantoni Hospital","correspondingAuthor":false,"prefix":"","firstName":"Ilaria","middleName":"","lastName":"Giaquinto","suffix":""},{"id":145209593,"identity":"18b1c81a-35db-4914-8df4-a43b04df2b35","order_by":5,"name":"Emanuela Spagnolo","email":"","orcid":"","institution":"Hospital Universitario La Paz, Paseo de la Castellana","correspondingAuthor":false,"prefix":"","firstName":"Emanuela","middleName":"","lastName":"Spagnolo","suffix":""},{"id":145209600,"identity":"4bbc0dd2-5f52-4c81-b247-ce983c8365f9","order_by":6,"name":"Antonio Travaglino","email":"","orcid":"","institution":"University of Naples Federico II","correspondingAuthor":false,"prefix":"","firstName":"Antonio","middleName":"","lastName":"Travaglino","suffix":""},{"id":145209604,"identity":"63be4164-2f52-46af-ac5c-86f9a32f44c8","order_by":7,"name":"Federico A. Galatolo","email":"","orcid":"","institution":"University of Pisa","correspondingAuthor":false,"prefix":"","firstName":"Federico","middleName":"A.","lastName":"Galatolo","suffix":""},{"id":145209605,"identity":"6e98b705-04d3-41df-b743-f5b60f6be5e3","order_by":8,"name":"Mario G. C. A. Cimino","email":"","orcid":"","institution":"University of Pisa","correspondingAuthor":false,"prefix":"","firstName":"Mario","middleName":"G. C. A.","lastName":"Cimino","suffix":""},{"id":145209608,"identity":"82091275-d44d-46c2-9581-d1c20bbdfddf","order_by":9,"name":"Gabriele Centini","email":"","orcid":"","institution":"University of Siena","correspondingAuthor":false,"prefix":"","firstName":"Gabriele","middleName":"","lastName":"Centini","suffix":""},{"id":145209609,"identity":"e19f08ef-58af-41e4-9c11-f68b71e8362c","order_by":10,"name":"Lucia Lazzeri","email":"","orcid":"","institution":"University of Siena","correspondingAuthor":false,"prefix":"","firstName":"Lucia","middleName":"","lastName":"Lazzeri","suffix":""},{"id":145209610,"identity":"33569823-5eca-4879-bc79-57736d5b1a45","order_by":11,"name":"Antonio Mollo","email":"","orcid":"","institution":"University of Salerno","correspondingAuthor":false,"prefix":"","firstName":"Antonio","middleName":"","lastName":"Mollo","suffix":""},{"id":145209611,"identity":"94a88334-daa3-4402-b40f-27101841a1a5","order_by":12,"name":"Renato Seracchioli","email":"","orcid":"","institution":"IRCCS Azienda Ospedaliero-Universitaria di Bologna","correspondingAuthor":false,"prefix":"","firstName":"Renato","middleName":"","lastName":"Seracchioli","suffix":""},{"id":145209612,"identity":"5a8ceedd-1538-4d78-942a-e7b826cdaf07","order_by":13,"name":"Paolo Casadio","email":"","orcid":"","institution":"IRCCS Azienda Ospedaliero-Universitaria di Bologna","correspondingAuthor":false,"prefix":"","firstName":"Paolo","middleName":"","lastName":"Casadio","suffix":""}],"badges":[],"createdAt":"2022-10-17 18:59:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-2176240/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-2176240/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":28057903,"identity":"ce9b83d7-924b-4969-99dd-5996584bce75","added_by":"auto","created_at":"2022-10-20 18:19:11","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":170462,"visible":true,"origin":"","legend":"\u003cp\u003eConfusion matrix of the DL diagnosis of the testing set\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-2176240/v1/d56977b0f77c953795ab6020.png"},{"id":28246053,"identity":"24e35461-cf8a-4f31-8a6a-5b680f5b4fc5","added_by":"auto","created_at":"2022-10-25 19:29:23","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":553477,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-2176240/v1/1b4de758-0bcf-4e98-8b02-ac76bf830f4a.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Application of deep learning model in the sonographic diagnosis of uterine adenomyosis","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eAdenomyosis is a benign gynecological disease described by the presence of endometrial glands and stroma within the myometrium, as well as reactive hyperplasia and hypertrophy of the muscular layer [1]. Adenomyosis can cause symptoms like heavy menstrual bleeding, dysmenorrhea and infertility [2\u0026ndash;6].\u003c/p\u003e \u003cp\u003ePathological examination of myometrial specimen remains the gold standard for the diagnosis of adenomyosis which has an estimated prevalence ranging from 21\u0026ndash;36% among hysterectomized women. However, only a small and selected percentage of symptomatic women with adenomyosis undergoes hysterectomy, therefore the real prevalence of the disease is underestimated [3, 7\u0026ndash;9].\u003c/p\u003e \u003cp\u003eConversely, transvaginal ultrasound (TVUS) is the method of choice for the non-invasive diagnosis of adenomyosis because it is cheap and easily accessible [10\u0026ndash;12]. Unfortunately, although standardization of terminology for the description of myometrium with Morphological Uterus Sonographic Assessment (MUSA) allows universal recognition and assessment of typical adenomyotic ultrasound features [2\u0026ndash;3, 13], TVUS is an operator-dependent technique, with adequate diagnostic performance and inter-operator reproducibility only if it is performed by expert sonographers [10, 14\u0026ndash;15].\u003c/p\u003e \u003cp\u003eRecently, the need to improve efficiency in all clinical settings using technological advances led to the development of powerful instruments such as artificial intelligence (AI) [16]. AI is defined as the use of several complex algorithm-based applications that can solve problems by simulating human cognitive functions, including data learning and processing, problem solving and decision making [17]. Machine learning (ML) and deep learning (DL) can be accounted among the newest developed technologies in this area. Deep learning is a subfield of machine learning, able to consistently add new data with self-learning ability, thus increasing the performance of the application itself and able to find correlations that humans cannot [18].\u003c/p\u003e \u003cp\u003eAI applicability for healthcare purposes has already been partially investigated and has shown promising results in several medical fields, including gynecology. In particular, AI in gynecological studies was tested for several tasks on medical images, including discriminating malignancy or benignity of ovarian masses, diagnosing cervical cancer, staging endometrial cancer, or diagnosing rectosigmoid endometriosis [17, 19\u0026ndash;20].\u003c/p\u003e \u003cp\u003eTo the best of our knowledge, no study has ever assessed the accuracy of DL in the diagnosis of adenomyosis using TVUS. Therefore, the aims of this study are to first evaluate the diagnostic performance of DL in the diagnosis of adenomyosis on uterine ultrasonographic images and compare it to that of intermediate ultrasound skilled trainees.\u003c/p\u003e"},{"header":"2. Materials And Methods","content":"\u003cp\u003e2.1.\u0026nbsp;Study Protocol and Selection Criteria\u003c/p\u003e\n\u003cp\u003eThis was a proof-of-concept, single-center, observational, prospective, cohort study, carried out\u0026nbsp;in a tertiary academic centre.\u0026nbsp;The whole study followed an \u003cem\u003ea priori\u003c/em\u003e protocol previously drawn up according to the STROBE guidelines and checklist, which were also followed for reporting the study [21].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe enrolled all consecutive women referring for gynecological ultrasound\u0026nbsp;to the Division of Gynecology and Human Reproduction Physiopathology, IRCCS Azienda Ospedaliero-Universitaria di Bologna, 40138 Bologna, Italy,\u0026nbsp;from April 1, 2022 to April 30, 2022. \u003cem\u003eA priori\u0026nbsp;\u003c/em\u003edefined exclusion criteria were the following: age less than 18 years old,\u0026nbsp;virgo\u0026nbsp;status, ongoing or recent (less than 6 months) pregnancy, suspicion of gynecological malignancy, previous hysterectomy, menopausal status, coexistence of adenomyosis and fibroids at TVUS. 2D gray scale mode TVUS uterine images were extracted from ultrasound video clips by an automatic system to develop an end-to-end DL model for the classification of uterine images. Accuracy metrics of this DL model in diagnosing adenomyosis were calculated and compared with those of intermediate skilled TVUS trainees.\u003c/p\u003e\n\u003cp\u003e2.2. Study Outcomes\u003c/p\u003e\n\u003cp\u003ePrimary outcome was the diagnostic performance of the DL machine for the diagnosis of adenomyosis at 2D gray scale mode TVUS images.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSecondary outcomes were the following:\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e-\u0026nbsp;the comparison between the accuracy of the DL machine and that of the intermediate skilled TVUS trainees for the diagnosis of adenomyosis;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e-\u0026nbsp;the comparison of recall (sensitivity), precision (positive predictive value, PPV), F1- score (harmonic mean of precision and sensitivity), specificity and negative predictive value (NPV) between the DL model and the intermediate skilled trainees for the diagnosis of adenomyosis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.3. Patient assessment\u003c/p\u003e\n\u003cp\u003eFor each patient, anamnestic and clinical data were acquired, as follows: age, body mass index (BMI), parity, history of infertility, previous endometriosis surgery, moderate-to-severe pain symptoms defined as numerical rating scale (NRS) equal or superior to 5 [22], heavy menstrual bleeding referred to as pictorial blood loss analysis chart ≥100 [23], and use of hormonal therapy.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.4. Ultrasound Details\u003c/p\u003e\n\u003cp\u003eVoluson\u0026nbsp;E8 ultrasound machine (GE Healthcare, Zipf, Austria) with a 4-9 MHz volumetric vaginal probe was used for all acquisitions. Ultrasound scans were obtained with patients in a modified lithotomic position. During 2D gray scale mode TVUS examination, an expert sonographer classified uteruses in three groups: homogenous myometrial echogenicity, fibroids or adenomyosis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAdenomyosis was diagnosed when two or more of the following sonographic criteria were present: globular uterus appearance, asymmetrical thickening of uterine walls, hypoechogenic myometrial cysts, hyperechoic islands, fan-shaped shadowing, echogenic\u0026nbsp;subendometrial\u0026nbsp;lines and buds, junctional zone\u0026nbsp;irregolarities\u0026nbsp;[2-3, 8]. Otherwise, uterine fibroids were diagnosed as well-defined round lesions of the myometrium, frequently with shadows at the edge or an internal fan-shaped shadow [2, 24].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor each patient, presence of deep endometriotic lesions and endometrioma was also investigated according to the IDEA consensus [25, 26].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.5. Deep Learning\u0026nbsp;(DL)\u003c/p\u003e\n\u003cp\u003eAn end-to-end DL model was developed for the classification of uterine images. Sequential ultrasound images including uterine corpus and cervix were extracted from ultrasound video clips by an automatic system. These ultrasound images were used for the construction, validation and testing of the DL system.\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe dataset of ultrasound video clips was divided in a random and balanced way into three parts: training (30%), validation (30%) and testing set (40%). \u0026nbsp;The training set was used to train the network by teaching it the parameters of the models. Two architectures were considered: ResNet and\u0026nbsp;Vgg. Among these, Vgg13, Vgg19, ResNet 18 and ResNet 34 models were used.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe validation set was used for early stopping, which saves the network weights at the point of best performance, and for optimizing the hyper-parameters. The hyper-parameters used were as follows:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- pre-trained and un-trained networks on Microsoft Common Objects in Context;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- Batch size: 8-16-32;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e- Patience: 3-5-10;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eL1 and L2 regularizations have been implemented in the network.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo find the best combination of hyperparameters, Tree\u0026nbsp;Parzen\u0026nbsp;Estimator (TPE) was used as a sampler and Successive Halving Pruner (SHP) as a pruner. To reduce over-fitting, data augmentation was also applied, which generates additional training models using random image transformations. To this end, the captured images were extracted with the resolution reduced from 300x300 pixels to 224x224 pixels, random horizontal flips and vertical flips were employed and Gaussian blur was applied.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe test set was used to independently assess the generalization error for the final models chosen.\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eDiagnostic performance of each DL models was acquired.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.6. Diagnostic performance of trainees\u003c/p\u003e\n\u003cp\u003eFor each patient, uterine images were acquired for storage using short video clips (8-10 sec). The uterus (cervix and corpus) was filmed in a sagittal plane and with a lateral left-to-right movement of the probe, using a grayscale mode. Videoclips were downloaded in Mp4 format from the hospital electronic database system and then de-identified prior to get reanalyzed by three intermediate ultrasound skilled trainees. These trainees were 4th year residents in O\u0026amp;G with intermediate ultrasound skills (consisting of more than 500 gynecologic ultrasound cases) doing their postgraduate studies in endometriosis management [27]. The trainees blinded to clinical data were asked separately to make their own diagnosis reviewing uterine images of testing set. Diagnostic performance of each trainee was acquired.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.7. Statistical analysis\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNumerical variables were summarized as mean ± SD or median (95% CI); categorical variables were summarized as counts and percentages. \u0026nbsp;Chi-squared test, Fisher's exact test and variance analysis were used for comparison of categorical and numerical variables, where appropriate.\u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTo compare the performance of the best DL model with that of the best trainee in diagnosing adenomyosis, accuracy, sensitivity, positive predictive value (PPV), F1- score (harmonic mean of positive predictive value and sensitivity), specificity and negative predictive value (NPV) were calculated.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAnalyses were conducted using Stata 15 software (StataCorp. 2017. Stata Statistical Software: Release 15. College Station, TX:\u0026nbsp;StataCorp\u0026nbsp;LLC). The significance level was set at 5%.\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e2.8. Ethical Statement and Informed Consent\u003c/p\u003e\n\u003cp\u003eThe study protocol received approval by the local Ethics Committee (114/2022/Oss/AOUBo). All patients signed an informed consent before entering the study, and all data were anonymized.\u003c/p\u003e"},{"header":"3. Results","content":"\u003cp\u003eDuring the study period, 100 eligible patients were enrolled. Ultrasound diagnosis by expert operator were as follows: 45 patients with homogeneous echogenicity of the myometrium, 30 with fibroids and 25 women with adenomyosis.\u003c/p\u003e\n\u003cp\u003eBaseline and clinical characteristics are summarized in Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. Mean (\u0026plusmn;\u0026thinsp;SD) age and BMI of the study sample were 35.4\u0026thinsp;\u0026plusmn;\u0026thinsp;8.0 years and 22.5\u0026thinsp;\u0026plusmn;\u0026thinsp;2.5 kg/m\u0026sup2;, respectively. There was no significant difference in terms of baseline data among the three study groups, except for age, rate of spontaneous delivery, and heavy menstrual bleeding, which were higher in the fibroids group, while previous surgery for endometriosis was more frequent in the adenomyosis group.\u0026nbsp;\u003c/p\u003e\n\u003ctable border=\"1\" id=\"Tab1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eBaseline characteristics of the study sample and of the three types of diagnosis by expert sonographers\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003eAll\u003c/p\u003e\n \u003cp\u003e(\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;100)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"3\"\u003e\n \u003cp\u003eDiagnosis by expert sonographers\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"3\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAdenomyosis\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eFibroids\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eHomogenous echogenicity\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e(\u003c/strong\u003e\u003cspan class=\"BoldItalic\" name=\"Emphasis\" type=\"BoldItalic\"\u003en\u003c/span\u003e\u0026thinsp;\u003cstrong\u003e=\u0026thinsp;25)\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e(\u003c/strong\u003e\u003cspan class=\"BoldItalic\" name=\"Emphasis\" type=\"BoldItalic\"\u003en\u003c/span\u003e\u0026thinsp;\u003cstrong\u003e=\u0026thinsp;30)\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003e(\u003c/strong\u003e\u003cspan class=\"BoldItalic\" name=\"Emphasis\" type=\"BoldItalic\"\u003en\u003c/span\u003e\u0026thinsp;\u003cstrong\u003e=\u0026thinsp;45)\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAge, years\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e35.4\u0026thinsp;\u0026plusmn;\u0026thinsp;8.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e35.1\u0026thinsp;\u0026plusmn;\u0026thinsp;7.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e42.1\u0026thinsp;\u0026plusmn;\u0026thinsp;6.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e31.1\u0026thinsp;\u0026plusmn;\u0026thinsp;6.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eBMI, kg/m\u0026sup2;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e22.5\u0026thinsp;\u0026plusmn;\u0026thinsp;2.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e22.0\u0026thinsp;\u0026plusmn;\u0026thinsp;1.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e22.9\u0026thinsp;\u0026plusmn;\u0026thinsp;2.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e22.5\u0026thinsp;\u0026plusmn;\u0026thinsp;2.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.395\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpontaneous delivery\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e24 (24%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 (20%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e12 (40%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (16%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.045\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eCaesarean section\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 (9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 (8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 (13%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3 (7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.608\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eInfertility\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e22 (22%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 (16%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (23%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11 (24%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.700\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ePrevious surgery for endometriosis\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e21 (21%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 (36%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (23%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 (11%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.046\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eHormonal therapy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e37 (37%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e14 (56%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e10 (33%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e13 (29%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.070\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eModerate to severe pain symptoms (NRS equal or superior to 5)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDysmenorrhea\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e21 (21%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 (32%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 (7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11 (24%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.053\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eChronic pelvic pain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 (8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2 (8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1 (3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5 (11%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.574\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDyspareunia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e19 (19%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 (36%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 (13%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e6 (13%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.063\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eHeavy menstrual bleeding\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11 (11%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3 (12%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 (27%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e0 (0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eCoexistence of endometriosis at TVUS\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDeep endometriosis\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11 (11%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 (16%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3 (10%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4 (9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.655\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEndometrioma\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e27 (27%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e9 (36%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (23%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e11 (24%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.501\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\"\u003e*P-value\u0026thinsp;\u0026le;\u0026thinsp;0.05. All characteristics were expressed as number (percentage). Abbreviations: BMI: body mass index; kg, kilograms; m, meters; NRS: numerical rating scale; TVUS: transvaginal ultrasound.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003c/p\u003e\n\u003cp\u003eSonographic signs suggestive for adenomyosis in the \u0026ldquo;adenomyosis group\u0026rdquo; diagnosed by expert operator are reported in Table \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e. \u0026ldquo;Globular uterus\u0026rdquo; was the most frequent sonographic sign (72%), followed by \u0026ldquo;asymmetrical thickening of uterine walls\u0026rdquo; and \u0026ldquo;fan shaped shadowing\u0026rdquo; (60%).\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable border=\"1\" id=\"Tab2\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eSonographic features of adenomyosis in the adenomyosis group (25 patients), assessed by an expert sonographer\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCharacteristics\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ePrevalence\u003c/p\u003e\n \u003cp\u003e\u003cem\u003en\u003c/em\u003e (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGlobular uterus\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e18 (72%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAsymmetrical thickening\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (60%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFan-shaped shadowing\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e15 (60%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMyometrial Cysts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e12 (48%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eJunctional zone irregularities\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8 (32%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eHyperechoic islands\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (28%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEchogenic subendometrial lines and buds\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (28%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQuestion mark sign\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e7 (28%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003eAll characteristics were expressed as number (percentage)\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eAfter the application of data augmentation, number of uterine images were as follows:\u003c/p\u003e\n\u003cp\u003e- training set: n\u0026thinsp;=\u0026thinsp;1645 homogeneous echogenicity, n\u0026thinsp;=\u0026thinsp;1071 fibroids, n\u0026thinsp;=\u0026thinsp;836 adenomyosis;\u003c/p\u003e\n\u003cp\u003e- validation set: n\u0026thinsp;=\u0026thinsp;481 homogeneous echogenicity, n\u0026thinsp;=\u0026thinsp;336 fibroids, n\u0026thinsp;=\u0026thinsp;252 adenomyosis;\u003c/p\u003e\n\u003cp\u003e- testing set: n\u0026thinsp;=\u0026thinsp;495 homogeneous echogenicity, n\u0026thinsp;=\u0026thinsp;359 fibroids, n\u0026thinsp;=\u0026thinsp;336 adenomyosis.\u003c/p\u003e\n\u003cp\u003eConfusion matrix of the DL for the testing set is shown in Fig. \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. As reported in Table \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e, accuracy of DL and intermediate ultrasound skilled trainees for the diagnosis of adenomyosis were 0.51 (95% CI, 0.48\u0026ndash;0.54) and 0.70 (95% CI, 0.60\u0026ndash;0.79), respectively. Sensitivity, specificity and F1-score of DL were 0.43 (95% CI, 0.38\u0026ndash;0.48), 0.82 (95% CI, 0.79\u0026ndash;0.85) and 0.46 (0.42\u0026ndash;0.50), whereas intermediate ultrasound skilled trainees had sensitivity of 0.72 (95% CI, 0.52\u0026ndash;0.86), specificity of 0.69 (95% CI, 0.58\u0026ndash;0.79) and F1-score of 0.55 (95% CI, 0.43\u0026ndash;0.66).\u0026nbsp;\u003c/p\u003e\n\u003ctable border=\"1\" id=\"Tab3\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eDiagnostic performance of intermediate ultrasound skilled trainees and DL in the diagnosis of homogeneous echogenicity, fibroid and adenomyosis in the testing set\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eAdenomyosis\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eFibroids\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eHomogeneous echogenity\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eIntermediate ultrasound skilled trainees\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eDL\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eIntermediate ultrasound skilled trainees\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eDL\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eIntermediate ultrasound skilled trainees\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eDL\u003c/strong\u003e\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSensitivity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.72 (0.52\u0026ndash;0.86)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.43 (0.38\u0026ndash;0.48)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.63 (0.46\u0026ndash;0.78)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.57 (0.52\u0026ndash;0.62)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.58 (0.44\u0026ndash;0.72)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.53 (0.49\u0026ndash;0.57)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eSpecificity\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.69 (0.58\u0026ndash;0.79)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.82 (0.79\u0026ndash;0.85)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.82 (0.72\u0026ndash;0.89)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.73 (0.70\u0026ndash;0.76)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.84 (0.74\u0026ndash;0.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.71 (0.68\u0026ndash;0.74)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003ePPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.44 (0.30\u0026ndash;0.59)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.49 (0.43\u0026ndash;0.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.61 (0.43\u0026ndash;0.76)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.47 (0.42\u0026ndash;0.52)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.74 (0.59\u0026ndash;0.89)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.57 (0.52\u0026ndash;0.62)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eNPV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.88 (0.77\u0026ndash;0.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.78 (0.75\u0026ndash;0.81)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.84 (0.74\u0026ndash;0.91)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.80 (0.77\u0026ndash;0.83)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.71 (0.60\u0026ndash;0.82)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.68 (0.65\u0026ndash;0.71)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eAccuracy\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.70 (0.60\u0026ndash;0.79)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.51 (0.48\u0026ndash;0.54)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.77 (0.67\u0026ndash;0.84)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.68 (0.65\u0026ndash;0.71)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.72 (0.63\u0026ndash;0.81)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.64 (0.61\u0026ndash;0.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e\u003cstrong\u003eF1-score\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.55 (0.43\u0026ndash;0.66)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.46 (0.42\u0026ndash;0.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.61 (0.49\u0026ndash;0.72)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.52 (0.48\u0026ndash;0.56)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.65 (0.52\u0026ndash;0.78)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.70 (0.67-0-73)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\"\u003eValues are expressed as median (95%, CI). Abbreviations: PPV, positive predictive value; NPV, negative predictive value; DL, deep learning.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Main findings and interpretation\u003c/h2\u003e \u003cp\u003eDespite recently AI has gained popularity in the field of medical imaging and has increased its applications in gynecology, no study has ever used this tool in the diagnosis of uterine adenomyosis. Therefore, this study can be considered a proof-of-concept for this issue.\u003c/p\u003e \u003cp\u003eThe development of standardized and shared diagnostic criteria for the ultrasound diagnosis of adenomyosis has made it possible to achieve adequate levels of accuracy (around 83%), but only in the hands of experienced sonographers or after an intensive and supervised training of O\u0026amp;G trainees [9, 27].\u003c/p\u003e \u003cp\u003eRecently, deep learning (DL) based on artificial neural networks with representation learning has been adopted to help operators to untangle among differential diagnoses.\u003c/p\u003e \u003cp\u003eIn the present study, the DL model showed a low accuracy in the diagnosis of uterine adenomyosis (51%). This observation may reflect the complexity of the disease. Indeed, adenomyosis is a heterogeneous disease that can have several phenotypes, varying per extension (diffuse, focal or adenomyoma) and location (internal myometrial or junctional invasion) within the myometrium [4, 11].\u003c/p\u003e \u003cp\u003eAs a secondary finding, the accuracy of intermediate ultrasound skilled trainees (70%) resulted higher than that of the DL. Moreover, these trainees showed a higher sensitivity (72% \u003cem\u003evs\u003c/em\u003e 43%), but a lower specificity (69% \u003cem\u003evs\u003c/em\u003e 82%) compared to those of the DL. This over-diagnosis could be explained by the tertiary center setting in which frequency of adenomyosis is estimated higher than general population and the offline assessment of uterine images instead of personal execution of TVUS. Conversely, the DL model showed a higher specificity, being more effective in identifying healthy uteruses, with low false positive values. Indeed, the DL model could be a useful tool to exclude adenomyosis where it is not present and disprove the over diagnosis of less experienced operators, avoiding unnecessary second-level examinations or over treatment cases.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e4.2. Limitations\u003c/h2\u003e \u003cp\u003eA limitation of the study is the small sample size and the monocentric design of the study. Larger multicentric studies might be needed to better evaluate the potential clinical aid of AI in the diagnosis of adenomyosis. Although the lack of histological confirmation of adenomyosis may be considered another limitation of the study, pathological examination is unethical in patients without any indication for surgery. On the other hand, to date, an experienced sonographer must be considered an adequate alternative to histological diagnosis in women who are asymptomatic or have not completed their reproductive plan.\u003c/p\u003e \u003cp\u003eImpossibility to fully investigate the JZ by using 3D-TVUS examination and to evaluate translesional vascularity through Power Doppler mode may have affected the diagnostic performance of the expert sonographer firstly and then that of the trainees and the DL machine.\u003c/p\u003e \u003c/div\u003e"},{"header":"5. Conclusions","content":"\u003cp\u003eIn this proof-of-concept study, DL model achieved a low diagnostic performance for the detection of adenomyosis. Compared to intermediate skilled trainees, DL showed lower accuracy but higher specificity in diagnosing adenomyotic uteri.\u003c/p\u003e \u003cp\u003eIn order to improve the diagnostic performance of the DL in diagnosing adenomyosis, future research might be focalized to specific training of the DL machine on the recognition of each of ultrasound criteria suggestive for adenomyosis. Moreover, more studies are needed to evaluate any improvement of DL performance adding other sonographic signs (i.e. translesional vascularity using Power Doppler and junctional zone thickness or irregularities at 3D TVUS) and/or clinical data (i.e. presence and severity of pain symptoms and uterine tenderness).\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments:\u003c/strong\u003e No acknowledgement to provide.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval:\u0026nbsp;\u003c/strong\u003eThe study was conducted in accordance with the Declaration of Helsinki and its later amendments, and approved by the local Ethics Committee of Azienda Ospedaliero-Universitaria di Bologna (protocol code\u0026nbsp;114/2022/Oss/AOUBo\u0026nbsp;and date of approval 24/03/2022).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e\u0026nbsp; The authors did not receive support from any organization for the submitted work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability:\u0026nbsp;\u003c/strong\u003eThe data presented in this study are available on request from the corresponding author. The data are not publicly available due to the need of privacy maintenance of patients.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; Contribution Statement:\u003c/strong\u003e Conceptualization and design of the study: DR and ES; Methodology: GC, MG, AR, AT, AM and ACA; Deep learning model development (software), FAG and MGCAC; Validation, AR, AT, AM, PC and RS; Formal Analysis, FAG and LL; Investigation, IG; Data Curation IG; Writing \u0026ndash; Original Draft Preparation, MG and ACA; Writing \u0026ndash; Review \u0026amp; Editing, DR, GC, LL, AM, PC, AT and AR; Supervision, AM, PC, LL, GC, RS; Project Administration, DR. All authors have approved the final version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInformed Consent:\u0026nbsp;\u003c/strong\u003eWritten informed consent has been obtained from the patients to publish this paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDisclosure of Potential\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eConflict of Interes\u003c/strong\u003e\u003cstrong\u003et\u003c/strong\u003e\u003cstrong\u003e:\u003c/strong\u003e The authors have no relevant financial or non-financial interests to disclose.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003e\u003cspan\u003eCunningham, R.K., Horrow, M.M., Smith, R.J., Springer, J.: Adenomyosis: A Sonographic Diagnosis. RadioGraphics (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1148/rg.2018180080\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003evan den Bosch, T., Dueholm, M., Leone, F.P.G., Valentin, L., Rasmussen, C.K., Votino, A., Van Schoubroeck, D., Landolfo, C., Install\u0026eacute;, A.J., Guerriero, S., 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). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/uog.14806\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eVan den Bosch, T., de Bruijn, A.M., de Leeuw, R.A. Dueholm, M., Exacoustos, C., Valentin, L., Bourne, T., Timmerman, D., Huirne, J.A.F.: Sonographic classification and reporting system for diagnosing adenomyosis. Ultrasound Obstet. Gynecol. (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/uog.19096\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eExacoustos, C., Morosetti, G., Conway, F., Camilli, S., Martire, F.G., Lazzeri, L., Piccione, E., Zupi, E.: New Sonographic Classification of Adenomyosis: Do Type and Degree of Adenomyosis Correlate to Severity of Symptoms? J. Minim. Invasive Gynecol. (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jmig.2019.09.788\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHarada, T., Khine, Y.M., Kaponis, A., Nikellis, T., Decavalas, G., Taniguchi, F.: The Impact of Adenomyosis on Women\u0026rsquo;s Fertility. Obstet. Gynecol. Survey. (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1097/OGX.0000000000000346\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eZannoni, L., Ambrosio, M., Raimondo, D., Arena, A., Del Forno, S., Borghese, G., Paradisi, R., Seracchioli, R.: Question Mark Sign and Transvaginal Ultrasound Uterine Tenderness for the Diagnosis of Adenomyosis. J. Ultrasound Med. (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/jum.15237\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eNaftalin, J., Hoo, W., Pateman, K., Mavrelos, D., Holland, T., Jurkovic, D.: How common is adenomyosis? A prospective study of prevalence using transvaginal ultrasound in a gynaecology clinic. Hum. Reprod. (2012). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/humrep/des332\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003edi Donato, N., Montanari, G., Benfenati, A., Leonardi, D., Bertoldo, V., Monti, G., Raimondo, D., Seracchioli, R.: Prevalence of adenomyosis in women undergoing surgery for endometriosis. Eur. J. Obstet. Gynecol. Reprod. Biol. (2014). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ejogrb.2014.08.016\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eExacoustos, C., Zupi, E.: A new era in diagnosing adenomyosis is coming. Fertil. Steril. (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.fertnstert.2018.07.005\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eAndres, M.P., Borrelli, G.M., Ribeiro, J., Baracat, E.C., Abr\u0026atilde;o, M.S., Kho, R.M.: Transvaginal Ultrasound for the Diagnosis of Adenomyosis: Systematic Review and Meta-Analysis. J. Minim. Invasive Gynecol. (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jmig.2017.08.653\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eTellum, 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). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jmig.2019.11.001\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eLiu, L., Li, W., Leonardi, M., Condous, G., Da Silva Costa, F., Mol, B.W., Wong, L.: Diagnostic Accuracy of Transvaginal Ultrasound and Magnetic Resonance Imaging for Adenomyosis. J. Ultrasound Med. (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/jum.15635\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHarmsen, M.J., Van den Bosch, T., de Leeuw, R.A., Dueholm, M., Exacoustos, C., Valentin, L., Hehenkamp, W.J.K., Groen-man, F., De Bruyn, C., Rasmussen, C., et al.: Consensus on revised definitions of Morphological Uterus Sonographic Assess-ment (MUSA) features of adenomyosis: Results of modified Delphi procedure. Ultrasound Obstet. Gynecol. (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/uog.24786\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eRasmussen, C.K., Van den Bosch, T., Exacoustos, C., Manegold-Brauer, G., Benacerraf, B.R., Froyman, W., Landolfo, C., Condorelli, M., Egekvist, A.G., Josefsson, H., et al.: Intra- and Inter‐Rater Agreement Describing Myometrial Lesions Using Morphologic Uterus Sonographic Assessment: A Pilot Study. J. Ultrasound Med. (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/jum.14971\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eLazzeri, L., Morosetti, G., Centini, G., Monti, G., Zupi, E., Piccione, E., Exacoustos, C.: A sonographic classification of adeno-myosis: Interobserver reproducibility in the evaluation of type and degree of the myometrial involvement. Fertil. Steril. (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.fertnstert.2018.06.031\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eShen, J., Zhang, C.J.P., Jiang, B., Chen, J., Song, J., Liu, Z., He, Z., Wong, S.Y., Fang, P.H., Ming, W.K.: Artificial Intelligence Versus Clinicians in Disease Diagnosis: Systematic Review. JMIR Med Inform. (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2196/10010\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGuerriero, S., Pascual, M., Ajossa, S., Neri, M., Musa, E., Graupera, B., Rodriguez, I., Alcazar, J.L.: Artificial intelligence (AI) in the detection of rectosigmoid deep endometriosis. Eur J Obstet Gynecol Reprod Biol. (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ejogrb.2021.04.012\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eAkkus, Z., Cai, J., Boonrod, A., Zeinoddini, A., Weston, A.D., Philbrick, K.A., Erickson, B.J.: A Survey of Deep-Learning Applications in Ultrasound: Artificial Intelligence-Powered Ultrasound for Improving Clinical Workflow. J Am Coll Radiol. (2019). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jacr.2019.06.004\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eSone, K., Toyohara, Y., Taguchi, A., Miyamoto, Y., Tanikawa, M., Uchino-Mori, M., Iriyama, T., Tsuruga, T., Osuga, Y.: Application of artificial intelligence in gynecologic malignancies: A review. J Obstet Gynaecol Res. (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/jog.14818\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eChristiansen, F., Epstein, E.L., Smedberg, E., \u0026Aring;kerlund, M., Smith, K., Epstein, E.: Ultrasound image analysis using deep neural networks for discriminating between benign and malignant ovarian tumors: comparison with expert subjective assessment. Ultrasound Obstet Gynecol. (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/uog.23530\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003evon Elm, E., Altman, D.G., Egger, M., Pocock, S.J., G\u0026oslash;tzsche, P.C., Vandenbroucke, J.P., STROBE Initiative.: The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies. Ann. Intern. Med. 2007, 147, 573\u0026ndash;577. Erratum in Ann. Intern. Med. (2008). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.7326/0003-4819-147-8-200710160-00010\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eBourdel, N., Alves, J., Pickering, G., Ramilo, I., Roman, H., Canis, M.: Systematic review of endometriosis pain assessment: how to choose a scale? Hum Reprod Update. (2015). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/humupd/dmu046\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eHigham, J.M., O\u0026apos;Brien, P.M., Shaw, R.W.: Assessment of menstrual blood loss using a pictorial chart. Br J Obstet Gynaecol. (1990). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1471-0528.1990.tb16249.x\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eFreytag, D., G\u0026uuml;nther, V., Maass, N., Alkatout, I.: Uterine Fibroids and Infertility. Diagnostics (Basel). (2021). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/diagnostics11081455\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eGuerriero, S., Condous, G., van den Bosch, T., Valentin, L., Leone, F.P., Van Schoubroeck, D., Exacoustos, C., Install\u0026eacute;, A.J., Martins, W.P., Abrao, M.S., et al.: Systematic approach to sonographic evaluation of the pelvis in women with suspected endometriosis, including terms, definitions and measurements: a consensus opinion from the International Deep Endometriosis Analysis (IDEA) group. Ultrasound Obstet Gynecol. (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/uog.15955\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eExacoustos, C., De Felice, G., Pizzo, A., Morosetti, G., Lazzeri, L., Centini, G., Piccione, E., Zupi, E.: Isolated Ovarian Endometrioma: A History Between Myth and Reality. J Minim Invasive Gynecol. (2018). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jmig.2017.12.026\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n \u003cli\u003e\u003cspan\u003eIndrielle-Kelly, T., Fischerova, D., Hanu\u0026scaron;, P., Fr\u0026uuml;hauf, F., Fanta, M., Dundr, P., Lavu, D., Cibula, D., Burgetova, A.: Early Learning Curve in the Assessment of Deep Pelvic Endometriosis for Ultrasound and Magnetic Resonance Imaging. Biomed Res Int. (2020). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1155/2020/8757281\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"artificial intelligence, deep learning, adenomyosis, endometriosis, trainee, ultrasound","lastPublishedDoi":"10.21203/rs.3.rs-2176240/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-2176240/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003ePurpose:\u003c/strong\u003e To evaluate the diagnostic performance of Deep Learning (DL) machine for the detection of adenomyosis on uterine ultrasonographic images and compare it to intermediate ultrasound skilled trainees.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e Prospective observational study conducted between 1st and 30th April 2022. Transvaginal ultrasound (TVUS) diagnosis of adenomyosis was investigated by an experienced sonographer on 100 fertile-age patients. Videoclips of the uterine corpus were recorded and sequential ultrasound images were extracted. Intermediate ultrasound skilled trainees and DL machine were asked to make a diagnosis reviewing uterine images. We evaluated and compared the accuracy, sensitivity, positive predictive value, F1- score, specificity and negative predictive value of the DL model and the trainees for adenomyosis diagnosis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Accuracy of DL and intermediate ultrasound skilled trainees for the diagnosis of adenomyosis were 0.51 (95% CI, 0.48-0.54) and 0.70 (95% CI, 0.60-0.79), respectively. Sensitivity, specificity and F1-score of DL were 0.43 (95% CI, 0.38-0.48), 0.82 (95% CI, 0.79-0.85) and 0.46 (0.42-0.50), whereas intermediate ultrasound skilled trainees had sensitivity of 0.72 (95% CI, 0.52-0.86), specificity of 0.69 (95% CI, 0.58-0.79) and F1-score of 0.55 (95% CI, 0.43-0.66).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003e In this preliminary study DL model showed a lower accuracy but a higher specificity in diagnosing adenomyosis on ultrasonographic images compared to intermediate skilled trainees.\u003c/p\u003e","manuscriptTitle":"Application of deep learning model in the sonographic diagnosis of uterine adenomyosis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2022-10-20 18:19:09","doi":"10.21203/rs.3.rs-2176240/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0d5a2f21-46e8-467c-a544-c7470437711d","owner":[],"postedDate":"October 20th, 2022","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2022-10-25T19:29:18+00:00","versionOfRecord":[],"versionCreatedAt":"2022-10-20 18:19:09","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-2176240","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-2176240","identity":"rs-2176240","version":["v1"]},"buildId":"B-jG_2CBjPDmsCi4Wdhf-","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}