Early identification of women with endometriosis by means of a simple patient-completed questionnaire screening tool: a diagnostic study

Utilizing Artificial Intelligence: Machine Learning Algorithms to Develop a Preoperative Endometriosis Prediction Model

OBJECTIVE

To evaluate the predictive value of clinical features in the diagnosis of endometriosis by utilizing machine learning algorithms (MLAs), aiming to develop an accurate, explainable prediction model.

DESIGN

Retrospective case-control study from 2011 to 2022.

SETTING

Tertiary referral center specializing in pelvic pain and minimally invasive gynecologic surgery.

PARTICIPANTS

All women aged 18-55 undergoing laparoscopic or robot-assisted excision of lesions of the ovary, pelvic viscera, or peritoneal surface by a single surgeon from 2011-2022. Exclusion criteria included women who required emergent surgery, as well as those lacking surgical specimens submitted for pathological analysis or a documented pre-operative pelvic examination.

INTERVENTIONS

209 clinical features, including demographics, presenting symptoms, gynecologic/obstetric history, and physical exam findings, were analyzed as predictors of endometriosis. The primary outcome was model performance in predicting endometriosis, evaluated using accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (ROC-AUC). Feature importance was assessed using SHAP values.

RESULTS

Among 788 participants, 654 (83%) had pathology-confirmed endometriosis. The MLA, XGBoost, achieved an accuracy of 83%, sensitivity of 96%, and ROC-AUC of 0.81. SHAP analysis identified key predictors, including emesis (141 [21.56%] vs 10 [7.46%], p<.001), crampy pain (325 [49.69%] vs 38 [28.36%], p<.001), regular periods (429 [65.60%] vs 60 [44.78%], p<.001), severity of dysmenorrhea (0-3 Likert scale) (3 [2,3] vs 3 [2,3], p=.02), and retrocervical tenderness on rectovaginal exam (126 [19.27%] vs 7 [5.22%], p<.001).

CONCLUSION

This study demonstrates that MLAs have potential to predict endometriosis preoperatively utilizing clinical features. Identified predictors, such as retrocervical tenderness, crampy pain, and regular periods, can aid primary care providers in early recognition and referral. Further validation in diverse populations is necessary to develop a widely applicable clinical prediction tool.

Consensus on Symptom Selection for Endometriosis Questionnaires: A Modified e-Delphi Study

OBJECTIVE

To build consensus on most important symptoms and related consequences for use in questionnaires to characterise individuals with suspected and confirmed endometriosis in the general population.

DESIGN

A questionnaire of 107 symptoms and related consequences of endometriosis was collaboratively developed by patients, medical doctors and researchers and further assessed in a two-round e-Delphi study. Participants assessed the relevance of the symptoms, and a priori it was decided that 70% was the threshold for inclusion of a symptom.

SETTING

Participants represented 7 countries, including Australia, Denmark, France, Hungary, the United Kingdom, the United States, and Turkey.

POPULATION

Individuals with endometriosis, medical doctors and researchers with expertise in endometriosis.

METHODS

A modified e-Delphi study.

MAIN OUTCOME MEASURES

Consensus-based selection of symptoms for endometriosis questionnaires.

RESULTS

Seventy-six participants completed the first Delphi round and 65 completed the second round. Four symptoms met consensus in the first round (menstrual pain, pain during sexual intercourse, cyclic pain during defecation, and infertility), with two additional symptoms reaching consensus in the second round (cyclic pain and increased doctor/health care contacts for abdominal/pelvic pain).

CONCLUSION

This study highlighted six symptoms relevant for inclusion in endometriosis research questionnaires: menstrual pain, pain during sexual intercourse, cyclic pain during defecation, cyclic pain, infertility, and a high number of doctor/health care visits due to abdominal/pelvic pain. Recognising a broad range of potential symptoms is essential for raising awareness and supporting early detection efforts.

Single-Cell RNA Sequencing of PBMCs Identified Junction Plakoglobin (JUP) as Stratification Biomarker for Endometriosis

This study aimed to identify unique characteristics in the peripheral blood mononuclear cells (PBMCs) of endometriosis patients and develop a non-invasive early diagnostic tool. Using single-cell RNA sequencing (scRNA-seq), we constructed the first single-cell atlas of PBMCs from endometriosis patients based on 107,964 cells and 25,847 genes. Within CD16+ monocytes, we discovered JUP as a dysregulated gene. To assess its diagnostic potential, we measured peritoneal fluid (PF) and serum JUP levels in a large cohort of 199 patients including 20 women with ovarian cancer (OC). JUP was barely detectable in PF but was significantly elevated in the serum of patients with endometriosis and OC, with levels 1.33 and 2.34 times higher than controls, respectively. Additionally, JUP was found in conditioned culture media of CD14+/CD16+ monocytes aligning with our scRNA-seq data. Serum JUP levels correlated with endometriosis severity and endometrioma presence but were unaffected by dysmenorrhea, menstrual cycle, or adenomyosis. When combined with CA125 (cancer antigen 125) JUP enhanced the specificity of endometriosis diagnosis from 89.13% (CA125 measured alone) to 100%. While sensitivity remains a challenge at 19%, our results suggest that JUP's potential to enhance diagnostic accuracy warrants additional investigation. Furthermore, employing serum JUP as a stratification marker unlocked the potential to identify additional endometriosis-related genes, offering novel insights into disease pathogenesis.

Assessment of the Pelvic Pain Experienced by Infertile Women is of Prime Importance for Diagnosing Endometriosis

OBJECTIVE

To provide evidence regarding the significance of painful symptoms among women suffering from infertility.

DESIGN

An observational retrospective cross-sectional study.

SETTINGS

University hospital-based research center.

PATIENTS

Infertile patients aged between 18 and 42 years surgically explored for benign gynecological conditions between 01-2004 and 12-2020.

INTERVENTIONS

For each patient, a standardized questionnaire was completed during a face-to-face interview conducted by the surgeon in the month preceding the surgery. Preoperative assessment the pain symptoms was recorded. Pain intensity was assessed with a 10cm visual analog scale (VAS). The pain was considered to be severe when the VAS score was ≥ 7.

MEASUREMENTS AND MAIN RESULTS

Surgery was performed in 839 infertile women. 451 women had severe pelvic pain. Infertile patients with severe pain significantly more often had endometriosis (67.4% versus 30.7% respectively; p <.001) than infertile women without severe pelvic pain, and especially deep infiltrating lesions (43.2% versus 8.5% respectively; p <.001). Moreover, these women more often had intestinal endometriosis lesions (28.4% vs 1.8%; p <.001). After multivariable regression analysis, the presence of endometriosis, irrespective of the phenotype (superficial lesions (OR1.84 [1.19-2.86] and/or ovarian endometrioma OR 2.79 [1.70-4.59] and/or deep infiltrating endometriosis OR 4.49 [2.69-7.51]), and the presence of at least one intestine endometriosis lesion (OR6.49 [2.69-7.51] were significantly associated with severe pelvic pain.

CONCLUSION

Severe pelvic pain is significantly associated with endometriosis and especially deep infiltrating lesions in a population of infertile women. These results demonstrate the importance of thorough questioning regarding pelvic pain symptoms during the initial management of infertile patients.

Questionnaire-based screening of adolescents and young adult women can identify markers associated with endometriosis

STUDY QUESTION

Do adolescents and young adult women (YAW) with histologically proven endometriosis present a specific clinical history?

SUMMARY ANSWER

Questionnaire screening of adolescents and YAW can identify clinical markers associated with histologically proven endometriosis.

WHAT IS KNOWN ALREADY

Some validated questionaries can contribute to an earlier endometriosis diagnosis in adults. None of these scores, however, have been validated for adolescents or YAW.

STUDY DESIGN, SIZE, DURATION

This was an observational cross-sectional study using prospectively recorded data performed between January 2005 and January 2020 in a single university tertiary referral centre for endometriosis diagnosis and management. After a thorough surgical examination of the abdomino-pelvic cavity, women with histologically proven endometriosis were allocated to the endometriosis group, and symptomatic women without evidence of endometriosis were allocated to the endometriosis-free control group. The endometriotic patients were allocated into two sub-groups according to their age: adolescent (≤20 years) and YAW (21-24 years).

PARTICIPANTS/MATERIALS, SETTING, METHODS

Adolescents and YAW ≤24 years of age were operated for a symptomatic benign gynaecological condition with signed informed consent. A standardized questionnaire was prospectively completed in the month before the surgery and included epidemiological data, pelvic pain scores, family history of endometriosis, and symptoms experienced during adolescence. The study searched for correlations by univariate analysis to determine clinical markers of endometriosis in adolescents and YAW compared with endometriosis-free control patients.

MAIN RESULTS AND THE ROLE OF CHANCE

Of the 262 study participants, 77 women were adolescents (≤20 years of age) and 185 patients (70.6%) were YAW. The endometriosis group included 118 patients (45.0%) and 144 (55.0%) were assigned to the control group. A family history of endometriosis, absenteeism from school during menstruation, history of fainting spells during menstruation, and prescription of oral contraceptive pills for intense dysmenorrhea were significantly more frequently observed in the endometriotic patients. The prevalence and mean pain scores for dysmenorrhea, deep dyspareunia, non-cyclic chronic pelvic pain and gastrointestinal and lower urinary tract symptoms were significantly greater in the endometriosis group, as was experienced rectal bleeding.

LIMITATIONS, REASONS FOR CAUTION

The study was performed in a single referral centre that treats patients with potentially more severe disease. This questionnaire was evaluated on a population of patients with an indication for endometriosis surgery, which can also select patients with more severe disease. Women with asymptomatic endometriosis were not considered in this study. These factors can affect the external validity of this study.

WIDER IMPLICATIONS OF THE FINDINGS

Patient interviews are relevant to the diagnosis of endometriosis in adolescents and YAW. Combined with imaging and clinical examination, this approach will enable earlier diagnosis and treatment, while remaining non-invasive and rapid.

STUDY FUNDING/COMPETING INTEREST(S)

The study received no funding from external sources. There are no conflicts of interest.

TRIAL REGISTRATION NUMBER

N/A.

Developing a Predictive Model for Minimal or Mild Endometriosis as a Clinical Screening Tool in Infertile Women: Uterosacral Tenderness as a Key Predictor

STUDY OBJECTIVE

To develop a noninvasive predictive model based on patients with infertility for identifying minimal or mild endometriosis.

DESIGN

A retrospective cohort study.

SETTING

This study was conducted at a tertiary referral center.

PATIENTS

A total of consecutive 1365 patients with infertility who underwent laparoscopy between January 2013 and August 2020 were divided into a training set (n = 910) for developing the predictive model and a validation set (n = 455) to confirm the model's prediction efficiency. The patients were randomly assigned in a 2:1 ratio.

INTERVENTIONS

Sensitivities, specificities, area under the curve, the Hosmer-Lemeshow goodness of fit test, Net Reclassification Improvement index, and Integrated Discrimination Improvement index were evaluated in the training set to select the optimum model. In the validation set, the model's discriminations, calibrations, and clinical use were tested for validation.

MEASUREMENTS AND MAIN RESULTS

In the training set, there were 587 patients with minimal or mild endometriosis and 323 patients without endometriosis. The combination of clinical parameters in the model was evaluated for both statistical and clinical significance. The best-performing model ultimately included body mass index, dysmenorrhea, dyspareunia, uterosacral tenderness, and serum cancer antigen 125 (CA-125). The nomogram based on this model demonstrated sensitivities of 87.7% and 93.3%, specificities of 68.6% and 66.4%, and area under the curve of 0.84 (95% confidence interval 0.81-0.87) and 0.85 (95% confidence interval 0.80-0.89) for the training and validation sets, respectively. Calibration curves and decision curve analyses also indicated that the model had good calibration and clinical value. Uterosacral tenderness emerged as the most valuable predictor.

CONCLUSION

This study successfully developed a predictive model with high accuracy in identifying infertile women with minimal or mild endometriosis based on clinical characteristics, signs, and cost-effective blood tests. This model would assist clinicians in screening infertile women for minimal or mild endometriosis, thereby facilitating early diagnosis and treatment.

Do women with suspected endometriosis benefit from pelvic examination to improve diagnostic and management strategy?

OBJECTIVE

To analyze the literature and expose best evidence available regarding the benefit of pelvic examination for women with suspected endometriosis METHODS: the AGREE II and GRADE systems for grading scientific evidence.

RESULTS

Endometriosis is characterized by the heterogeneity in its clinical presentation with many different symptoms reported by patients. In the literature, questioning for each symptom has a high sensitivity, reaching 76-98 %, but lacks specificity (20 - 58 %). The symptom-based approach is limited by its low specificity, the absence of external validation for most of the models developed and the inability to characterize the extent of the disease, which could have major implications in the decision - making process. The latest systematic review and meta-analysis included a total of 30 studies with 4,565 participants, compared the diagnostic performance of several modalities for endometriosis. Physical examination had a pooled sensitivity of 71 % and a specificity of 69 %, with an average diagnostic accuracy of 0.76. Overall, the value of pelvic examination is conferred by its high positive likehood ratio and specificity. Besides its diagnostic value, pelvic examination improves patients' management by allowing the identification of a possible myofascial syndrome as a differential diagnosis. It also increases the quality of the preoperative workup and influences the quality of surgical excision and decreases the time to diagnosis.

CONCLUSION

Despite the lack of studies in the primary care context, pelvic examination (vaginal speculum and digital vaginal examination) increases the diagnostic value for suspected endometriosis in association with questioning for symptoms.

Update on the management of endometriosis-associated pain in France

The French National College of Obstetricians and Gynecologists (CNGOF) published guidelines for managing endometriosis-associated pain in 2018. Given the development of new pharmacological therapies and a review that was published in 2021, most national and international guidelines now suggest a new therapeutic approach. In addition, a novel validated screening method based on patient questionnaires and analysis of 109-miRNA saliva signatures, which combines biomarkers and artificial intelligence, opens up new avenues for overcoming diagnostic challenges in patients with pelvic pain and for avoiding laparoscopic surgery when sonography and MRI are not conclusive. Dienogest (DNG) 2 mg has been a reimbursable healthcare expense in France since 2020, and, according to recent studies, it is at least as effective as combined hormonal contraception (CHC) and can be used as an alternative to CHC for first-line treatment of endometriosis-associated pain. Since 2018, the literature concerning the use of DNG has grown considerably, and the French guidelines should be modified accordingly. The levonorgestrel intrauterine system (LNG IUS) and other available progestins per os, including DNG, or the subcutaneous implant, can be offered as first-line therapy, gonadotropin-releasing hormone (GnRH) agonists with add-back therapy (ABT) as second-line therapy. Oral GnRH antagonists are promising new medical treatments for women with endometriosis-associated pain. They competitively bind to GnRH receptors in the anterior pituitary, preventing native GnRH from binding to GnRH receptors and from stimulating the secretion of luteinizing hormone and follicle-stimulating hormone. Consequently, estradiol and progesterone production is reduced. Oral GnRH antagonists will soon be on the market in France. Given their mode of action, their efficacy is comparable to that of GnRH agonists, with the advantage of oral administration and rapid action with no flare-up effect. Combination therapy with ABT is likely to allow long-term treatment with minimal impact on bone mass. GnRH antagonists with ABT may thus be offered as second-line treatment as an alternative to GnRH agonists with ABT. This article presents an update on the management of endometriosis-associated pain in women who do not have an immediate desire for pregnancy.

[Pelvic exam in gynecology and obstetrics: Guidelines for clinical practice]

OBJECTIVE

To provide guidelines for the pelvic clinical exam in gynecology and obstetrics.

MATERIAL AND METHODS

A multidisciplinary experts consensus committee of 45 experts was formed, including representatives of patients' associations and users of the health system. The entire guidelines process was conducted independently of any funding. The authors were advised to follow the rules of the Grading of Recommendations Assessment, Development and Evaluation (GRADE®) system to guide assessment of quality of evidence. The potential drawbacks of making strong recommendations in the presence of low-quality evidence were emphasized.

METHODS

The committee studied 40 questions within 4 fields for symptomatic or asymptomatic women (emergency conditions, gynecological consultation, gynecological diseases, obstetrics, and pregnancy). Each question was formulated in a PICO (Patients, Intervention, Comparison, Outcome) format and the evidence profiles were produced. The literature review and recommendations were made according to the GRADE® methodology.

RESULTS

The experts' synthesis work and the application of the GRADE method resulted in 27 recommendations. Among the formalized recommendations, 17 present a strong agreement, 7 a weak agreement and 3 an expert consensus agreement. Thirteen questions resulted in an absence of recommendation due to lack of evidence in the literature.

CONCLUSIONS

The need to perform clinical examination in gynecological and obstetrics patients was specified in 27 pre-defined situations based on scientific evidence. More research is required to investigate the benefit in other cases.

Self-report symptom-based endometriosis prediction using machine learning

Endometriosis is a chronic gynecological condition that affects 5-10% of reproductive age women. Nonetheless, the average time-to-diagnosis is usually between 6 and 10 years from the onset of symptoms. To shorten time-to-diagnosis, many studies have developed non-invasive screening tools. However, most of these studies have focused on data obtained from women who had/were planned for laparoscopy surgery, that is, women who were near the end of the diagnostic process. In contrast, our study aimed to develop a self-diagnostic tool that predicts the likelihood of endometriosis based only on experienced symptoms, which can be used in early stages of symptom onset. We applied machine learning to train endometriosis prediction models on data obtained via questionnaires from two groups of women: women who were diagnosed with endometriosis and women who were not diagnosed. The best performing model had AUC of 0.94, sensitivity of 0.93, and specificity of 0.95. The model is intended to be incorporated into a website as a self-diagnostic tool and is expected to shorten time-to-diagnosis by referring women with a high likelihood of having endometriosis to further examination. We also report the importance and effectiveness of different symptoms in predicting endometriosis.

Accuracy of combined physical examination, transvaginal ultrasonography, and magnetic resonance imaging to diagnose deep endometriosis

OBJECTIVES

To assess the ability of physical examination (PE), transvaginal ultrasonography (TVUS), and magnetic resonance imaging (MRI) alone and combined to diagnose deep infiltrating endometriosis (DIE).

DESIGN

We retrospectively queried our pelvic MRI database to identify women who underwent PE, TVUS, and pelvic MRI for DIE up to 12 months before surgery between January 1, 2016 and August 31, 2020. The presence of uterosacral ligaments (USL), vaginal, rectosigmoid (RS), parametrial, or sacrorectogenital septum (lateral) DIE shown by PE, TVUS, and MRI were correlated with surgical and histological findings.

SETTING

Academic hospital.

PATIENT(S)

We included 178 patients.

INTERVENTION(S)

Clinical and imaging evaluation of women who were diagnosed at surgery with deep pelvic endometriosis.

MAIN OUTCOME MEASURE(S)

The sensitivity, specificity, positive and negative predictive values, and accuracy of each technique separately and combined were assessed for each location. When the 3 techniques were combined, 2 models were tested as follows: all 3 techniques positive and concordant; and ≥2 techniques positive and concordant.

RESULT(S)

The prevalence of USL, vaginal, RS, and lateral DIE were 94.4%, 20.2%, 34.3%, and 32.6%, respectively. In addition, MRI was more sensitive than PE, TVUS or any combination to detect DIE. Moreover, MRI and model B were the most accurate for detecting USL and RS locations with an accuracy of 90.4% and 82.6%, a sensitivity of 91.1% and 50%, and a specificity of 77.8% and 90.9%, respectively. Model B was the most accurate for the vaginal location with an accuracy, sensitivity, and specificity of 82.6%, 50%, and 90.9%, respectively. Finally, MRI was more accurate than any combination for identifying a lateral location with an accuracy, sensitivity, and specificity of 75.1%, 36%, and 93.8%, respectively.

CONCLUSION(S)

A combination of PE, TVUS, and MRI was more accurate than each technique separately to diagnose DIE because of the equally high sensitivity of each, as well as the high specificity of PE and TVUS.

Diagnostic delay of endometriosis in adults and adolescence-current stage of knowledge

PURPOSE

The purpose of this article is to provide a modern perspective on the diagnosis of endometriosis with particular attention to the role of ultrasound examination. In the present study, we highlight the problem of endometriosis in teenage girls and discuss the patients' perspective on the diagnostic process.

METHODS

In order to present the most recent reports on the diagnosis of endometriosis, the PubMed database was searched. Articles published within the last 3 years (2019-2021) and those considered relevant during the bibliographic review were analyzed.

RESULTS

The role of ultrasound examination and assessment of patients' perspective related to delayed and incorrect diagnosis were considered to be the most important in the recent reports. Attention was also paid to the problem of endometriosis diagnosis in adolescent girls.

CONCLUSIONS

Appropriately constructed and used questionnaires help to determine the risk of endometriosis in a particular patient. The primary method for diagnosis is extended ultrasound examination, which should be performed especially in patients with a high risk of developing the disease. This procedure is applicable to both adult and adolescent women. Awareness of the possibility of developing the disease in a particular patient, combined with appropriate use of ultrasound examination, can contribute to the decrease in diagnostic delay.

Patient-Reported Outcome Measures in Endometriosis

Patient Reported Outcome Measures (PROM) evoke measurements that allow capturing patients’ perspectives on their condition. In endometriosis care, physicians’ understanding of the effect of the disease and the treatment on patients is often poor. The use of PROMs in endometriosis clinical practice can facilitate patient-provider communication and the implementation of patient-centered care, improve patients’ quality of life, as well as provide a tool for patients’ self-management of the disease. Today, PROMs are extensively used in research and clinical trials, however they are barely used in clinical practice. The development of digital tools facilitating capturing PROMs can contribute to their use by physicians in routine endometriosis care. However, all PROMs are not adapted to be used in routine care in the context of endometriosis. The objective of this study was to present a catalogue of available PROMs for routine endometriosis care and evaluate them according to selected criteria. To do so, we explored the different PROMs currently in the literature. Consequently, 48 PROM were identified as tools used to evaluate various dimensions of the impact of endometriosis on patients. The selected PROMs were evaluated for their potential to be used as a standard in clinical practice in endometriosis. The selected catalogue of PROMs is the starting point for the integration of digital tools to capture PROMs and the development of patient-centered dashboards to be used by patients and clinicians in endometriosis care and self-management to improve care processes, patient satisfaction, quality of life, and outcomes.