Diagnostic discrepancies and clinical value of second medical opinions (SMO) for endometriosis: a nationwide study analysis

The SMO platform deuxiemeavis.fr received approval from the French national commission for data privacy (Commission Nationale Informatique et Libertés, CNIL) and patients provided informed consent that their anonymous data may be used for research and quality improvement purposes. No additional approvals were required according to the French legislation as this was an observational study based on medical records. All methods were performed in accordance with the relevant guidelines and regulations (Declarations of Helsinki). The SMO request process via deuxiemeavis.fr consists of five main steps carried out by the patient, as recently described by Sanchez et al. [ 13 ]. The process begins with the patient selecting the category of disease and completing a medical questionnaire related to their condition, including their primary inquiries for the physician. A specialist is then chosen from a list of SMO experts. Once notified, the designated expert can either accept or decline the patient’s request. If accepted, the physician performs a detailed medical analysis of the case and generates a written medical report within seven days. With the patient’s consent, this SMO report can be shared with their general practitioner or the specialist who provided the initial opinion. All cases were evaluated exclusively through telemedicine, with no direct in-person consultation; therefore, diagnostic conclusions were based on a comprehensive synthesis of patient history, symptomatology, and available imaging, rather than on physical examination or surgical confirmation. The expert panel for endometriosis comprises 26 specialists, including radiologists proficient in analyzing endometriosis imaging (pelvic ultrasound and magnetic resonance imaging) and gynecologists specializing in treatment recommendations. These experts were carefully selected based on their academic standing, clinical expertise, and international contributions, and were approved by a scientific committee. The panel’s composition reflects a multidisciplinary approach, ensuring comprehensive evaluation and management of endometriosis cases. An observational cohort study was conducted using nationwide data retrospectively extracted from the French SMO platform deuxiemeavis.fr ( https://www.deuxiemeavis.fr/ ) between January 1 st, 2020, and March 31 st, 2024. The study population is based on anonymous data from patients who submitted a SMO request through the platform for an issue relating to endometriosis diagnosis. Inclusion criteria were all patients who requested a SMO for endometriosis. Exclusion criteria were patients with incomplete records or missing MRI images required for analysis. The following variables were abstracted from patient questionnaires, including endometriosis-specific questionnaires assessing symptoms and quality of life: height and weight, professional activity, age, ECOG score [ 15 ] (Table 1 ), evaluation of pain using the Visual Analogue Scale (VAS) pain scale [ 16 ], family and patient medical history, gynecological history, history of surgical and medical treatment, and if available, a report of the patient’s first medical opinion received prior to their SMO request. Patients are required to upload their MRI and sonography reports and images, as well as a drawing of the areas on the body where they experience pain. They can also upload the results of laboratory tests, surgical reports and other non-specific reports. This study was conducted in a general population setting, and there was no standardized diagnostic protocol imposed beyond the uploading of MRI images by the patients. Table 1 ECOG - Performance status scale 0 Fully active, able to carry on all pre-disease performance without restriction 1 Restricted in physically strenuous activity but ambulatory and able to carry out work of a light or sedentary nature, e.g., light house work, office work 2 Ambulatory and capable of all selfcare but unable to carry out any work activities; up and about more than 50% of waking hours 3 Capable of only limited selfcare; confined to bed or chair more than 50% of waking hours 4 Completely disabled; cannot carry on any selfcare; totally confined to bed or chair 5 Dead ECOG - Performance status scale The diagnostic convergence index (DCI) was calculated as the rate of diagnostic concordance between initial physician assessments and SMO evaluations DCI is a proportion of patients whose diagnosis changed between the first and second medical opinions, defined mathematically as : \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$DCI\;=\;(N_{(divergent\;opinion\;between\;first\;and\;second\;medical\;opinion)}\;/\;N_{(total)}\;)\ast100$$\end{document} Five distinct clinical profiles were identified in the database: Confirmed endometriosis diagnosis • Established Physical/histological/imaging confirmation from initial workup Negative endometriosis status • Explicit exclusion of diagnosis through prior investigations Suspected endometriosis • Suggestive symptomatology (dysmenorrhea, dyspareunia, infertility) Imaging discrepancies • Discordance between radiological findings and clinical interpretation Undetermined diagnostic status • Non-specific pelvic symptoms without supporting clinical evidence Confirmed endometriosis diagnosis • Established Physical/histological/imaging confirmation from initial workup Negative endometriosis status • Explicit exclusion of diagnosis through prior investigations Suspected endometriosis • Suggestive symptomatology (dysmenorrhea, dyspareunia, infertility) Imaging discrepancies • Discordance between radiological findings and clinical interpretation Undetermined diagnostic status • Non-specific pelvic symptoms without supporting clinical evidence Data management was performed using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA), and statistical analyses were conducted using R version 4.3.2 (R Foundation for Statistical Computing, Vienna, Austria). Non-normally distributed quantitative variables were presented as median with interquartile range (IQR: Q1-Q3), while qualitative variables were expressed as frequencies and percentages. Statistical analysis was based on χ2 test and Fisher’s exact test for ordinal variables. For continuous variables, Student’s t-test or Mann–Whitney test were used ( p  < 0.05 was considered significant).

Fig. 1 Flow chart of patient diagnosis after the first medical opinion Flow chart of patient diagnosis after the first medical opinion Among the 3,070 requests for SMOs on endometriosis diagnostic, 97.3% (N = 2,987) were complete for analysis. The main reason for exclusion was incomplete or absent data concerning the first medical opinion in 83 requests (2.7%), specifically cases where no MRI data were available. Among the 2987 patients included in the analysis, at first medical opinion 41.3% of patients (N = 1233) had confirmed endometriosis diagnosis, 34.1% (N = 1018) had been diagnosed without endometriosis (although presenting suggestive symptoms), 13.5% (N = 402) were suspected to have endometriosis. Moreover, 4.8% of patients (N = 143) presented imaging discrepancy, and 6.4% (N = 191) were considered with an undetermined diagnostic status. Figure  1 represents the flow chart of patient diagnosis after the first medical opinion. Table  2 describes the characteristics of patients ( N  = 2987). The mean age of patients was 29.3 ± 7.3 years old with the majority of patients at reproductive age and between 18 and 43 years old. Mean BMI was 23.6 ± 5.3. A noTable 19% of the population were smokers. The majority (73.7%) were nulliparous, while 12.2% had one child and 14.1% had two or more children. The distribution by professional categories shows that 40.5% of the participants were employees, 11.9% were in higher intellectual professions and executive positions, whereas 27.7% were students or unemployed. Regarding physical performance status, assessed through the ECOG scale, 14.7% had a score of two or higher meaning they were unable to carry out any work-related activities and 46.5% had a score of one, meaning they were restricted in physical activity, but still able to perform light or sedentary activities. A total of 50.7% of patients declared suffering dysmenorrhea and having an average VAS pain scale in daily life of 6.1 ± 2. Moreover, 69.4% of patients declared regular interruption of sexual intercourse due to pain, with an average VAS pain scale during intercourse of 5.7 ± 2.3. A history of sexual abuse was declared by 13.5% of patients, while 18.7% reported infertility. A family history of endometriosis was found in 20.3% of patients, and 43.1% had previously undergone abdominal or gynecologic surgery, including 8.8% with endometriosis surgery. The average delay for SMOs was 3.6 days. Baseline population characteristics following SMOs are presented in Additional files 1 and 2. Table 2 Baseline population characteristics before SMO Variables N = 2,987 (100%) Patient’s characteristics Mean age (years old) (+/- SD*) 29.3 ± 7.3 Mean BMI* (+/- SD*) 23.6 ± 5.3 Smoker 566 (19%) Professional categories  Farmers 9 (0.3%)  Craftsmen, shopkeepers and company directors 142 (4.8%)  Executives and higher intellectual professions 356 (11.9%)  Employees 1,209 (40.5%)  Workers 45 (1.5%)  Intermediate occupations 107 (3.6%)  Others 286 (9.7%)  Unemployed, students 827 (27.7%)  Retirees 1 (0.03%) ECOG score*   0 1,160 (38.8%)   1 1,382 (46.5%)   2 290 (9.7%)   ≥ 3 149 (5%) Patient’s history  History of sexual abuse 402 (13.5%)  History of infertility 560 (18.7%) Parity  0 2,201 (73.7%)  1 357 (12,2%)  ≥ 2 421(14.1%) History of Endometriosis surgery 263 (8.8%) Pain Characteristics  Dysmenorrhea 1,515 (50.7%)  Average VAS*pain scale in daily life (+/- SD*) 6.1 ± 2  Interruption of sexual intercourse due to pain 2,072 (69.4%)  Average VAS pain scale during sexual intercourse (+/- SD*) 5.7 ± 2.3 * BMI Body Mass Index, SD Standard Deviation, ECOG score Eastern Cooperative Oncology Group score, VAS Visual Analogue Scale Baseline population characteristics before SMO * BMI Body Mass Index, SD Standard Deviation, ECOG score Eastern Cooperative Oncology Group score, VAS Visual Analogue Scale In France, seven regional healthcare networks were established to support the diagnosis, triage, and coordination of care pathways for patients with confirmed or suspected endometriosis, in alignment with the French National Strategy for Endometriosis published in 2020 by the “Agence Régionale de Santé (ARS)” [ 17 ]. The seven regions were defined arbitrarily, covering nearly half of mainland France. Figure 2 illustrates the number of SMO requests per region, ranging from 5.7 requests per 100,000 women in Provence-Alpes-Côte d'Azur to 16 requests per 100,000 women in Bourgogne-Franche-Comté. Notably, the presence of a dedicated endometriosis healthcare network did not significantly influence the number of SMO requests (p = 0.22). Contrary to expectations, the presence of dedicated healthcare networks for endometriosis did not significantly reduce the diagnostic delay. Patients in regions with established endometriosis networks experienced similar delays (3.8 ± 1.5 Years) compared to those in regions without such networks (3.8 ± 1.5 Years, p = 0.6). Fig. 2 Number of support requests by region and location across seven endometriosis healthcare networks in total Number of support requests by region and location across seven endometriosis healthcare networks in total Figure 3 presents a heat map for the Diagnostic Convergence Index (DCI) for each diagnostic category, comparing first and second diagnostic opinions (SMOs). The analysis revealed significant diagnostic modifications in 51.1% of cases (1,523 out of 2,987 patients) following SMOs. Notably, SMOs overturned the initial diagnosis of endometriosis in 28.3% of cases, while diagnosing endometriosis in 22.7% of cases where the initial opinion did not. After SMOs, 55.4% of patients (n=1,654) received a final diagnosis of endometriosis, whereas 39.6% (n=1,184) were diagnosed with conditions other than endometriosis, despite presenting suggestive symptoms. The SMOs demonstrated an improvement in diagnostic precision. Initially, 24.7% of cases (n=736) lacked a definitive diagnosis. Following SMOs, this percentage decreased significantly, with experts unable to reach a conclusion in only 5.0% of cases (n=149). Specifically, 3.9% of SMOs (n=115) maintained uncertainty regarding endometriosis, while 1.1% (n=34) remained without a diagnosis due to insufficient key elements in the medical record or imaging limitations preventing a conclusive diagnosis. Fig. 3 Heatmap of diagnostic reclassification after SMO. Rows represent diagnoses provided after SMO, and columns represent initial diagnoses provided to patients. Each cell indicates the number of patients with a given combination of first and second opinion diagnoses.The color gradient from light to dark corresponds to increasing patient counts, as shown in the scale bar Heatmap of diagnostic reclassification after SMO. Rows represent diagnoses provided after SMO, and columns represent initial diagnoses provided to patients. Each cell indicates the number of patients with a given combination of first and second opinion diagnoses.The color gradient from light to dark corresponds to increasing patient counts, as shown in the scale bar A significant proportion of patients reported long-standing symptoms: 34.5% had been experiencing pain for 1 to 5 years, while 54.5% had been suffering for more than 5 years (Additional file 2).

This study highlights the significant impact of SMOs in improving the diagnosis and management of endometriosis. SMOs led to diagnostic changes in over half of cases and reduced diagnostic uncertainty from 24.7% to 5%. The findings emphasize the need for equitable access to expert care, particularly in underserved regions like overseas territories. SMO, clinical judgment, patient context, and multidisciplinary collaboration remain essential to ensure balanced and individualized care to address the complex challenges posed by endometriosis.

This observational cohort study provides compelling evidence for the significant clinical value of SMO in improving the diagnosis of endometriosis. The study, which included 2,987 patients, highlights several key findings that underscore both the complexity of endometriosis and the critical role of expert evaluation in its management. The patient demographic and clinical profile data offer valuable insights into the population seeking SMOs for suspected endometriosis. With a mean age of 29.3 years and a high rate of nulliparity (73.7%), the study population primarily consists of young women of reproductive age, many of whom may have concerns about fertility. This reflects the reproductive challenges associated with endometriosis, particularly in patients with deep infiltrating endometriosis or ovarian endometriomas. The impact of the condition on patients’ quality of life is also evidenced, with high pain scores (average VAS 6.1/10): this study describes the characteristics of patients consulting for chronic pain, with or without a confirmed diagnosis of endometriosis. Chronic pain is a central feature (dysmenorrhea: 51.9%, dyspareunia: 71%, non-menstrual pelvic pain: 43.0%), significantly impacting quality of life: chronic fatigue (88.2%), taking painkillers (88.5%), limiting movements (75.3%), stopping professional activity (56.4%) or school absence (61.2%) are frequently affected. The underestimation of symptoms such as dysmenorrhea remains a key contributor to delayed endometriosis diagnosis. Often perceived as a normal menstrual experience, dysmenorrhea is frequently overlooked by both patients and healthcare providers. This contributes to prolonged diagnostic delays and may worsen clinical outcomes. Promoting greater clinical awareness of the pathological significance of such symptoms is essential to improving early detection and care. These findings highlight the need for a multidisciplinary approach that includes pain management, fertility preservation strategies, psychological support and supportive care. They underscore the need for comprehensive evaluation and personalized care in managing suspected endometriosis cases. One of the most striking results is the high rate of diagnostic modifications following SMOs. Over half (51.1%) of the cases reviewed led to significant changes in the initial diagnosis, emphasizing the challenges inherent in accurately diagnosing endometriosis and the importance of specialized expertise. Additionally, SMOs markedly improved diagnostic precision. Initially, nearly a quarter (24.7%) of cases lacked a definitive diagnosis, but this uncertainty was reduced to just 5.0% after SMOs. To address the diagnostic uncertainty observed in these 5% of cases, patients could be invited for a face-to-face consultation with a specialist, allowing for a thorough clinical examination likely to provide essential information not captured during the remote assessment. This improvement underscores the ability of SMO providers to clarify complex cases and provide more conclusive answers for patients. To do this, radiologists can rely on guidelines developed by a focused panel of experts to elaborate a lexicon of image interpretation, and a standardized region-based reporting of deep infiltrating endometriosis with magnetic resonance imaging [ 18 ]. However, it is important to acknowledge that, given the absence of direct clinical examination or surgical confirmation, there remains a potential risk of misclassification. This limitation is particularly relevant for subtle lesions or cases of peritoneal-only endometriosis, where imaging and patient-reported symptoms may not fully capture disease extent, possibly leading to underdiagnosis or misinterpretation of lesion severity. In addition, diagnostic delays are notable, averaging seven years from symptom onset to diagnosis wich is in accordance with previous studies. The mean time to diagnosis, thanks to SMO, was 45 months ± 18 months after the onset of symptoms, i.e. 3.8 ± 1.5 years. This shorter delay compared with literature reports may be explained by the fact that these patients are proactive in their care and were often already engaged in a diagnostic pathway before the SMO. The diagnostic timeline could be shortened by involving radiologists with expertise in endometriosis at an earlier stage. Furthermore, as diagnostic difficulties frequently originate during adolescence, expanding access to SMOs for younger patients may support earlier identification of the disease and contribute to improved long-term clinical outcomes. Moreover, once the request has been submitted by the patient, experts respond on average within 3.6 days, which is very short given the time it takes to consult a specialist in person. Those results suggest that integrating SMOs into the diagnostic pathway could significantly enhance patient care, particularly in complex cases where initial diagnoses may be uncertain or incomplete. The study also highlights specific diagnostic challenges, particularly when imaging fails to detect peritoneal isolated endometriosis. In adolescents, diagnosis remains especially difficult due to the limitations of pelvic examination and transvaginal imaging, even though SMOs may support earlier identification. In such complex cases, emerging non-invasive tools such as saliva-based tests could, in the future, complement SMOs by helping reduce diagnostic uncertainty. The study also sheds light on specific diagnostic challenges, particularly in cases where imaging fails to identify peritoneal isolated endometriosis. In such complex cases, emerging non-invasive tools such as saliva-based tests could, in the future, complement SMOs by helping reduce diagnostic uncertainty [ 19 ]. This approach could be particularly beneficial for the percentage (5.0%) of patients who remain undiagnosed even after SMOs. The authors argue that these two approaches—expert imaging analysis and advanced molecular diagnostics—are complementary. While imaging by radiologists specialized in endometriosis is often sufficient to confirm a diagnosis and determine lesion localization for deep endometriosis and endometrioma, saliva-based testing could serve as a supportive tool to offers an accessible alternative for challenging cases or early detection. Finally, Benbassat et al., discussed the significance of obtaining a SMO and its implications for healthcare inequalities [ 20 ]. Indeed, studies reveal frequent discrepancies between expert interpretations of imaging, histopathology, and clinical assessments, with second opinions often differing from initial evaluations. These differences can lead to changes in diagnosis or treatment, highlighting the value of second opinions in improving patient care and reducing both over- and under-treatment. Despite the benefits, access to second opinions is uneven. Older and less-educated patients are less likely to be informed about this option, contributing to healthcare disparities. Surveys indicate that many patients seek second opinions due to doubts about their diagnosis or dissatisfaction with the initial consultation, with a significant proportion preferring the second opinion [ 21 ]. In the present study, endometriosis emerges as the primary concern among women’s health requests on the SMO platform. The platform addresses three major issues: diagnosis, hormonal treatment, and surgical indications. Requests for these services are evenly distributed across France, ensuring coverage in rural areas and facilitating access to specialized care nationwide. This equitable distribution is crucial for improving healthcare accessibility and expertise for women suffering from endometriosis. The only limitation observed concerns the overseas territories, likely due to insufficient awareness about the availability of the service or inadequate communication with healthcare professionals. These regions face persistent healthcare inequalities, including a shortage of specialists, poorly organized medical structures, and limited access to care in remote areas. Geographic isolation and resource disparities exacerbate these issues, as highlighted in reports emphasizing the need for innovative solutions like telemedicine and decentralized care centers. Improved communication strategies targeting healthcare providers in these territories are essential to bridge the gap and ensure equitable access to expertise across all regions. The digital aspect of SMOs, providing facility to reach in distance radiologists specialized in endometriosis, makes it possible to overcome geographical and disability constraints. Some of this study’s limitations should be highlighted. Firstly, the criteria defining an endometriosis expert within a single private healthcare system such as deuxiemeavis.fr are not clearly established in the international literature. Although stringent criteria were applied to reduce the risk of selecting non-experts, the 26 appointed specialists still represent a relatively large group, which may affect the reproducibility of their conclusions. The digital nature of the SMO platform may also introduce selection biases. Patients with limited digital access could be underrepresented, while those experiencing more severe pain or a strong suspicion of endometriosis are more likely to seek a second opinion. While findings appear consistent in the French context, caution is needed when extrapolating to other healthcare systems, where care pathways, specialist access, and radiologist expertise vary. In settings with limited access to expert imaging or multidisciplinary teams, SMOs might add greater value, whereas in structured referral networks, their impact could differ. Evaluating SMO implementation across diverse systems would help assess broader applicability. Similarly, an endometriosis diagnosis is defined in several ways across literature. Each definition specifying a different measure of the diagnostic timeframe, as well as a different degree of certainty, ranging from the suspicion of endometriosis to a surgically confirmed diagnosis [ 3 ]. These variations influence our understanding of the disease itself, as well as our assessment of medical ignorance or progress in treating it. Moreover, data were lacking to provide a clearer understanding of patient profiles, particularly regarding the number of practitioners consulted prior to the SMO. Additionally, the study relies on self-reported prior diagnoses, which may introduce reporting bias and limit the accuracy of the baseline diagnostic information, as patients’ recollections or interpretations of previous assessments could be incomplete or imprecise. While primary verification of diagnostic was not accessible, there is a remaining risk of bias that could explain part of the large number of diagnostic discrepancies observed. Furthermore, the efficacy of hormonal therapy was unknown in around 33% of patients. This information is pivotal, as hormonal therapy is often the first line of treatment for endometriosis and can significantly influence diagnostic assessments. Another limitation is the absence of a standardized diagnostic protocol or staging system (rASRM or ENZIAN), which may introduce variability in lesion interpretation and affect the consistency and comparability of diagnostic conclusions. There is also a potential risk of overconfidence in remote imaging interpretation, as diagnoses rely solely on MRI review without direct clinical examination, which may occasionally result in misclassification or overlooked subtle lesions. Lastly, follow-up is crucial for managing patients diagnosed with endometriosis via the SMO process, as it enables continuous assessment of treatment response and timely adjustment of therapeutic strategies. However, the current SMO approach lacks a standardized, systematic long-term follow-up, which may hinder effective monitoring of disease progression and optimal care delivery. Implementing structured follow-up protocols within the SMO framework is therefore essential to improve patient outcomes and to evaluate the durability of diagnostic modifications and the overall impact of SMOs on care.

Endometriosis is a chronic gynecological condition affecting approximately 10% of women of reproductive age worldwide, impacting an estimated 190 million women globally [ 1 ]. In France, between 1.5 and 2.5 million women, or 1 in 10 women of reproductive age, are affected by this disease [ 2 ]. Recent studies have revealed that the average diagnostic delay for endometriosis ranges from 5 to 12 years internationally [ 3 ], measured from the onset of symptoms to confirmed diagnosis. However, this delay can vary significantly depending on several factors, including the type of endometriosis, patient profile, physician specialty, and healthcare system differences [ 4 ]. The prolonged diagnostic delay, coupled with non-specific symptoms and the absence of definitive biomarkers, not only contributes to the emotional burden and diminished quality of life in affected women, but also leads to increased medical consultations, strained personal relationships, social withdrawal, and prolonged use of pain medications, including opioids [ 5 ]. Furthermore, endometriosis patients face increased risks of mental health burdens, with anxiety prevalence ranging from 10% to 87.5% compared to 6% in the general population, and depression rates of 9.8% to 98.5% compared to 6.6% to 9.3% in controls [ 6 ]. Precise questionnaires assessing detailed patient symptoms have proven to be crucial in helping experts diagnose endometriosis [ 7 – 9 ]. These tools enable a comprehensive understanding of the patient’s condition, often leading to more accurate conclusions. Over the past decade, SMOs have emerged as a valuable tool in healthcare, offering significant benefits to both patients and healthcare systems. SMOs provide access to additional expertise, potentially improving treatment trajectories and outcomes, particularly in complex medical cases where conventional approaches have failed. This includes improvements in diagnosis, treatment, and patient satisfaction [ 10 – 12 ]. The advent of online SMO platforms has increased accessibility to physicians, allowing patients to seek opinions from a broad network of specialists without geographical constraints. Sanchez et al. recently focused on predictive factors that influence diagnostic and therapeutic divergence in SMOs, in order to understand their impact on healthcare decision-making processes suggesting its value to solve diagnostic discrepancies and therapeutic divergence [ 13 ]. In a field as complex as endometriosis, characterized by diagnostic challenges and varied treatment options, SMO is hypothesized to be an interesting component of comprehensive care. Especially, in the context of internet-based applications like ChatGPT (Open AI, San Francisco, CA USA) which have demonstrated the ability to accurately and satisfactorily respond to over 90% of patient questions [ 14 ], the apparent value, of SMOs and the physician expertise remain under-assessed as a complementary solution. To our knowledge no formal studies have been reported on this topic to date, indicating a significant gap in research that warrants further investigation. Therefore, this study aimed to assess the effectiveness of a diagnostic strategy for patients seeking a second medical opinion for endometriosis, using a database from a specialized French network dedicated to SMOs.

Supplementary Material 1. Additional file 1: Baseline population characteristics after SMOs. Additional file 2 : Patient pain profile after SMOs Supplementary Material 1. Additional file 1: Baseline population characteristics after SMOs. Additional file 2 : Patient pain profile after SMOs