Development and validation of a nomogram model for predicting natural pregnancy after hysteroscopy and laparoscopy in patients with tubal infertility complicated with pelvic endometriosis

This study shows that age is an important independent risk factor for predicting postoperative natural pregnancy (OR = 1.294, P  < .001). This result is consistent with the consensus in the field of reproductive medicine: the ovarian reserve function of women decreases with age, especially after 35 years old; the number and quality of oocytes decrease; and the fertilization rate and embryo implantation rate decrease significantly. [ 4 , 5 ] For patients with endometriosis, age-related ovarian function decline may be superimposed with the chronic damage of endometriosis itself to ovarian tissue, further weakening the fertility potential. [ 6 – 9 ] In clinical practice, for elderly patients (such as ≥ 35 years old), ovarian reserve (such as anti-Mullerian hormone AMH) should be evaluated more actively after operation, and pregnancy assistance plans should be made as soon as possible in combination with the prediction results of the model to avoid delaying the best pregnancy opportunity. Prolonging the period of infertility will significantly reduce the probability of natural pregnancy after operation (OR = 1.065, P  < .01), suggesting that the duration of the disease is closely related to the reproductive prognosis. Long-term infertility may reflect the persistent destruction of pelvic microenvironment by endometriosis lesions (such as increased pelvic adhesion and impaired tubal function), accompanied by indirect effects such as psychological pressure accumulation and endocrine disorders. [ 10 ] In addition, patients who have been infertile for a long time may have unidentified combined factors (such as the decline of male semen quality), so it is necessary to conduct a comprehensive fertility assessment before operation to avoid relying solely on surgery to improve the prognosis. For patients who have been infertile for a long time, even if the anatomical structure is restored after operation, we should be alert to the possibility of low natural pregnancy rate and timely combine assisted reproductive technology (such as intrauterine insemination IUI or in vitro fertilization-embryo transfer IVF-ET) to improve the pregnancy success rate. [ 11 ] In this study, the increase of ASRM stage (that is, the enlargement of the focus of endometriosis and the aggravation of adhesion) is an independent risk factor for postoperative natural pregnancy (OR = 1.773, P  < .05). ASRM staging, as the most commonly used clinical grading standard of endometriosis, is mainly based on the location, size, and degree of pelvic adhesion, which directly reflects the anatomical severity of the disease. Patients in stages Ⅲ and Ⅳ are often accompanied by adhesion at the umbrella end of fallopian tube and extensive pelvic adhesion. Even if the adhesion is released and the focus is removed by hysteroscopy and laparoscopy, the pregnancy may be affected by irreversible damage of tubal peristalsis function or persistent abnormality of pelvic microenvironment (such as inflammatory factors and oxidative stress level). [ 12 ] This suggests that clinicians should thoroughly remove the focus and restore the pelvic anatomical structure during the operation, at the same time, strengthen postoperative follow-up for patients with high stages and intervene as soon as possible in combination with the predicted results of the model to avoid waiting for natural pregnancy excessively. The increase in CA125 level (OR = 2.431, P  < .01) was identified as the strongest independent risk factor in this study, suggesting that CA125 plays a significant role in predicting postoperative pregnancy prognosis. As a well-known serum marker of endometriosis, CA125 levels are closely associated with the activity and inflammatory status of ectopic lesions. When the lesions are active, peritoneal mesothelial cells and ectopic endometrial glands are stimulated to secrete more CA125. Elevated CA125 may contribute to infertility by inducing peritoneal immune dysregulation (e.g., overactivation of macrophages, increased cytokine secretion) and by impairing endometrial receptivity through altered expression of integrins and adhesion molecules. [ 13 ] However, CA125 is a nonspecific biomarker, and its elevation can also occur in various benign and inflammatory conditions, such as pelvic inflammatory disease, adenomyosis, and even during menstruation, as well as in malignant tumors like ovarian cancer. These confounding influences limit its specificity as a stand-alone predictive factor. Therefore, the interpretation of CA125 levels should always be combined with comprehensive clinical and imaging findings. Clinically, preoperative CA125 can still serve as a convenient and sensitive indicator for assessing endometriosis activity and inflammation. For patients with markedly elevated CA125 levels, short-term postoperative GnRH-a therapy may help suppress lesion recurrence, improve the pelvic inflammatory microenvironment, and increase the likelihood of natural conception. Future mechanistic studies focusing on the immunological and molecular pathways linking CA125 to impaired fertility are warranted to clarify its biological role and enhance its clinical interpretability. Focus on the specific population of “tubal infertility complicated with pelvic endometriosis” for the first time. Previous studies mostly discussed the pregnancy prediction of endometriosis or tubal infertility separately, but the model was built for the complicated situation of the 2 diseases, which is more suitable for the clinical diagnosis and treatment needs of such patients. Four independent influencing factors, namely, female age, infertility years, ASRM stage (endometriosis stage), and CA125 (tumor marker), were integrated, and the basic characteristics of patients (age, infertility years), disease severity (ASRM stage), and biological indicators (CA125) were taken into account, thus realizing multi-dimensional accurate prediction of pregnancy outcome. Least absolute shrinkage and selection operator algorithm is used to screen key variables, and multivariate logistic regression is used to determine independent influencing factors, and then ROC curve, calibration curve, and decision curve analysis curve are used to comprehensively verify the distinguishing ability, accuracy, and clinical value of the model, which is more scientific and reliable than traditional single factor analysis or simple regression model. Previous studies have proposed predictive models for endometriosis-related pregnancy outcomes, but most of them focused on general endometriosis populations or assisted reproductive technology (ART) outcomes rather than natural conception. For instance, Guo et al. (Front Med, 2020) [ 9 ] developed preoperative nomograms to predict the presence and severity of endometriosis in infertile women, but these models did not assess postoperative natural fertility potential. Compared with those models, our nomogram specifically targets tubal infertility complicated with pelvic endometriosis and predicts natural pregnancy after hysteroscopy and laparoscopy. By integrating anatomical (ASRM stage and infertility duration) and biochemical (CA125) factors, this model provides a more practical and individualized tool for postoperative fertility evaluation. Future multicenter comparative studies are warranted to further validate the external applicability and comparative performance of this model. The model can quantitatively predict the postoperative natural pregnancy probability by inputting the patient’s age, infertility years, ASRM stage, and preoperative CA125 level, which helps doctors objectively judge the prognosis of patients and avoid the evaluation bias caused by subjective experience. For patients with low pregnancy prediction probability, intervention measures such as assisted reproductive technology (such as IVF) can be suggested in advance; for patients with high probability, expectant treatment can be given, and follow-up plans can be made to reduce unnecessary medical intervention and optimize the diagnosis and treatment path. The clinical decision-making curve (decision curve analysis) shows that the model can bring significant net benefits in a wide threshold probability range (019–0.65 in the model group and 0.12–0.68 in the verification group), which shows that it has practical application value in different clinical scenarios and can help doctors to quickly formulate reasonable treatment plans and improve the pregnancy outcome of patients. It is convenient for clinical popularization: the indicators included in the model are all routine clinical detection items (age, medical history, ASRM stage, CA125), which are convenient to obtain and can be easily applied in primary hospitals without additional complicated examination, with wide applicability. This hierarchical management not only conforms to the concept of “individualized medical care” but also optimizes the allocation of medical resources. Our standardized surgical and laboratory protocols ensure that this nomogram can be readily adopted by other centers using similar techniques. This study has several limitations. As all participants were recruited from a single center – Beijing Shijitan Hospital – the findings may be affected by selection bias. The model’s performance and generalizability in hospitals from other regions or with different surgical expertise levels have not yet been validated. Therefore, external validation using multicenter cohorts and diverse medical settings is necessary to ensure its broader applicability and robustness. Potential influencing factors such as lifestyle variables (e.g., BMI, smoking history) and genetic polymorphisms were not included, which may reduce the comprehensiveness of the model. The follow-up duration was limited to 12 months; long-term reproductive outcomes such as cumulative pregnancy rate and live birth rate were not assessed. Future studies should aim to conduct prospective, multicenter investigations to externally validate and refine this nomogram, incorporating broader demographic, biological, and environmental factors to enhance its predictive power and clinical utility across various populations.

In this study, the nomogram model of natural pregnancy after hysteroscopy and laparoscopy in patients with tubal infertility and pelvic endometriosis was constructed and verified, and it was clear that age, infertility years, ASRM stage, and preoperative CA125 were independent influencing factors. The model is easy to operate and has good prediction efficiency, which can provide an objective prognostic evaluation tool for clinic and help to formulate personalized diagnosis and treatment strategies. In the future, it is necessary to further improve the model through multi-center research and explore its application in clinical transformation and finally improve the reproductive outcome of such patients.

Tubal infertility and pelvic endometriosis are common causes of female infertility, and the incidence of their coexistence is high, reaching as high as 25% to 50% among infertile women, which seriously affects reproductive function. [ 1 , 2 ] As the main method to treat this disease at present, hysteroscopy combined with laparoscopy can improve pregnancy conditions by removing tubal adhesion and removing ectopic endometrial lesions, but there are still significant individual differences in pregnancy outcome after operation. Accurate prediction of postoperative natural pregnancy rate is of great significance for formulating individualized treatment strategies and optimizing fertility guidance. Nomogram, as a visual prediction model, can transform the results of multi-factor analysis into intuitive charts and provide quantitative reference for clinical decision-making. [ 3 ] At present, there is still a blank in the nomogram study of postoperative pregnancy outcome of patients with tubal infertility complicated with pelvic endometriosis. The purpose of this study is to construct and verify the nomogram model of natural pregnancy after hysteroscopy and laparoscopy in such patients and to provide scientific basis for clinical practice.

Conceptualization: Cong Yin, Wenpei Bai. Data curation: Cong Yin, Wenpei Bai. Formal analysis: Cong Yin, Wenpei Bai. Funding acquisition: Wenpei Bai. Investigation: Cong Yin, Wenpei Bai. Methodology: Cong Yin, Wenpei Bai. Supervision: Cong Yin, Wenpei Bai. Validation: Cong Yin, Wenpei Bai. Visualization: Cong Yin, Wenpei Bai. Writing – original draft: Cong Yin, Wenpei Bai. Writing – review & editing: Cong Yin, Wenpei Bai.

The clinical data of patients with tubal infertility complicated with pelvic endometriosis who underwent hysteroscopy combined with laparoscopy in the gynecology department of Beijing Shijitan Hospital, Capital Medical University, from January 2018 to January 2023 were retrospectively collected. Five hundred and twenty-three patients were randomly divided into model group (n = 366) and verification group (n = 157) according to the ratio of 7:3 to ensure that the baseline data of the 2 groups were balanced and comparable. Inclusion criteria: It meets the diagnostic criteria of female infertility of the World Health Organization (WHO); normal menstruation, diagnosed as tubal infertility by clinical diagnosis, hysterosalpingography, or laparoscopy; pelvic endometriosis was diagnosed during operation; and the postoperative follow-up time was ≥ 12 months when GnRH-a was stopped to treat. Exclusion criteria: Combined with other serious diseases of reproductive system (such as congenital uterine malformation, hysteromyoma, premature ovarian failure, ovarian endometriosis, intrauterine adhesions, etc.); receiving assisted reproductive technology (ART) after operation; and lack of clinical data. Finally, 523 patients were included. The research pathway is illustrated in Figure 1 . This study was examined and approved by the Ethics Committee of Beijing Shijitan Hospital, Capital Medical University [Ethical approval number: Sjtkyll-LX-2023 (062)]. Research pathway diagram. AMH = anti-Müllerian hormone, BMI = body mass index, CA125 = cancer antigen 125. Baseline data of patients were collected, including age, body mass index (BMI), infertility time (months), infertility type (primary infertility/secondary infertility), pregnancy times, anti-Mullerian hormone (AMH) level, CA125, whether there is a history of pelvic inflammatory disease, tubal patency before operation, endometriosis staging (r-ASRM staging of American Society of Reproductive Medicine), pelvic adhesion during operation, endometrial polyps, tubal patency after operation (assessed by laparoscopic fluid drainage), etc. Previous history of lower abdominal pain with or without fever, diagnosed as pelvic inflammatory disease by doctors, and treated with antibiotics or surgery. Contrast-enhanced ultrasound of fallopian tubes in our hospital before operation revealed tubal obstruction (including left fallopian tube obstruction, right fallopian tube obstruction and bilateral fallopian tube obstruction). Contrast-enhanced ultrasound examination of fallopian tubes is completed by a senior and experienced sonographer. According to the adhesion grading standard (MCASM) improved by American Society of Reproductive Medicine (ASRM), the presence of pelvic adhesion was determined during operation. Hysteroscopic surgery found, removed and pathologically confirmed. After hysteroscopy and laparoscopy, methylene blue was introduced through the cervix, and the patency of bilateral fallopian tubes was observed under laparoscopy. Tubal patency was graded as:Obstructed: No spillage after 2 minutes or pressure > 300 mm Hg. Serum CA125 levels were measured using chemiluminescent immunoassay. AMH was quantified by ELISA. All assays were performed in duplicate, and values exceeding the coefficient of variation threshold (>15%) were re-tested. Laboratory personnel were blinded to clinical outcomes. All patients were operated within 3 to 7 days after menstruation was clean. The indications for operation are women of childbearing age who have been diagnosed with primary infertility or secondary infertility. Before operation, contrast-enhanced ultrasound of fallopian tubes in our hospital showed that the fallopian tubes were obstructed or blocked. Contrast-enhanced ultrasound examination of fallopian tubes is completed by a senior and experienced sonographer, and all fallopian tube recanalization operations are completed by the same operator, so as to ensure the uniformity of measurement of clinical indicators and the controllability of operation quality. Surgical procedures: First, laparoscopic fertility-preserving surgery, including deep invasive endometriosis lesion resection, pelvic adhesion release, tubal repair and plastic surgery, to completely remove the visible endometriosis lesions in the pelvic cavity and restore the anatomical position of the pelvic cavity as much as possible. At the same time, pay attention to the protection of ovarian function and avoid damage to normal ovarian tissue as much as possible. Hysteroscopy was performed again to exclude or remove endometrial lesions, and methylene blue was introduced through the cervix. If there was one or both fallopian tube obstruction, tubal intubation was performed under hysteroscopy. After the operation, a large amount of normal saline was used to wash the pelvic cavity. All laparoscopic procedures were performed using a standard 10-mm 30° laparoscope (Karl Storz, Germany) with CO₂ pneumoperitoneum maintained at 12 to 14 mm Hg. Pelvic adhesiolysis was achieved using bipolar forceps (Martin, Germany) and harmonic scalpel. Endometriotic lesions were excised with cold scissors for superficial implants and harmonic scalpel for deep infiltrating lesions. Tubal patency was assessed by transcervical injection of methylene blue (10 mL, 0.5% solution) under laparoscopic visualization. Patency was confirmed if dye spillage from both fimbrial ends was observed within 2 minutes of injection at a steady pressure of 150 to 200 mm Hg (controlled by a manual pressure pump). Patients with 1) ASRM stage I-II were treated with GnRH-a (3.75 mg/mo, AbbVie) for 3 months after operation; Patients with stage Ⅲ and Ⅳ were treated with drugs for 6 months. Expect to conceive naturally after stopping taking medicine, and give birth guidance to patients. The postoperative time was calculated from the time when GnRH-a was stopped for the treatment of menorrhagia, and regular follow-up was conducted by outpatient follow-up and telephone follow-up. Natural pregnancy is defined as intrauterine pregnancy, which can be diagnosed as intrauterine pregnancy if ultrasound can see intrauterine pregnancy sac and cardiac pulsation, and ectopic pregnancy must be confirmed by pathology. The natural pregnancy within 12 months was calculated from the time when GnRH-a was stopped to treat. Statistical analysis was conducted utilizing R 4.1, SPSS 26.0, GraphPad 8.0, and Stata 15.0 software packages for data processing and graphical representation. Firstly, the Kolmogorov–Smirnov test and Levene’s test were utilized to analyze the normality and homogeneity of variances in the data. Non-normally distributed continuous variables were presented as M (P25, P75), while categorical variables were expressed as n/%. The influencing factors of natural pregnancy after hysteroscopy and laparoscopy in patients with tubal infertility complicated with pelvic endometriosis were screened by univariate variable analysis. Lasso regression with ten-fold cross-validation and lambda 1se as the criterion was employed to select the optimal combination of influencing factors. Subsequently, multiple forward stepwise logistic regression was then conducted to further examine the selected factors from Lasso regression. The purpose of this analysis is to determine the independent influencing factors of natural pregnancy after hysteroscopy and laparoscopy in patients with tubal infertility complicated with endometriosis. Additionally, a nomogram for the model group was established. The sample was randomly split into model and validation groups in a 7:3 ratio. The area under the receiver operating characteristic (ROC) curve (AUC) assessed the discrimination of the model group and validation group. The Hosmer-Lemeshow test was utilized to examine the statistical difference between predicted and actual probabilities. The clinical efficacy was evaluated using clinical decision curves and clinical impact curves. P  < .05 was deemed as a statistically significant value.

There was no significant difference in baseline data such as age, infertility period, BMI, and AMH level between the model group and the verification group ( P  > .05, Table 1 ), which was comparable. Comparison of baseline data between model group and validation group. AMH = anti-Müllerian hormone, ASRM = American Society for Reproductive Medicine, CA125 = cancer antigen 125. The univariate analysis of factors influencing natural pregnancy in patients with tubal infertility complicated with pelvic endometriosis demonstrated significant associations with the following variables: Age, BMI, infertility period, ASRM stage, pelvic adhesion, history of pelvic inflammatory disease, AMH, and CA125 were related to postoperative natural pregnancy ( P  < .1, Table 2 ). Univariate analysis of inffuencing factors of natrual pregnancy. AMH = anti-Müllerian hormone, ASRM = American Society for Reproductive Medicine, CA125 = cancer antigen 125. P < .005. Lasso regression was employed to further analyze the factors mentioned above. The model identified the following as the best matching factors: age, infertility period, ASRM stage, AMH, and CA125 (a total of 5 factors), as shown in Figure 2 . Best match factor screening by lasso regression. A is the Lasso regression path diagram; B shows the plot of the best matching factors screened by the 10-fold cross validation method, and the best matching factors were selected using lambda.1se as the criterion. Through multivariable logistic regression analysis, age (odds ratio [OR] = 1.294, 95% confidence interval = 1.119–1.502, P  < .001), infertility period (OR = 1.065, 95% confidence interval = 1.020–1.112, P  < .01), ASRM stage (OR = 1.773, 95% confidence interval = 1.069–2.949, P  < .05), and CA125 (OR = 2.431, 95% confidence interval = 1.309–4.669, P  < .01) were determined as independent factors influencing the natural pregnancy rate in patients with tubal infertility complicated with pelvic endometriosis (Fig. 3 ). Forest plots of independent influencing factors for natural pregnancy by multivariate analysis univariate. AMH = anti-Müllerian hormone, ASRM = American Society for Reproductive Medicine, CA125 = cancer antigen 125. As per the findings of multivariate logistic regression, a nomogram was developed to predict the natural pregnancy in patients with tubal infertility complicated with pelvic endometriosis. In age, infertility period, ASRM stage, AMH, and CA125 can be used to determine the corresponding points on the nomogram. By summing up the points from each variable, the nomogram estimates the natural pregnancy probability for patients with tubal infertility complicated with pelvic endometriosis, as shown in Figure 4 . Nomogram of the prediction model for natural pregnancy in patients with tubal infertility complicated with pelvic endometriosis. AMH = anti-Müllerian hormone, ASRM = American Society for Reproductive Medicine, CA125 = cancer antigen 125. A random sampling was performed, with a ratio of 7:3 between the model group (n = 366) and the validation group (n = 157). The accuracy and discrimination of the nomogram were validated using ROC and calibration curves. The results showed that AUC was 0.771 (95% CI = 0.718–0.823) for the model group and 0.731 (95% CI = 0.639–0.823) for the validation group, suggesting an enhanced discrimination ability of the prediction model. The Hosmer-Lemeshow test revealed no statistically significant difference between the predicted and observed clinical pregnancy probabilities in both the model group ( R 2  = 0.268, P  = .810 > .05) and the validation group ( R 2  = 0.174, P  = .657 > .05), suggesting that the model accurately predicted natural pregnancy probability for patients with tubal infertility complicated with pelvic endometriosis (Fig. 5 ). Discriminative power and accuracy of nomogram of the prediction model for natural pregnancy in patients with tubal infertility complicated with pelvic endometriosis. A and B show the receiver operating curve of the model group and the validation group, respectively; C and D are calibration curves of the model group and the validation group, respectively. The clinical efficacy of the model was analyzed using clinical decision curve analysis. The outcomes revealed that both the model and the validation groups achieved maximum benefits at threshold probability values ranging from 0.19 to 0.65 and 0.12 to 0.68 (Fig. 6 ). Discriminative power and accuracy of nomogram of the prediction model. A and B show the clinical decision curves of the model group and the validation group, respectively.