Between July 2021 and July 2024, 51 patients with endometriosis who received treatment at Xiamen Maternal and Child Health Care Hospital were selected for the observation group. The ages of the patients ranged from 26 to 54 years, with all cases confirmed by histopathological examination. Inclusion criteria required that patients who had laparoscopic surgery and were pathologically confirmed to have endometriosis. Exclusion criteria comprised patients with missing clinical data, those with hormone-dependent diseases (such as adenomyosis and uterine fibroids), individuals taking any hormonal medications, those with a history of pregnancy within the past six months, patients with other endocrine, immune, or metabolic diseases, and those with severe liver and kidney dysfunction, hematological disorders, or malignant tumors (Fig. 1). Furthermore, 52 healthy women who underwent physical examinations at the same hospital during the same period were selected as the control group (Fig. 1), aged between 25 and 60 years. There was no statistically significant difference in age between the two groups (p $>$ 0.05). This study has been approved by the Ethics Committee of Women and Children’s Hospital, School of Medicine, Xiamen University (KY-2025-104-K01).

Fig. 1.

Flowchart of study population inclusion and exclusion criteria. n, number of samples.

For serum tumor marker testing, 4 mL of peripheral venous blood was drawn from patients in the observation group preoperatively during non-menstrual periods and under fasting conditions in the morning. Similarly, the control group had 4 mL of peripheral venous blood drawn on the morning of the physical examination, also while fasting. After standing at room temperature for 30 minutes, the samples were centrifuged at 3000 rpm for 10 minutes. The serum was collected and stored in sterile EP tubes, which were then stored at –80 °C until testing. The serum (20153401967, Innovax, Xiamen, Fujian, China) and HE4 (20163400698, Innovax, Xiamen, Fujian, China) expression levels of both groups were measured using a chemiluminescence analyzer (I2000, Abbott, Lake Forest, IL, USA), with all testing procedures strictly adhering to the kit instructions.

The ultrasound examination was performed by multiple operators using the Philips (IU22, Amsterdam, NH, Netherlands), Aloka (ProSound Alpha 5, Hitachi Aloka, Tokyo, Japan) and GE (Voluson E8, Boston, MA, USA) color Doppler ultrasound diagnostic system, with meticulous observation of the patient’s pelvic area, including the uterine and ovarian spaces. Key aspects assessed include the thickness of the endometrium, uterine size, internal echo, and the position of the uterus. Additionally, detailed observations and recordings were made regarding the lesion’s location, marginal morphology, diameter, and whether a capsule was present. The blood flow signal distribution in the targeted lesion was also examined. The diagnosis of endometriosis at various sites was made using established clinical diagnostic criteria.

This study aims to evaluate the positive outcomes associated with endometriosis diagnosed through ultrasound using statistical methods. We compared and analyzed the serum levels of CA125 and HE4 between two groups of subjects. Additionally, we considered the histopathological examination results from surgery as the gold standard for diagnosing endometriosis. The diagnostic efficacy of ultrasound, CA125, and HE4 was statistically analyzed and compared both individually and in combination. Specificity was defined as the proportion of true negative cases to the sum of true negative and false positive cases, expressed as a percentage. Sensitivity was calculated as the ratio of true positive cases to the sum of true positive and false negative cases, also expressed as a percentage. Accuracy was calculated by dividing the sum of true positive and true negative cases by the total number of cases, then multiplying by 100%.

The sample size requirement was determined using PASS (version 2021, NCSS, Kaysville, UT, USA), specify the test direction, report power = 90%, anticipated AUC of 0.9, null AUC of 0.7, allocation ratio was set at 1:1, the lower limit of the false positive rate was 0.0, and the upper limit of the false positive rate was 1.0, resulting in a need for a minimum of 24 patients with endometriosis and 24 healthy controls. Ultimately, the study included a total of 51 patients with endometriosis and 52 controls. All statistical analyses were conducted using SPSS 27.0 software (IBM Corp., Armonk, NY, USA). The Shapiro-Wilk test was employed to analyze the type of data distribution, and the parameters distributed with normal distribution were explained as Mean $\pm$ SD. The t test was used when parameters were distributed with normally distribution, The Mann-Whitney U test was used when parameters distributed with non-normal distribution. Pearson’s $\chi$² test and Fisher’s exact test were applied to analyze the correlation of categorical data derived from ultrasound detection indicators. The Pearson’s $\chi$² test was used when all expected frequencies t $\ge$ 5 and the total number of cases N $\ge$ 40. The Fisher’s exact test was used when the expected frequency t 5 or the total number of cases N 40. Binary logistic regression analysis was utilized to construct the diagnostic model. We employed Prism 10 to plot the receiver operating characteristic (ROC) curves and calculate the AUC to evaluate the diagnostic efficiency of various diagnostic indicators. Statistical significance was set at p 0.05.

A total of 51 patients with histologically and pathologically confirmed endometriosis and 52 age-matched controls were enrolled for analysis. Comparative analysis of baseline characteristics demonstrated that the mean age of the endometriosis group was slightly higher than that of the control group (38.20 $\pm$ 7.30 years vs. 35.60 $\pm$ 7.50 years). Notably, serum CA125 levels were markedly elevated in endometriosis patients compared to controls (1.90 $\pm$ 0.39, log10 vs. 1.06 $\pm$ 0.22, log10). On the contrary, HE4 levels were significantly higher in the control group (1.67 $\pm$ 0.07 vs. 1.62 $\pm$ 0.17) (Table 1).

Analysis of endometriotic lesion distribution revealed that the most common isolated lesion type was ovarian endometriosis (25 cases, 49.02%), followed by ovarian combined with pelvic endometriosis (11 cases, 21.57%). Other isolated lesion types included pelvic endometriosis (2 cases, 3.92%), abdominal wall endometriosis (1 case, 1.96%), and fallopian tube endometriosis (1 case, 1.96%). Additionally, several multifocal lesion combinations were observed: ovarian with fallopian tube involvement (2 cases, 3.92%), ovarian with uterosacral ligament involvement (1 case, 1.96%), and ovarian with rectal involvement (1 case, 1.96%). Furthermore, seven cases (13.72%) exhibited triple-site involvement, including combinations of ovarian-cervical-pelvic, ovarian-vaginal wall-pelvic, ovarian-fallopian tube-pelvic, ovarian-fallopian tube-vaginal wall, ovarian-rectal-abdominal wall, ovarian-rectal-vaginal, and ovarian-pelvic-intestinal endometriosis (Table 2).

We analyzed the serum tumor marker expression levels between the endometriosis group and the control group. The results showed that the CA125 expression level in the endometriosis group (1.90 $\pm$ 0.39, log10) was significantly higher compared to the control group (1.06 $\pm$ 0.22, log10) (p 0.001) (Fig. 2A, Table 3). However, there was no significant difference in the HE4 expression levels between the endometriosis group (1.62 $\pm$ 0.17, log10) and the control group (1.67 $\pm$ 0.07, log10) (Fig. 2B, Table 3).

Fig. 2.

Comparative analysis of serum tumor marker levels between endometriosis patients and healthy controls. (A) Carbohydrate antigen 125 (CA125) levels in controls versus endometriosis patients. (B) HE4 levels in controls versus endometriosis patients. (C) CA125 levels across study groups: controls, isolated ovarian endometriosis, and ovarian endometriosis with pelvic involvement. (D) HE4 levels across study groups: controls, isolated ovarian endometriosis, and ovarian endometriosis with pelvic involvement. * represents p 0.05, ** represents p 0.01, **** represents p 0.0001; ns, not significant.

To further elucidate the characteristics of endometriosis based on different sites of disease onset, this study classified patients according to their disease location and focused on comparing the levels of CA125 and HE4 in the two most common subtypes—ovarian endometriosis and ovarian endometriosis combined with pelvic endometriosis—with those of the control group. The results showed that the CA125 levels in the control group were statistically different from those in both the ovarian endometriosis group and the ovarian endometriosis combined with pelvic endometriosis group. However, there was no statistically significant difference in CA125 levels between the ovarian endometriosis group and the ovarian endometriosis combined with pelvic endometriosis group (Fig. 2C, Table 3). Additionally, the HE4 levels in the control group were statistically different from those in the ovarian endometriosis group, but not from those in the ovarian endometriosis combined with pelvic endometriosis group. Notably, there was a statistically significant difference in HE4 levels between the ovarian endometriosis group and the ovarian endometriosis combined with pelvic endometriosis group (Fig. 2D, Table 3).

We further analyzed the individual diagnostic efficacy of CA125 and HE4 serum markers in the diagnosis of endometriosis. To this end, we used a healthy control group as a reference and drew ROC curves to evaluate the diagnostic accuracy of these two markers. The results showed that the AUC value of the CA125 diagnostic model was 0.97 (p 0.001, 95% CI: 0.94–1.00) (Fig. 3A, Table 4), indicating that it has very high diagnostic accuracy. In contrast, the AUC value of the HE4 diagnostic model was 0.58 (p = 0.156, 95% CI: 0.47–0.70) (Fig. 3B, Table 4), close to the level of random guessing, suggesting that its value in diagnosing endometriosis is limited.

Fig. 3.

Analysis of the diagnostic value of CA125 and HE4 for endometriosis. (A) ROC curve analysis of the diagnostic performance of CA125 for endometriosis. (B) ROC curve analysis of the diagnostic performance of HE4 for endometriosis. (C) Sensitivity, specificity, and accuracy of diagnosing endometriosis through CA125. (D) Sensitivity, specificity, and accuracy of diagnosing endometriosis through HE4. AUC, area under the curve; 95% CI, 95% confidence interval; PPV, positive predictive value; NPV, negative predictive value; ROC, receiver operating characteristic.

Furthermore, through the maximum Youden index, we calculated the sensitivity, specificity, and accuracy of CA125 and HE4 in the diagnosis of endometriosis. CA125 had a sensitivity of 94.23% (95% CI: 0.87–1.01), specificity of 92.16% (95% CI: 0.85–1.00), accuracy of 93.20% (95% CI: 0.88–0.98), positive predictive value (PPV) of 92.18% (95% CI: 0.85–1.00), and negative predictive value (NPV) of 94.21% (95% CI: 0.87–1.01) (Fig. 3C, Table 4). This means that CA125 can effectively identify the vast majority of actual patients and accurately exclude non-patients, with high overall diagnostic accuracy. HE4 had a sensitivity of 98.08% (95% CI: 0.94–1.02), specificity of 35.29% (95% CI: 0.21–0.48), accuracy of 66.38% (95% CI: 0.57–0.75), PPV of 59.78% (95% CI: 0.49–0.70), and NPV of 94.93% (95% CI: 0.84–1.06) (Fig. 3D, Table 4). Although HE4 has high sensitivity, meaning that it can detect most actual patients, its low specificity indicates that there may be more false-positive results, i.e., misdiagnosing some non-patients as patients. This could lead to unnecessary further examinations and treatments. Therefore, based on the AUC values and various diagnostic indicators, CA125 performs better than HE4 in the diagnosis of endometriosis.

We used pathological diagnosis as the gold standard to evaluate the diagnostic value of ultrasound in endometriosis. A total of 51 patients diagnosed with endometriosis and 52 controls were included in the study. Ultrasound examination results showed that among the 51 patients, 14 were missed (ultrasound did not detect the disease), while in the 52 controls, ultrasound did not detect any abnormalities.

To explore the diagnostic value of ultrasound in endometriosis, we used the Chi-Square test to analyze the association between ultrasound results and disease status. The results showed that ultrasound could significantly distinguish patients with endometriosis from controls (Chi-Square test; $\chi$² = 58.875; p 0.01) (Fig. 4A), indicating that ultrasound has statistical significance in diagnosing endometriosis.

Fig. 4.

Analysis of the diagnostic value of ultrasound for endometriosis. (A) Pearson’s Chi-Square test analysis of whether ultrasound can differentiate between patients with endometriosis and controls. (B) Sensitivity, specificity, and accuracy of ultrasound in diagnosing endometriosis.

Furthermore, through the maximum Youden index, we calculated the sensitivity, specificity, accuracy, PPV, and NPV of ultrasound detection to be 72.55% (95% CI: 0.60–0.85), 100.00% (95% CI: 1.00–1.00), 86.41% (95% CI: 0.80–0.93), 100.00% (95% CI: 1.00–1.00), and 78.79% (95% CI: 0.69–0.89), respectively (Fig. 4B, Table 5). These metrics indicate that ultrasound detection has high specificity and accuracy in the diagnosis of endometriosis, although the sensitivity is slightly lower, meaning that a certain proportion of patients may be missed. However, the 100% specificity means that ultrasound can very reliably exclude those without the disease, which helps reduce unnecessary further examinations. In conclusion, ultrasound detection has important clinical value in the diagnosis of endometriosis.

Previous studies have shown that serum tumor marker CA125 and ultrasound imaging have certain diagnostic value in the diagnosis of endometriosis. However, the diagnostic accuracy of these single-detection methods still requires improvement. Therefore, this part of the study aims to explore whether the combination of CA125 and ultrasound detection can improve the diagnostic accuracy.

To achieve this goal, we used logistic regression model to combine the numerical values of CA125 and ultrasound detection results to establish a combined diagnostic model and evaluated its diagnostic performance by drawing ROC curves (Fig. 5A, Table 6). The results showed that the AUC of the ROC curve for the combined diagnostic model was 0.96 (95% CI: 0.92–1.00), indicating that it has high accuracy in distinguishing patients with endometriosis from non-patients. Further analysis through the maximum Youden index showed that the sensitivity, specificity, accuracy, PPV, and NPV of the combined diagnostic model were 92.16% (95% CI: 0.85–1.00), 98.08% (95% CI: 0.94–1.01), 95.15% (95% CI: 0.91–1.00), 97.92% (95% CI: 0.94–1.02), and 92.73% (95% CI: 0.86–1.00), respectively. These metrics indicate that the model can not only effectively identify true patients (high sensitivity) but also accurately exclude non-patients (high specificity), with high overall diagnostic accuracy.

Fig. 5.

Analysis of the diagnostic value of combined CA125 and ultrasound for endometriosis. (A) ROC curve analysis of the diagnostic performance of combined CA125 and ultrasound for endometriosis. (B) Sensitivity, specificity, and accuracy of combined CA125 and ultrasound in diagnosing endometriosis.

Notably, compared with the single use of CA125 or ultrasound detection, the accuracy of the combined diagnostic model increased by 2.05% and 8.84%, respectively (Fig. 5B, Table 6). These results indicate that the combined application of CA125 and ultrasound detection outperforms single detection methods in the diagnosis of endometriosis. This may provide an effective approach for earlier and more accurate diagnosis of endometriosis in clinical practice.

This study has several limitations that should be considered. The modest sample size (n = 103) restricted in-depth subgroup analyses, particularly for rare lesion subtypes, and may have increased the risk of selection bias. Furthermore, although healthy controls were included, the absence of pathological confirmation means that asymptomatic endometriosis cannot be definitively ruled out in the control group. Future studies with larger, prospectively recruited cohorts and gold-standard verification for all participants are needed to validate these findings and facilitate the translation of this combined diagnostic strategy into routine clinical practice.

In summary, this study developed a combined diagnostic model integrating CA125 and ultrasound detection, offering a novel method for the precise non-invasive diagnosis of endometriosis. To strengthen clinical applicability, future efforts should focus on multi-center validation using larger, prospective cohorts. Further refinement of diagnostic accuracy could also be achieved by incorporating additional novel biomarkers alongside advanced ultrasound characteristics.