{"paper_id":"d52e334e-a706-46f3-9583-b5ad3c9c93f7","body_text":"Clin. Exp. Obstet. Gynecol. 2025; 52(12): 45947\nhttps://doi.org/10.31083/CEOG45947\nCopyright: © 2025 The Author(s). Published by IMR Press.\nThis is an open access article under the CC BY 4.0 license .\nPublisher’s Note: IMR Press stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.\nOriginal Research\nThe Diagnostic Value of the CA125 and Ultrasound Combination for\nEndometriosis: A Retrospective Cohort Study\nY anhong Zhang1,2,*,†\n , Jiayin Wu3,4,†\n , Shujiao Lin 1,2\n , Hongli Liu 2,5\n , Y anqiu Zhong1,2\n ,\nMaoqi Wu1,2\n , Xiaohong Zhong 1,2,*\n1Department of Ultrasound, Women and Children’s Hospital, School of Medicine, Xiamen University, 361102 Xiamen, Fujian, China\n2Department of Obstetrics and Gynecology, Women and Children’s Hospital, School of Medicine, Xiamen University, 361102 Xiamen, Fujian, China\n3Department of Laboratory Medicine, Women and Children’s Hospital, School of Medicine, Xiamen University, 361102 Xiamen, Fujian, China\n4Fujian Key Clinical Specialty of Laboratory Medicine, Women and Children’s Hospital, School of Medicine, Xiamen University, 361102 Xiamen,\nFujian, China\n5Department of Gynecology, Women and Children’s Hospital, School of Medicine, Xiamen University, 361102 Xiamen, Fujian, China\n*Correspondence: 25185127@qq.com (Y anhong Zhang);1369946030@qq.com (Xiaohong Zhong)\n†These authors contributed equally.\nAcademic Editors: Michael Friedrich and Michael H. Dahan\nSubmitted: 21 August 2025 Revised: 24 September 2025 Accepted: 9 October 2025 Published: 22 December 2025\nAbstract\nBackground: Endometriosis affects an estimated 10% of women of reproductive age and is characterized by chronic pelvic pain, infertil-\nity, and an increased risk of ovarian cancer. Laparoscopy remains the diagnostic gold standard for endometriosis; however, this technique\nis an invasive procedure. Therefore, non-invasive alternatives, such as serum biomarkers (e.g., carbohydrate antigen 125 (CA125), car-\nbohydrate antigen 199 (CA199), and human epididymis protein 4 (HE4)) and ultrasound, represent promising techniques but currently\nlack consistent accuracy. Thus, this study aimed to evaluate the diagnostic efficacy of CA125, HE4, and ultrasound, individually and in\ncombination, for detecting endometriosis. Methods: A retrospective analysis was conducted involving 51 patients with endometriosis\nand 52 healthy controls. Serum levels of CA125 and HE4 were measured by chemiluminescence, and ultrasound examinations were\nperformed. Diagnostic performance was evaluated using sensitivity, specificity, accuracy, and receiver operating characteristic (ROC)\ncurves. Results: CA125 levels were significantly elevated in patients with endometriosis (1.90 ± 0.39, log10) compared to controls\n(1.06 ± 0.22, log10; p < 0.001), yielding an area under the curve (AUC) of 0.97 (95% confidence interval (CI): 0.94–1.00), with a\nsensitivity of 94.23% (95% CI: 0.87–1.01), specificity of 92.16% (95% CI: 0.85–1.00), positive predictive value (PPV) of 92.18% (95%\nCI: 0.85–1.00), negative predictive value (NPV) of 94.21% (95% CI: 0.87–1.01), and accuracy of 93.20% (95% CI: 0.88–0.98). In\ncontrast, HE4 did not exhibit a significant difference ( p = 0.156). Ultrasound demonstrated a sensitivity of 72.55% (95% CI: 0.60–0.85),\na specificity of 100.00% (95% CI: 1.00–1.00), PPV of 100.00% (95% CI: 1.00–1.00), NPV of 78.79% (95% CI: 0.69–0.89), and an\naccuracy of 86.41% (95% CI: 0.80–0.93). Combining CA125 and ultrasound enhanced the diagnostic accuracy to 95.15% (95% CI:\n0.91–1.00), achieving a sensitivity of 92.16% (95% CI: 0.85–1.00), PPV of 97.92% (95% CI: 0.94–1.02), NPV of 92.73% (95% CI:\n0.86–1.00), and specificity of 98.08% (95% CI: 0.94–1.01) (AUC: 0.96, 95% CI: 0.92–1.00). Conclusions: Combining CA125 and\nultrasound significantly improves the non-invasive diagnosis of endometriosis, providing high sensitivity and specificity. This approach\nmay reduce the reliance on laparoscopy, thereby facilitating earlier intervention. However, as this is a small-scale retrospective cohort\nstudy, larger sample cohorts are required for future validation to confirm these findings.\nKeywords: biomarkers; CA125; diagnostic; endometriosis; HE4; ultrasound\n1. Introduction\nEndometriosis is a prevalent gynecological disorder\ncharacterized by the ectopic implantation of functional en-\ndometrial tissue outside the uterine cavity, predominantly\nin the ovaries, pelvic peritoneum, and rectovaginal septum\n[1]. Under hormonal influence, these ectopic lesions un-\ndergo cyclic bleeding, eliciting chronic inflammatory re-\nsponses, fibrosis, and adhesion formation, which lead to\nclinical symptoms such as chronic pelvic pain, dysmen-\norrhea, dyspareunia, and infertility [ 2,3]. Globally, en-\ndometriosis affects approximately 10% of reproductive-\naged women [ 4], with its prevalence rising to 30%–50%\namong infertile women. Endometriosis not only severely\nimpacts the quality of life of patients but also increases the\nrisk of ovarian cancer [ 5]. The chronic pain and infertil-\nity associated with this condition impose significant eco-\nnomic and psychological burdens on families and society\n[6]. Therefore, early diagnosis and intervention are crucial\nfor delaying disease progression, improving reproductive\noutcomes, and reducing the risk of malignant transforma-\ntion.\nCurrently, laparoscopy combined with pathological\nbiopsy is considered the gold standard for diagnosing en-\ndometriosis, enabling direct visualization of lesions and\nstaging [ 7]. However, this technique is invasive, costly,\n\nrequires general anesthesia, and carries the risk of surgi-\ncal complications, which makes it difficult to promote it\nas a screening method. In recent years, serum biomarker\ntesting and imaging techniques have garnered widespread\nattention due to their convenience. Carbohydrate antigen\n125 (CA125) is one of the earliest studied markers, with el-\nevated levels associated with pelvic inflammation and peri-\ntoneal irritation [8]. Multiple studies have reported that the\nsensitivity of CA125 for diagnosing endometriosis ranges\nfrom 24% to 94%, while its specificity ranges from 83% to\n93% [9,10]. However, CA125 is susceptible to interference\nfrom the menstrual cycle, ovarian cysts, and other gyneco-\nlogical conditions [ 11]. A decrease in hemoglobin (HGB)\nis commonly observed in patients with endometriosis who\nexperience excessive menstrual bleeding or the deep infil-\ntrating type [ 12], but this marker lacks specificity. Carbo-\nhydrate antigen 199 (CA199) may be elevated in some en-\ndometriosis patients [ 13], but its sensitivity is low. Human\nepididymis protein 4 (HE4) is a novel biomarker; in recent\nyears, HE4 has been introduced into endometriosis research\nas an ovarian cancer marker [ 14], with the advantage of be-\ning unaffected by the menstrual cycle [ 15]. Studies have\nshown that HE4 is significantly elevated in endometriosis\npatients, with a sensitivity of about 66% [ 16,17], but its di-\nagnostic efficacy is limited when used alone. In imaging\ntechniques, transvaginal ultrasound (TVUS) demonstrates\na high diagnostic value for ovarian endometriosis, partic-\nularly for chocolate cysts [ 18], with a sensitivity ranging\nfrom 70% to 92% [ 19,20]. However, its sensitivity for su-\nperficial endometriosis is relatively low, and this sensitiv-\nity is highly dependent on the operator’s experience, which\nmay introduce diagnostic bias. Previous studies have in-\ndicated that the combination of multiple markers can en-\nhance diagnostic accuracy. For instance, the AUC for en-\ndometriosis diagnosis can reach 0.90 when CA125 is com-\nbined with HE4, HGB, and CA199 [ 16], although this still\nrequires validation with larger sample sizes.\nIn this study, we conducted a retrospective analysis of\nserum tumor marker levels and ultrasound findings in both\npatients with endometriosis and healthy controls. The di-\nagnostic performance of CA125, HE4, ultrasound, and their\ncombinations was assessed to identify an optimal combined\nindicator for the detection of endometriosis.\n2. Materials and Methods\n2.1 General Materials\nBetween July 2021 and July 2024, 51 patients with\nendometriosis who received treatment at Xiamen Maternal\nand Child Health Care Hospital were selected for the ob-\nservation group. The ages of the patients ranged from 26 to\n54 years, with all cases confirmed by histopathological ex-\namination. Inclusion criteria required that patients who had\nlaparoscopic surgery and were pathologically confirmed to\nhave endometriosis. Exclusion criteria comprised patients\nwith missing clinical data, those with hormone-dependent\ndiseases (such as adenomyosis and uterine fibroids), in-\ndividuals taking any hormonal medications, those with a\nhistory of pregnancy within the past six months, patients\nwith other endocrine, immune, or metabolic diseases, and\nthose with severe liver and kidney dysfunction, hematolog-\nical disorders, or malignant tumors (Fig. 1). Furthermore,\n52 healthy women who underwent physical examinations\nat the same hospital during the same period were selected\nas the control group (Fig. 1), aged between 25 and 60 years.\nThere was no statistically significant difference in age be-\ntween the two groups ( p > 0.05). This study has been\napproved by the Ethics Committee of Women and Chil-\ndren’s Hospital, School of Medicine, Xiamen University\n(KY -2025-104-K01).\n2.2 Sample Testing\nFor serum tumor marker testing, 4 mL of peripheral\nvenous blood was drawn from patients in the observation\ngroup preoperatively during non-menstrual periods and un-\nder fasting conditions in the morning. Similarly, the con-\ntrol group had 4 mL of peripheral venous blood drawn on\nthe morning of the physical examination, also while fast-\ning. After standing at room temperature for 30 minutes,\nthe samples were centrifuged at 3000 rpm for 10 minutes.\nThe serum was collected and stored in sterile EP tubes,\nwhich were then stored at –80 °C until testing. The serum\n(20153401967, Innovax, Xiamen, Fujian, China) and HE4\n(20163400698, Innovax, Xiamen, Fujian, China) expres-\nsion levels of both groups were measured using a chemi-\nluminescence analyzer (I2000, Abbott, Lake Forest, IL,\nUSA), with all testing procedures strictly adhering to the\nkit instructions.\nThe ultrasound examination was performed by mul-\ntiple operators using the Philips (IU22, Amsterdam, NH,\nNetherlands), Aloka (ProSound Alpha 5, Hitachi Aloka,\nTokyo, Japan) and GE (V oluson E8, Boston, MA, USA)\ncolor Doppler ultrasound diagnostic system, with metic-\nulous observation of the patient’s pelvic area, including\nthe uterine and ovarian spaces. Key aspects assessed in-\nclude the thickness of the endometrium, uterine size, in-\nternal echo, and the position of the uterus. Additionally,\ndetailed observations and recordings were made regarding\nthe lesion’s location, marginal morphology, diameter, and\nwhether a capsule was present. The blood flow signal dis-\ntribution in the targeted lesion was also examined. The di-\nagnosis of endometriosis at various sites was made using\nestablished clinical diagnostic criteria.\n2.3 Observation Indicators\nThis study aims to evaluate the positive outcomes as-\nsociated with endometriosis diagnosed through ultrasound\nusing statistical methods. We compared and analyzed the\nserum levels of CA125 and HE4 between two groups of\nsubjects. Additionally, we considered the histopathologi-\ncal examination results from surgery as the gold standard\n2\n\n\nFig. 1. Flowchart of study population inclusion and exclusion criteria. n, number of samples.\nfor diagnosing endometriosis. The diagnostic efficacy of\nultrasound, CA125, and HE4 was statistically analyzed and\ncompared both individually and in combination. Specificity\nwas defined as the proportion of true negative cases to the\nsum of true negative and false positive cases, expressed as\na percentage. Sensitivity was calculated as the ratio of true\npositive cases to the sum of true positive and false negative\ncases, also expressed as a percentage. Accuracy was cal-\nculated by dividing the sum of true positive and true nega-\ntive cases by the total number of cases, then multiplying by\n100%.\n2.4 Statistical Analysis\nThe sample size requirement was determined using\nPASS (version 2021, NCSS, Kaysville, UT, USA), spec-\nify the test direction, report power = 90%, anticipated AUC\nof 0.9, null AUC of 0.7, allocation ratio was set at 1:1, the\nlower limit of the false positive rate was 0.0, and the up-\nper limit of the false positive rate was 1.0, resulting in a\nneed for a minimum of 24 patients with endometriosis and\n24 healthy controls. Ultimately, the study included a total\nof 51 patients with endometriosis and 52 controls. All sta-\ntistical analyses were conducted using SPSS 27.0 software\n(IBM Corp., Armonk, NY , USA). The Shapiro-Wilk test\nwas employed to analyze the type of data distribution, and\nthe parameters distributed with normal distribution were ex-\nplained as Mean ± SD. The t test was used when parame-\nters were distributed with normally distribution, The Mann-\nWhitney U test was used when parameters distributed with\nnon-normal distribution. Pearson’s χ2 test and Fisher’s ex-\nact test were applied to analyze the correlation of categori-\ncal data derived from ultrasound detection indicators. The\nPearson’s χ2 test was used when all expected frequencies\nt ≥ 5 and the total number of cases N ≥ 40. The Fisher’s\nexact test was used when the expected frequency t < 5 or\nthe total number of cases N < 40. Binary logistic regres-\nsion analysis was utilized to construct the diagnostic model.\nWe employed Prism 10 to plot the receiver operating char-\nacteristic (ROC) curves and calculate the AUC to evaluate\nthe diagnostic efficiency of various diagnostic indicators.\nStatistical significance was set at p < 0.05.\n3. Results\n3.1 Basic Clinical Data Between the Endometriosis Group\nand the Control Group\nA total of 51 patients with histologically and patholog-\nically confirmed endometriosis and 52 age-matched con-\ntrols were enrolled for analysis. Comparative analysis of\nbaseline characteristics demonstrated that the mean age of\nthe endometriosis group was slightly higher than that of the\ncontrol group (38.20 ± 7.30 years vs. 35.60 ± 7.50 years).\nNotably, serum CA125 levels were markedly elevated in\nendometriosis patients compared to controls (1.90 ± 0.39,\nlog10 vs. 1.06 ± 0.22, log10). On the contrary, HE4 levels\nwere significantly higher in the control group (1.67 ± 0.07\nvs. 1.62 ± 0.17) (Table 1).\nAnalysis of endometriotic lesion distribution revealed\nthat the most common isolated lesion type was ovarian\nendometriosis (25 cases, 49.02%), followed by ovarian\ncombined with pelvic endometriosis (11 cases, 21.57%).\nOther isolated lesion types included pelvic endometrio-\nsis (2 cases, 3.92%), abdominal wall endometriosis (1\ncase, 1.96%), and fallopian tube endometriosis (1 case,\n1.96%). Additionally, several multifocal lesion combi-\n3\n\nTable 1. Basic clinical data and blood parameters between the endometriosis group and the control group.\nParameters Control group (n = 52) Endometriosis group (n = 51) Differential analysis\nAge (years, mean ± SD) 35.60 ± 7.50 38.20 ± 7.30 t = 1.75, p = 0.080\nMenopause 11.54% 1.96% Z = 2.14, p < 0.050\nDysmenorrhea 7.69% 76.47% Z = 18.43, p < 0.001\nCA125 (log10, mean ± SD) 1.06 ± 0.22 1.90 ± 0.39 t = 1.85, p < 0.001\nHE4 (log10, mean ± SD) 1.67 ± 0.07 1.62 ± 0.17 t = 13.79, p = 0.070\nSD, standard deviation; CA125, cancer antigen 125; HE4, human epididymis protein 4.\nTable 2. Statistics on the proportion of endometriosis patients\nwith different lesion sites.\nLesion sites Cases Proportion (%)\nOvary 25 49.02%\nOvary and pelvic cavity 11 21.57%\nOvary and fallopian tube 2 3.92%\nPelvic cavity 2 3.92%\nAbdominal wall 1 1.96%\nFallopian tube 1 1.96%\nOvary and intrinsic ligament 1 1.96%\nOvary and rectum 1 1.96%\nOvary, cervix, and pelvic cavity 1 1.96%\nOvary, pelvic cavity, and vaginal wall 1 1.96%\nOvary, fallopian tube, and pelvic cavity 1 1.96%\nOvary, fallopian tube, and vaginal wall 1 1.96%\nOvary, rectum, and abdominal wall 1 1.96%\nOvary, rectum, and vagina 1 1.96%\nOvary, pelvic cavity, and intestine 1 1.96%\nnations were observed: ovarian with fallopian tube in-\nvolvement (2 cases, 3.92%), ovarian with uterosacral lig-\nament involvement (1 case, 1.96%), and ovarian with rec-\ntal involvement (1 case, 1.96%). Furthermore, seven cases\n(13.72%) exhibited triple-site involvement, including com-\nbinations of ovarian-cervical-pelvic, ovarian-vaginal wall-\npelvic, ovarian-fallopian tube-pelvic, ovarian-fallopian\ntube-vaginal wall, ovarian-rectal-abdominal wall, ovarian-\nrectal-vaginal, and ovarian-pelvic-intestinal endometriosis\n(Table 2).\n3.2 Quantitative Comparison of Serum Tumor Biomarker\nLevels Between the Endometriosis and the Control Group\nWe analyzed the serum tumor marker expression lev-\nels between the endometriosis group and the control group.\nThe results showed that the CA125 expression level in the\nendometriosis group (1.90 ± 0.39, log10) was significantly\nhigher compared to the control group (1.06 ± 0.22, log10)\n(p < 0.001) (Fig. 2A, Table 3). However, there was no sig-\nnificant difference in the HE4 expression levels between the\nendometriosis group (1.62 ± 0.17, log10) and the control\ngroup (1.67 ± 0.07, log10) (Fig. 2B, Table 3).\nTo further elucidate the characteristics of endometrio-\nsis based on different sites of disease onset, this study clas-\nsified patients according to their disease location and fo-\ncused on comparing the levels of CA125 and HE4 in the two\nmost common subtypes—ovarian endometriosis and ovar-\nian endometriosis combined with pelvic endometriosis—\nwith those of the control group. The results showed that\nthe CA125 levels in the control group were statistically dif-\nferent from those in both the ovarian endometriosis group\nand the ovarian endometriosis combined with pelvic en-\ndometriosis group. However, there was no statistically sig-\nnificant difference in CA125 levels between the ovarian en-\ndometriosis group and the ovarian endometriosis combined\nwith pelvic endometriosis group (Fig. 2C, Table 3). Ad-\nditionally, the HE4 levels in the control group were sta-\ntistically different from those in the ovarian endometrio-\nsis group, but not from those in the ovarian endometrio-\nsis combined with pelvic endometriosis group. Notably,\nthere was a statistically significant difference in HE4 levels\nbetween the ovarian endometriosis group and the ovarian\nendometriosis combined with pelvic endometriosis group\n(Fig. 2D, Table 3).\n3.3 Analysis of the Diagnostic V alue of Serum Tumor\nBiomarkers for Endometriosis\nWe further analyzed the individual diagnostic efficacy\nof CA125 and HE4 serum markers in the diagnosis of en-\ndometriosis. To this end, we used a healthy control group as\na reference and drew ROC curves to evaluate the diagnos-\ntic accuracy of these two markers. The results showed that\nthe AUC value of the CA125 diagnostic model was 0.97 ( p\n< 0.001, 95% CI: 0.94–1.00) (Fig. 3A, Table 4), indicat-\ning that it has very high diagnostic accuracy. In contrast,\nthe AUC value of the HE4 diagnostic model was 0.58 ( p\n= 0.156, 95% CI: 0.47–0.70) (Fig. 3B, Table 4), close to\nthe level of random guessing, suggesting that its value in\ndiagnosing endometriosis is limited.\nFurthermore, through the maximum Y ouden index,\nwe calculated the sensitivity, specificity, and accuracy of\nCA125 and HE4 in the diagnosis of endometriosis. CA125\nhad a sensitivity of 94.23% (95% CI: 0.87–1.01), specificity\nof 92.16% (95% CI: 0.85–1.00), accuracy of 93.20% (95%\nCI: 0.88–0.98), positive predictive value (PPV) of 92.18%\n(95% CI: 0.85–1.00), and negative predictive value (NPV)\nof 94.21% (95% CI: 0.87–1.01) (Fig. 3C, Table 4). This\nmeans that CA125 can effectively identify the vast major-\nity of actual patients and accurately exclude non-patients,\nwith high overall diagnostic accuracy. HE4 had a sensitiv-\n4\n\n\nFig. 2. Comparative analysis of serum tumor marker levels between endometriosis patients and healthy controls.(A) Carbohydrate\nantigen 125 (CA125) levels in controls versus endometriosis patients. (B) HE4 levels in controls versus endometriosis patients. (C)\nCA125 levels across study groups: controls, isolated ovarian endometriosis, and ovarian endometriosis with pelvic involvement. (D) HE4\nlevels across study groups: controls, isolated ovarian endometriosis, and ovarian endometriosis with pelvic involvement. * represents p\n< 0.05, ** represents p < 0.01, **** represents p < 0.0001; ns, not significant.\nity of 98.08% (95% CI: 0.94–1.02), specificity of 35.29%\n(95% CI: 0.21–0.48), accuracy of 66.38% (95% CI: 0.57–\n0.75), PPV of 59.78% (95% CI: 0.49–0.70), and NPV of\n94.93% (95% CI: 0.84–1.06) (Fig. 3D, Table 4). Although\nHE4 has high sensitivity, meaning that it can detect most ac-\ntual patients, its low specificity indicates that there may be\nmore false-positive results, i.e., misdiagnosing some non-\npatients as patients. This could lead to unnecessary further\nexaminations and treatments. Therefore, based on the AUC\nvalues and various diagnostic indicators, CA125 performs\nbetter than HE4 in the diagnosis of endometriosis.\n3.4 The Diagnostic V alue of Ultrasound in Endometriosis\nWe used pathological diagnosis as the gold standard to\nevaluate the diagnostic value of ultrasound in endometrio-\nsis. A total of 51 patients diagnosed with endometriosis and\n52 controls were included in the study. Ultrasound exami-\nnation results showed that among the 51 patients, 14 were\nmissed (ultrasound did not detect the disease), while in the\n52 controls, ultrasound did not detect any abnormalities.\nTo explore the diagnostic value of ultrasound in en-\ndometriosis, we used the Chi-Square test to analyze the\nassociation between ultrasound results and disease status.\nThe results showed that ultrasound could significantly dis-\ntinguish patients with endometriosis from controls (Chi-\nSquare test; χ2 = 58.875; p < 0.01) (Fig. 4A), indicating\nthat ultrasound has statistical significance in diagnosing en-\ndometriosis.\nFurthermore, through the maximum Y ouden index,\nwe calculated the sensitivity, specificity, accuracy, PPV ,\nand NPV of ultrasound detection to be 72.55% (95% CI:\n0.60–0.85), 100.00% (95% CI: 1.00–1.00), 86.41% (95%\nCI: 0.80–0.93), 100.00% (95% CI: 1.00–1.00), and 78.79%\n(95% CI: 0.69–0.89), respectively (Fig. 4B, Table 5). These\nmetrics indicate that ultrasound detection has high speci-\nficity and accuracy in the diagnosis of endometriosis, al-\nthough the sensitivity is slightly lower, meaning that a cer-\ntain proportion of patients may be missed. However, the\n100% specificity means that ultrasound can very reliably\nexclude those without the disease, which helps reduce un-\nnecessary further examinations. In conclusion, ultrasound\ndetection has important clinical value in the diagnosis of\nendometriosis.\n5\n\nTable 3. Comparative analysis of serum tumor marker levels between endometriosis patients and healthy controls.\nSerum tumor\nmarker\nGroup (number) Quantitative level\n(log10, mean ± SD)\np-value\nCA125\nHealthy people (n = 52) (1) 1.06 ± 0.22\n1 vs. 2: t = 1.85,\np < 0.0010\n1 vs. 3: t = 13.36,\np < 0.0010\n1 vs. 4: t = 9.75,\np < 0.0010\n3 vs. 4: t = 0.27,\np = 0.7900\nEndometriosis (n = 51) (2) 1.90 ± 0.39\nOvarian endometriosis (n = 25) (3) 1.91 ± 0.34\nOvarian and pelvic endometriosis (n = 11) (4) 1.95 ± 0.47\nHE4\nHealthy (n = 52) (1) 1.67 ± 0.07\n1 vs. 2: t = 13.79,\np = 0.0700\n1 vs. 3: t = 3.22,\np = 0.0019**\n1 vs. 4: t = 0.94,\np = 0.3500\n3 vs. 4: t = 2.05,\np = 0.0480*\nEndometriosis (n = 51) (2) 1.62 ± 0.17\nOvarian endometriosis (n = 25) (3) 1.57 ± 0.18\nOvarian and pelvic endometriosis (n = 11) (4) 1.69 ± 0.09\n* represents p < 0.05; ** represents p < 0.01.\n6\n\n\nFig. 3. Analysis of the diagnostic value of CA125 and HE4 for endometriosis. (A) ROC curve analysis of the diagnostic performance\nof CA125 for endometriosis. (B) ROC curve analysis of the diagnostic performance of HE4 for endometriosis. (C) Sensitivity, specificity,\nand accuracy of diagnosing endometriosis through CA125. (D) Sensitivity, specificity, and accuracy of diagnosing endometriosis through\nHE4. AUC, area under the curve; 95% CI, 95% confidence interval; PPV , positive predictive value; NPV , negative predictive value; ROC,\nreceiver operating characteristic.\nTable 4. Analysis of the diagnostic value of CA125 and HE4\nfor endometriosis.\nSerum tumor marker Indicator Result (95% CI)\nCA125\nAUC 0.97 (0.94–1.00)\nSensitivity 94.23% (0.87–1.01)\nSpecificity 92.16% (0.85–1.00)\nAccuracy 93.20% (0.88–0.98)\nPPV 92.18% (0.85–1.00)\nNPV 94.21% (0.87–1.01)\nHE4\nAUC 0.58 (0.47–0.70)\nSensitivity 98.08% (0.94–1.02)\nSpecificity 35.29% (0.21–0.48)\nAccuracy 66.38% (0.57–0.75)\nPPV 59.78% (0.49–0.70)\nNPV 94.93% (0.84–1.06)\n3.5 The Diagnostic V alue of CA125 Combined With\nUltrasound in Endometriosis\nPrevious studies have shown that serum tumor marker\nCA125 and ultrasound imaging have certain diagnostic\nvalue in the diagnosis of endometriosis. However, the di-\nagnostic accuracy of these single-detection methods still re-\nquires improvement. Therefore, this part of the study aims\nto explore whether the combination of CA125 and ultra-\nsound detection can improve the diagnostic accuracy.\nTable 5. Analysis of the diagnostic value of ultrasound for\nendometriosis.\nDetection method Indicator Result (95% CI)\nUltrasound\nSensitivity 72.55% (0.60–0.85)\nSpecificity 100.00% (1.00–1.00)\nAccuracy 86.41% (0.80–0.93)\nPPV 100.00% (1.00–1.00)\nNPV 78.79% (0.69–0.89)\nTo achieve this goal, we used logistic regression\nmodel to combine the numerical values of CA125 and ultra-\nsound detection results to establish a combined diagnostic\nmodel and evaluated its diagnostic performance by draw-\ning ROC curves (Fig. 5A, Table 6). The results showed\nthat the AUC of the ROC curve for the combined diagnos-\ntic model was 0.96 (95% CI: 0.92–1.00), indicating that\nit has high accuracy in distinguishing patients with en-\ndometriosis from non-patients. Further analysis through the\nmaximum Y ouden index showed that the sensitivity, speci-\nficity, accuracy, PPV , and NPV of the combined diagnostic\nmodel were 92.16% (95% CI: 0.85–1.00), 98.08% (95% CI:\n0.94–1.01), 95.15% (95% CI: 0.91–1.00), 97.92% (95% CI:\n0.94–1.02), and 92.73% (95% CI: 0.86–1.00), respectively.\nThese metrics indicate that the model can not only effec-\ntively identify true patients (high sensitivity) but also ac-\n7\n\nFig. 4. Analysis of the diagnostic value of ultrasound for endometriosis . (A) Pearson’s Chi-Square test analysis of whether ultrasound\ncan differentiate between patients with endometriosis and controls. (B) Sensitivity, specificity, and accuracy of ultrasound in diagnosing\nendometriosis.\nFig. 5. Analysis of the diagnostic value of combined CA125 and ultrasound for endometriosis . (A) ROC curve analysis of the\ndiagnostic performance of combined CA125 and ultrasound for endometriosis. (B) Sensitivity, specificity, and accuracy of combined\nCA125 and ultrasound in diagnosing endometriosis.\nTable 6. Analysis of the diagnostic value of combined CA125\nand ultrasound for endometriosis.\nCombined detection Indicator Result (95% CI)\nCA125 and ultrasound\nAUC 0.96 (0.92–1.00)\nSensitivity 92.16% (0.85–1.00)\nSpecificity 98.08% (0.94–1.01)\nAccuracy 95.15% (0.91–1.00)\nPPV 97.92% (0.94–1.02)\nNPV 92.73% (0.86–1.00)\ncurately exclude non-patients (high specificity), with high\noverall diagnostic accuracy.\nNotably, compared with the single use of CA125 or\nultrasound detection, the accuracy of the combined diag-\nnostic model increased by 2.05% and 8.84%, respectively\n(Fig. 5B, Table 6). These results indicate that the combined\napplication of CA125 and ultrasound detection outperforms\nsingle detection methods in the diagnosis of endometriosis.\nThis may provide an effective approach for earlier and more\naccurate diagnosis of endometriosis in clinical practice.\n4. Discussion\nEndometriosis is an estrogen-dependent chronic in-\nflammatory disease characterized by the ectopic implanta-\ntion of endometrial stromal cells outside the uterine cavity\n[21]. This condition affects approximately 10% of women\nof reproductive age and manifests as severe pelvic pain,\ndysmenorrhea, and infertility, which significantly impair-\ning quality of life [ 2,3,22]. Additionally, endometriosis is\nlinked to the formation of adhesions and potential ovarian\nmalignancies, highlighting its long-term health risks and\ncontributing to a substantial socioeconomic burden due to\ndelayed diagnosis and repeated interventions [ 5,6]. Tradi-\ntional diagnostic methods for endometriosis have signifi-\ncant limitations. Currently, the gold standard for diagnosis\nis laparoscopic examination combined with histopatholog-\nical confirmation [ 7]. However, this invasive procedure,\nposes a risk of surgical complications and is expensive,\nlimiting its practicality for early screening. Non-invasive\nimaging techniques, such as TVUS and magnetic reso-\nnance imaging (MRI), exhibit limited sensitivity in detect-\ning early-stage diseases [ 23]. Meanwhile, serum biomark-\ners like CA199 lack specificity, as their elevated levels\n8\n\n\nare frequently observed in other gynecological conditions,\nleading to diagnostic ambiguity. Therefore, identifying re-\nliable diagnostic indicators is crucial to bridging these clini-\ncal gaps. A combination of biomarkers with high sensitivity\nand specificity can enable timely diagnosis, reduce reliance\non surgical confirmation, and facilitate the implementation\nof personalized treatment strategies.\nIn this study, we performed a retrospective analysis of\nserum inflammatory markers (including CA125, HE4) and\nultrasound indicators in patients with endometriosis. We\naimed to evaluate the differences between patients with en-\ndometriosis and healthy controls, as well as to assess the\ndiagnostic efficacy of these markers to identify an optimal\ncombined diagnostic indicator for precise diagnosis of en-\ndometriosis.\nBased on the theory that immune system dysfunction\nplays a role in the pathogenesis of endometriosis [ 24], im-\nmune molecules and inflammatory cytokines have been ex-\ntensively studied as potential biomarkers for this condition\n[25]. However, most results remain controversial. Despite\nthe conflicting evidence regarding the efficacy of CA125 as\na biomarker for endometriosis, meta-analyses indicate that\nCA125 remains one of the most used markers for this con-\ndition [10]. Our study found that CA125 levels in patients\nwith endometriosis are higher than those in healthy indi-\nviduals. This increase is associated with the inflammatory\nresponse and fibrotic processes, where ectopic endometrial\ntissue induces chronic localized inflammation, leading to\nincreased production and release of CA125 [ 26]. Addition-\nally, angiogenesis and cellular proliferation in ectopic le-\nsions may also promote the expression of CA125, resulting\nin a significant increase in its concentration in the blood.\nHE4 has been reported to be associated with endometrio-\nsis, suggesting it’s a hematological marker for diagnosing\n[14]. In this study, we explored the diagnostic potential of\nHE4; however, we found no statistically significant differ-\nence in HE4 levels between patients with endometriosis and\nhealthy individuals. This lack of significance may be at-\ntributed to variations in sample population used in the study\nas well as the relatively small sample size. Therefore, addi-\ntional research involving a larger sample size is necessary to\nassess the diagnostic efficiency of the HE4 as a biomarker\nfor endometriosis.\nUltrasound, due to its high resolution, real-time imag-\ning capability, and non-invasive nature, offers substantial\nadvantages in the diagnosis of endometriosis, particularly in\nidentifying ovarian endometriosis [18–20,27]. In this study,\nthe sensitivity of ultrasound for diagnosing endometrio-\nsis was found to be 72.55%, with a specificity of 100%.\nThese findings are consistent with previous reports, sug-\ngesting that the lower resolution of ultrasound in detect-\ning small cysts and early lesions largely accounts for the\nreduced sensitivity observed [ 28]. Notably, the combined\nuse of ultrasound and CA125 in diagnosing endometriosis\ncan increase sensitivity to 92.16% while maintaining high\nspecificity, thereby improving diagnostic accuracy. Fur-\nthermore, our findings revealed a superior AUC when com-\npared to prior studies utilizing multiple diagnostic indica-\ntors for endometriosis detection [ 24]. This enhancement\nmay be attributed to the complexity of endometriosis, which\npresents with varied symptoms and affects multiple organs.\nA collaborative diagnostic approach can facilitate a com-\nprehensive assessment, further enhancing diagnostic effi-\ncacy.\nLimitations\nThis study has several limitations that should be con-\nsidered. The modest sample size (n = 103) restricted in-\ndepth subgroup analyses, particularly for rare lesion sub-\ntypes, and may have increased the risk of selection bias.\nFurthermore, although healthy controls were included, the\nabsence of pathological confirmation means that asymp-\ntomatic endometriosis cannot be definitively ruled out in\nthe control group. Future studies with larger, prospectively\nrecruited cohorts and gold-standard verification for all par-\nticipants are needed to validate these findings and facilitate\nthe translation of this combined diagnostic strategy into rou-\ntine clinical practice.\nIn summary, this study developed a combined diag-\nnostic model integrating CA125 and ultrasound detection,\noffering a novel method for the precise non-invasive diag-\nnosis of endometriosis. To strengthen clinical applicabil-\nity, future efforts should focus on multi-center validation\nusing larger, prospective cohorts. Further refinement of di-\nagnostic accuracy could also be achieved by incorporating\nadditional novel biomarkers alongside advanced ultrasound\ncharacteristics.\n5. Conclusions\nThis study demonstrates that the combined use of\nCA125 and ultrasound shows a trend toward improvement\nof the non-invasive diagnosis of endometriosis, achieving a\nsensitivity of 92.16%, specificity of 98.08%, and an overall\ndiagnostic accuracy of 95.15%. While CA125 alone ex-\nhibited high diagnostic performance (AUC: 0.97, 95% CI:\n0.94–1.00), its integration with ultrasound further improved\nsensitivity, particularly in detecting early or atypical lesions\nthat may evade detection by either method alone. The non-\ninvasive nature of this combined approach offers a promis-\ning alternative to laparoscopy, potentially reducing surgical\nrisks and enabling earlier intervention for symptom man-\nagement and fertility preservation. However, the retrospec-\ntive design and limited sample size underscore the need for\nvalidation in larger, prospective cohorts. These findings ad-\nvocate for the adoption of CA125 and ultrasound as a first-\nline diagnostic strategy, bridging the gap between clinical\nsuspicion and definitive diagnosis in endometriosis care.\n9\n\nAvailability of Data and Materials\nAll the original contributions presented in the study\nare included in the article, further inquiries can be directed\nto the corresponding authors.\nAuthor Contributions\nYHZ and XZ designed the research study. YQZ, SL\nand JW performed the research. HL, YQZ and MW ana-\nlyzed the data. All authors contributed to editorial changes\nin the manuscript. All authors read and approved the fi-\nnal manuscript. All authors have participated sufficiently\nin the work and agreed to be accountable for all aspects of\nthe work.\nEthics Approval and Consent to Participate\nThe study was carried out in accordance with the\nguidelines of the Declaration of Helsinki. This study has\nbeen approved by the Ethics Committee of Women and\nChildren’s Hospital, School of Medicine, Xiamen Univer-\nsity (KY -2025-104-K01). All patients meet the criteria for\nwaiver of informed consent.\nAcknowledgment\nNot applicable.\nFunding\nThis research received no external funding.\nConflict of Interest\nThe authors declare no conflict of interest.\nDeclaration of AI and AI-Assisted\nTechnologies in the Writing Process\nDuring the preparation of this work the authors used\nDeepSeek in order to check spell and grammar. After us-\ning this tool, the authors reviewed and edited the content as\nneeded and takes full responsibility for the content of the\npublication.\nReferences\n[1] Taylor HS, Kotlyar AM, Flores V A. Endometriosis is a\nchronic systemic disease: clinical challenges and novel inno-\nvations. Lancet. 2021; 397: 839–852. https://doi.org/10.1016/\nS0140-6736(21)00389-5 .\n[2] Janssen EB, Rijkers ACM, Hoppenbrouwers K, Meuleman C,\nD’Hooghe TM. Prevalence of endometriosis diagnosed by la-\nparoscopy in adolescents with dysmenorrhea or chronic pelvic\npain: a systematic review. Human Reproduction Update. 2013;\n19: 570–582. https://doi.org/10.1093/humupd/dmt016.\n[3] de Ziegler D, Borghese B, Chapron C. Endometriosis and in-\nfertility: pathophysiology and management. Lancet. 2010; 376:\n730–738. https://doi.org/10.1016/S0140-6736(10)60490-4 .\n[4] Ghiasi M, Kulkarni MT, Missmer SA. Is Endometriosis More\nCommon and More Severe Than It Was 30 Y ears Ago? Journal\nof Minimally Invasive Gynecology. 2020; 27: 452–461. https:\n//doi.org/10.1016/j.jmig.2019.11.018.\n[5] Kim HS, Kim TH, Chung HH, Song YS. Risk and prognosis of\novarian cancer in women with endometriosis: a meta-analysis.\nBritish Journal of Cancer. 2014; 110: 1878–1890. https://doi.or\ng/10.1038/bjc.2014.29.\n[6] Simoens S, Dunselman G, Dirksen C, Hummelshoj L, Bokor A,\nBrandes I, et al . The burden of endometriosis: costs and qual-\nity of life of women with endometriosis and treated in refer-\nral centres. Human Reproduction. 2012; 27: 1292–1299. https:\n//doi.org/10.1093/humrep/des073.\n[7] Pascoal E, Wessels JM, Aas-Eng MK, Abrao MS, Condous G,\nJurkovic D, et al . Strengths and limitations of diagnostic tools\nfor endometriosis and relevance in diagnostic test accuracy re-\nsearch. Ultrasound in Obstetrics & Gynecology. 2022; 60: 309–\n327. https://doi.org/10.1002/uog.24892.\n[8] Garzetti GG, Ciavattini A, Tranquilli AL, Arduini D, Ro-\nmanini C. Serum CA-125 concentration in endometriosis pa-\ntients: role of pelvic and peritoneal irritation. Gynecologi-\ncal Endocrinology. 1994; 8: 27–31. https://doi.org/10.3109/\n09513599409028454.\n[9] Bedaiwy MA, Falcone T. Laboratory testing for endometrio-\nsis. Clinica Chimica Acta. 2004; 340: 41–56. https://doi.org/10.\n1016/j.cccn.2003.10.021.\n[10] Hirsch M, Duffy J, Davis CJ, Nieves Plana M, Khan KS, Interna-\ntional Collaboration to Harmonise Outcomes and Measures for\nEndometriosis. Diagnostic accuracy of cancer antigen 125 for\nendometriosis: a systematic review and meta-analysis. BJOG:\nAn International Journal of Obstetrics and Gynaecology. 2016;\n123: 1761–1768. https://doi.org/10.1111/1471-0528.14055.\n[11] Kafali H, Artuc H, Demir N. Use of CA125 fluctuation dur-\ning the menstrual cycle as a tool in the clinical diagnosis of en-\ndometriosis; a preliminary report. European Journal of Obstet-\nrics, Gynecology, and Reproductive Biology. 2004; 116: 85–88.\nhttps://doi.org/10.1016/j.ejogrb.2004.02.039.\n[12] Nelson AL, Ritchie JJ. Severe anemia from heavy menstrual\nbleeding requires heightened attention. American Journal of Ob-\nstetrics and Gynecology. 2015; 213: 97.e1–97.e6. https://doi.or\ng/10.1016/j.ajog.2015.04.023.\n[13] Harada T, Kubota T, Aso T. Usefulness of CA19-9 versus CA125\nfor the diagnosis of endometriosis. Fertility and Sterility. 2002;\n78: 733–739. https://doi.org/10.1016/s0015-0282(02)03328-9 .\n[14] Huhtinen K, Suvitie P , Hiissa J, Junnila J, Huvila J, Kujari H,\net al . Serum HE4 concentration differentiates malignant ovar-\nian tumours from ovarian endometriotic cysts. British Journal\nof Cancer. 2009; 100: 1315–1319. https://doi.org/10.1038/sj.b\njc.6605011.\n[15] Xu XR, Wang X, Zhang H, Liu MY , Chen Q. The clinical\nsignificance of the combined detection of serum Smac, HE4\nand CA125 in endometriosis-associated ovarian cancer. Can-\ncer Biomarkers. 2018; 21: 471–477. https://doi.org/10.3233/CB\nM-170720.\n[16] Chen T, Wei JL, Leng T, Gao F, Hou SY . The diagnostic value\nof the combination of hemoglobin, CA199, CA125, and HE4 in\nendometriosis. Journal of Clinical Laboratory Analysis. 2021;\n35: e23947. https://doi.org/10.1002/jcla.23947.\n[17] Zapardiel I, Gorostidi M, Ravaggi A, Allende MT, Silveira M,\nAbehsera D, et al. Utility Serum Marker HE4 for the Differential\nDiagnosis Between Endometriosis and Adnexal Malignancy. In-\nternational Journal of Gynecological Cancer. 2016; 26: 52–55.\nhttps://doi.org/10.1097/IGC.0000000000000579.\n[18] Baușic A, Coroleucă C, Coroleucă C, Comandașu D, Matasariu\nR, Manu A, et al . Transvaginal Ultrasound vs. Magnetic Res-\nonance Imaging (MRI) V alue in Endometriosis Diagnosis. Di-\nagnostics. 2022; 12: 1767. https://doi.org/10.3390/diagnostic\ns12071767.\n[19] Alborzi S, Rasekhi A, Shomali Z, Madadi G, Alborzi M, Kazemi\nM, et al . Diagnostic accuracy of magnetic resonance imaging,\ntransvaginal, and transrectal ultrasonography in deep infiltrating\n10\n\n\nendometriosis. Medicine. 2018; 97: e9536. https://doi.org/10.\n1097/MD.0000000000009536.\n[20] Zhang X, He T, Shen W. Comparison of physical examina-\ntion, ultrasound techniques and magnetic resonance imaging for\nthe diagnosis of deep infiltrating endometriosis: A systematic\nreview and meta-analysis of diagnostic accuracy studies. Ex-\nperimental and Therapeutic Medicine. 2020; 20: 3208–3220.\nhttps://doi.org/10.3892/etm.2020.9043.\n[21] McCallion A, Nasirzadeh Y , Lingegowda H, Miller JE, Kha-\nlaj K, Ahn S, et al . Estrogen mediates inflammatory role of\nmast cells in endometriosis pathophysiology. Frontiers in Im-\nmunology. 2022; 13: 961599. https://doi.org/10.3389/fimmu.\n2022.961599.\n[22] Guo SW, Wang Y . The prevalence of endometriosis in women\nwith chronic pelvic pain. Gynecologic and Obstetric Investiga-\ntion. 2006; 62: 121–130. https://doi.org/10.1159/000093019.\n[23] Tian Z, Zhang YC, Sun XH, Wang Y , Zhao Y , Chang XH,\net al . Accuracy of transvaginal ultrasound and magnetic reso-\nnance imaging for diagnosis of deep endometriosis in bladder\nand ureter: a meta-analysis. Journal of Obstetrics and Gynaecol-\nogy. 2022; 42: 2272–2281. https://doi.org/10.1080/01443615.\n2022.2040965.\n[24] Abramiuk M, Grywalska E, Małkowska P , Sierawska O,\nHrynkiewicz R, Niedźwiedzka-Rystwej P . The Role of the Im-\nmune System in the Development of Endometriosis. Cells. 2022;\n11: 2028. https://doi.org/10.3390/cells11132028.\n[25] He Y , Li J, Qu Y , Sun L, Zhao X, Wu H, et al . Identifica-\ntion and Analysis of Potential Immune-Related Biomarkers in\nEndometriosis. Journal of Immunology Research. 2023; 2023:\n2975581. https://doi.org/10.1155/2023/2975581.\n[26] Schneider MP , Vitonis AF, Fadayomi AB, Charlton BM, Miss-\nmer SA, DiV asta AD. Quality of Life in Adolescent and Y oung\nAdult Women With Dyspareunia and Endometriosis. The Jour-\nnal of Adolescent Health. 2020; 67: 557–561. https://doi.org/10.\n1016/j.jadohealth.2020.02.024.\n[27] Nguyen PN, Nguyen VT. Additional value of Doppler ultra-\nsound to B-mode ultrasound in assessing for uterine intracav-\nitary pathologies among perimenopausal and postmenopausal\nbleeding women: a multicentre prospective observational study\nin Vietnam. Journal of Ultrasound. 2023; 26: 459–469. https:\n//doi.org/10.1007/s40477-022-00732-w .\n[28] V an den Bosch T, V an Schoubroeck D. Ultrasound diagnosis of\nendometriosis and adenomyosis: State of the art. Best Practice &\nResearch. Clinical Obstetrics & Gynaecology. 2018; 51: 16–24.\nhttps://doi.org/10.1016/j.bpobgyn.2018.01.013.\n11","source_license":"CC0","license_restricted":false}