The predictive diagnostic value of combined serum detection of CXCL12, VEGF, and CA125 for endometriosis

We performed the analysis via IBM SPSS Statistics, Version 27.0 software (IBM Corp., Armonk). The quantitative data used in this study followed a normal distribution and were analyzed via a t test; the results are expressed as the means ± standard deviations ( X̄  ±  S ). Comparisons between different rASRM stages and VAS groups were performed via ANOVA. Count data were compared via the χ 2 test. To assess the independent diagnostic value of each biomarker while controlling for potential confounders, we performed multivariate logistic regression analysis to identify independent predictors and variables with P  < .1 in differential analysis were entered into the multivariate model. Receiver operating characteristic (ROC) curves and the area under the curve (AUC) were plotted to evaluate the predictive value of serum markers for EMs. A P value < .05 was considered statistically significant.

The study also has potential limitations: the small sample size in this study may have caused biases in the statistical results. Moreover, this is a single-center study, with a relatively small sample size, targeting only Chinese people in certain regions, without external validation. Efforts are needed to collect multicenter case data and expand sample to validate the study in the future. In addition, Longitudinal follow-up can also be performed to explore the utility of these indicators for monitoring disease progression and recurrence. Besides, we did not account for all potential confounders such as specific comorbidities or hormonal use, which may give rise to selection bias and fail to detect differences across groups, further analysis is required to enable generalizability of the findings to other populations. Additionally, the delay in time between sample collection and processing could affect biomarker concentrations observed in this study. On the other hand, the detection of serum biomarkers only has certain limitations, it is better to correlate with more clinical symptoms, signs, and other imaging examinations to achieve better functionality. Future prospective large-sample studies and more available information should be conducted to further confirm their diagnostic value.

Our clinical study measured the levels of serum CXCL12, VEGF, and CA125 from women with and without EMs. By comparing the differences between the 2 groups, we identified differential biomarkers. Additionally, through ROC curve analysis, we evaluated the diagnostic efficacy of these biomarkers both individually and in combination. This study has laid a theoretical foundation for the development of noninvasive tests for EMs. Future work should focus on expanding the sample size to further confirm their practicality and provide more reliable evidence-based medical evidence for clinicians.

Endometriosis (EMs) is a common disease in women of reproductive age that causes significant psychological and physiological trauma due to infertility and chronic pelvic pain. Various inflammatory and immune factors are currently believed to be associated with EMs and play pivotal roles in its occurrence and development. [ 1 , 2 ] Although a diagnosis can be confirmed by surgery along with postoperative pathology, patients do not seek surgical intervention until the disease has advanced, leading to delay in treatment. Otherwise, surgical procedure is invasive and associated with risks such as trauma, adhesion, and decreased fertility, early assessment and seeking noninvasive biomarkers of EMs have significant effects on alleviating disease progression, reducing complications, and improving patient prognosis. [ 3 ] As far as we know, there is no reliable single serum biomarker for accurate diagnosis currently, but there is a general consensus that a combination of specific biomarkers may enhance diagnostic accuracy for EMs. [ 4 , 5 ] Carbohydrate antigen 125 (CA125) is an inflammatory marker, and studies have indicated that the destruction and erosion of the peritoneum by endometriotic lesions release CA125 into the bloodstream, leading to a significant increase in serum CA125 levels. [ 6 ] Although elevated CA125 levels is common in patients with EMs, it is not a definitive diagnostic marker. [ 7 ] Consequently, the identification of novel and effective biomarkers for EMS diagnosis remains a pressing need to facilitate clinical management. Stromal cell-derived factor (CXCL12), a proinflammatory factor, promotes chemotaxis, adhesion, and migration by binding to high-affinity CXCR4, activating immune cells and driving them to inflammatory sites to induce an inflammatory response. [ 8 ] Vascular endothelial growth factor (VEGF), an important angiogenic factor, plays a major regulatory role in the process of blood vessel formation. Abnormal overexpression of VEGF can accelerate the establishment, growth, and persistence of endometriotic lesions. [ 9 ] However, the combined diagnostic value of the 3 indicators for EMs remains unexplored. This study investigated the changes in serum CXCL12, CA125, and VEGF levels in EMs patients and explored their diagnostic value for EMS through differential and diagnostic performance analysis, assessing their potential clinical utility and aiming to provide reference for the diagnosis and treatment of EMs patients.

Funding acquisition: Hong Xu. Investigation: Chenxiang Wang. Resources: Hong Xu, Zhijia Xie, Yu Zheng, Chenxiang Wang. Supervision: Chenxiang Wang. Writing – original draft: Hong Xu. Writing – review & editing: Zhijia Xie.

Patients who were diagnosed with pelvic masses by ultrasound at Suzhou Ninth People’s Hospital during the period from October 2023 to February 2025 were enrolled in this trial. The inclusion criteria were as follows: met the diagnostic criteria for EMs according to the guidelines and were confirmed by surgery and pathology; between 20 and 45 years of age; and had complete preoperative examinations and case records at the hospital. The exclusion criteria were as follows: oral contraceptive pills (COCs) or other hormonal agents within the past 6 months; significant infectious diseases within the past month, extensive use of antibiotics, anticoagulants, or long-term use of nonsteroidal anti-inflammatory drugs; recurrent EMs patients; patients with cardiovascular, liver, lung, kidney diseases, or mental disorders; patients with a history of autoimmune or tumor diseases; patients with genital malformations, adenomyosis, or unexplained abnormal uterine bleeding; patients with acute abdominal conditions such as cyst rupture or torsion; pregnant or lactating women; and patients with incomplete case records. Blood samples of all patients who were diagnosed with pelvic masses by ultrasound at our hospital and met the criteria during the period were collected, and no data imputation was needed. Patient information: we collected data on age, BMI, parity, gravidity, age of menarche and menstrual period, the results revealed no significant differences ( P  > .05) between groups (Table 1 ). CA125 was detected via the chemiluminescence method by the hospital laboratory, while the other 2 indicators were measured via the ELISA method (with a reagent kit from Wuhan Elabscience). Fasting blood samples of 5ml from all patients were collected into serum tubes and immediately sent to the laboratory. The samples were allowed to clot naturally at room temperature (23°C ± 2°C) for 30 minutes until solidification. The samples were then centrifuged at 3000 rpm for 5 minutes, and the supernatant (serum) was stored at −80 °C for later use. The storage conditions and duration for the serum samples were the same for both the observational and control groups. Serum samples were coded with unique Arabic numerals to ensure blinding during laboratory analysis. The laboratory technicians were blinded to the groups, VAS scores and staging information of all samples. Patients with a postoperative pathological diagnosis of EMs were included in the observational group, while those meeting the inclusion and exclusion criteria were included in the control group. This study was approved by the hospital’s ethics committee, and informed consent was obtained from the participants. Comparison of basic clinical characteristics between the 2 groups. BMI = body mass index. Pain was assessed via the visual analog scale (VAS) [ 10 ] : no pain, 0 points; mild pain, 1 to 3 points; moderate pain, 4 to 6 points; and severe pain, 7 to 10 points. The scoring was completed 1 to 2 days before discharge.

The general data of patients in 2 groups were compared and revealed no differences (Table 1 ). The serum levels of CXCL12, VEGF and CA125 in the observational group were significantly greater than those in the control group ( P  < .001; Table 2 ). Comparison of the serum levels of 3 indicators between the 2 groups. CA125 = Carbohydrate antigen 125, CXCL12 = stromal cell-derived factor, VEGF = vascular endothelial growth factor. The serum level of CXCL12 at stage IV was not different from that at the previous stage ( P  = .398), whereas compared with stages I and II, the serum level at stage III was significantly elevated ( P  < .001). VEGF and CA125 were significantly different at different stages ( P  < .001; Table 3 ). Comparison of the levels of serum indices at different disease stages. CA125 = Carbohydrate antigen 125, CXCL12 = stromal cell-derived factor, VEGF = vascular endothelial growth factor. compared with the previous stage, P  < .001. compared with the previous stage, P  = .398. The serum level of CXCL12 in the severe dysmenorrhea group was not different from that in the mild dysmenorrhea group ( P  = .405), whereas compared with that in the mild dysmenorrhea group, the serum level in the moderate dysmenorrhea group was significantly elevated ( P  < .001). VEGF and CA125 were significantly different ( P  < .001; Table 4 ). Comparison of the levels of serum indices among different dysmenorrhea degree. CA125 = Carbohydrate antigen 125, CXCL12 = stromal cell-derived factor, VEGF = vascular endothelial growth factor. compared with the previous stage, P  < .001. compared with the previous stage, P  = .405. In the multivariate logistic regression analysis, patients with EMs was taken as the dependent variable, and BMI, CXCL12, VEGF, and CA125 were taken as independent variables. The results demonstrated that CXCL12, VEGF, and CA125 were independent predictors for EMs (OR = 5..328 [3.132, 10.673], 3.712 [1.512, 9.872], 1.174 [1.038, 1.329)]), and the differences were statistically significant ( P  < .05; Table 5 ). Multivariate logistic regression analysis of patients with EMs. BMI = body mass index. CA125 = Carbohydrate antigen 125, CI = confidence interval, CXCL12 = stromal cell-derived factor, EMs = endometriosis, OR = odds ratio, VEGF = vascular endothelial growth factor. The results showed that the AUC of the ROC for diagnosing the occurrence of EMs with serum CXCL12, VEGF and CA125 alone were 0.8044 [0.7417, 0.863], 0.7224 [0.6412, 0.7931] and 0.6260 [0.5471, 0.6998], respectively. Compared with individual detection, combined detection has greater diagnostic value (AUC = 0.8823 [0.8272, 0.9334]). The combined model significantly improved diagnostic accuracy, achieving a sensitivity of 84.44%, and a specificity of 85.71%, compared with CXCL12 (sensitivity of 75.56%, specificity of 77.55%), VEGF (sensitivity of 79.08%, specificity of 52.04%) and CA125 (sensitivity of 54.44%, specificity of 68.37%) separately (Table 6 ). Diagnostic value of the combined detection of 3 serum indices in EMs. AUC = area under the curve, CI = confidence interval, EMs = endometriosis, VEGF = vascular endothelial growth factor. ROC curves were plotted using the serum CXCL12, VEGF and CA125 levels alone and in combination to predict probability as test variables and the occurrence of EMs as the state variable. As shown in the curves, the combination of the 3 markers yielded a higher AUC compared to any single marker alone (Fig. 1 ). ROC curve analysis was used to evaluate the diagnostic value of individual and combined detection of 3 serum biomarkers for endometriosis. CA125 = carbohydrate antigen 125, CXCL12 = stromal cell-derived factor, ROC = receiver operating characteristic, VEGF = vascular endothelial growth factor.

EMs refers to the presence of endometrial glands and stroma outside the uterine cavity. EMs not only causes chronic pelvic pain and infertility but also imposes a significant economic burden on society. [ 11 ] According to the staging system of the American Society for Reproductive Medicine, EMs can be divided into 4 stages. Incomplete statistics show that EMs affects more than 10% of women of reproductive age, especially those aged 25 to 29 years. Changes in the pelvic microenvironment and anatomy lead to infertility rates as high as 30% to 40%. [ 12 , 13 ] Although EMs is benign, it exhibits malignant biological behavior characterized by local tissue destruction and adhesion, similar to that of malignant tumors. Studies have shown that it takes an average of 4 to 11 years from the onset of EMs to surgical intervention and pathological confirmation, during which the disease subtly changes the anatomical structure and physiological function of the female pelvis. [ 12 ] Therefore, identifying noninvasive biomarkers associated with the diagnosis of EMs and improving the accuracy of early diagnosis are highly important for reducing the physical and psychological burdens on patients. Chemokines are multifunctional small-molecule proteins that can be divided into 4 subfamilies (C, CC, CXC, and CX3C) on the basis of the position and number of cysteine residues at the amino terminus. [ 14 ] CXCL12 of the CXC family can regulate inflammatory responses and promote cell migration, participate in embryonic development and immune regulation, and are associated with stem cell migration, homing, angiogenesis, and extracellular matrix remodeling. [ 15 - 17 ] CXCL12 exerts its biological functions by binding to its specific 7-transmembrane G-protein-coupled receptors CXCR4/CXCR7, activating the downstream PI3K/AKT/mTOR, p38 MAPK and Ras/Raf/MEK/ERK and JAK/STAT pathways. [ 18 , 19 ] In various inflammatory diseases, such as rheumatoid arthritis, inflammatory bowel disease, and systemic sclerosis, the levels of CXCR4 and CXCL12 are also significantly elevated. [ 20 ] Moridi reported that the expression of CXCL12 is significantly upregulated in endometriotic lesions and enhances the chemotactic activity of bone marrow stem cells. [ 21 ] Uzelli Şimşek et al revealed that, compared with normal endometrial stromal stem cells, those derived from endometriotic tissues expressed higher levels of CXCL12, which significantly decreased after treatment with dienogest, approximately in line with the levels in normal tissue. [ 22 ] By measuring serum CXCL12 levels in patients with and without EMs, we found that CXCL12 levels were significantly elevated in patients with EMs ( P  < .001), with significantly higher levels in stage III patients than in stages I and II patients, but there was no significant difference between stages IV and III patients ( P  = .398). In terms of the VAS score, the CXCL12 level was significantly greater in patients with moderate pain than in those with mild pain ( P  < .001), but no significant difference was observed between severe and moderate pain patients ( P  = .405). These findings suggest that CXCL12 levels are correlated with disease progression to some extent and can serve as a biological marker for early disease progression. CA125 is the most extensively studied glycoprotein in EMs. [ 23 ] Previous studies have investigated serum CA125 levels in women with EMs and reported that they are significantly higher than those in healthy women. [ 24 ] Tang et al compared serum CA125 levels between EMs patients and non-EMs patients and reported a significant increase in the EMs group, with stages III and IV representing higher levels than stages I and II, suggesting an association with disease progression. [ 25 ] Szubert et al further confirmed through LASSO regression that CA125 levels are significantly higher in women with EMs than in those without EMs ( P  < .0001). [ 26 ] Chen’s meta-analysis of 14 randomized controlled studies revealed that acupuncture therapy significantly alleviated dysmenorrhea and chronic pelvic pain in EMs patients, with a corresponding decrease in serum CA125 levels, indicating a positive correlation between CA125 levels and pain scores. [ 27 ] Our study revealed that CA125 levels in the EMs group were significantly higher than those in the non-EMs group, showing an increase trend with the progression of the disease. Similarly, CA125 levels gradually increased from mild to severe dysmenorrhea, with statistically significant differences. It is theorized that the elevated serum CA125 levels in EMs patients may be related to retrograde diffusion of the active endometrium through the pelvic and peritoneal cavities, stimulating the proliferation and adhesion of mesothelial cells and promoting lesion spread and disease progression, leading to increased disease stages and severe dysmenorrhea .[ 28 ] VEGF is a dimeric glycoprotein secreted by cells. As important angiogenic factors, it initiates cell proliferation and endothelial cell angiogenesis by activating the downstream PI3K/AKT/mTOR, p38 MAPK, and Ras/Raf/MEK/ERK signaling pathways, increasing vascular permeability and playing a major regulatory role in blood vessel formation. [ 29 , 30 ] VEGF also has immunomodulatory properties, directly or indirectly inhibiting the antitumor activity of immune cells. Its expression and structural changes are associated with cancer progression, overall patient survival, and treatment response. [ 31 , 32 ] In cervical cancer tissues, VEGF expression is significantly elevated and positively correlated with disease stage, lymph node metastasis, and prognosis. [ 30 ] A meta-analysis of 11 studies involving 1251 endometrial cancer patients revealed that patients with high VEGF expression had shorter overall survival, lower cell differentiation, and a greater likelihood of lymph node metastasis, indicating a significant association with disease progression and prognosis. [ 33 ] Given that EMs exhibit biological behaviors similar to those tumors, the abnormal overexpression of VEGF in ectopic endometrial tissues and the peritoneal fluid of EMs patients is widely accepted. [ 34 ] Elevated VEGF expression accelerates the establishment, growth, and persistence of EMs lesions. [ 9 ] Compared with those in pretreatment tissues, VEGF-mRNA expression levels in EMs tissues significantly decreased after GnRH treatment, further confirming the positive correlation between disease progression and VEGF expression. [ 35 ] Our comparison of this biomarker revealed that VEGF levels in the EMs group were significantly greater than those in the non-EMs group, presenting an increasing trend as the disease advanced. Similarly, VEGF levels gradually increased from mild to severe dysmenorrhea, with statistically significant differences, indicating an association with disease progression. For years, clinicians have made great effort to explore noninvasive biomarkers for EMs, including blood, peritoneal fluid, urine, tissue, noncoding ribonucleic acids, small-molecule metabolites, immune cells, and inflammatory factors. [ 36 - 38 ] Compared to the indicators we mentioned, other biomarkers, such as proteomics, metabolomics and genomics, which have a broad information dimension that can reveal disease mechanisms, are considered as the novel perspectives of noninvasive biomarkers, but these technologies need better standardization and are cost and time-intensive. As for miRNA, which has shown promising results, it remains controversial and still equivocal, and it requires complex experimental procedures (such as extraction and reverse transcription first). [ 39 ] In comparison, the detection technology of serum indicators in our study is mature and less expensive, with higher clinical accessibility, making them more favorable for promotion in hospitals. This study investigated the differences in the expression of CXCL12, VEGF, and CA125 between the observational and control groups. The ROC curves indicated that the serum levels of these 3 markers have certain diagnostic value for this disease. Combined detection showed greater diagnostic efficacy for EMs than individual ones. However, not all of the indicators (such as CXCL12) can be performed at all levels of hospitals, some even rely on third-party laboratories, which increases the overall cost and limits the clinical application of this method. Nevertheless, this method provides clinicians with a diagnostic mindset and approach, enabling them to closely monitor the serum expression of these 3 markers, intervene in patients in a timely manner, and improve patient prognosis.

We thank the clinic staff at the Suzhou Ninth People’s Hospital and everyone who helped with recruitment and all the women who participated in the study.