Construction method and diagnosis system of combined diagnosis model of ovarian cancer based on multi-modal ultrasound parameters and serum st2 and fibrinogen

Abstract

The application discloses a kind of based on the construction method and diagnostic system of combined diagnostic model of ovarian cancer of multi-modal ultrasound parameter and serum ST2 and fibrinogen, it is related to multi-modal ultrasound medical technical field, the construction method of combined diagnostic model of ovarian cancer based on multi-modal ultrasound parameter and serum ST2 and fibrinogen includes the following steps: S1, sample data is collected, the sample data includes multi-modal ultrasound parameter and corresponding serum ST2, fibrinogen detection result and disease malignant degree diagnostic information;S2, the sample data is statistically analyzed, and the parameter independently related with ovarian cancer malignant degree is screened out;S3, based on the independent correlation parameter screened out, construct the combined diagnostic model for diagnosing ovarian cancer malignant degree;By collecting multi-modal ultrasound parameter and serum ST2 and fibrinogen detection result comprehensive data, physiological information of ovarian space-occupying lesion patient is comprehensively covered, the accuracy and reliability of the construction method are improved.

Description

Technical Field

[0001] This invention relates to the field of multimodal ultrasound medical technology, and in particular to a method and diagnostic system for constructing a combined diagnostic model of ovarian cancer based on multimodal ultrasound parameters and serum ST2 and fibrinogen. Background Technology

[0002] Ovarian cancer, one of the deadliest malignant tumors of the female reproductive system, ranks third in incidence among gynecological malignancies, but its mortality rate remains the highest. Approximately 70% of patients are diagnosed at stage III-IV, with a 5-year survival rate of only 30%-40%, while early-stage patients have a 5-year survival rate exceeding 90%. This significant difference in prognosis highlights the crucial value of early diagnosis. However, the anatomical characteristics of ovarian cancer, located deep within the pelvic cavity, make its early symptoms insidious, difficult to detect during routine gynecological examinations. Patients often delay diagnosis and treatment due to nonspecific symptoms such as abdominal distension and indigestion, leading to the clinical dilemma of "diagnosis at an advanced stage." Current clinical diagnosis mainly relies on tumor markers such as serum CA125 and HE4, as well as ultrasound examination. However, traditional markers have insufficient specificity; approximately 20% of patients with epithelial ovarian cancer have normal CA125 levels. Conventional ultrasound has limited ability to differentiate between small lesions and benign / malignant lesions, resulting in an early diagnostic sensitivity of only about 60%. Recent studies have found that growth-stimulating gene 2 protein (serum ST2), acting as an IL-33 receptor, is highly expressed in ovarian cancer tissues and participates in disease progression by promoting tumor cell proliferation and inhibiting apoptosis. Its serum level is closely related to clinical stage and prognosis. Fibrinogen, as a coagulation function indicator, is significantly elevated in ovarian cancer patients, with levels in stage III-IV patients being significantly higher than in benign lesions, and is associated with adverse prognostic factors such as ascites formation and distant metastasis. Meanwhile, multimodal ultrasound technology, by integrating tumor morphological characteristics and hemodynamic information, can significantly improve diagnostic sensitivity. Elastography can differentiate tumors of different differentiation degrees by quantifying tissue stiffness, while ultramicrovascular imaging can clearly show the distribution of tumor neovascularization, providing a new dimension for malignancy assessment. Therefore, based on this understanding, this paper proposes a method and system for constructing a combined diagnostic model for ovarian cancer based on multimodal ultrasound parameters, serum ST2, and fibrinogen. By integrating imaging and molecular biological information, this study explores its application value in ovarian cancer malignancy grading, aiming to improve patient prognosis. Summary of the Invention

[0003] In view of this, the present invention addresses the deficiencies of existing technologies, and its main objective is to provide a method and diagnostic system for constructing a combined diagnostic model of ovarian cancer based on multimodal ultrasound parameters and serum ST2 and fibrinogen levels. By collecting comprehensive data such as multimodal ultrasound parameters and serum ST2 and fibrinogen test results, it comprehensively covers the physiological information of patients with ovarian space-occupying lesions, providing rich evidence for accurate diagnosis. Multimodal ultrasound examination assesses lesions from different angles, and combined with serum index detection, it improves the accuracy and reliability of the construction method.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: A method for constructing a combined diagnostic model for ovarian cancer based on multimodal ultrasound parameters and serum ST2 and fibrinogen levels includes the following steps: S1. Collect sample data, which includes multimodal ultrasound parameters of patients with ovarian space-occupying lesions, corresponding serum ST2 and fibrinogen test results, and disease malignancy diagnosis information. S2. Perform statistical analysis on the sample data to screen out parameters that are independently correlated with the malignancy of ovarian cancer. The independently correlated parameters include at least: the mean elastic modulus of shear wave elastography, the contrast-enhanced ultrasound mode, serum ST2 concentration, and serum fibrinogen concentration. S3. Based on the selected independent correlation parameters, construct a joint diagnostic model for diagnosing the malignancy of ovarian cancer.

[0005] As a preferred embodiment, the multimodal ultrasound examination parameters in step S1 are obtained through the following examination modes: color Doppler flow imaging, shear wave elastography, and contrast-enhanced ultrasound; in shear wave elastography, the average elastic modulus of the lesion's solid region is recorded; in contrast-enhanced ultrasound, it is determined and recorded whether the enhancement pattern of the lesion is "fast in and fast out".

[0006] As a preferred embodiment: the statistical analysis in step S2 includes univariate analysis and multivariate analysis; the univariate analysis is used to initially screen candidate indicators that show statistical differences between the ovarian cancer group and the benign lesion group; the multivariate analysis is a multivariate logistic regression analysis, used to identify independent risk factors from the candidate indicators.

[0007] As a preferred approach: the parameters that are independently associated with the malignancy of ovarian cancer in step S2 are specifically screened out by including parameters with a p-value less than 0.05 in the univariate analysis into a multivariate logistic regression model, with the pathological results as the dependent variable, to screen out independent risk factors.

[0008] As a preferred option, the construction of the joint diagnostic model for diagnosing the malignancy of ovarian cancer in step S3 specifically involves using the selected independent relevant parameters as independent variables and establishing a mathematical model to predict the malignancy risk of ovarian cancer through the Logistic regression algorithm.

[0009] As a preferred embodiment, after step S3, a model verification step is also included: S4. Using test set data, evaluate the diagnostic efficacy of the combined diagnostic model through receiver operating characteristic (ROC) curves, wherein the diagnostic efficacy includes at least the area under the curve, sensitivity, and specificity.

[0010] As a preferred option, the combined diagnostic model constructed in step S3 for diagnosing the malignancy of ovarian cancer integrates and evaluates the following indicators: the average tissue elastic modulus obtained based on shear wave elastography, the enhancement pattern determined by ultrasound contrast imaging, serum ST2 concentration, and serum fibrinogen concentration, so as to output the diagnostic result of ovarian cancer malignancy risk.

[0011] As a preferred embodiment: the serum growth-stimulating gene 2 protein concentration in step S2 is obtained by enzyme-linked immunosorbent assay (ELISA), and the serum fibrinogen concentration is obtained by coagulation analyzer.

[0012] A combined diagnostic system for the malignancy degree of ovarian cancer, used to perform the aforementioned construction method, includes: The data acquisition module is used to acquire the data to be evaluated of the target object, including: multimodal ultrasound parameters, serum ST2 concentration and serum fibrinogen concentration; The model processing module is used to receive the data to be evaluated and process and analyze it based on the joint diagnostic model. The results output module is used to output the risk assessment results of the malignancy of ovarian cancer in the target subject.

[0013] As a preferred embodiment, the data acquisition module includes: The ultrasound imaging unit is used to perform multimodal ultrasound examinations and generate average elastic modulus and ultrasound contrast enhancement pattern data for shear wave elastography. The serum detection unit is used to detect serum samples and generate serum ST2 concentration and serum fibrinogen concentration data.

[0014] Compared with the prior art, the present invention has obvious advantages and beneficial effects. Specifically, as can be seen from the above technical solution: First, by collecting comprehensive data including multimodal ultrasound parameters and serum ST2 and fibrinogen test results, the physiological information of patients with ovarian space-occupying lesions was fully covered, providing rich evidence for accurate diagnosis. Multimodal ultrasound examination assesses lesions from different angles, and combined with serum index detection, it improves the accuracy and reliability of the construction method.

[0015] Secondly, statistical analysis was used to screen independently relevant parameters, eliminating the interference of irrelevant factors and making the model more accurate. The combination of univariate analysis and multivariate logistic regression analysis scientifically determined parameters closely related to the malignancy of ovarian cancer, such as the mean elastic modulus of shear wave elastography and the contrast-enhanced ultrasound pattern, laying a solid foundation for the construction method.

[0016] Third, the constructed joint diagnostic model integrates multiple key indicators, effectively outputting diagnostic results regarding the malignancy risk of ovarian cancer. The model validation step further ensures its diagnostic efficacy, providing a more accurate method for constructing diagnostic models by evaluating indicators such as area under the curve, sensitivity, and specificity.

[0017] Fourth, the diagnostic system is highly efficient and accurate: the data acquisition module can comprehensively and accurately acquire the evaluation data of the target object; the combination of multimodal ultrasound examination and serum testing provides rich and reliable information for subsequent analysis. The ultrasound imaging unit and serum testing unit have clearly defined functions, ensuring the accuracy and completeness of the data.

[0018] To more clearly illustrate the structural features and effects of the present invention, a detailed description is provided below in conjunction with the accompanying drawings and specific embodiments. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the ROC curve for diagnosing benign and malignant ovarian cancer using multimodal ultrasound parameters combined with ST2 and FIB according to the present invention. Figure 2 This is a flowchart of the diagnostic system of the present invention. Detailed Implementation

[0020] The present invention is as follows Figures 1 to 2 As shown, a method for constructing a combined diagnostic model for ovarian cancer based on multimodal ultrasound parameters and serum ST2 and fibrinogen includes the following steps: S1. Collect sample data, which includes multimodal ultrasound parameters of patients with ovarian space-occupying lesions, corresponding serum ST2 and fibrinogen test results, and disease malignancy diagnosis information; S2. Perform statistical analysis on the sample data to screen out parameters that are independently correlated with the malignancy of ovarian cancer. Independently correlated parameters include at least: the mean elastic modulus of shear wave elastography, the contrast-enhanced ultrasound pattern, serum ST2 concentration, and serum fibrinogen concentration. S3. Based on the selected independent correlation parameters, construct a joint diagnostic model for diagnosing the malignancy of ovarian cancer.

[0021] In step S1, the multimodal ultrasound examination parameters are obtained through the following examination modes: color Doppler flow imaging, shear wave elastography, and contrast-enhanced ultrasound. In shear wave elastography, the average elastic modulus of the solid region of the lesion is recorded. In contrast-enhanced ultrasound, it is determined and recorded whether the enhancement pattern of the lesion is "fast in and fast out".

[0022] The statistical analysis in step S2 includes univariate analysis and multivariate analysis. The univariate analysis is used to initially screen candidate indicators that show statistical differences between the ovarian cancer group and the benign lesion group. The multivariate analysis is a multivariate logistic regression analysis, which is used to identify independent risk factors from the candidate indicators.

[0023] In step S2, the parameters independently associated with the malignancy of ovarian cancer are screened as follows: parameters with a p-value less than 0.05 in the univariate analysis are included in the multivariate logistic regression model, with the pathological results as the dependent variable, to screen for independent risk factors. The independent risk factors are: the mean elastic modulus of shear wave elastography ≥ 89.5 kPa; the contrast-enhanced ultrasound mode is "fast in, fast out"; serum ST2 concentration ≥ 42.3 ng / mL; and serum fibrinogen concentration ≥ 4.62 g / L.

[0024] In step S3, the construction of a joint diagnostic model for diagnosing the malignancy of ovarian cancer specifically involves using the selected independent relevant parameters as independent variables and establishing a mathematical model to predict the malignancy risk of ovarian cancer through a Logistic regression algorithm.

[0025] Following step S3, a model validation step is also included: S4. Using test set data, evaluate the diagnostic efficacy of the combined diagnostic model through receiver operating characteristic (ROC) curves. The diagnostic efficacy includes at least the area under the curve, sensitivity, and specificity.

[0026] The combined diagnostic model constructed in step S3 for diagnosing the malignancy of ovarian cancer integrates and evaluates the following indicators: the average tissue elastic modulus obtained based on shear wave elastography, the enhancement pattern determined by ultrasound contrast imaging, serum ST2 concentration, and serum fibrinogen concentration, in order to output a diagnostic result of the malignancy risk of ovarian cancer.

[0027] The serum growth-stimulating gene 2 protein concentration in step S2 was obtained by enzyme-linked immunosorbent assay (ELISA), and the serum fibrinogen concentration was obtained by coagulation analyzer.

[0028] A combined diagnostic system for the malignancy of ovarian cancer, and a method for constructing such a system, comprising: The data acquisition module is used to acquire the data to be evaluated of the target object, including: multimodal ultrasound parameters, serum ST2 concentration and serum fibrinogen concentration; The model processing module is used to receive the data to be evaluated and to process and analyze it based on the joint diagnostic model. The results output module is used to output the risk assessment results of the malignancy of ovarian cancer in the target subject.

[0029] The data acquisition module includes: The ultrasound imaging unit is used to perform multimodal ultrasound examinations and generate average elastic modulus and ultrasound contrast enhancement pattern data for shear wave elastography. The serum detection unit is used to detect serum samples and generate serum ST2 concentration and serum fibrinogen concentration data.

[0030] First, the data acquisition module starts operating. The ultrasound imaging unit performs multimodal ultrasound examination on the target object, including color Doppler flow imaging, shear wave elastography, and contrast-enhanced ultrasound, recording the average elastic modulus of the lesion's solid region and determining whether the enhancement pattern of the lesion is "fast in, fast out," generating corresponding multimodal ultrasound parameter data. The serum testing unit tests the serum sample of the target object, detecting serum ST2 concentration using enzyme-linked immunosorbent assay (ELISA) and serum fibrinogen concentration using a coagulation analyzer, generating serum indicator data.

[0031] Next, the data acquisition module transmits the acquired data to be evaluated—namely, multimodal ultrasound parameters, serum ST2 concentration, and serum fibrinogen concentration—to the model processing module. Upon receiving this data, the model processing module performs processing and analysis based on the joint diagnostic model. This joint diagnostic model is constructed by collecting sample data, performing statistical analysis to screen for independently relevant parameters, and using the indicators in the data to be evaluated as independent variables, calculating the results using a logistic regression algorithm.

[0032] Finally, the results output module receives the analysis results from the model processing module and outputs the risk assessment results of the malignancy of the target object's ovarian cancer.

[0033] Example: A method for constructing a combined diagnostic model for ovarian cancer based on multimodal ultrasound parameters and serum ST2 and fibrinogen levels. Materials and Methods: Study subjects: 102 patients with ovarian space-occupying lesions who visited the Third Clinical Medical College (Affiliated Cancer Hospital) of Xinjiang Medical University from April 2024 to June 2025 were selected. All patients underwent postoperative pathological examination for a definitive diagnosis. Based on the pathological results, they were divided into an ovarian cancer group (n=68) and a benign ovarian lesion group (n=34). This application was approved by the hospital's ethics committee, and informed consent was obtained.

[0034] Inclusion criteria: ① Postoperative pathological diagnosis of ovarian cancer or benign lesions, meeting the diagnostic criteria of the "Guidelines for the Diagnosis and Treatment of Malignant Ovarian Tumors (2021 Edition)"; ② Completion of multimodal ultrasound examination and serum ST2 and fibrinogen testing within 1 week before surgery; ③ Complete clinical data (including age, medical history, surgical records, pathology reports, etc.).

[0035] Exclusion criteria: ① Comorbid malignancies or severe heart, liver or kidney dysfunction; ② Previous radiotherapy, chemotherapy, immunotherapy or targeted therapy; ③ Poor ultrasound image quality or hemolysis of serum samples; ④ Missing clinical data or withdrawal from the study.

[0036] Instruments and Methods: A GE LOGIQE20 color Doppler ultrasound diagnostic system (GE Healthcare, USA) was used, equipped with a transvaginal probe (model IC5-9-D, frequency 3-10 MHz) and a transabdominal probe (model C1-6VN, frequency 1-6 MHz). The probe was selected based on the lesion location and patient condition: a transvaginal probe (higher resolution) was preferred for deep pelvic lesions or married patients; a transabdominal probe was used for larger lesions (diameter >8 cm) or unmarried / sexually devoid patients. Before the examination, the bladder should be full (500-800 mL of water intake, approximately 300 mL of urine). The examination was divided into transvaginal and transabdominal examinations. For transvaginal examination: the patient was placed in the lithotomy position, and the probe was adjusted after applying coupling gel to fully visualize the lesion and surrounding tissues. For transabdominal examination: the patient was placed in a supine position, exposing the lower abdomen. After applying coupling gel, a fan-scan technique was used from the pubic symphysis upwards to the umbilicus to ensure clear visualization of both adnexa and the entire uterus.

[0037] Conventional two-dimensional ultrasound (2D-US): The lesion is scanned in multiple planes (transverse, longitudinal, oblique) and the following parameters are recorded: (1) Morphological characteristics: lesion shape (round / elliptical / irregular), boundary (clear / blurred / lobed), internal echo (anechoic / hypoechoic / isoechoic / hyperechoic / mixed echo), presence or absence of calcification (punctate / clustered / septate calcification) and septa (thickness uniformity, presence or absence of papillary protrusions). (2) Lesion size: Measure the three-dimensional diameter (long diameter × transverse diameter × anteroposterior diameter), take the average of 3 measurements, and calculate the volume (V=π / 6×length×transverse×anteroposterior diameter). (3) Ascites assessment: Record the depth of ascites (none / small amount <3 cm / medium amount 3-5 cm / large amount >5 cm) and the distribution range (pelvic cavity / abdominal cavity).

[0038] 1.2.2 Color Doppler flow imaging (CDFI) The color gain was adjusted to a noise-free level, the pulse repetition frequency (PRF) was 0.6-1.0 kHz, the wall filter was moderate (50-100 Hz), and the sampling frame covered the lesion and the surrounding 1 cm area. (1) Blood flow grading standard: Adler semi-quantitative method was used: Grade 0: no blood flow signal in the lesion; Grade I: small amount of blood flow, 1-2 punctate or short rod-shaped vessels; Grade II: moderate amount of blood flow, 3-4 vessels or a single long vessel penetrating the lesion; Grade III: abundant blood flow, ≥5 vessels or diffuse vascular distribution. (2) Hemodynamic parameters: For grade II-III blood flow signals, pulsed Doppler (PW) was enabled, with a sampling volume of 2-3 mm and an angle between the sound beam and the blood flow of <60°. The peak systolic velocity (PSV) and end-diastolic velocity (EDV) were measured, and the resistance index (RI=(PSV-EDV) / PSV) was calculated. Three consecutive cardiac cycles were measured, and the average value was taken.

[0039] Shear wave elastography (SWE): For lesions ≥2 cm in diameter, avoid areas with calcification, liquefaction, and intestinal gas interference. Switch to SWE mode, instruct the patient to hold their breath for 3-5 seconds, and wait for the elastography image to stabilize (quality index ≥80). Draw a region of interest (ROI) in the solid area of ​​the lesion, with a size of 1 / 3-1 / 2 of the lesion area, avoiding necrotic or cystic components. Record the maximum, minimum, mean, and standard deviation (SD) of the elastic modulus in kPa. Repeat the measurement 3 times and take the average.

[0040] 1.2.4 Contrast-enhanced ultrasound (CEUS) Contrast agent preparation: SonoVue (Bracco, Italy) lyophilized powder was dissolved in 5 mL of normal saline and shaken until a uniform suspension was formed with a concentration of 8 μL / mL microbubbles. 2.4 mL of contrast agent was injected via the antecubital vein, followed by a rapid injection of 5 mL of normal saline to flush the tubing. The timer and dynamic storage function (frame rate 15-30 frames / s) were started simultaneously and the data was continuously collected for 3 minutes. (1) Time-intensity curve (TIC): QLAB 10.0 software (Philips) was used to select the lesion solid area and the ipsilateral normal ovarian tissue (control) to generate TIC and record the time to peak (TTP), peak intensity (PI), ascending limb slope (α) and clearance time (WT). (2) Enhancement pattern: According to the enhancement sequence, it is divided into "fast in and fast out" (enhancement earlier than normal tissue, rapid clearance), "fast in and slow out" (early enhancement, slow clearance), and "slow in and slow out" (late enhancement, slow clearance); according to the enhancement uniformity, it is divided into uniform enhancement, uneven enhancement, ring enhancement and no enhancement area.

[0041] All images were reviewed in a double-blind manner by two attending physicians with more than 5 years of experience in gynecological ultrasound. If opinions differed, the images were reviewed and confirmed by an associate chief physician. Image quality grading: Grade I (Excellent): Lesion boundaries are clear, and internal structures and blood flow signals are fully displayed; Grade II (Good): Lesion boundaries are partially blurred, but key features are identifiable; Grade III (Poor): Image artifacts are severe, and the image should be discarded and re-examined (re-examination ≤ 2 times).

[0042] Serum biomarker detection: 5 mL of fasting venous blood was collected from the patient, centrifuged, and the serum was stored at -80℃ (3000 r / min, 10 min). ST2 levels were detected using enzyme-linked immunosorbent assay (ELISA) (kit purchased from R&D Systems, USA), fibrinogen (FIB) levels were detected using a coagulation analyzer (Sysmex CA-7000, Japan), and CA125 was detected using electrochemiluminescence immunoassay (Roche Cobas e601, Switzerland). All procedures were strictly performed according to the kit instructions.

[0043] Statistical analysis: SPSS 26.0 software and GraphPad Prism 9.0 were used for plotting. Normally distributed continuous data were expressed as (x̅±s), and independent samples t-tests were used for comparisons between groups; non-normally distributed data were analyzed using M(P25, P75), and Mann-Whitney U tests were used for comparisons. Categorical data were expressed as percentages (%), and chi-square tests were used for comparisons. Spearman analysis was used to analyze parameter correlation, multivariate logistic regression was used to screen for independent risk factors, and ROC curves were used to assess diagnostic efficacy (AUC, sensitivity, and specificity were calculated). P < 0.05 was considered statistically significant.

[0044] Discussion of Results: Comparison of clinicopathological data: There was no significant difference in the incidence of comorbidities between the two groups. The average age of the ovarian cancer group was significantly higher than that of the benign lesion group. In terms of pathological types, high-grade serous carcinoma was the most common in the ovarian cancer group, while ovarian cysts were the most common in the benign group. Regarding the degree of differentiation, poorly differentiated ovarian cancer accounted for 60.3% and moderately to well differentiated ovarian cancer accounted for 39.7% (see Table 1 for details).

[0045] Table 1: Comparison of clinicopathological data and comorbidities between the two groups of patients index Ovarian cancer group (n=68) Benign lesion group (n=34) Statistical value p-value Age (years, x±s) 54.2±8.7 42.6±7.3 t=6.832 <0.001 hypertension 23（33.8） 9（26.5） χ²=0.725 0.394 diabetes 15（22.1） 6（17.6） χ²=0.328 0.567 Coronary heart disease 8（11.8） 3（8.8） χ²=0.221 0.638 Chronic kidney disease 4（5.9） 1（2.9） χ²=0.556 0.456 FIGO installments (e.g., %) Phase I-II: 21 (30.9) - - - Stages III-IV: 47 (69.1) - - - Pathological type (e.g., %) Serous carcinoma: 50 (73.5) Ovarian cysts: 22 (64.7) χ²=42.157 <0.001 Mucinous carcinoma: 8 (11.8) Teratomas: 8 (23.5) Endometrioid carcinoma: 6 (8.8) Chocolate cysts: 4 (11.8) Other: 4 (5.9) - Degree of differentiation (e.g., %) High-to-medium differentiation: 27 (39.7) - - - Low differentiation: 41 (60.3) - - - Comparison of multimodal ultrasound parameters between the two groups: Table 2 shows the comparison results of multimodal ultrasound parameters between the ovarian cancer group and the benign lesion group. Except for RI value and Emean, the peak systolic velocity (PSV) and end-diastolic velocity (EDV) of the ovarian cancer group were significantly higher than those of the benign lesion group, while the maximum and minimum elastic moduli (Emax and Emin) were also significantly increased. In contrast-enhanced ultrasound (CEUS) parameters, the ovarian cancer group had a shorter time to peak (TTP), higher peak intensity (PI), and the enhancement pattern was predominantly "rapid in-and-out" and "heterogeneous enhancement."

[0046] Table 2: Comparison of two sets of multimodal ultrasound parameters Ultrasound parameters Ovarian cancer group (n=68) Benign lesion group (n=34) Statistical value p-value Color Doppler flow imaging (CDFI) Adler flow grading (e.g., %) Level II-III: 57 (83.8) Level II-III: 8 (23.5) χ²=38.251 <0.001 PSV (cm / s, x̅±s) 35.6±8.2 18.3±5.7 t=10.257 <0.001 EDV (cm / s, x̅±s) 12.4±3.6 5.2±2.1 t=9.832 <0.001 RI value (x̅±s) 0.48±0.07 0.72±0.09 t=-14.362 <0.001 Shear wave elastography (SWE) Emax(kPa,M(P25,P75)) 125.3（108.6，142.5） 45.2（38.7，52.6） Z=-6.894 <0.001 Emin(kPa,M(P25,P75)) 62.4（53.8，71.2） 22.5（18.3，26.7） Z=-5.983 <0.001 Emean(kPa,M(P25,P75)) 89.5（76.3，102.4） 32.6（25.1，40.8） Z=-7.215 <0.001 Contrast-enhanced ultrasound (CEUS) TTP(s, x̅±s) 18.5±4.2 28.3±5.6 t=-8.762 <0.001 PI (dB, x̅±s) 25.6±6.3 15.2±4.8 t=7.934 <0.001 Enhanced CEUS mode (e.g., %) Quick in and quick out 43（63.2） 3（8.8） χ²=42.857 <0.001 Uneven reinforcement 55（80.9） 5（14.7） χ²=58.362 <0.001 Comparison of serum marker levels between the two groups: Serum ST2, FIB, and CA125 levels in the ovarian cancer group were significantly higher than those in the benign lesion group (Table 3). Specifically, the ST2 level in the ovarian cancer group [42.3 (35.6, 51.8) ng / mL] was 1.88 times that in the benign lesion group [22.5 (18.7, 26.9) ng / mL] (Z=-6.934, P<0.001); the FIB level (4.62±0.85 g / L) was significantly higher than that in the benign lesion group (2.85±0.53 g / L) (t=11.274, P<0.001); and the CA125 level [486.5 (213.8, 752.6) U / mL] was also significantly higher than that in the benign lesion group [35.2 (22.6, 48.7) U / mL] (Z=-7.582, P<0.001).

[0047] Table 3: Comparison of serum marker levels between the two groups Serum markers Ovarian cancer group (n=68) Benign lesion group (n=34) Statistical value p-value ST2 (ng / mL, M (P25, P75)) 42.3（35.6，51.8） 22.5（18.7，26.9） Z=-6.934 <0.001 FIB (g / L, x̅±s) 4.62±0.85 2.85±0.53 t=11.274 <0.001 CA125 (U / mL, M (P25, P75)) 486.5（213.8，752.6） 35.2（22.6，48.7） Z=-7.582 <0.001 Multivariate logistic regression analysis for ovarian cancer diagnosis: With pathological results as the dependent variable (ovarian cancer = 1, benign lesion = 0), indicators with P < 0.05 in univariate analysis (Emean, ST2, FIB, CEUS enhancement pattern) were included in the multivariate logistic regression model. Results showed that Emean ≥ 89.5 kPa, ST2 ≥ 42.3 ng / mL, FIB ≥ 4.62 g / L, and the CEUS "fast in, fast out" pattern were independent risk factors for ovarian cancer (all P < 0.05), with the CEUS enhancement pattern showing the highest OR (9.826).

[0048] Table 4: Multivariate Logistic Regression Analysis for Ovarian Cancer Diagnosis project b SE Wald χ2 value p-value OR 95% CI Emean (≥89.5 kPa) 1.832 0.526 12.368 0.001 6.257 2.183-17.925 ST2 (≥42.3 ng / mL) 1.564 0.487 10.425 0.001 4.773 1.782-12.764 FIB (≥4.62 g / L) 1.258 0.432 8.572 0.003 3.516 1.452-8.517 CEUS Enhanced Mode (Fast Forward & Fast Out) 2.284 0.615 13.826 <0.001 9.826 3.057-31.562 ROC curve analysis: The ROC curve analysis results for each indicator and the combined model are shown in Table 5. The AUC of the combined model (Emean+ST2+FIB+CEUS enhancement mode) is 0.963, which is significantly higher than any single indicator. The sensitivity corresponding to its optimal cutoff value combination is 92.6%, the specificity is 91.2%, and the diagnostic efficacy is optimal.

[0049] Table 5: Diagnostic value analysis of multimodal ultrasound parameters combined with ST2 and FIB variable Optimal cutoff value AUC Standard error p-value 95% CI Sensitivity (%) Specificity (%) Emean (kPa) ≥89.5 0.876 0.032 <0.001 0.813-0.939 83.8 79.4 ST2 (ng / mL) ≥42.3 0.852 0.035 <0.001 0.784-0.920 80.9 76.5 FIB (g / L) ≥4.62 0.825 0.038 <0.001 0.751-0.899 77.9 73.5 CEUS Enhanced Mode (Fast Forward & Fast Out) - 0.883 0.030 <0.001 0.824-0.942 85.3 82.4 Joint model - 0.963 0.018 <0.001 0.928-0.998 92.6 91.2 This application integrates multimodal ultrasound parameters with serum ST2 and FIB indicators to construct a joint diagnostic model for the malignancy of ovarian cancer. The results show that the diagnostic efficacy of this model is significantly better than that of a single indicator, providing a new approach for the early non-invasive diagnosis of ovarian cancer.

[0050] In this application, multimodal ultrasound technology integrates morphological, hemodynamic, and tissue stiffness information to achieve a comprehensive assessment of the biological characteristics of ovarian cancer. Color Doppler flow imaging (CDFI) results showed that the proportion of Adler flow grades II-III in the ovarian cancer group was significantly higher than in the benign group, and the RI value was significantly lower. This result is closely related to the tumor angiogenesis mechanism: ovarian cancer cells induce angiogenesis by highly expressing pro-angiogenic factors such as vascular endothelial growth factor (VEGF) and fibroblast growth factor (bFGF). These vessels are characterized by irregular lumens and incomplete basement membranes, leading to decreased blood flow resistance. Furthermore, the significantly elevated PSV and EDV in the ovarian cancer group indicated a high perfusion state of tumor tissue, consistent with clinical observations of advanced-stage patients frequently developing ascites and distant metastasis. Shear wave elastography (SWE), as a technique for quantifying tissue stiffness, demonstrated excellent differential diagnostic capabilities in this application. The mechanism involves two aspects: ① Excessive proliferation of tumor cells leads to increased cell density and deposition of components such as collagen and fibronectin in the extracellular matrix, increasing tissue stiffness; ② Upregulation of hypoxia-inducible factor (HIF-1α) in the tumor microenvironment promotes the expression of fibrosis-related genes (such as TGF-β1), further increasing tissue elastic modulus. The "rapid in-and-out" and "heterogeneous enhancement" modes of contrast-enhanced ultrasound (CEUS) accounted for 63.2% and 80.9% of the ovarian cancer group, respectively. The pathological basis lies in the heterogeneity of tumor angiogenesis: the microvessel density (MVD) of malignant tumors is significantly higher than that of benign lesions, and the vascular course is disordered, with arteriovenous shunts forming, leading to rapid filling and clearing of the contrast agent. In this application, the TTP in the ovarian cancer group was significantly shorter than that in the benign group, while the PI was significantly increased. This is consistent with the characteristics of high perfusion and high metabolism in tumor tissue and can serve as a potential indicator for assessing the efficacy of chemotherapy—studies have confirmed that prolonged TTP and decreased PI after chemotherapy indicate decreased tumor cell activity.

[0051] Serum ST2, as an IL-33 receptor, plays a dual role in the ovarian cancer microenvironment: on the one hand, serum ST2 promotes tumor cell proliferation and inhibits apoptosis by activating the NF-κB pathway. In this application, the serum ST2 level in the ovarian cancer group was 1.88 times that in the benign group, and the level in stage III-IV patients (48.7 ng / mL) was significantly higher than that in stage I-II, confirming its correlation with disease progression; on the other hand, serum ST2 can inhibit anti-tumor immunity by recruiting M2 macrophages and regulating Treg cell activity, and blocking the IL-33 / ST2 pathway can reduce the number of tumor-infiltrating lymphocytes and inhibit tumor growth. Fibrinogen (FIB), a key component of the coagulation-fibrinolysis system, has multiple clinical implications when elevated in ovarian cancer: ① In this study, FIB levels in the ovarian cancer group were significantly higher than in the benign group. The mechanism may be related to the release of tissue factor (TF) by tumor cells, which activates the extrinsic coagulation pathway, leading to the conversion of FIB into fibrin, promoting ascites coagulation and adhesion formation; ② FIB can enhance tumor cell adhesion by binding to integrin αvβ3. Multivariate regression analysis in this study showed that FIB ≥ 4.62 g / L was an independent risk factor for ovarian cancer (OR = 3.516); ③ The complementarity of FIB with traditional biomarkers. Compared with CA125, FIB has higher sensitivity in early ovarian cancer, especially providing a diagnostic model construction method for CA125-negative patients (approximately 20% of epithelial ovarian cancer).

[0052] Multimodal ultrasound provides intuitive information on tumor morphology, blood flow, and stiffness, while serum ST2 and FIB reflect the tumor immune microenvironment and coagulation abnormalities. Combining these two methods allows for a comprehensive assessment of disease progression. The innovations of this approach are: ① It is the first to integrate SWE elasticity parameters with ST2 / FIB; ② Quantitative analysis of CEUS enhancement modes improves the objectivity of subjective judgment; ③ Multivariate regression screening of independent variables reduces collinearity interference.

[0053] The method for constructing a combined diagnostic model for ovarian cancer based on multimodal ultrasound parameters and serum ST2 and fibrinogen provided in this application offers a non-invasive and efficient tool for early screening of ovarian cancer. In clinical practice, for patients with pelvic masses detected by ultrasound, Emean, CEUS mode, and serum ST2 / FIB can be prioritized for testing. If the combined model is positive, it is recommended to perform laparoscopic exploration and pathological biopsy as early as possible to achieve "early detection and early intervention" and improve patient prognosis.

[0054] The key design focus of this invention is: First, by collecting comprehensive data including multimodal ultrasound parameters and serum ST2 and fibrinogen test results, the physiological information of patients with ovarian space-occupying lesions was fully covered, providing rich evidence for accurate diagnosis. Multimodal ultrasound examination assesses lesions from different angles, and combined with serum index detection, it improves the accuracy and reliability of the construction method.

[0055] Secondly, statistical analysis was used to screen independently relevant parameters, eliminating the interference of irrelevant factors and making the model more accurate. The combination of univariate analysis and multivariate logistic regression analysis scientifically determined parameters closely related to the malignancy of ovarian cancer, such as the mean elastic modulus of shear wave elastography and the contrast-enhanced ultrasound pattern, laying a solid foundation for the construction method.

[0056] Third, the constructed joint diagnostic model integrates multiple key indicators, effectively outputting diagnostic results regarding the malignancy risk of ovarian cancer. The model validation step further ensures its diagnostic efficacy, providing a more accurate method for constructing diagnostic models by evaluating indicators such as area under the curve, sensitivity, and specificity.

[0057] Fourth, the diagnostic system is highly efficient and accurate: the data acquisition module can comprehensively and accurately acquire the evaluation data of the target object; the combination of multimodal ultrasound examination and serum testing provides rich and reliable information for subsequent analysis. The ultrasound imaging unit and serum testing unit have clearly defined functions, ensuring the accuracy and completeness of the data.

[0058] The above description is merely a preferred embodiment of the present invention and does not constitute any limitation on the technical scope of the present invention. Therefore, any minor modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A method for constructing a combined diagnostic model for ovarian cancer based on multimodal ultrasound parameters and serum ST2 and fibrinogen, characterized in that: Includes the following steps: S1. Collect sample data, which includes multimodal ultrasound parameters of patients with ovarian space-occupying lesions, corresponding serum ST2 and fibrinogen test results, and disease malignancy diagnosis information. S2. Perform statistical analysis on the sample data to screen out parameters that are independently correlated with the malignancy of ovarian cancer. The independently correlated parameters include at least: the mean elastic modulus of shear wave elastography, the contrast-enhanced ultrasound mode, serum ST2 concentration, and serum fibrinogen concentration. S3. Based on the selected independent correlation parameters, construct a joint diagnostic model for diagnosing the malignancy of ovarian cancer.

2. The construction method according to claim 1, characterized in that: In step S1, the multimodal ultrasound examination parameters are obtained through the following examination modes: color Doppler flow imaging, shear wave elastography, and contrast-enhanced ultrasound. In shear wave elastography, the average elastic modulus of the lesion's solid region is recorded. In contrast-enhanced ultrasound, it is determined and recorded whether the enhancement pattern of the lesion is "fast in and fast out".

3. The construction method according to claim 1, characterized in that: The statistical analysis in step S2 includes univariate analysis and multivariate analysis; the univariate analysis is used to initially screen candidate indicators that show statistical differences between the ovarian cancer group and the benign lesion group; the multivariate analysis is a multivariate logistic regression analysis, used to identify independent risk factors from the candidate indicators.

4. The construction method according to claim 3, characterized in that: The specific steps in step S2 for screening parameters that are independently associated with the malignancy of ovarian cancer are as follows: parameters with a p-value less than 0.05 in the univariate analysis are included in the multivariate logistic regression model, with the pathological results as the dependent variable, to screen out independent risk factors.

5. The construction method according to claim 1, characterized in that: The construction of the joint diagnostic model for diagnosing the malignancy of ovarian cancer in step S3 specifically involves: using the selected independent relevant parameters as independent variables, and establishing a mathematical model to predict the malignancy risk of ovarian cancer through the Logistic regression algorithm.

6. The construction method according to claim 1, characterized in that: Following step S3, a model validation step is also included: S4. Using test set data, evaluate the diagnostic efficacy of the combined diagnostic model through receiver operating characteristic (ROC) curves, wherein the diagnostic efficacy includes at least the area under the curve, sensitivity, and specificity.

7. The construction method according to claim 2, characterized in that: The combined diagnostic model constructed in step S3 for diagnosing the malignancy of ovarian cancer integrates and evaluates the following indicators: the average tissue elastic modulus obtained based on shear wave elastography, the enhancement pattern determined by ultrasound contrast imaging, serum ST2 concentration, and serum fibrinogen concentration, in order to output a diagnostic result of the malignancy risk of ovarian cancer.

8. The construction method according to claim 1, characterized in that: The serum growth-stimulating gene 2 protein concentration in step S2 was obtained by enzyme-linked immunosorbent assay (ELISA), and the serum fibrinogen concentration was obtained by coagulation analyzer.

9. A combined diagnostic system for the malignancy of ovarian cancer, used to perform the construction method according to any one of claims 1-8, characterized in that: include: The data acquisition module is used to acquire the data to be evaluated of the target object, including: multimodal ultrasound parameters, serum ST2 concentration and serum fibrinogen concentration; The model processing module is used to receive the data to be evaluated and process and analyze it based on the joint diagnostic model. The results output module is used to output the risk assessment results of the malignancy of ovarian cancer in the target subject.

10. The combined diagnostic system for ovarian cancer malignancy according to claim 9, characterized in that: The data acquisition module includes: The ultrasound imaging unit is used to perform multimodal ultrasound examinations and generate average elastic modulus and ultrasound contrast enhancement pattern data for shear wave elastography. The serum detection unit is used to detect serum samples and generate serum ST2 concentration and serum fibrinogen concentration data.

Images