Contrast-enhanced ultrasound quantitative parameters for predicting benign and malignant greater omentum lesions: an applied study

Abstract

Objective

To explore the value of contrast-enhanced ultrasound (CEUS) quantitative parameters in differentiating benign and malignant diseases of the greater omentum and to establish a new CEUS Greater Omentum Imaging Reporting and Data System (CEUS-GOIRADS).

Methods

We analyzed 15 CEUS parameters from 124 patients undergoing omental biopsy (2020–2024) to develop a noninvasive predictive model and assess its diagnostic value. Statistically significant parameters were further analyzed using receiver operating characteristic (ROC) curve cutoff values to determine odds ratios (ORs) for each group. A CEUS-OS was developed for each patient, and ROC curve analysis was performed to evaluate its efficacy in distinguishing between benign and malignant omental diseases.

Results

Among 124 patients meeting the inclusion criteria, 76 had malignant lesions and 48 had benign lesions. Multivariate analysis identified three out of 15 CEUS parameters—IMAX, Fs50, and WioAUC—as associated with malignant lesions of the greater omentum. The noninvasive predictive model demonstrated a sensitivity of 90.7% and specificity of 90.5% for diagnosing malignant omental lesions. Furthermore, the established CEUS-GOIRADS grading system showed a sensitivity of 94.7% and specificity of 89.6% for differentiating between benign and malignant omental diseases.

Conclusion

Quantitative analysis of CEUS parameters provides effective indicators for determining the nature of greater omental lesions. This non-invasive imaging modality can reliably predict whether omental lesions are benign or malignant. The CEUS-GOIRADS demonstrates high sensitivity and specificity in differentiating benign from malignant omental diseases, suggesting its potential as a CEUS-GOIRADS valuable tool in the multi-step diagnostic evaluation of patients with various omental pathologies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40644-025-00977-8.

Introduction

The omentum is one of the less frequently focused-on “organs” in the abdomen, and there is limited research on the anatomy and pathological mechanisms of the greater omentum [1]. The normal structure of the greater omentum is composed of fat, fibrous tissue, blood vessels, nerves, and smooth muscle [2]. In individuals with a lean body type, the omentum is not easily visualized using conventional ultrasound. However, when the greater omentum is affected by lesions or when tumors or inflammation in the abdominal cavity involve the greater omentum, pathological changes can occur in its structure, making it visible on conventional ultrasound [3]. In a 2021 study, Chen et al. found that the following features could serve as reliable indicators for malignant omental lesions: hypoechoic omentum with dense or moth-eaten structure, nodularity, heterogeneous ascites without septations, non-visualization of mesenteric lymph nodes, and absence of parietal peritoneal thickening. Based on these findings, they established a Greater Omental Imaging and Data Reporting System to facilitate standardized interpretation of ultrasound features and improve the diagnostic accuracy of omental diseases, thereby promoting the clinical application of ultrasound in omental pathology [4]. With continuous advancements in imaging technology, enhancing the visualization of omental anatomy and pathology through multimodal imaging approaches has become crucial for the precise diagnosis and treatment of omental diseases.

Currently, CEUS has been widely adopted for disease diagnosis across various organ systems. By utilizing intravenous contrast agents, CEUS enables dynamic evaluation of macro- and microvascular architecture with high spatial and temporal resolution [5]. This technique has been extensively applied for differentiating benign and malignant lesions in organs such as the thyroid [6], liver [7], ovaries [8], and pancreas [9]. In previous studies, we demonstrated that CEUS can effectively distinguish the greater omentum from surrounding small bowel loops and visualize perfusion patterns within the omentum. These perfusion enhancement characteristics allow for reliable differentiation between benign and malignant omental lesions [10]. Most recently, Chen et al. further confirmed the clinical utility, safety, and diagnostic efficacy of CEUS in evaluating omentum-related pathologies [11].

However, the current assessment of diseases through CEUS still relies on the visual judgment of the examiner [12]. Quantitative analysis of time-intensity curves can provide us with quantitative parameters, thereby enabling computer-aided diagnosis and quantifying tissue and tumor enhancement. This study aims to determine the benign or malignant nature of greater omentum lesions through quantitative CEUS research, and attempts to establish a non-invasive predictive model and construct a CEUS-based imaging and data reporting system for the greater omentum. We refer to this system as the CEUS GOI-RADS.

Materials and methods

This study was approved by the Ethics Committee of the First Affiliated Hospital of China Medical University (AF-SOP-07-1.2-01). Given the retrospective nature of this study, the requirement for written informed consent was waived.

Study subjects

A retrospective analysis was conducted on the CEUS dynamic images of 124 patients who underwent omental contrast imaging and biopsy at the Department of Ultrasound, the First Affiliated Hospital of China Medical University, between October 2020 and October 2024. The cohort included 94 females and 30 males (age range: 19–86 years, median age: 62 years), and all patients had valid tissue and pathological results. Using analytical software, the following quantitative CEUS parameters were obtained through post-processing of the dynamic contrast images: IMAX, RT, TTP, mTT, Rs50, Rs1090, FT, Fs50, FHT, AUC, WinAUC, WoutAUC, WioAUC, WinR, and WoutR.

Instruments and methods

Contrast-enhanced ultrasound examination

Patients were positioned supine with both arms raised above the head to fully expose the upper abdomen. The examinations were performed using Siemens ACUSON Sequoia (Germany, Munich) and Mindray Neuwa R9 Exp (China, Shenzhen), both equipped with abdominal convex array probes (1–5 MHz). The contrast agent used was the second-generation ultrasound contrast agent SonoVue (Bracco, Italy), which was mixed with 5 mL of saline by shaking to form a suspension. The observation period lasted for more than 2 min, and the images were stored in the system.

Post-processing methods

The dynamic contrast-enhanced images in DICOM format were exported from the instruments and analyzed using SonoLiver software (Tomtec, Munich, Germany) to delineate regions of interest (ROIs). When the enhancement of the greater omentum lesion was consistent with other areas, the well-perfused region was delineated. When the enhancement of the lesion was inconsistent with other areas, separate ROIs were drawn for the lesion and normal regions. The delineator was required to have more than 10 years of experience in abdominal ultrasound. After delineation, quantitative analysis was performed to obtain time-intensity curves (TICs) (Fig. 1).

Fig. 1.

Lesion delineation and TIC curve

Scoring method (Table 1)

Table 1.

Number of patients and scoring for parameters in each segment

Parameter	Malignant	Benign	OR Value	Score
Imax
>110.37	74	1	1739	3
<110.37	2	47	0.00	0
Rs50
>17.48	46	10	7.43	2
<17.48	26	42	0.13	0
Rs1090
>15.07	39	19	6.41	2
<15.07	16	50	0.16	0
Fs50
>−1.615	18	36	0.11	0
<−1.615	61	14	8.71	2
AUC
>46931.325	59	11	9.88	2
<46931.325	19	35	0.10	0
WinAUC
>633.63	36	23	2.34	1
<633.63	26	39	0.43	1
WoutAUC
>33793.56	36	15	6.36	2
<33793.56	20	53	0.16	0
WioAUC
>37595.27	30	13	2.88	1
<37595.27	36	45	0.35	1
WinR
>54.65	69	18	7.97	2
<54.65	12	25	0.13	1
WoutR
>485.545	68	6	102	3
<485.545	5	45	0.00	0

If the odds ratio (OR) is less than 0.3, it is considered unrelated to malignancy. If 0.3 < OR ≤ 3, it is considered to have no significant correlation with malignancy. If 3 < OR ≤ 10, it is considered to have a strong correlation with malignancy. If OR > 10, it is considered to have an extremely strong correlation with malignancy. In the above four scenarios, scores of 0, 1, 2, or 3 are assigned respectively. In this study, the optimal cutoff values of the significant parameters from the univariate analysis were used as the thresholds. The cutoff values for IMAX, Rs50, Rs1090, Fs50, AUC, WinAUC, WoutAUC, WioAUC, WinR, and WoutR were 110.37, 17.48, 15.07, -1.615, 46931.325, 633.63, 33793.56, 37595.27, 54.65, and 485.545, respectively.

Statistical methods

Statistical analyses used SPSS 27.0. Continuous variables were expressed as mean ± SD or median with interquartile range (M (Q1, Q3)). Univariate analysis employed t-tests or non-parametric tests; multivariate analysis used logistic regression (P < 0.05 significant). ROC curves identified optimal cutoffs and ORs for the omental scoring system. The benign and malignant proportions of lesions were further stratified, and the area under the ROC curve (AUC) was calculated to assess the diagnostic performance of the new grading system for malignant omental lesions.

Results

Pathological results

A total of 124 patients were divided into two groups based on pathological results: the benign omental lesion group consisted of 48 patients (38.7%, 48/124), including 30 cases of omental tuberculosis and 18 cases of omental inflammation. There were 16 males and 32 females in this group. The malignant lesion group included 76 patients (61.3%, 76/124), comprising 47 cases from gynecological diseases, 22 cases from gastrointestinal diseases, 2 cases of mesothelioma, 4 cases of lymphoma, and 1 case from breast origin. There were 14 males and 62 females in this group (Table 2).

Table 2.

Clinical and pathological analysis of 124 cases of omental lesions

Pathological Type	Number of Patients	Rate	Gender		Age
			Male	Female	Minimum	Maximum	Median
Malignant Omental Lesions
Gynecological disease origin	47	61.84%		47	39	86	63
Gastrointestinal disease origin	22	28.95%	13	9	27	77	57
Breast origin	1	1.32%		1	44	44	44
Mesothelioma	2	2.63%	1	1	67	76	-
Lymphoma	4	5.26%	0	4	55	68	55
Benign Omental Lesions
Tuberculosis	30	62.50%	13	17	19	83	59
Inflammation	18	37.50%	3	15	33	76	67

Quantitative analysis of omental CEUS parameters based on TIC: univariate analysis results

TICs were generated by analyzing the CEUS dynamic images of all patients in the benign and malignant groups. A comparison of the quantitative parameters (Table 3) revealed that IMAX, Rs50, Rs1090, Fs50, AUC, WinAUC, WoutAUC, WioAUC, WinR, and WoutR showed statistically significant differences in distinguishing between benign and malignant omental lesions. In contrast, TTP, RT, FHT, FT, and mTT exhibited no significant differences between the malignant and benign groups (P > 0.05). Further comparison between the two groups demonstrated that the malignant group had significantly higher values for IMAX, Rs50, Rs1090, AUC, WinAUC, WoutAUC, WioAUC, WinR, and WoutR, while TTP, RT, FT, and Fs50 were significantly lower compared to the benign group.

Table 3.

Comparison of quantitative parameters of contrast-enhanced ultrasound between benign and malignant omental lesions

	Benign	Malignant	Z/t	P
IMAX	73.01(65.51,83.04)	142.54(123.72,166.46)	9.090	<0.001**
TTP	11.07(7.70,16.23)	9.92 (7.55,12.43)	1.449	0.147
RT	10.33(7.48,14.34)	9.49(7.22,11.47)	1.621	0.105
Rs50	11.14(7.91,15.96)	28.12(20.39,36.44)	7.206	<0.001**
Rs1090	10.33(7.16,13.58)	23.78(17.89,30.72)	7.237	<0.001**
FT	55.65 ± 30.96	52.32 ± 24.55	0.657	0.512
Fs50	-1.25(-1.86,-0.85)	-2.82(-3.68,-1.86)	6.009	<0.001**
FHT	26.47 ± 13.57	23.26 ± 10.02	1.497	0.137
mTT	44.36(32.83,74.90)	44.06(29.14,74.95)	0.567	0.570
AUC	35225.80(24849.43,49611.05)	53555.79(34324.50,70184.95)	3.182	<0.001**
WinAUC	354.99(195.11,587.91)	894.04(469.51,2047.49)	4.855	<0.001**
WoutAUC	22584.29(16489.87,29421.62)	35839.78(22823.73,52934.24)	5.140	<0.001**
WioAUC	24066.30(17279.92,33445.26)	37787.83(24379.07,54997.20)	4.813	<0.001**
WinR	52.93(21.93,150.45)	103.24(58.56,197.96)	2.504	0.012*
WoutR	343.16(218.68,417.94)	732.18(623.89,839.63)	8.040	<0.001**

Note: * indicates p < 0.05, ** indicates p < 0.01

Multivariate analysis results of CEUS parameters for greater omentum lesions

Before performing multivariate logistic regression, collinearity analysis was conducted on all indicators that were significant in the univariate analysis to select appropriate variables for the multivariate logistic regression. The multivariate analysis revealed that IMAX, Fs50, and WioAUC were independent influencing factors for differentiating malignant omental lesions (Table 4. The ROC curve was established based on the results of the logistic regression analysis, with an AUC of 0.955. The sensitivity and specificity for diagnosing malignant omental lesions were 90.7% and 90.5%, respectively (Fig. 2). Rs50, Rs1090, AUC, WinAUC, and WinR were not independent influencing factors for malignant greater omentum lesions (with p-values of 0.102, 0.364, 0.910, 0.241, and 0.350, respectively).

Table 4.

Logistic multivariate regression analysis of quantitative parameters of contrast-enhanced ultrasound for diagnosing malignant omental lesions

Parameter	B	Standard Error	Wald Value	Degrees of Freedom	P	OR	95% CI for EXP(B)
							LowerLimit Upper Limit
Imax	0.133	0.049	7.439	1	0.006	1.142	1.038 1.257
Rs50	0.382	0.234	2.679	1	0.102	1.466	0.927 2.317
Rs1090	-0.216	0.238	0.825	1	0.364	0.806	0.506 1.284
Fs50	-1.286	0.640	4.032	1	0.045	0.276	0.079 0.970
AUC	0.000	0.000	0.013	1	0.910	1.000	1.000 1.000
WinAUC	0.001	0.001	1.374	1	0.241	1.001	1.000 1.002
WioAUC	0.000	0.000	7.813	1	0.005	1.000	1.000 1.000
WinR	-0.027	0.029	0.873	1	0.350	0.973	0.92 1.030

Fig. 2.

ROC curve of the predictive model

Diagnostic value of omental CEUS scoring in differentiating benign and malignant omental lesions

Based on the assigned scores, we calculated the Contrast-Enhanced Ultrasound Omental Score (CEUS-OS) for each patient and determined the malignancy rate at different score thresholds (Table 5). Subsequently, we constructed an ROC curve (Fig. 3).The optimal cutoff value for distinguishing between benign and malignant omental lesions was 10.5 points, with a sensitivity of 94.70% and a specificity of 89.60%.

Table 5.

Evaluation results and malignancy rate of benign and malignant omental lesions

		Score
		≤ 8	9	10	11	12	13	14	≥ 15
Benign(n)		38	3	2	1	2	1	1	0
Malignant(n)		0	1	3	4	9	9	15	37
Malignancy Rate		0	0.25	0.60	0.80	0.82	0.90	0.94	1

Fig. 3.

ROC curve of the contrast-enhanced ultrasound scoring for omental lesions

CEUS greater omentum imaging and data reporting system (CEUS GOI-RADS)

Based on the aforementioned study results, a new scoring system has been established: Grade 1: CEUS-OS ≤ 8, with a malignancy rate of 0%; Grade 2: CEUS-OS = 9, with a malignancy rate ≤ 25%; Grade 3: CEUS-OS = 10, with a malignancy rate ≤ 60.0%; Grade 4: CEUS-OS = 11–13, with a malignancy rate of 80.0%–90.0%; Grade 5: CEUS-OS ≥ 14, with a malignancy rate ≥ 94.0%.

Discussion

As a central structure of the peritoneal cavity, the greater omentum can be affected by a variety of conditions, including inflammatory diseases, tuberculosis, and metastatic tumors [3]. Studies have shown that the greater omentum is often the initial site of peritoneal metastasis for cancer cells [13], with microvascular angiogenesis and tumor blood perfusion considered critical rate-determining steps in tumor growth and metastasis [14]. Due to its deep anatomical location and nonspecific symptoms, conventional imaging techniques exhibit low detection rates for small omental nodules (< 1 cm). Early diagnosis of malignant lesions (e.g., peritoneal metastatic tumors) is crucial for cancer staging and treatment decision-making. However, current clinical reliance on CT and MRI has certain limitations: CT provides insufficient hemodynamic assessment, while dynamic contrast-enhanced MRI is susceptible to respiratory motion artifacts. Neither modality offers quantitative analysis of microvascular abnormalities [15]. Consequently, some patients require repeated biopsies, increasing the risks of bleeding and tumor seeding. Therefore, the development of noninvasive, quantitative functional imaging techniques has become an urgent clinical need.

CEUS enables real-time dynamic monitoring of capillary-level perfusion through microbubble contrast agents, revealing the perfusion patterns of microcirculatory blood flow. Moreover, the contrast agents used in CEUS are non-nephrotoxic and can be safely administered to patients with renal insufficiency, thereby avoiding the risk of nephrotoxicity associated with iodinated contrast agents used in CT scans [16]. Quantitative parameters in CEUS are generated by calculating the changes in the average signal intensity within a defined ROI in the contrast images to produce a time–intensity curve. The average signal intensity within the ROI is calculated in linear units and displayed as a function of time. Thus, with the use of a contrast agent bolus, the TIC describes the wash-in and wash-out of the contrast agent within the ROI [5], and the resulting parameters quantitatively reflect the perfusion in the ROI by generating 15 objective parameters. These parameters indirectly assess the formation of neovascularization in greater omentum lesions, significantly surpassing the qualitative limitations of CT/MRI, which rely on subjective descriptions such as “mild/moderate enhancement.” Previous CEUS research has primarily focused on solid organs such as the liver and pancreas, with a long-standing lack of a standardized assessment system for the greater omentum. This study, through a systematic exploration of the application of CEUS quantitative parameters in omental lesions, has for the first time established a CEUS-GOIRADS, filling the diagnostic gap in this important anatomical region.

In our study, we conducted a systematic analysis of the quantitative parameters of CEUS in 124 patients with omental lesions and established for the first time a non-invasive predictive model based on multi-parameter fusion and a CEUS-GOIRADS. The results showed that IMAX, Fs50, and WioAUC are key independent predictors of malignant omental lesions. The combined model achieved a diagnostic sensitivity of 90.7% and specificity of 90.5%. Furthermore, the constructed CEUS-GOIRADS scoring system, with a cutoff value of 10.5 points, can effectively distinguish between benign and malignant lesions (sensitivity 94.7%, specificity 89.6%). It also quantifies the risk of malignancy through grading, providing a visual tool for clinical decision-making.

In this study, we first conducted univariate analysis on the 15 CEUS parameters obtained. The results showed that IMAX, Rs50, Rs1090, Fs50, AUC, WinAUC, WoutAUC, WioAUC, WinR, and WoutR had statistically significant differences between the malignant and benign groups (P < 0.05) and exhibited relatively high sensitivity (97.6%, 85.4%, 82.9%, 80.5%, 95.1%, 92.7%, 87.8%, and 84.0%, respectively). In contrast, TTP, RT, FHT, FT, and mTT showed no statistically significant differences between the two groups (P > 0.05). These findings revealed the specific hemodynamic changes in omental malignancies: Firstly, the significant increase in IMAX in the malignant group is directly related to the high-density perfusion of neovessels induced by pro-angiogenic factors such as VEGF in the tumor microenvironment, reflecting the hyperperfusion state of malignant tumors [17]. Secondly, the increase in Rs50 and Rs1090, combined with the decrease in Fs50, forms a “rapid inflow-slow outflow” pattern, which is consistent with the characteristic of rapid contrast agent extravasation and delayed clearance caused by the disordered vascular basement membrane structure in malignant tumors [18]. The elevated WinAUC in the malignant group intuitively reflects the highly active neovascularization within the tumor, with dense and disordered neovessels leading to a significant increase in blood flow into the tumor core per unit time, which is a direct manifestation of the high metabolic demands of malignant tumors. The increase in WoutAUC indicates that the tumor is invasive, with its secreted pro-angiogenic factors stimulating the surrounding normal omentum to form an abundant but structurally abnormal neovascular network, providing additional blood supply channels for the infiltrative growth of the tumor [19]. Additionally, the significant difference in WioAUC suggests the presence of perfusion heterogeneity within the lesion, which is closely related to regional necrosis and fibrosis foci in pathology, leading to uneven microcirculation distribution [20]. WinR reveals the structural defects of neovessels within the tumor, which often lack a complete smooth muscle layer and neural regulation, with immature and tortuous dilated vessel walls forming low-resistance shunts similar to arteriovenous fistulas, resulting in reduced blood flow resistance and increased velocity. In contrast, WoutR may exhibit relatively higher or variable values, reflecting that although the vessels generated around the tumor are more numerous, their structure may be slightly more organized compared to the tumor core, or there may be changes in resistance due to factors such as vessel compression and tissue edema [21]. Notably, the lack of intergroup differences in parameters such as TTP may be due to the superficial anatomical location of the omentum and significant interference from respiratory movements and intestinal peristalsis [22], which reduces the stability of time-related parameters. This phenomenon suggests that time-related indicators should be used cautiously in the assessment of omental lesions. Subsequently, through multivariate logistic regression analysis, this study identified three CEUS parameters independently associated with omental malignancy: IMAX (OR = 1.142, 95% CI: 1.038–1.257, P = 0.006), Fs50 (OR = 0.276, 95% CI: 0.079–0.970, P = 0.045), and WioAUC (OR = 1.0, 95% CI: 1.00–1.00, P = 0.005). IMAX reflects the abnormal hyperperfusion state mediated by VEGF (OR > 1 indicates that the risk of malignancy increases with the intensity), the negative association of Fs50 (OR < 1) confirms the reduced outflow rate due to contrast agent retention in malignant lesions, and WioAUC quantifies microcirculatory heterogeneity (such as the coexistence of low perfusion in necrotic areas and high perfusion in richly vascularized areas) (Table S2). The combined model constructed from these three parameters demonstrated excellent diagnostic performance, with an AUC of 0.955 under the ROC curve and sensitivity and specificity of 90.7% and 90.5%, respectively, for diagnosing omental malignancy.

Subsequently, this study established a grading system by calculating scores for each patient, enabling objective prediction of the likelihood of malignant omental lesions based on CEUS parameters. The malignancy rate increased with higher CEUS-OS scores, demonstrating a strict hierarchical correlation between the score and the probability of malignancy—ranging from 0% in the ≤ 8-point group to 94% in the ≥ 14-point group. For CEUS-GOIRADS grades 1–2 (score ≤ 9, malignancy rate ≤ 25%): Imaging follow-up (CT or CEUS re-examination) at 6-month intervals is recommended to avoid unnecessary biopsies. CEUS-GOIRADS grade 3 (score 10, malignancy rate 60%): Comprehensive evaluation combining tumor markers (e.g., CA125, CEA) and clinical history is required. CEUS-GOIRADS grades 4–5 (score ≥ 11, malignancy rate ≥ 80%): Immediate ultrasound-guided omental biopsy is indicated, with grade 4 (11–13 points) showing an 80%-90% malignancy rate and grade 5 (≥ 14 points) exhibiting ≥ 94% malignancy rate. The priority for pathological confirmation escalates with higher grades. In terms of diagnostic performance, using a cutoff score of 10.5, the CEUS-GOIRADS system achieved a sensitivity of 94.7% and specificity of 89.6% (AUC 0.96, 95% CI 0.92–0.99). Its efficacy not only surpassed the joint predictive model in this study (sensitivity 90.7%, specificity 90.5%, AUC 0.955) but also significantly outperformed the sensitivity of CT (78.3%) and the specificity of MRI (82.1%) in diagnosing peritoneal metastasis [23]. From a standardization perspective, CEUS-GOIRADS provides a quantifiable and unified framework for multicenter research, advancing the standardization of imaging reports for omental diseases.

The quantitative CEUS model (IMAX + Fs50 + WioAUC) and the CEUS-GOIRADS grading system constructed in this study provide innovative solutions for the diagnosis and treatment of omental lesions. Overall, the pattern of changes in a series of indicators obtained through quantitative CEUS analysis collectively constitutes the hemodynamic “fingerprint” of omental malignancies. This fingerprint characterizes a tumor with abundant internal vessels, low resistance, and high-speed blood flow perfusion, which actively induces neovascularization in surrounding tissues and presents an overall state of high vascularity [24]. These features not only aid in radiological differential diagnosis but also provide important functional information for assessing tumor invasiveness. In terms of technical reliability, the parameters of this model can be automatically generated by post-processing software, effectively avoiding operator bias and ensuring the reproducibility of the results [25]. For non-invasive diagnosis, the model achieves a sensitivity of 90.7% and a specificity of 90.5% in identifying malignant lesions, offering a safe alternative for patients with coagulopathy or ECOG ≥ 3. In terms of optimizing clinical pathways, a CEUS-OS score ≥ 11 (malignancy rate ≥ 80%) can serve as a pre-screening tool for biopsy, increasing the positive rate of puncture from 62% to 91% and reducing unnecessary procedures by 35%. Moreover, as the first dedicated CEUS reporting system for the omentum, this study lays the foundation for the establishment of the Peritoneal Imaging Reporting and Data System (PERADS) by standardizing parameter definitions (e.g., calculation of WioAUC) and implementing a five-tier stratification (malignancy rate 0%-94%). This study marks the transition of omental disease assessment from empirical morphology to standardized functional quantification. The established diagnostic paradigm of “high perfusion - slow outflow - heterogeneity” provides a new pathway for the early screening and precise intervention of peritoneal metastatic tumors.

This study has certain limitations. Currently, there is a lack of research on the quantitative parameter analysis of omental CEUS. In this study, the selection of ROIs, including their size, shape, and location, still needs to be further standardized. Among the 15 parameters, only three were included in the final model. Future work could explore machine learning algorithms to identify the optimal combination of parameters. Additionally, the sample in this study was sourced from a single center and was retrospective in nature. Future validation of the CEUS-GOIRADS system should involve larger sample sizes and prospective, multicenter studies to assess its generalizability. Future research could also investigate the correlation between quantitative parameters and histological features, integrate artificial intelligence to achieve automatic parameter extraction and scoring, and explore the predictive ability of CEUS parameters for the primary lesions of peritoneal metastatic tumors.

Conclusion

This study confirms that CEUS quantitative parameters can noninvasively differentiate between benign and malignant omental lesions. The constructed predictive model and CEUS-GOIRADS grading system demonstrate high diagnostic efficacy and clinical practicality. This system is expected to serve as a benchmark for the application of contrast-enhanced ultrasound quantitative analysis in omental diseases, promoting the transition of CEUS from morphological assessment to functional quantitative diagnosis, ultimately integrating it into the imaging reporting standards for peritoneal diseases.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

CEUS

Contrast-Enhanced Ultrasound

CEUS-GOIRADS

CEUS Greater Omentum Imaging Reporting and Data System

ORs

odds ratios

CEUS-OS

Contrast-Enhanced Ultrasound Omental Score

ROC

Receiver Operating Characteristic curve

IMAX

Maximum intensity

RT

Rise Time

TTP

Time To Peak

mTT

Mean Transit Time

Rs50

Rise Slope 50%

Rs1090

Rise Slope 10%-90%

FT

Falling Time

Fs50

Falling Slope 50%

FHT

Falling Half Time

WinAUC

Wash-in Area Under the Curve

WoutAUC

Wash-out Area Under the Curve

WioAUC

Wash-in and out Area Under Curve

WinR

Wash-in Rate

WoutR

Wash-out Rate

TIC

Time-Intensity Curve

QOF

Quality of Fit

ROIs

Regions of interest

AUC

Area Under the ROC Curve

CT

Computerized Tomography

MRI

Magnetic Resonance Imaging

VEGF

Vascular Endothelial Growth Factor

PERADS

Peritoneal Imaging Reporting and Data System

Author contributions

CY performed the final data processing and wrote the manuscript, CZG gave technical guidance, YYT conducts data collation and preliminary analysis, WXM conducts data collection and mapping, LYJ provides resources as well as technical guidance and article corrections. All authors read and approved the final manuscript.

Funding

China Ultrasound Physician Technology Rising Star Program (No. KJXX2024006).

Data availability

Data will be made available on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of the First Affiliated Hospital of China Medical University (AF-SOP-07-1.2-01). Date:2024.11.19.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.