Abstract
Background
Regional lymph node metastasis (RLNM) is a crucial prognostic factor for rectal cancer (RC). Accurate preoperative assessment of lymph node status assists clinicians in identifying high-risk patients and formulating treatment plans. Transrectal contrast-enhanced ultrasound (CEUS) can be used to evaluate the degree of microcirculation perfusion in RC, and provide information about tumor heterogeneity. The aim of this study was to develop a nomogram based on CEUS for the accurate assessment of RLNM in patients with RC.
Methods
CEUS data, routine ultrasound parameters, and clinical and pathological data of patients with RC who underwent surgery between April 2020 and December 2023 were collected. Univariate analysis was performed to identify relevant RLNM factors, which were included in binary logistic regression to determine independent risk factors and generate a nomogram. To evaluate the model, receiver operating characteristic curve (ROC), calibration curve, and decision curve analysis (DCA) were plotted.
Results
A total of 195 patients were included in this study, with 138 and 57 patients in the training and validation groups, respectively. The independent risk factors for RLNM in RC identified through the analysis were Contrast-enhanced ultrasound inhomogeneity grade (CEUS-IG), endoscopic tumor morphology, and carbohydrate antigen 199 (CA199) level. The nomogram constructed based on these variables demonstrated good discrimination, with area under the curve (AUC) values of 0.832 in the training group and 0.792 in the validation group. The calibration curve indicated good consistency between the predicted values and the actual observations, and DCA demonstrated good clinical utility of the model.
Conclusions
This study combined CEUS-IG with endoscopic tumor morphology and CA199 levels to develop a predictive model for RLNM in rectal cancer, providing valuable predictive information for preoperative evaluation. This model can enhance diagnostic accuracy.
Keywords: Regional lymph node metastasis (RLNM), Contrast-enhanced ultrasound inhomogeneity grade (CEUS-IG), Contrast-enhanced ultrasound (CEUS), Rectal cancer (RC)
Introduction
Colorectal cancer (CRC), as one of the most common malignant tumors, ranks third in global cancer incidence and second in cancer mortality [1]. Rectal cancer (RC) accounts for approximately 40% of all CRC cases. The early methods of screening for RC include stool-based tests, blood tests, radiological examinations, and endoscopy; however, the effectiveness is still not ideal [2], with approximately one-quarter of RC patients already presenting with metastasis at the time of diagnosis [3]. Regional lymph node metastasis (RLNM) is the primary risk factor for local recurrence and distant metastasis after surgery in RC. Underestimating the N stage and omitting preoperative neoadjuvant chemoradiotherapy increase the risk of recurrence, while overestimating it may lead to overtreatment and associated complications [4], particularly in T1–T2 stages. Therefore, accurately assessing lymph node metastasis is crucial for selecting the appropriate preoperative surgical modality and determining the necessity of neoadjuvant therapy [5].
In patients with rectal cancer, if metastatic lymph nodes are present in the lymphatic drainage area, they are staged as N (1–2); if tumor deposits are present in the lymphatic drainage area, they are staged as N1c [6]. In this study, both were classified as RLNM-positive. This study focused on regional lymph nodes, which primarily include areas such as the mesorectum, internal iliac vessels, and obturator regions [7]. The National Comprehensive Cancer Network guidelines recommend commonly used imaging methods for diagnosing RLNM in RC, including magnetic resonance imaging (MRI), endorectal ultrasound (ERUS), and computed tomography (CT) [8]. However, the current conventional imaging methods have limited value in assessing RLNM. The diagnostic efficacies of MRI, ERUS, and CT are low. Comparatively, MRI has higher sensitivity (0.69), whereas transrectal ultrasound has higher specificity (0.80) [9]. Meta-analyses have indicated that even when ERUS, CT, and MRI are used in combination, their diagnostic efficacy for lymph node staging remains suboptimal [10]. Consequently, developing a novel method for precise evaluation of lymph node condition is crucial.
In the images from the aforementioned imaging methods, lymph nodes displayed are difficult to correspond one-to-one with the locations of the lymph nodes in the gross pathological specimens post-surgery; therefore, it is challenging to verify through pathology whether a specific lymph node indicated in the images has metastasis. Some studies have found potential correlations between the primary tumor characteristics of RC and lymph node status [11]. Therefore, this study aims to predict RLNM based on the ultrasound image characteristics of primary RC lesions. Studies on predicting lymph node status using primary lesion images from MR and CT scans have already been conducted, albeit with varying results [12, 13]. The contrast agents used in these imaging techniques differ; CEUS uses microbubble contrast agents, CT relies on iodine, and MRI uses gadolinium-based contrast agents, each with different pharmacokinetics. Ultrasound contrast agents are confined to the vascular system as pure blood pool contrast agents, whereas most approved CT and MRI contrast agents are rapidly cleared from the blood pool into the extracellular space. Despite some similarities in the enhancement modes among these methods, the enhancement mode of specific lesions in ultrasound imaging may differ from that observed on CT or MRI [14]. A recent study reported that in contrast-enhanced ultrasound of rectal cancer lesions, calculating differences using ultrasound contrast parameters showed a certain correlation with lymph node metastasis [15]. We still need to further validate this correlation and explore whether there are more direct and straightforward methods.
A nomogram is a predictive model that transforms complex data into a graphical format, providing a more intuitive explanation of the impact of each independent risk factor on the outcomes. This allows clinicians to accurately quantify and predict the probability of individual risk events [16]. Therefore, the aim of this study is to develop a nomogram based on CEUS for the accurate assessment of RLNM in patients with RC.
Methods
Population
A total of 195 patients with RC were included (Fig. 1). These patients underwent surgery at our hospital between April 2020 and December 2023 and were pathologically confirmed to have RC postoperatively. The inclusion criteria were: patients who had contrast-enhanced transrectal ultrasonography at our hospital within a month prior to surgery. The exclusion criteria were: (1) patients who underwent chemotherapy or radiotherapy prior to surgery, (2) The tumor volume being excessively large and located too high in the upper rectum, resulting in the inability to display a complete image of the tumor through ERUS. (3) patients with a history of other malignant tumors or recurrent RC, (4) patients who did not undergo lymph node dissection or had fewer than 12 lymph nodes dissected, and (5) patients whose clinical data were incomplete.
Fig. 1.
Flowchart of participant enrollment process
Instrument selection
The instruments and probes used in this study were as follows: (1) BK 5000 Color Doppler Ultrasound Diagnostic System (BK Medical, Herlev, Denmark), E14CL4b biplanar transrectal ultrasound probe; (2) MyLab Twice Ultrasound Diagnostic System (Esaote, Genoa, Italy), TRT33 biplanar transrectal ultrasound probe.
Ultrasound acquisition
To obtain clear ultrasound images, it is necessary to instruct the patient to empty their bowels before the examination. During the examination, the patients were positioned on their left side, took deep breaths to relax the anal canal, and an ultrasound probe was slowly inserted through the anus. A comprehensive scan of the rectal and anal canal regions was performed, and detailed parameters of the lesions were recorded. Subsequently, an ultrasound contrast examination was conducted by selecting the contrast mode and administering 2.4 mL of the prepared sulfur hexafluoride contrast agent through the median cubital vein. The lesion was observed for approximately one minute under the longitudinal section of the tumor, the patients’ contrast-enhanced ultrasound video was stored, and the image of the frame when the contrast-enhanced ultrasound reached its peak was selected for CEUS inhomogeneity grading (CEUS-IG), (the examiners were blinded to all clinical and pathological information). All aforementioned examination steps and image acquisition were conducted by two ultrasound specialists with more than 7 years of experience.
Data collection
The clinical data collected included sex; age; carcinoembryonic antigen (CEA); carbohydrate antigen 199 (CA199); carbohydrate antigen 125 (CA125); and endoscopic appearance of the tumor (ulcerative, elevated, others). Ultrasound metrics comprised the distance from the tumor’s lower margin to the white line (also known as the intersphincteric groove), tumor length, thickness, circumferential ratio of tumor invasion to the rectum, peak systolic velocity (PSV) of the tumor artery, resistance index (RI) of the tumor artery, ultrasound T stage, ultrasound N stage, depth of tumor invasion into the muscularis propria, and CEUS-IG. CEUS-IG involves selecting an image of the largest longitudinal section of the tumor at peak enhancement and semiquantitatively grading the level of enhancement homogeneity. Grade I indicates homogeneous enhancement, with no significant differences in the enhancement degree and distribution; Grade II indicates moderately homogeneous enhancement, with some areas of low enhancement or slightly unenhanced regions; and Grade III indicates heterogeneous enhancement with substantial unenhanced areas [17] (Fig. 1). RLNM status was determined using postoperative pathology as the gold standard. RLNM-negative indicates no tumor metastasis in lymph nodes, whereas RLNM-positive indicates at least one RLNM or the presence of any tumor nodules [6].
Fig. 2.
Grade of inhomogeneity in contrast-enhanced ultrasonography. Figures a, b and c on the left: Grayscale ultrasound images of 3 rectal cancer patients. A: CEUS-IG Grade (I) B: CEUS-IG Grade (II) C: CEUS-IG Grade III
Statistical analysis
The IBM SPSS Statistics software (version 26.0, Armonk, NY, USA) was used for statistical analysis.Comparisons of normally distributed data between groups were performed using the independent sample t-test. For non-normally distributed data, the Mann–Whitney U test was employed. Categorical data comparisons between groups were performed using the chi-square test or Fisher’s exact test. Univariate factors related to RLNM were selected and included in binary logistic regression for multivariate analysis. The Kappa consistency test was used to evaluate intra-observer and inter-observer agreement for CEUS-IG. A P value < 0.05 was considered statistically significant. The development of a predictive nomogram was performed using the rms package in R version 4.3.2 (The R Foundation of Statistical Computing, Vienna, Austria) [16]. To assess the predictive ability of the nomogram, we utilized the pROC package to plot the receiver operating characteristic (ROC) curves.The nomogram can convert each regression variable into a score ranging from 0 to 100 based on regression coefficients, then sum these converted numbers to obtain a total score, and finally find the corresponding predicted probability value in the nomogram.To assess the consistency between the model’s forecasted and actual probabilities, calibration graphs were created. To evaluate the net clinical benefit of the nomogram, we utilized the rmda package to construct a decision curve analysis (DCA).
Results
Analyze the intra-observer and inter-observer consistency of CEUS-IG
To evaluate the intra-observer and inter-observer consistency of CEUS-IG, images from 100 randomly selected patients were analyzed.CEUS image observers were blinded to postoperative pathological results. Intra-observer consistency was determined by having the same observer (Observer 1) reassess the images after three months later. The intra-observer agreement was excellent (κ = 0.86, P < 0.05), as shown in Table 1. Inter-observer consistency was evaluated by comparing the grading of the same images by a second observer (Observer 2). The inter-observer agreement was good ( κ = 0.704, P < 0.05) , as shown in Table 2.
Table 1.
Intra-observer consistency of CEUS-IG
| First time | Second time | Total | Kappa | P | ||
|---|---|---|---|---|---|---|
| Grade Ⅰ | Grade Ⅱ | Grade Ⅲ | ||||
| Grade I | 29 | 3 | 0 | 32 | 0.860 | <0.001 |
| Grade II | 1 | 41 | 3 | 45 | ||
| Grade III | 0 | 2 | 21 | 23 | ||
| Total | 30 | 46 | 24 | 100 | ||
Table 2.
Inter-observer consistency of CEUS-IG
| Observer 1 | Observer 2 | Total | Kappa | P | ||
|---|---|---|---|---|---|---|
| Grade I | Grade II | Grade III | ||||
| Grade Ⅰ | 28 | 4 | 0 | 32 | 0.704 | <0.001 |
| Grade Ⅱ | 6 | 35 | 4 | 45 | ||
| Grade Ⅲ | 0 | 5 | 18 | 23 | ||
| Total | 34 | 44 | 22 | 100 | ||
Independent risk factors
A total of 195 patients were included, and randomly divided into training and validation groups at a ratio of 7:3 using IBM SPSS Statistics version 26.0 software. There were no significant differences in clinical characteristics between the training and validation groups (P > 0.05, Table 3). Based on the pathological results, the training group data were categorized into RLNM-positive and RLNM-negative groups. Univariate analysis was performed to identify potential predictors of RLNM. Variables with a P-value < 0.05 in the univariate analysis were deemed statistically significant and associated with RLNM. Univariate analysis showed that variables such as the circumferential ratio of tumor invasion to the rectum, ultrasound T stage, ultrasound N stage, morphological type, CEA and CA199 levels, and CEUS-IG were associated with RLNM (P < 0.05). Variables such as age, sex, PSV, RI, CA125 level, tumor length, tumor thickness, and distance to the white line were not associated (P > 0.05, Table 4).
Table 3.
Characteristics of patients
| Variable | Training group (n = 138) | Validation group (n = 57) | P |
|---|---|---|---|
| Age(years), mean ± SD | 64.81 ± 13.02 | 64.37 ± 12.82 | 0.828 |
| Length(cm), mean ± SD | 3.79 ± 1.42 | 3.35 ± 1.36 | 0.050 |
| Distance(cm), mean ± SD | 6.03 ± 2.96 | 6.66 ± 3.23 | 0.204 |
| RI, mean ± SD | 0.74 ± 0.12 | 0.73 ± 0.11 | 0.447 |
| Thickness(cm), median(IQR) | 1.40 (0.70) | 1.20 (0.80) | 0.447 |
| PSV(cm/s), median (IQR) | 13.95 (10.0) | 13.40 (7.78) | 0.784 |
| CEA(ng/mL), median (IQR) | 2.74 (3.19) | 2.36 (3.16) | 0.922 |
| CA199(ng/mL), median(IQR) | 10.50 (13.94) | 10.50 (10.37) | 0.691 |
| CA125(ng/mL), median(IQR) | 9.32 (5.17) | 10.90 (6.59) | 0.086 |
| Sex, n (%) | 0.123 | ||
| Female | 70 (50.7) | 22 (38.6) | |
| Male | 68 (49.3) | 35 (61.4) | |
| Circumferential ratio, n (%) | 0.151 | ||
| < 1/4 | 17 (12.3) | 11 (19.3) | |
| 1/4–1/2 | 70 (50.7) | 31 (54.4) | |
| 1/2–3/4 | 33 (23.9) | 6 (10.5) | |
| 3/4–1 | 18 (13.0) | 9 (15.8) | |
| Ultrasound stage T, n (%) | 0.742 | ||
| uT1 | 12 (8.8) | 6 (10.5) | |
| uT2 | 37 (27.0) | 14 (24.6) | |
| uT3 | 51 (37.2) | 25 (43.9) | |
| uT4 | 37 (27.0) | 12 (21.1) | |
| Ultrasound stage N, n (%) | 0.362 | ||
| uN– | 70 (50.7) | 33 (57.8) | |
| uN+ | 68 (49.3) | 24 (42.2) | |
| Morphological type, n (%) | 0.549 | ||
| Ulcerative | 90 (65.2) | 34 (59.6) | |
| Elevated | 46 (33.3) | 23 (40.4) | |
| Others | 2 (1.4) | 0 (0) | |
| CEUS-IG, n (%) | 0.125 | ||
| Grade I | 39 (28.3) | 21 (36.8) | |
| Grade II | 68 (49.3) | 19 (33.3) | |
| Grade III | 31 (22.5) | 17 (29.8) | |
| Postoperative pathological lymph node metastasis, n (%) | 0.976 | ||
| pN+ | 82 (59.4) | 34 (59.6) | |
| pN– | 56 (40.6) | 23 (40.4) | |
Table 4.
Comparisons of clinical characteristics between lymph node metastasis positive and lymph node metastasis negative group
| Variable | Training group (n = 138) | P | |
|---|---|---|---|
| RLNM (–) (n = 82) | RLNM (+) (n = 56) | ||
| Age(years), mean ± SD | 64.2 ± 12.6 | 65.7 ± 13.6 | 0.495 |
| Length(cm), mean ± SD | 3.62 ± 1.53 | 4.03 ± 1.24 | 0.097 |
| Distance(cm), mean ± SD | 5.6 ± 2.9 | 6.6 ± 2.9 | 0.051 |
| RI, mean ± SD | 0.75 ± 0.11 | 0.73 ± 0.13 | 0.297 |
| Thickness(cm), median(IQR) | 1.4 (0.8) | 1.5 (0.5) | 0.341 |
| PSV(cm/s), median (IQR) | 14.0 (9.4) | 13.5 (11.0) | 0.718 |
| CEA(ng/mL), median(IQR) | 2.3 (2.2) | 3.6 (6.8) | 0.001 |
| CA199(ng/mL), median(IQR) | 8.9 (11.0) | 17.5 (23.0) | < 0.001 |
| CA125(ng/mL), median(IQR) | 9.3 (5.2) | 9.3 (5.2) | 0.907 |
| Sex, n (%) | 0.626 | ||
| Female | 43 (52.4%) | 27 (48.2%) | |
| Male | 39 (47.6%) | 29 (51.8%) | |
| Circumferential ratio, n (%) | 0.006 | ||
| < 1/4 | 14 (17.1%) | 3 (5.4%) | |
| 1/4–1/2 | 47 (57.3%) | 23 (41.1%) | |
| 1/2–3/4 | 13 (15.9%) | 20 (35.7%) | |
| 3/4–1 | 8 (9.8%) | 10 (17.9%) | |
| Ultrasound stage T, n (%) | 0.014 | ||
| uT1 | 11 (13.6%) | 1 (1.8%) | |
| uT2 | 26 (32.1%) | 11 (19.6%) | |
| uT3 | 27 (33.3%) | 24 (42.9%) | |
| uT4 | 17 (21.0%) | 20 (35.7%) | |
| Ultrasound stage N, n (%) | 0.026 | ||
| uN– | 48 (58.5%) | 22 (39.3%) | |
| uN+ | 34 (41.5%) | 34 (60.7%) | |
| Morphological type, n (%) | < 0.001 | ||
| Ulcerative | 43 (52.4%) | 47 (83.9%) | |
| Elevated | 38 (46.3%) | 8 (14.3%) | |
| Others | 1 (1.2%) | 1 (1.8%) | |
| CEUS-IG, n (%) | < 0.001 | ||
| Grade I | 32 (39.0%) | 7 (12.5%) | |
| Grade II | 42 (51.2%) | 26 (46.4%) | |
| Grade III | 8 (9.8%) | 23 (41.1%) | |
SD: standard deviation; IQR: interquartile range; RI: resistance index (blood flow resistance index of the tumor); PSV: Peak systolic velocity, peak systolic blood flow in the tumor arteries; CEA: carcinoembryonic antigen; CA199: carbohydrate antigen 199; CA125: carbohydrate antigen 125; Circumferential ratio: circumferential ratio of tumor invasion to the rectum; CEUS-IG: contrast-enhanced ultrasound inhomogeneity grade
SD: standard deviation; IQR: interquartile range; RI: resistance index (blood flow resistance index of the tumor); PSV: Peak systolic velocity, peak systolic blood flow in the tumor arteries; CEA: carcinoembryonic antigen; CA199: carbohydrate antigen 199; CA125: carbohydrate antigen 125; Circumferential ratio: circumferential ratio of tumor invasion to the rectum; CEUS-IG: contrast-enhanced ultrasound inhomogeneity grade.
To further investigate their predictive significance, the seven variables related to RLNM mentioned above were included in a binary logistic regression analysis (Table 5), after analysis, variables with P < 0.05 were deemed statistically significant and considered independent risk factors, which were subsequently used to construct the nomogram.
Table 5.
Binary logistic regression results
| Variable | OR | 95% CI | P |
|---|---|---|---|
| CEA | 1.043 | 0.963–1.131 | 0.302 |
| CA199 | 1.053 | 1.009–1.099 | 0.018 |
| Circumferential ratio | |||
| < 1/4 | Reference | ||
| 1/4–1/2 | 0.757 | 0.145–3.953 | 0.741 |
| 1/2–3/4 | 1.768 | 0.259–12.063 | 0.561 |
| 3/4–1 | 1.060 | 0.131–8.580 | 0.956 |
| CEUS-IG | |||
| Grade I | Reference | ||
| Grade II | 3.884 | 1.232–12.246 | 0.021 |
| Grade III | 18.711 | 4.462–78.460 | < 0.001 |
| Ultrasound stage T | |||
| uT1 | Reference | ||
| uT2 | 2.825 | 0.250–31.910 | 0.401 |
| uT3 | 2.438 | 0.208–28.646 | 0.478 |
| uT4 | 3.413 | 0.276–42.170 | 0.339 |
| Ultrasound stage N | |||
| uN– | Reference | ||
| uN+ | 1.187 | 0.459–3.074 | 0.724 |
| Morphological type | |||
| Ulcerative | Reference | ||
| Elevated | 0.277 | 0.095–0.811 | 0.019 |
| Others | 0.199 | 0.009–4.270 | 0.302 |
CEA: carcinoembryonic antigen; CA199: carbohydrate antigen 199; Circumferential ratio: circumferential ratio of tumor invasion to the rectum; CEUS-IG: contrast-enhanced ultrasound inhomogeneity grade
The results showed the performance of tumor CEUS as follows: compared with Grade I (reference), Grade II had a higher risk of RLNM, with the difference was statistically significant ( P = 0.021). Similarly, compared with Grade I, the difference for Grade III was also statistically significant (P < 0.001). Homogeneous enhancement (Grade I) at the peak of CEUS could serve as a protective factor against RLNM. Additionally, when comparing Grades II and III, using Grade III as the reference, the risk of metastasis was lower for Grade II (P = 0.006) than for Grade III. Grade III has a particularly strong predictive ability for RLNM (OR = 18.711). Increased CA199 levels were correlated with a greater likelihood of RLNM (P = 0.018). Tumors that appeared elevated on endoscopy had a significantly lower risk of RLNM compared to ulcerative lesions (P = 0.018). In contrast, the circumferential ratio of tumor invasion to the rectum, ultrasound T stage, ultrasound N stage, and CEA level were not statistically significant (P > 0.05). In summary, CEUS-IG, morphological type, and CA199 level were independent risk factors for RLNM .
Construction and evaluation of the nomogram
We incorporated three independent risk factors and used R version 4.3.2 to plot the nomogram (Fig. 3). The ROC curve (Fig. 4) showed that the AUC was 0.832 (95% CI, 0.766–0.898) for the training group and 0.792 (95% CI, 0.673–0.910) for the validation group, indicating that the nomogram had good predictive capability. The calibration curve (Fig. 5) showed good agreement between predicted and actual observed values in both the training and validation groups. Additionally, a DCA curve (Fig. 6) demonstrated that using this predictive model to assess RLNM provides a net clinical benefit across a wide range of thresholds.
Fig. 3.
Nomogram to predict the risk of lymph node metastasis in patients with rectal cancer. CEUS-IG: contrast-enhanced ultrasound inhomogeneity grade; Morphological type: the type of tumor under the endoscope
Fig. 4.
Receiver operating characteristic (ROC) curve of the nomogram for predicting lymph node metastasis. A: ROC curve for the training set; B: ROC curve for the validation set. AUC: area under the curve
Fig. 5.
Calibration curve for the nomogram in the cohort. A: Calibration curve of the training cohort, B: Calibration curve of the validation cohort
Fig. 6.
Decision curve analysis of the nomogram. The Y-axis represents the net income. The pink curve indicates the training set, whereas the purple curve indicates the validation set. DCA: decision curve analysis
Discussion
CRC is one of the most common malignant tumors worldwide [1]. RLNM is one of the most important features of CRC staging, indicating that clinical treatment requires adjuvant chemoradiotherapy, in addition to local resection and postoperative lymph node dissection [5]. RLNM is associated with poor overall survival rates [18]; therefore, accurate preoperative assessment of RLNM is crucial. However, in clinical practice, it is difficult to match each lymph node identified on ERUS images with postoperative pathological lymph nodes when evaluating N staging. This study aimed to predict RLNM based on the characteristics of primary RC lesions. This research involved patients diagnosed with RC who underwent total mesorectal excision and lymph node dissection, and comprehensively analyzed the risk factors for RLNM. We found that patients with RC and positive RLNM had higher CEUS-IG grades, were more likely to have ulcerative lesions on endoscopy, and had higher CA199 levels.
In this study, we graded the heterogeneity of RC lesions at the peak of ultrasound contrast enhancement to objectively evaluate the degree of microcirculatory perfusion uniformity within the lesions. This grading method is a static semiquantitative diagnostic approach, whereas previous ultrasound contrast studies often focused on the dynamic observation of tumor vascular heterogeneity through time–intensity curves of a specific region of interest (ROI) [19], only a few studies focus on the uneven degree of tumor enhancement at peak time during contrast-enhanced ultrasound [20]. However, for larger lesions, issues such as selecting an ROI that is significantly small to represent the entire lesion may arise, leading to an inadequate overall lesion assessment. Lesions were graded as Grade I if they showed uniform enhancement, indicating uniform microcirculatory perfusion within the tumor. Grade II was assigned to lesions with overall uneven enhancement, indicating heterogeneity and possible necrosis [21]. Grade III lesions showed significant uneven enhancement and large areas without enhancement, corresponding to necrotic regions without microcirculatory perfusion.
Our findings showed that the risk of RLNM increased sequentially across CEUS-IG grades I, II, and III, whereas the uniformity and intensity of tumor enhancement decreased. Based on this, we concluded that in CEUS-IG, the grading level positively correlated with the risk of RLNM. Uniform enhancement, particularly highly uniform enhancement, can be a protective factor against RLNM, indicating a lower risk; whereas uneven enhancement with non-enhancing areas correlates with a higher risk. A similar negative correlation was observed in dynamic contrast-enhanced-MRI studies, where advanced N-stage primary tumors showed weaker enhancement intensity [22, 23].
Malignant tumor neovasculature is characterized by disorganized, abundant blood vessels, resulting in significant arteriovenous shunts and high enhancement on CEUS [24]. The outer regions of the tumor are densely vascularized, whereas the central regions exhibit perfusion deficiencies and necrosis owing to abnormal vascular distribution [25, 26]. When necrosis occurs within the tumor, the hypoxic environment can induce the production of various cell growth factors such as vascular endothelial growth factor (VEGF) [27–29]. These factors can promote the growth of new lymphatic vessels towards the tumor margin, while tumor cells also move towards the capillary lymphatic vessels, subsequently infiltrate the lymphatic system, and proliferating or metastasizing to other lymph nodes [30]. This explains the observed negative correlation.
The morphological appearance of tumors is mainly classified as polypoid elevation or ulcerative infiltration [31]. Tumor morphology is related to tumor characteristics and serves as an independent prognostic predictor [32]. Early-stage RC (with only submucosal invasion) is predominantly polypoid and well-differentiated and has a nearly 100% 5-year survival rate with minimal RLNM [33]. Ulcerative RC has a worse prognosis, increased opportunities for lymphatic invasion, and higher lymph node involvement [34]. When endoscopic examination reveals an ulcerative tumor lesion, it indicates that the tumor has undergone localized necrosis and has a higher risk of RLNM than do elevated lesions. Elevated RC lesions exhibit outward expansive growth with slow infiltration into the intestinal wall, whereas ulcerative RC can infiltrate the intestinal wall once tumor growth causes ulceration. Ulcerative RC is associated with an earlier and higher risk of RLNM than elevated RC. In this study, two cases involved lesions identified during endoscopic examination that were neither elevated nor ulcerative.
CEA is a widely used tumor marker for CRC [35] and is recommended as a postoperative prognostic biomarker [36, 37]. Patients with elevated preoperative serum CEA levels are at greater risk of RLNM [38–40]. However, in our study, CEA levels did not provide additional value in the multivariate analysis for predicting RLNM, which may be related to the study population’s characteristics, and does not imply that CEA levels are unimportant. Therefore, these findings should be interpreted with caution.
CA199 is an oligosaccharide tumor-associated antigen first identified by Koprowski et al. in 1979 [41]. In addition to CEA, CA199 can serve as a tumor marker and as an alternative marker in patients with normal CEA levels [42]. Serum CA199 levels differ significantly across various TNM stages of CRC, with markedly higher levels in patients with RLNM [38]. Elevated preoperative serum CA199 levels are associated with a greater likelihood of poorly differentiated tumors and an increased risk of RLNM [40]. Baseline CA199 levels are negatively correlated with CRC prognosis and are independent risk factors for metastasis [43]. Moreover, elevated CA199 levels are independent risk factors for poor disease-free survival and overall survival in patients with CRC [44].
This study assessed the risk factors for RLNM in rectal cancer patients and included CEUS-IG, endoscopic tumor morphology, and CA199 as three variables to construct a clinical nomogram, which effectively predicts lymph node metastasis. The AUC for the training set is 0.832 (95% CI: 0.766–0.898), and for the validation set, it is 0.792 (95% CI: 0.673–0.910) (Fig. 4). Given that these three variables are easily obtainable from preoperative examinations in rectal cancer patients, the use of this nomogram will aid in the preoperative assessment of regional lymph node metastasis, providing a reference for clinical diagnosis and treatment.However, this study did not include patients who had undergone chemoradiotherapy. If rectal cancer patients have lymph node-positive metastasis, the status of their lymph nodes might change due to chemoradiotherapy, introducing confounding factors that could affect the stability of the model. Therefore, the nomogram from this study is not applicable for assessing rectal cancer patients who have received chemoradiotherapy.
Currently, MRI remains the preferred imaging modality for diagnosing rectal cancer due to its superior soft-tissue contrast and ability to assess the depth of tumor invasion and lymph node involvement.On the other hand, CEUS has unique value in clinical practice, such as lower cost, bedside capability, especially in resource-limited settings. This study found that CEUS can capture biological information about microcirculatory perfusion within rectal cancer lesions, providing functional imaging information that can complement MRI diagnosis.In the future, combining the two examination methods could enhance diagnostic accuracy and provide valuable prognostic information for rectal cancer patients.
This study has some limitations. First, the contrast images we analyzed were static, focusing solely on the frame at peak enhancement, and we did not analyze the time-series characteristics of the CEUS images. In the future, we will further explore dynamic infusion data (time-intensity curves) to assess predictive accuracy, and we will also attempt ultrasound imaging radiomics analysis to further enhance the value of the research.Second, the ultrasound section we selected was the largest longitudinal or cross-section of the tumor, which is not representative of the overall tissue perfusion of the tumor. Third, this study was retrospective, and the results may have been biased, future prospective studies or external validation are needed.Fourth, CEUS is an operator-dependent technique, its diagnostic performance may vary substantially across institutions and personnel, standardized procedures are needed to improve reproducibility and generalizability.
Conclusions
This study combined CEUS-IG with endoscopic tumor morphology and CA199 levels to develop a predictive model for RLNM in rectal cancer, providing valuable information for preoperative evaluation. This model can enhance diagnostic accuracy.
Author contributions
Author contribution: Conception and design: Qin Huang, Zhongfan Liao, Hua Zhuang; Administrative support: Hua Zhuang; Provision of study materials or patients: Qin Huang, Zhongfan Liao, Hua Zhuang; Collection and assembly of data: Qin Huang, Yin Yang, Dilimire Abuliezi, Anqi Tao; Data analysis and interpretation: Qin Huang, Yuan Luo, Zhongfan Liao; Manuscript writing: All authors; Final approval of manuscript: All authors.
Funding
This study was supported by The Science and Technology Planning Project of Sichuan Province, China (grant number 2021YJ0243).
Data availability
The data generated during the current study are available from the corresponding author upon reasonable request. And all the data supporting the findings of this study are included in this published article and its supplementary information files.
Declarations
Ethical approval and consent to particpate
The authors are accountable for all aspects of the work and assure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This retrospective study was approved by the Medical Ethics Committee of West China Hospital, Sichuan University (NO.2024 − 191) and adhered to the tenets of the Declaration of Helsinki (as revised in 2013). The institutional review board waived the requirement for informed consent owing to the retrospective nature of the study.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Bray F, Laversanne M, Sung H, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries.CA-CANCER. J CLIN. 2024;74(3):229–63. [DOI] [PubMed] [Google Scholar]
- 2.Tonini V, Zanni M. Why is early detection of colon cancer still not possible in 2023? WORLD J GASTROENTERO. 2024;30(3):211–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Van Cutsem E, Nordlinger B, Adam R, et al. Towards a pan-European consensus on the treatment of patients with colorectal liver metastases. EUR J CANCER. 2006;42(14):2212–21. [DOI] [PubMed] [Google Scholar]
- 4.Kapiteijn E, Marijnen CA, Nagtegaal ID, et al. Preoperative radiotherapy combined with total mesorectal excision for resectable rectal cancer. NEW ENGL J MED. 2001;345(9):638–46. [DOI] [PubMed] [Google Scholar]
- 5.Nagtegaal ID, Schmoll HJ. Colorectal cancer: what is the role of lymph node metastases in the progression of colorectal cancer? NAT REV GASTRO HEPAT. 2017;14(11):633–4. [DOI] [PubMed] [Google Scholar]
- 6.Weiser MR. AJCC 8th edition: colorectal cancer. ANN SURG ONCOL. 2018;25(6):1454–5. [DOI] [PubMed] [Google Scholar]
- 7.Chinese protocol of diagnosis and treatment of colorectal cancer of the National Health Commission. (2023 edition). Zhonghua Wei Chang Wai Ke Za Zhi. 2023; 26 (6): 505–528. [DOI] [PubMed]
- 8.National Comprehensive Cancer Network. Rectal cancer (version 2.2025). https://www.nccn.org/guidelines/guidelines-detail?category=1&id=1461
- 9.Gao Y, Li J, Ma X, et al. The value of four imaging modalities in diagnosing lymph node involvement in rectal cancer: an overview and adjusted indirect comparison. CLIN EXP MED. 2019;19(2):225–34. [DOI] [PubMed] [Google Scholar]
- 10.Glynne-Jones R, Wyrwicz L, Tiret E, et al. Rectal cancer: ESMO clinical practice guidelines for diagnosis, treatment and follow-up. ANN ONCOL. 2017;28(suppl4):iv22–40. [DOI] [PubMed] [Google Scholar]
- 11.Yang L, Liu D, Fang X, et al. Rectal cancer: can T2WI histogram of the primary tumor help predict the existence of lymph node metastasis? EUR RADIOL. 2019;29(12):6469–76. [DOI] [PubMed] [Google Scholar]
- 12.Yang H, Jiang P, Dong L, et al. Diagnostic value of a radiomics model based on CT and MRI for prediction of lateral lymph node metastasis of rectal cancer. UPDATES SURG. 2023;75(8):2225–34. [DOI] [PubMed] [Google Scholar]
- 13.Xu H, Zhao W, Guo W, et al. Prediction model combining clinical and MR data for diagnosis of lymph node metastasis in patients with rectal cancer. J MAGN RESON IMAGING. 2020;53(3):874–83. [DOI] [PubMed] [Google Scholar]
- 14.Lencioni R, Piscaglia F, Bolondi L. Contrast-enhanced ultrasound in the diagnosis of hepatocellular carcinoma. J HEPATOL. 2008;48(5):848–57. [DOI] [PubMed] [Google Scholar]
- 15.Su S, Huang X, Li X, et al. Predictive value of TRUS and CEUS parameters for lymph node metastasis in rectal cancer: A retrospective study. Int J Gen Med. 2025;18:3335–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Park SY. Nomogram: an analogue tool to deliver digital knowledge. J THORAC CARDIOV SUR. 2018;155(4):1793. [DOI] [PubMed] [Google Scholar]
- 17.Luo Y, Qin LK, Yan JW, et al. Classification of contrast-enhanced ultrasonograms in rectal cancer according to tumor inhomogeneity using machine learning-based texture analysis. TRANSL CANCER RES. 2022;11(5):1053–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kogay M, Evesque L, Falcoz A, et al. Sidedness in metastatic colorectal carcinoma: which are the factors which influence the prognosis? J CLIN ONCOL. 2019;37:497. [Google Scholar]
- 19.Averkiou MA, Bruce MF, Powers JE, et al. Imaging methods for ultrasound contrast agents. ULTRASOUND MED BIOL. 2019;46(3):498–517. [DOI] [PubMed] [Google Scholar]
- 20.Abuliezi D, She YF, Liao ZF, et al. Combined transrectal ultrasound and radiomics model for evaluating the therapeutic effects of neoadjuvant chemoradiotherapy in locally advanced rectal cancer. INT J COLORECTAL DIS. 2025;40(1):7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Yao WW, Zhang H, Ding B, et al. Rectal cancer: 3D dynamic contrast-enhanced MRI; correlation with microvascular density and clinicopathological features. RADIOL MED. 2011;116(3):366–74. [DOI] [PubMed] [Google Scholar]
- 22.Hong HS, Kim SH, Park HJ, et al. Correlations of dynamic contrast-enhanced magnetic resonance imaging with morphologic, angiogenic, and molecular prognostic factors in rectal cancer. YONSEI MED J. 2013;54(1):123–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Grøvik E, Redalen KR, Storås TH, et al. Dynamic multi-echo DCE- and DSC-MRI in rectal cancer: low primary tumor Ktrans and ∆R2* peak are significantly associated with lymph node metastasis. J MAGN RESON IMAGING. 2016;46(1):194–206. [DOI] [PubMed] [Google Scholar]
- 24.Zhuang Z, Fan G, Yuan Y, et al. A fuzzy clustering based color-coded diagram for effective illustration of blood perfusion parameters in contrast-enhanced ultrasound videos. COMPUT METH PROG BIO. 2019;190:105233. [DOI] [PubMed] [Google Scholar]
- 25.Kluza E, Kleijnen JP, Martens MH, et al. Non-invasive MR assessment of macroscopic and microscopic vascular abnormalities in the rectal tumour-surrounding mesorectum. EUR RADIOL. 2015;26(5):1311–9. [DOI] [PubMed] [Google Scholar]
- 26.Mao Y, Mu J, Zhao J, et al. The comparative study of color doppler flow imaging, Superb microvascular imaging, contrast-enhanced ultrasound micro flow imaging in blood flow analysis of solid renal mass. Cancer Imaging. 2022;22(1):21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Catalano V, Turdo A, Di Franco S, et al. Tumor and its microenvironment: a synergistic interplay. SEMIN CANCER BIOL. 2013;23:522–32. [DOI] [PubMed] [Google Scholar]
- 28.Kong D, Zhou H, Neelakantan D et al. VEGF-C mediates tumor growth and metastasis through promoting EMT-epithelial breast cancer cell crosstalk. ONCOGENE.2020;40(5):964–979. [DOI] [PMC free article] [PubMed]
- 29.Rak J. VEGF-D(ilated) lymphatics as gateways to metastasis. Cancer Cell. 2012;21(2):139–40. [DOI] [PubMed] [Google Scholar]
- 30.Nathanson SD. Insights into the mechanisms of lymph node metastasis. CANCER-AM CANCER SOC. 2003;98(2):413–23. [DOI] [PubMed] [Google Scholar]
- 31.Bjerkeset T, Morild I, Mørk S, et al. Tumor characteristics in colorectal cancer and their relationship to treatment and prognosis. DIS COLON RECTUM. 1987;30(12):934–8. [DOI] [PubMed] [Google Scholar]
- 32.Chambers WM, Khan U, Gagliano A, et al. Tumour morphology as a predictor of outcome after local excision of rectal cancer. BRIT J SURG. 2004;91(4):457–9. [DOI] [PubMed] [Google Scholar]
- 33.Hermanek P. Evolution and pathology of rectal cancer. WORLD J SURG. 1982;6(5):502–9. [DOI] [PubMed] [Google Scholar]
- 34.Golia Pernicka JS, Bates DDB, Fuqua JL, et al. Meaningful words in rectal MRI synoptic reports: how polypoid May be prognostic. CLIN IMAG. 2021;80:371–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Gold P, Freedman SO. Specific carcinoembryonic antigens of the human digestive system. J EXP MED. 1965;122(3):467–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Locker GY, Hamilton S, Harris J, et al. ASCO 2006 update of recommendations for the use of tumor markers in Gastrointestinal cancer. J CLIN ONCOL. 2006;24(33):5313–27. [DOI] [PubMed] [Google Scholar]
- 37.Duffy MJ, Lamerz R, Haglund C, et al. Tumor markers in colorectal cancer, gastric cancer and Gastrointestinal stromal cancers: European group on tumor markers 2014 guidelines update. INT J CANCER. 2013;134(11):2513–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Stojkovic Lalosevic M, Stankovic S, Stojkovic M, et al. Can preoperative CEA and CA19-9 serum concentrations suggest metastatic disease in colorectal cancer patients? HELL J NUCL MED. 2017;20(1):41–5. [DOI] [PubMed] [Google Scholar]
- 39.Thirunavukarasu P, Talati C, Munjal S, et al. Effect of incorporation of pretreatment serum carcinoembryonic antigen levels into AJCC staging for colon cancer on 5-Year survival. JAMA SURG. 2015;150(8):747–55. [DOI] [PubMed] [Google Scholar]
- 40.Gao Y, Wang J, Zhou Y, et al. Evaluation of serum CEA, CA19-9, CA72-4, CA125 and ferritin as diagnostic markers and factors of clinical parameters for colorectal cancer. Sci Rep. 2018;8(1):2732. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Koprowski H, Steplewski Z, Mitchell K, et al. Colorectal carcinoma antigens detected by Hybridoma antibodies. Somatic Cell Genet. 1979;5(6):957–71. [DOI] [PubMed] [Google Scholar]
- 42.Stiksma J, Grootendorst DC, van der Linden PW. CA 19 – 9 as a marker in addition to CEA to monitor colorectal cancer. CLIN COLORECTAL CANC. 2014;13(4):239–44. [DOI] [PubMed] [Google Scholar]
- 43.Yu Z, Chen Z, Wu J et al. Prognostic value of pretreatment serum carbohydrate antigen 19 – 9 level in patients with colorectal cancer: A meta-analysis. PLoS ONE.2017;12(11). [DOI] [PMC free article] [PubMed]
- 44.Shin JK, Kim HC, Lee WY, et al. High preoperative serum CA 19 – 9 levels can predict poor oncologic outcomes in colorectal cancer patients on propensity score analysis. ANN SURG TREAT RES. 2018;96(3):107–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data generated during the current study are available from the corresponding author upon reasonable request. And all the data supporting the findings of this study are included in this published article and its supplementary information files.