Computed tomography with carcinoembryonic antigen and carbohydrate antigen 19-9 in diagnosing lymph node metastasis of early gastric cancer

Abstract

BACKGROUND

Gastric cancer (GC) is the fifth most prevalent and fourth most lethal malignancy globally. China bears a disproportionately high burden, accounting for 44.0% of new cases and 48.6% of deaths worldwide. In early GC (EGC), the presence of lymph node metastasis (LNM) is a critical prognostic determinant that directly guides therapeutic strategy. While multi-detector computed tomography (CT) and serum biomarkers carcinoembryonic antigen (CEA)/carbohydrate antigen 19-9 (CA19-9) are established diagnostic tools, each demonstrates limited efficacy when used independently. This study therefore aims to verify whether a combined diagnostic approach integrating multi-detector CT (MDCT) with serum CEA/CA19-9 can significantly improve the accuracy of LNM detection in EGC patients.

AIM

To investigate the diagnostic value of CT combined with CEA or CA19-9 for detecting LNM in EGC.

METHODS

This retrospective study included 120 patients with EGC confirmed by gastroscopic biopsy at our institution (Huai’an Hospital of Huai’an City) between February 2024 and August 2024. Based on postoperative pathological findings, participants were categorized into a LNM group (n = 60) and a non-metastasis group (n = 60). All patients underwent MDCT scanning and serum CEA and CA19-9 level measurements. The diagnostic efficacy of CT, CEA, and CA19-9 alone and in combination was evaluated using receiver operating characteristic (ROC) curve and Kappa consistency analysis.

RESULTS

Serum analysis showed significantly elevated CEA and CA19-9 levels and higher positivity rates in the metastasis group (P < 0.0001). ROC analysis yielded area under the curves of 0.9443 (CEA) and 0.9292 (CA19-9), with Kappa values of 0.683 and 0.650, respectively. CT revealed significantly greater short-axis diameter, CT attenuation, blood volume, and permeability in metastatic nodes (P < 0.05), whereas blood flow and mean transit time showed no significant differences. CT alone demonstrated 85.00% sensitivity and 95.00% specificity (Kappa = 0.800). Combined diagnosis improved sensitivity to 91.67% (CT + CEA) and 90.00% (CT + CA19-9), with specificities of 90.00% and 88.33%, respectively.

CONCLUSION

The combination of CT with CEA or CA19-9 improves sensitivity for detecting LNM in EGC, supporting personalized treatment planning and demonstrating clinical value.

Key Words: Computed tomography; Carcinoembryonic antigen; Carbohydrate antigen 19-9; Gastric cancer; Lymph node metastasis

Core Tip: This study identified that elevated computed tomography (CT) values and increased short-axis diameter of lymph nodes, along with raised serum carcinoembryonic antigen and carbohydrate antigen 19-9 levels, are key indicators for lymph node metastasis in early gastric cancer. Combining anatomical imaging with tumor biomarker analysis improved diagnostic sensitivity and reduced the missed diagnosis rate by 6.67%-8.33% compared to using CT alone. These findings highlight the value of integrating morphological and biological data for more accurate metastasis detection, filling a gap in current diagnostic approaches and highlights the need for multimodal strategies in optimizing individualized treatment decisions.

INTRODUCTION

Gastric cancer (GC), as a group of primary malignant tumors developing within the gastric mucosal epithelium, occupies a prominent position in the global cancer burden. According to the GLOBOCAN 2020 annual report, GC is the fifth most prevalent and the fourth most lethal cancer worldwide. China accounts for a large proportion of the distribution of this disease, accounting for 44.0% of new cases and 48.6% of deaths worldwide[1]. Early GC (EGC) is considered as a malignant tumor of the stomach in which the lesion has not breached the mucosal layer (stage T1a) or has invaded only the submucosal layer (stage T1b). The prognosis of this disease is significantly correlated with a variety of pathologic features, including regional lymph node metastasis (LNM), depth of infiltration, gross staging, and histologic differentiation grade[2,3]. Studies have shown that the risk of LNM in patients with EGC ranges from 5.7% to 19.1%[4]. Multifactorial analysis by Wang et al[5] confirmed that regional lymph node involvement serves as an independent prognostic indicator for EGC patients. In clinical diagnosis, carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9) have been widely used in assisting the diagnosis and prognosis of cancer due to their extremely low levels in normal tissues and abnormal elevation in the process of occurrence and development of many cancers. CEA is a glycoprotein initially detected in colorectal cancer, and later shown to be highly expressed in gastric tumor, lung tumor and other tumors, while CA19-9 is a glycoconjugate antigen closely associated with pancreatic carcinoma, cholangiocarcinoma and other digestive tumors. The detection of these two markers can not only help cancer screening and early diagnosis, but also monitor the changes of the disease during the treatment process, providing an important basis for clinical decision-making[6-8]. Meanwhile, multi-detector computed tomography (CT) is a state-of-the-art imaging technique that has extensively utilized for clinical diagnosis due to its rapid, non-invasive and high accuracy. The core advantage of multi-detector CT (MDCT) is its sub-millimeter thin-layer scanning and multi-phase dynamic enhancement technologies, which enable MDCT to provide high-resolution images to effectively evaluate the depth of gastric wall involvement and LNM. With these advantages, MDCT has become a routine clinical examination for cancer, especially being a key factor in the diagnosis and staging of GC and other digestive tumors[9]. In view of the importance of accurate preoperative determination of LNM in the therapeutic decision-making in EGC and the potential value of CEA, CA19-9 and MDCT in the evaluation, this study aimed to investigate the clinical efficacy of the combined application of MDCT and serum CEA/CA19-9 assay in the diagnosis of LNM in EGC. By integrating imaging and serological tests, it is expected to provide more powerful support for accurate diagnosis and individualized treatment of EGC patients, and improve their prognosis.

MATERIALS AND METHODS

Baseline data

We selected 120 patients diagnosed with EGC who visited the Huai’an Hospital of Huai’an City for treatment and follow-up between February 2024 and August 2024, comprising 63 male and 57 female participants, with an average age of 52.06 ± 12.81 years. The lesions were concentrated in the three regions of gastric antrum, gastric body and cardia. Based on LNM status in the pathological examination, was stratified into metastasis group (n = 60) and non-metastasis group (n = 60) groups. The differences between baseline characteristics of the two groups patients were not statistically significant and had good comparability (P > 0.05; Table 1). Tumor marker levels, CT parameters, and the diagnostic efficacy of CT, CEA, and CA19-9 alone and in combination for the diagnosis of LNM in EGC were compared between the two groups. All patients included in this study were fully informed about the purpose, methods, potential risks and benefits of the study and voluntarily signed an informed consent form. The study protocol has passed the approval by the Medical Ethics Committee of the First Affiliated Hospital of Bengbu Medical University and was formally approved, Approval No.[2024]KY033.

Table 1 Baseline characteristics of the two groups, mean ± SD.

Groups	n	Gender		Age (years)	Site of tumor
		Male	Female		Gastric sinus	Gastric body	Cardia
Non-metastasis group	60	31	29	52.87 ± 12.55	37	13	10
Metastasis group	60	32	28	51.25 ± 13.13	40	14	6
χ2/t		0.33		0.6895
P value		0.855		0.4919

Selection and exclusion standards

Inclusion criteria: (1) EGC diagnosed by gastroscopic biopsy and subsequent pathological examination; (2) All patients were found for the first time and had not been treated with radiotherapy or chemotherapy; (3) Patients’ age ranged from 33 years to 76 years old; (4) Patients did not have serious dysfunctions of heart, lungs, liver, kidneys and other major organs; (5) Medical records were complete and traceable; and (6) Patients were aware of all the details of the study and signed informed consent.

Exclusion criteria: (1) Patients with a history of other primary malignant tumors; (2) Patients with mental disorders that prevented them from cooperating with the relevant examinations; (3) Patients with active autoimmune diseases, active infections, or serious hematological diseases; (4) Patients with incomplete data or missing key information; and (5) Patients who did not cooperate with the diagnosis.

Examination methods

Tumor marker detection: All subjects underwent standardized serum tumor marker testing: All subjects had 5 mL of elbow vein blood collected by a professional nurse in the morning fasting state, and preserved in a standardized manner using certified non-pyrogenic vacuum blood collection tubes. The samples were placed in a 4 °C constant temperature transfer box immediately after collection to ensure that the pretreatment was completed within 2 hours. The serum was centrifuged at 3000 rpm (relative centrifugal force 2000 × g) for 15 minutes in a biosafety cabinet using a pre-cooled centrifuge to achieve efficient separation of the serum. After separation, the supernatant was divided into two 1.8 mL sterile cryopreservation tubes, labeled with a unique identification code and stored in a gradient cryopreservation: Firstly, it was flash frozen at -80 °C for 30 minutes, and then transferred to -20 °C for long-term storage. The assay was performed by electrochemiluminescence immunoassay (Roche Cobas e601 system), and the standardized procedure was strictly followed, in which the double-antibody sandwich assay was used for CEA (positive threshold > 5.0 ng/mL), and the competitive binding assay was used for CA19-9 (positive threshold > 35 U/mL)[10]. Strict quality control is implemented throughout the entire testing process, including daily instrument calibration, three levels of quality control product monitoring, and operator competency certification.

Imaging examination: A standardized MDCT examination protocol was used, and all subjects underwent a uniform and standardized examination procedure. Patients were required to strictly fast for more than 6 hours before the examination in order to minimize the interference of gastrointestinal contents and the risk of aspiration. Thirty minutes before the formal scanning, patients were required to take 1000 mL of warm boiled water at 37 °C to 40 °C orally in divided doses, so that the gastric lumen and the proximal small intestine would be evenly filled to form a good natural contrast. Patients without contraindications (e.g., glaucoma, prostatic hypertrophy, tachycardia) were injected intramuscularly with 20 mg of scopolamine hydrochloride 15 minutes prior to scanning to effectively inhibit gastrointestinal peristalsis and avoid motion artifacts. During the examination, the midline of patients’ body was strictly aligned with the centerline of the scanning bed, and the scanning range was covering the volume from the diaphragmatic apex to the iliac crest inferior margin. The scanning procedure consisted of a full abdominal plain scan and three-phase enhancement scan: Firstly, a plain scan was performed, followed by a high-pressure regimen injection of non-ionic iodine contrast agent iohexol (350 mg I/mL) via an 18-gauge indwelling needle in the anterior elbow vein at a dosage of 1.5-2.0 mL/kg, with the injection rate of 3.0-4.0 mL/second, and then a follow-up injection of 30 mL of physiological saline to rinse the line at the same flow rate immediately after the injection was completed. After the injection, 30 mL of saline was delivered immediately to flush the line. The enhancement scans were performed at three time points, arterial phase (35 seconds), venous phase (65 seconds), and delayed phase (180 seconds) after contrast injection.

After the completion of scanning, all image data were automatically transferred to the professional workstation accompanying the CT equipment, and systematic post-processing was performed by a dedicated computer processing system. The post-processing process includes image noise reduction, layer thickness optimization, artifact correction and other pre-processing steps to ensure that the image quality based on the subsequent analysis is optimal. Quantitative analysis uses professional CT perfusion analysis software to calculate and generate a number of important parameters including blood volume (BV), permeability (PS), blood flow (BF), and mean transit time (MTT), providing an objective quantitative basis for clinical assessment. To ensure the objectivity and accuracy of the diagnostic images, all images were independently analyzed and evaluated by two senior radiologists with more than 10 years of experience in abdominal imaging. These physicians have undergone rigorous training and are well versed in the imaging features of various abdominal diseases. To eliminate potential bias, the physicians were blinded to key information. If the initial diagnostic opinions of the two physicians differ, a final consensus will be reached by jointly reviewing the films and discussing the points of difference one by one, thus minimizing the influence of subjective factors on the diagnostic results and ensuring the reliability and consistency of the study data. The whole examination process strictly follows the international radiology operation standard to ensure that high quality imaging data are obtained for subsequent analysis and research.

Observation indicators: (1) Compare the levels of tumor markers and their positivity rates between the two groups: Established cutoff values for tumor markers were defined as follows: CEA concentrations ≤ 5.0 ng/mL and CA19-9 levels ≤ 35 U/mL represented the reference thresholds employed in this analysis. Those with values higher than the corresponding thresholds were categorized into the positive group, and those with values lower than or equal to the thresholds were categorized into the negative group[10]. The specific concentration values for each sample were recorded in detail and a statistical difference analysis between groups was implemented; (2) Evaluate the results of LNM on CT: LNM was defined as present if one of the following conditions is met: (a) CT images show that maximum short-axis diameter of lymph node ≥ 6 mm , and the CT value of the scan ≥ 25 Hu; and (b) The lymph nodes appear to be enlarged, fused together or enlarged with abnormal enhancement; and (3) Systematic comparison of the diagnostic efficacy of CT, CEA, CA19-9 alone and in combination: The diagnostic value of CT, CEA single test, CA19-9 single test, and the combination of these tests for the diagnosis of regional lymphatic metastasis of GC was systematically evaluated using postoperative pathologic-histologic diagnosis as the final criterion (gold standard). The efficacy indexes of each testing program were precisely calculated including: (1) Sensitivity: (Number of identified positive cases/number of positive cases diagnosed by the gold standard) × 100%; (2) Specificity: (Number of identified negative cases/number of negative cases diagnosed by the gold standard) × 100%; (3) False positive rate: 100% - specificity; and (4) False negative rate: 100% - sensitivity. For combined tests where CT is combined with CEA and/or CA19-9, the combination was considered positive if any of the CT, CEA, or CA19-9 indicators are positive; the combination is considered negative only if all of the indicators are negative.

Statistical analysis

Statistical analyses were conducted using SPSS 25.0 software. Quantitative variables concentrations (CEA, CA19-9) underwent normalization before being reported as mean ± SD. Group comparisons for these continuous parameters employed independent samples t-tests. Categorical data were expressed as n (%), analyzed via χ² tests. Diagnostic performance of serum biomarkers for detecting LNM in early gastric carcinoma was evaluated through receiver operating characteristic (ROC) curve methodology. Meanwhile, the consistency of CT, CEA, CA19-9 individual and combined test results with pathological diagnosis (gold standard) was further analyzed using the Kappa consistency test. Statistical significance was defined at P < 0.05 throughout all analyses.

RESULTS

Comparison of serum tumor marker levels of CEA and CA19-9 between the two groups

As shown in Tables 2 and 3, in terms of serum CEA level, the detection value of patients in the metastatic group was 7.51 ± 2.49 ng/mL, and 3.27 ± 1.18 ng/mL in the non-metastatic group; in the comparison of CA19-9 level, it was 51.29 ± 17.38 U/mL in the metastatic group, and 23.61 ± 8.33 U/mL in the non-metastatic group, and the levels of the two tumor markers in the metastatic group were all The difference was statistically significant (P < 0.0001); in terms of positivity rate, the positive rate of CEA in the metastatic group was 76.67%, while that of the non-metastatic group was only 8.33%; the positive rate of CA19-9 was 75.00%, while that of the non-metastatic group was 10.00%; the positive rates of the two indexes were higher than that of the non-metastatic group, which further confirmed the correlation between the tumor marker levels and the metastasis of lymph nodes (P < 0.0001). ROC curve analysis showed that the area under the curve (AUC) of CEA for diagnosing LNM in EGC was 0.9443, and the AUC of CA19-9 was 0.9292 (Figure 1), indicating that both CEA and CA19-9 have high diagnostic efficacy for LNM in EGC. In addition, the consistency of the two indexes with pathological diagnosis was analyzed by Kappa consistency test (Table 4), and the Kappa values of CEA and CA19-9 were 0.683 and 0.650, respectively, both of which were in the range of 0.6 < Kappa value ≤ 0.8, which indicated that the two test results were in high consistency with pathological diagnosis (the gold standard), and all the diagnostic efficacy indexes showed a good clinical application value.

Figure 1 Receiver operating characteristic curves of carcinoembryonic antigen and carbohydrate antigen 19-9 predicting the occurrence of lymph node metastasis in early gastric cancer. CA19-9: Carbohydrate antigen 19-9; CEA: Carcinoembryonic antigen; AUC: Area under the curve.

Table 2 Comparison of carcinoembryonic antigen and carbohydrate antigen 19-9 levels between the two groups, mean ± SD.

Groups	n	CEA (ng/mL)	CA19-9 (U/mL)
Non-metastasis group	60	3.27 ± 1.18	23.61 ± 8.33
Metastasis group	60	7.51 ± 2.49	51.29 ± 17.38
t value		11.94	11.13
P value		< 0.0001	< 0.0001

CA19-9: Carbohydrate antigen 19-9; CEA: Carcinoembryonic antigen.

Table 3 Positive rates of carcinoembryonic antigen and carbohydrate antigen 19-9 detection in the two groups, n (%).

Tumor markers	Non-metastasis group (n = 60)	Metastasis group (n = 60)	χ2	P value
CEA	5 (8.33)	46 (76.67)	57.323	< 0.0001
CA19-9	6 (10.00)	45 (75.00)	51.867	< 0.0001

CA19-9: Carbohydrate antigen 19-9; CEA: Carcinoembryonic antigen.

Table 4 Comparison of diagnostic results of carcinoembryonic antigen and carbohydrate antigen 19-9 with pathological diagnosis results.

		Pathological diagnosis results		Total	Kappa value	P value
		Positive	Negative
CEA	Positive	46 (TP)	5 (FP)	51	0.683	< 0.0001
	Negative	14 (FN)	55 (TN)	69
Total		60	60	120
CA19-9	Positive	45 (TP)	6 (FP)	51	0.650	< 0.0001
	Negative	15 (FN)	54 (TN)	69
Total		60	60	120

CA19-9: Carbohydrate antigen 19-9; CEA: Carcinoembryonic antigen; TP: True positive; FP: False positive; FN: False negative; TN: True negative.

Analysis of patients' CT diagnostic results

According to Table 5, in the comparison of CT parameters between the two groups, the BV was 7.87 ± 2.33 mL/100 mg in the metastasis group and 6.81 ± 2.35 mL/100 mg in the non-metastasis group, and in terms of the vascular PS, the metastasis group had a PS of 34.37 ± 10.65 mL/100 mg/minute, while the non-metastasis group had 23.49 ± 11.24 mL/100 mg/minute. Both parameters were higher than those of the non-metastasized group (P < 0.05), while BF and MTT did not show statistical differences between the two groups (P > 0.05). In terms of morphological features, the maximum short diameter of lymph nodes in the metastasis group was 7.82 ± 1.06 mm, and 5.09 ± 0.41 mm in the non-metastasis group, and the CT value was 32.5 ± 6.89 HU in the metastasis group, and 18.33 ± 4.23 HU in the non-metastasis group, and the maximum short diameter and CT value of the metastasis group were larger than those of the non-metastasis group (P < 0.0001), which suggests that after the occurrence of metastasis of lymph nodes, the patients, the volume of GC increased significantly, and at the same time, the density of its internal structure also increased significantly. As shown in Table 6, by comparing the results of CT diagnosis with the gold standard of pathological diagnosis, 51 cases of CT diagnosis were true positive, 3 cases of false positive, 9 cases of false negative, and 57 cases of true negative in the 120 study samples, and the value of Kappa’s consistency test was 0.800 (P < 0.0001), which indicated that there was a high degree of consistency between the CT diagnosis and the gold standard of pathological diagnosis, and that there was a high degree of consistency in the Kappa’s consistency test between CT diagnosis and the gold standard of pathological diagnosis. The Kappa consistency test value was 0.800 (P < 0.0001), indicating that the CT diagnostic results had a high degree of consistency with the pathologic gold standard. The above results confirmed that CT examination has reliable diagnostic performance in the diagnosis of LNM of EGC.

Table 5 Comparison of computed tomography scan parameters between the two groups, mean ± SD.

Groups	n	Maximum short diameter	CT value	BF (mL/100 mg/minute)	BV (mL/100 mg)	MTT (second)	PS (mL/100 mg/minute)
Non-metastasis group	60	5.09 ± 0.41	18.33 ± 4.23	63.98 ± 12.44	6.81 ± 2.35	12.65 ± 5.43	23.49 ± 11.24
Metastasis group	60	7.82 ± 1.06	32.51 ± 6.89	60.44 ± 9.95	7.87 ± 2.33	11.58 ± 5.07	34.37 ± 10.65
t value		18.57	13.60	1.721	2.479	1.117	5.442
P value		< 0.0001	< 0.0001	0.0879	0.0146	0.2663	< 0.0001

CT: Computed tomography; BF: Blood flow; BV: Blood volume; MTT: Mean transit time; PS: Permeability.

Table 6 Patients’ computed tomography diagnostic findings against pathologic findings.

CT diagnostic findings	Pathologic findings		Total
	Positive	Negative
Positive	51 (TP)	3 (FP)	54
Negative	9 (FN)	57 (TN)	66
Total	60	60	120
Kappa value	0.800
P value	< 0.0001

CT: Computed tomography; TP: True positive; FP: False positive; FN: False negative; TN: True negative.

Comparison of diagnostic results of CT, CEA/CA19-9 alone and in combination on LNM of EGC

As shown in Table 7, when diagnosed alone, CEA had a sensitivity of 76.67% (46/60), a specificity of 91.67% (55/60), a false negative rate of 23.33% (14/60), and a false positive rate of 8.33% (5/60); CA19-9 had a sensitivity of 75.00% (45/60), a specificity of 90.00% (54/60), a 25.00% (15/60), and 10.00% (6/60) for misdiagnosis; CT had a sensitivity of 85.00% (51/60), a specificity of 95.00% (57/60), a miss rate of 15.00% (9/60), and a false positive rate of 5.00% (3/60). For individual diagnostic methods, the sensitivity and specificity of CT diagnosis were relatively high compared with CEA and CA19-9, and the false negative rate and false positive rate were relatively low. When combined testing was used, the diagnostic efficacy of each diagnostic method was significantly improved. The sensitivity of CT combined with CEA reached 91.67% (55/60), the specificity was 90.00% (54/60), the false negative rate was reduced to 8.33% (5/60), and the false positive rate was 10.00% (6/60); the sensitivity of CT combined with CA19-9 was 90.00% (54/60), specificity was 88.33% (53/60), false negative rate was 10.00% (6/60), and false positive rate was 11.67% (7/60). Although the specificity of the combined diagnosis of LNM in EGC was slightly lower and the risk of misdiagnosis was increased compared with that of the separate diagnosis, the risk of leakage of the single method was significantly reduced. Specifically, among the 9 patients missed by CT alone, 7 were identified as positive by either CEA or CA19-9. Similarly, among the 14 patients missed by CEA, 9 showed abnormalities on CT; likewise, among the 15 missed by CA19-9, 9 were detected by CT. The results fully confirmed the synergistic effect of anatomical morphology examination (CT) and tumor biological indexes (CEA/CA19-9). CT can clearly show the morphological characteristics of lymph nodes, while CEA and CA19-9 reflect the biological behavior of tumors. The combination of the two can make the overall diagnostic sensitivity break through 90%, which greatly improved the detection rate of LNM in EGC, and provided a good opportunity for the precise screening of LNM of EGC in the clinic. It provides a reliable and efficient solution for the accurate screening of EGC LNM.

Table 7 Comparison of the efficacy of different methods in diagnosing lymph node metastasis of early gastric cancer, n (%).

Diagnostic methods, n = 60	Sensitivity	Specificity	False negative rate	False positive rate
CEA	46 (76.67)	55 (91.67)	14 (23.33)	5 (8.33)
CA19-9	45 (75.00)	54 (90.00)	15 (25.00)	6 (10.00)
CT	51 (85.00)	57 (95.00)	9 (15.00)	3 (5.00)
CT combined CEA	55 (91.67)	54 (90.00)	5 (8.33)	6 (10.00)
CT combined CA19-9	54 (90.00)	53 (88.33)	6 (10.00)	7 (11.67)

CT: Computed tomography; CA19-9: Carbohydrate antigen 19-9; CEA: Carcinoembryonic antigen.

DISCUSSION

GC constitutes a major malignancy posing a severe threat to human health, and its treatment strategy and prognosis are highly dependent on the tumor-node-metastasis staging classification[11]. LNM represents a critical prognostic determinant for individuals diagnosed with EGC[12], moreover, performing sufficient lymphadenectomy enables these patients to achieve a high probability of cure[13]. However, subjecting EGC patients confirmed to lack LNM to unnecessary gastrectomy combined with D1 or D2 lymph node dissection might not only lead to over-treatment of non-metastatic lymph nodes, which may damage the immune barrier of the lymph node system, reduce the patient’s immunity to the tumor, and affect the efficacy of the surgery, but also increase the mortality rate and the risk of surgery-related complications[14,15]. For EGC cases without LNM, endoscopic submucosal dissection (ESD) offers a viable treatment approach. This less invasive technique provides precise outcomes with minimal tissue damage. For EGC patients with minimal risk of LNM, endoscopic treatment is the appropriate choice, including both endoscopic mucosal resection and ESD. If the patient has LNM, a comprehensive treatment plan should be adopted with close follow-up; among them, individuals diagnosed with EGC exhibiting a significant risk of LNM require radical gastrectomy plus concurrent D2 lymphadenectomy to effectively minimize recurrence potential. In addition, the lymph node status of EGC patients receiving non-surgical treatment also has an important impact on the development of chemotherapy and radiotherapy regimens and the assessment of efficacy[5,16,17]. Therefore, precise preoperative identification of LNM status and its distribution range in GC patients is essential for the development of individualized and optimized treatment plans.

However, precisely establishing the nodal metastasis status in patients suffering from early-stage GC is still a difficult problem in clinical diagnosis and treatment. In this field, imaging technology plays an irreplaceable role, in which CT is the main means of preoperative assessment of regional lymph nodes, and its diagnostic basis focuses on the morphological parameters of lymph nodes (including the maximum diameter, geometric contour) and contrast-enhanced imaging features[18]. CT has been widely used and demonstrated its value in LNM discrimination in a variety of malignant tumors. For example, CT with magnetic resonance imaging incorporating artificial intelligence algorithms has shown good diagnostic efficacy in predicting lymph node involvement in patients with breast cancer, which is potentially promising for clinical applications[19]. Similarly, dual-energy CT has demonstrated utility in diagnosing LNM among esophageal cancer patients[13]. For some EGC cases that are negative for LNM by preoperative imaging, CT is valuable for detecting occult LNM, which is directly related to clinical decision-making and prognostic evaluation of patients[20]. However, at this stage, there is still a lack of uniform consensus on the imaging determination criteria. Clinical judgment is routinely based on the short diameter of the lymph nodes or the ratio of the short diameter to the long diameter, but the accuracy of these morphological criteria is not satisfactory, and their diagnostic efficacy needs to be further improved[20]. Therefore, a synergistic diagnostic scheme that integrates CT with other assessment tools (such as endoscopic ultrasound, molecular imaging techniques, or serum marker testing) is expected to break through the bottleneck of a single technique, and may significantly improve the identification of LNM in EGC, which may help to optimize the selection of individualized treatment strategies. Relevant studies have demonstrated the superiority of this strategy, and Song and other scholars have pointed out[21] that multilayer spiral CT (MSCT) has certain limitations in identifying soft tissue structures, especially for lymph nodes with a diameter of less than 0.3 cm. Its detection efficacy is insufficient, which may lead to underdiagnosis and is not conducive to accurately assessing the metastatic status of lymph nodes. In contrast, the combined application of ultrasonography and MSCT is superior to either method alone for the detection of LNM in gastric carcinoma. The advantage of this combined strategy is that it can complement each other’s technical characteristics, thus effectively improving the accuracy of preoperative judgment of LNM status in GC patients.

This study revealed elevated CT values in metastatic lymph nodes compared to non-metastatic counterparts, likely reflecting histopathological features of metastatic deposits - specifically, heightened cellular density due to tumor infiltration and active neovascularization within involved nodes. Standalone CT imaging demonstrated 85.00% sensitivity and 95.00% specificity for detecting LNM, suggesting that there is still room for optimization.

The limitations of CT scanning in the detection of LNM may lie in the following aspects. Firstly, microscopically detectable metastases in GC frequently involve small lymph nodes, posing a significant challenge for CT imaging; this inherent limitation often results in diagnostic uncertainty, comprising both false-negative and false-positive assessments[18]. Secondly, the interpretation of CT images is highly dependent on the radiologist’s professional experience, and there may be significant differences in interpretation between different observers. In addition, benign lesions such as inflammatory hyperplasia and fibrosis of lymph nodes may also lead to elevated CT values, which may overlap with metastatic foci.

Tumor markers are of great value in the screening, clinical staging and prognostic system of malignant tumors. One of the distinguishing features of the cancerous process is aberrant glycosylation modification, and this post-translational modification abnormality is closely related to tumorigenesis and development. Representative glycoprotein markers such as HER2, AFP and CEA have been established as key diagnostic biomarkers[22]. Among them, CEA and CA19-9, as routine clinical testing indicators, have been confirmed by studies to be important in the development and progression of cancer as an aid to diagnosis[23]. For example, the combination of serum CEA, glycan antigen 72-4 and CA19-9 can significantly improve the sensitivity and discriminative accuracy of GC screening[24]. CEA and CA19-9 are independent indicator of advanced pancreatic cancer[25]. CEA and CA19-9 demonstrate significant diagnostic utility when integrated with glycan antigen 125. Specifically, elevated preoperative serum concentrations of these three tumor markers collectively enable effective discrimination among benign, borderline, and malignant pathological subtypes within mucinous ovarian neoplasms[26]. Analysis revealed elevated circulating concentrations of CEA and CA19-9 in the serum of GC patients with confirmed LNM relative to non-metastatic cases, and the positivity rate was also higher, which suggests that the two markers have a potential application for the assessment of lymphatic metastasis status in EGC. However, the diagnostic efficacy of a single tumor marker is still limited: The data in this study showed that the sensitivity and specificity of CEA alone were 76.67% and 91.67%, while those of CA19-9 were 75.00% and 90.00%, respectively. Although positive results can indicate metastatic risk, single-indicator testing still faces a false negative rate of about 25% and a false-positive risk of nearly 10%. Therefore, tumor marker testing combined with other methods may be a more effective diagnostic strategy.

In this study, the clinical value of the synergistic diagnostic modality of CT and serum tumor markers in the discrimination of LNM in EGC was evaluated by combining CT with CEA or CA19-9. The results showed that the diagnosis of CT combined with CEA increased the sensitivity to 91.67%, and the diagnosis of CT combined with CA19-9 increased the sensitivity to 90.0%; the specificity of the combined test, although slightly decreased compared with that of the test alone, was still maintained at a high level (90.00% and 88.33%). This result indicates that the use of CT combined with CEA/CA19 diagnostic strategy can effectively compensate for the shortcomings of a single method, which can improve the diagnostic sensitivity. This enhancement stems from the fact that CT and serum tumor markers provide information from two different dimensions, anatomical morphology and tumor biology, respectively, and the combination of the two can cover a wider range of potential metastatic cases. This combined diagnostic paradigm is essential to reduce underdiagnosis and ensure more aggressive therapeutic evaluation of patients with potential metastases, especially in micrometastases that are morphologically unremarkable on CT but biologically active, where elevated tumor markers can provide important early warning signals.

In summary, serum CEA and CA19-9 concentrations can be effective predictors of LNM in EGC. Although multislice spiral CT is the current core imaging tool for assessing lymph node status, its application alone still carries a false-negative risk of approximately 15%. A multimodal diagnostic strategy integrating MDCT with tumor markers significantly improves the sensitivity of the test. This combined advantage stems from the complementary nature of anatomical imaging and tumor biomarkers, which provides a more reliable evidence-based basis for preoperative precision staging and individualized treatment decision-making, and has clear clinical translational value. However, the sample size of this study is relatively limited. In the future, with the expansion of sample size and multicenter studies, the diagnostic efficacy of CT combined with CEA/CA19-9 is expected to be further validated and optimized, and combined with other advanced imaging techniques and molecular diagnostic tools, it is expected to further improve the diagnostic accuracy of LNM in EGC, and to provide stronger support for accurate treatment and prognosis evaluation of patients.

CONCLUSION

Combining CT with serum CEA or CA19-9 significantly improves the diagnostic sensitivity for LNM in EGC, effectively reducing the missed diagnosis risk of single imaging or serological tests. This combined method integrates anatomical morphological information (from CT) and tumor biological information (from CEA/CA19-9), covers more potential metastatic cases, and provides a reliable basis for formulating individualized treatment plans for EGC patients. It has important clinical application value and lays a foundation for optimizing EGC treatment strategies and improving patient prognosis.