Establishment and validation of a risk score model based on EUS: assessment of lymph node metastasis in early gastric cancer.

ABSTRACT

BACKGROUND AND AIMS: Lymph node metastasis significantly affects the prognosis of early gastric cancer patients. EUS plays a crucial role in the preoperative assessment of early gastric cancer. This study evaluated the efficacy of EUS in identifying lymph node metastasis in early gastric cancer patients and developed a risk score model to aid in choosing the best treatment options. METHODS: We retrospectively analyzed the effectiveness of EUS for detecting lymph node metastasis in early gastric cancer patients. A risk score model for predicting lymph node metastasis preoperatively was created using independent risk factors identified through binary logistic regression analysis and subsequently validated. Receiver operating characteristic curves were generated for both the development and validation cohorts. RESULTS: The overall accuracy of EUS in identifying lymph node metastasis was 85.3%, although its sensitivity (29.2%) and positive predictive value (38.7%) were relatively low. Patients were categorized based on preoperative risk factors for lymph node metastasis, including tumor size of ≥20 mm, lymph nodes of ≥10 mm, body mass index of ≥24 kg/m2, and lymph node metastasis on CT scans. A 7-point risk score model was developed to assess the likelihood of lymph node metastasis. The areas under the receiver operating characteristic curve for the development and validation sets were 0.842 and 0.837, respectively, with sensitivities of 64% and 79%, respectively. CONCLUSIONS: We developed a practical risk score model based on preoperative factors to help EUS predict lymph node metastasis in early gastric cancer patients, guiding the selection of optimal treatment approaches for these patients.