Development of a Predictive Model for Large-Volume Pleural Effusion After Separation Surgery in Patients With Spinal Metastatic Tumors.

BACKGROUND: Separation surgery followed by radiotherapy has emerged as a prevalent approach for managing spinal metastatic tumors. However, large-volume postoperative pleural effusion (POPE) represents a challenging complication, as it potentially delays subsequent treatments and increases morbidity. This study aims to identify risk factors for large-volume POPE and develop a predictive model for early identification to improve patient prognosis. METHODS: This retrospective study analyzed 443 patients who underwent separation surgery for spinal metastases at our center between January 2014 and January 2022. High-resolution CT-based 3D modeling was utilized for accurate pleural effusion (PE) volume quantification. Variables including patient demographics, surgical details, and laboratory results were examined to identify risk factors associated with large-volume POPE (≥ 1000 mL). A predictive nomogram was developed based on the multivariate logistic regression analysis. RESULTS: Our findings indicated that advanced age, increased intraoperative blood loss, and decreased levels of preoperative serum albumin, postoperative serum protein, and hemoglobin were significant independent risk factors for large-volume POPE. The predictive nomogram demonstrated high accuracy, with a mean AUC value of 0.953 for the training dataset and 0.927 for the testing dataset, indicating reliable predictability for identifying patients at high risk for large-volume POPE. CONCLUSION: Our study identified independent risk factors for large-volume PE following separation surgery in patients with spinal metastasis. The developed nomogram offers a practical tool for early identification of high-risk groups, enabling timely and targeted interventions. By reducing the risk of large POPE, this approach may shorten hospitalization and accelerate the resumption of postoperative treatment, ultimately improving patient prognosis.