Nomograms to Predict Tumor Regression Grade (TRG) and YpTNM Staging in Patients with Locally Advanced Esophageal Cancer Receiving Neoadjuvant Therapy

Background: Neoadjuvant therapy (NT) has increased survival rates for patients with locally advanced esophageal cancer (EC), but estimating the impact of NT treatment prior to surgery is still very difficult.

Methods: A retrospective study of the clinical information of 150 patients with locally advanced EC who got NT at Qilu Hospital of Shandong University between June 2018 and June 2023. Patients were randomized into training and internal validation groups at a 3:1 ratio. Furthermore, an external validation cohort comprised 38 patients who underwent neoadjuvant therapy at Qianfoshan Hospital in the Shandong Province between June 2021 and June 2023. Independent risk factors were identified using univariate and multivariate logistic regression (forward stepwise regression). Predictive models and dynamic web nomograms were developed by integrating these risk factors.

Results: A total of 188 patients with locally advanced EC were enrolled, of whom 118 achieved stage I of neoadjuvant pathologic TNM (ypTNM) after receiving NT and 129 achieved grades 0-1 in the tumor regression grade (TRG). Logistic regression analysis identified five independent predictors of TRG grades 0-1: pulmonary function tests (PFT), prognostic nutritional index (PNI), triglyceride (TG) levels, squamous cell carcinoma antigen (SCC-Ag) levels, and combination immunotherapy. The areas under the receiver operating characteristic (ROC) curves for the training, internal validation, and external validation groups were 0.87, 0.75, and 0.80, respectively. Meanwhile, two independent predictors of stage I of ypTNM were identified: prealbumin (PA) and SCC antigen. The areas under the ROC curves for the training, internal validation, and external validation groups were 0.78, 0.67, and 0.70, respectively. The Hosmer-Lemeshow test for both predictive models showed excellent calibration, with well-fitted calibration curves. Decision curve analysis (DCA) and clinical impact curves (CIC) have demonstrated that nomograms are of clinical utility.

Conclusion: The nomograms performed well in predicting the likelihood of stage I of ypTNM and TRG grade 0-1 after NT in patients with locally advanced EC. It helps thoracic surgeons to predict the sensitivity of patients to NT before surgery, which enables precise treatment of patients with locally advanced EC.